Traffic Related Air Pollution and Internal Combustion Engines [1 ed.] 9781612098869, 9781607411451

201 27 20MB

English Pages 438 Year 2009

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Traffic Related Air Pollution and Internal Combustion Engines [1 ed.]
 9781612098869, 9781607411451

Citation preview

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

AIR, WATER AND SOIL SCIENCE AND TECHNOLOGY SERIES

TRAFFIC RELATED AIR POLLUTION AND INTERNAL COMBUSTION ENGINES

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in rendering legal, medical or any other professional services.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

AIR, WATER AND SOIL SCIENCE AND TECHNOLOGY SERIES Trends in Air Pollution Research James, V. Livingston (Editor) 2005. ISBN: 1-59454-326-7 Agriculture and Soil Pollution: New Research James, V. Livingston (Editor) 2005. ISBN: 1-59454-310-0 Water Pollution: New Research A.R. Burk (Editor) 2008. ISBN: 1-59454-393-3 Air Pollution: New Research James, V. Livingston (Editor) 2007. ISBN: 1-59454-569-3

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Air Pollution Research Advances Corin G. Bodine (Editor) 2007. ISBN: 1-60021-806-7 Marine Pollution: New Research Tobias N. Hofer (Editor) 2008. ISBN: 978-1-60456-242-2 Complementary Approaches for Using Ecotoxicity Data in Soil Pollution Evaluation M. D. Fernandez and J. V. Tarazona 2008. ISBN: 978-1-60692-105-0 Lake Pollution Research Progress Franko R. Miranda and Luc M. Bernard 2008. ISBN: 978-1-60692-106-7

Lake Pollution Research Progress (Online Book) Franko R. Miranda and Luc M. Bernard (Editors) 2008. ISBN: 978-1-60741-905-1 River Pollution Research Progress Mattia N. Gallo and Marco H. Ferrari (Editors) 2009. ISBN: 978-1-60456-643-7 Heavy Metal Pollution Samuel E. Brown and William C. Welton (Editors) 2008. ISBN: 978-1-60456-899-8 Cruise Ship Pollution Oliver G. Krenshaw (Editor) 2009. ISBN: 978-1-60692-655-0 Industrial Pollution including Oil Spills Harry Newbury and William De Lorne (Editors) 2009. ISBN: 978-1-60456-917-9 Environmental and Regional Air Pollution Dean Gallo and Richard Mancini 2009. ISBN: 978-1-60692-893-6 Traffic Related Air Pollution and Internal Combustion Engines Sergey Demidov and Jacques Bonnet (Editors) 2009. ISBN: 978-1-60741-145-1

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

AIR, WATER AND SOIL SCIENCE AND TECHNOLOGY SERIES

TRAFFIC RELATED AIR POLLUTION AND INTERNAL COMBUSTION ENGINES

SERGEY DEMIDOV AND

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

JACQUES BONNET EDITORS

Nova Science Publishers, Inc. New York

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009 by Nova Science Publishers, Inc. All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers’ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA Traffic related air pollution and internal combustion engines / [compiled by] Sergey Demidov and Jacques Bonnet. p. cm. Includes bibliographical references and index. ISBN 978-1-61209-886-9 (eBook) 1. Smog--Prevention. 2. Air--Pollution. 3. Automobiles--Environmental aspects. 4. Motor fuels-Environmental aspects. 5. Alternative fuel vehicles. I. Demidov, Sergey, 1958- II. Bonnet, Jacques, 1959TD884.3.T73 2009 628.5'32--dc22 2009008043

Published by Nova Science Publishers, Inc. 

New York

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

CONTENTS

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Preface

vii

Chapter 1

Traffic- RelatedAir Pollution T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez, N. López-Valdes, M. Ustarroz-Cano and L.F. Montaño

Chapter 2

Road Traffic Emission and Fuel Consumption Modelling: Trends, New Developments and Future Challenges Robin Smit, Hussein Dia and Lidia Morawska

29

Chapter 3

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application to Transport Modelling and Health Impact Assessments of Urban Fleets Diane U. Keogh, Joe Kelly, Kerrie Mengersen, Rohan Jayaratne, Luis Ferreira and Lidia Morawska

69

Chapter 4

Snowmobile Pollution in North America: Annual Flux Estimates of Air Toxics and Implications for Potential Personal Exposure in Snowmobile Dominated Communities David Shively, Yong Zhou and Barkley Sive

103

Chapter 5

Modeling of Traffic-Related Environmental Pollution in the GIS Lubos Matejicek and Zbynek Janour

121

Chapter 6

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution in Taiwan From 1994 to 2006 Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu and Shuenn-Chin Chang

135

Chapter 7

Mobile Laboratories for Particle and Gaseous Pollutants U. Wa Tang, Ni Sheng and Zhishi Wang

149

Chapter 8

Urban Trees and Air Amelioration Capability Loretta Gratani, Laura Varone and Maria Fiore Crescente

161

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

1

vi

Contents

Chapter 9

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis John Shrimpton and Agissilaos Kourmatzis

179

Chapter 10

Recent Progress in Hydrogen-Fueled Internal Combustion Engines Sebastian Verhelst and Roger Sierens

211

Chapter 11

In Search of Improvements for the Computational Simulation of Internal Combustion Engines Ezequiel J. López, Norberto M. Nigro and Mario A. Storti

251

Chapter 12

Thermal Efficiency of Hydrogen Combustion Engine Toshio Shudo

339

Chapter 13

Dynamics of Pressure Fluctuations in Internal Combustion Engines Asok K. Sen and Grzegorz Litak

351

Chapter 14

Syngas Production by Partial Oxidation Using a Compression Ignition Engine Young Nam Chun

367

Chapter 15

A Possible Effect of Exhaust Emissions from Vehicles Using Unleaded Petrol on House Sparrow Populations J.D. Summers-Smith

379

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Short Communications

385

Benzene Exposure and the Story of Carcinogenesis: Experience of Traffic Policemen in Bangkok Viroj Wiwanitkit

387

Traffic Benzene Pollution in Bangkok: Air Benzene Level and Implication Viroj Wiwanitkit

395

Index

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

401

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

PREFACE Atmospheric pollution has been a major problem in human technological development and motor vehicles are one of the major sources of particulate matter pollution. This book investigates current models designed to predict air pollutant emissions and fuel consumption for road traffic and presents the outputs of statistical models developed to derive emission factors. Information on the use of Geographic Information Systems and traffic area air pollution monitoring stations is presented in order to comprehend the variations of trafficrelated air pollution. Furthermore, this book reports the pros and cons of hydrogen-fueled internal combustion engines, a study of the new technology to produce syngas from methane with a compression ignition engine. An overview of the characteristics of the factors influencing the thermal efficiency of spark ignition engines fueled with hydrogen is given as well. As discussed in Chapter 1, atmospheric pollution has been a nightmare in human technological development, which is being reflected worldwide. The worst scenarios are located in developing countries such as those located in Latin America, Asia and Africa. An interesting phenomena has been reported by Campell and Campell (2007) that they defines as “flattening of the cities” which is related with the spread of human settlements in periurban regions far from the city and health care facilities; also these settlements promote erosion, which facilitates particulate air pollution. This changes increase the need for transportation and the increase in traffic jams, as well as the accretion of atmospheric pollutants. Chapter 2 investigates current models designed to predict air pollutant emissions and fuel consumption for road traffic. It will consist of two parts: 1) a review of current road traffic emission modelling around the world, and 2) expected direction of further model development (outlook). The review will use a model classification framework that facilitates a structured discussion of model features, complexity, model application and prediction accuracy. The outcomes from the review are then discussed in light of current developments with respect to emission measurements, traffic control and in-vehicle technology. Motor vehicles are a major source of particulate matter pollution in urban areas and therefore estimates relating to the extent of this pollution are critically needed for urban and transport planning, scenario modelling, and development of mitigation strategies, air quality assessments and air quality regulation. More than 900 particle emission factors are available in the international published literature relating to motor vehicle tailpipe emissions, however it remained unclear which are the most appropriate to use in transport modelling and health impact assessments.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

viii

Sergey Demidov and Jacques Bonnet

A very large body of published data on emission factors was reviewed within the scope of this work, and based on a statistical analysis a comprehensive set of tailpipe particle emission factors for motor vehicles was derived. Chapter 3 presents the most appropriate emission factors derived to use in transport modelling and health impact assessments for particle number, particle volume, PM1, PM2.5 and PM10 to develop comprehensive, size-resolved inventories of tailpipe particle emissions for urban fleets in developed countries, covering the full size range of particles emitted. This chapter presents the outputs of statistical models developed to derive these emission factors, their explanatory variables, and the average particle emission factors, corresponding 95% confidence intervals and standard errors produced by the statistical models. Results of statistical tests which investigated the statistical relationship between sub-classes of categorical model variables for different particle metrics examined in the analysis are also presented. Yellowstone National Park’s (YNP) winter use planning and oversnow vehicle use issues have spawned the most intensive studies of snowmobile emissions and associated ambient air quality anywhere. These issues and the research they have engendered have served to improve our understanding of the dynamics of snowmobile use in winter landscapes and the attending exposure of people to air toxics, and can be seen as having contributed to the development of cleaner engine technologies and policies designed to improve the personal safety and health of those who might be exposed to snowmobile emissions. In Chapter 4 we examine the results of studies conducted in YNP prior to the promulgation of policies mandating the use of best available technology (BAT) equipped snowmobiles (i.e., cleaner burning 4-stroke machines), and existing data concerning snowmobile use in North America (i.e., the United States and Canada) to estimate annual fluxes of air toxics (benzene, toluene, ethyl benzene, xylenes, and n-hexane) for this region. While we do not attempt to present our results in a spatially disaggregated manner beyond national units, our results show that annual air toxic emissions from snowmobile use appear to be significantly higher (i.e., ~16-29%) than the USEPA’s estimates of air toxics emitted by snowmobiles in the United States. Additionally, we find that emissions associated with Canadian snowmobile use are likely to increase the U.S. figures. The implications of these fluxes are discussed in the context of agency air toxic assessment programs, and personal exposure to air toxics using data pertaining to the community of West Yellowstone, Montana. Our results provide an initial set of baseline data for potential personal exposure in this small mountain community to ambient air toxics coming from this non-regulated mobile source. Lastly, these data are discussed in the context of other North American communities in which snowmobiles represent an important component of the motorized vehicle fleet. As explained in Chapter 5, the numerical models are based on dispersion modeling and statistical analysis. In case of dispersion modeling, the ISC-AEROMOD View is used for modeling multiple pollutants with the U.S. EPA modeling tool ISCST3. The Mobile View assists as an interface for the U.S. EPA MOBILE6 model that predict arterial street emissions focused on hydrocarbons, carbon monoxide, nitrogen oxides, carbon dioxide, particulate matter, and toxics from cars, motorcycles, light- and heavy-duty trucks under various conditions. The potential impacts of accidental releases are solved by SLAB View that complements the modeling tools by analysis of emissions from accidental releases of toxic gases. Analysis of urban traffic-induced noise pollution is assessed by U.S. FHWA-TNM tools. The GIS is finally used to serve as a common analysis framework for individual

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Preface

ix

modeling tools. In order to display the numerical simulation outputs together with urban area map layers, numerical modeling based on U.S. EPA software tools is integrated into the GIS for spatial interpolations and spatial analysis. It assists to evaluate high levels of air pollution and noise pollution together with the thematic map layers of residential zones, business centers, schools, and hospitals. Finally, finding alternative routes can decreases air pollution and noise pollution in selected zones. As a case study, the city of Prague sample data set helps to demonstrate data processing and modeling of traffic-related environmental pollution. The ESRI’s geodatabase is used for implementation a comprehensive information model and a transaction model in the GIS environment. It is also the common application logic used in ArcGIS for accessing and working with all spatial thematic data and simulation inputs/outputs. Spatial interpolation for prediction maps and probability maps complement the existing thematic map layers, which enable cell based modeling for spatial multi criteria decision analysis. The synthesis of environmental models and GISs creates a more complex base for environmental simulation that can support decision-making processes in a more straightforward way. The Air Pollution Control Act (APCA) of Taiwan signed in 1975 prescribes the maximum permissible limits of motor vehicle exhausts as well as the monitoring of air pollution, etc. In Chapter 6, the concentrations of air pollutants including SO2, CO, O3, PM10, NO2, and non-methane hydrocarbons (NMHC) from background air pollution monitoring stations (BAQMSs) and traffic area air pollution monitoring stations (TAQMSs) were evaluated to comprehend the variations of traffic-related air pollution in Taiwan from 1994 to 2006. The results indicated that the background concentrations of SO2 increased from 1994, peaked at 1997 and decreased to 4.31 ppb at 2006. The concentrations of traffic area SO2 decreased from 16.09 ppb to 7.19 ppb during this period. The background concentrations of CO varied between 0.35 ppm and 0.51 ppm, while the concentrations of traffic area CO decreased from 5.20 ppm to 1.17 ppm from 1994 to 2006. The background concentrations of O3 increased from 1994, peaked at 2004 (34.10 ppb) and maintained at 33.51 ppb at 2006. The concentrations of traffic area O3 varied between 18.70 ppb and 25.44 ppb during this period. From 1994 to 2006, the background concentrations of PM10 varied between 41.55 μ g/m3 and 60.73 μ g/m3, while the concentrations of traffic area PM10 decreased from 119.39

μ g/m3 to 69.62 μ g/m3. The background concentrations of NO2 varied between

13.93 ppb and 16.54 ppb, while the concentrations of traffic area NO2 decreased from 55.85 ppb to 31.68 ppb from 1994 to 2006. The background concentrations of NMHC decreased from 0.61 ppm to 0.11 ppm. The concentrations of traffic area NMHC decreased from 2.52 ppb to 0.72 ppb from 1994 to 2006. From 1994 to 2006, SO2, CO, PM10, NO2 and NMHC from TAQMSs decreased by 55.3 %, 77.5 %, 41.7 %, 43.3 % and 71.4 %, respectively. Contrarily, the concentrations of O3 increased by 17.2 %. The traffic area air quality was improved as the transportation loading increased from 1994. Actually, many efforts including vehicle exhaust emission regulation, fuel and mechanical improvement for cars have been made after the signature of APCA in 1975. But according to the analysis, transportation was the major source of SO2, CO, PM10, NO2 and NMHC. For further improving the air quality in Taiwan, TEPA shall adopt stricter regulations for traffic area. Due to limited budget, installation space and labour resource, permanent monitoring sites are very scattered and thus complex distributions of particulates in various streets and urban environments are neglected. Recently, a number of mobile laboratories have been assembled

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

x

Sergey Demidov and Jacques Bonnet

for a variety of purposes such as monitoring the spatial and temporal distribution, investigating the emission and dispersion characteristics of tailpipe exhaust and aggregating fleet emission characteristics. Chapter 7 reviews their assembling of mobile laboratories, experimental designs and measurements. The authors’ work on the determination of gaseous emission factors for individual on-road vehicles is also introduced. Urbanisation processes have increased pollution levels in urban areas, and vehicular traffic is one of the major sources of air pollution. Nevertheless, understanding the urban ecosystem pollution dynamics requires long-term researches because of several factors involved (i.e. climate, pollution sources, urban characteristics, trees species, traffic density, vehicle type). In the last years, investigators are themselves concentrating on the possibility to develop models that can quantify the role of urban trees in removing pollutants from atmosphere. Nevertheless, these models need many data on the species composition, age, structure, and health. In Chapter 8 we analysed the main factors affecting atmospheric CO2 and heavy metal concentration in Rome in the long term, and the trees air amelioration capability, considering the most important deciduous and evergreen species widely distributed in the city. These species have a different role in carbon dioxide sequestration and in bioaccumulation of heavy metals. Such a role depends on many factors, including trees species, plant size and leaf longevity. In particular, information on air pollution could be deduced from heavy metal concentration in plant tissues, offering a low-cost information about urban environment quality. Moreover, crown volume is a discriminant factor for carbon sequestration, and leaf longevity is the most important leaf trait changing in response to heavy metal pollution. Nevertheless, incorrect pruning practices can reduce or undo the trees air amelioration role. Moreover, trees contribute to air temperature mitigation by shading and transpiration, thus lowering the energy consumption for air conditioning during summer. It is important to identify the tree species mostly contributing to air amelioration, and the most discriminant plant traits helping long-time monitoring and city management. The results might be exported into other urban areas to improve air quality enhancing social benefits. Downsizing direct injection spark ignition engines presents several challenges to the engine designer, but is a necessary requirement if significant savings in terms of fuel economy and CO2 emissions are to be realised. This challenge becomes more acute if we wish to employ a flexible fuel supply, for instance a range of biological fuel blends. These typically require more mechanical and thermal effort to provide good fuel vapour-air mixture preparation at ignition. In Chapter 9 these issues are investigated by comparing engine timescales (a function of engine size, speed and injection timing) against droplet timescales (a function of drop diameter, liquid physical properties and local air thermodynamic conditions). The analysis is used to make predictions of the target drop diameter required for a given engine size, speed and injection timing (load). Finally, we briefly explore the possibility of employing electrostatic charging to reduce the mass transfer timescale, since this is the limiting timescale for small direct injection spark ignition operation. Greenhouse gas emission by the transport sector is a hot topic these days. There is a strong drive towards legislation limiting the fleet average CO2 emissions. The use of hydrogen as an energy carrier is one option with the potential of lowering CO2 emissions investigated by the vehicle manufacturers. Governments from regions that are pollutant hot spots are also looking into hydrogen as a potential solution.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Preface

xi

Hydrogen is mostly associated with fuel cells. However, affordable fuel cell vehicles seem to be a long way off, both for technological as for economic reasons. An interesting alternative is using hydrogen in internal combustion engines (ICEs). Next to being less expensive, hydrogen-fueled ICEs offer a number of other benefits of which the most practical one is the ability to run in bi-fuel or flex-fuel operation. Research on hydrogen-fueled ICEs has been reported from the 1930’s onwards. In the past decade, the research has shifted from general laboratory-type ”proof of concept” work to efforts focused on achieving practical vehicles through thorough optimization. This includes modern concepts such as direct injection, supercharging, exhaust gas recirculation, variable valve timing etc. Chapter 10 reports the pros and cons of hydrogen-fueled ICEs, the properties of hydrogen relevant to its application as a fuel in engines and the resulting engine hardware and software features. The current state of the art is discussed, for achieving maximum power output with minimal emissions and minimal fuel consumption. Results from numerous engine test bench experiments in the authors’ department are shown, and results obtained by others are reviewed. The modeling of internal combustion engines is a multidisciplinary subject that involves chemical thermodynamics, fluid mechanics, turbulence, heat transfer, combustion, and numerical methods. In this chapter, we focus on some aspects of the computational resolution of the fluid dynamic problem. We present strategies designed in order to improve the simulation tools available today. In Chapter 11, a Computational Mesh Dynamics (CMD) to solve the movement of the mesh is presented. This kind of techniques are useful when an Arbitrary Lagrangian Eulerian (ALE) method is applied in the resolution of flows on moving domains. For in-cylinder flows in internal combustion engines, the domain has a very high relative deformation and even changes on its topology. This demands a CMD strategy with great robustness to avoid the deterioration of the grid quality and to reduce at minimum possible the number of remeshing needed in the whole simulation. The CMD strategy proposed is based on an optimization problem solved in a global way. The strategy can handle meshes with inverted elements, being a simultaneous mesh untangling and smoothing method. The flow inside of an internal combustion engine is characterized by a low Mach number, except in the early moments in which the exhaust valve (or port) is opened. The numerical methods for compressible flow based on the density fail when they are applied to flows with low Mach numbers, which is due to the ill-conditioning of the system of equations. For this reason, it is necessary to apply a technique that allows the resolution of compressible flows in all the range of Mach numbers, especially in the low Mach limit. Here, we apply the method of preconditioning of the equations in conjunction with the dual time stepping technique. The preconditioning matrix used was originally designed by Choi and Merkle to solve steady compressible flows with the Finite Volume Method. We adapt this matrix to the unsteady flows case via an eigenvalues analysis of the system. A stabilized Finite Element formulation for the preconditioned system of equations is presented. The dynamic boundary conditions on inlet/outlet of the 3D problem are obtained by using a code with thermodynamic (0D) and gas-dynamic (1D) models. Therefore, the need to couple appropriately the solutions obtained in the computational 1D and 3D domains arises. We propose a coupling strategy of 1D/multi-D domains for compressible flows based on constraints of the state at the coupling section.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

xii

Sergey Demidov and Jacques Bonnet

Finally, these tools are applied to solve the fluid flow in the novel rotative engine MRCVC. Hydrogen is in many ways an attractive clean alternative to fossil fuel. Hydrogen can be readily used as a fuel for spark ignition type internal combustion engines. There are drawbacks however, as hydrogen has a higher burning velocity and a shorter quenching distance than hydrocarbons, and these characteristics of hydrogen influence the degree of constant volume and the cooling loss by the heat transfer from burning gas to the combustion chamber walls. The cooling loss fraction and the degree of constant volume are two of the major factors influencing the thermal efficiency of internal combustion engines. So the investigation of thermal efficiency in hydrogen combustion is an important issue. Chapter 12 is an overview of the characteristics of the various factors influencing the thermal efficiency of spark ignition engines fuelled with hydrogen. In Chapter 13 we have studied the dynamics of cycle-to-cycle pressure fluctuations in internal combustion engines. Using wavelet analysis, we have examined the time series of maximum pressure variations under different loading. In particular, we have employed a continuous wavelet transform to identify the long, intermediate and short-term periodicities in the pressure signal. It is found that depending on the load, the long and intermediate-term periodicities may persist over several engine cycles, whereas the short-term periodicities occur intermittently. Results are given for both spark-ignition (SI) and diesel engines. Knowledge of these periodicities may be useful to develop effective control strategies for efficient combustion. An advantage of wavelet analysis is that in addition to detecting the various periodicities, it can delineate the number of engine cycles over which these periodicities may persist. It is essential to develop the environment-friendly alternative energies urgently considering the limited fossil fuel and the global warming caused by environmental destruction. In Chapter 14 the new technology was studied to produce syngas from methane with a compression ignition engine. This experiment was conducted on syngas production according to the variations of oxygen/methane ratio, total flow rate, air intake temperature and oxygen enrichment, with a commercial diesel engine without modification as a reformer with partial oxidation. Results showed that the concentration of hydrogen and carbon monoxide was 20.84% and 13.36%, respectively, under the optimal standard condition of oxygen/methane ratio 0.26, total flow rate 106.5 L/min and intake preheating temperature 355℃. Under the same condition, the concentration of hydrogen became 20.31% when the oxygen enrichment ratio was 55.6%, while that of carbon monoxide became 20.85% when the oxygen enrichment ratio was 50.33%. As presented in Chapter 15, the House Sparrow Passer domesticus is one of the most familiar wild birds. It is common, with an immense natural range that stretches from the west coast of Ireland to the east coast of Siberia, north to the Arctic Circle and south to North Africa, the Middle East, India and Sri Lanka. In addition, it has also been introduced to America, both North and South, South Africa, Australia, New Zealand and many points between, making it one of the most widespread of all birds; furthermore it is currently expanding its range by colonising new areas in the north of South America, West Africa and the Far East. Its success, as its name implies, is through its association with man, both in farmland and built-up areas, even extending into the centres of large towns. This association began about 10,000 years ago when man developed from a hunter/gatherer to a sedentary agriculturist, the bird taking advantage of the cereals that man developed from the large-

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Preface

xiii

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

seeded grasses, which had previously provided its diet, supplemented by the food put out for domesticated animals. Benzene is of particular concern because recent research indicates that benzene exposure can result in chronic toxicity including induction of hematological cancer. Benzene exposure and the story of carcinogenesis is a present focus in occupational health. In the first Short Communication, the author will present and comment on experience involving traffic policemen in Bangkok. Data on chromosome and flow cytometry study with correlation to biomarker monitoring will be reported and presented in this article. Benzene exposure is of particular concern because of recent research indicating that benzene exposure can result in several chronic toxicities including carcinogenesis. Long-term benzene exposure is hematotoxic, genotoxic and immunotoxic. Exposure to benzene from automobile exhaust can be a significant occupational problem for urban populations. Here, the author assesses the seasonal pattern of air benzene levels in an urban area of Bangkok. In the second Short Communication, the author also calculated the cancer risk for different occupations exposed to benzene vapor based on risk classification.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 1-28

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 1

TRAFFIC- RELATEDAIR POLLUTION T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez, N. López-Valdes, M. Ustarroz-Cano and L.F. Montaño Cellular and Tissular Biology Department, School of Medicine, National University of Mexico (UNAM) Mexico City CP 04510, Mexico

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Health Impact of Air Pollution Atmospheric pollution has been a nightmare in human technological development, which is being reflected worldwide. The worst scenarios are located in developing countries such as those located in Latin America, Asia and Africa. An interesting phenomena has been reported by Campell and Campell (2007) that they defines as “flattening of the cities” which is related with the spread of human settlements in periurban regions far from the city and health care facilities; also these settlements promote erosion, which facilitates particulate air pollution. This changes increase the need for transportation and the increase in traffic jams, as well as the accretion of atmospheric pollutants.

The Vehicular Traffic Factor The health impact that inhaled particulate matter has on human health and its association with myocardial infarctions, thromboembolic diseases and other related health problems has increase. Peters et al. (2004) identified the association of traffic-related air pollution and an increase in myocardial infarction events. Some factors increased de risk such as strenuous exercise, and there were no differences if the subject was riding a bicycle, or being transported in public transportation, made no difference. The report mentioned that women and subjects over 60 years of age or older, and patients with diabetes were at higher risk for the onset of an infarction when compared with men, non diabetic or subjects younger 60 years old, after the exposure to traffic. Stress, air pollution, and noise, as well as, individual

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

2

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

susceptibility are a dangerous cocktail which increases the chances of having an ischemic event on the road. As it was described above, usually epidemiological and clinical studies have helped to identify the association of air pollution and health events. Others have reported the increase in mortality and morbidity with lower life expectancy and greater disease severity in those exposed to vehicular traffic (Kappos et al., 2004; Makri and Stilianakis, 2008). Diverse pollutants induce a variety of effects in different organs, for example in the respiratory system nose and throat irritation will be associated to PM10, PM2.5, O3, CO, SO2 and NO2(Linker et al, 2000; Tolbert et al, 2000; Yu et al, 2000; Wong et al, 2001). In patients with different lung conditions, PM that enters the alveolar epithelium can initiate an inflammatory process that may deteriorate their condition. An interesting study conducted by Peel and co-workers (2005) showed a statistically significant association between the number of visits to the emergency unit for upper respiratory infections and Chronic Obstructive Pulmonary Disease, and higher outdoor levels of PM10, NO2 and CO. Peters and co-workers (2000) established that cardiac arrhythmias were associated with higher levels of NO2, CO and particulate matter. Later, this same author and others found that air pollutants such as PM10 or PM2.5, O3, and NO2 have been related with cardiac mortality and morbidity (Peters et al, 2001; Curtis et al, 2006), and that these pollutants may increase the risk of heart diseases at levels below standards set by US EPA or WHO (Figure 1).

Figure 1. Sources of air pollutants.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

3

A great amount of evidence has emerged indicating that the respiratory system is not the only target affected by the exposure to air pollution. A link between exposure to air pollution and insults on the cardiovascular system has now been established. A study made by the Massachusetts Department of Public Health (Boston, U.S.A.) recorded 107, 925 deaths from which they found that traffic particles were associated with increased deaths in the city of Boston, this finding is in agreement with other studies suggesting that traffic related particles may be connected with myocardial infarction, alteration in heart rate , and impaired endothelial function (Peters et al, 2001; Mann et al, 2002; Dominici et al, 2003). Other effect that many studies have observed is that air pollution, especially particulate matter increases fibrinogen and blood coagulability, facts that may lead to trombothic episodes and other cardiac events. Another consequence of outdoor air pollution is that it may increase the incidence of strokes. Neurological disorders are also related with air pollution. Neurotoxic effects such as memory disturbances, sleep disorders, anger, fatigue, hand tremors, blurred vision, and slurred speech had been reported (Ewan and Pamphlett, 1996; Garza et al, 2006). Among other traffic related air pollutants, carbon monoxide and nitrogen oxides have been implicated in higher rates of headaches and migraines among adults ; also carbon monoxide has been related with schizophrenia (Pedersen et al., 2004; Nattero and Enrico, 1996). Other pollutants originated from combustion processes, are known as carcinogens such as benzo(a)anthracene, benzo(k)fluoranthene, benzo(a)-pyrene, benzo(b)fluoranthene, indeno(1,2,3-cd)pyrene and dibenzo(ah)anthracene. Several studies indicate that these compounds are highly mutagenic and are capable to form bulky DNA adducts after metabolic activation and when unrepaired or missrepaired DNA , adducts may lead to mutations, that may ultimately, induce cancer formation (Pallia et al, 2008). Recent studies showed that bulky DNA adducts may be associated to increased lung cancer risk (Peluso et al, 2005; Bak et al, 2006; Pallia et al, 2008). Also, DNA adducts were detected in different sections of the heart including smooth muscle cells of human abdominal aorta affected by atherosclerotic lesions, atrial appendages from open heart surgery patients and human thoracic aorta samples from autopsy, so these may be additional evidence that links traffic related air pollution to increase cardiovascular disease and cancer (Lewtas, 2007).

Hydrocarbons These pollutants are very simple organic compounds with a shared chemical structure. Their components are carbon and hydrogen, and may be assembled in open or close rings (aromatic or aliphatic) with one or more ramifications; taking into account their hydrogencarbon proportion, hydrocarbons (HC) could be saturated or unsaturated (aliphatic, heterocyclic) (Figure 2). The unsaturated are the main components of fossil fuels and the principal products from their incomplete combustion are carbon monoxide and some hydrocarbons – for cars-, and vehicles which burn diesel produce suspended particles and black smoke derived from the incomplete burning of the hydrocarbons.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

4

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

HYDROCARBON CLASSIFICATION

HYDROCARBONS

ALIPHATIC (OPEN CHAINS)

ALICYCLIC OR CYCLIC

AROMATIC

HETEROCYCLIC

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 2. Schematic representation of hydrocarbons.

Some reports mention arrhythmias associated with diesel exposure, as well as the changes associated with PM 2.5 on which HC are adhered, mainly the aromatic and heterocyclic ones (Holguin M, 2007; Gotschi T, 2008). The most important association might be with cancer. The association of medullar aplasia with benzene has been stressed (Molina M, 2004). More recently some studies associate low birth weight with HC –polycyclic aromatic compounds - exposure (Slama R, 2007). The finding of new energy sources will reduce the atmospheric presence of these compounds, which until now have been moving the human world.

Nitrogen Oxides Nitrogen oxides are emitted as NO to the atmosphere, in which rapidly reacts with ozone or radicals in the atmosphere forming NO2. The main anthropogenic sources are mobile and stationary combustion sources. Moreover, ozone in the lower atmospheric layers is formed by a series of reactions involving NO2 and volatile organic compounds, a process initiated by sun light. The major contributor to nitrogen oxides (NOx ¼ NO+NO2) concentrations in urban areas is road traffic and NOx can be used as a tracer for road traffic emissions at monitoring sites located in urban areas. Nitrogen dioxide (NO2) is mainly a secondary pollutant since in most ambient situations nitrogen oxide (NO) is emitted and transformed into NO2 in the atmosphere through photochemical processes. NO and NO2 are major precursors of photochemical oxidant air pollution resulting in tropospheric ozone and smog formation (WHO, 1997). Ozone (O3) is a secondary pollutant, produced (via ultraviolet light) by the photochemically and temperature driven reactions of NOx and volatile organic compounds (VOCs) (Patz and Balbus, 2001). The major sources of

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

5

nitrogen oxides in North America are electric utilities and transportation vehicles fossil fuel combustion. Biochemically, NO2 acts as a strong oxidant and may lead to toxicity in lung cells through peroxidation of membrane lipids or membrane proteins, each of which increases cell permeability, leading to injury and death. NO2 may decrease the bactericidal activity of alveolar macrophages that are significant in preventing pneumonia. It also increases susceptibility to both bacterial and viral respiratory infections (Wolfe and Patz, 2002). Animal study findings suggest that humoral and cell-mediated defense systems are affected, possibly interfering with the ability of the lungs to remove inhaled particles. Other studies indicate that commonly encountered domestic NO2 concentrations may potentiate the specific airway response of mild asthmatics, leading to increased broncho-constriction, especially in young children (Anderson et al., 1998; Hajat et al., 1999; Schierhorn et al., 1999; Smith et al., 2000). Some studies found stronger associations between exposure and symptoms during the cumulative lags following exposure that immediately upon exposure (Anderson et al., 1999; Schierhorn et al., 1999). Objective lung injury findings include increased airway resistance and reduced diffusion capacity. An association has been shown between increasing outdoor air NO2 concentrations and upper respiratory infections, asthma, and mortality from respiratory illness among children less than 5-years old throughout the world. Low levels of NO2 (ranging from 0 to 40 ppb) showed no association with respiratory illness in healthy infants and toddlers (Samet et al., 1993). Recent studies have shown that chronic exposure to the levels of air pollutants, such as NO2, currently observed may have even higher impact on mortality than acute exposure. Furthermore, recent population studies document that NO2 is consistently associated with adverse health effects at relatively low levels of long-term average exposure, even when the annual average NO2 concentration complied with the annual guideline value suggested by the WHO, which is also the EU annual standard. Persons suffering from respiratory diseases, such as asthma, are very sensitive to NO2 at high concentrations, while several risk assessment studies have shown that both short- and long-term exposure to NO2 can induce effects to the human health and that given the role of NO2 as a precursor of other pollutants and as a marker of traffic-related pollution, there should be benefits for the public health from keeping low NO2 levels in the atmospheric air (European Commission, 1997) In Europe, some cities like Athens have problems with the NO2 pollution. In the urban area of Athens, it was shown that both mortality due to respiratory problems and hospital admissions for cardiac and respiratory causes are related with observed air pollution levels (Katsouyanni et al., 1990). The possible health effects from the observed NO2 levels in the Athens area were also examined using a dose–response relation, both for the urban background sites—corresponding to concentration levels at residential areas at which the population is usually exposed—but also for the traffic-affected sites in order to examine the risk of health effects from possible extreme cases of long-term exposure to concentration levels observed at ‘‘hot spot’’ locations. Exposure to high concentrations of NO2 may lead to development of acute pulmonary edema and, occasionally, to fatality. Survivors frequently develop lifelong chronic pulmonary fibrosis (Speizer et al., 1980). The WHO estimates that NO2 concentrations of 560–940 mg m

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

6

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

are needed to produce fatal pulmonary edema or asphyxia, and 47–140 mg /m3 for development of bronchitis or pneumonia (WHO, 1997).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Sulphur Oxides Sulfur dioxide (SO2) is a gas generated primarily in the process of burning fossil fuels containing sulfur. SO2 remains in the ambient air for 1-7 days, during which time it can be converted to sulfates and sulfuric acid by sunlight, photochemical oxidants, or by the catalytic effect of certain particulates in the air. These processes are complex and not quantitatively understood (Rall, 1974). The chemical form of sulphur oxides in the ambient air, which is associated in epidemiological studies with morbidity and mortality has not yet been clearly identified. The sulphur-containing products which have been implicated include SO2, sulfuric acid, and inorganic sulfates (Rall, 1974). SO2 can also be oxidized under photochemical conditions but the sulfur-oxygen bond is very strong so that sulphur dioxide cannot undergo the photo-dissociation. The oxidation process involves the hydroxyl radical and in combination with humid air (water aerosols) it reacts to sulfuric acid (H2SO3, H2SO4), that, if inhaled, exerts corrosive properties to the nasal mucosa, the trachea and the alveolar tissue, resulting in respiratory problems and severe attacks of coughing (Yip y Madl, 2002). There is a general agreement that the inhalation of sulphur dioxide exerts its primary toxicity on the respiratory system, specifically in the lung (Melhman, 1983). Also that the inhalation of 1 ppm sulphur dioxide for approximately 2 h may produce alterations in pulmonary ventilatory function, both in normal and asthmatic subjects (Snell and Luchsinger, 1969; Bates et al., 1973; Andersen et al., 1974; Lawther et al., 1975; Kreisman et al., 1977; Koenig et al., 1979, 1980). Some studies by Laskin et al, (1976) were negative in terms of lung carcinogenicity, when rats were exposed to 10 ppm sulfur dioxide alone. However, 24% of rats exposed to a combination of sulfur dioxide and benzo(a)pyrene developed lung neoplasms. The third series consisted of lifetime exposures with the following results: negative for lung carcinogenicity following exposure to sulfur dioxide alone, and the presence of lung neoplasm in 20% or fewer of rats exposed to both sulfur dioxide gas and benzo(a)pyrene. The positive observations by Laskin et al., have been interpreted as suggesting that sulphur dioxide gas is a cancer-promoting agent or a cocarcinogen to benzo(a)pyrene. Furthermore, a relationship has been demonstrated between outdoor SO2 levels and hospital admissions for asthma (Walters et al., 1994; Bates et al., 1990), bronchitis (Sunyer et al., 1993), and upper-respiratory-tract infections that does not depend on the occurrence of smog episodes (Ponka 1990). Experimental exposure studies have suggested that healthy subjects may show airway obstruction after inhaling high concentrations of SO2, and that asthmatic subjects respond at much lower concentrations (Sheppard et al., 1980). Mixtures of SO2 and particles are often more toxic than SO2 alone, depending on the SO2 concentration, the nature and size of the particles, and the ambient relative humidity. Particles can absorb SO2, and facilitate its reactions. Airborne particulate metals (vanadium, manganese, iron, etc.) catalyze the conversion of SO2 to sulfuric acid and sulfates. Sulfuric acid and acid sulfates have proved to be particularly toxic in animal experiments (Rall, 1974).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

7

Some negative effects of the sulphur oxides are linked to DNA damage. It has been reported that the frequencies of chromosomal aberrations (CA), sister chromatid exchanges (SCE), and micronuclei (MN) in peripheral blood lymphocytes of workers chronically exposed to SO2 in factories were higher than in unexposed controls (Beckman and Nordenson, 1986; Meng and Zhang, 1990a, b; Yadav and Kaushik, 1996). These results suggest that SO2 and its derivatives are clastogenic and genotoxic agents (Meng and Zhang, 1992). Other studies have shown that the SO2 derivatives (bisulfite and sulfite) may induce CA, SCE, and MN in cultured human blood lymphocytes in vitro (Meng and Zhang, 1992) and MN in mouse bone marrow cells in vivo (Meng et al., 2002a,b; Meng and Zhang, 2002). Meng et al (2004) made a study where they evaluated sodium bisulfite and sodium sulfite for their ability to induce DNA damage in various organs of mice using the SCGE (Single Cell Gel Electrophoresis) technique. They treated cells from various organs of mice with SO2 derivatives, at different doses. The damage was investigated with the alkaline SCGE assay. They evaluated DNA damage on cells of brain, lung, heart, liver, stomach, spleen, thymus, kidney and marrow. The ratio of cells with comet tails reached 450% even after low SO2 derivative exposure (125 mg/kg body wt), except for thymus and bone marrow. This implies that DNA in these cells, is very sensitive to the toxicological effects of SO2 and its derivatives. The alkaline comet assay shows that SO2 derivates might be strong DNAbreaking agents at the various doses tested. The results imply that SO2 and its derivates are systemic DNA-damaging agents. SO2 derivatives can cross the alveolar--capillary barrier, reach the blood and all organs and then damage DNA in the cells in various organs (Meng and Zhang, 2002; Meng, 2003). The mechanism of DNA damage induced by SO2 and its derivates is not clear and further studies are needed. The genetic effects of the SO2 derivatives sulfite and bisulfite have been examined in bacterial and mammalian cells. It has been suggested that DNA damage induced by SO2 and its derivatives might involve the following factors. At high concentrations and at pH values between 5 and 6 bisulfite modifies DNA in vitro and deaminates cytosine to uracil (Fishbein, 1976; Shapiro, 1977). Transitional mutations (G:C-A:T) induced by bisulfite were reported in Escherichia coli and bacteriophage T4 (Mukai et al., 1970; Summers and Drake, 1971). Furthermore, a lot of free radicals are produced during sulfite and bisulfite oxidation, such as SO3, SO4 and SO5 (Singh and Pathak, 1990; Shi and Mao, 1994). These free radicals can damage DNA and induce mutations (Singh and Pathak, 1990; Shi and Mao, 1994; Meng and Zhang, 1999). It has been shown that SO2 inhalation results in a significant increase in lipid peroxidation levels in multiple organs from mice of both sexes (Meng, 2003). Gaseous emissions from fossil-fuel power plants generally contribute more material to the atmosphere than other sources (Natusch, 1978). Sulphur oxides are not, themselves, carcinogenic. They are quite reactive and are known to interact with, for example, polycyclic aromatic species (National Academy of Sciences, 1972) and to promote lung damage when associated with airborne particles.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

8

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Ozone This pollutant is a secondary gas that is not directly emitted into the atmosphere, since it is the result of a chemical reaction between NOx and Volatile Organic Compounds (VOCs) in the presence of sunlight (Thrurston, 2007). As we mentioned previously, Ozone (O3) is generated at ground level by photochemical reactions between ultraviolet radiation and atmospheric mixtures of NO2 and VOCs derived from motor vehicle emissions. O3 levels depend on NO2 emitted by cars, and particularly on sunny weather that transforms NO2 into O3. Ozone is the most important factor for summer smog, because it accounts for up to 90% of total oxidant levels in cities with a mild sunny climate. Short-term O3 inhalation first modifies the ciliated cells, which seem to be the most sensitive, also Clara cells then underwent degranulation and destruction, while the reorganization of the epithelium takes place over seven days. O3 is a very strong oxidant, which has the ability to overwhelm the natural defences of the lungs. It induces lipid peroxidation and inactivation of biomolecules. This compound also activates stress signaling pathways in epithelial cells and interacts with the nuclear factor (NF)-kB. Ascorbic acid, uric acid and glutathione, are the first line antioxidants in the epithelial lining fluid when ozone exposure occurs. It also increases the permeability of epithelial cells, favoring the entry of inhaled allergens, toxins and a subsequent release of inflammatory cytokines such as: IL-1, IL-6, IL-8 and tumor necrosis factor (TNF). Mucociliary clearance decreased as O3 increased causing susceptibility to bacterial respiratory infections. Other studies suggest that O3 increases asthma morbidity by enhancing airway inflammation, as demonstrated by the increase of granulocyte macrophage colony stimulating factor (GM-CSF) and fibronectin in bronchoalveolar lavage fluid (Yang, 2005). Furthermore, O3 induces decrement in pulmonary function, increasing airway responsiveness and resistance and altering lung volumes and flow (Olivieri, 2005). Additional evidence demonstrate associations between the levels of this pollutant and adverse respiratory effects, such as decrements in lung function, aggravation of pre-existing respiratory disease, increase in hospital admission and premature deaths for respiratory causes.

Suspended Particles Suspended Particles (also called particulate matter or PM), usually consist of a carbon core to which complex mixtures of compounds, such as transition metals, polyaromatic hydrocarbons, ions, reactive gases, dust and biological material like endotoxins, and pollen adhere (Poschl, 2005). The particulate matter is produced by a wide variety of natural and anthropogenic activities, but the major sources are anthropogenic activities like fuel and derivates, combustion by factories, power plants, motor vehicles (heavy vehicles are the major contributors, due to the production of 50–100 times more fine and ultra-fine particles per km travelled) trash incinerators, and others like construction activity, fires, and natural windblown dust (Cariñanos et al., 2001; WHO 2003). Particulate matter (PM) varies in size. In general, particle pollution includes: "inhalable coarse particles," with aerodynamic diameters larger than 2.5 micrometers and smaller than

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

9

10 micrometers (PM10) and "fine particles," with aerodynamic diameters that are 1 to 2.5 micrometers (PM 2.5). The smallest particles, with aerodynamic diameter less than 0.1 μm, contain the secondarily formed aerosols (gas-to-particle conversion), combustion particles and recondensed organic and metal vapors (WHO 2003). The larger particles are mechanically produced by the breakup of larger solid particles and usually contain earth crust materials, wind-blown dust from agricultural processes, uncovered soil, unpaved roads or mining operations. Pollen grains, mould spores, plant and insect parts are all in this larger size range (WHO 2003). On the other hand, the fine fraction or PM 2.5 contains most of the acidity (hydrogen ion) and mutagenic activity of particulate matter. This matter is mainly produced by particles emitted directly into the atmosphere, and particles formed in the air from the chemical transformation of gaseous pollutants (secondary particles) (WHO 2003). The principal types of directly emitted particles are soil related, elemental/organic carbon (EC/OC) and other particles from the combustion of fossil fuels and biomass materials. However, PM2.5 have a low level of soil particle components, and the main anthropogenic source is a product of the combustion of fossil fuels. The main sources of PM2.5 include combustion of coal, oil, gasoline, diesel or wood, atmospheric transformation products of NOx, SO2 and organic compounds, natural and anthropogenic (WHO 2000; EPA 2002). The atmospheric lifetime of PM2.5 is in the order of days to weeks and it has been shown to travel hundreds to thousands of kilometers. These characteristics can result in prolonged exposure, promoting or aggravating health problems (WHO, 2000). The smallest particles, less than 0.1 μm, are formed by nucleation, that is, condensation of low-vapor-pressure substances formed by high-temperature vaporization or by chemical reactions in the atmosphere to form new particles (nuclei). Four major classes of sources with equilibrium pressures low enough to form nuclei mode particles can yield particulate matter: heavy metals (vaporized during combustion), elemental carbon, organic carbon, sulfates and nitrates (Ohlstrom et al., 2000; EPA 2002).

Figure 3. Deposition on the respiratory tract is related with the size of the particle. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

10

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

The size of the particles determines the site in the respiratory tract that they will deposit: PM10 particles deposit mainly in the upper respiratory tract , while fine and ultra fine particles are able to reach lung alveoli (Figure 3). Particle size is also directly linked to their potential for causing health problems. Small particles less than 10 micrometers in diameter cause the greatest problems, because they can get deep into lungs, and some may even get into bloodstream. Particulate air pollutants have been associated with increased respiratory, cardiovascular and cancer mortality and morbidity, and with other health problems (Dockery et al., 1993). The fine fraction of atmospheric particulate matter is of great concern because it is predominantly deposited in the alveolar region of the lung where absorption efficiency is higher and the overall removal of particles is relatively inefficient. Only about 20% of particles deposited in the alveolar region are cleared in the first day, and the remaining portion is cleared very slowly (Lippmann et al., 1980; WHO 2003; Kim et al., 2005). Fine fraction has shown a closer association with human adverse health effects than other particles (PM10) or total suspended particles (TSP) (Reichhardt, 1995). In addition several studies suggest an association between motorized traffic-related air pollution and diminished pulmonary function and/or increased respiratory symptoms such as irritation of the airways, coughing, or difficulty breathing, for example; decreased lung function; aggravated asthma; and development of chronic bronchitis, mainly in children, age and susceptible people (Kim et al., 2004; Weiland et al., 1994). Furthermore, particle pollution exposure produces irregular heartbeat; nonfatal heart attacks; and premature death in people with heart or lung diseases. Additionally, in vivo and in vitro studies demonstrate that ambient urban particulates may be more toxic than some surrogate particles such as iron oxide or carbon particles (Clarke et al., 1999; Clarke et al., 2000). For animal models of chronic bronchitis, cardiac impairment, or lung injury, increased susceptibility to PM has been established (Killingsworth et al., 1997; Clarke et al., 1999; Costa and Dreher 1997; Godleski et al., 2000). Animal studies have also shown that fine particulate matter recovered from cities can cause lung inflammation and injury (Clarke et al., 1999). In healthy and asthmatic volunteers, airborne particles increase bronchial responsiveness, airway resistance, and bronchial tissue mast cell, neutrophil, and lymphocyte counts. A specific role for ultrafine particles and metallic content of PM (especially iron) has been advocated (Wilson et al., 2002). Concentrations of PM that are somewhat higher than those common in ambient air in cities, are necessary to induce toxic effects in very short-term clinical experimental studies. Exposure to concentrated ambient air particles (23–311 µg/m3) for two hours induced transient, mild pulmonary inflammatory reactions in healthy human volunteers exposed to the highest concentrations, with an average of 200 μg/m3 PM2.5 (Gio et al., 2000). However, no other indicators of pulmonary injury, respiratory symptoms or decrements in pulmonary function were observed in association with acute exposure. In another study, exposure to ambient air particles (23–124 µg/m3) for 2 hours did not induce any observed inflammation in healthy volunteers (Petrovic et al., 1999). Particulate matter pollution may decrease life expectancy by long-term exposure to PM due to increase cardio-pulmonary and lung cancer mortality. This have been supported by cohort studies and by new analyses of time-series studies that have shown death being advanced by periods of at least a few months, for causes of death such as cardiovascular and chronic pulmonary disease ( Lagorio et al., 2006).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

11

About toxicity of particulate matter, there is strong evidence to conclude that fine particles (< 2.5 µm, PM2.5) are more hazardous than larger ones (coarse particles) in terms of mortality and cardiovascular and respiratory endpoints in panel studies. This does not imply that the coarse fraction of PM10 is innocuous. In toxicological and controlled human exposure studies, several physical, biological and chemical characteristics of particles have been found to elicit cardiopulmonary responses. Amongst the characteristics found to be contributing to toxicity in epidemiological and controlled exposure studies are metal content, presence of organic components, endotoxin and both small (< 2.5 µm) and extremely small size (< 0.100 µm) (Adamson et al., 1999) . Few epidemiological studies have addressed interactions of PM with other pollutants. Toxicological and controlled human exposure studies have shown additive and in some cases, more than additive effects, especially for combinations of PM and ozone, and of PM (especially diesel [exhaust] particles) and allergens. Finally, studies of atmospheric chemistry demonstrate that PM interacts with gases which alter its composition and hence its toxicity (Vincet et al., 1997). In relation to susceptibility, several studies have shown that elderly people and those with pre-existing heart and lung disease tend to be more susceptible to effects of ambient PM on mortality and morbidity. In panel studies, asthmatics have also been shown to be more vulnerable to ambient PM compared to non-asthmatics. Responses of asthmatics to PM exposure include increased symptoms, larger lung function changes, and increased medication use. In long-term studies, it has been suggested that socially disadvantaged and poorly educated populations respond more strongly in terms of mortality. PM exposure is also related to reduced lung growth in children. In cohort studies, no consistent differences have been found between men and women or between smokers and non-smokers in PM responses. People with heart or lung diseases, children and older adults are the most likely to be affected by particle pollution exposure. However, person healthy may experience temporary symptoms from exposure to elevated levels of particle pollution (Bateson and Schwartz 2004; Lagorio et al., 2006). So far, no single component has been identified that could explain most of the PM effects. Animal exposure studies have generally supported many of the findings reported in human studies and have provided additional information about mechanisms of toxicity. Studies considering the way that different particles deposit in the lungs, their chemical composition, and their toxicity provide further evidence of adverse health effects of PM. For example, some effects that are seen with the coarse particles may be due to the presence of microbial structures and toxins which are less frequently found associated with fine particles. Therefore, there is sufficient concern about the health effects of coarse particles to justify their control. In addition, the metal content, the presence of PAHs, mainly contribute to PM toxicity (Kampa and Castanas 2008). One of the proposed mechanisms to the observed adverse health effects is that PM can induce oxidative stress mediated by transition metals on the particle surface, by PM inducing inflammation causing phagocytes to release reactive oxygen species (ROS), and/or by quinines in the particles that produce ROS through redox cycling. Furthermore, oxidative stress might up-regulate redox sensitive transcription factors (via nuclear factor kappa B, NFkB) in airway epithelial cells, thus increasing the synthesis of proinflammatory cytokines and resulting in cell and tissue injury. Different PM fractions have been found to generate ROS. Soluble transition metals are abundant in the water-soluble PM fraction and in vitro studies

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

12

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

have shown that this fraction is able to induce ROS, specifically hydroxyl radicals, through the metal-dependent Fenton reaction both in cell free systems and biological systems. In the Fenton reaction, hydroxyl radicals are generated through a transition metal –mediated reduction of hydrogen peroxide. Studies have shown that particle suspensions from which the soluble fraction has been removed also can generate ROS, suggesting that other fractions of PM than the transition metals can induce oxidative stress (EPA 2002; Donaldson et al., 2003; Tao et al., 2003).

Heavy Metals Heavy metals are the most toxic component of particulate matter. It is well known that fine particles have high concentrations of many potentially toxic trace metals, such as cadmium (Cd), chromium (Cr), copper (Cu), iron (Fe), manganese (Mn), nickel (Ni), lead (Pb), vanadium (V) and zinc (Zn), that can be incorporated into the body through inhalation (Singh et al., 2002). The metal content on fine particles has been suggested as causative agents associated with adverse respiratory health effects (Dreher et al., 1997; Ghio et al., 1996). Most of the toxic metals in the air are in the form of fine particles, with a size distribution equivalent to that of aerosols (1.0 mm or less in diameter). It is suggested that these metals can produce lung tissue damage by catalyzing oxidant formation and promoting the release of inflammatory mediators and cytotoxicity (Frampton et al., 1999).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Lead Since leaded gasoline additives were phased out beginning in the 1970s, and control measures were implemented in industries, which have reduced air emissions, inhalation is no longer the major exposure pathway for the general population in many countries. In some foreign countries, however, leaded gasoline is still used, and the resulting emissions pose a major public health threat. Inhalation may be the primary route of exposure to some workers in industries that involve lead as well as for adults involved in home renovation activities (ATSDR 2005). The absorption and biologic fate of lead, once it enters the human body, depends on a variety of factors including nutritional status, health, and age. Adults typically absorb up to 20% of ingested lead. Most inhaled lead in the lower respiratory tract is absorbed, the lead that enters the body is excreted in urine or through biliary clearance (ultimately, in the feces) (ATSDR 2005). The chemical form of lead or lead compounds, entering the body is also a factor for the absorption and biologic fate of lead. Inorganic lead is not metabolized in the liver. Absorbed lead that is not excreted is exchanged primarily among three compartments: Blood, mineralizing tissues (bones and teeth), which typically contain the vast majority of the lead body burden, soft tissue (liver, kidneys, lungs, brain, spleen, muscles, and heart) (Kehoe et al. 1961; ATSDR 2005). On a molecular level, proposed mechanisms for toxicity involve fundamental biochemical processes. These include lead's ability to inhibit or mimic the actions of calcium

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

13

(which can affect calcium-dependent or related processes) and to interact with proteins (including those with sulfhydryl, amine, phosphate and carboxyl groups) (ATSDR, 2005). Although blood generally carries only a small fraction of total lead body burden, it does serve as the initial receptacle of absorbed lead and distributes it throughout the body, making it available to other tissues (or for excretion). The half-life of lead in adult human blood has been estimated to be from 28 days (Griffin et al. 1975) to 36 days. (Rabinowitz et al., 1976). Approximately 99% of the lead in blood is associated with red blood cells; the remaining 1% resides in blood plasma. (DeSilva 1981; EPA 1986a; Everson and Patterson 1980). The bones and teeth of adults contain about 94% of their total lead body burden; in children is approximately 73% (Barry 1981). Lead in mineralizing tissues is not uniformly distributed. It tends to accumulate in bone regions undergoing the most active calcification at the time of exposure. Known calcification rates of bones in childhood and adulthood suggest that lead accumulation will occur predominately in trabecular bone during childhood, and in both cortical and trabecular bone in adulthood (AufderHeide and Wittmets 1992). Under certain circumstances, however, this apparently inert lead will leave the bones and reenter the blood and soft tissue organs. Bone-to-blood lead mobilization increases during periods of pregnancy, lactation, menopause, physiologic stress, chronic disease, hyperthyroidism, kidney disease, broken bones, and advanced age, all which are exacerbated by calcium deficiency. Because lead from past exposures can accumulate in the bones (endogenous source), symptoms or health effects can also appear in the absence of significant current exposure. Lead toxicity can affect every organ or system. The nervous system is the most sensitive target of lead exposure. Neurological effects of lead in children have been documented at exposure levels once thought to cause no harmful effects ( 60 µg/dL). Lead can induce anemia, often accompanied by basophilic stippling of the erythrocytes. (ATSDR, 1999). Acute high-level lead exposure has been associated with hemolytic anemia. The anemia of lead intoxication is hypochromic, and normo- or microcytic with associated reticulocytosis. On other hand, lead exposure is one factor of many that may contribute to the onset and development of hypertension. Recent reproductive function studies in humans suggest that current occupational exposures decrease sperm counts and increase abnormal sperm frequencies (Alexander et al. 1996; Lin et al. 1996; Telisman et al. 2000).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

14

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Cadmium It is estimated that about 25,000 to 30,000 tons of cadmium are released to the environment each year, about half from the weathering of rocks into river water and then to the oceans. Forest fires and volcanoes also release some cadmium to the air. Release of cadmium from human activities is estimated at from 4,000 to 13,000 tons per year, with major contributions from mining activities, and burning of fossil fuels. Cadmium can enter the air from the burning of fossil fuels (e.g., coal fired electrical plants) and from the burning of household waste. Cadmium that is in or attached to small particles can enter the air and travel a long way before coming down to earth as dust, or in rain or snow. Air levels greater than 40 ng/m³ may occur in urban areas with high levels of air pollution from the burning of fossil fuels (ATSDR 1999). There is evidence that inhalation of cadmium, especially cadmium oxide, cause prostate and kidney cancer in humans; cadmium has been shown to cause lung and testicle cancer in animals. It is classified by the National Occupational Health and Safety Commission (NOHSC) as a category 2 carcinogen (substance which should be regarded as carcinogenic to humans). Studies on workers exposed to cadmium in the air have not resulted in convincing evidence that cadmium can cause lung cancer in humans. In animals studies, mice or hamsters that breathed in cadmium did not get lung cancer, but rats that breathed in cadmium did develop lung cancer (Verougstraete and Lison 2003; Waalkes 2003; Sahmoun et al., 2005). On the other hand, breathing air with very high levels of cadmium can severely damage the lungs and may cause death or permanent lung damage that occurs due to rapid lung damage, shortness of breath, chest pain, and a build-up of fluid in the lungs. High exposures also can causedeath. Also, inhalation of smoke from burning cadmium or from cadmium oxide is toxic to the respiratory system. Breathing air with lower levels of cadmium over long periods of time (for years) results in a build-up of cadmium in the kidney, and if sufficiently high, may result in kidney disease (Buchet et al., 1990). Other effects that may occur after breathing cadmium for a long time are lung damage and fragile bones. Cadmiun is also a teratogen, and may cause reproductive damage. It is unlikely that this sort of exposure would occur, except in cases of unusual industrial accidents. Long-term exposures can cause anemia, fatigue and loss of the sense of smell. High exposure may also cause nausea, vomiting, cramps, and diarrhea (Satho et al., 2002). Breathing cadmium has also been shown to cause liver damage and changes in the immune system in rats and mice. There is no reliable information on people to indicate that breathing cadmium harms peoples' liver, heart, nervous system, or immune system (ATSDR 1999).

Chromium Air emissions of chromium are predominantly of trivalent chromium, and in the form of small particles or aerosols (Eighmy et al., 1995; Johnson and Furrer 2002). The most important industrial sources of chromium in the atmosphere are those related to ferrochrome production. Ore refining, chemical and refractory processing, the metal finishing industry, cement-producing plants (cement contains chromium), automobile brake lining and catalytic converters for automobiles, leather tanneries, the combustion of oil and coal, motor vehicle exhaust (crude oil contains traces of chromium (III) compounds, which may oxidize to the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

15

chromium (VI) state during fuel combustion in vehicle engines), and chrome pigments also contribute to the atmospheric burden of chromium (Eighmy et al., 1997). Airborne chromium (VI) particles will settle in less than 10 days, depending on particle size, and will stick strongly to soil particles. Rain will remove chromium (VI) particles from the atmosphere, depositing them in the ground, or they may be transported over long distances by wind. The general population is exposed to chromium (generally chromium [III]) by eating food, drinking water, and inhaling air that contains the chemical. The average daily intake from air, water, and food is estimated to be less than 0.2 to 0.4 micrograms (µg), 2.0 µg, and 60 µg, respectively (Johnson and Furrer 2002). Chromium III is much less toxic than chromium (VI). Chromium (III) is an essential element in humans. The body can detoxify some amount of chromium (VI) to chromium (III). Chromium exposure may occur from natural or industrial sources of chromium. Occupational exposure to chromium occurs from chromate production, stainless-steel production, chrome plating, and working in tanning industries; occupational exposure can be two orders of magnitude higher than exposure to the general population. People who live in the vicinity of chromium waste disposal sites or chromium manufacturing and processing plants have a greater probability of elevated chromium exposure than the general population. These exposures are generally to mixed chromium (VI) and chromium (III) (Johnson and Furrer 2002). The respiratory tract is the major target organ for chromium (VI) toxicity, for acute (short-term) and chronic (long-term) inhalation exposures. Shortness of breath, coughing, and wheezing were reported from a case of acute exposure to chromium (VI), while perforations and ulcerations of the septum, bronchitis, decreased pulmonary function, pneumonia, and other respiratory effects have been noted from chronic exposure. Human studies have clearly established that inhaled chromium (VI) is a human carcinogen, resulting in an increased risk of lung cancer. Also animal studies have shown chromium (VI) to cause lung tumors via inhalation exposure (ATSDR 1998). Other effects noted from acute inhalation exposure to very high concentrations of chromium (VI) include gastrointestinal (including abdominal pain, vomiting, and hemorrhage) and neurological effects, while dermal exposure causes skin burns, contact dermatitis, sensitivity, and ulceration in humans. Additionally chronic human exposure to high levels of chromium (VI) by inhalation or oral exposure may produce effects on the liver, kidney, immune systems, and possibly the blood (Hjelmar et al., 2001; Lundtorp et al., 2002; Johnson and Furrer 2002). About reproductive effects of chromium (VI), it was reported that humans exposed by inhalation to chromium (VI) might result in complications during pregnancy and childbirth (Johnson and Furrer 2002). However, animal studies have not reported reproductive or developmental effects from inhalation exposure to chromium (VI). Oral studies have reported severe developmental effects in mice such as gross abnormalities and reproductive effects including decreased litter size, reduced sperm count, and degeneration of the outer cellular layer of the seminiferous tubules (Hjelmar et al., 2001) No information is available on the reproductive or developmental effects of chromium (III) in humans (Eighmy et al., 1997). On the other hand, epidemiological studies of workers have clearly established that inhaled chromium is a human carcinogen, resulting in an increased risk of lung cancer. Although chromium-exposed workers were in contact to both chromium (III) and chromium

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

16

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

(VI) compounds, only chromium (VI) has been found to be carcinogenic in animal studies, so EPA has concluded that only chromium (VI) should be classified as a human carcinogen (Cai et al., 2003; Johnson and Furrer 2002; Lundtorp et al., 2002). EPA has classified chromium (VI) as a Group A, known human carcinogen by the inhalation route of exposure (Cai et al., 2003). No data are available on the carcinogenic potential of chromium (III) compounds alone. EPA has classified chromium (III) as a Group D, not classifiable as to carcinogenicity in humans (Lechner et al., 1997; Johnson and Furrer 2002).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Vanadium Vanadium pentoxide is the most common commercial form of vanadium and the primary compound found in industrial situations (Dill et al., 2004). Also, it is the most toxic compound (Barceloux, 1999). Vanadium pentoxide is considered as the marker element of air pollution emitted from the combustion of fossil fuels, particularly residual fuel oils, which constitute the single largest overall release of vanadium to the atmosphere (Wang et al., 1999). This compound has gained recent attention due to its status as an atmospheric pollutant principally in megacities like Mexico City (Fortoul et al., 2002). Atmospheric emissions of vanadium from natural sources have been estimated at 8.4 tons per annum globally (range 1.5-49.2 tons). Natural sources, in order of importance, are continental dusts, volcanoes, seasalt spray, forest fires, and biogenic processes (Nriagu, 1990). By far the most important source of environmental vanadium contamination is combustion of oil, with coal combustion as the second most important. Of the estimated total global emissions from both natural and anthropogenic sources of 64 000 tons per annum to the atmosphere, 58 500 tonnes come from oil combustion, with more than 33 500 tons of this accounted for by the developing economies in Asia and just under 14 500 tons by Eastern Europe and the former USSR. There are considerable regional variations in vanadium emissions (Nriagu and Pirrone, 1998). Inhalation is the principal route of entry into the body. In lungs approximately 90% of vanadium is absorbed. Absorbed vanadium is transported mainly in the plasma, bound to transferrin. Vanadium is widely distributed in body tissues; principle organs of vanadium retention are lungs, kidneys, liver, testicles, spleen and bones. A major fraction of vanadium from cellular vanadium was found retained in nuclei (Sabbioni et al., 1978). In pregnant rats the injected vanadium was found in the fetus (Soremark et al., 1962). Vanadium is excreted mainly in the urine, but also in the faces. Bile is probably not an important pathway for excretion into the faces, but the existence of alternative routes for excretion into the gut (salivary excretion or direct transfer across the intestinal wall) has been suggested (Sabbioni et al., 1981). Ambient vanadium pentoxide dust produces irritation of eyes, nose and throat (Hauser et al., 1995). Over long periods, inhalation may potentiate chronic bronchitis, eczematous skin lesions, fine tremors of the extremities, greenish discoloration of the tongue and gastrointestinal disturbances (Aw TC et al., 2007). Inhalation of vanadium pentoxide may result acutely in severe pneumonitis with associated mucus membrane irritation. The lungs are a significant site of entry of vanadium in the case of community exposure, and the distribution pattern of particles and the solubility of

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

17

vanadium compounds, as well as alveolar and mucociliar clearance, are important factors that determine the rate of absorption in the respiratory tract (WHO 2000). In addition vanadium pentoxide can injure the lungs and bronchial airways, (MSDS 2005) possibly involving acute chemical pneumonotis, pulmonary edema and/or acute tracheobronchitis (Nemery 1990). Symptoms include irritation and inflammation of the mucus membranes, nasal passages and pharynx. Clinical complications include a persistent cough, shortness of breath, bronchiolar constriction, tightness of the chest and a pseudoasthmatic inflammation (MSDS 2005). Alveolar/bronchiolar neoplasms developed in male rats exposed to 0.5 and 2 mg/m3 vanadium pentoxide, with only a marginal increase thereof in female rats exposed to 0.5 mg/m3. There were associated increase in inflammation, fibrosis and alveolar and bronchiolar hyperplasia/metaplasia and squamous metaplasia in both male and female rats (Ress et al., 2003). In an investigation of cynomolgus monkeys exposed to vanadium pentoxide dust for six hours per day, five days per week for 26 weeks, it demonstrated that airway obstruction accompanied a significant influx of inflammatory cells into alveolar tissue (Knecht et al., 2006). In an earlier study, it was suggested that vanadium-induced pulmonary inflammatory changes involving polymorphonuclear leukocytes may play an important role in air-flow limitation (Vandenplas et al., 2002). Respiratory tract inflammation following inhalation of vanadium particles is characterized by abundant neutrophilia initiated by alveolar macrophages and release of proinflammatory cytokines (Grabowski et al., 1999). Short-term, repeated inhalation of occupationally relevant levels of vanadium by rats results in pulmonary immunocompetence via cytokine production and pulmonary macrophage induction (Cohen et al., 1997). Its cumulative effect in lung tissue possibly contributes to the development lung cancer. Also vanadium exposure induces thrombocytosis and may be associated with various thromboembolic diseases. Fortoul et al., (2008) reported the influence of vanadium on megakaryopoyesis, suggesting that inhaled vanadium could induce megakaryocytic proliferation, which may result in increased platelet production and increased risk for thromboembolic events. Also Gonzalez-Villalva et al., (2006) reported that vanadium inhalation produce an increase in platelets, as well as the presence of megaplatelets. In addition, vanadium exposure produces alterations in the immune system. Spleens of V2O5 exposed animals showed an increased number of very large and non-clearly delimited germinal centers. In addition, their red pulp was poorly delimited and had an increase in CD19+ cells with hyperplasic germinal nodes. Vanadium pentoxide induces histological changes and functional damage to the spleen, each of which appear to result in severe effects on the humoral immune response (Piñon-Zarate et al., 2008). Moreover, vanadium induces alteration in liver. Chronically, histopathological changes observed in the liver following the higher level of inhalation exposure (27μg/m3 for 70 days) included central vein congestion with scattered small hemorrhages and granular degeneration of hepatocytes (WHO 2000). Severe acute exposure to vanadium pentoxide has major pathophysiological manifestations on the nervous system (WHO 2000). Inhalation produces a time-dependent loss of dendritic spines, necrotic-like cell death and considerable alterations of the hippocampus CA1 neurophile, all associated with spatial memory impairment (Avila-Costa et al., 2006). Additionally, there is a decrease in the number of tyrosine hydroxylase immunreactive neurones in the substatia nigra pars compacta (Avila-Costa et al., 2004).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

18

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Within the ependymal epithelium, cilia loss, cell sloughing and cell layer detachment occur after vanadium pentoxide inhalation (Avila-Costa et al., 2005). The damage results in disrupted permeability of the epithelium and promotes access of inflammatory mediators to the underlying neuronal tissue causing injury and neuronal death (Avila-Costa et al., 2005). In humans, severe chronic exposure results in general symptomatology including nervous disturbances, neurasthenic or vegetative symptoms (WHO 2000). Severe acute exposure (tens of mg/m3) is responsible for aberrations in renal function. Chronically in experimental animals, histopathological changes observed in the kidneys following the higher level of inhalation exposure (27μg/m3 for 70 days) included marked granular degeneration of the epithelium of the convoluted tubules. Dose-dependent histological changes included corticomedullary microhemorrhagic foci in the kidneys (Domingo et al., 1985). Finally chronic ingestion of vanadium may have significant consequences for fertility by damaging spermatogenesis. Studies in mice have demonstrated that inhalation of vanadium pentoxide results in necrosis of spermatogonium, spermatocytes and Sertoli cells (Fortoul et al., 2007). Vanadium accumulates in the testes and attenuates the percentage of gammatubulin in all analyzed testicular cells, suggesting changes in the microtubules used in cell division (Mussali et al., 2005). Vanadium also induces DNA damage (AltamiranoLozano et al., 1996). Leydig cells may not be affected by vanadium pentoxide as testosterone levels remain unchanged (Li et al., 1995).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Final Remarks The wide range of health effects that the myriad of components that could be found in the air polluted cities, could fill the same amount of pages. Here we summarized the most common or studied pollutants, but there should be all possible combinations which are waiting to be analyzed. Everyday new compounds are emitted into the atmosphere, and more complex reactions might occur. New alloys are used in vehicle manufacture; recent discovered compounds are added to gasoline in order to modify its efficiency. Maybe traffic is not the main problem to solve, but more investment in engine development and in cleaner gasoline and fuel derivates should be encourage. Also, more and better roads, highways and freeways must be developed. But more important is to make people understand that they are a very important part of the problem. If humans do not assume their part in the pollution chain, our environment will maintain its degradation, and the cost will be in more chronic diseases and a worse life quality. We will pay in health affection our indifference and negligence.

References Adamson IY, Prieditis H and Vincent R. (1999). Pulmonary toxicity of an atmospheric particulate sample is due to the soluble fraction. Toxicol. Appl. Pharmacol. 157:43-50 Agency for Toxic Substances and Disease Registry (ATSDR). (1998). Toxicological Profile for Chromium. U.S. Public Health Service, U.S. Department of Health and Human Services, Atlanta, GA.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

19

Agency for Toxic Substances and Disease Registry (ATSDR). (1999). Toxicological Profile for Cadmium. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service. Agency for Toxic Substances and Disease Registry. (2005). Toxicological profile for lead. Atlanta: US Department of Health and Human Services, Public Health Service. Alexander H., Checkoway H., van Netten C., et al. (1996). Semen quality of men employed at a lead smelter. Occup. Environ. Med. 53:411-416. Altamirano-Lozano M, Alvarez-Barrera L, Basurto-Alcántara F, Valverde M and Rojas E. (1996). Reprotoxic and genotoxic studies of vanadium pentoxide in male mice. Teratog. Carcinog. Mutagen. 16:7-17. Andersen I., Lundqvist CR., Jensen PL., and Proctor DF. (1974). Human response to controlled levels of sulfur dioxide. Arch. Environ. Health 28: 31. Anderson HR., Ponce de Leon A., Bland J.M., Bower J.S., Emberlin J. and Strachan D.P. (1998). Air pollution, pollens, and daily admissions for asthma in London 1987– 1992. Thorax. 53: 842–848. Aufderheide AC., Wittmers LE. Jr. (1992). Selected aspects of the spatial distribution of lead in bone. Neurotoxicol. 13:809-820. Avila-Costa MR., Colín-Barenque L., Zepeda-Rodríguez A., Antuna SB., Saldivar OL., Espejel-Maya G., Mussali-Galante P., del Carmen Avila-Casado M., Reyes-Olivera A., Anaya-Martinez V., Fortoul TI. (2005). Ependymal epithelium disruption after vanadium pentoxide inhalation. A mice experimental model. Neurosci. Lett. 381(1-2):21-5. Avila-Costa MR., Fortoul TI., Niño-Cabrera G., Colín-Barenque L., Bizarro-Nevares P., Gutiérrez-Valdez AL., Ordóñez-Librado JL., Rodríguez-Lara V., Mussali-Galante P., Díaz-Bech P., Anaya-Martínez V. (2006). Hippocampal cell alterations induced by the inhalation of vanadium pentoxide (V2O5) promote memory deterioration. Neurotoxicology 27(6):1007-12. Avila-Costa MR., Montiel Flores E., Colin-Barenque L., Ordoñez JL., Gutiérrez AL., NiñoCabrera HG., Mussali-Galante P., Fortoul TI. (2004). Nigrostriatal modifications after vanadium inhalation: an immunocytochemical and cytological approach. Neurochem Res. 29(7):1365-9. Aw TC, Gardiner K, Harrington JM. Chapter 5. (2007). Occupational toxicology. In: Pocket consultant. Occupational Health. 5th ed. Blackwell Publishing: Oxford; p. 71-114. Bak H., Autrup H., Thomsen BL., Tjonneland A., Overvad K., Vogel U. (2006). Bulky DNA adducts as risk indicator of lung cancer in a Danish case–cohort study. Int. J. Cancer . 118:1618–22. Barceloux D.G. (1999). Vanadium. Clinical Toxicology. 37:265-78. Barry PSI. (1981). Concentrations of lead in the tissues of children. Br. J. Ind. Med. 38:61-71. Bates DV., and Hazucha M. (1973). The short-term effects of ozone on the human lung. In: Proceedings, Assembly of Life Sciences, Health Effects Air Pollution Conference, National Academy of Sciences and National Research Council, Washington, DC. pp. 507-540. Bates DV., Baker-Anderson M., and Sizto R. (1990). Asthma attack periodicity: a study of hospital emergency visits in Vancouver. Environ. Res. 51:51–70. Bateson TF., Schwartz J. (2004). Who is sensitive to the effects of particulate air pollution on mortality? A case crossover analysis of effect modifiers. Epidemiology 15:143-149.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

20

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Beckman L., Nordenson I. (1986). Interaction between some common genotoxic agents. Hum. Hered. 36: 397-401. Buchet J., Lauwerys R., et al. (1990). "Renal effects of cadmium body burden of the general population." Lancet 336: 699-702. Cai Z., Jensen DL., Christensen TH. and Bager DH. (2003). Reuse of stabilized flue gas ashes from solid waste incineration in cement-treated base layers for pavements. Waste Management and Research 21: 42–53 Campbell C and Campbell A. (2007). Emerging Disease Burdens and the Poor in Cities of the Developing World. J Urban Health: Bull New York Acad of Medicine. 84: i54-164. Cariñanos P., Prieto JC., Galán C and Domínguez E. (2001). Solid Suspended Particles Affecting the Quality of Air in Urban Environments. Bull. Environ. Contam. Toxicol. 67:385–391. Clarke RW. Et al. (1999).Urban air particulate inhalation alters pulmonary function and induces pulmonary inflammation in a rodent model of chronic bronchitis. Inhalation toxicology, 11: 637–656. Clarke RW. Et al. (2000) Inhaled concentrated ambient particles are associated with hematologic and bronchoalveolar lavage changes in canines. Environ Health perspect. 108: 1179–1187. Cohen MD, Becker S, Devlin R, Schlesinger RB and Zelikoff JT. (1997). Effects of vanadium upon poly: C-induced responses in rat lung and alveolar macrophages. J. Toxicol. Environ. Health. 51:591-608. Costa DL. and Dreher KL. (1997). Bioavailable transition metals in particulate matter mediate cardiopulmonary injury in healthy and compromised animal models. Environmental health perspectives. 105: 1053–1060. Cuadrat J.M y Pita MF. (1997). Climatología. Cátedra. 496-502. Curtis L., Rea W., Smith-Willis P., Fenyves E and Pan Y. (2006). Adverse health effects of outdoor air pollutants. Environment International. 32: 815–830. DeSilva PE. (1981). Determination of lead in plasma and studies on its relationship to lead in erythrocytes. Br. J. Ind. Med. 38:209-217. Dill A.J., Lee M.K and Mellinger H.K. (2004). Lung deposition and clearance of inhaled vanadium pentoxide in chronically exposed F344 rats and B6C3F1 mice. Toxicol Sci. 77: 6-18. Dockery WD., Pope III CA., Xu X., Spengler JD., Ware JH., Marthia EF., Ferris BG., Speizer RE. (1993). An association between air pollution and mortality in six US cities. N. Engl. J. Med. 329: 1753–1759. Domingo JL, Llobet JM, Tomas JM and Corbella J. (1985). Short-term toxicity studies of vanadium in rats. J. Appl. Toxicol. 5:418-21. Dominici F, McDermott A, Daniels D. Mortality among residents of 90 cities. In: Special Report: Revised Analyses of Time-Series Studies of Air Pollution and Health, Boston, MA: Health Effects Institute. 2003: 9-24. Donaldson K., Stone V., Borm PJ., Jimenez LA., Gilmour PS., Schins RP., Knaapen AM., Rahman I., Faux SP., Brown DM., MacNee W. (2003). Oxidative stress and calcium signaling in the adverse effects of environmental particles (PM10). Free Radic Biol Med 34:1369-1382.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

21

Dreher KL., Jaskot RH., Lehmann JR., Richards JH., McGee JK., Ghio AJ. and Costa DL. (1997) Soluble transition metals mediate residual oil fly ash induced acute lung injury; J. Toxicol. Environ Health 50: 285–305. Eighmy TT., Eusden JD., Krzanowski JE., Domingo DS., Stämpfli DM., Martin JR and Erickson PM. (1995): Comprehensive approach toward understanding element speciation and leaching behavior in municipal solid waste incineration electrostatic precipitator ash. Environ. Sci. Technol. 29:629-646. Eighmy TT., Crannell, BS., Butler LG., Cartledge FK., Emery EF., Oblas D., Krzanowski JE., Eusden JD Jr., Shaw EL and Francis CA. (1997). Heavy metal stabilization in municipal solid waste combustion dry scrubber residue using soluble phosphate. Environ. Sci. Technol., 31, 3330-3338 EPA. (1986a). Air quality criteria for lead. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research and Development, Office of Health and Environmental Assessment. Environmental Criteria and Assessment Office. EPA 600/883-028F. EPA (1992). Prescribed Burning Background Document and Technical Information Document for Prescribed Burning Best vailable Control Measures, EPA-450/2-92-003, Office of ir Quality Planning and tandards, Research Triangle Park, C, p. 317. Everson J., Patterson CC. (1980). “Ultra-clean” isotope dilution/mass spectrometric analyses for lead in human blood plasma indicate that most reported values are artificially high. Clin. Chem. 26:1603-1607. Ewan K.B and Pamphlett R. (1996). Increased inorganic mercury in spinal motor neurons following chelating agents. Neurotoxicology. 17: 343-350. Fishbein L. (1976). Atmospheric mutagens I. Sulfur oxides and nitrogen oxides. Mutat. Res. 32: 309-330. Fortoul TI., Bizarro-Nevares P., Acevedo-Nava S., Piñón-Zárate G., Rodríguez-Lara V., Colín-Barenque L., Mussali-Galante P., Avila-Casado M del C., Avila-Costa MR., Saldivar-Osorio L. (2007). Ultrastructural findings in murine seminiferous tubules as a consequence of subchronic vanadium pentoxide inhalation. Reprod. Toxicol. 23(4):58892. Fortoul TI., Piñón-Zarate G., Diaz-Bech ME., González-Villalva A., Mussali-Galante P., Rodriguez-Lara V., Colin-Barenque L., Martinez-Pedraza M., Montaño LF. (2008). Spleen and bone marrow megakaryocytes as targets for inhaled vanadium. Histol. Histopathol. 23(11):1321-6. Fortoul TI., Quan-Torres A., Sánchez I., López IE., Bizarro P., Mendoza ML., Osorio LS., Espejel-Maya G., Avila-Casado M del C., Avila-Costa MR., Colin-Barenque L., Villanueva DN., Olaiz-Fernandez G. (2002). Vanadium in ambient air: concentrations in lung tissue from autopsies of Mexico City residents in the 1960s and 1990s. Arch. Environ. Health 57(5):446-9. Fortoul TI., Saldivar LO., Espejel-Maya G., Mussali-Galante P., Ávila-Casado MC., ColinBarenque L and Ávila-Costa MR. (2005). Inhalation of cadmium, lead or its mixture. Its effects on the bronchiolar structure and its relation with metals tissue concentrations. Environ. Toxicol. and Pharmacol. 19: 329-334. Frampton MW., Andrew J. and Ghio AJ. et al. (1999). Effects of aqueous extracts of PM10 filters from the Utah valley on human airway epithelial cells. American Journal of Physiology 277:L960-67.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

22

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Garza A., Vega R and Soto E. (2006). Cellular mechanisms of lead neurotoxicity. Med. Sci. Monit. 12: 57-65. Ghio AJ., Stonehuerner J. and Pritchard RJ. et al. (1996). Humic-like substances in air pollution particulates correlate with concentrations of transition metals and oxidant generation. Inhalation Toxicology 8: 479-94. Ghio AJ.Et al. (2000). Concentrated ambient air particles induce mild pulmonary inflammation in healthy human volunteers. American journal of respiratory and critical care medicine, 162: 981-988. González-Villalva A., Fortoul TI., Avila-Costa MR., Piñón-Zarate G., Rodriguez-Lara V., Martínez-Levy G., Rojas-Lemus M., Bizarro-Nevarez P., Díaz-Bech P., Mussali-Galante P., Colin-Barenque L. (2006). Thrombocytosis induced in mice after subacute and subchronic V2O5 inhalation. Toxicol. Ind. Health 22(3):113-6. Gotschi T, Heinrich J, Sunyer J and Kunzli N. (2008). Long-term effects of ambient air pollution on lung function: a review. Epidemiology. 19: 690-701. Grabowski GM., Paulauskis JD., Godleski JJ. (1999). Mediating phosphorylation events in the vanadium-induced respiratory burst of alveolar macrophages. Toxicol. Appl. Pharmacol 156:170-178. Griffin TB., Couiston F., Wills H. (1975). Biological and clinical effects of continuous exposure to airborne particulate lead. Arh .Hig. Toksikol 26:191-208. Hajat S., Haines A.., Goubet SA., Atkinson RW. and Anderson HR. (1999). Association of air pollution with daily GP consultations for asthma and other lower respiratory conditions in London. Thorax . 54: 597–605. Hauser R, Elreedy S, Hoppin JA and Christiani DC. (1995). Upper airway response in workers exposed to fuel oil ash:nasal lavage analysis. Occup. Environ. Med.. 52:353-8. Hjelmar O., Birch H., Hansen JB. In Proceedings Sardinia (2001) Eighth International Waste Management and Landfill Symposium, S. Margherita di Pula, Cagliari, Italy; Christensen, T. H., Cossu, R., Stegman, R., Eds.; CISA, 2001; Vol. 1: 667-675. http://cfpub.epa.gov/ncea/cfm/partmatt.cfm?ActType=default. (2002). Holguin F, Flores S, Ross Z, Cortez M, Molina M, Molina L, Rincon C, Jerrett M, Berhane K, Granados A and Romieu I. (2007). Traffic-related exposures, airway function, inflammation, and respiratory symptoms in children. Am. J. Respir. Crit. Care Med. 15: 1236-1242. Jarup L., (2003). Hazards of heavy metal contamination. Br. Med. Bull. 68:167-182. Johnson CA., Furrer G. (2002). Influence of biodegradation processes on the duration of CaCO3 as a pH buffer in municipal solid waste incinerator bottom ash. Environ. Sci. Technol. 36: 215-220. Kagawa J. (1985). Evaluation of biological significance of nitrogen oxides exposure. Tokai J. Exp. Clin. Med. 10: 348-353. Kampa M and Castanas E. (2008). Vulnerability to air pollution health effects. Environ. Pollution. 151: 362-367. Kappos AD., Bruckmann P., Eikmann T., Englert N., Heinrich U., Hoppe P., Koch E., Krause GH., Kreyling WG., Rauchfuss K., Rombout P., Schulz- Klemp V., Thiel WR and Wichmann HE. (2004). Health effects of particles in ambient air. Int. J. Hyg. Environ. Health. 207: 399-407. Katsouyanni K., Karakatsani A., Messari I., Touloumi G., Hatzakis A., Kalandidi A and Trichopoulos D. (1990). Air pollution and cause specific mortality in Athens. Journal of Epidemiology and Community Health . 44: 321-324.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

23

Kehoe RA. (1961). The metabolism of lead in man in health and disease: Present hygienic problems relating to the absorption of lead: The Harben lectures, 1960. J. R. Inst. Public Health Hyg. 24:177-203. Killingsworth CR. Et al. (1997). Inflammation, chemokine expression, and death in monocrotaline-treated rats following fuel oil fly ash inhalation. Inhalation toxicology 9: 541–565. Kim JJ., Smorodinsky S. and Lipsett M. et al. (2004). Traffic related air pollution near busy roads/ the East Bay children’s respiratory health study. Am. J. Resp Crit. Care Med. 170: 520-/26. Kim E., Hopke PK., Kenski DM and Koerber M. (2005). Sources of fine particles in a rural Midwestern U.S. area. Environ. Sci. Technol. 39(13): 4953-/60. Knecht EA, Moorman WJ, Clark JC, Delon Hull R, Biagini RE and Lynch DW. (2006). Pulmonary reactivity to vanadium pentoxide following subchronic inhalation exposure in a non-human primate animal model. J. Appl. Toxicol. 12:427-34. Koenig JQ., Pierson WE and Frank R. (1979). Acute effects of inhaled sulfur dioxide and exercise on pulmonary function in asthmatic adolescents. J. Allergy Clin. Immunol. 63: 154. Koenig JQ., Pierson, WE., and Frank R. (1980). Acute effects of inhaled SO2 plus NaCl droplet aerosol on pulmonary function in asthmatic adolescents. Environ. Res. 22: 145153. Kreisman H., Mitchell CA., Hosein HR and Bouhuys A. (1977). Effect of low concentrations of sulfur dioxide on respiratory function in man. Lung. 154: 25-34 Lagorio S., Forastiere F., Pistelli R., Iavarone I., Michelozzi P., Fano V., Marconi A., Ziemacki G and Ostro B.D. (2006). Air pollution and lung function among susceptible adult subjects: a panel study. Environmental Health: A Global Access Science Source 5:11 Laskin S., Kuschner M., Sellakumar A., and Katz GV. (1976). Combined carcinogen-irritant animal inhalation studies. In: Air Pollution and the Lung (E. F. Aharonson, A. BenDavid and M. A. Klingberg, Eds.), Wiley, New York. pp. 190-213. Lawther PJ., MacFarlane AJ., Waller RE., and Brooks AG. (1975). Pulmonary function and sulphur dioxide, some preliminary findings. Environ. Res. 10: 355-367. Lechner, P., Huber, H.; Japos, M. In Proceedings Sardinia (1997) Sixth International Waste Management and Landfill Symposium, S. Margherita di Pula, Cagliari, Italy; Christensen, T. H., Cossu, R., Stegman, R., Eds.; CISA, Vol. 5, pp 553-563. Lewtas J. (2007). Air pollution combustion emissions: Characterization of causative agents and mechanisms associated with cancer, reproductive, and cardiovascular effects. Mut. Res. 636: 95–133. Li H, Yang Z and Zhang T. (1995).Toxicity of vanadium to Leydig cells in vitro. Hua Xi Yi Ke Da Xue Xue Bao. 26:433-5. Lin S., Hwang S., Marshall EG., et al. (1996). Fertility rates among lead workers and professional bus drivers: A comparative study. Ann. Epidemiol. 6:201-208. Linaker CH., Coggon D., Holgate ST., Clough J., Josephs L and Chauhan AJ. (2000). Personal exposure to nitrogen dioxide and risk of airflow obstruction in asthmatic children with upper respiratory infection. Thorax. 55:930–933.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

24

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Lippmann M. Et al. (2000). Association of particulate matter components with daily mortality and morbidity in urban populations. HEI Health Effects Institute, Cambridge Massachusetts 95: 1–86. http//:www. healtheffects.org. Lundtorp K., Jensen DL., Sorensen MA., Mogensen EPB. and Christensen TH. (2002). Treatment of waste incinerator air-pollution-control residues with FeSO4: Concept and product characterization. Waste Management and Research 20: 69-79. Makri A and Stilianakis NI. (2008). Vulnerability to air pollution health effects. Int. J. Hyg. Environ. Health. 211: 326–336. Maheswaran R., Haining RP., Brindley P., Law J., Pearson T and Fryers PR. (2005). Outdoor air pollution and stroke in Sheffield, United Kingdom. A small-area level geographic study. Stroke. 36:239–43. Mann JK, Tager IB, Lurmann F. Air pollution and hospital admissions for ischemic heart disease in persons with congestive heart failure or arrhythmia. Environ. Health Perspect. 2002; 110:1247-1252. Meng Z. and Zhang L. (1990). Chromosomal aberrations and sister chromatid exchanges in lymphocytes of workers exposed to sulphur dioxide. Mutat. Res.241: 15-20. Meng Z. and Zhang L. (1990). Observation of frequencies of lymphocytes with micronuclei in human peripheral blood cultures from workers in a sulphuric acid factory. Environ. Mol. Mutagen. 15: 218-220. Meng Z. and Zhang L. (1992). Cytogenetic damage induced in human lymphocytes by sodium bisulfite. Mutat. Res. 298: 63-69. Meng Z and Zhang B. (1999). Polymerase chain reaction-based deletion screening of bisufite (sulfur dioxide)—enhanced gpt-mutants in CHO-AS52 cells. Mutat. Res. 425: 81-85. Meng Z and Zhang B. (2002). Induction effects of sulfur dioxide inhalation on chromosomal aberrations in mouse bone marrow cells. Mutagen. 17: 215-217. Meng Z. Zhang,B. Ruan A. Sang N and Zhang J. (2002). Micronuclei induced by sulfur dioxide inhalation in mouse bone-marrow cells in vivo. Inhal. Toxicol. 14: 303-309. Meng Z., Sang N and Zhang J. (2002). Effects of derivatives of sulfur dioxide on micronuclei formation in mouse bone marrow cells in vivo. Bull. Environ. Contam. Toxicol. 69: 257264. Meng Z. 2003. Oxidative damage of sulfur dioxide on various organs of mice: sulfur dioxide is a systemic oxidative damage agent. Inhal. Toxicol. 15: 181-195. Meng Z., Qin G., Zhang B and Bai J. (2004). DNA damaging effects of sulfur dioxide derivatives in cells from various organs of mice. Mutagen. 19(6). 465—468. Molina M. (2004). Propuesta para Limpiar el Aire en México en 10 años. Reporte del Taller sobre la Contaminación del Aire en México. 1-20. MSDS. (2005). Vanadium pentoxide. Environmental Health and Safety: Phillipsburg, New Jersey; p. 5. Mukai,F. Hawryluk,I. and Shapiro,T. (1970). The mutagenic specificity of sodium bisulfite. Biochem. Biophys. Res. Commun. 39: 983-998. Mussali-Galante P, Rodriguez-Lara V, Hernández-Tellez B, Avila-Costa MR, ColinBarenque L and Bizarro-Nevares P. (2005). Inhaled vanadium pentoxide decrease gamma-tubulin of mouse testes at different exposure times. Toxicol. Ind. Health. 21:21522. Nattero G and Enrico A. (1996). Outdoor pollution and headaches. Headache. 36: 243–5.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

25

National Academy of Sciences (U.S). Vapor phase organic pollutants. Commitee on biological effects of atmospheric pollutants, Division of Medical Sciences, National Research Council, National Academy of Sciences, Washington, D.C., 1976. Natusch DFS. (1978). Potentially Carcinogenic Species Emitted to the Atmosphere by FossilFueled Power. Environ Health Perspect. 22: 79-90 Nemery B. (1990). Metal toxicity and the respiratory tract. Eur. Respir. J. 3:202-19. Nriagu J. (1990). Global metal pollution: Poisoning the biosphere. Environment. 32:7–11. Nriagu J and Pirrone N. (1998) Emission of vanadium into the atmos phere. In: Nriagu J, ed. Vanadium in the environment. Part 1: Chemistry and biochemistry. New York, NY, John Wiley and Sons, pp. 25–36. Ohlstrom MO., Lehtinen K. and Moisio M. et al. (2000). Fine-particle emissions of energy production in Finland. Atmospheric Environment 34: 3701-11. Olivieri D., Scoditti E. (2005) Impact of environmental factors on lung defences. Eur. Respir. Rev. 14:51–56. Pallia D., Saievaa C., Munniab A., Peluso M., Grechic D., Zannaa I., Cainia S., Decarlid A., Seraa F and Masalaa G. (2008). DNA adducts and PM10 exposure in traffic-exposed workers and urban residents from the EPIC-Florence City study. Sci. Total Environ. 403: 105 – 112. Patz JA and Balbus JM. (2001). In: Ecosystem Change and Public Health: a Global Perspective. Aron, J.L. and Patz, J.A. (eds). Johns Hopkins University Press, pp. 379– 408. Paz-Elizur T., Krupsky M. and Blumenstein S. et al. (2003). DNA repair activity for oxidative damage and risk of lung cancer. Journal of the National Cancer Institute 95: 1312-19. Pedersen CB., Raaschou-Nielsen O., Hertel O and Mortenson PB. Air pollution from traffic and schizophrenia. Schizophr. Res. 2004; 66:83–5. Peel J., Paige T., Klein M., Metzger KB., Flanders WD and Todd K. Ambient air pollution and respiratory emergency department visits. Epidemiology. 2005; 16:164–74. Peluso M., Munnia A., Hoek G., Krzyzanowski M., Veglia F and Airoldi L. (2005). DNA adducts and lung cancer risk: a prospective study. Cancer Res. 65: 8042–8. Peters A., Liu E., Verrier RL., Schwartz J., Gold DR and Mittleman M. (2000). Air pollution and risk of cardiac arrhythmia. Epidemiology. 11:11–7. Peters A, Dockery DD, Miller JF, Mittelman MA. (2001). Increased particular matter and the triggering of myocardial infarction. Circulation. 103:2810–2815. Peters A., von Klot, Heier M., Trentinaglia I., Hörmann A., Wichmann HE., Löwel, H. (2004). Exposure to Traffic and the Onset of Myocardial Infarction. New. Eng. J. Med. 351:1721-1730. Petrovic S. Et al. (1999). Cardiorespiratory effects of concentrated ambient PM2.5: a pilot study using controlled human exposures. Inhalation toxicology 12: 173-188. Piñon-Zarate G., Rodriguez-Lara V., Rojas-Lemus M., Martinez-Pedraza M., GonzalezVillalva A., Mussali-Galante P., Fortoul TI., Barquet A., Masso F., Montano LF. (2008). Vanadium pentoxide inhalation provokes germinal center hyperplasia and suppressed humoral immune responses. J. Immunotoxicol. 5(2):115-22. Ponka A. (1990). Absenteeism and respiratory disease among children and adults in Helsinki in relation to low-level air pollution and temperature. Environ. Res. 52:34–36. Pöschl U. (2005). Atmospheric aerosols: composition, transformation, climate and health effects. Angew. Chem. Int. Ed. Engl. 44(46):7520-40.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

26

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Rabinowitz MB., Wetherill GW., Kopple JD. (1976). Kinetic analysis of lead metabolism in healthy humans. J. Clin. Invest. 58:260-270. Rall DP. (1974). Review of the health effects of sulfur oxides. Environ. Health Perspect. 8: 97-121. Reichhardt, T. (1995). Weighting the health risks of airborne particulates. Environment and Science Technology 29: 360-64. Ress NB, Chou BJ, Renne RA, Dill JA, Miller RA and Roycroft JH. (2003). Carcinogenicity of inhaled vanadium pentoxide in F344/N rats and B6C3F1 mice. Toxicol. Sci . 74:28796. Sabbioni E, Marafante E, Amantini L, Ubertalli L and Birattari C. (1978). Similarity in metabolic patterns of different chemical species of vanadium in the rat. Bioinorg. Chem. 8:503-15. Sabbioni E, Marafante E, Rade J, Gregotti C, Di Nucci A and Manzo L. (1981). Biliary excretion of vanadium in rats. Toxicol. Eur. Res. 3:93-8. Sahmoun, A., L. Case, et al. (2005). Cadmium and prostate cancer: a critical epidemiological analysis. Cancer Investigation 23: 256-263. Samet JM., Lambert WE., Skipper BJ., Cushing AH., Hunt WC., Young SA., McLaren LC, Schwab M and Spengler JD. (1993). Nitrogen dioxide and respiratory illnesses in infants. Am. Rev. Resp. Dis. 148: 1258–1265. Satoh M., Koyama H., et al. (2002). Perspectives on cadmium toxicity research. Tohoku J. Exp. Med. 196: 23-32. Schierhorn K., Zhang M., Matthias C and Kunkel G. (1999). Influence of ozone and nitrogen dioxide on histamine and interleukin formation in a human nasal mucosa culture system. Am. J. Resp. Cell Molec. Biol. 20, 1013–1019. Shapiro R. (1977). Genetic effects of bisulfite (sulfur dioxide). Mutat. Res. 38: 149-176. Sheppard D., Wong WS., Uehara CF.,. Nadel JA, and Boushey HA.. (1980). Lower threshold and greater bronchomotor responsiveness of asthmatic subjects to sulfur dioxide. Am. Rev. Respir. Dis. 122:873–878. Shi X. and Mao Y. (1994) 8-Hydroxy-20-deoxyguanosine formation and DNA damage induced by sulfur trioxide anion radicals. Biochem. Biophys. Res. Commun. 205: 141147. Singh, M., Jaques, P.A. and Sioutas, C. (2002). Size distribution and diurnal characteristics of particle-bound metals in source and receptor sites of the Los Angeles Basin. Atmospheric Environment 36: 1675-/89. Singh R and Pathak,D.N. (1990) Lipid peroxidation and glutathione peroxidase, glutathione reductase, superoxide dismutase, catalase and glucose-6-phosphate dehydrogenase activities in FEC13- induced epileptogenic foci in the rat brain. Epilepsia. 31: 15-26. Slama R, Morgenstern V, Cyrys J, Zutavern A, Herbarth O and Wichmann H. (2007). TrafficRelated Atmospheric Pollutants Levels during Pregnancy and Offspring’s Term Birth Weight: A Study Relying on a Land-Use Regression. Environ. Health Perspec. 115: 1284-1292. Smith BJ., Nitschke M., Pilotto LS., Ruffin RE., Pisaniello DL and Willson KJ. (2000). Health effects of daily indoor nitrogen dioxide exposure in people with asthma. Eur. Respir. J. 16: 879–885.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic- Related Air Pollution

27

Snell RE., and Luchsinger PC. (1969). Effects of sulfur dioxide on expiratory flow rates and total respiratory resistance in normal human subjects. Arch. Environ. Health. 18: 693698. Soremark R, Ullberg S and Appelgren LE. (1962). Autoradiographic localization of V-48labelled vanadium pentoxide (V2O5) in developing teeth and bones of rats. Acta. Odontol. Scand. 20:225 32. Speizer FE., Ferris B., Bishop YMM and Spengler J. (1980). Respiratory disease rates and pulmonary function in children associated with NO2 exposure. Am. Rev. Resp. Dis. 121: 3–11. Summers GW and Drake JW. (1971). Bisulfite mutagenesis in bacteriophage T4. Genetics. 68: 603--607. Sunyer JM., Murillo SC., Castellsague J., Martinez F., and Anto JM.. 1993. Air pollution and emergency room admissions for chronic obstructive pulmonary disease: a 5-year study. Am. J. Epidemiol. 137: 701–705. Tao F., Gonzalez-Flecha B., Kobzik L. (2003). Reactive oxygen species in pulmonary inflammation by ambient particulates. Free Radic. Biol. Med. 35:327-340. Telisman S., Cvitkovic P., Jurasovic J., Pizent A., Gavella M., Rocic B. (2000). Semen quality and reproductive endocrine function in relation to biomarkers of lead, cadmium, zinc, and copper in men. Environ. Health Perspect 108:45-53 Tolbert PE., Mulholland JA., MacIntosh DL., Xu F., Daniels D and Devine OJ. (2000). Air quality and pediatric emergency room visits for asthma in Atlanta, Georgia, USA. Am. J. Epidemiol. 151:798–810. Thurston GD and Wallace LA. Air pollution: outdoor and indoor sources. Chap 93. In Ron WN and Markowitz SB. Environmental and Occupational Medicine. Philadelphia. Wolters Kluwer/Lippicott Williams and Wilkins. 2007. Uysal N. and Schapira RM. (2003). Effects of ozone on lung function and lung diseases. Curr. Opin. Pulm. Med. 9, 144-154. Vandenplas O, Binard-Van Caugh F, Gregoire J, Brumagne A and Larbanois A. (2002). Fever and neutrophilic alveolitis caused by a vanadium based catalyst. Occup. Environ. Med.. 59:785-7. Verougstraete V. and Lison D. (2003). "Cadmium, lung, and prostate cancer: a systematic review of recent epidemiological data." Journal of Toxicology and Environmental Health 6(Part B): 227-255. Vincent R., Bjarnason SG., Adamson IY., Hedgecock C., Kumarathasan P., Guenette J., Potvin M., Goegan P., Bouthillier L. (1997). Acute pulmonary toxicity of urban particulate matter and ozone. Am. J. Pathol. 151:1563-1570. Waalkes M. (2003). "Cadmium carcinogenesis." Mutation Research 533: 107-120. Walters S., Griffiths RK and. Ayres JG. (1994). Temporal association between hospital admissions for asthma in Birmingham and ambient levels of sulphur dioxide and smoke. Thorax. 49:133–140. Wang C., Chang C. Y., Chin C. J. and Men L. C. (1999). Determination of arsenic and vanadium in airborne related reference materials by inductively coupled plasma-mass spectrometry. Anal. Chim. Acta. 392: 299–306. Weiland SK., Mundt KA. and Ruckmann A. et al. (1994). Selfreported wheezing and allergic rhinitis in children and traffic density on street of residence. Annals of Epidemiology 4: 243- 47.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

28

T.I. Fortoul, V. Rodríguez-Lara, C.I. Falcón-Rodríguez et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

WHO Air Quality Guidelines for Europe. (2000). Second edition. Copenhagen, WHO Regional WHO. World health report (2002). Geneva, World Health Organization. WHO. World health report (2003). Geneva, World Health Organization. WHO. (2000). Vanadium. Chapter 6.12. WHO Regional Office for Europe, Copenhagen: Denmark; p. 9. WHO (1997). Environmental Health Criteria, No. 188, 2nd edition. WHO, Geneva. 550-570. Wilson MR. Et al. (2002). Interactions between ultrafine particles and transition metals in vivo and in vitro. Toxicology and applied pharmacology 184: 172-179. Wong GW, Ko FW, Lau TS., Li ST., Hui D and Pang SW. (2001). Temporal relationship between air pollution and hospital admissions for asthmatic children in Hong Kong. Clin. Exp. Allergy. 31:565–569. Wolfe AH. and Patz JA. (2002). Reactive Nitrogen and Human Health: Acute and Long-term Implications. Ambio. 31(2): 120-125. Yang Z., Ballinger S.W. (2005) Environmental contributions to cardiovascular disease: Particulates and ozone. Drug Discovery Today: Disease Mechanisms. 2:71-75. Yadav JS. and Kaushik VK. (1996). Effects of sulphur dioxide exposure on human chromosomes. Mutat. Res. 359: 25-29. Yip M and Madl P. (2002). Air Pollution in Mexico City. Project-Study paper, University of Salzburg, Austria, Department of Biophysics and in Cooperation with the Afro-Asian Institute (Salzburg, Austria) and International Laboratory for Air Quality and Health at QUT (Australia). http://www.sbg.ac.at/ipk/avstudio/pierofun/mexico/air.html/Effects Yu O., Sheppard C., Lumley T., Koenig JQ and Shapiro GG. (2000). Effects of ambient air pollution on symptoms of asthma in Seattle-Area children enrolled in the CAMP study. Environ. Health Perspect. 106:1209–14. Websites: European Commission, 1997. Position paper on air quality: nitrogen dioxide. Working Group on Nitrogen Dioxide, EC. http://ec.europa.eu/environment/air/pdf/pp_no2.pdf.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 29-68

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 2

ROAD TRAFFIC EMISSION AND FUEL CONSUMPTION MODELLING: TRENDS, NEW DEVELOPMENTS AND FUTURE CHALLENGES Robin Smit1, Hussein Dia2 and Lidia Morawska3 1

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

PAEHolmes, 59 Melbourne Street, South Brisbane, Australia, QLD 4101, [email protected] 2 ITS Research Laboratory, School of Engineering, The University of Queensland, Brisbane, Australia, QLD 4072 3 School of Physical and Chemical Sciences, Queensland University of Technology, International Laboratory for Air Quality and Health, 2 George Street, Brisbane, Australia, QLD 4001

Abstract This chapter investigates current models designed to predict air pollutant emissions and fuel consumption for road traffic. It will consist of two parts: 1) a review of current road traffic emission modelling around the world, and 2) expected direction of further model development (outlook). The review will use a model classification framework that facilitates a structured discussion of model features, complexity, model application and prediction accuracy. The outcomes from the review are then discussed in light of current developments with respect to emission measurements, traffic control and in-vehicle technology.

Introduction Transport is a major source of air pollution and greenhouse gas emissions around the world, and its significance in this respect is increasing. The problems and issues relating to traffic are (perhaps) surprisingly similar in the affluent nations around the world. From an air quality perspective, road traffic is particularly significant since it emits large quantities of harmful chemicals close to populated areas. In fact, around the world, road traffic is the dominant anthropogenic source of air pollution in urban areas (e.g. Fenger 1999). We can

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

30

Robin Smit, Hussein Dia and Lidia Morawska

expect this to remain the case as reductions in emissions from individual vehicle (e.g. due to stricter vehicle emissions standards) are at least partially offset by continued growth in traffic and elevated congestion levels. Emissions from individual vehicles are a function of many, often interacting, variables. First, air pollutant emissions and fuel consumption vary substantially with vehicle design characteristics, which include, but are not limited to: • • • • • • •

vehicle size and weight; engine type; type of fuel; transmission type; presence and type of emission control technology; presence of auxiliary equipment such as air conditioning; and aerodynamic characteristics.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

In addition to these vehicle-related factors, the way a vehicle is being driven (driving behaviour) also affects vehicle emissions and fuel consumption. Driving behaviour itself is the result of many factors such as the interaction with road features (speed limits, intersections, road width, road condition, gradient, and so forth), level of congestion, road grade, personal driving style (aggressive, gentle, etc.), gear shift behaviour, use of auxiliary equipment and weather related conditions (ambient temperature, humidity, rain, fog, etc.). Finally, various other factors impact on emission levels, such as: • • • • •

driving mode (cold, hot; running, idling, parked); deterioration of engine and emission control components (ageing effects); engine tuning and maintenance; fuel composition and characteristics; and geographic location reflected in e.g. air density and altitude.

Importantly, these factors are often interrelated and may influence different “types of emissions”, namely: •





Hot running emissions: exhaust emissions that occur under "hot stabilized" conditions, which means that the engine and the emission control system (e.g. catalytic converter) have reached their typical operating temperatures. Start emissions: exhaust emissions that occur in addition to hot running emissions because engines and catalysts are not (fully) warmed up and operate in a non-optimal manner. Evaporative emissions: non-exhaust hydrocarbon losses through the vehicle's fuel system.

For instance, ambient temperature and humidity affect air conditioning use, and thus hot running emissions, but also affect the magnitude of evaporative emissions. Clearly, the many factors that influence individual vehicle emissions make the relationships between road traffic, air pollution and greenhouse gas emissions quite complex.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

31

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Furthermore, impact assessment is a multidisciplinary exercise, which adds to the complexity of the subject. This is illustrated in Figure 1, which shows how road traffic exerts its effects through different multidisciplinary steps. The intensity and spatial distribution of human and economic activity within an area creates a demand for transport. How, where and to what extent this demand is met depends on the supply of transport facilities, i.e. the transport infrastructure. Demand and supply of transport are both reflected in, and are a function of land use. The interaction between supply and demand determines the level and distribution of transport activity in an area, which is the starting point for the modelling of road transport impacts on air quality. The magnitude of emissions and its spatial and temporal distribution are determined by fleet characteristics, traffic activity and the quality of traffic flow (traffic performance). Once vehicular emissions are released into the atmosphere, dispersion processes transport and dilute these emissions. In addition to dispersion, pollutants can also undergo physical and chemical transformation and deposition. Depending on location, these processes results in certain ambient concentration levels, the extent of which is a function of meteorological conditions, topographical characteristics and distance between source and receptor. The level of exposure to air pollutants depends on ambient concentration levels and where sensitive receptors (e.g. population) are situated in time and place. For instance, health effects depend on exposureeffect relationships, which may be obtained from epidemiological or clinical studies. The magnitude of the combined effects on health, structures and the environment subsequently determine the economic effects (costs) of air pollution.

Figure 1. Multidisciplinary Relationships Between Road Transport and its Effects.

This chapter will focus on the second and third step in the assessment of air quality and greenhouse gas impacts from road traffic, namely the generation of air pollutant emissions and greenhouse gases from road traffic. It is noted that in the remainder of this chapter, for Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

32

Robin Smit, Hussein Dia and Lidia Morawska

readability purposes, the term “emissions” is used as a generic term for emissions of air pollutants and greenhouse gases, the latter which is strongly correlated to fuel consumption.

Generic Structure of Emission Models Due to the complex relationships between road traffic and air pollution, models (Smit et al. 2008a) are commonly used to predict and evaluate impacts and determine solutions. As will be seen later, this occurs at different scales, ranging from local road projects (e.g. hot spot analysis, traffic management) to entire urban or regional transport networks and even national or global emission inventories. Before current emission models are discussed, this section discusses relevant terminology and some generic aspects of emission models: 1. 2. 3. 4.

empirical base of emission models; generic computation procedure; vehicle classification scheme; and modelling of driving behaviour.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Empirical Base of Emission Models Emission models are developed from emission measurements. There are different emission measurement methods available, namely laboratory engine bench testing, laboratory chassis dynamometer testing, on-board measurements, near-road measurements and tunnel studies. A review of emission models (Smit, 2006a) revealed that the majority of current emission models are based on laboratory emission testing using driving cycles1, although different driving cycles are used in different models. An advantage of laboratory measurements is that they are conducted under controlled conditions. This enables investigation of specific aspects that influence emissions such as driving pattern and ambient temperature. A disadvantage of this method is the limitation on the number of vehicles or engines that can be tested due to time and budget constraints. As a consequence, certain types of vehicles may not be adequately reflected in the test data. For instance, owners of high emitting vehicles2 tend not to register their vehicles, thereby reducing the chance for inclusion in a test programme. In addition, emissions from high-emitting vehicles are much more variable than emissions from normal emitters and thus require a large sampling fraction to obtain reasonably accurate emission estimates (Schulz et al., 2000). So, models based on laboratory test data are potentially biased and may significantly underestimate traffic emissions. This issue generally seems to receive more attention in the USA than in Europe, and only US models explicitly model high emitting vehicles. 1 2

A driving cycle is a speed-time profile that is assumed to be representative of (a mixture of) certain traffic and road conditions in a particular geographic area, and which has been synthesized from a set of driving patterns. These vehicles have very high emissions due to equipment malfunctioning (e.g. less-robust emission controls, neglect of maintenance), tampering or incorrect repairs (e.g. fifty times higher than normal emitters as reported by Barth et al., 2000). This appears to be a significant issue as several (remote sensing) studies (e.g. Zhang et al, 1995; Singh and Huber, 2000) have shown that on-road vehicle emissions are highly skewed, which effectively means that a small proportion of the fleet (20%) make up a large proportion of the total traffic emissions (60-80%).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

33

Laboratory vehicle exhaust emission testing may be conducted using tedlar sample bags (denoted as “bag measurement”) that are analysed after completion of the driving cycle, or may be conducted using continuous measurement at a high time resolution (typically 1-10 Hz) . As it is the method prescribed by emission legislation around the world, bag sampling has traditionally been the dominant approach. As a consequence, a large body of bag test data is available and these data have traditionally been used in the development of emission models. The second method of continuous measurements has become increasingly common for emission rate calculation and engine development. There are some additional aspects that specifically concern this second method such as correction for the time lag and mixing dynamics in the sampling and analysis system before measured emission values can be correctly correlated with driving conditions (Atjay and Weilenmann, 2004). Continuous test data is required for the development of emission models that operate at a high temporal and spatial resolution, as will be discussed later. Laboratory measurements were traditionally based on “standard” (vehicle) driving or (engine) test cycles (e.g. FTP, NEDC, ESC), but in time have included driving or test cycles that are believed to better reflect real-world driving conditions (“off-cycle”), and therefore emissions (e.g. CADC, AUC, ETC). It has been reported that emission factors based on the standard driving cycles for light-duty vehicles underestimate emissions in “real-world” driving by up to 50-60% (Joumard et al., 2000), or even higher (Watson, 1995). On the difference between standard (steady-state) engine tests and real-world driving, Rexeis et al. (2005) report errors for NOx emissions of heavy-duty vehicles between about -30% to +190%. Thus, use of standard cycles only in the development of emissions models may lead to biased emission models. In addition to dynamometer testing, researchers have used vehicles with on-board measurement systems to collect emissions and driving pattern data while they are driving on the road. Although research of this sort has provided valuable insights in real world vehicle emissions, it has traditionally not been used in the development of traffic emission models. This may have been due to the costs, the size and weight of these on-board systems, detection limits and possible data quality issues (Elst, Smokers and Koning, 2004). This, however, appears to be changing now and developments of improved new systems are being reported (North et al., 2005). Further advances in measurement technology such as portable on-board instruments (Frey, Unal and Chen, 2002) show increased use in general (El-Shawarby et al., 2005; Qiao et al., 2007) and increased use of on-board test data in emission models, particularly over the last few years (US EPA, 2002; Panis et al., 2006; North et al., 2006; Krishnamurthy et al., 2007; ISSRC, 2008). On-board testing is particularly interesting for developing nations as a relatively cost-effective way to generate real-world test data and calibrate models (ISSRC, 2008). However, measurement on a large number of vehicles can still be restricted by labour time and costs, particularly for older vehicles (North et al., 2005). Other methods such as remote sensing (e.g. Ekström et al., 2004), tunnel studies (Staehelin et al., 1998) and on-road or near-road modelling (e.g. Morawska et al., 2001) are commonly used for emission model validation purposes and have contributed to an increased understanding of model accuracy and real world emission behaviour of vehicles (e.g. high emitters). There are, however, some issues with these approaches that complicate its direct use in emission models such as a limited range of operating or traffic conditions.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

34

Robin Smit, Hussein Dia and Lidia Morawska

Generic Computation Procedure Three levels of analysis can be distinguished for prediction of vehicle emissions: • • •

vehicle; traffic stream (road link); and network.

Irrespective of the level of analysis, total emissions (E, unit mass) for moving vehicles are predicted by multiplication of an emission factor or emission rate with appropriate traffic activity data for each vehicle class. Emission factors quantify the amount of pollutant emitted and they are usually expressed as mass per unit distance (ex) at link and network level, although others such as mass per kg of fuel burned (ef) may be used instead. Emission rates are expressed as mass per unit time (et) at vehicle level, as will be seen later in this chapter. The corresponding traffic activity data follows from these units, i.e. vehicle kilometres travelled (VKT3), total fuel consumption or total time spent in particular driving conditions (e.g. idling). As will be discussed later, the spatial resolution for ex is typically restricted to a section of road (link) or an entire network area and et is typically used at higher spatial resolutions. For parked vehicles (with engines off), start and evaporative emissions are computed using emission factors that align with the required input data, i.e. mass per start (cold and hot start), mass per engine shutdown event (hot soak) and mass per hour of day (diurnal, resting loss).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Vehicle Classification We have seen before that vehicle design characteristics significantly impact on emissions and fuel consumption. A vehicle classification scheme is normally used in emission modelling to take differences in vehicle design characteristics into account. Given the large number of vehicle design characteristics, an almost infinite number of vehicle classes could be defined. However, the level of detail in vehicle classification should be comprehensive (i.e. include all important classes4) as well as practical (i.e. level of detail that matches available data on fleet or traffic composition). In addition, a vehicle classification scheme should be up-to-date. In practice, vehicle technology classes are usually defined by a taking into account a limited number of factors such as “main vehicle type” (e.g. passenger car, light-commercial vehicle, motorcycle, articulated trucks, buses, etc.), “fuel type” (e.g. diesel, petrol, LPG) and “emission standards”. However, more detailed vehicle classification schemes are sometimes used. For instance, the VERSIT+ model (Smit et al., 2007a) also considers vehicle weight, fuel injection technology, emission reduction technology and type of transmission and the 3 4

At road level, VKT is computed as the product of road length (km) and traffic volume (vehicles/hour). It is noted that in reality only a few specific vehicle classes dominate traffic emissions due to their relatively high emission levels and/or high usage. For instance, Smit (2006b) illustrated that about 70% of mean traffic NOx emissions are caused by only three main vehicle classes, i.e. diesel and petrol passenger cars and articulated trucks. This implies that a vehicle classification scheme should at least consider the most important classes.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

35

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

CMEM model (Barth et al., 2000) also uses power-to-weight ratio, mileage, model year, after-treatment technology and emitter type (normal, high emitter) as classification variables. The choice for classification variables generally appears to reflect an established but arbitrary decision process, although in some cases it may actually involve statistical analysis to group vehicles according to their emission characteristics (Fomunung et al., 2000; Rakha et al., 2004). With respect to practicality it is noted that the more detail that is used in the vehicle classification scheme, the larger are the demands that are imposed on the traffic input data. This is because a similar breakdown of VKTs by vehicle class is needed to be able to compute road traffic emissions. In practice, detailed information is, however, often not available. Traffic models, for instance, commonly generate traffic volume data for just two vehicle classes (“light”, “heavy”). In addition, there is a need to make future predictions on fleet composition. As a consequence, fleet composition models are needed to generate information on the proportion of total travel for each individual vehicle class in an emission model. Fleet composition models take into account vehicle sales, scrapage rates, growth rates and vehicle activity (e.g. mean annual VKT), which are a function of vehicle age, calendar year or (vehicle) model year (e.g. Keller and Kljun, 2007). These proportions or “weighting factors” are then used to compute composite emission factors or emission rates reflecting a less detailed vehicle classification scheme that matches the available traffic input data. As will be discussed later, emission models can be “incomplete” because they predict emissions for specific vehicle categories only (e.g. passenger cars) or because they are outdated (e.g. based on test data that do not reflect the latest developments in vehicle technology). Use of incomplete models introduces errors. It effectively restricts emission prediction to a specific part of a traffic stream, and additional models would be needed to estimate total traffic emissions. So, from a practical point of view, models with a comprehensive vehicle classification scheme are most useful for model users.

Modelling of Driving Behaviour (Model Classification) As was discussed before, the way in which a vehicle is being driven on the road is affected by many factors. Given the almost infinite number of combinations of driving conditions and vehicle characteristics (e.g. power to mass ratio, vehicle size), a large variety in driving behaviour is observed in the real world (Holmén and Niemeier, 1998; Smit, 2007b), which is very specific in terms of time and location (Brundell-Freij and Ericsson, 2005). Driving Behaviour is usually measured and quantified through a speed-time profile, which can be either a driving pattern5 or a driving cycle. An example of a driving pattern is provided in Figure 2.

5

A driving pattern is defined as a speed-time profile that has been recorded (measured) on the road or generated by a traffic model. In this chapter the difference between a driving pattern and a driving cycle is not relevant.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

36

Robin Smit, Hussein Dia and Lidia Morawska

Figure 2. Example of a Driving Pattern.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The characteristics of a particular speed-time profile are commonly described (i.e. quantified) by statistics called “cycle variables” such as average speed, number of stops per kilometre, mean acceleration, percentage idle time, mean acceleration power, etc. (e.g. André et al., 2006). It is noted that other relevant aspects of driving behaviour such as gear shift profiles exist as well, but they are not commonly available, as will be discussed later. In general, the level of model complexity, i.e the extent to which influencing factors are incorporated in an emission model, depend on the actual (type of) emission model that is considered. For a structured review of emission models, it is useful to consider a classification framework. In line with the classification used by Smit et al. (2008b), emission models are classified in terms of the way driving behaviour is incorporated in the model. The following categories of emission models can then be distinguished: • • •

models that have incorporated speed-time profiles in the development phase of the model (Type 1); models that generate speed-time profiles as part of the emission modelling process (Type 2); and models that require speed-time profiles data as input (Type 3).

These three main model types will be used and discussed in the next section. Typically, Type 2 and 3 models are well suited for analysing changes to emission levels at the local level (e.g. intersections, small networks), whereas Type 1 models are better suited to operate at a larger scale (e.g. urban regions, small areas) (Smit et al., 2007a).

Review of Current Traffic Emission and Fuel Consumption Models Possibly the earliest emission model was developed by Rose et al. (1965) who found that the empirical relationship between HC and CO emissions (pounds mile-1) and mean journey speed less than about 80 km/h was best satisfied by using a power function. Since then,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

37

considerable work has been conducted in the field of emission and fuel consumption modelling, resulting in a large number of models available to model users today. The development of emission and fuel consumption models has undergone significant changes over the last forty years or so. Not only have models become increasingly complex and detailed, as will be shown in this section, also the number of models has increased significantly. In time, specific types of emission models have become more comprehensive and detailed with respect to simulation of driving behaviour (type 3A) and the number of pollutants and vehicle classification (all types). The last factor is in response to an ongoing diversification in fuels (e.g. biofuels) and vehicle types (e.g hybrids, SCR trucks) in on-road fleets. As will be seen later, there is at this time no emission model that provides a detailed simulation of all aspects of traffic emissions at all scales, although some models are close to achieving this. This section reviews current emission and fuel consumption models. As models quickly become outdated due to ongoing changes in fleet composition, the focus of this chapter will be on models that are either under development or that are used in practise, regularly updated and actively maintained. Although it is not possible to discuss all specific models that are developed around the world, this chapter is comprehensive in that it discusses all relevant model types. As mentioned in the previous section, the model types follow from earlier work (Smit et al., 2008b).

Type 1 Models

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Driving cycles are fundamental building blocks in the development of Type 1 models. As was discussed before, they form the basis for emissions testing data, which is subsequently used to develop distance-based emission factors (g km-1) or fuel-based emission factors (g kg1 of fuel). Three subtypes can be distinguished: • • •

1A Model “aggregate model”: emission factors are constant, as is the case for areawide and fuel-based models. 1B model “average speed model”: emission factors are a function of one (continuous) driving pattern variable, i.e. average speed. 1C model “traffic situation model”: emission factors are defined in terms of discrete quantitative or qualitative descriptions of a particular traffic situation.

The Type 1A Model The Type 1A aggregate models use a top-down approach and are applied at low spatial (e.g. national, state, area) and temporal resolution (e.g. year). VKT-based models combine data on total VKT (derived from for example national statistics) and a single distance-based emission factor to compute total area emissions (e.g. AGO, 2003; Lyons et al., 2003; Gerard et al., 2007). Similarly, fuel-based models use data on total fuel consumption and fuel-based emission factors to compute total area emissions (e.g. Singer and Harley, 1996; IPCC, 1996; Pokharel, Bishop and Stedman, 2002; Davies et al., 2007; Guo et al., 2007).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

38

Robin Smit, Hussein Dia and Lidia Morawska

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The Type 1B Model The Type 1B model is commonly referred to as average speed emission model. Emission factors (g km-1) are a function of average speed, where average speed is defined as the overall speed on a section of road or for an entire journey. Average speed shows a significant correlation with emissions and fuel consumption (e.g. André and Hammarström, 2000). In addition, information on average speed is relatively easy to obtain as it can be sourced from traffic models or travel time surveys. This (at least partly) explains its common use in emission modelling for a long time (e.g. Kurtzweg, 1973; Evans, 1978), as will be seen later. Although several average speed models exist around the world (e.g. Singh and Huber, 2000; Reynolds and Broderick, 2000; Park et al., 2001; Pekol et al., 2003; Noland and Quddus, 2006), the primary Type 1B emission models currently in use are MOBILE (US EPA, 2008a) and EMFAC (CARB, 2008) in the United States and COPERT in Europe (LAT, 2008). These models are also used, possibly in a modified or augmented form, in other countries. For instance, COPERT is used in South-America (Corvalán et al., 2002) and MOBILE is used in Canada (Scott et al., 1997) and Asia (Hao et al., 2000; Mukherjee and Viswanathan, 2001). In addition, (part of) these models are sometimes incorporated as submodels in other models (Mellios et al., 2006) or (partly) used in the development of new models (QGEPA, 2002; ISSRC, 2008). One characteristic of average speed models is that they have become comprehensive in the number of pollutants, vehicle categories, influencing aspects and types of emissions that are covered (i.e. increased complexity). For instance, original average speed models (e.g. Evans, 1978) predicted emissions for a specific vehicle class (e.g. catalyst petrol car), emission type (e.g. hot running) and certain regulated pollutants (CO, HC, NOx) only. In contrast, COPERT IV now considers 218 vehicle classes including new technology vehicles (hybrids) and emerging fuels (biodiesel), 122 pollutants and greenhouse gases including emerging ones (e.g. active particulate surface area, particle number by size range), correction for several aspects (e.g. deterioration, fuel quality, road gradient, loading) and all types of emissions (hot running, start, evaporative) in its modelling process. However, treatment of driving behaviour effects remains somewhat simplistic in average speed models. To a certain extent, average speed models take driving dynamics into account as lower mean speeds are naturally the result of, for example, more speed fluctuation and idle time. But driving dynamics are not explicitly modelled and this may introduce substantial errors in the emission predictions at the local scale, as will be discussed later. For instance, Smit et al. (2007a) showed that NOx emissions from an average Euro 3 petrol car could vary between about – 80% to + 200% around the COPERT estimate for an average speed of 60 km/h, depending on the driving pattern. Others (Negrenti, 1999) reported a difference in emissions of up to a factor of 4 for the same mean speed with different levels of speed fluctuation. This implies that application of current travel speed models is valid as long as level of speed fluctuation does not substantially depart from the values inherently used by the model. However, information on the level of speed fluctuation for each mean speed in average speed models is not readily available from model documentation and requires additional driving cycle analysis (e.g. Smit et al., 2008b). Another issue is the common use of a single (mean) speed for all vehicles on a section of road. In reality a distribution of average speeds would apply to a traffic stream. Smit et al. (2008a) showed that this can potentially lead to substantial errors (up to 75%) in road link

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

39

emissions. This issue can be addressed by considering average speeds of individual vehicles, either by using a mean speed post-processing method, traffic field data or an appropriate traffic model.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The Type 1C Model The type 1C traffic situation models use discrete emission factors (g km-1) for certain “traffic situations”, which are defined either quantitatively (set of quantitative variables) or qualitatively (verbal descriptions). For instance, TNO (Veurman et al., 2002) developed a quantitative traffic situation model for passenger cars in freeway driving conditions. Emission factors are provided for nine congestion classes, which are defined in terms of (ranges of) traffic volume, space mean speed and speed limit. Smit et al. (2008a) extended this model and developed a discrete quantitative emission model for both urban and freeway conditions and for all major vehicle classes. The majority of traffic situation models are, however, qualitative models (e.g. Neylon and Collins, 1982; Anyon et al., 2000; Keller, 2004; Klein et al., 2006; INFRAS, 2007). Similar to type 1B models, type 1C models have become more complex in time. For instance, early models only used a few traffic situations, whereas others the latest models are quite comprehensive. For instance, Neylon and Collins (1982) and Carnovale et al. (1997) used only four traffic situations (freeway, arterial, residential and minor roads, congested), whereas the ARTEMIS model (Keller and Kljun, 2007) uses 280 traffic situations6. Traffic situation models require information on vehicle kilometres travelled and determination of which particular traffic situation apply to which specific road link(s). Application of qualitative traffic situation models may present some difficulties with respect to the last point, as the boundaries between traffic situations are in some cases not clearly established. For instance, the extent to which “stop-and-go” conditions apply to particular links in a network, or the decision that perhaps another traffic situation would be more suitable, is a matter of opinion of the model user. This is not an issue for quantitative traffic situation models because traffic situation boundaries are clearly defined.

Type 2 Models Type 2 “traffic variable” emission models generate (simplified) driving pattern data as a function of a number of traffic variables relating to traffic characteristics (e.g. traffic volume, average speed, traffic density, queue length) and road infrastructure characteristics (e.g. link length, number of lanes, free-flow speed, type of intersection, signal settings). These driving patterns are then combined with instantaneous (e.g. Matzoros, 1990; Coelho et al., 2005) or fundamental driving mode (Akçelik, 2006) emission rates (g s-1), possibly derived from other (Type 3) emission models, or are used to compute correction algorithms for incorporated emission factor tables (g km-1) derived from other (Type 1) emission models (e.g. Negrenti, 1999).

6

These are defined in terms of road type (motorway, trunk road, distributor, collector, etc.), area type (urban, rural), speed limit (30, 40, …, > 130 km h-1), and congestion level (“free-flow”, “heavy”, “saturated”, “stop and go”).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

40

Robin Smit, Hussein Dia and Lidia Morawska

Type 3 Models Type 3 models have been developed since the 1970s. Over recent years, international research has focused more strongly on the development of these emission models as it is generally believed that this will lead to more accurate emission assessments at local scale compared to more aggregate models. However, emission model prediction accuracy is an area that requires more research, as will be discussed later. Two main types of Type 3 models can be distinguished: •

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.



3A Model “cycle variable model”: emission factors or rates are a function of specific driving pattern variables at a high resolution (several seconds to several minutes). 3B model “modal model”: emission rates are a function of specific engine or vehicle operating modes at the highest resolution (typically one to several seconds).

As will be seen later, Type 3 emission models have generally developed from relatively simple models to more complex ones. This is not only the case with respect to their computation of driving behaviour effects on emissions, but for some particular models also with respect to other aspects. For instance, early modal models typically predicted hot running emissions of regulated pollutants from light-duty vehicles only (e.g. Kunselman et al. 1974; André and Pronello, 1997). Some recent modal models are still limited in a similar way (e.g. Krishnamurthy et al., 2007), but others are more comprehensive. For instance, CMEM now considers 31 vehicle classes (cars, trucks, petrol, diesel), correction for several aspects (deterioration, high emitters) and both hot running and start emissions in its modelling process. Although modal emission models have become substantially more comprehensive, they generally do not match current average speed and traffic situation models on certain aspects such as number of pollutants and comprehensiveness with respect to types of fuels and emissions included. An exception is the anticipated modal MOVES model (US EPA, 2008b), which will replace the average speed model MOBILE 6.2. MOVES will have the same functionality as MOBILE plus additional capabilities (energy consumption, GHGs, well-to-wheel). It will include many vehicle types (e.g. conventional, hybrids, electric) and fuels (e.g. conventional, E85, M85, electricity).

The Type 3A Model Type 3A “cycle variable” models require driving pattern data to quantify selected cycle variables and possibly other variables, which are then used to estimate emission factors or emission rates (expressed as g km-1 or g s-1) for different vehicle classes. The temporal resolution depends on the length of the driving patterns on which the model is based, which could be short driving pattern segments of several seconds, stop-go-stop segments or entire driving cycles. They could be described as average speed models with additional model variables to take speed fluctuation into account. Type 3A models are not new, although they have become more sophisticated in time. Following earlier work by Evans, Herman and Lam (1976), the PKE-average speed model was developed by Watson and co-workers (Watson et al., 1982). The model computes emissions as a function of average speed, idle time and “positive acceleration kinetic energy” or PKE. Recent examples of Type 3A models are MEASURE (Fomunung et al., 2000;

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

41

Hallmark et al., 2002), NEMO (Rexeis and Hausberger, 2005) and VERSIT+ (Smit et al., 2007a). MEASURE predicts instantaneous emission rates (g s-1) for different light-duty vehicle classes as a function of 114 cycle and vehicle-related variables (e.g. transmission, mileage), which depend on the pollutant considered. NEMO uses five cycle variables (average speed, maximum and minimum speed, average acceleration and deceleration) and vehicle specifications to compute the cycle average engine power, which is subsequently used to compute emission factors (g km-1) for different model classes. VERSIT+ computes emission factors (g km-1) using multivariate regression functions that incorporate several cycle and vehicle-related variables (e.g. mass) for different vehicle classes. It uses statistical optimisation software to select the best combination of a pool of fifty driving cycle variables and also predicts confidence intervals. The IVE model (ISSRC, 2008) uses instantaneous data on speed, acceleration and road grade to estimate vehicle specific power (VSP) for each second of driving as well as another variable called engine stress. Engine stress reflects mean power over the last 20 seconds of driving, and uses derived instantaneous engine speed in its computation. VSP and engine stress are used to correct for (binned) driving behaviour effects on average “base” emission levels. The input data requirements for Type 3A models are similar to Type 3B models. The advantage of Type 3A models is believed to be enhanced prediction accuracy as Type 3A models are typically based on substantially larger bodies of emission test data then Type 3B models, as will be shown later. This large body of test data is believed to better reflect the large variability in vehicle emissions.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The Type 3B Model Classical “driving mode” emission models have been around for a long time (e.g. Ludwig, Sandys and Moon, 1973; Benson, 1989; Taylor and Young, 1996). Here, emissions are not estimated for each second of driving, but rather for each “fundamental driving mode” (e.g. cruise, idle, acceleration, deceleration). The emissions module in the Type 2 SIDRA Intersection model (Akçelik, 2006) and other models (Zito and Taylor, 2001; Midenet et al., 2004) are current examples of such models. In the 1970s and later, instantaneous emission rates were computed using (continuous) empirical analytical functions that used instantaneous speed and acceleration as variables (e.g. Kunselman et al., 1974; Kent and Mudford, 1979; Cernuschi et al., 1995). More recent examples of this type of model are the VT-micro model (Rakha et al., 2004) and a model developed by Panis et al. (2006). A few decades ago modal models often consisted of discrete two-dimensional empirical lookup tables for emission rates (g s-1) with rows representing a “speed” interval and the columns representing an “acceleration interval7” (e.g. Kent et al., 1982; St. Denis and Winer, 1994; André and Pronello, 1997). From this work, a more theoretical approach emerged where instantaneous emissions are modelled as a function of the engine power needed to propel a vehicle (e.g. Post et al., 1985). Model variables include instantaneous speed and acceleration and possibly vehicle variables such as engine capacity, vehicle mass, road gradient, aerodynamic drag coefficient and frontal 7

Expressed as “acceleration” or “acceleration times speed”.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

42

Robin Smit, Hussein Dia and Lidia Morawska

area. Examples of current power-based are the Australian CVEM (Leung and Williams, 2000) and SIDRA TRIP (Açkelik and Besley, 2003). Another example is the MOVES model which uses instantaneous speed and vehicle specific power (VSP), the latter which is computed as a function of instantaneous speed, acceleration and road grade (Sonntag and Gao, 2007). VSP is also used in other models to estimate instantaneous emissions directly (e.g. Coelho et al., 2006). Another recent modelling approach involves the use of artificial neural network models. Dia and Boongrapue (2008) developed a neural network model to predict fuel consumption and emissions for individual vehicles using speed, acceleration, air fuel ratio and torque as explanatory variables. These workers compared the neural network model with simple/multiple/non-linear and hybrid regression techniques, based on the same laboratory test data, and demonstrated superior prediction accuracy of the neural networks over the other regression-based approaches tested. Smit and McBroom (2009a; 2009b) are currently developing a high-resolution traffic emission model that is based on (n-order) autoregressive multivariate regression functions for individual vehicles in a traffic stream. Model variables include traditional variables such as instantaneous speed, acceleration and power, but also newly developed variables that quantify the change in power and oscillation in either speed or power over a pre-defined period of time prior to the point in time for which the prediction is made. These variables aim to quantify and include “operating history effects” into the model. Preliminary results show that the new modelling approach delivers satisfactory results in terms of model accuracy, reliability and robustness (e.g. R2 ranging between 0.62-0.85 for NOx and 0.88-0.94 for CO2).A more complex approach that has emerged more recently computes instantaneous (engine-out or tailpipe) emission rates (g s-1) as a function of engine variables such as engine speed (rpm) and engine load (torque) or brake mean effective pressure and change in manifold pressure and possibly other variables such as injection timing, oil temperature, air-to-fuel ratio and operating mode (soak time, stoichiometric operation, enrichment, enleanment)8. These relationships can be incorporated in the model as (discrete) emission matrices (e.g. Koupal and German, 1995) or as continuous algorithms (e.g. Zalinger, Ahn and Hausberger, 2005). Additional modules are usually required to simulate gear shift behaviour and to compute instantaneous engine load and engine speed from input data such as vehicle parameters (e.g. gear ratios, wheel size), speed-time data, road gradient data and use of accessories (e.g. air conditioning)9. In addition, modules that simulate catalyst reduction efficiency may be included to account for the effects of emission control technology. Examples are the European PHEM (Hausberger et al. 2003a; Zalinger et al. 2005; 2008) and DIVEM (Atjay et al. 2005) models and the US CMEM model (Barth et al., 2000; Barth et al., 2004). A characteristic feature of modal models is that they require a substantial amount of input data, such as instantaneous (usually second-by-second) speed-time data and information on vehicle-specific parameters. The implications of this will be discussed later.

8

9

It is noted that complex models such as ADVISOR (Markel et al., 2002) exist that provide detailed simulations of predefined vehicle configurations (e.g. conventional, hybrid, fuel cell) in different driving conditions. However, these models are not used to estimate traffic emissions, but rather are vehicle system analysis tools used for vehicle system optimisation and impact assessment of changes to a particular vehicles aspects. There are also models that directly require engine-related variables as input (e.g. Krishnamurthy et al., 2007), making their application to estimate traffic emissions difficult.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

43

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Overview of Model Features Table 1 (next page) presents an overview of various modelling aspects for 35 different models that have been developed over time. The information has been gathered from available literature and from personal correspondence with model developers. A number of interesting observations can be made from table 1. First of all it is clear that the complexity of emission models has clearly increased in time, and this is the case for all model types. This increasing complexity is reflected in the increasing number of vehicles types, fuel types, pollutants and emission types considered in the models. For instance, emission models developed in the 1970s and 1980s generally predict hot running emissions of criteria pollutants from light-duty petrol vehicles only, using test data from standard driving cycles. In contrast, current models generally include many more vehicle types including heavy-duty vehicles and motorcycles, other fuels such as diesel, additional pollutants and fuel consumption, other types of emissions and are (at least partly) based on real-world driving behaviour. Similarly, the size of the test database has generally increased in time with models of the pre 1990s typically based on less than 100 vehicles, whereas more recent Type 1B, 1C and 3A models are typically based on thousands of test vehicles. This is graphically shown in the following chart where size of test database and year of publication are plotted for the different model types. The trend of increasing test databases is less evident for Type 3A models, which are still typically based on a few tens to a few hundred test vehicles. For the other model types the available data are too limited to reveal any trends. One interesting aspect of the fuelbased type 1A models is the very large test database for these models. These models are based on remote sensing data. This method is able to efficiently measure many thousands of vehicles at specific locations in the road network. A few other observations that emerge from table 1 are that average speed models include the largest number of pollutants and that average speed and traffic situation models typically include all emission types. In particular, evaporative emissions are often not included in Type 2 and 3 models, with a few exceptions.

Model Application Modelling Objectives Traffic emission models operate at different scales, depending on the modelling objective. Smit (2006a) distinguished three main modelling objectives: • • •

modelling to verify compliance with air quality standards; development of emission inventories; and evaluation of transport policies (scenario testing).

Assessment may be conducted at the local, regional, national or even global scale. For instance, it may involve local air quality impacts (e.g. Nagendra and Khare, 2002) due to new road projects or the implementation of (new) traffic management measures (e.g. lower speed limits, traffic signal coordination, metering signals).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009. ProQuest

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 1. Overview of Various Model Features Main Vehicle Types LCVs MCs Trucks Buses Petrol

Main Fuel Types Diesel LPG CNG

Pollutants Emission Type Hot Start Evap- Criteria NonSpec. Others Run orative (CO, HC, Exhaust HCs (GHGs, -ning NOx, PM NO2, PM) SO2, NH3, EC, etc.)

Code Number of Tested Vehicles (* )

Type of Driving Cycles

-

9 9

RW -

1 4 7 4 29

1 1

2 2 7 9 6

RW ST RW/ ST RW/ ST RW/ ST RW

15 5

1 11 4

1 1

3 2 8 5

ST ST RW RW RW

-

-

1

1 1

4 2

RW ST/ RW

Negrenti, 1999 A&A, 2008, p.c. R. Akcelik 26-8-2008

3 4 4 4 4

1 1 -

>1 5 5

1 3 6 2 5

1 1 -

1 9 8 6 8

RW/ ST RW/ ST RW/ ST RW/ ST RW/ ST

Watson et al, 1982 Fomunung et al, 1999; 2000; p.c. R. Guensler 22-8-2008/ S. Kimbrough 9-9-2008 p.c. M. Rexeis 28-7-2008 Smit et al, 2007; p.c. N. Ligterink 8-7-2008 ISSRC, 2008; p.c. J. Lents 23-8-2008

3 3 3 1 3 3 1 3 3 3 3 4 3 4 4

1 -

12 -

1 1 1 1 7 1 1 2

1 1 1 1 1 1 1 1 1

7 3 4 4 3 1 1 4 6 2 3 5 4 5

ST ST ST ST RW/ ST RW/ ST ST RW RW/ ST RW RW RW RW/ ST RW/ ST RW/ ST

Kunselman et al., 1974 Kent & Mudford, 1979 Post et al, 1985 Benson, 1989 Koupal & German, 1995 Cernushi et al., 1995 Taylor & Young, 1996 André & Pronello, 1996; 1997 Leung & Williams, 2000 Atjay and Weilenmann, 2004 Panis et al., 2006 Sonntag & Gao, 2007 p.c. G. Scora 5-8-2008 Rakha et al., 2004; p.c. Rakha 22-8-2008 p.c. M. Rexeis 28-7-2008

Model Type

Model Name

Number of Vehicle Classes

PCs, LDTs

1A 1A

Fuel Based Model Fuel Based Model

36 48

x x

x x

-

-

-

x x

-

-

-

-

x x

-

-

2 2

-

-

-

1B 1B 1B 1B 1B 1B

Average Speed Model Average Speed Model QGEPA MOBILE 6.2 EMFAC 2006 COPERT IV

1 1 73 28 13 218

x x x x x x

x x x x

x x x x

x x x x

x x x x

x x x x x x

x x x x

x x

x x x

x

x x x x x x

x x x x

x x x x

3 1 4 4 4 4

1 1

90 10 88

1C 1C 1C 1C 1C

Aggregate Modal Model Traffic Situation Model Traffic Situation Model ARTEMIS0.4D HBEFA 2.1

2 5 2 505 295

x x x x x

x x x

x x x

x x x

x x

x x x x x

x x x x

x x -

x -

x -

x x x x x

x x x

x x x

1 3 4 4 4

-

2 2

TEE-KCF SIDRA Intersection

44 2

x x

x x

x -

x x

x x

x x

x x

x -

x -

-

x x

x -

-

1 3

3A 3A 3A 3A 3A

PKE model MEASURE NEMO 1.7 VERSIT+ 3 IVE model

1 up to 120 Var 103 1372

x x x x x

x x x x

x x

x x x x

x x x x

x x x x x

x x x x

x x

x x

x

x x x x x

x x x x

x x

3B 3B 3B 3B 3B 3B 3B 3B 3B 3B 3B 3B 3B 3B 3B

Modal Analysis Model Modal Model Matrix Model CALINE 4 VEMISS Analytical Model Aggregate Modal Model DRIVE MODEM CVEM DIVEM Analytical Model MOVES CMEM 3.01 VT-Micro PHEM 08

11 6 2 1 1 4 5 3 5 13 31 (Var) 10 Var

x x x x x x x x x x x x x x x

x x x

-

x x x x x x

x x x

x x x x x x x x x x x x x x x

x x x x x x x x x x

x -

x x

-

x x x x x x x x x x x x x x x

x x x x x x x

x -

Others

FC

Reference (* * )

Singer & Harley, 1996 Singer & Harley, 2000 Rose et al, 1965 Evans, 1978 QGEPA, 2002 USEPA, 2008; Sonntag & Gao, 2007 p.c. B. Hancock 12-8-2008 LAT, 2008; p.c. L. Ntziachristos 20-8-2008 Ludwig et al. 1973 Neylon & Collins, 1982 Veurman et al,2002 Keller & Kljun, 2007; p.c. M. Keller 12-8-2008 Keller, 2004; INFRAS, 2007; p.c. M. Keller 12-8-2008

(*) Number of Vehicles Tested: “1” 0-10 vehicles, “2” 11-20 vehicles, “3” 21-50 vehicles, “4” 51-200 vehicles, “5” 201-500 vehicles, “6” 501-1000 vehicles, “7” 10012000 vehicles, “8” 2001-5000 vehicles, “9” > 5001 vehicles. (**) p.c. means personal correspondence.

Road Traffic Emission and Fuel Consumption Modelling

8 7 6 5 4 3 1

2

Size of Test Database (Code)

8 7 6 5 4 3 2 1

Size of Test Database (Code)

9

Type 1B

9

Type 1A

1965 1970 1975 1980 1985 1990 1995 2000 2005

1965 1970 1975 1980 1985 1990 1995 2000 2005

Type 1C

Type 2

8 7 6 5 4 3 1

2

Size of Test Database (Code)

8 7 6 5 4 3 2 1

Size of Test Database (Code)

9

Year of Publication

9

Year of Publication

1965 1970 1975 1980 1985 1990 1995 2000 2005

1965 1970 1975 1980 1985 1990 1995 2000 2005

Type 3A

Type 3B

8 7 6 5 4 3 1

2

Size of Test Database (Code)

8 7 6 5 4 3 2

Size of Test Database (Code)

9

Year of Publication

9

Year of Publication

1

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

45

1965 1970 1975 1980 1985 1990 1995 2000 2005 Year of Publication

1965 1970 1975 1980 1985 1990 1995 2000 2005 Year of Publication

Figure 3. Size of Test Database (Number of Vehicles Tested), Data taken from Table 1, Database Code: “1” = 0-10 vehicles, “2” = 11-20 vehicles, “3” = 21-50 vehicles, “4” = 51-200 vehicles, “5” = 201-500 vehicles, “6” = 501-1000 vehicles, “7” = 1001-2000 vehicles, “8” = 2001-5000 vehicles, “9” > 5001 vehicle.

Assessment of regional air quality are commonly based on emission inventories and may involve modelling of key pollutants such as NOx, SO2 and particulate matter (PM10 or PM2.5) (e.g. Owen et al., 2000), but is often directed towards the analysis and prediction of photochemical smog levels (e.g. Harley et al., 1993). National emission inventories are needed to verify compliance with international agreements (e.g. national emission ceilings, greenhouse gas emissions). The actual study objective determines the size of the study area (and thus size of the modelled road network, i.e. the number of links and nodes) that needs to be considered. For instance, modelling of photochemical smog formation requires consideration of a large

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

46

Robin Smit, Hussein Dia and Lidia Morawska

regional area (e.g. Helali and Hutchinson, 1994), whereas an air quality assessment in the direct vicinity of a new road requires consideration of only a small road network or possibly a single road (e.g. Meng and Niemeier, 1998).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Model Input The use of field data as input to emission modelling has a clear advantage in terms of accuracy when compared to modelled data1. Smit (2006a) distinguished four levels of availability with respect to input field data. Infrastructure-related input data are (in principle) available for every road in a network and include speed limit, type of road (e.g. arterial, freeway, residential), road length, type of intersection (e.g. signalised, roundabout) and signal settings. Certain types of traffic data such as traffic count data are commonly collected at a representative grid of points in the road network by urban traffic control systems. Others such as (average) speed2 and vehicle classification3 are less commonly measured and typically restricted to major roads. Finally, some types of traffic data, such as speed-time profiles and queue data are rarely measured, although the advent of GPS and video sensor and image processing technology is expected to greatly enhance data availability of these types of data, as will be discussed later. As modelling of emissions requires input data for preferably all relevant roads in the network, traffic models are usually needed to fill in the data gaps. In addition, traffic models are needed to make predictions for future situations. Two main types of traffic models may be distinguished, i.e. macroscopic and microscopic models. Macroscopic traffic models simulate the performance and behaviour of a traffic stream and typically generate output for road links such as traffic volume, mean speed and mean vehicle delay. Examples are, with increasing complexity, strategic planning models such as EMME2 and TRIPS (Brindle et al., 2000), dense network models such as SATURN (Van Vliet, 1982) and CONTRAM (Taylor, 2003) and traffic performance models such as SIDRA (Açkelik and Besley, 2003) and the Highway Capacity Manual (TRB, 2000). Microscopic traffic models simulate the movement of individual vehicles in space and time, and are thus able to generate detailed data on driving behaviour. Examples are AIMSUN, VISSIM and PARAMICS (EC, 2000). It is noted that input data for emission models are not restricted to actual output data from traffic models (e.g. link VKT, link travel speed), but may also concern input data to traffic models (e.g. link length, free-flow speed). From the perspective of emission modelling, traffic models can be regarded as a data source from which all relevant input data can be conveniently extracted.

1

It is noted that it may be quite acceptable (e.g. from the point of view of cost-effectiveness), depending on the modelling purposes (e.g. screening study), to use less accurate output from other sources such as traffic models. 2 (Mean) travel times, and thus (mean) travel speeds, can be measured on specific segments of road or entire routes using travel time studies. These studies are commonly conducted on major roads, for example, to measure the effectiveness of a transport system. Space mean speed data is directly measured by dual-loop detectors (i.e. two closely spaced induction loop detectors), which may be used on major roads (e.g. freeways). In case of single-loop detectors, space mean speed may be estimated from measured traffic volume and occupancy data. Speed data for short sections of road may be collected manually using e.g. radar guns or video analysis. 3 Basic vehicle classification data (e.g. light vehicle, heavy vehicle, perhaps a few heavy vehicle sub classes) is usually available for major roads. More comprehensive classification data is usually collected by manual classified counting surveys or video image surveys.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

47

As discussed in the previous section, the size of the study area (i.e. road network) is determined by the modelling objectives. However, the size of the modelled network affects the availability of emission model input data. The demand for resources to generate and process input data for emission models from either traffic models, field data or both, increases with network size. As a consequence, the extent and the level of detail of available input data are effectively reduced in practise when network size increases. On the other hand, the amount and types of required input data are a function of emission model complexity. More complex emission models impose a larger demand for input data on the model user. The trade-off between model complexity and network size leads to the following observation: in practise, complex traffic and emission models are usually applied at small networks and large networks require less complex traffic and emission models. The next table shows this and is based on a literature review of 55 air quality studies in which road traffic emission models were used. Table 2. Number of Traffic and Air Quality Studies by Traffic Input Data Source, Emission Model and Scale (N = National, R = Regional, L = Local) (Source: Smit, 2006a) Type of Input Data

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Field Data Macroscopic Model & Field Data Macroscopic Model Microscopic Traffic Model National Statistics Unspecified Model Total

1A. Aggregate Model

1B. Average Speed Model

1 (R)

4 (R) 2 (N) 5 (R) 15 (R) 2 (L)

Emission Model Type 3A. Cycle 2. Traffic 1C. Traffic Variable Variable Situation Model Model Model

8 (R) 3 (R) 1 (R)

3B. Modal Model

1 (L) 1 (R) 1 (L)

5 (L) 3 (L)

1

8

1 (R) 2

1 (R) 29

1 (R) 13

2

Total

15 10 22 5 1 2 55

The table shows that in particular average speed models are often used in air emission modelling (53%) at a regional scale, generally using input traffic data from macroscopic traffic models. Traffic situation models are also regularly used in air emission modelling (24%) at a regional scale and regularly use traffic field data as input. Type 3 emission models have been used in air emission modelling (16%) in combination with both macroscopic and microscopic traffic models, and they are always applied to small urban networks. Microscopic models are always applied at the local scale. Other emission model types such as the traffic variable models and the aggregate models have found some limited practical application in emission modelling (together 7%). It is noted that that these published studies do not necessarily accurately reflect actual usage of various emission model types in practise. For instance, it is possible that large scale air quality studies would tend to publish their results more than small scale local studies with relatively limited budgets, so the table may present biased results. Also there is an increasing interest in local effects of road traffic, as will be seen later, that will likely lead to higher application of Type 2 and 3 models4. 4

A current example of the detailed microscopic approach as applied to Intelligent Transport Systems (ITS) is provided by Dia and Gondwe (2008). The authors conducted a simulation study which aimed to quantify the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

48

Robin Smit, Hussein Dia and Lidia Morawska

The definitions used with respect to traffic data need to be considered carefully. A difference in definitions could lead to significant errors in emission predictions. An example is the use of “speed” for which five different definitions (space mean speed, time mean speed, running speed, travel speed, instantaneous speed) were identified by Smit (2006a). For instance, the correct speed definition for average speed emission models would correspond to “travel speed”, which is defined as the overall speed between two points of “sufficient” distance (i.e. distance corresponding to driving cycles used in emission model development), including all delays, for either an individual vehicle or a traffic stream. In contrast, use of space mean speed measured by dual loop detectors measures speed over only a short distance (a few meters) and would only be appropriate to use in average speed models if this speed is relatively constant over a sufficient distance (homogeneous traffic flow, e.g. on a section of freeway). To illustrate this point, Figure 4 shows that driving cycles on which commonly used average speed models are based, have a cycle length that varies between 130 m to 98 km and a median value of 6 km. Shorter driving cycles typically occur at lower more congested mean speeds, whereas longer cycles tend to occur at higher mean speeds. This has implications for the appropriate spatial resolution at which these average speed models should be applied. For instance, a driving cycle with an average speed of 70 kmph that represents a journey through a road network may involve driving on residential, arterial and freeways (as is shown in Figure 5) and would be expected to have significantly different driving characteristics, and thus emissions, than a driving pattern with the same mean speed for a 100 m stretch of arterial road that would represent free-flowing driving conditions. As a consequence, depending on the model, the correct spatial application of certain average speed models that are (partly) developed from journey-based driving cycles such as EMFAC and COPERT is likely to be at network level5. This is indeed suggested in model documentation (e.g. Ntziachristos and Samaras, 2000), but not always followed in practise where average speed models are applied at link level (e.g. Carslaw and Beevers, 2002).

5

impacts of incidents and evaluate the benefits of selected ITS and incident management strategies. The evaluation was based on the development of a large scale micro-simulation model covering an area approximately 122 kilometres squared, including 43 kilometres of Motorway and about 85 kilometres of arterial roads on the Gold Coast, Australia. This is a very large network for microscopic modelling. More than 10,000 vehicles were simulated during the AM and PM peak periods. The simulator collected second-bysecond data on each vehicle as they traversed the road network. A type 3B emissions model was interfaced to the simulator. The inputs to the emissions model comprised speed and acceleration and the output comprised fuel consumption and pollutant emissions. The results showed that the incidents resulted in an average increase of 1.5 percent in CO emissions and fuel consumption, and 5.0 percent increase in operating costs. The simulation approach and the interfacing of emissions models to data from the simulator was successful in quantifying the environmental impacts of incidents and evaluating the benefits of ITS management strategies terms of reduction in travel times, emissions and operating costs. It is noted that some average speed models such as MOBILE (6) are link-based and not journey-based. Here, driving cycles have been explictly developed to represent driving on certain types of roads and levels of congestion. Application of link-based models at link level is the main aim of these models.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

49

Figure 4. Variation in Cycle Length with Mean Speed for Three Average Speed Models.

Speed (km/h)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

150

100

50

0 0

250

500

750

1000

1250

1500

1750

2000

Elapsed Time (sec)

Figure 5. Example of a Journey Driving Cycle (UCC50) used in EMFAC 2000.

To complicate the issue further, cycle length may not be the only variable to examine correct spatial resolution for a particular emission model. A driving cycle may be long, but could reflect relatively homogeneous traffic conditions. For instance, it could represent low speed stop-and-go conditions or typical arterial or freeway driving. An example of typical arterial driving behaviour is given in Figure 6.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

50

Robin Smit, Hussein Dia and Lidia Morawska

Speed (km/h)

150

100

50

0 0

100

200

300

400

500

600

700

800

900

1000

Elapsed Time (sec)

Figure 6. Example of a Driving Cycle (UFLUI2) used in COPERT III.

It can be conceived that these cycles are a combination of driving behaviour samples from different vehicles on the same section of road. So in this case, it may be acceptable to apply the average speed model at road level. However, as Type 1B models are often developed from emissions test data based on various driving cycles, using statistical techniques (regression), it is difficult to arrive at a general recommendation on appropriate spatial resolution, so the issue remains unresolved and requires further research.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Spatial and Temporal Resolution Traffic flows in urban areas and on roads are by nature highly variable in time and space (Taylor et al., 2000). Because of this large variability, emissions from road traffic are highly variable as well and, in order to capture this variability, emission modelling requires a high temporal and spatial resolution. In practice, a time resolution of one hour is generally regarded as being sufficient because it aligns with air quality standards for which averaging times typically ranging from 1 hour to 1 year are of practical interest (Reynolds and Broderick, 2000)6. This time resolution also aligns with the time period in which highest congestion levels are commonly observed in road networks, i.e. peak hour or peak period. All emission models can be applied at this temporal resolution as long as emission model input data is provided at this resolution. It is noted that higher temporal resolutions may become more important for (emerging) pollutants with perhaps more relevant health impacts at peak exposures less than one hour such as PM2.5 (e.g. Greaves et al., 2008). The appropriate spatial resolution depends on the study objectives and scale of interest. For local level studies, and particular when near-road concentration levels are predicted using dispersion models, a high spatial resolution appears appropriate. An emission model should at least distinguish between traffic conditions that are relatively homogeneous such as stop-and6

It is noted that higher resolutions may be required for specific components, such as 15-minute averaged WHO standards for CO (WHO, 2002).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

51

go queuing and mid-block free-flow conditions. In this respect, Hallmark et al. (2002) observed that arterial road segments of about 60 m show relatively homogeneous driving behaviour. Higher resolutions, e.g. 10 m road segments, may seem even better, but prediction accuracy of emission models at this resolution may become questionable, as will be discussed shortly. For regional level studies, emissions data can be generated for area grid cells or for individual road links. Link level predictions would reflect a network segmentation that represents road environment conditions that are relatively constant, e.g. in terms of road capacity. For national level studies, national or perhaps large regional areas (States) would seem appropriate.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Prediction Accuracy To the knowledge of the authors, emission model prediction accuracy is not subject to specific criteria, e.g. as established by legislation or codes of practise. In practise, the required level of prediction accuracy would depend on the type of study. An emission model used for “screening” purposes would not be required to be accurate, but instead would be required to be conservative. For other "non-screening" applications, accurate emission predictions are important. For instance, accurate emission predictions would be necessary in cases with poor air quality close to guideline values (e.g. at critical locations such as a new residential area near a busy highway, hotspots) or in cases where policy measures are likely to cause relatively small impacts on emissions and fuel consumption (e.g. specific traffic management measures7). The choice for a specific model (type) clearly affects prediction accuracy and Figure 7 illustrates this. It shows composite emission factors predicted by a type 1A average speed model (COPERT IV, line) and a Type 3A cycle variable model (VERSIT+ 2B, dots representing specific driving cycles). It can be seen that for traffic situations with different dynamics but similar average speeds VERSIT + predicts different emission factors, whereas COPERT IV predicts the same emission factors for all these situations. Suppose now that the implementation of a particular local traffic management measure (e.g. improved signal timing) has smoothed the flow of traffic (i.e. reduced dynamics) and has increased the average speed from 20 to about 55 km h1 . For this particular situation, COPERT IV would always predict a decrease in emissions. In contrast, VERSIT+ could predict either a decrease or an increase in emissions, or even no effect, depending on input driving pattern data in both the reference and the new traffic situation. This effect is shown by the dashed arrows. Clearly, for policy makers and transport planners the correct direction and magnitude of these effects is vital information for imposing the right (i.e. effective and cost-effective) measures in order to improve on local air quality and reduce greenhouse gas emissions.

7

For instance, Midenet et al. (2004) report a CO2 emission reduction of 4% due to an improved signal control strategy.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

52

Robin Smit, Hussein Dia and Lidia Morawska

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 7. NOx Emission Factors for an Average Euro 3 Petrol Car as Predicted by COPERT IV and VERSIT+ (Source: Smit, 2008).

Type 2 and 3 emission models are able to discriminate between various traffic situations with similar average speeds but different dynamics and are appropriate for applications where the response of emissions to actual traffic conditions is an issue. It is noted that certain Type 1C models such as the ARTEMIS model with many traffic situations would also, to a certain extent, be capable to distinguish between these situations. However, other Type 1 models such as the average speed models do not account for different levels of speed fluctuation at a particular average speed , and the model would usually only provide accurate predictions for large (urban) road networks, or perhaps substantially large parts of a road network (e.g. 1 km2 grid cells). On first sight it appears that application of the most detailed and complex emission model (Type 3B modal) would lead to the most accurate prediction of traffic emissions. However, the sensitivity towards many factors would also create additional uncertainty because more (types of) input data are needed, each with its own inherent uncertainty. So, a point of discussion is how accurate a particular emission model can be in practice. There are several aspects that influence this, namely: • • • •

inherent variability in vehicle emissions, input data availability (modelling assumptions), input data quality, and model development aspects.

These aspects are discussed below. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

53

Inherent Variability in Vehicle Emissions

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Due to increasingly complex engine (management) and emission control technology, emissions of modern vehicles exhibit increasing (inherent) variability. Typically, modern spark-ignition vehicles show low base line emission rates (g s-1), with various short-duration emission peaks due to vehicle operation outside the “window” of optimum emission control. An example of this erratic emissions behaviour is shown in Figure 8 for a three-way catalyst equipped passenger car. Emissions from diesel vehicles have (traditionally) been modelled quite accurately using existing modal models (e.g. Hausberger et al. 2003), but emission controls in diesel vehicles are now also becoming more sophisticated (e.g. SCR, NOx storage catalysts) in response to stricter emission standards. As a consequence, accurate modelling of diesel vehicle emissions is likely to become more challenging. In addition, there is substantial variability among road vehicles, even if they are of similar type, make and model (e.g. Bishop et al., 1996). This is due to vehicle-specific engine management (e.g. fuel injection strategy), emission control technology (e.g. size of the catalyst, catalyst material) and other factors (e.g. ageing effects, presence of malfunctioning equipment). This large variance in emissions has implications for prediction accuracy. De Haan and Keller (2000), for instance, found it impossible to construct a speed-acceleration modal emission model that could accurately simulate this irregular emissions behaviour. Also, it is not clear if it is actually possible to accurately predict emissions at very short time intervals, and there are indications that this might not be the case (e.g. Hickman et al., 1999; Barth et al., 2000).

Figure 8. Example of Measured Second-by-Second Speed (Urban Driving Cycle) and NOx Emission Rates for a Euro 3 Petrol Car.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

54

Robin Smit, Hussein Dia and Lidia Morawska

Finally, inter-vehicle variability is further enhanced by differences in driving style. There is evidence that personal driving style (intensity of vehicle operation, gear shift behaviour) is quite important with respect to emissions. For instance it has been reported that its effect on vehicle emissions is larger than e.g. the frequency of time spent in a particular driving mode, which primarily reflects traffic conditions (Holmén and Niemeier, 1998). In order to adequately reflect the large inter-vehicle variability in emissions in traffic streams, it appears that emission models need to be based on a large body of test data for as many (representative) vehicles as possible. As was shown in Table 1, there is a positive trend towards inclusion of large bodies of test data in models and current Type 1B and 3A models are based on relatively large test databases in terms of the number of vehicles when compared to the most complex Type 3B models.

Input Data Availability (Modelling Assumptions)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Another important issue is the trade-off between model accuracy and input accuracy. As was discussed, availability of detailed input data decreases with network size. If a complex emission model were run, requiring more detailed input than were available, simplifying assumptions would need to be made, leading to reduced accuracy. For instance, recent modal models simulate gear shift behaviour as part of the modelling process or offer the possibility to use this information as input. However, location-specific data on gear shift behaviour in traffic streams are rarely (if ever) available. This means that assumptions need to be made in order to run the model, introducing unknown errors to model predictions, possibly offsetting accuracy gains over less complex models. So, there may be an optimum level of modelling detail for a certain application. This hypothetical curve is shown in Figure 9.

Figure 9. Hypothetical Relationship Between Input Accuracy, Model Accuracy and Level of Modelling Detail. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Road Traffic Emission and Fuel Consumption Modelling

55

It shows that level of (overall) input accuracy (due to availability) decreases with level of modelling detail (model complexity), whereas (potential) model accuracy increases with model complexity. Prediction accuracy is a function of both input and model accuracy, and a cost-effective optimum level occurs where both curves cross. Beyond this point (more complexity), an increase in prediction accuracy is either small or does not exist. Even a small increase in prediction accuracy is not interesting, as the costs to run the model (data collection, computer runtime) will increase disproportionately.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Input Data Quality In addition to data availability, the quality of the input data needs to be considered because errors in input data propagate through the emission modelling process. Similar to emission models, more complex traffic models are not necessarily generating more accurate input data. Each (type of) traffic model would have its own accuracy issues. Examples are the accuracy of predicted link speeds by macroscopic strategic planning models (Grant et al., 2000) and the validity of predicted vehicle trajectories by microscopic simulation models (Hallmark and Guensler, 1999). More research (e.g. model validation, sensitivity analysis, model comparison) is needed to establish which traffic-emission model combinations can achieve which levels of accuracy (in a relative and absolute sense) in field applications at various scales. Efforts to improve traffic emission modelling should focus on variables that have shown to have a large effect on emission predictions. In this respect, Smit (2006a) showed that the amount of travel (VKT) is a particularly important input variable. This is because errors in traffic volume will cause proportional errors in emission estimates, which may be amplified when other traffic variables such as average speed are derived from traffic volumes, as is the case in traffic models that use congestion functions (Smit, 2008d). For instance, an error in traffic volumes of 10-20% may result in errors in emissions of 50% (Reynolds and Broderick, 2000). As a consequence, particular attention should be directed at obtaining accurate information on traffic volumes. A sensitivity analysis using two average speed models (Smit, 2008d) showed large errors in NOx predictions (up to a factor of about 3.5) due to (in order of importance) variation in traffic composition, average speed and model choice, the actual magnitude depending on the level of congestion. The external errors due to input data quality were of the same order of magnitude as internal errors that have been reported from (partial) road validation studies. This implies that in terms of further improvements of traffic emission modelling, focus should be on both the quality of input data (application) and the quality of the actual emission models (model development).

Model Development Aspects As was discussed before, there are several other aspects of emission models that affect model accuracy such as the use of standard and/or real world cycles, comprehensiveness of the model in terms of vehicle classification, the use of up-to-date emissions test data and other factors such as the use of local emissions test data. In this respect, Table 1 showed that less

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

56

Robin Smit, Hussein Dia and Lidia Morawska

complex models such as average speed models may actually be more accurate than more complex emission models, because they use the largest empirical database, use real-world cycles and are regularly updated.

Summary It is clear that determining whether one emission model is more accurate than another is not an easy task, and would require information from different validation studies (tunnel studies, on-road, etc.). To date, model validation studies that involve several types of emission model are limited and provide inconclusive evidence that certain model types provide substantially better results over others (Smit, 2006a). For instance, it has been found (Fomunung et al., 2000; Park et al., 2001; Lacour et al., 2001) that, depending on the pollutant, average speed models sometimes perform better and sometimes perform worse than more complex (Type 3) models in terms of prediction accuracy. Even if a particular study shows that one model performs significantly better than another (e.g. EC, 1995), the results are necessarily based on assumptions that are known to strongly influence emissions (e.g. percent cold vehicles), casting some doubt on the claims that are made. Finally, application of complex emission models such as CMEM (Rakha et al., 2004) or TEE (Smit, 2006a) has revealed peculiar prediction behaviour in certain traffic conditions, which indicates the need for further emissions testing, fine-tuning and model testing of complex emission models to ensure robust and valid model predictions in all situations. In summary, prediction accuracy of emission predictions is an important area that will require more research.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Outlook It is expected that demand for detailed and comprehensive emission models will increase in the future. This is due to a number of developments. For example, there is a growing application of traffic control to improve traffic flow and reduce accidents. This is combined with an increasing interest around the world in the effects of local scale traffic measures on traffic emissions, air pollution and fuel consumption (e.g. Coelho et al., 2005; Akçelik, 2006). Similarly ongoing developments of in-vehicle technology will, when applied in practise, affect driving behaviour and hence emissions. The extent to which these technologies affect emissions and fuel consumption would be of interest to policy makers and perhaps the general public itself. An example is the ISA system (Intelligent Speed Adaptation system) investigated by Panis et al. (2006), which regulates (caps) vehicle speed by comparing GPS derived real-time speeds with the local speed limits, thereby reducing the number of accidents. There is also an increasing interest in the effects of road traffic on greenhouse gases and air pollutants other than the traditional criteria pollutants (CO, HC, NOx, PMmass). An example is the increasing importance of different ways to quantify particulate emissions such as particle number and particle area. Several issues with accurate modelling of particulate emissions (mass, number, etc.) at a high resolution in time and space are currently being addressed (Zalinger et al., 2008). Probably one of the most important aspects that may improve current emission modelling is the development of input data. Ongoing developments in sensor, communications and

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

57

computing technology create exciting opportunities to substantially improve both traffic emission model input data quality (e.g. field data) and to enhance availability (e.g. high temporal and spatial resolution). There is now an increasing use of automated data collection systems such as GPS (Global Positioning System) and video imaging technology, which are cost-effective ways of collecting large bodies of (real-time) data on vehicle behaviour in time and space (Midenet et al., 2004; Nesami and Subramanian, 2005; Brundell-Freij and Ericsson, 2005; Ahn and Rakha, 2007). At this stage, certain issues such as differences between GPS receivers and improving positional accuracy and spatial coverage (missing data due to e.g. buildings) need to be resolved before these data can be used to generate reliable highresolution (1 Hz) vehicle speed and acceleration data (Jackson and Aultman-Hall, 2007; Toledo et al., 2007). One good example of these developments is the MESSAGE project (Hoose et al., 2008). Here, a real-time computing environment for high resolution (30 m, 1 minute) road networks has been developed and tested that extracts information from various sources using wireless communication such as mobile air quality sensors (e.g. attached to vehicles with positioning devices), fixed pollution monitoring networks, field data on traffic control and weather information and traffic, emission and dispersion models. Improvements in the amount and quality of traffic data will certainly lead to more accurate emission predictions. The developments with respect to input data will also facilitate application of more complex emission models leading to a finer spatial and temporal allocation of traffic emissions. It is therefore expected that application of more complex emission models, such as the Type 3 modal and traffic variable models, will increase in time. Whether this will actually lead to (substantial) improvements in emission prediction accuracy, compared to less complex emission models, is, however, a matter that requires further investigation – and this is an important but overlooked issue in the opinion of the authors. Relevant research questions in this respect include how accurate a particular (type of) emission model should be and can be in practise, and under which circumstances. One advantage of increased use of complex models is that they can be used to develop a modelling framework (bottom-up), including separate models of various levels of complexity, that provides consistent results at all modelling scales. A modelling framework is necessary to choose the appropriate modelling approach in terms of modelling objectives (scale), study type (screening, other) and other factors (costs). This advantage is acknowledged in the development of the MOVES model in the USA (Sonntag and Gao, 2007), which is intended to predict traffic emissions at local, regional and national scale. Given the large inter-vehicle variability in emissions, emission models should ideally be based on a large number of emission tests. As laboratory testing is expensive, current emission models are necessarily based on a relatively limited number of tests. With ongoing developments in the field of and application of on-board emission measurements, there are opportunities to create large and (complementary) databases in a cost-effective way that can then be used for emission model development. Due to ongoing diversification of the on-road vehicle fleet in terms of fuel and technology types, there will be a strong need to include these new types of vehicles into emission testing programs to allow for subsequent inclusion in model updates. As a consequence, it is expected that emission models will become increasingly complex and comprehensive in their modelling of vehicle types, fuel types, pollutants and greenhouse gases.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

58

Robin Smit, Hussein Dia and Lidia Morawska

As was shown in this chapter, there are several issues with respect to the interaction between traffic models and emission models. Examples are issues with respect to spatial and temporal resolution, definition of model variables and balancing of prediction accuracy of both types of models. It seems a good way forward to further integrate specific traffic and emission models taking those issues into account. This may involve development of new emission models or modification of existing emission models that would match particular traffic models. It is expected that emission models will be applied to address increasingly complex research questions and study objectives. A recent example of this is the investigation of the effects of changes to current road network configuration (e.g. longer merge sections, signal timing) and increases in road capacity on traffic emissions and fuel consumption, taking into account feedback mechanism such as induced traffic generation (Noland and Quddus, 2006). Complex type 3 emission models in combination with either microscopic traffic models or high resolution field traffic data in these cases seem the best way forward as they imply to increase the probability of correct outcomes, in both a relative and positive sense. However, the outstanding issues with respect prediction accuracy will require further research in order to make this probability a reality.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Conclusion This chapter investigated current models designed to predict air pollutant emissions and fuel consumption for road traffic. Given the many factors that influence emissions and fuel consumption from individual vehicles, modelling of these emissions for many vehicles in road networks (traffic streams) is an involved, multidisciplinary and complex exercise. A review of current road traffic emission modelling around the world revealed that these models share various common features such as empirical base, generic computation procedure, vehicle classification and modelling of driving behaviour. Emission models were then classified and discussed in terms of the way driving behaviour is incorporated in the model and 6 (sub)types of models were identified. When development of emission models is examined over time, a few trends emerge. Emission models have become increasingly complex (both more detailed and more comprehensive) with respect to various modelling aspects such as number of pollutants, type of emissions, type of vehicles and modelling of driving behaviour. Also, the number of test vehicles on which models are based have generally increased in time, which provides more confidence in predictions given the large inter-vehicle variability in emissions behaviour. The most commonly used emission model is the average speed model. Although current average speed models are generally comprehensive, simulation of driving behaviour is somewhat simplistic. Therefore, these models are generally not suitable for local area impact assessment, which is a scale that attracts increasing interest around the world, and should be used at network level. However, the appropriate spatial resolution of a particular average speed model will depend on the driving cycles on which it is based, which is a subject that requires further examination. It is (logically) assumed that emission models that can discriminate between various traffic situations with similar average speed can be used for local scale studies, and they include (comprehensive) traffic situation models, traffic variable models, cycle variable models and modal models.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

59

There may also be the perception that a more detailed and complex model will lead to more accurate predictions. There are, however, many factors to influence prediction accuracy such as type and source of emissions test data on which a model is based, size of the emissions database, input data availability and input data quality. With respect to available (partial) validation studies, there is inconclusive evidence at this stage that certain models perform better than others. As a consequence, model prediction accuracy is an area that requires further research. It is envisaged that that emission models will be applied to address increasingly complex research questions and study objectives. As a consequence, it is expected that demand for detailed and comprehensive emission (type 3) models will increase in the future, leading to a finer spatial and temporal allocation of traffic emissions. This expectation is based on a number of ongoing developments that facilitate the development and application of high resolution emission models. Firstly, there is a growing application of (high-tech) traffic control measures and ongoing developments of in-vehicle technology. These developments create new opportunities to substantially improve both the quality and to enhance availability of emission model input data. Similarly, with ongoing developments in the field of and application of high-resolution on-board emission measurements, there are also opportunities to create large and (complementary) databases in a cost-effective way that can then be used for emission model development. Secondly, there is increasing interest around the world in the effects of local scale traffic measures on traffic emissions, air pollution and fuel consumption. Whether these developments will actually lead to (substantial) improvements in emission prediction accuracy, compared to less complex emission models, is, however, a matter that requires further investigation. Relevant research questions in this respect include how accurate a particular (type of) emission model should be and can be in practise, and under which circumstances. Due to ongoing diversification of the on-road vehicle fleet, there will be a strong need to include new types of vehicles and fuels into emission testing programs to allow for subsequent inclusion in model updates. As a consequence, it is expected that emission models will become increasingly complex and comprehensive in their modelling of vehicle types, fuel types, (number of) pollutants and greenhouse gases. As was shown in this chapter, there are some issues with respect to the interaction between traffic models and emission models. Examples are correct spatial and temporal resolution, definition of model variables and balancing of prediction accuracy of both types of models. It seems a good way forward to further integrate specific traffic and emission models taking those issues into account. This may involve development of new emission models or modification of existing emission models that would match particular traffic models.

References [1] [2]

Australian Greenhouse Office, National Greenhouse Gas Inventory Committee. Australian Methodology for the Estimation of Greenhouse Gas Emissions and Sinks 2003 – Energy (Transport). 2003. ISBN 1 928040 16 8. Akçelik, R. and Besley, M. Operating cost, fuel consumption and emission models in aaSIDRA and aaMOTION. Proceedings of the 25th Conference of Australian Institutes of Transport Research. Adelaide, Australia. 2003, 1-15.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

60 [3] [4] [5] [6] [7] [8] [9] [10] [11] [12]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[13] [14] [15] [16] [17] [18] [19] [20]

Robin Smit, Hussein Dia and Lidia Morawska Akçelik, R. Operating cost, fuel consumption and pollutant emission savings at a roundabout with metering signals. Proceedings of the 22nd Australian Road Research Board Conference. Canberra, Australia. 2006. Akçelik and Associates. 2008. Sidra Solutions. http://www.sidrasolutions.com/ default.html. André, M. and Pronello, C. Speed and acceleration impact on pollutant emissions. SAE Technical Paper Series. 1996, 961113, 1-8. André‚ M. and Pronello, C. Relative influence of acceleration and speed on emissions under actual driving conditions. International Journal of Vehicle Design. 1997, 18, 3, 340-353. André, M. and Hammarström, U. Driving speeds in Europe for pollutant emission estimation. Transportation Research Part D. 2000, 5, 321-335. André, M.; Joumard, R.; Vidon, R.; Tassel, P. and Perret, P. Real-world European driving cycles for measuring pollutant emissions from high- and low-powered cars. Atmospheric Environment. 2006. 40, 5944-5953. Ahn, K. and Rakha, H. Energy and environmental impacts of route choice decisions. Transportation Research Board Annual Meeting. 2007, 1-19. Anyon, P.; Brown, S.; Pattison, D.; Beville-Anderson, J.; Walls, G. and Mowle, M. InService Emissions Performance - Phase 2 Vehicle Testing. National Environment Protection Council Service Corporation, Australia. 2000. ISBN 0 642 323 348. Atjay, D.E. and Weilenmann, M. Static and dynamic instantaneous emission modelling. International Journal of Environment and Pollution. 2004, 22, 3, 226-239. Atjay, D.E.; Weilenmann, M. and Soltic, P. Towards accurate instantaneous emission models. Atmospheric Environment. 2005, 39, 2443-2449. Barth, M.; An, F.; Younglove, T.; Scora, R.; Levine, C.; Ross, M. and Wenzel, T. Development of a Comprehensive Modal Emissions Model, National Corporate Highway Research Project. Transportation Research Board. April, 2000. Barth, M. and Scora, R.; Modal emissions for heavy-duty diesel vehicles. Transportation Research Record. 2004, 1880, 10-20. Bishop, G.A.; Stedman, D.H. and Ashbaugh, L. Motor vehicle emissions variability. Journal of the Air and Waste Management Association. 1996, Vol. 46, pp. 667-675. Brindle, R.; Ashley, D.; Johnston, D.; Milthorpe, F.; Smith, N.; Taylor, M.A.P. and Young, W. Improving Urban Transport Demand Models and Their Use. Sydney, Australia: Austroads. Report AP-R164/00. ISBN 0 85588 551 3. Brundell-Freij, K. and Ericsson, E. Influence of street characteristics, driver category and car performance on urban driving patterns. Transportation Research Part D. 2005, 10, 213-229. CARB. 2008. EMFAC 2007. http://www.arb.ca.gov/msei/onroad/latest_version.htm Carnovale, F.; Tilly, K.; Stuart, A.; Carvalho, C.; Summers, M. and Eriksen, P. Metropolitan Air Quality Study - Air Emissions Inventory. NSW Environment Protection Authority and Coffey. Sydney, Australia. 1997. Carslaw, D.C. and Beevers, S.D.; The efficacy of low emission zones in central London as a means of reducing nitrogen dioxide concentrations. Transportation Research Part D. 2002, 7, 49–64.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

61

[21] Cernuschi, S.; Giugliano, M.; Cemin, A. and Giovannini, I. Modal analysis of vehicle emission factors. Science of the Total Environment, Elsevier Science BV. 1995, 169, 175-183. [22] Corvalán, R.M.; Osses, M. and Urrutia, C.M. Hot emission model for mobile sources: Application to the metropolitan region of the city of Santiago, Chile. Journal of the Air and Waste Management Association. 2002, 52, 2, 167-174. [23] Coelho, M.C.; Farias, T.L. and Rouphail, N.M. Impact of speed control traffic signals on pollutant emissions. Transportation Research Part D. 2005, 10, 323-340. [24] Coelho, M.C.; Farias, T.L. and Rouphail, N.M. Effect of roundabout operations on pollutant emissions. Transportation Research Part D. 2006, 11, 333-343. [25] Davies, J.; Grant, M.; Venezia, J. and Aamidor, J. U.S. Transportation sector greenhouse gas emissions: trends, uncertainties and methodological improvements. Transportation Research Board Annual Meeting. 2007, 1-17. [26] De Haan, P. and Keller, M. Emission factors for passenger cars: application of instantaneous emission modelling. Atmospheric Environment. 2000, 34, 4629-4638. [27] Dia, H. and Boongrapue, N. Development and evaluation of neural network emission models using the Second National In-Service Emission (NISE) Data. Department of Civil Engineering - The University of Queensland. 2008. Internal Report. [28] Dia, H. and Gondwe. Evaluation of Incident Impacts on Integrated Motorway and Arterial Networks using Traffic Simulation. 31st Australasian Transport Research Forum. Gold Coast, Australia. October 2008. [29] EC. ESTEEM - European Scenarios on Transport-Energy-Environment for Metropolitan Areas - Final Report. 1995. The European Commission. Contract No. JOE3-CT95-0016. [30] EC. SMARTEST – Programme of the 4th Framework Programme - Final Report. The European Commission. 2000. Contract No. RO-97-SC.1059. [31] Ekström, M.; Sjödin, Å. and Andreasson, K. Evaluation of the COPERT III emission model with on-road optical remote sensing measurements. Atmospheric Environment. 2004, 38, 6631-6641. [32] El-Shawarby, I.; Ahn, K. and Rakha, H. Comparative field evaluation of vehicle cruise speed and acceleration level impacts on hot stabilized emissions. Transportation Research Part D. 2005, 10, 13-30. [33] Elst, D.A.M.M.; Smokers, R.T.M. and De Koning, J.W. Evaluation of the Capabilities of On-Board Emission Measurement System for the Purpose of Generating Real-Life Emission Factors. TNO Report. 2004, Report 04.OR.VM.039.1/DE. [34] Evans, L. Urban traffic, fuel economy and emissions – consistency of various measurements. SAE Technical Paper Series. 1978, 780934, 1-11. [35] Evans, L.; Herman, R. and Lam, T. Multivariate analysis of traffic factors related to fuel consumption in urban driving. Transportation Science. 1976, 10, 205-215. [36] Fenger, J. Urban air quality. Atmospheric Environment. 1999, 33, 4877-4900. [37] Benson, P.E. CALINE 4 - A Dispersion Model for Predicting Air Pollutant Concentrations Near Roadways, FHWA. California Department of Transportation. Sacramento, California. 1989, Report No. FHWA/CA/TL-84/15. [38] Fomunung, I.; Washington, S. and Guensler, R. A statistical model for estimating oxides of nitrogen emissions from light-duty vehicles. Transportation Research Part D. 1999, 4, 5, 333-352.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

62

Robin Smit, Hussein Dia and Lidia Morawska

[39] Fomunung, I.; Washington, S.; Guensler, R. and Bachman, W. Validation of the MEASURE automobile emissions model: a statistical analysis. Journal of Transportation and Statistics. September, 2000, 65-84. [40] Frey, H.C.; Unal, A. and Chen, J. Recommended Strategy for On-Board Emission Data Analysis and Collection for the New Generation Model. Computational Laboratory for Energy, Air and Risk, North Carolina State University. US EPA, Office of Transportation and Air Quality, February 2002. [41] Gerard, D.; Fischbeck, P. and Rice, D. An interactive tool for the impacts of new engine technologies on fuel consumption and emissions. Transportation Research Board Annual Meeting. 2007, 07-2967, 1-16. [42] Grant, C.; Gillis, B. and Guensler, R. Collection of vehicle activity data by video detection for use in transportation planning. ITS Journal. 2000, 5, 4, 346-361. [43] Greaves, S.; Issarayangyun, T. and Liu, Q. Exploring variability in pedestrian exposure to fine particulates (PM2.5) along a busy road. Atmospheric Environment. 2008, 42, 1665-1676. [44] Guo, H.; Zhang, Q. and Wang, D. On-road sensing measurements and fuel-based motor vehicle emission inventory in Hangzhou, China. Atmospheric Environment. 2007, 41, 3095-3107. [45] Hallmark, S.L. and Guensler, R. Comparison of speed/acceleration profiles from field data with NETSIM output for modal air quality analysis of signalised intersections. Transportation Research Record. 1999, 1664, 40-46. [46] Hallmark, S.L.; Guensler, R. and Fomunung, I. Characterizing on-road variables that affect passenger vehicle modal operation. Transportation Research Part D. 2002, 7, 8198. [47] Hao, J.; He, D.; Wu, Y.; Fu, L. and He, K. A study of the emissions and concentrations distribution of vehicular pollutants in the urban area of Beijing. Atmospheric Environment. 2000, 34, 453-465. [48] Harley, R.A.; Russell, A.G.; McRae, G.J.; Cass, G.R. and Seinfeld, J.H. Photochemical smog modelling of the Southern California Air Quality Study. Environmental Science and Technology. 1993, 27, 378-388. [49] Hausberger, S.; Rodler, J.; Sturm, P. and Rexeis, M. Emission factors for heavy-duty vehicles and validation by tunnel measurements. Atmospheric Environment. 2003, 37, 5237-5245. [50] Helali, K. and Hutchinson, B. Improving Road link speed estimates for air quality models. Transportation Research Record Part D. 1994, 1444, 71-78. [51] Hickman, J. (Ed.); Hassel, D.; Joumard, R.; Samaras, Z. and Sorenson, S. Methodology for Calculating Transport Emissions and Energy Consumption. Transport Research Laboratory, UK. 1999, Deliverable No. 22 for the Project MEET, EC-DGVII, Brussels, Belgium. Report No. SE/491/98. [52] Holmén, B.A. and Niemeier, D.A. Characterising the effects of driver variability on real-world vehicle emissions. Transportation Research Record Part D. 1998, 3 (2), 117-128. [53] Hoose, N.; North, R.; Polak, J.W.; Cohen, J. and Richards, M. The use of mobile sensors and GRID computing managing the environmental impact of traffic. Proceedings of the 7th European Congress on Intelligent Transport Systems and Services. Geneva, 2008.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

63

[54] INFRAS. 2007. Handbook Emission Factors for Road Transport – HBEFA http://www.hbefa.net/ [55] Houghton J.T.; Meira Filho, L.G.; Lim, B.; Treanton, K.; Mamaty, I.; Bonduki, I.; Griggs, D.J. and Callender, B.A. (Eds.). 1996. Revised IPCC Guidelines for National Greenhouse Gas Inventories –Reference Manual, 3. http://www.ipccnggip.iges.or.jp/public/gl/invs6.htm. [56] ISSRC. 2008. International Vehicle Emissions Model. http://www.issrc.org/ive/ [57] Jackson, E. and Aultman-Hall, L.; The accuracy of GPS based acceleration for vehicle emissions modelling. Transportation Research Board Annual Meeting. 2007, 1-22. [58] Joumard, R.; André, M.; Vidon, R.; Tassel, P. and Pruvost, C. Influence of driving cycles on unit emissions from passenger cars. Atmospheric Environment. 2000, 34, 4621-4628. [59] Kurtzweg, J.A. Urban planning and air pollution: a review of recent research. Journal of the American Institute of Planners. 1973, 39, 2, 82-92. [60] LAT. 2008. COPERT IV. http://lat.eng.auth.gr/copert/ [61] Lyons, T.J.; Kenworthy, J.R.; Moy, C. and Dos Santos, F. An international urban air pollution model for the transportation sector. Transportation Research Part D. 2003, 8, 159-167. [62] Keller, M. 2004. Handbook Emission Factors for Road Transport 2.1. INFRAS, April 2004. [63] Keller, M. and Kljun, N. Assessment and reliability of transport emission models and inventory systems: Road Emission Model – Model Description. The European Commission. 2007, 1999-RD.10429, Workpackage 1100, Deliverable No. 13. Berne, Switzerland. [64] Kent, J.H. and Mudford, N.R. Motor vehicle emissions and fuel consumption modelling. Transportation Research Part A. 1979, 13, 395-406. [65] Kent, J.H.; Post, K. and Tomlin, J.A. Fuel consumption and emission modelling in traffic links. Proceedings of SAE-A/Australian Road Research Board 2nd Conference on Traffic Energy and Emissions. 1982, 82140, 10.1-10.20. [66] Klein, J.; Hoen, A.; Hulskotte, J.; Van Duynhoven, N.; Smit, R.; Hensema, A. and Broekhuizen, D. Methoden voor de Berekening van de Emissies door Mobiele Bronnen in Nederland (Methods for the Computation of Emissions from Mobile Sources in the Netherlands), in Dutch. Dutch Central Bureau for Statistics (CBS), http://www.cbs.nl/ . November 2006. [67] Koupal, J. and German, J. Real-time simulation of vehicle emissions using VEMISS. Proceedings of the 5th Co-ordinating Research Council On-Road Vehicle Emissions Workshop. San Diego, California, USA. April 1995, 2/65-2/94. [68] Krishnamurthy, M.; Chen, X.; Schmid, N.A. and Gautam, M. Development of a NOx predictive model using multiple adaptive regression splines.’ Transportation Research Board Annual Meeting. 2007, 1-14. [69] Kunselman, P.; McAdams, H.T.; Domke, C.J. and Williams, M. Automobile Exhaust Emissions Modal Analysis Model. US Environment Protection Agency, Office of Air and Water Programs, Office of Mobile Source and Air Pollution. 1974, Report No. EPA-460/3-74-005.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

64

Robin Smit, Hussein Dia and Lidia Morawska

[70] Leung, D.Y.C. and Williams, D.J. Modelling of motor vehicle fuel consumption and emissions using a power-based model. Environmental Monitoring and Assessment. 2000, 65, 21-29. [71] Lacour, S.; Joumard, R. and Andr‚ M. Exploring ways to improve instantaneous emission models for passenger cars. International Journal of Vehicle Design. 2001, 27, 1-4, 76-85. [72] Ludwig, F.L.; Sandys, R.C. and Moon, A.E. Preliminary studies of the modelling of air pollution effects from traffic engineering alternatives. Journal of the Air Pollution Control Association, 1973, 23, 6, 499-504. [73] Markel, T.; Brooker, A.; Hendricks, T.; Johnson, V.; Kelly, K.; Kramer, B.; O’Keefe, M.; Sprik, S. and Wipke, K. ADVISOR: a systems analysis tool for advanced vehicle modelling. Journal of Power Sources. 2002, 110, 255-266. [74] Matzoros, A. Results from a model of road traffic air pollution, featuring junction effects and vehicle operating modes. Traffic Engineering and Control. 1990, 31, 1, 2435. [75] Mellios, G.; Van Aalst, R. and Samaras, Z. Validation of road traffic urban emission inventories by means of concentration data measured at air quality monitoring stations. Atmospheric Environment. 2006, 40, 7362-7377. [76] Meng, Y. and Niemeier, D.A. Project level carbon monoxide hot-spot analysis for Level of Service D intersections. Transportation Research Record. 1998, 1641, 980679, 73-80. [77] Midenet, S.; Boillot, F. and Pierrelée, J.C. Signalized intersection with real-time adaptive control: on-field assessment of CO2 and pollutant emission reduction. Transportation Research Part D. 2004, 9, 29-47. [78] Sucharov, L.J. and Brebbia, C.A. (Eds.); Morawska, L.; Ferreira, L.; Jamriska, M.; Thomas, S. and McGregor, F. Estimation of motor vehicle emission factors from road measurements of pollutant concentrations. Urban Transport and the Environment VII in 21st Century. WIT press, Southampton. 2001, 455-465. [79] Mukherjee, P. and Viswanathan, S. Carbon monoxide modeling from transportation sources. Chemosphere. 2001, 45, 1071-1083. [80] Nagendra, S.M.S. and Khare, M. Review - line source emission modelling. Atmospheric Environment. 2002, 36, 2083-2098. [81] Negrenti, E. The 'corrected average speed' approach in ENEA's TEE model: an innovative solution for the evaluation of the energetic and environmental impacts of urban transport policies. Science of the Total Environment. 1999, 235, 411-413. [82] Neylon, M. and Collins, B. Mobile sources emissions inventory for the Melbourne airshed study. Proceedings of the 2nd Conference on Traffic, Energy and Emissions, SAE Australasia and Australian Road Research Board, Australia. May 1982, 82144, 14.1-14.16. [83] Nesami, K.S. and Subramanian, K.P. Impact of real-world driving characteristics on vehicular emissions. Institute of Transport Studies, University of California, USA. December 2005. [84] Noland, R.B. and Quddus, M. Flow improvements and vehicle emissions: effects of trip generation and emission control technology. Transportation Research Part D. 2006, 11, 1-14.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

65

[85] North, R.J.; Ochieng, W.Y.; Quddus, M.A.; Noland, R.B. and Polak, J.W. Development of a vehicle emissions monitoring system. Transportation Research Part D. 2005, 13662, 167-177. [86] North, R.J.; Noland, R.B.; Ochieng, W.Y. and Polak, J.W. Modelling of particulate matter mass emissions from a light-duty vehicle. Transportation Research Part D. 2006, 11, 344-357. [87] Ntziachristos, L. and Samaras, Z. COPERT III Computer Programme to Calculate Emissions from Road Transport: Methodology and Emission Factors (Version 2.1). European Environment Agency. 2000, Technical Report No. 49, Copenhagen. [88] Owen, B.; Edmunds, H.A.; Carruthers, D.J. and Singles, R.J. Prediction of total oxides of nitrogen and nitrogen dioxide concentrations in a large urban area using a new generation urban scale dispersion model with integral chemistry model. Atmospheric Environment. 2000, 34, 397-406. [89] Panis, L.; Broekx, S. and Liu, R. Modelling instantaneous traffic emission and the influence of traffic speed limits. Science of the Total Environment. 2006, 371, 270-285. [90] Park, J.Y.; Noland, R.B. and Polak, J.W. Microscopic model of air pollutant concentrations – Comparison of simulated results with measured and macroscopic estimates. Transportation Research Record. 2001, 1773, 32-38. [91] Pekol, A.; Anyon, P.; Hulbert, M.; Smit, R.; Ormerod, R. and Ischtwan, J. South East Queensland motor vehicle fleet air emissions inventory 2000, 2005, 2011. Linking Air Pollution Science, Policy and Management. Proceedings of National Clean Air Conference. Newcastle, Australia. 2003. [92] Pokharel, S.S.; Bishop, G.A. and Stedman, D.H. An on-road motor vehicle emissions inventory for Denver: an efficient alternative to modelling. Atmospheric Environment. 2002, 36, 5177-5184. [93] Post, K.; Kent, J.H.; Tomlin, J. and Carruthers, N. Vehicle characterization and fuel consumption prediction using maps and power demand models. International Journal of Vehicle Design. 1985, 6, 1, 72-92. [94] Queensland Government Environmental Protection Agency. 2002. South-East Queensland Motor Vehicle Emissions Inventory – Final Report. Adam Pekol Consulting, Parsons Australia, Pacific Air and Environment, Computing in Transportation and Resource Coordination Partnership. [95] Qiao, F.; Wang, Z. and Lei Yu, P.E. On-road vehicle emissions in Beijing, China: an experimental study using PEMS. Transportation Research Board Annual Meeting. 2007, 1-25. [96] Rakha, H.; Ahn, K. and Trani, A. Development of VT-Micro model for estimating hot stabilized light-duty vehicle and truck emissions. Transportation Research Part D. 2004, 9, 49-74. [97] Rexeis, M. and Hausberger, S. Calculation of vehicle emissions in road networks with the model “NEMO”. Proceedings of the 14th Symposium Transport and Air Pollution Conference. Graz, Austria. June 2005, 1, 118-127. ISBN 3 902465 16 6. [98] Rexeis, M.; Hausberger, S.; Riemersma, I.; Tartakovsky, L.; Zvirin, Y.; Van Poppel, M. and Cornelis, E. 2005 Heavy duty vehicle emissions - Final Report for ARTEMIS WP 400. European Commission. Report No: I 02/2005/Hb 20/2000 I680.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

66

Robin Smit, Hussein Dia and Lidia Morawska

[99] Rose, A.H.; Smith, R.; McMichael, W.F. and Kruse, R.F. Comparison of auto exhaust emissions in two major cities. Journal of the Air Pollution Control Association. 1965, 15, 8, 362-371. [100] Reynolds, A.W. and Broderick, B.M. Development of an emissions inventory model for mobile sources. Transportation Research Part D. 2000, 5, 2, 77-101. [101] Schulz, D.; Younglove, T. and Barth, M. Statistical analysis and model validation of automobile emissions. Journal of Transportation and Statistics. September 2000, 2938. [102] Scott, D.M.; Kanaroglou, P.V. and Anderson, W.P. Impacts of commuting efficiency on congestion and emissions: case of the Hamilton CMA, Canada. Transportation Research Part D. 1997, 2, 4, 245-257. [103] Singer, B.C. and Harley, R.A. A fuel based motor vehicle emission inventory. Journal of the Air and Waste Management Association. 1996, 46, 581-593. [104] Singer, B.C. and Harley, R.A. A fuel-based inventory of motor vehicle exhaust emissions in the Los Angeles area during Summer 1997. Atmospheric Environment. 2000, 34, 1783-1795. [105] Singh, R.B. and Huber, A.H. Development of a microscale emission factor model for CO for predicting real-time motor vehicle emissions. Journal of the Air and Waste Management Association. 2000, 50, 1980-1991. [106] Smit, R.; Bawden, K. and Ormerod, R. Speciated hydrocarbon emissions from road traffic. Proceedings of the 16th Clean Air and Environment Conference, CASANZ. Christchurch, New Zealand. 2002, 746-757. [107] Smit, R. An Examination of Congestion in Road Traffic Emission Models and Their Application to Urban Road Networks. Ph.D. Dissertation Griffith University, Brisbane, Australia. 2006. [108] Smit, R. Engine Technology and Emissions of Air Pollutants. Course prepared for Stichting Postacademisch Onderwijs (Organisation for Post-Academic Education). Delft, Netherlands. 2006. [109] Smit, R.; Smokers, R. and Rabé, E. A new modelling approach for road traffic emissions: VERSIT+. Transportation Research Part D. 2007, 12, 414-422. [110] Smit, R. 2007. Statistical Driving Pattern Analysis: An Extrapolation Assessment Tool for VERSIT + LD – Proof of Concept. Report TNO-033-DTS-2007-02502. [111] Smit, R.; Poelman, M. and Schrijver, J. Improved road traffic emission inventories by adding mean speed distributions. Atmospheric Environment. 2008, 42, 916-926. [112] Smit, R.; Brown, A.L. and Chan, Y.C. Do air pollution emissions and fuel consumption models for roadways include the effects of congestion in the roadway traffic flow? Environmental Modelling and Software. 2008, 23, 10, 1262-1270. [113] Smit, R. Developments in road traffic emission and fuel consumption modelling: some recent experiences from Europe. Clean Air and Environmental Quality. 2008, 42, 3, 1821. [114] Smit, R. Errors in model predictions of NOx traffic emissions at road level – Impacts of input data quality. 16th International Conference on Modelling, Monitoring and Management of Air Pollution. Skiathos, Greece. 2008. [115] 5.Smit, R., McBroom, J. Development of a New High Resolution Traffic Emissions and Fuel Consumption Model for Australia and New Zealand – Why is it needed? Air Quality and Climate Change, 2009a, 43, 1, February 2009a, 14-16.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Road Traffic Emission and Fuel Consumption Modelling

67

[116] Smit, R., McBroom, J. 2009b, Development of a New High Resolution Traffic Emissions and Fuel Consumption Model for Australia and New Zealand – Data Quality Considerations, Air Quality and Climate Change, 2009b, 43, 2, May 2009, 13-16. [117] Sonntag, D. and Gao, H.O. The MOVES from MOBILE: preliminary comparison of EPA’s current and future emissions models. Transportation Research Board Annual Meeting. 2007, 1-17. [118] St. Denis, M.J. and Winer, A.M. Prediction of on-road emissions and comparison of modeled on-road emissions to the FTP emissions. The Emission Inventory: Perception and Reality –Journal of the Air and Waste Management Association. 1994, 495-505. [119] Staehelin, J.; Keller, C.; Stahel, W.; Schläpfer, K. and Wunderli, S. Emission factors from road traffic from a tunnel study (Gubrist Tunnel, Switzerland). Part 3: results of organic compounds, SO2 and speciation of organic exhaust emissions. Atmospheric Environment. 1998, 32, 6, 999-1009. [120] Taylor, N.B. The CONTRAM dynamic traffic assignment model. Networks and Spatial Economics. 2003, 3, 297-322. [121] Taylor, M.A.P. and Young, T.M. Fuel consumption and emissions models for traffic engineering and transport planning applications: some new results. Proceedings of 18th ARRB Transport Research Conference/ Roads 96 Conference. Christchurch, New Zealand. 1996, 6, 189-204. [122] Toledo, T.; Koutsopoulos, H.N.and Ahmed, K.I. Estimation of vehicle trajectories with locally weighted regression. Transportation Research Board Annual Meeting. 2007, 118. [123] Transportation Research Board. 2000. Highway Capacity Manual. ISBN 0 309 06681 6. Washington D.C., USA: National Research Council. [124] US EPA. 2008. MOBILE6 Vehicle Emission Modeling Software. http://www.epa.gov/ omswww/m6.htm. [125] US EPA. 2008. MOVES (Motor Vehicle Emission Simulator), http://www.epa.gov /oms/ngm.htm. [126] Koupal, Hart, C., Brzezinski, D., Giannelli, R. and Bailey, C. 2002. Draft Emission Analysis Plan for MOVES GHG. Report No. EPA420-P-02-008. US EPA Office of Transportation and Air Quality. [127] Van Vliet, D. SATURN - A modern assignment model. Traffic Engineering and Control. 1982, 23, 12, 578-581. [128] Veurman, J.; Gense, N.L.J.; Wilmink, I.R. and Baarbé, H.I. Emissions at different conditions of traffic flow. Proceedings of the 8th International Conference on Urban Transport and the Environment in the 21st Century. Malta. 2002, 571-580. Wit Press, U.K. ISBN 1 85312 905 4. [129] Watson, H.C.; Milkins, E.E.; Preston, M.O.; Beardsley, P. and Chittleborough, C.W. Further application of the PKE average speed fuel consumption and emissions model. Proceedings of Joint SAE-A/Australian Road Research Board 2nd Conference on Traffic, Energy and Emissions. Melbourne, Australia. 1982, 82139, 9.1-9.14. [130] Watson, H.C.; Milkins, E.E.; Preston, M.O.; Chittleborough, C. and Alimordian, B. Predicting fuel consumption and emissions, transferring chassis dynamometer results to real driving conditions. SAE Technical Paper Series. 1983, 830435, 1-24. [131] Watson, H.C. Effects of a wide range of drive cycles on the emissions from vehicles of three levels of technology. SAE Technical Paper Series. 1995, 950221, 119-132.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

68

Robin Smit, Hussein Dia and Lidia Morawska

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[132] WHO. Guidelines for concentration and exposure-response measurement of fine and ultrafine particulate matter for use in epidemiological studies. 2002, World Health Organisation, Geneva. [133] Weilenmann, M.; Bach, C. and Rüdy, C. Aspects of instantaneous emission measurement. Proceedings of the 9th International Symposium Transport and Air Pollution Conference. Avignon. 2000, 119-126. [134] Zalinger, M.; Ahn, T.L. and Hausberger, S. Improving an instantaneous emission model for passenger cars. Proceedings of the 14th Symposium Transport and Air Pollution. Graz, Austria. 2005, 1, 167-176. ISBN 3 902465 16 6. [135] Zalinger, M.; Tate, J. and Hausberger, S. An instantaneous emission model for the passenger car fleet. 16th International Transport and Air Pollution Conference. Graz, Austria. 2008, 1-12. [136] Zhang, Y.; Stedman, D.H.; Bishop, G.A.; Guenther, P.L. and Beaton, S.P. Worldwide on-road vehicle exhaust emissions study by remote sensing. Environmental Science and Technology. 1995, 29, 9, 2286-2294. [137] Zito, R. and Taylor, M.A.P. Irvine, L. (Ed.), Modelling Emissions from an Instrumented Car. Proceedings of the 8th World Congress on Intelligent Transport Systems. Sydney, Australia. 2001.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 69-101

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 3

TAILPIPE PARTICLE EMISSION FACTORS DERIVED FOR MOTOR VEHICLES FOR APPLICATION TO TRANSPORT MODELLING AND HEALTH IMPACT ASSESSMENTS OF URBAN FLEETS Diane U. Keogh1, Joe Kelly2, Kerrie Mengersen2, Rohan Jayaratne2, Luis Ferreira3 and Lidia Morawska1

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1

International Laboratory for Air Quality and Health, Queensland University of Technology, Gardens Point, Brisbane, Australia 2 School of Mathematical Sciences, Queensland University of Technology, Gardens Point, Brisbane, Australia 3 School of Urban Development, Queensland University of Technology, Gardens Point, Brisbane, Australia

Abstract Motor vehicles are a major source of particulate matter pollution in urban areas and therefore estimates relating to the extent of this pollution are critically needed for urban and transport planning, scenario modelling, and development of mitigation strategies, air quality assessments and air quality regulation. More than 900 particle emission factors are available in the international published literature relating to motor vehicle tailpipe emissions, however it remained unclear which are the most appropriate to use in transport modelling and health impact assessments. A very large body of published data on emission factors was reviewed within the scope of this work, and based on a statistical analysis a comprehensive set of tailpipe particle emission factors for motor vehicles was derived. This chapter presents the most appropriate emission factors derived to use in transport modelling and health impact assessments for particle number, particle volume, PM1, PM2.5 and PM10 to develop comprehensive, size-resolved inventories of tailpipe particle emissions for urban fleets in developed countries, covering the full size range of particles emitted. This chapter presents the outputs of statistical models developed to derive these emission factors, their explanatory variables, and the average particle emission factors, corresponding 95% confidence intervals and standard errors produced by the statistical models. Results of

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

70

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al. statistical tests which investigated the statistical relationship between sub-classes of categorical model variables for different particle metrics examined in the analysis are also presented.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Introduction Particulate matter generated by motor vehicle fleets in urban areas is a major source of pollution, and these particle emissions span diameters ranging from 0.003–10 µm. Most of these particles are ultrafine size and measured in terms of particle number (number concentration of particles with diameters < 0.1 µm) (Harrison et al. 1999; Shi and Harrison 1999; Shi et al. 1999; Shi et al. 2001; Morawska 2003; Wahlin et al. 2001). To date a comprehensive inventory of motor vehicle tailpipe particle emissions that covers this broad size range is not available anywhere in the world; nor is a detailed emission inventory available for particle number concentration (Jones and Harrison 2006). This means that our current knowledge about the full extent of particulate matter pollution emitted by motor vehicle fleets is substantially incomplete. A wide range of different methods are used to estimate motor vehicle emission inventories, from mobile emission models such as the average speed models MOBILE (USEPA 1993), EMFAC (CARB 2002), COPERT (Ahlvik et al. 1997; Ntziachristos et al. 2000; Bellasio et al. 2007); and VERSIT+ LD (Smit et al. 2007), which employ performancebased emission factors (emissions generated per vehicle per kilometre derived from measurement data); to more indirect methods such as basing estimates on remotely sensed data or estimated total fuel consumed (Goodwin et al. 1999; Kittelson et al. 2004; Shifter et al. 2005). Indirect methods are often used in developing countries as access to land use and transport network data is often rare (Walker et al. 2008). Inventories based on remotely sensed data may not represent a typical trip in a region because they often only provide a snapshot of emissions relating to a limited number of locations (Frey et al. 2002a). The accuracy of fuel-based models can be dependent upon how well vehicle type, age distribution and driving modes, from which the emission factors were derived, represent the region being modelled (Frey et al. 2002 a,b). As most motor vehicle particle emissions are ultrafine size and measured in terms of particle number, it is critical that inventories and health impact assessments include estimates for particle number. The majority of motor vehicle emission models provide estimates which are limited to PM10, and to a lesser extent PM2.5 (mass concentration of particles with aerodynamic diameters < 10 µm and < 2.5 µm respectively). One average speed model COPERT IV, however, has available a small suite of solid particle number emission factors for different vehicle types derived from dynamometer measurements (Samaras et al. 2005). Current knowledge concerning which are the most appropriate emission factors to use in transport modelling and health impact assessments is patchy and ill-defined, and a very large body of data is available in the international literature on published tailpipe particle emission factors for motor vehicles under different driving conditions. These have been derived in measurement studies conducted on dynamometers in laboratories, measured on or near roads and in tunnels, but it remains unclear: Which emission factors are the most appropriate to use in transport modelling and health impact assessments?

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

71

Many different measurement methods have been used in different parts of the world that have measured different particle size ranges, and in order to derive emission factors a multiplicity of issues need to be considered. These can range from vehicle type, fuel type, vehicle age, technologies fitted, vehicle speed and load, road environment characteristics, driving patterns and driving cycles, size range measured to method and instrumentation used, to name a few. To date this very large body of data on particle emission factors for motor vehicles has not been comprehensively analysed. This chapter presents a comprehensive set of tailpipe particle emission factors for different Vehicle and Road Type combinations that can be used in transport modelling and health impact assessments to estimate motor vehicle fleet emissions. As only limited data is available in the international literature for emission factors for brake and tyre wear, road dust and particle surface area emissions these were not considered in this study.

Method An extensive review was undertaken of the international literature in relation to published emission factors for motor vehicle tailpipe emissions for particle number, particle volume, total particle mass, PM1, PM2.5 and PM10. The publications examined are outlined in Appendix 1 (and referenced in Appendix 4). Based on this review, statistical models were developed and emission factor data classified and grouped into relevant sub-classes within each model variable class. Outputs from these statistical models were analysed and a rationale developed to identify the most appropriate average emission factors to use in modelling urban fleet emissions.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Model Variables Examined More than 900 tailpipe particle emission factors were identified in the international published literature, and based on variables commonly mentioned in these publications, model variables were developed for the statistical analysis. Data relating to a total of 667 particle emission factors were examined in the statistical analysis, which was grouped into relevant sub-classes within each model variable class. The emission factors examined in this statistical analysis were derived from measurement studies conducted in developed countries, and hence the final set of emission factors selected as the most appropriate have particular application for urban areas in developed countries. From this point on in this chapter, tailpipe particle emission factors for motor vehicles will simply be referred to as motor vehicle particle emission factors. For the statistical analysis, two categories of model variables were developed. Categorical variables, which were Particle Metric, Country of Study, Study Location, Instrumentation, Vehicle Type, Fuel Type, Road Type, Road Class; and continuous variables which were Size Range Measured, Average Vehicle Speed, Speed Limit on the Road, Average Number of Vehicles travelling in a fleet per day, Drive Cycles, Engine Power, Heavy Duty Vehicle (HDV) Share, Number of HDVs travelling in a fleet per day. These model variables are described in Appendix 2. Appendix 3 presents the sample sizes of emission factors examined for the different model variables and particle metrics.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

72

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Statistical Analysis of Model Variables Categorical variables were analysed by fitting a univariate general linear model (a multifactor ANOVA), and a stepwise technique was used to select the best model. The continuous variables were analysed using simple linear regression, with the variables added as independent explanatory covariates in the general linear model. ANOVA tests and post-hoc Scheffe multiple comparisons (Scheffe, 1959) were employed to identify the model variables which were statistically significant. All analysis was undertaken at a 5% level of significance. The post-hoc Scheffe multiple comparison tests were conducted for all categorical variables and their sub-classes to determine if, within each class of categorical variable, there were statistically significant differences between the average published emission factor values for different sub-classes of variable, irrespective of whether the categorical variable was found to be statistically significant. The software package SPSS (Version 14.0) was used to analyse the data, and separate statistical models were developed for the six particle metrics examined. The model coefficients of determination (R2) for these statistical models provided information on the fraction of variability in the dependent variable (the emission factor) that was accounted for by the variation in the independent variable/s. The statistical models produced average particle emission factors, and their associated standard error and 95% confidence intervals. The standard error provided an indication of the reliability of the model as a means of predicting the average particle emission factor for each combination of values of the independent variables it related to. The lower the standard error, in relation to the average emission factor it is associated with, the more reliable the predicted average emission factor may be considered. The 95% confidence interval lower and upper bound values produced by the statistical models for each average emission factor indicated the range within which it can be assumed we might be 95% confident the true value will lie. Some combinations of dependent and independent variables in the statistical models produced high standard error values, and a lower bound 95% confidence interval value which, although physically uninterpretable, can be obtained as a consequence of the normal assumptions underlying the model. Hence, where these lower bound values were obtained, they were not reported.

The Basis for Selecting the Most Appropriate Particle Emission Factors The wide range of different capabilities of Instrumentation used to derive emission factors were not evaluated as an aim of this study. The basis for selection of the most appropriate emission factors related to examination of the statistical robustness of the statistical model outputs. These included consideration of conservative average particle emission factors with the lowest standard errors, narrowest 95% confidence intervals and largest sample sizes. Other factors taken into account in selection were the relevant explanatory variables for each particle metric and, in the case of the model variable Size Range Measured, the focus was on Instrumentation which measured the broadest size ranges, including down to the lowest size range.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

73

Results and Discussion The emission factors considered the most appropriate to use in transport modelling and health impact assessments are summarized in Table 2, and appear in Tables 3-7 shaded, in bold italics. Tables 3-7 present the outputs of the five statistical models developed for particle number, particle volume, PM1, PM2.5 and PM10. Outputs for the sixth model, developed for total particle mass, are not presented in this study as this was found to be null model because no explanatory variables were found. The results of post-hoc multiple comparison statistical tests that examined differences in mean values of published emission factors for sub-classes of all categorical variables examined in this study are shown in Figure 1, and are discussed below. All emission factors derived by the statistical models are expressed in terms of particle emissions generated per vehicle per kilometre driven. One single emission factor value may represent an individual vehicle tested on a dynamometer (or a group of vehicles), or be the average emission factor derived for a Vehicle Type (such as a heavy duty vehicle) travelling in a fleet in a tunnel or on a road. Therefore, the total sample size of 667 emission factors examined in this study represent a relatively very large sample of motor vehicles. The extensive literature review conducted in this study revealed that at the time of this study the majority of light duty vehicles (LDV) were petrol-fuelled and heavy duty vehicles (HDV) diesel-fuelled.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Variation in Published Emission Factors Explained by the Statistical Models In the statistical analysis six separate statistical models were proposed for particle number, particle volume, total particle mass, PM1, PM2.5 and PM10 . Table 1 lists the sample sizes of emission factors examined in this study, the explanatory variables and coefficients of determination (R2) identified for each of the statistical models, and also summarizes the rationale used to select the most appropriate emission factors. The particle number, particle volume, PM1 and PM2.5 statistical models were found to be robust and explained 86%, 93%, 87% and 65% respectively of the variation in published emission factors (Table 1). The PM10 model, on the other hand, was less robust and explained only 47% of the variation in published emission factor values (Table 1). The inability of the PM10 statistical model to explain a higher degree of variation may relate to the confounding effects of resuspended road dust at the PM10 size range, and the fact that PM10 emission factors derived from studies conducted on or near roads and in tunnels are likely to have been affected by various quantities of resuspended road dust.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009. ProQuest

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 1. Sample size of emission factors examined in each statistical model, the explanatory variables and coefficients of determination; and rationale for selection of the most appropriate tailpipe particle emission factors to use in transport modelling and health impact assessments

a

Number of emission factors examined in the statistical analysis. One emission factor may represent a single vehicle measurement, or be an average value derived to represent an entire vehicle fleet (or vehicle category such as LDV or HDV) travelling in a vehicle fleet on a road or in a tunnel, or represent the average of a group of vehicles tested on a dynamometer. b From an original population of over 900 emission factors, the final sample size obtained was 667. This occurred due to the high number of missing values in the studies, as not all studies reported the same information in their studies. c The inability to identify relationships in this model may stem from the fact that the studies measured a broad range of different particle sizes, and emission factors were not generally derived segregated by particle size. The sample size was 199 and overall mean from this null model was 158 mg/km for all combined vehicle types; 158 mg/km for Bus, and 91mg/km for Fleet 380 mg/km for HDV and 32 mg/km for LDV.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009. ProQuest

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 2. Tailpipe particle emission factors for motor vehicles considered the most appropriate to use in transport modelling and health impact assessments of urban fleets in developed countries

a

The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. These values, although physically uninterpretable, can be obtained as a consequence of the normal assumptions underlying the models, and hence are not reported. b Diesel buses. c Buses – Fuel not specified (can be assumed to be Diesel-fuelled due to the timing and location of the studies), principally Diesel-fuelled buses. d Condensation Particle Counter (CPC), Scanning Mobility Particle Sizer (SMPS), Tapered Element Oscillating Microbalances(TEOM) and Differential Mobility Analyzer (DMA). e The average dynamometer emission factor for buses for PM10 is also presented; as the on-road boulevard and urban Road Type studies were reported to be affected by very high levels of resuspended road dust and the influence of variation in acceleration and speed (Abu-Allaban et al. 2003).

76

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

The total particle mass model was found to be a null model as no explanatory variables were identified, and this is likely to be attributed to the fact that most of the studies measured total particle mass and not subsets of different particle mass size fractions, that typically have varying proportions of particle mass associated with them. Table 2 presents the average emission factors considered the most appropriate to use in transport modelling and health impact assessments for different Vehicle and Road Type combinations, and their 95% confidence interval values. These emission factors are also shown in Tables 3-7, shaded in bold italics, and these tables present the outputs of the five statistical models for different combinations of statistically significant model variables.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Particle number model explanatory variables and average particle number emission factors, their standard errors and 95% confidence intervals

(a) The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. (b) The minimum and maximum size range measured by the Instrumentation. (c) Includes 1 HDV and 1 Bus Dynamometer measurement. (d) Based on modified population marginal mean. (e) 300 Bus trips were measured in a tunnel and 12 Buses tested on a dynamometer. Instrumentation: CPC - Condensation Particle Counter; DMA - Differential Mobility Analyser ; DMPS - Differential Mobility Particle Sizer ; EAA – Electrical Aerosol Analyser ; ELPI – Electrical Low Pressure Impactor ; SMPS – Scanning Mobility Particle Sizer; UCPC – Ultrafine Condensation Particle Counter. *ALL – Instrumentation and Vehicle Types. ** ALL emission factors Combined.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

77

Particle number model: Table 3 presents the statistical model outputs for 10 different Instrumentation. This statistical model explained 86% of the variation in published emission factors, and the explanatory variables were Vehicle Type and Instrumentation. It was important when selecting the most appropriate emission factors for particle number (shaded in bold italics in Table 3) to select those that related to Instrumentation measuring the lowest possible size range, including down to 0.003 µm where particle number tends to be very prolific. This lower limit size range is commonly measured by the Condensation Particle Counter (CPC). The CPC estimates particle count, and emission factors were available in the international literature based on CPC measurements for Fleet, light duty vehicles (LDV) and heavy duty vehicles (HDV).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 4. Particle volume model explanatory variables and average particle volume emission factors, their standard errors and 95% confidence intervals

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

78

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 4. Continued

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

79

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 4. Continued

(a) The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. (b) This study measured a vehicle fleet that comprised 60% HDVs. (c) Based on modified population marginal mean. *All emission factors for Vehicle Types and Size Ranges Measured Combined. ** All emission factors Combined.

Particle number emission factors for Diesel buses were restricted to those derived from Scanning Mobility Particle Sizer (SMPS) measurements. The SMPS focuses on estimating particle size distribution, as opposed to total particle count, and does not measure the lower size range of the nucleation mode < 10 µm. The lower size window of the SMPS is usually set higher than that for the CPC, generally in the range 0.010-0.02 µm, whereas for the CPC the range is usually 0.002-0.01 µm. This means that the CPC usually measures the lower size range of the nucleation mode and the SMPS does not. Particle volume model: This statistical model explained 93% of the variation in published emission factors, and the explanatory model variables were Vehicle Type, Speed Limit on the Road and Size Range Measured. The statistical model outputs are shown in Table 4 and consideration was given to selecting emission factors for different reported Speed Limits on the Road and which related to the broadest size ranges measured, including down to the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

80

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

lowest possible size range. In Table 4 the most appropriate emission factors are shaded in bold italics. Most of the average particle volume emission factors produced by this statistical model, and their 95% confidence interval values, were less than 1 cm3 per vehicle per kilometre. PM1 model: Table 5 presents the statistical model outputs for PM1, and the most appropriate emission factors are shaded in bold italics. The explanatory variables were Vehicle Type and Fuel Type, which explained 87% of the variation in published emission factors. Emission factors examined included those derived for diesel vehicles measured on a dynamometer; and from studies conducted in tunnels or on or near roads where the Fuel Type was not specified. Very few emission factor data are available in the published literature for PM1, and as most motor vehicle particle emissions are < 1 µm (dominated by ultrafine particles) this is an important size range to develop a database for. Recent research by Morawska et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 5. PM1 model explanatory variables and average PM1 emission factors, their standard errors and 95% confidence intervals

(a) The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. (b) Based on modified population marginal mean. * ALL - Diesel and Fuel not Specified.

(2008) found a combination of PM1 and PM10 mass ambient air quality standards are likely to be more suitable for controlling combustion and mechanically-generated sources, such as motor vehicles, than the current standards of PM2.5 and PM10, which emphasizes the importance of deriving emission factors for PM1 for different Vehicle Types. PM2.5 model: This statistical model examined emission factors for 8 different Instrumentation, and explained 65% of the variation in published emission factors. The explanatory variables were Vehicle Type and Instrumentation. Table 6 presents the statistical model outputs, and the most appropriate emission factors are shaded in bold italics.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

81

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 6. PM2.5 model explanatory variables and average PM2.5 emission factors, their standard errors and 95% confidence intervals

(a) The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. (b) Based on modified population marginal mean. (c) Relate to 300 Diesel Bus trips measured in a tunnel, and Buses, Fuel Not Specified, tested in the vicinity of the road. (d) Buses included 3 hybrid buses (2 fitted with catalysed particulate filters); 3 Buses fuelled with Diesel (fitted with oxidation catalysts) and 1 Bus fuelled with liquified natural gas. (e) TEOM equivalent data, where the correlation between TEOM and DustTrak response to diesel emissions was assessed and the DustTrak results were recalculated into TEOM equivalent data. Instrumentation: APS – Aerodynamic Particle Sizer; DMPS – Differential Mobility Particle Sizer; TEOM – Tapered Element Oscillating Microbalances. * ALL – Instrumentation and Vehicle Types.

PM10 model: The explanatory variables for this statistical model were Vehicle Type and Road Type, which explained 47% of the variation in published emission factors. The statistical model’s low R2 value is reflected in large values for standard errors (in relation to

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

82

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

the predicted average emission factor) and high values for upper bound 95% confidence intervals produced by the statistical model. Resuspended road dust at the PM10 size range is likely to have confounded the ability of the statistical model to explain the variation in published emission factors. Few methods are available for discriminating road dust from motor vehicle tailpipe emissions, particularly at the PM2.5 and PM10 size ranges, and quantities of road dust can vary depending on the construction material of road surfaces and their maintenance, climatic conditions, and other factors such as vehicle speed and traffic volumes. Few emission factors are available for buses that have been derived from on-road measurements, and those available and included in the statistical model related to measurements on urban and boulevard Road Types in the USA (Abu-Allaban et al. 2003). Abu-Allaban et al. (2003), the authors of this study, reported that they considered their high PM10 emission factors were influenced by significantly high contributions from resuspended road dust and, within each vehicle category, by the effects of speed and acceleration. For this reason the average emission factor for buses derived from dynamometer measurements is also presented as an appropriate emission factor in Table 2 and in Table 7, in addition to the average emission factors for bus for urban and boulevard Road Types. This was included as the dynamometer emission factor for bus was considered more conservative than those derived for the urban and boulevard Road Types, and was unlikely to be affected by high rates of resuspended road dust. This average dynamometer emission factor for bus included consideration of emission factors for a wide range of different urban bus Drive Cycles. It can be seen from Table 7 that differences exist between the values of average emission factors produced by the statistical models for PM10 for LDVs, HDVs and buses for different Road Types, as compared to tunnel and dynamometer average emission factors. This table shows that the LDV statistical model had a sample size of 49 emission factors and produced average emission factors that ranged from 14 mg/km for tunnel (n=6) to 47 mg/km (n=10) for dynamometer studies; and for different Road Types ranged from 46 mg/km for a rural area road (n=1) to 63 mg/km for motorway (n=2); 141 mg/km (n=6) for highway; 156 mg/km (n=16) for urban road; 285 mg/km (n=4) for freeway; and 454 mg/km (n=4) for boulevard. Even greater variation can be seen between PM10 average emission factors derived by the statistical model for HDVs for tunnel, dynamometer and studies of different Road Types (Table 7), which suggest that generally higher values of emission factors are derived in studies conducted on different Road Types, which may be affected by the influence of varying quantities of resuspended road dust at the PM10 size range. Total particle mass model: No statistically significant variables were identified for the total particle mass statistical model, which had a sample size of 199 emission factors. The overall mean from this null model was 158 mg/km for all combined Vehicle Types; 158 mg/km for bus, and 91mg/km for Fleet, 380 mg/km for heavy duty vehicles and 32 mg/km for light duty vehicles. The inability of the statistical model to identify relationships may be related to the fact that most of the studies derived emission factors for total particle mass, and not for different subsets of particle mass fractions, which typically have varying quantities of mass associated with them.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

83

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 7. PM10 model explanatory variables and average PM10 emission factors, their standard errors and 95% confidence intervals

(a) The lower bound 95% confidence interval value calculated to be negative and therefore is not valid. (b) Based on modified population marginal mean. (c) Fuel not specified by the studies (can be assumed to be Diesel-fuelled due to the timing and location of the study). (d) Five of the 6 buses were tested on the CBD (Central Business District) Urban bus Drive Cycle, and 1 bus (Fuel not specified, can be assumed to be Diesel-fuelled due to the timing and location of the study) was tested on an Urban road. In this analysis CBD Drive Cycle emission factors were classed as Urban Road Type, due to the scarcity of studies available that have measured PM10 for buses in on-road measurement campaigns, and as this Drive Cycle closely emulates urban driving conditions. Of the 5 buses tested on the CBD Drive Cycle - 3 were Dieselfuelled (1 Low Sulphur Diesel (LSD) with an oxidation catalyst and 2 Ultralow Sulphur Diesel (ULSD) one with an oxidation catalyst and 1 with both an oxidation catalyst and a particle filter) and 2 Diesel Hybrids (with catalysed particle filters). (e) These 19 buses emulated a mixed bus fleet. They comprised 9 buses where the fuel was not specified (these may have been Diesel-fuelled, however the Fuel Types were not reported); 5 buses were Diesel-fuelled fitted with oxidation catalysts (of these 1 LSD; and 1 ULSD with a particle filter); 2 Diesel Hybrids (with oxidation catalysts; one also had a catalysed particle filter); and 3 LNG-fuelled with no aftertreatment devices. Although removal of the 3 LNG bus emission factors would have increased the overall average emission factor produced by the PM10 statistical model for buses for dynamometer from 313 to 371 mg/km, these LNG emission factors were not removed because for 9 of the 19 buses tested the fuel used was not specified and was unable to be determined.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

84

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Statistical Relationships between Categorical Model Variables Figure 1 presents a multiple comparison plot depicting the statistical relationships identified between the average values of published emission factors in terms of categorical variables examined in the statistical analysis. These related to the results of post-hoc Scheffe’s multiple comparison statistical tests (Scheffe 1959) that investigated the differences in means between levels corresponding to all categorical variables irrespective of whether they had a significant effect on the response variable (the published emission factor value). In Figure 1 variables whose mean values are statistically similar are connected by joinedlines, and those also annotated with an ‘X’ indicate the variable marked ‘X’ is statistically similar to the variables to which it is joined. Variables without joined-lines between them have statistically significant relationships at a 95% confidence level. Figure 1 shows that in terms of the statistical model variables: •

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.







• •



Country of Study and Study Location (dynamometer, on or near the road, tunnel) were not statistically significant in explaining the variation in the means of published emission factors for most particle metrics. No statistically significant difference was found between published emission factors for different Road Types compared with those derived from dynamometer measurements, and where differences were found these were between dynamometer and motorway (PM1) and dynamometer and boulevard Road Types (PM10). However, these differences are likely to have been influenced by high speed scenarios - the PM1 study measured emissions on a motorway in Switzerland with a speed limit of 120km/hr (Imhof et al., 2005) and the PM10 study in the USA attributed the significantly high PM10 emission rates to contributions from resuspended road dust and to the influence of variation in acceleration and speed (Abu-Allaban et al. 2003). For Vehicle Type statistically significant differences were found between the means for Fleet and HDV for particle number, PM1, PM2.5; and between the means for Fleet and LDV for PM2.5. The means for LDV and HDV were found to be statistically significantly different for all particle metrics. No statistically significant differences were found between the means of different Fuel Types for particle number, or between the means for different Fuel Types for total particle mass; but were found between the means for petrol and diesel Fuel Types for PM10. No statistically significant differences were found between mean values measured by different Instrumentation for PM2.5 and total particle mass. However, a statistically significant difference was found between the mean value for published emission factors for particle number derived from Condensation Particle Counter (CPC) measurements of 22.69 x 1014 particles per vehicle per km and Scanning Mobility Particle Sizer (SMPS) measurements of 2.08 x 1014 particles per vehicle per km, highlighting a major difference between the results of these two measurement techniques, which requires investigation as a broader issue. Statistically significant differences were found for PM1 between the means of Instrumentation for the Aerodynamic Particle Sizer (APS) and Betameter and

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…



85

between the APS and Beta-ray absorption monitors, however these differences are likely to be influenced by the fact that the PM1 measurements related exclusively to diesel vehicles (LDVs and HDVs) tested on dynamometers in Australia. Diesel vehicles typically emit higher particle emission rates than petrol and other fuelled vehicles. In relation to the Size Range Measured for particle number, no statistically significant differences were found for the lower and upper size ranges measured for particle number between the average emission factors for the various levels of each of the categorical variables, after accounting for the associated variability of these estimates. Emission factors derived using the CPC for total particle count which reported only the lower size ranges measured (and did not report the upper size range measured) were unable to be included in these statistical tests. Their inclusion may have led to a different result as the CPC generally measures down to 0.002 µm, where particle numbers are very prolific.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A general conclusion from examination of the results of the post-hoc multiple comparison tests depicted in Figure 1 and discussed above, is that these findings support the universal application of the average emission factors derived in this study for modelling urban fleet emissions in developed countries. Where statistically significant differences were found these generally related to emission factors for diesel-fuelled vehicles (higher emission rates would be expected from diesel-fuelled vehicles, as compared to petrol and other fuelled vehicles), high speed scenarios or road environments affected by significantly high levels of resuspended road dust.

(a) Post-hoc multiple comparison tests were not performed for PM1 as there were fewer than 3 Country of Study groups. (b) Australian PM2.5 emission factors related mainly to diesel-fuelled vehicles, which could generally be expected to produce higher values than other fuelled vehicles.

Figure 1a. Country of Study.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

86

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

(a) Vicinity of the Road relates to measurements on or near the road (near a curb, upwind and downwind, downwind only, using vehicle chasing or on-road mobile laboratories). (b) The PM1 dynamometer measurements related exclusively to diesel LDVs and HDVs tested in Australia. Diesel vehicles would generally produce higher emission factors than other fuelled vehicles. (c) No dynamometer values were available for particle volume, and there were less than 3 groups so post-hoc multiple comparison tests were not able to be performed.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1b. Study Location.

Figure 1c. Dynamometer and Road Types.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

87

(a) Total particle mass was not tested as sample sizes were too small. (b) No dynamometer emission factors were available for particle volume.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1d. Dynamometer and Road Classes.

(a) The statistical similarity between average emission factors for bus and HDV, bus and LDV and bus and Fleet may be influenced by the fact that bus measurements for particle number related exclusively to measurements undertaken using a Scanning Mobility Particle Size (SMPS), whereas the sample of emission factors for HDV, LDV and Fleet included measurements undertaken using a Condensation Particle Counter (CPC). The CPC measures the nucleation mode (where particle number tend to be very prolific), and the SMPS does not, which may result in lower value measurements. (b) In terms of particle mass, PM1 is considered more relevant for buses than PM2.5 and PM10.

Figure 1e. Vehicle Type.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

88

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

(a) Post-hoc multiple comparison tests were not performed for PM1 and PM2.5 as there were fewer than 3 groups. (b) Fuel Types were not reported for particle volume. Fuel Types: CNG – Compressed Natural Gas, LNG – Liquified Natural Gas, ULSD – Ultralow Sulphur Diesel, LSD – Low Sulphur Diesel.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1f. Fuel Type.

(a) The small sample size of emission factors (n=7) and very small mean value for APS in particle number of 0.002 x 1014 particles per vehicle per kilometre hampered a meaningful comparison with mean values for CPC and SMPS, which both had substantially larger sample sizes and larger mean values of CPC 22.69 (n=18) x 1014 particles per vehicle per kilometre and SMPS 2.083 x 1014 particles per vehicle per kilometre (n=96). The APS also measures a vastly different particle size range to CPC and SMPS, as shown in Table 3. (b) Post-hoc multiple comparison tests were not performed as there were fewer than 3 groups. Instrumentation - APS Aerodynamic Particle Sizer; CPC - Condensation Particle Counter; DMA - Differential Mobility, DMPS – Differential Mobility Particle Sizer; EAA – Electrical Aerosol Analyser ; ELPI – Electrical Low Pressure Impactor; SMPS – Scanning Mobility Particle Sizer. TEOM – Tapered Element Oscillating Microbalances. (c) PM10 Instrumentation sample sizes were too small to test for significant differences between the means.

Figure 1g. Instrumentation. Variables connected by a joined-line are statistically similar and those marked X show the variable marked X and the variables to which it is joined are statistically similar. Variables without a joined-line are statistically significantly different at a 95% confidence level.

Figure 1. Multiple comparison plot depicting the nature of the statistical relationship between the categorical model variables for different particle metrics.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

89

Conclusion

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A very large number of particle emission factors are available in the international literature relating to motor vehicle tailpipe emissions that have been derived from measurement studies on or near roads, in tunnels and conducted on dynamometers in laboratories. This study derived a comprehensive set of tailpipe particle emission factors for different Vehicle and Road Type combinations, based on statistical analysis of this large body of data. These emission factors enable the full size range of particles generated by motor vehicle fleets to be modelled, including particle number and different particle mass size fractions. The statistical models developed in this study were found to explain 86%, 93%, 87% and 65% of the variation in published emission factors for particle number, particle volume, PM1, and PM2.5 respectively, and present as robust models. The statistical model for PM10 was able to account for only 47% of the variation in published emission factors, and difficulty explaining a larger proportion of the total variation may be related to the confounding effects of resuspended road dust at the PM10 size range. The explanatory variables identified in these statistical models represent variables which are important to consider in the design and interpretation of emission factor studies conducted for different particle metrics. These explanatory variables were Vehicle Type (all particle metrics), Instrumentation (particle number and PM2.5), Fuel Type (PM1), Road Type (PM10) and Size Ranged Measured and Speed Limit on the Road (particle volume). The comprehensive set of tailpipe particle emission factors presented in this study are suitable for modelling urban fleets in developed countries, for example, for: • • •

developing road-link based inventories of vehicle fleet emissions; estimating the spatial distribution of particle concentrations; and developing health impact assessments.

They have particular application for regions which may not have sufficient funding to conduct measurement studies, or have limited data available from which to derive emission factors for their local region. The relevance and universal application of these derived emission factors for modelling urban fleet emissions in developed countries is supported by the results of statistical tests conducted in this study which found that:•



few statistically significant differences were found between the mean values for different particle metrics for Country of Study and Study Location (dynamometer, on or near a road, tunnel); nor between dynamometer studies and different Road Types; statistically significant differences were found, however, between LDVs and HDVs for all particle metrics; and between petrol and diesel-fuelled vehicles for PM10.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

90

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al. •

Where statistically significant differences were found these were generally attributed to high speed scenarios, very high levels of resuspended road dust, or due to the influence of higher emission rates from diesel-fuelled vehicles, as compared to petrol and other fuelled vehicles.

The statistical analysis revealed that variation existed between different scientific techniques, indicating that better scientific tools need to be developed for deriving emission factors. An example of this was the statistically significant differences found between the mean values of published emission factors for particle number measured by the Condensation Particle Counter of 22.69 x 1014 particles per vehicle per km, as compared to Scanning Mobility Particle Sizer (SMPS) Instrumentation of 2.08 x 1014 particles per vehicle per km. This significant difference requires further investigation as a broader issue. The extensive literature review undertaken in this study found gaps in our current knowledge about particle emission factors for motor vehicles, specifically related to:•





Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.



Particle number emission factors for buses are rare and restricted to those derived from SMPS measurements, which do not usually measure the lower nucleation size range, where particle number is prolific. More studies are needed to derive emission factors for different Vehicle Types for different subsets of particle number < 1 µm, such as for ultrafine and nanoparticles (diameters < 0.05 µm), where particle number tends to be prolific. Speed-related particle emission factors derived from studies conducted on or near roads and in tunnels, particularly for speeds less than 50 km/hr, are needed to enable modelling of congestion. Few data are available for particle volume, particle surface area, PM1, brake and tyre wear, road grade, engine power, and for buses measured on different Road Types.

The comprehensive set of particle emission factors presented in this study are suitable for developing comprehensive, size-resolved inventories and health impact assessments for motor vehicle fleets, covering the full size range of particles generated by fleets. These estimates are critical for future urban and transport planning, scenario modelling, identification of particulate matter pollution hotspots, as well as for development of air quality regulation in urban areas around the world.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Appendix 1. Source of tailpipe particle emission factors examined in the statistical analysis to derive average emission factors for different Vehicle and Road Type combinations (reference details are listed in Appendix 4)

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

91

92

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Appendix 1. Continued

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

93

Appendix 1. Continued

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

a

1000 nm is equivalent to 1 µm. These units refer to particle diameter. b Instrumentation (in alphabetical order) Aerodynamic Particle Sizer (APS), Berner low pressure Impactor, Beta-ray absorption monitors, Betameter, Chemical Mass Balance, Condensation Particle Counter (CPC), Differential Mobility Analyzer (DMA), Differential Mobility Particle Sizer (DMPS), Dynamometer, DustTrak, Electrical Aerosol Analyser (EAA), Electrical Low Pressure Impactor (ELPI), Filters (Fibreglass, Glass fibre, Teflon, Quartz), Kleinfiltergerate, LIDAR-based VERSS and remote sensing, Mass Single Stage Multidilutor, MOUDI (Micro-Orifice Uniform Deposit Impactor), Samplers (IMPROVE, high volume, medium volume), Scanning Mobility Particle Sizer (SMPS), Tapered Element Oscillating Microbalances(TEOM) and Ultrafine Condensation Particle Counter (UCPC). c Fit log-normal functions to extrapolate concentrations beyond > 220nm. Statistical analysis of in-stack pollution monitoring data and hourly vehicle counts. d Not reported – dynamometer studies which did not provide further information on Instrumentation used. e Vicinity of the road studies refer to studies conducted on or near the road, near a kerb, upwind or downwind of the road. f sm – refers to Size Range Measured and relates to particles with diameters < 1 µm, < 2.5 µm and < 10 µm (known as PM1, PM2.5 and PM10 respectively). g LDV (Light duty vehicles), HDV (Heavy duty vehicles) – refer Appendix 2 for further detail.

Appendix 2. Model variables examined in the statistical analysis to derive average emission factors to use in transport modelling and health impact assessments, to estimate tailpipe particle emissions emitted from motor vehicle fleets Model Variable Name Particle Metric Country of Study Study Location

Model Variable Sub-classes Particle number, particle volume, total particle mass, PM1, PM2.5, PM10 Australia; USA/Canada; Other Countries (Austria, Belgium, Denmark, Germany, Sweden, Switzerland, UK) a Dynamometer (in a laboratory), tunnel or in the vicinity of a road b

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

94

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al. Appendix 2. Continued

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Model Variable Name Road Type Speed Limit on the Road Road Class Average Number of Vehicles Per Day Heavy Duty Vehicle Share Number of HDVs Per Day Vehicle Type Fuel Types Drive Cycles Average Vehicle Speed

Model Variable Sub-classes Boulevard, freeway, highway, motorway, rural area, tunnel, urban c The reported Speed Limit on the Road d Urban and Non-Urban roads; Highway and Non-Highways roads e The average number of vehicles travelling in a vehicle fleet per day f Percentage of HDVs travelling in a vehicle fleet per day g Number of HDVs travelling in a vehicle fleet per day h

Fleet, LDV, HDV, Bus i Diesel, Gasoline, Compressed Natural Gas, Liquefied Natural Gas j Drive Cycles for Buses, Trucks and Other vehicles k Average Vehicle Speed tested on a dynamometer or reported in a tunnel or vicinity of the road study l Engine Power Reported for two bus studies m Instrumentation 20 different types of Instrumentation n Size Range Measured Size Range Measured by Instrumentation o a Groups based on numbers of studies found. b Vicinity of the road - on or near the road, near a curb, upwind, downwind or a road. c Urban Drive Cycle data classed as urban Road Type. d Few studies reported, where reported was Boulevard 82, highway 82 and 100, freeway 100, motorway 120, tunnel 60, 64, 80, 89, urban 50 and 57 km/hr. e Road Class based on either the reported Speed Limit on the Road, or the speed limit that would most likely be associated with the Road Type. < 60 road classed Urban; ≥ 60 nonUrban; ≥ 80 Highway; < 80 km/hr non-Highway. Insufficient data were available to examine individual speeds or other specific speed ranges. f Ranges 13,128-103,080 per day particle number; 23,000-30,000 particle volume; 12,540-12,900 total particle mass; 20,000-69,816 PM1; 20,000-69,816 per day PM10. 5 buses/minute particle number and PM2.5. g Ranges 5-100% particle number, 7-60% particle volume; 1100% total particle mass, PM2.5; 6.1-18% PM1; 2.6-83% for PM10. h Derived where data for both Average Number of Vehicles Per Day and Heavy Duty Vehicle Share (%) were available. i Based on author classifications, including HDV (number of axles, gross vehicle mass or length); LDV (wheel pair distance, vehicle length or weight). LDVs included cars and trucks with specified vehicle weights; and HDVs with gross vehicle mass ranging from 3.5-12 tonne to > 25 tonne. j Few reported diesel fuel sulphur content, where reported was < 15ppm, < 30 ppm Ultralow sulphur diesel (ULSD) HDV; 300ppm Low sulphur diesel (LSD) for Bus, 24-480ppm for LDV and HDV. Diesel, ULSD and LSD classed as diesel Fuel Type. k Buses - Bus Route, Central Bus District, Central Business District – Aggressive Driving, Composite, CUEDC cycle, Manhattan, New York Bus, Orange County Transit Authority, Route 22, Route 77, UDDS and Urban. Other vehicles - CUEDC cycle, FTP, HHDDT; Hot UC, Hot Cycle, Cold Cycle, REP05, Steady State, UC and Urban. Trucks - CBD–CBD14, HDCC. l Ranges < 50, 50-120 particle number, 86-113 particle volume; 80-120 total particle mass; 30-90 PM1; 45-91 PM2.5; < 65 and 45-91 km/hr for PM10. m Engine Power: Reported in two diesel bus studies (Jamriska et al. 2004; Ristovski et al. 2002). n Instrumentation (in alphabetical order) Aerodynamic Particle Sizer, Berner low pressure Impactor, Betameter, Beta-ray absorption monitors, Chemical Mass Balance, Condensation Particle Counter, Differential Mobility Analyzer, Differential Mobility Particle Sizer, DustTrak, Electrical Aerosol Analyser, Electrical Low Pressure Impactor, Filters (Fibreglass, Glass fibre, Teflon, Quartz), Kleinfiltergerate, LIDAR-based VERSS and remote sensing, Mass Single Stage Multidilutor, Micro-Orifice Uniform Deposit Impactor, Samplers, Scanning Mobility Particle Sizer, Tapered Element Oscillating Microbalances, Ultrafine Condensation Particle Counter. o Particle number 0.003-1 µm (dynamometer), 0.01-0.9 µm (tunnel), 0.003-20 µm (vicinity of the road); particle volume 0.018-10 µm. Ranges particle number 0.003-1 µm (dynamometer), 0.01-0.9 µm (tunnel studies), 0.003-20 µm (vicinity of the road), total particle number count > 3 nm; 0.018-10 µm (particle volume). Few size ranges reported in total particle mass studies, where reported 0.008-16 µm (dynamometer), 0.017-0.7 µm (tunnel), 0.008-0.3 µm, > 0.22 µm vicinity of the road.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

95

Appendix 3. Sample size of tailpipe particle emission factors for different model variables examined in the statistical analysis, listed by particle metric

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

a

Country of Study is considered to have limited relevance for dynamometer measurements, except for Urban Drive Cycles, which were classed Urban Road Type (see c below). b Other Countries included studies from Austria, Belgium, Denmark, Germany, Sweden, Switzerland and the United Kingdom. c Within these total Road Type sample sizes, 92 emission factors related to total particle mass, 16 to PM2.5 and 23 to PM10 which were dynamometer measurements using an Urban Drive Cycle. These data were classified in the statistical models as Urban Road Type. d Average Number of Vehicles Per Day and Heavy Duty Vehicle Share sample sizes related to on-road vehicle fleets, and where data was available in studies for both these variables, the additional model variable Number of HDVs Per Day was derived. e Not all studies reported vehicle Fuel Type, particularly studies of on-road vehicle fleets. f Some particle number studies reported only the lower Size Range Measured, such as where total particle count was measured.

Appendix 4. References for international studies listed in Appendix 1 Abu-Allaban, M., 2002. Exhaust particle size distribution measurements at the Tuscarora Mountain tunnel. Aerosol Science and Technology 36(6), 771-789. Abu-Allaban, M., Gillies, J.A., Gertler, A.W., 2003a. Application of a multi-lag regression approach to determine on-road PM10 and PM2.5 emission rates. Atmospheric Environment 37(37), 5157-5164. Abu-Allaban, M., Gillies, J.A., Gertler, A.W., Clayton, R., Proffitt, D., 2003b. Tailpipe, resuspended road dust, and brake-wear emission factors from on-road vehicles. Atmospheric Environment 37(37), 5283-5293. ARB's, 2002. Study of Emissions from Two "Late Model" Diesel and CNG Heavy-Duty Transit Buses. California Air Resources Board, 12th CRC On-Road Vehicle Emissions Workshop, April 15-17, San Diego. Ayala, A., Kado, N.Y., Okamoto, R.A., 2002. Diesel and CNG Heavy-duty Transit Bus Emissions over Multiple Driving Schedules: Regulated Pollutants and Project Overview. Society of Automotive Engineers SAE 2002-01-17221-13.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

96

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Bradley, M.J., 2000. Hybrid-Electric Drive Heavy-Duty Vehicle Testing Project; Final Emissions Report. Northeast Advanced Vehicle Consortium, Defense Advanced Research Projects Agency, West Virginia University, USA. Cadle, S.H., Mulawa, P.A., Ball, J., Donase, C., Weibel, A., Sagebiel, J. C., Knapp, K. T., Snow, R., 1997. Particulate emission rates from in use high emitting vehicles recruited in Orange County, California. Environmental Science and Technology 31(12), 3405-3412. Cadle, S.H., Mulawa, P., Groblicki, P., Laroo, C., Ragazzi, R. A., Nelson, K., Gallagher, G., Zielinska, B., 2001. In-use light-duty gasoline vehicle particulate matter emissions on three driving cycles. Environmental Science and Technology 35(1), 26-32. CARB., 2001. Heavy-Duty Emissions Laboratory, Heavy Duty Testing and Field Support Section, California Air Resources Board. Report No. 01-01. Chatterjee, S., Conway, R., Lanni, T., Frank, B., Tang, S., Rosenblatt, D., Bush, C., Lowell, D., Evans, J., McLean, R., Levy, S., 2002. Performance and Durability Evaluation of Continuously Regenerating Particulate Filters on Diesel Powered Urban Buses at NY City Transit - Part II. Society of Automotive Engineers SAE 2002-01-0430. Clark, N.N., Lyons, D.W., Bata, R.M., Gautam, M., Wang, W.G., Norton, P., Chandler, K., 1997. Natural Gas and Diesel Transit Bus Emissions: Review and Recent Data. Society of Automotive Engineers Tech. Pap. No. 973203. Clark, N.N., Lyons, D.W., Rapp, B.L., Gautam, M., Wang, W.G., Norton, P., White, C., Chandler, C., 1998. Emissions from Trucks and Buses Powered by Cummins L-10 Natural Gas Engines. Society of Automotive Engineers Tech. Pap. No. 981393. Clark, N.N., Gautam, M., Rapp, B.L., Lyons, D.W., Graboski, M.S., McCormick, R. L., Alleman, T. L., Norton, P., 1999. Diesel and CNG Transit Bus Emissions Characterization by Two Chassis Dynamometer Laboratories: Results and Issues. Society of Automotive Engineers SAE 1999-01-1469. CONCAWE., 1998. A study of the number, size and mass of exhaust particles emitted from european diesel and gasoline vehicles under steady-state and european driving cycle conditions. CONCAWE, Brussels Report no. 98/51. Corsmeier, U., Imhof, D., Kohler, M., Kuhlwein, J., Kurtenbach, R., Petrea, M., Rosenbohm, E., Vogel, B., Vogt, U., 2005. Comparison of measured and model-calculated real-world traffic emissions. Atmospheric Environment 39(31), 5760-5775. DOEH., 2003. Technical Report No. 1: Toxic Emissions from Diesel Vehicles in Australia, Department of the Environment and Heritage, Canberra. Gehrig, R., Hill, M., Buchmann, B., Imhof, D., Weingartner, E., Baltensperger, U., 2004. Separate determination of PM10 emission factors of road traffic for tailpipe emissions and emissions from abrasion and resuspension processes. International Journal of Environment and Pollution 22(3), 312-325. Gertler, A.W., Gillies, J.A., Pierson, W.R., Rogers, C.F., Sagebiel, J. C., Abu-Allaban, M., Coulombe, W., Tarnay, L., Cahill, T.A., 2002. Real-World Particulate Matter and Gaseous Emissions from Motor Vehicles in a Highway Tunnel. Health Effects Institute Research Report 107. Gidhagen, L., Johansson, C., Strom, J., Kristensson, A., Swietlicki, E., Pirjola, L., Hansson, H.C., 2003. Model simulation of ultrafine particles inside a road tunnel. Atmospheric Environment 37(15), 2023-2036.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

97

Gidhagen, L., Johansson, C., Langner, J., Olivares, G., 2004a. Simulation of NOx and ultrafine particles in a street canyon in Stockholm, Sweden. Atmospheric Environment 38(14), 2029-2044. Gidhagen, L., Johansson, C., Omstedt, G., Langner, J., Olivares, G., 2004b. Model simulations of NOx and ultrafine particles close to a Swedish highway. Environmental Science and Technology 38(24), 6730-6740. Gillies, J.A., Gertler, A.W., Sagebiel, J.C., Dippel, W.A., 2001. On-road particulate matter (PM2.5 and PM10) emissions in the Sepulveda Tunnel, Los Angeles, California. Environmental Science and Technology 35(6), 1054-1063. Gramotnev, G., Brown, R., Ristovski, Z., Hitchins, J., Morawska, L., 2003. Determination of average emission factors for vehicles on a busy road. Atmospheric Environment 37(4), 465-474. Gramotnev, G., Ristovski, Z.D., Brown, R.J., Madl, P., 2004. New methods of determination of average particle emission factors for two groups of vehicles on a busy road. Atmospheric Environment 38(16), 2607-2610. Hibberd, M.F., 2005. Vehicle NOx and PM10 Emission Factors from Sydney's M5-East Tunnel. 17th International Clean Air and Environment Conference proceedings, Hobart. Clean Air Society of Australia and New Zealand. Holmen, B., Chen, Z., Davila, A., Gao, O., Vikara, D.M., 2005. Particulate matter emissions from Hybrid Diesel-electric and Conventional Diesel Transit Buses: Fuel and Aftertreatment Effects. The University of Connecticut Report No. JHR 05-304. Hueglin, C., Buchmann, B., Weber, R. O., 2006. Long-term observation of real-world road traffic emission factors on a motorway in Switzerland. Atmospheric Environment 40(20), 3696-3709. Imhof, D., Weingartner, E., Ordonez, C., Gehrigt, R., Hill, N., Buchmann, B., Baltensperger, U., 2005a. Real-world emission factors of fine and ultrafine aerosol particles for different traffic situations in Switzerland. Environmental Science and Technology 39(21), 83418350. Imhof, D., Weingartner, E., Prevot, A., Ordonez, C., Kurtenbach, R., Wiesen, P., Rodler, J., Sturm, P., McCrae, I., Sjodin, A., Baltersperger, U., 2005b. Aerosol and NOx Emission Factors and Submicron Particle Number Size Distributions in Two Road Tunnels with Different Traffic Regimes. Atmospheric Chemistry and Physics Discussions 5512755166. Imhof, D., Weingartner, E., Vogt, U., Dreiseidler, A., Rosenbohm, E., Scheer, V., Vogt, R., Nielsen, O.J., Kurtenbach, R., Corsmeier, U., Kohler, M., Baltensperger, U., 2005c. Vertical distribution of aerosol particles and NOx close to a motorway. Atmospheric Environment 39(31), 5710-5721. Jamriska, M., Morawska, L., 2001. A model for determination of motor vehicle emission factors from on-road measurements with a focus on submicrometer particles. Science of the Total Environment 264(3), 241-255. Jamriska, M., Morawska, L., Thomas, S., Congrong, H., 2004. Diesel Bus Emissions Measured in a Tunnel Study. Environmental Science and Technology 38(24), 6701-6709. Jones, A.M., Harrison, R.M. 2006. Estimation of the emission factors of particle number and mass fractions from traffic at a site where mean vehicle speeds vary over short distances. Atmospheric Environment 40(37), 7125-7137.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

98

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Kado, N.Y., Okamoto, R.A., Kuzmicky, P.A., Kobayashi, R., Ayala, A., Gebel, M. E., Rieger, P.L., Maddox, C., Zafonte, L., 2005. Emissions of toxic pollutants from compressed natural gas and low sulfur diesel-fueled heavy-duty transit buses tested over multiple driving cycles. Environmental Science and Technology 39(19), 7638-7649. Ketzel, M., Wahlin, P., Berkowicz, R., Palmgren, F., 2003. Particle and trace gas emission factors under urban driving conditions in Copenhagen based on street and roof-level observations. Atmospheric Environment 37(20), 2735-2749. Kittelson, D.B., Watts, W.F., Johnson, J.P., 2004. Nanoparticle emissions on Minnesota highways. Atmospheric Environment 38(1), 9-19. Kristensson, A., Johansson, C., Westerholm, R., Swietlicki, E., Gidhagen, L., Wideqvist, U., Vesely, V., 2004. Real-world traffic emission factors of gases and particles measured in a road tunnel in Stockholm, Sweden. Atmospheric Environment 38(5), 657-673. Lanni, T., Frank, B. P., Tang, S., Rosenblatt, D., Lowell, D., 2003. Performance and Emissions Evaluation of Compressed Natural Gas and Clean Diesel Buses at New York City's Metropolitan Transit Authority. SAE 2003-01-0300. Lowell, D.M., Parsley, W., Bush, C., Zupo, D., 2003. Comparison of Clean Diesel buses to CNG Buses. 9th Diesel Engine Emissions Reduction (DEER) Workshop, Newport, RI, USA, 24-28 August. Mazzoleni, C., Kuhns, H.D., Moosmuller, H., Keislar, R.E., Barber, P.W., Robinson, N. F., Watson, J.G., 2004. On-road vehicle particulate matter and gaseous emission distributions in Las Vegas, Nevada, compared with other areas. Journal of the Air and Waste Management Association 54(6), 711-726. Morawska, L., Bofinger, N.D., Kocis, L., Nwankwoala, A., 1998. Submicrometer and supermicrometer particles from diesel vehicle emissions. Environmental Science and Technology 32(14), 2033-2042. Morawska, L., Ristovski, Z., Ayoko, G.A., Jayaratne, E.R., Lim, M., 2001. Report of a short investigation of emissions from diesel vehicles operating on low and ultralow sulphur content fuel. Prepared for BP Australia by Queensland University of Technology, Brisbane. Morawska, L., Jamriska, M., Thomas, S., Ferreira, L., Mengersen, K., Wraith, D., McGregor, F., 2005. Quantification of particle number emission factors for motor vehicles from onroad measurements. Environmental Science and Technology 39(23), 9130-9139. Ristovski, Z.D., Morawska, L., Ayoko, G.A., Jayaratne, E.R., Lim, M., 2002. Final report of a comparative investigation of particle and gaseous emissions from twelve in-service B.C.C. buses operating on 50 and 500 ppm sulphur diesel fuel. Queensland University of Technology, Brisbane. Romilly, P., 1999. Substitution of bus for car travel in urban Britain: an economic evaluation of bus and car exhaust emission and other costs. Transportation Research Part DTransport and Environment 4(2), 109-125. SAE., 2001. Performance and Durability Evaluation of Continuously Regenerating Particulate Filters on Diesel powered Urban Transit Buses at NY City Transit. Society of Automotive Engineers SAE 2001-01-0511. SAE., 2002a. Performance and Durability of Continuously Regenerating Particulate Filters on Diesel powered Urban Transit Buses at NY City Transit - Part II. Society of Automotive Engineers SAE 2002-01-0430.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

99

SAE., 2002b. Year-Long Evaluation of Trucks and Buses Equipped with Passive Diesel Diesel Particulate Filters. Society of Automotive Engineers SAE 2002-01-0433. SAE., 2003a. Oxidation catalyst effect on CBG Transit Bus Emissions. Society of Automotive Engineers SAE 2003-01-1900. SAE., 2003b. Performance and Emissions Evaluation of Compressed Natural Gas and Clean Diesel Buses at New York City's Metropolitan Transit Authority. Society of Automotive Engineers SAE 2003-01-0300. Schmid, H., Pucher, E., Ellinger, R., Biebl, P., Puxbaum, H., 2001. Decadal reductions of traffic emissions on a transit route in Austria - results of the Tauerntunnel experiment 1997. Atmospheric Environment 35(21), 3585-3593. Shah, S.D., Cocker, D.R., Miller, J.W., Norbeck, J.M., 2004. Emission rates of particulate matter and elemental and organic carbon from in-use diesel engines. Environmental Science and Technology 38(9), 2544-2550. Tran, T. V., Ng, Y. L., Denison, L., 2003. Emission Factors for In-Service Vehicles Using Citylink Tunnel. Proceedings of the National Clean Air Conference, Newcastle. Ubanwa, B., Burnette, A., Kishan, S., Fritz, S.G., 2003. Exhaust particulate matter emission factors and deterioration rate for in-use motor vehicles. Journal of Engineering for Gas Turbines and Power-Transactions of the Asme 125(2), 513-523. Venkatram, A., Fitz, D., Bumiller, K., Du, S.M., Boeck, M., Ganguly, C., 1999. Using a dispersion model to estimate emission rates of particulate matter from paved roads. Atmospheric Environment 33(7), 1093-1102. Wayne, W.S., Clark, N.N., Nine, R.D., Elefante, D., 2004. A comparison of emissions and fuel economy from hybrid-electric and conventional-drive transit buses. Energy and Fuels 18(1), 257-270. Zhang, K.M., Wexler, A.S., Niemeier, D.A., Zhu, Y.F., Hinds, W. C., Sioutas, C., 2005. Evolution of particle number distribution near roadways. Part III: Traffic, analysis and on-road size resolved particulate emission factors. Atmospheric Environment 39(22), 4155-4166. Zhu, Y. F., Hinds, W. C., 2005. Predicting particle number concentrations near a highway based on vertical concentration profile. Atmospheric Environment 39(8), 1557-1566.

References Abu-Allaban, M., Gillies, J.A., Gertler, A.W., 2003. Application of a multi-lag regression approach to determine on-road PM10 and PM2.5 emission rates. Atmospheric Environment 37(37), 5157-5164. Ahlvik, P., Eggleston, S., Goriben, N., Hassel, D., Hickman, A. J., Joumard, R., Ntziachristos, L., Rijkeboer, R., Samaras, Z., Zierock, K. H., 1997. COPERT II Computer programme to calculate emissions from road transport: methodology and emission factors. Technical report prepared by the European Environment Agency, Copenhagen. Report No. 6. Bellasio, R., Bianconi, R., Corda, G., Cucca, P., 2007. Emission inventory for the road transport sector in Sardinia (Italy). Atmospheric Environment 41, 677-691. CARB., 2002. EMFAC2001/EMFAC200. Calculating emissions inventories for vehicles in California, User’s Guide, California California Air Resources Board.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

100

Diane U. Keogh, Joe Kelly, Kerrie Mengersen et al.

Frey, H.C., Unal, A., Chen, J., Li, S., Xuan, C., 2002a. Methodology for developing modal emission rates for EPA's multi-scale motor vehicle and equipment emission estimation system, North Carolina State University for the Office of Transportation and Air Quality, US Environmental Protection Agency. Frey, H.C., Unal, A., Chen, J., 2002b. Recommended strategy for on-board emission data analysis and collection for the new generation model. Prepared for Office of Transportation and Air Quality, US Environmental Protection Agency. Goodwin, J. W. L., Salway, A. G., Eggleston, H. S., Murrells, T. P., Berry, J.E., 1999. National Atmospheric Emissions Inventory, UK Emissions of Air Pollutants 1970 to 1996, National Environmental Technology Centre on behalf of the Department of the Environment, Transport and the Regions. Harrison, R., Jones, M., Collins, G., 1999. Measurements of the Physical Properties of Particles in the Urban Atmosphere. Atmospheric Environment 33, 309-321. Imhof, D., Weingartner, E., Ordonez, C., Gehrigt, R., Hill, N., Buchmann, B., Baltensperger, U., 2005. Real-world emission factors of fine and ultrafine aerosol particles for different traffic situations in Switzerland. Environmental Science and Technology 39(21), 83418350. Jones, A.M., Harrison, R.M. 2006. Estimation of the emission factors of particle number and mass fractions from traffic at a site where mean vehicle speeds vary over short distances. Atmospheric Environment 40(37), 7125-7137. Kittelson, D. B., Watts, W. F.Johnson, J. P. 2004. Nanoparticle emissions on Minnesota highways. Atmospheric Environment 38(1), 9-19. Morawska, L., Salthammer, T., 2003. Chapter 3: Motor Vehicle Emissions as a Source of Indoor Particles in, Morawska-Salthammer (eds). Indoor Environment, Wiley-VCH297318. Morawska, L., Bofinger, N. D., Kocis, L., Nwankwoala, A., 1998. Submicrometer and supermicrometer particles from diesel vehicle emissions. Environmental Science and Technology 32(14), 2033-2042. Morawska, L., Keogh, D.U., Thomas, S.B., Mengersen, K., 2008. Modality in ambient particle size distributions and its potential as a basis for developing air quality regulation. Atmospheric Environment 42 (7), 1617-1628. Morawska, L., Moore, M. R., Ristovski, Z.D., 2004. Health Impacts of Ultrafine Particles Desktop Literature Review and Analysis, Department of the Environment and Heritage, September, Canberra. Ntziachristos, L., Samaras, Z., Eggleston, S., Goriben, N., Hassel, D., Hickman, A. J., Joumard, R., Rijkeboer, R., White, L., Zierock, K. H., 2000. COPERT III Computer programme to calculate emissions from road transport: methodology and emission factors (version 2.1). Technical report prepared by the European Environment Agency, Copenhagen, Report 49. Samaras, Z., Ntziachristos, L., Thompson, N., Hall, D., Westerholm, R., Boulter, P., 2005. Characterisation of Exhaust Particulate Emissions from Road Vehicles, PARTICULATES program, European Commission. Contract No 2000-RD.11091, source http://lat.eng.auth.gr/particulates/downloads.htm. Scheffe, H., 1959. The Analysis of Variance, John Wiley and Sons, Inc. Shi, J., Harrison, R.M., 1999. Investigation of ultrafine particle formation during diesel exhaust dilution. Environmental Science and Technology 33, 3730-3736.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Tailpipe Particle Emission Factors Derived for Motor Vehicles for Application…

101

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Shi, J. P., Khan, A. A., Harrison, R.M., 1999. Measurements of ultrafine particle concentration and size distribution in the urban atmosphere. The Science of the Total Environment 235, 51-64. Shi, J., Evans, D., Khan, A., Harrison, R., 2001. Sources and Concentration of Nanoparticles (< 10 nm Diameter) in the Urban Atmosphere. Atmospheric Environment 35, 1193-1202. Shifter, I., Diaz, L., Mugica, V., Lopez-Salinas, E., 2005. Fuel-based motor vehicle emission inventory for the metropolitan area of Mexico city. Atmospheric Environment 39(5), 931940. Smit, R., Smokers, R., Rabe, E., 2007. A new modelling approach for road traffic emissions: VERSIT+. Transportation Research Part D-Transport and Environment 12, 414-422. USEPA., 1993. User's Guide to MOBILE5A, Mobile source emissions factor model, U.S. Environmental Protection Agency. Wahlin, P., Palmgren, F., Van Dingenen, R., 2001. Experimental studies of ultrafine particles in streets and the relationship to traffic. Atmospheric Environment 35, S63-S69. Walker, J. L., Li, J., Srinivasan, S., Bolduc, D. 2008. Travel Demand Models in the Developed World: Correcting for Measurement Errors Transportation Research Board 87th Annual Meeting Washington.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 103-119

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 4

SNOWMOBILE POLLUTION IN NORTH AMERICA: ANNUAL FLUX ESTIMATES OF AIR TOXICS AND IMPLICATIONS FOR POTENTIAL PERSONAL EXPOSURE IN SNOWMOBILE DOMINATED COMMUNITIES David Shively1, Yong Zhou2 and Barkley Sive2

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1

Department of Geography, The University of Montana, Missoula, MT 59812 2 Climate Change Research Center, Institute for the Study of Earth, Oceans and Space, University of New Hampshire, Durham, NH 03824

Abstract Yellowstone National Park’s (YNP) winter use planning and oversnow vehicle use issues have spawned the most intensive studies of snowmobile emissions and associated ambient air quality anywhere. These issues and the research they have engendered have served to improve our understanding of the dynamics of snowmobile use in winter landscapes and the attending exposure of people to air toxics, and can be seen as having contributed to the development of cleaner engine technologies and policies designed to improve the personal safety and health of those who might be exposed to snowmobile emissions. In this chapter we examine the results of studies conducted in YNP prior to the promulgation of policies mandating the use of best available technology (BAT) equipped snowmobiles (i.e., cleaner burning 4-stroke machines), and existing data concerning snowmobile use in North America (i.e., the United States and Canada) to estimate annual fluxes of air toxics (benzene, toluene, ethyl benzene, xylenes, and n-hexane) for this region. While we do not attempt to present our results in a spatially disaggregated manner beyond national units, our results show that annual air toxic emissions from snowmobile use appear to be significantly higher (i.e., ~16-29%) than the USEPA’s estimates of air toxics emitted by snowmobiles in the United States. Additionally, we find that emissions associated with Canadian snowmobile use are likely to increase the U.S. figures. The implications of these fluxes are discussed in the context of agency air toxic assessment programs, and personal exposure to air toxics using data pertaining to the community of West Yellowstone, Montana. Our results provide an initial set of baseline data for potential personal exposure in this small mountain community to ambient

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

104

David Shively, Yong Zhou and Barkley Sive air toxics coming from this non-regulated mobile source. Lastly, these data are discussed in the context of other North American communities in which snowmobiles represent an important component of the motorized vehicle fleet.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Introduction Coming from Yellowstone National Park’s (YNP) winter use planning activities in the late 1990s and early years of the twenty-first century are the most intensive studies of snowmobile emissions and associated ambient air quality anywhere. These studies focused on pollutants that have implications for human health, and for visibility in a designated Class I airshed. Air toxics (also referred to as toxic air pollutants, or hazardous air pollutants or HAPs) “are those pollutants that are known or suspected to cause cancer or other serious health effects, such as reproductive effects or birth defects, or adverse environmental effects” [USEPA 2008a]. One hundred and eighty-seven different HAPs are listed on the U.S. Clean Air Act Amendments of the 1990 list of HAPs [USEPA 2007]. The United States’ Environmental Protection Agency (USEPA) utilizes a comprehensive program, the National Air Toxics Assessment (NATA) to estimate the levels and implications of HAPs for human health in the United States. Inclusive within this program is the National Emissions Inventory (NEI), which estimates the emission into the atmosphere of air toxics and other compounds. Specifically, the assessments accomplish the following: (1) they “estimate the risk of cancer and other serious health effects associated with breathing (inhaling) air toxics;” and (2) they generate “estimates of cancer and non-cancer health effects based on chronic exposure from outdoor sources” [USEPA 2008b]. Canada, through its national agency Environment Canada, has a National Pollutants Release Inventory (NPRI) program that is similar, to the USEPA NATA program [Environment Canada 2008a]. Important in these assessments are air toxic emission inventories for pollutants coming from point and non-point sources; for the USEPA, point sources include those activities that require permits to discharge pollutants into the atmosphere from state agencies (i.e., coal-fired thermoelectric plants, coke ovens, dry cleaning facilities, etc.), and non-point sources include mobile (i.e., onroad and nonroad) and non-permitted diffuse source (i.e., biomass burning) estimates obtained by the USEPA and state, local, and tribal (S/L/T) governments [Driver 2004]. Environment Canada takes a similar approach. It is important to note that the inventories and associated assessments are based largely on modeling rather than on monitoring or in situ measurements of ambient air quality, and that many assumptions and simplifications are necessarily incorporated into the models that are utilized [USEPA 2008c]. Snowmobiles are among the nonroad vehicle types considered by the USEPA in its assessments; others include farm/construction/mining equipment, and recreational vehicles such as all terrain vehicles, personal watercraft, and motorboats. Canada does not include snowmobiles or any other mobile sources in their NPRI [Environment Canada 2008b]. The models used by the U.S. (i.e., NONROAD) and Canadian agencies in their inventories, however, model only a few pollutants. The USEPA models a number of different hydrocarbons (HCs) or volatile organic compounds (VOCs), HAPs that are VOCs, NOX (oxides of nitrogen), carbon monoxide (CO), carbon dioxide (CO2), sulfur oxides (SOX), and PM (particulate matter) [USEPA 2005]; Environment Canada models three different classes of PM, CO, NOx, SOx, HC, and ammonia (NH3) [Environment Canada 2008b]. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Snowmobile Pollution in North America

105

In this chapter, we attempt to gauge the modeling efforts of the USEPA in particular, since it does model snowmobile air toxic emissions, by presenting the results of our own modeling of U.S. snowmobile air toxic emissions based on measurements of volatile organic compounds (VOCs) made in the field in Yellowstone National Park (YNP, or the Park) and its gateway community of West Yellowstone, Montana. Furthermore, we extrapolate our findings to Canada to develop a continental scale assessment of air toxics coming from snowmobiles. The measurements on which our findings are based were obtained during a 2003 field campaign that was designed to assess the spatiotemporal variation in ambient air quality in the Park during periods of heavy oversnow vehicle (i.e., snowmobiles and snowcoaches) use. That research [Sive et al. 2002; Sive et al. 2003; Shively et al. 2008; Zhou et al. Submitted], which is described further below in the methods section and elsewhere in this chapter, complemented other research that focused on direct emissions of VOCs (as well as CO and PM) and ambient air at a limited number of locations in the Park and implications for human health. All of those research efforts (as well as ongoing studies) contributed to Yellowstone National Park’s (YNP) winter use planning activities which have been ongoing since 1968, and particularly intense since 1997 [NPS 2008a], and represent the most intensive studies of snowmobile emissions and associated ambient air quality anywhere in the world. Furthermore, they have served to improve our understanding of the dynamics of snowmobile use in winter landscapes and the attending exposure of people to air toxics, and can be seen as having contributed to the development of cleaner engine technologies and policies designed to improve the personal safety and health of those who might be exposed to snowmobile emissions.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Background Snowmobile Use in North America Snowmobiles are used for recreational, occupational, and subsistence purposes in Canada and the U.S.; most use occurs in the Rocky Mountain and northern tier states of the U.S., and throughout Canada [ISMA 2008]. Actual levels of snowmobile use in North America’s landscapes and communities are difficult to ascertain owing to a number of factors. First, the number of hours spent riding is likely to differ considerably among different user groups. Occupational and subsistence-based users are likely to spend much more time riding and in the vicinity of running or idling machines than recreational users. Second, while most U.S. states require snowmobiles to be registered if they are to be used, registration requirements vary among them. For example, Alaska doesn’t require snowmobiles used exclusively on private property to be registered [DMV.ORG 2008], and Montana snowmobiles need only be registered once unless they are sold to a new owner [MTDFWP 2008]. And last, it is quite likely that many snowmobile users in more remote communities never register their vehicles, though this cannot be confirmed. Despite these difficulties, snowmobile registration data provide the best means to estimate snowmobile use. Another point that bears on the issue of snowmobiles and human exposure to air toxics is the composition of the North American snowmobile fleet. Both 2-stroke and 4-stroke machines are manufactured, sold, and utilized today, however, four-stroke snowmobiles were very uncommon before 2000. It was the YNP snowmobile war (still in progress) that

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

106

David Shively, Yong Zhou and Barkley Sive

prompted the four major snowmobile manufacturers to begin mass-producing 4-stroke machines that are billed as being more environmentally friendly than earlier models or existing 2-stroke technology [MTDEQ 2008]; 2-stroke machines produce much more pollution than 4-stroke machines [Sive et al. 2003; MTDEQ 2008; Shively et al. 2008; Zhou et al. Submitted]. Despite this fact and owing to their generally superior performance, the majority of new snowmobile models introduced by the manufacturers use 2-stroke technology. Data that differentiate between the sales and use of 2-stroke and 4-stroke machines in the U.S. or North America in general are hard to come by, but Williford [2001] estimated that some 1.8 million 2-stroke snowmobiles were in use in the U.S. in 2001; though this estimate essentially predates the stronger manufacturing and marketing efforts corresponding to 4-stroke machines, it is assumed that the majority of snowmobiles in use in North America today are 2-stroke models. Importantly, the USEPA hasn’t yet set emission standards for snowmobiles or other off highway vehicles as it has for highway motorcycles.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Air Toxics and Human Health Risks There are various environmental and health impacts associated with snowmobile use including air quality impairment from exhaust emissions [Sive et al. 2003; Shively et al. 2008], stress to wildlife [Creel et al. 2002; Andersen and Aars 2008], human exposure to air toxics [NPS 2000], and trauma and mortality associated with accidents [Aiken 2003; Hoey 2003; CDC 2004; Stewart and Black 2004; Nayci et al. 2006]. Exhaust from snowmobiles, especially those with 2-stroke engines, contains numerous air toxics including benzene, toluene, ethylbenzene, xylenes (p-, m- and o-xylene), and n-hexane, all of which are USEPA classified air toxics. These toxic compounds are present in gasoline and are released to the atmosphere via evaporation and/or by passing through the engine as unburned fuel. Most air toxics have been identified through toxicological laboratory experiments in which animals are exposed to very high levels of the compound being studied, and while humans almost never encounter such levels, lower exposure levels still have the potential to pose severe health risks including numerous respiratory and neurological effects [Sive et al. 2003]. Exposure to benzene, toluene, ethlybenzene, xylenes, and n-hexane may lead to eye irritation, irritation of the upper respiratory tract, dizziness, headaches, and respiratory arrest [USEPA 2007b]. Acute and chronic exposures to toluene can result in central nervous system depression and loss of memory [Irwin et al. 1998]. Benzene has been classified as a human carcinogen by both the USEPA and the International Agency for Research on Cancer (IARC) based on evidence from epidemiological studies [IARC 1987; USEPA 1994]. Additionally, occupational exposure to benzene has been linked to the increased incidence of leukemia in humans.. Based on the presence of air toxics in snowmobile exhaust, and considering the quantities in which they are emitted, the cumulative negative health effects of exposure to highly concentrated, multi-pollutant laden snowmobile exhaust are likely to be significant. This is particularly relevant for those individuals who spend a considerable amount of time in the vicinity of snowmobile exhaust, whether in areas in which snowmobiles are used primarily for recreation (e.g., YNP and the winter landscapes of the northern U.S. and Canada) or in and around remote northern communities in which snowmobiles are a primary means of transportation. Of concern are areas with significant traffic congestion (e.g., YNP entrance stations and parking lots), and locations having relatively high levels of snowmobile use that are also characterized by poor emission dispersion.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Snowmobile Pollution in North America

107

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Aside from the health issues associated with snowmobile accidents, the longer-term health effects of snowmobile emissions have not been much studied. Studies that have focused on personal exposure to air toxics coming from snowmobile emissions were produced in conjunction with the YNP Winter Use Planning efforts of the National Park Service beginning in about 2000. Prior to the regulation of snowmobile use in YNP by the NPS, which resulted in limits on the number of entries and exclusive use of cleaner 4-stroke machines, Kado et al. [2001] found that NPS employee exposure levels to air toxics at sites characterized by close proximity to high levels of snowmobile traffic were measurable, were highest for toluene, and that these could sometimes approach the concentration of benzene for some employees could approach the Recommended Exposure Levels (RELs) as established by the National Institute for Occupational Safety and Health (NIOSH). Their results indicated that levels of individual air toxics, including carcinogens such as benzene, resulting from snowmobile exhaust, can be high enough to be a threat to human health. More recently, during the 2004-05 YNP winter use season that was characterized by the lowest snowmobile use levels (4-stroke machines only) since the rise in the popularity of such activity in the 1960s (a maximum of 279 machines were observed to have entered YNP during their sampling days, which is nearly an order of magnitude less that of the maxima of the unregulated years as noted in the NPS’s 1999 State of the Park Report), Spear and Stevenson [2005] measured air toxic exposure levels that were well below regulatory (OSHA) or recommended levels (NIOSH and the Agency for Toxic Substances and Disease Registry or ATSDR) at similar sites as used by Kado et al. [2001]. This study demonstrated that the adoption of regulations that would drastically reduce snowmobile use in the Park and restrict this to machines that are equipped with best available technology (i.e., cleaner 4-stroke models) can significantly reduce the potential health risks associated with long-term exposure of Park employees and nearby communities to snowmobile emissions.

US and Canadian National Air Toxics Assessments The air toxic assessments conducted by the USEPA include four steps: (1) the compilation of “a national emissions inventory of air toxics emissions from outdoor sources; (2) estimating ambient concentrations of air toxics across the United States; (3) estimating population exposures across the United States; and (4) characterizing potential public health risk due to inhalation” [USEPA 2008d]. To date, assessments were completed in 1996, 1999, and there is an ongoing assessment that commenced in 2002 and will be based on the emissions inventory completed that year (the inventories are completed every three years). Population risk assessments are estimated for census tracts. The emissions inventories included data for 33 different pollutants in 1996, and 177 in 1999; the numbers for the 2002 and 2005 inventories have not yet been disseminated but should not exceed the 187 pollutants listed on the Clean Air Act Amendments of the 1990 list of HAPs [USEPA 2007a] plus diesel particulate matter. Canada’s NPRI program is broader than the USEPA’s NATA program in that it inventories pollutants in both air and water, and includes substances that are criteria pollutants, toxic substances, and those which are not regarded as toxic substances [Environment Canada 2008b]. As is the case for the inventorying of pollutants in the U.S., Canada’s inventories rely on self-reporting for point sources and modeling of non-point

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

108

David Shively, Yong Zhou and Barkley Sive

sources. Interestingly, Canada does not include snowmobiles as a discrete recreational vehicle class in its inventories, while it does include activities such as cigarette smoking, bakeries, residential fuel-wood combustion, and even meat cooking [Environment Canada 2008c].

Methods Experimental

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The 2003 campaign that produced the data used in this work was broadly configured to study the spatial and temporal variability of 87 different VOCs (including a number of important air toxics) in and around YNP during periods of moderate and high snowmobile use. The results of that study are reported in Sive et al. [2003], Shively et al. [2008], and Zhou et al. [Submitted].

a) Figure 1. Continued. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Snowmobile Pollution in North America

109

b)

Figure 1. Sample sites within and around a) Yellowstone National Park and b) West Yellowstone, MT.

In assessing snowmobile effects on ambient air quality, we relied on mixed methods including: (1) the acquisition and analysis of many whole-air samples from systematically located sites in the Park (Figure 1) on a moderate use day (12 February 2003) and a high use day (15 February 2003); (2) the acquisition and analysis of random exhaust samples from four 2-stroke snowmobiles and three 4-stroke machines (Table 1); (3) the acquisition and analysis of 3 random exhaust samples from other oversnow vehicles (i.e., two snowcoaches and a YNP snowcat/groomer) (Table 1); (4) upper air soundings acquired using radiosonde equipped 2 meter diameter meteorological balloons to assess vertical stability of the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

110

David Shively, Yong Zhou and Barkley Sive

atmosphere and potential for constraints on vertical mixing and air transport; (5) alkyl nitrate/parent alkane ratios and ratios of trace gases to benzene (especially toluene/benzene ratios) to analyze air mass ages; and (6) spatial analysis of whole air sample data. To assess VOC production and/or transport outside the Park to nearby communities, whole-air samples were taken near Pahaska Teepee (a very small settlement with limited access) and also in and around the gateway community of West Yellowstone. West Yellowstone, in particular, we regard as being quite representative of snowmobile dependant communities. Table 1. Oversnow vehicles engine conditions for which exhaust samples were collected

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Sample Date 2/14/03 2/16/03 2/14/03 2/14/03 2/14/03 2/15/03 2/16/03 2/14/03 2/16/03 2/15/03

Sample Information 2001 Polaris 550 fan cooled (Park Ranger), idle, hot engine Polaris 550 Fan Cooled Super Sport, idle, hot engine 1998 Summit X Ski-doo, water cooled, 3500 RPM 2003 Polaris 500 Classic Touring Liquid Cooled, 3500 RPM 1978 Bombadier snowcoach w/ 350 Chevy, idle 2002 Polaris Fuel Injected 4-stroke; 3500 RPM Polaris Frontier 4 Stroke Edge 136, idle, hot engine Artic Cat 660, sample @ 3500 RPM NPS Ford Power Stroke V-8 Sno-Van, 4000 RPM Lake R.S. Snowcat (Groomer), idle

Engine Type 2-Stroke 2-Stroke 2-Stroke 2-Stroke 4-Stroke 4-Stroke 4-Stroke 4-Stroke Diesel Diesel

Billing itself as “the snowmobile capital of the world,” West Yellowstone, Montana, has relatively high levels of snowmobile use that are associated with the adjacent West Entrance to YNP (through which the majority of wintertime visitors to the Park pass), with the town’s location relative to other snowmobiling opportunities on nearby public lands, and because the majority of its streets are open to use by snowmobiles during winter months. Additional samples were taken under, upwind, and downwind of the YNP West Entrance roof enclosure to assess the effect of vehicle entry and idling in that area on ambient conditions (these samples included VOC contributions from West Yellowstone). To gain an understanding of the diurnal trends in VOCs and other compounds (including oxides of nitrogen or NOx, ozone or O3, and particulate matter or condensation nuclei – CN) we collected hourly samples over 24 hour periods on both sampling days at Lake Ranger Station (Lake R.S.). The station is situated near the middle of the park, is quite distant from the West Yellowstone-Old Faithful corridor which is the busiest part of the Park during the winter season, and receives low to moderate levels of traffic on the nearby road (outside of the Canyon to Tower road segment which is closed to all vehicle use and the North Entrance to Cooke City road segment which is open to wheeled vehicle use, YNP roads are groomed for oversnow vehicle use only). Four samples were collected at each sample site each day, the first set of two during the period before dawn and the second at or near solar noon. Both offroad (i.e., approximately ½ kilometer upwind of any road) and near-road (i.e., approximately 50 m upwind of any road) samples were acquired at each site during each sampling period. Diurnal samples from the Lake R.S. were timed to correspond to the corresponding to the other ambient samples collected across the Park. The geographic positions of sample sites

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Snowmobile Pollution in North America

111

were determined with GPS devices, and local environmental conditions (i.e., wind direction and velocity, cloud cover, etc.) were recorded with these devices. The exhaust samples were collected to characterize the emission signatures, with respect to the relative emissions of different VOCs, of the various oversnow vehicles used in the Park. These were acquired by placing canisters directly in the exhaust streams of the vehicles that were running at normal operating temperature, and filling them to ambient pressure. The majority of snowmobiles had just completed a transit from West Yellowstone to Old Faithful, and these were sampled in the Old Faithful parking lot. A portion of the snowmobiles were sampled while having their engines revved to 3500 revolutions per minute (RPM), to mimic the snowmobiles’ estimated average RPM during transit. Although engine loading is not necessarily accurately simulated in this fashion, vehicles are restricted to a maximum speed of 35 mph in this corridor and it is characterized by relatively level terrain in all but one section. Other samples were collected from snowmobiles that had just arrived at the sampling location, but were idling, in order to characterize exhaust emissions from machines that have been pulled over and stopped by their riders in order to view scenery and/or wildlife. The exhaust samples were collected in 1-liter silica-lined canisters, and the ambient (whole air) samples in 2-liter electropolished stainless-steel canisters. All samples were analyzed for non methaned hydrocarbons (NMHCs), halocarbons, and alkyl nitrates using gas chromatography at the University of New Hampshire’s Climate Change Research Center. Mixing ratios for these gases are reported as parts per trillion by volume (pptv). Detailed descriptions of the analytical procedures for all measurements are given by Sive et al. [2003, 2005], Zhou et al. [2005], and Zhou [2006].

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Estimating North American Air Toxic Emission Fluxes from Snowmobiles in YNP In order to estimate air toxic emission fluxes from snowmobiles utilized in North America, we first estimate the flux of emissions sampled directly from the exhaust streams of machines utilized in YNP in 2003, compare the results with flux estimates calculated using the box model method described in Finlayson-Pitts and Pitts [2000], and then extrapolate the results for selected air toxics to the North American snowmobile fleet. Emission estimates based direct exhaust measurements of snowmobiles (i.e., the “exhaust flux method”) were made using the following equation (Equation 1): F = E 2-stroke x C 2-stroke x N 2-stroke + E 4-stroke x C 4-stroke x N 4-stroke

(1)

where: F = emission rate; E = exhaust flow rate; C = the mixing ratio of a gas; N = the number of snowmobiles used in the Park. The box model method, which differs from the exhaust flux method in that it estimates the flux of a given gas in a localized context, is described by the equation below:

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

112

David Shively, Yong Zhou and Barkley Sive

(2)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

where: Fx = the flux of compound x (Gg yr-1); C1 = the final mixing ratio of the compound at the end of the sampling period (i.e., noon) (pptv); Co = the initial mixing ratio at the beginning of the sampling period (i.e., dawn) (pptv); As = the side surface area of the box (in YNP this is the north-south width of the Park, or 9.5 x 104 m, multiplied by the boundary layer height); As′ = the area of the box (i.e., the planar surface area of YNP); p = air pressure (in atm, adjusted for elevation); R = constant (8.2054 x 10-5) T = ambient air temperature (K); mwx = molecular weight of compound x; t = time (s). We use Equation 1 to estimate North American air toxics coming from snowmobiles, as opposed to the box model method, because it is based on direct measurements from snowmobiles and the samples would not be affected by emissions from other oversnow vehicle types. Furthermore, it allows us to consider and incorporate different vehicle use parameters (i.e., different proportions of 2-stroke and 4-stroke machines, engine loadings, periods of use, etc.), and to more easily extrapolate the results pertaining to YNP to the rest of the continent. The box model method is used as a means of quality control for the exhaust flux estimates because it effectively aggregates emissions coming from all snowmobile types in use in YNP during the sampling periods, and accounts for varying use parameters (i.e., running at various engine loadings and idling for various time periods).

Assumptions We’ve extended the analysis of air toxics coming from snowmobiles used in the Greater Yellowstone region in 2002-03 to the U.S. fleet [Zhou et al. Submitted] to the North American fleet (i.e., the U.S. and Canada) as a whole. Certain assumptions are built into this analysis: (1) the fleet of snowmobiles utilized in 2002-03 is representative of the North American fleet - the composition (in terms of vehicle types) and utilization of today’s North American snowmobile fleet has not changed significantly since 2002-03; (2) overall emission improvements for newer machines are counterbalanced by engine deterioration and attendant emission increases in older vehicles that are still in use; and (3) even given changes in the use of snowmobiles in and around YNP since the promulgation and implementation of best available technologies (BATs) by the National Park Service (NPS) in 2004, any decrease in the use of older machines and newer two-stroke models in and around YNP is offset by shifts in their use elsewhere – that is, their use was displaced. Examining the first assumption, we estimated that the majority (i.e., 50%–80%) of snowmobiles in use in YNP during our 2003 field campaign were two-stroke machines based on values reported by Ray [2005], visitor survey data reported by RTI International [2004],

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Snowmobile Pollution in North America

113

our observations of machines in use and those temporarily parked at Park attractions (i.e., hydrothermal features), and observations of the composition of West Yellowstone outfitters’ rental fleets. In regard to the following two assumptions, we simply maintain that, without evidence to the contrary, they are not unrealistic. While USEPA regulation of the U.S. automobile fleet for emissions has worked to maintain air quality even in the face of fleet expansion and increased vehicle use by American motorists, outside of the YNP and Grand Teton planning units the snowmobile fleet is unregulated. Lastly, displacement of snowmobile use that might occur in YNP in the absence of regulation is assumed to occur based on the reported continuing sales of snowmobiles in the U.S. and Canada by the International Snowmobile Manufacturers Association [ISMA 2008], and the generally steady number of registrations during the period 2003-2007.

Results

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Ambient Levels of Air Toxics in YNP and West Yellowstone Owing to some very high values for different compounds measured in ambient samples from the Park and West Yellowstone, ambient levels of air toxics are reported as median values. The mixing ratios of air toxics observed in the town of West Yellowstone on 16 February 2003 were generally considerably higher than the levels observed throughout the rest of the study area on both the moderate traffic (12 February) and high traffic (15 February) days. For benzene, for example, they ranged from ~87-22,318 pptv, though were generally above350 pptv. This is undoubtedly resulted from heavy levels of snowmobile use in and around the community. Importantly though, two ambient samples (Table 2) that were collected at the Park’s West Gate Entrance Station on February 16, 2003, adjacent to the main office (under the roof enclosure), showed extremely high levels of air toxics (~0.46-8.71 ppm). Table 2. Mixing ratios of air toxics, in ppmv, collected at the West Gate Entrance Station on February 16, 2003 Sample Time Compound Benzene Toluene p-xylene m-xylene ethylbenzene o-xylene

West Gate Entrance Station, 10:34 Mixing Ratio (ppmv) 2.88 8.71 1.72 3.55 1.35 1.72

West Gate EntranceStation, 15:44 Mixing Ratio (ppmv) 1.02 0.53 0.59 1.25 0.46 0.60

Although the time required to fill each of the 2-liter sample canisters was on the order of 1 minute, benzene levels may have approached the NIOSH recommended short term exposure limit (STEL, for a period of 15 minutes) of 1 ppm in both the morning and afternoon samples. We surmise that the area under the roof enclosure at the West Gate Entrance Station experienced elevated mixing ratios because of sustained traffic and

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

114

David Shively, Yong Zhou and Barkley Sive

impediments to mixing and dilution. These high levels of measured air toxics suggest that Park employees at or near this site were likely exposed to significantly higher levels of toxic compounds than might otherwise occur, increasing their risk of potential health effects. Of the samples collected throughout the park, the maximum values for the air toxics were found on the route between the West Gate Entrance and Old Faithful (Figure 1a). Mixing ratios ranged from ~5-30 ppbv for benzene, toluene, xylenes and ethylbenzene. While these mixing ratios are significantly higher than typical background levels for the ambient well-processed air sampled on the Pacific coastline during the same time of year and at similar latitudes as the Park (unpublished data, D. Blake, University of California, Irvine), they are considerably lower than the regulatory standards set by OSHA or NIOSH.

Exhaust Flux Estimates Visitor entry data from YNP show that, prior to the regulation of snowmobile use in the Park in 2004, an average of some 600 machines per day were run in the Park during the 120 day 2002-03 snowmobile season. Using the 80% 2-stroke fleet component discussed above, the average exhaust mixing ratios from Table 3, the flow rates of the different snowmobile engines [NeTT 2008], and assuming that each snowmobile was operated approximately 8 hours per day at ~30 miles per hour, we estimate emissions of 0.23, 0.77, 0.17, 0.70, and 0.23 Gg yr-1 in YNP for benzene, toluene, ethyl benzene, xylenes and n-hexane (Table 3).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Table 3. Average Emission ratios for air toxics, and emission fluxes of toxics from snowmobiles determined using the exhaust flux box model methods (Gg yr-1) Average mixing ratios in exhaust from 2 stroke snowmobile (ppm) Average mixing ratios in exhaust from 4 stroke snowmobile (ppm) Emission flux by direct measurement in YNP Emission flux using Box Model

Benzene

toluene

ethylbenzene

xylenes

n-hexane

267

753

140

591

240

8

17

3

14

3

0.23

0.77

0.17

0.7

0.23

0.35

1.12

0.24

1.45

0.36

Flux Estimates Based on Box Model Method Based on the vertical profiles of potential temperature as determined from upper air soundings acquired with the radiosonde equipped meteorological balloons, the boundary layer height was estimated to be approximately 1000 m during 1200-1600 hours on February 15, 2003. Using this value to determine AS and AS’, and substituting other values for the variables in Equation 1 as appropriate, the average emission flux estimates for benzene, toluene, ethyl benzene, xylenes, and n-hexane were 2.91, 9.31, 2.03, 12.06, and 3.01 tons day1 , respectively.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Snowmobile Pollution in North America

115

In our calculation, we assume that the variations of air toxics observed at Lake Ranger Station represented average conditions. However, the emission flux estimates represent the emission rates during high traffic days (approximately 1200 snowmobiles). Statistics of snowmobile use from YNP indicate that approximately 600 machines were run in the park per day during the 120 day 2002-2003 winter snowmobile season. By scaling the emission fluxes from the high traffic day to average conditions, the emission fluxes for benzene, toluene, ethyl benzene, xylenes and n-hexane are estimated to be 0.35, 1.12, 0.24, 1.45, and 0.36 Gg yr-1 (Table 3). These estimates are not that different from those calculated using the direct emission data (i.e, the exhaust flux results), demonstrating that those estimates are suitable for estimating continental-scale air toxics production from snowmobiles.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Estimating North American Snowmobile Air Toxic Production Here we use the results of the direct snowmobile emission flux estimates for air toxics in YNP to estimate those for the North American snowmobile fleet. Although snowmobile use has been limited to 4-stroke snowmobile in the YNP since the winter use season of 20042005, most of the snowmobiles used in the rest of this country are 2-stroke snowmobiles. The International Snowmobile Manufacturers Association, which is comprised of the four major manufacturing firms and provides data concerning snowmobile sales and registrations in the U.S., Canada, and elsewhere, reports that there were 1,625,695 snowmobiles registered in the U.S., and 708,490 in Canada in 2007-2008 [International Snowmobile Manufacturers Association 2008]. We feel, though, that these figures are very likely underestimating actual numbers of snowmobiles owned and operated during that winter season, primarily owing to the findings of Sylvester [2001] who found that the total number of snowmobiles in the state was ultimately found to be underestimated by ~65% because a large number of the snowmobiles were unregistered. However, we do not adjust our estimates for this potential underestimation of snowmobile use. Data for 2007-2008 indicate that the average snowmobile is run approximately 1,048 miles per year [ISMA 2008]. If a conservative estimate of 30 miles per hour is used for the average snowmobile speed, the total snowmobile emissions can be calculated based on the exhaust flow rates of 2-stroke and 4-stroke engines used in snowmobiles [NeTT 2008], and the emission measurements made during the 2003 YNP field campaign. Using these parameters and Equation 1 to estimate the high limits of exhaust fluxes depending on exhaust flow rates, using the ratio of the low flow rates to high flow rates for 2-stroke machines to estimate the low limits of exhaust fluxes, and converting units for comparison with the USEPA’s 1999 and 2002 estimates, we produce the values shown in Table 4. Based on these values, the aggregate snowmobile emissions in the U.S. for the air toxics benzene, toluene, ethylbenzene, the xylenes, and n-hexane are estimated to range between 0.070-0.41 Tg yr-1. The aggregate North American snowmobile emissions for these air toxics are estimated to range between 0.102-0.57 Tg yr-1. Table 4 also shows the USEPA’s 1999 (aggregates all non-road recreational vehicle use) and 2002 (for snowmobiles) National Emissions Inventories’ (the only years for which complete data are currently available) annual emissions estimates for the selected air toxics

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

116

David Shively, Yong Zhou and Barkley Sive

that include 29 states in the north and southwest, but not Alaska or California [USEPA 2007c; USEPA 2008e]. Table 4. Emission estimates of U.S. snowmobile air toxics by exhaust flux method, USEPA’s 1999 NONROAD model that does not distinguish snowmobiles, and USEPA’s 2002 NONROAD model that does distinguish snowmobiles (Tg yr-1)

US 2007* EPA (1999) NEI (2002)** North America 2007

benzene

Toluene

ethylbenzene

xylenes

n-hexane

0.008-0.05

0.026-0.15

0.006-0.04

0.02-0.14

0.008-0.04

0.32

0.9

0.14

0.65

0.22

0.0018

0.040

0.001

0.004

Not Reported

0.011-0.07

0.037-0.2

0.009-0.05

0.034-0.19

0.011-0.06

*

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Our low and high limits of emission fluxes are estimated based on exhaust flow rates at 1000 rpm and 6000 rpm, respectively. ** Emission estimates of annual emissions in 2002 by snowmobiles in NEI were based on 29 northern U.S. (not including AK), and western/southwestern states (not including CA).

The results shown in Table 4 clearly demonstrate that our snowmobile air toxic emission flux estimates for the U.S., based on direct sampling from the exhaust streams of 2-stroke and 4-stroke machines, and conditioned by differing flow rates, greatly exceed those of the USEPA’s 2002 NEI snowmobile estimates. Except for toluene, each of the low flow rate estimates exceeds those of the NEI by a factor of ~5.0. Our high flow rate estimations exceed those of the NEI by factors of 27-40 for benzene, ethylbenzene, and the xylenes, and by a factor of 3.75 for toluene. Furthermore, our high flow rate results suggest that snowmobile emissions might represent a significant fraction (~16-29%) of air toxics with respect to the 1999 USEPA estimates of annual emissions by non-road vehicles (Table 4). Because Canada does not estimate air toxics associated with snowmobile use there, we are not able to compare our results for North America to any published values.

Conclusion This study should be viewed as an effort to develop estimates of air toxic emissions that can be used to gauge those of the USEPA. This study produced results for U.S. snowmobile air toxic emissions that appear to be significantly higher, in general, than the values produced by the USEPAs NEI. This is important because of the potential implications associated with underestimates of compounds that have the potential to contribute to adverse affects on human health. It becomes more important in those communities that, because of choice or geographic circumstances, are more dependent on snowmobile use than others. Indeed, our results show that West Yellowstone and the West Entrance Station, during periods of high 2stroke snowmobile use, have the potential to experience air toxic mixing ratios that can approach NIOSH STELs. Longer term exposure to even more moderate levels of ambient air toxics cannot be considered to be in the positive health interests of people living in such environments.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Snowmobile Pollution in North America

117

Although the differences between our estimates and emission inventory need to be further studied. Reducing the overall number of 2-stroke snowmobiles, using clean fuel alternatives, and implementing state-of-the-art emission control strategies will likely reduce wintertime emissions of air toxics in the U.S. This will soon be important as the USEPA is in the process of developing regulations that will address benzene control technologies in oil refining, and controls on gasoline, passenger vehicles, and fuel cans to reduce emissions of benzene and other mobile source air toxics [USEPA 2008d]. While snowmobile emissions regulation is not likely to occur in the foreseeable future, efforts to regulate benzene, and likely other air toxics beyond that, demonstrate that these compounds are taken quite seriously by the agency. It is also hoped that the results presented here offer a point of departure for Environment Canada to begin taking the issue of measuring and modeling air toxics from mobile sources seriously, especially for those associated with snowmobile use. Canada contains a larger number of snowmobile dependent communities in its northern territories and portions of the provinces, and these are often First Nations communities whose populations are likely at greater risk for health impacts associated with exposure to snowmobile related air toxics than other Canadian populations. This is a public health issue that, without attention and action, could be viewed as an environmental justice issue.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

References Aiken, L. CMAJ. 2003, 168: 753. Andersen, M., Aars, J. Polar Biol. 2008, 31:501-507. Centers for Disease Control – CDC. JAMA. 2008, 291, 1314-15. Creel, S., Fox, J. E., Hardy, Al, Sands, J., Garrot, B, Peterson, R. Conser. Biol. 2002, 16:809814. DMV.ORG (2008). Snowmobile and ATV registrations [Internet]. http://www.dmv.org/akalaska/other-types-of-vehicles.php#Snow_Machines_and_ATV_Registrations. Driver, L. (2004). 2002 National Emission Inventory [Internet]. ftp://ftp.epa.gov/ EmisInventory/2002finalnei/general_information/2002nei_plan_briefing.pdf Environment Canada (2008a). Air pollutant emissions [Internet]. http://www.ec.gc.ca/pdb/ cac/cac_home_e.cfm. Environment Canada. (2008b). NPRI substance information [Internet]. http://www.ec.gc.ca/pdb/npri/npri_si_e.cfm. Environment Canada (2008c). 2005 air pollutant emissions [Internet]. http://www.ec.gc.ca/ pdb/cac/Emissions1990-2015/2005/2005_canada_e.cfm. Finlayson-Pitts, B., and J.N. Pitts Jr. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments, and Applications Academic Press, 2000. Hoey, J. CMAJ. 2003, 168: 739. International Agency for Research on Cancer – IARC. Overall Evaluation of Carcinogenicity: An Updating of IARC Monographs; IARC; Monographs on the Evaluation of Carcinogenic Risks to Humans; World Health Organization: Lyon, France, 1987; Vol 1-42, Supplement 7. International Snowmobile Manufacturers Association – ISMA (2008). Snowmobile statistics [Internet]. http://www.snowmobile.org/stats.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

118

David Shively, Yong Zhou and Barkley Sive

Irwin, R. J., VanMouwerik, M., Stevens, L., Seese, M. D., Basham, W. Environmental Contaminants Encyclopedi; National Park Service, Water Resources Division: Fort Collins, CO, 1998. Kado, N. Y., Kuzmicky, P. A., Okamoto, R. A. (2001). Environmental and Occupational Exposure to Toxic Air Pollutants from Winter Snowmobile Use in Yellowstone National Park - Final Report [Internet]. http://www.deq.state.mt.us/CleanSnowmobile/ publications/Reports/ExposureToxicAirPollutants_FinalRpt.pdf. Montana Department of Environmental Quality – MTDEQ (2008). Clean snowmobile facts [Internet]. http://deq.mt.gov/CleanSnowmobile/index.asp. Montana Department of Fish, Wildlife, and Parks – MTDFWP (2008). Register your vehicle or vessel [Internet] http://fwp.mt.gov/recreation/permits/register.html. National Park Service - NPS, Yellowstone National Park, 1999: The State of the Park; National Park Service: Mammoth Hot Springs, WY, 1999. National Park Service – NPS. Air Quality Concerns Related to Snowmobile Usage in National Parks; National Park Service: Denver, CO, 2000. National Park Service – NPS. Yellowstone Resources and Issues; National Park Service: Mammoth Hot Springs, WY, 2008. Nayci, A., Stavlo, P. L., Abdalla, E. Z., Zietlow, S. P., Moir, C. R., Rodeberg, D. A. Mayo Clin. Proc. 2006, 81: 39-44. NeTT (2008). Nett catalytic converter sizing program [Internet]. http://www.nett.ca/ tools/size/index.html). Ray, J.D. (2005). Winter air quality study 2004–2005 [Internet]. http://www.nps.gov/yell/ parkmgmt/upload/winteraqstudy04–05.pdf. RTI International (2004). Economic analysis of temporary regulations on snowmobile use in the Greater Yellowstone Area [Internet]. http://www.nps.gov/yell/parkmgmt/ upload/econ_analysis–04.pdf. Shively, D. D., Pape, B. M. C., Mower, R. N., Zhou, Y., Russo, R., Sive, B. C. (2008). Envir. Mgmt. 2008, 41: 183-199. Sive, B. C., Shively, D. D., Pape, B. M. C. Spatial variation and characteristics of volatile organic compounds associated with snowmobile emissions in Yellowstone National Park. Central Michigan University: Mount Pleasant, MI, 2002. Sive, B. C., Shively, D. D., Pape, B. M. C. (2003). Spatial variation of volatile organic compounds associated with snowmobile emissions in Yellowstone National Park [Internet]. http://www.nps.gov/yell/technical/planning/winteruse/plan/sive_report.htm. Spear, T. M., Stephenson, D. J. (2005). Yellowstone winter use personal exposure monitoring [Internet]. http://www.nps.gov/yell/parkmgmt/winterusetechnicaldocuments.htm. Stewart, R. L., Black, G. B. J. Can. Chir. 2004, 47, 90-94. Sylvester, J. T., Snowmobiling in Montana: A 1998 Update, October 1998. University of Montana Bureau of Business and Economic Research: Missoula, MT, 2001. United States Environmental Protection Agency – USEPA (1994). Air toxics from motor vehicles; EPA 400-F-92-004, U.S. EPA, Office of Mobile Sources: Ann Arbor, Michigan. United States Environmental Protection Agency – USEPA (2007a). The clean air act amendments of the 1990 list of hazardous air pollutants [Internet]. http://www.epa.gov/ttn/atw/orig189.html.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Snowmobile Pollution in North America

119

United States Environmental Protection Agency – USEPA (2007b). Health effects notebook for hazardous air pollutants [Internet]. http://www.epa.gov/ttn/atw/hlthef/hapindex.html. United States Environmental Protection Agency – USEPA (2007c). 1999 national emissions inventory documentation and data – final version 3.0. [Internet]. http://www.epa.gov/ttn/chief/net/1999inventory.html. United States Environmental Protection Agency – USEPA (2008a). About air toxics [Internet] http://www.epa.gov/ttn/atw/allabout.html. United States Environmental Protection Agency – USEPA (2008b). National air toxics assessments [Internet]. http://www.epa.gov/ttn/atw/natamain/index.html. United States Environmental Protection Agency – USEPA (2008c). Mobile source air toxics [Internet]. http://www.epa.gov/otaq/toxics.htm#regs. United States Environmental Protection Agency – USEPA (2008d). Technology transfer network: 1996 national-scale air toxics assessment frequently asked questions [Internet]. http://www.epa.gov/ttn/atw/nata/natsafaq.html#A11. United States Environmental Protection Agency – USEPA (2008e). Technology transfer network: 1999 national-scale air toxics assessment [Internet]. http://www.epa.gov/ ttn/atw/nata1999/. United States Environmental Protection Agency – USEPA (2008f). 2002 National Emissions Inventory data and documentation. [Internet]. http://www.epa.gov/ttn/chief/net/ 2002inventory.html#inventorydata. Williford, J. (2001). Status and prospects for two-stroke engines used in off-road recreational vehicles [Internet]. http://deq.mt.gov/CleanSnowmobile/montana/williford/FINALBU08.htm. Zhou, Y., Shively, D., Mao, H., Russo, R., Pape, B., Mower, R. N., Varner, R., and Sive, B. Air toxic emissions from snowmobiles in Yellowstone National Park, Environ. Sci. Technol., Submitted, 2009 Zhou, Y., Varner, R. K., Russo, R. S., Wingenter, O.W., Haase, K. B., Talbot, R.W., Sive, B. C. J. Geophys. Res. 2005, 110: D21302. Zhou, Y. Atmospheric Volatile Organic Compound Measurements: Distributions and Effects on Air Quality in Coastal Marine, Rural and Remote Continental Environments. Ph.D. Dissertation; University of New Hampshire, 2006.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 121-133

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 5

MODELING OF TRAFFIC-RELATED ENVIRONMENTAL POLLUTION IN THE GIS 1

Lubos Matejicek1 and Zbynek Janour2 Institute for Environmental Studies, Charles University in Prague Benatska 2, 128 01, Prague, Czech Republic 2 Institute of Thermomechanics, Academy of Sciences Dolejskova 5, 182 00, Prague, Czech Republic

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract The numerical models are based on dispersion modeling and statistical analysis. In case of dispersion modeling, the ISC-AEROMOD View is used for modeling multiple pollutants with the U.S. EPA modeling tool ISCST3. The Mobile View assists as an interface for the U.S. EPA MOBILE6 model that predict arterial street emissions focused on hydrocarbons, carbon monoxide, nitrogen oxides, carbon dioxide, particulate matter, and toxics from cars, motorcycles, light- and heavy-duty trucks under various conditions. The potential impacts of accidental releases are solved by SLAB View that complements the modeling tools by analysis of emissions from accidental releases of toxic gases. Analysis of urban traffic-induced noise pollution is assessed by U.S. FHWA-TNM tools. The GIS is finally used to serve as a common analysis framework for individual modeling tools. In order to display the numerical simulation outputs together with urban area map layers, numerical modeling based on U.S. EPA software tools is integrated into the GIS for spatial interpolations and spatial analysis. It assists to evaluate high levels of air pollution and noise pollution together with the thematic map layers of residential zones, business centers, schools, and hospitals. Finally, finding alternative routes can decreases air pollution and noise pollution in selected zones. As a case study, the city of Prague sample data set helps to demonstrate data processing and modeling of traffic-related environmental pollution. The ESRI’s geodatabase is used for implementation a comprehensive information model and a transaction model in the GIS environment. It is also the common application logic used in ArcGIS for accessing and working with all spatial thematic data and simulation inputs/outputs. Spatial interpolation for prediction maps and probability maps complement the existing thematic map layers, which enable cell based modeling for spatial multi criteria decision analysis. The synthesis of environmental models and GISs creates a more complex base for environmental simulation that can support decision-making processes in a more straightforward way.

Keywords: Road traffic; air pollution; noise assessment; spatial modeling; GIS. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

122

Lubos Matejicek and Zbynek Janour

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1. Introduction The increasing traffic in growing urban areas amplifies certain problems. It affect human daily life, because the traffic related emissions and noise are today’s leading cause of environmental pollution due to the overall success in reducing emissions of pollutants from other industrial sources in the past decades. Numerous studies have identified many associations between traffic-related air pollution merged with other negative phenomena (noise, traffic accidents, traffic jams), and adverse potential health effects indicated either by characterizing exposures to specific pollutants using measurements from a few ambient sites or by some estimates of traffic [1, 2]. Despite ignoring the local contribution of indoor sources, the local effect of residential ventilation and the accuracy to estimate personal exposures, home-specific outdoor measurements are often used for initial estimates. In lieu of using these measurements, which requires nearly labor-intensive steps as indoor monitoring, some factors can be generated from the geographic information system (GIS) [3, 4]. It includes using distance from roads, complex digital elevation models, land use, and population density in combination with central site monitoring data to derive estimates of ambient exposures. Questionnaire data focused on individual building characteristics (air conditioning usage, opening windows and local indoor emission sources) can be used to particularize local environmental pollution and potential health effects. Many current studies deal with publicly available data from monitoring networks and questionnaire responses managed by state agencies and local authorities. Some of them provide directions how publicly available data can be utilized, in order to predict residential indoor exposures in the absence of measurements. For example, information on traffic applied in the GIS framework in combination with ambient monitoring data focused on nitrogen dioxide, fine particulate matter and elemental carbon is used as a substitute for home-specific outdoor measurements and consecutively as a particular estimate for indoor exposures of outdoor dominated pollutants. Thus, models based on regression analysis and Bayesian approaches are often used for predictions [5, 6, 7]. Another research is represented by using methods of numerical modeling [8] or physical modeling [9]. Numerical models based on partial differential equations can solve dynamic phenomena of pollutant dispersion in dependence on wind flows above complex terrain. In addition to case oriented numerical models dedicated to special mathematical tasks, many software tools are accessible from U.S. EPA in the framework of the given guidelines. Physical models, exploring the relation between three-dimensional morphology and windiness in wind tunnel experiments, are usually used for prediction of pollutant dispersion in the idealized city models. A relatively new approach is represented by statistical tools and numerical models integrated in the GIS environment. The GIS offers many tools for exploratory spatial data analysis and management, which can support numerical models in a more efficient way [10]. Environmental pollution by noise brings health effects and behavioral problems. Noise pollution can cause hypertension, high stress levels, hearing loss and sleep disturbances. In case of animals, noise causes stress and increases risk of mortality. Mitigation of traffic noise is provided by noise barriers, limitation of vehicle speed and acceleration. Applying these strategies optimizes computer models based on physical rules and local topography [11]. Again, GISs offer a wide range of tools for data management and noise modeling support.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Modeling of Traffic-Related Environmental Pollution in the GIS

123

Other potential sources of environmental pollution are impacts of accidental releases. They can cause releases of toxic gases that may expose residential zones, schools and hospitals in urban areas. Thus, the potential accidents at selected sites of crossroads and arterial roads are also included in modeling of traffic-related environmental pollution [12].

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2. Traffic-Related Environmental Pollution Environmental pollution originated from traffic comes from a wide variety of individual moving sources. Thus, in order to provide computer simulations, the crossroads are mostly identified as the point sources, the highways and city streets as the line sources, and the parking places or arterial roads as the area sources. In case of this paper, the environmental pollution includes air pollution, air pollution from potential car accidents and noise. The air pollutants emitted directly into the atmosphere represent primary pollutants. The air pollutants formed in the air as a result of chemical reactions are known as secondary pollutants. For example, carbon monoxide (CO) and sulfur dioxide (SO2) are primary pollutants, while ozone is a secondary pollutant. Nitrogen dioxide (NO2) and some particulate matter represent both primary and secondary pollutants, because they are emitted directly into the atmosphere, and formed from other pollutants. In case of traffic emissions, some nitrogen dioxide is emitted directly from vehicle exhaust, but most is formed by oxidation of nitric oxide (NO) in the atmosphere. Similarly, fine particular matter is emitted directly by moving vehicles as well as formed in the air. In case of primary pollutants, the ambient concentrations indicate approximately proportional relationship to emissions. In case of secondary pollutants, the relationship is more complex. For example, reducing local emissions of nitrogen oxides can lead to an increase in local ozone concentration and other consequent reactions [13]. The emissions of carbon dioxide (CO2), SO2 and black smoke can be estimated from petrol and diesel consumption. Air pollution caused by SO2 usually varies in dependence on the sulfur content of the fuels. Black smoke emissions are estimated in accordance with soiling factors for different fuels and vehicle types. The estimates of emissions of NOx, CO and non-methane volatile organic compounds (NMVOCs) are often based on vehicle performance (emissions from cold starts, hot engines and evaporative losses) rather than fuel consumption. Thus, the emissions from individual motor vehicles mainly depend on the type of vehicle, the fuel used, and the vehicle performance that depends on configuration of the traffic network, especially in urban areas. In order to obtain required information, complex exploration and measurements of a number of various vehicles together with vehicle kilometers driven in roads are needed. In case of using the GIS, its spatial database can support mapping of traffic networks, classification of road types (urban, rural single or dual carriageway, motorway), mapping of vehicles categories (petrol or diesel cars, light goods or heavy duty vehicles, buses or coaches, and motorcycles) in dependence on spatial location. In addition to these categories, the vehicle type can be split according to the required regulation in the last years. The emission factors are mostly based on on-road measurements of emissions. Due to large variations from vehicle to vehicle, the emission factors used for calculations are represented by means of the measured emissions. Besides these uncertainties, a number of other implicit assumptions are needed like, for example, impacts of load, road gradient, etc. Particulate matters like PM10 need slightly different estimates in dependence on the type of vehicles and catalyst technology.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

124

Lubos Matejicek and Zbynek Janour

The prediction of future emissions based on specified factors also requires estimates of economic activity and demographic influence together with repeated terrain measurements. Information about emission forecasting models, annual reports and methodological recommendation are regularly published by U.S. EPA and by the European Environmental Agency (EMEP/CORINAIR Emission Inventory Guidebooks). Traffic noise represents other part of environmental pollution. Various methodologies are used for risk assessment in dependence on recommendations of government authorities. Traffic noise prediction models mostly assist in simulation of sound pressure levels, specified in terms of the equivalent continuous level (Leq) over a chosen period and under interrupted or varying flow conditions. The early models were focused on prediction of linear levels whereas the later models on prediction of A-weighted levels. For setting of the noise sources, single point sources, short line sources or multi point/line sources are taken into calculations, optionally, some with different spectra. Among many prediction models and case oriented studies [14, 15, 16], there is a few modeling tools recommended by national regulations (the FHWA model in the US, the CRTN model in the UK, the RLS90 model in Germany, the OAL model in Austria, the EMPA model in Switzerland, the ASJ model in Japan).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3. Description of Software Tools Used for Modeling of TrafficRelated Environmental Pollution In the framework of the attached case study a number of software tools are used for modeling of traffic-related environmental pollutants in the selected urban area. In order to merge all the needed data for risk assessment, the GIS is utilized for data management, spatial analyses and visualization. Sharing data between the GIS and the individual modeling software tools is mostly through the shared data files. Industrial Source Complex Short Term (ISCST), originally developed in the 1970’s, is the US EPA’s regulatory tool based on a steady-state Gaussian plume algorithm. The current version (ISCST3) contains many enhancements that include an improved area source algorithm, the complex terrain screening algorithms, a revised dry and wet deposition tools, and many revisions intended for air toxics applications. The ISCST3 is applicable for estimating ambient impacts from point, area and volume emission sources out to a distance of approximately 50 kilometers. In spite of that the ISCST3 is primarily dedicated to the dispersion modeling of the stationary sources, it can by applied together with other modeling tools for dominant traffic-related sources (sites with continuous traffic jams, larger parking places, and arterial roads during rush hours) [17, 18, 19]. The MOBILE6 includes a revised emission factor model for estimation of gram per mile emissions of hydrocarbons (HC), carbon monoxide (CO), nitrogen oxides (NOx), carbon dioxide (CO2), particulate matter (PM), and toxics from cars, trucks, and motorcycles under various conditions affecting in-use emission levels. Thus, the ambient temperatures and average traffic speeds can be specified by users to develop emission inventories and control strategies. Due to the new vehicles types and progress in traffic systems, the emission factor model has been repeatedly improved and revised since its initial version. Many case studies bring new ideas into research of the traffic-related emission estimates by adopting new technology [20, 21] and regional driving differences [22].

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Modeling of Traffic-Related Environmental Pollution in the GIS

125

The FHWA Traffic Noise Model (FHWA TNM) for highway traffic noise prediction and analysis was initially released in 1998. Actual version, FHWA TNM 2.5, contains many improvements and revisions. As sources of noise, it includes noise emission levels for automobiles, medium tracks, heavy trucks, buses, and motorcycles. Noise emission levels consist of A-weighted sound levels (one-third octave-band spectra), and subsource-height strengths for a few pavement types (dense-graded asphaltic concrete, Portland cement concrete, open-graded asphaltic concrete, and composite pavement types consisting of previous listed types). It includes full-throttle noise emission levels for vehicles on upgrades, and vehicles accelerating away from stop signs, toll booths, traffic signals and on-ramp start points. Upcoming case studies are mostly focused on evaluation the effects of noise barriers, and assessing the model sensitivity to parameter variations [23, 24]. The computer model focused on the potential impacts of accidental releases, SLAB, complements the list of the selected models for presented complex modeling of traffic-related environmental pollution. The SLAB model solves the atmospheric dispersion of denser-thanair releases over flat. All the source input conditions have to be determined externally by approximate estimates or terrain measurements. A few release types are treated. They include continuous release, finite duration release, instantaneous release, ground-level evaporating pool, elevated horizontal jet, stack or elevated vertical jet and instantaneous volume source. The SLAB model was developed in 1990 by D.L. Ermak of the Lawrence Livermore National Laboratory with support from U.S. Department of Energy, USAF Engineering and Services Center, and the American Petroleum Institute [25]. A wide range of various data based on terrain measurements, existing projects, and annual reports needs to be managed and integrated together with modeling tools, in order to enable the exploration of traffic related environmental pollution in dependence on spatial scales, time scales, and other attributes. Thus, the GISs are used for most of the tasks. The described U.S. EPA modeling tools (ISCST3, MOBILE6, FHWA-TNM 2.5, and SLAB) complement the risk assessment information system as the case oriented modeling tools. Integration of the GIS and spatially distributed environmental models is based on pre- and postprocessor linkage through shared data files. Despite building models as analytical functions into the fully functional GIS, the U.S. EPA modeling tools and the GIS form standalone software systems. The preprocessed GIS data are imported into U.S. EPA models, and simulation outputs are backward exported into the GIS to provide spatial analysis and visualization. Thus, the GIS project is finally used to serve as a common analysis framework for individual U.S. EPA modeling tools. In addition to standard GIS functionality (entering, storing, retrieving, transforming, measuring, combining, subsetting, and displaying spatial data that are registered to a common coordinate system) [26], managing 3D data for complex digital terrain models (DTMs extended by buildings, barriers, and other surface objects) and advanced spatial interpolation algorithms are needed to be carry out for final display and visualization. To accomplish all the needs, the ESRI’s geodatabase offers to implement a comprehensive information model and a transaction model in the GIS environment. Thus, the ArcGIS tools can assess and work with all spatial thematic data and simulation inputs/outputs. The GIS can also efficiently handle data from remote sensing and global positioning systems (GPSs). Thus, the aerial images and terrain observations complement the existing map layers in the framework of the final visualization. The satellite images can extend the map layers by data from other infrared spectral bands. GPSs, frequently used for car navigation, extend existing map layers by location of terrain measurements.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

126

Lubos Matejicek and Zbynek Janour

4. A Case Study for Modeling of Traffic-Related Environmental Pollution

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

In order to demonstrate modeling of the traffic-related environmental pollution, the GIS project is created to integrate dispersion of pollutants, noise assessment and potential accident releases together with terrain measurements, aerial images, satellite images, DTMs, annually reported data and other existing spatial data. The data processing and modeling in the framework of the GIS project is illustrated in Figure 1.

Figure 1. The flow diagram for data processing and modeling of the traffic-related environmental pollution in the GIS environment.

The upper part of the schema includes the main environmental data inputs. In case of air pollution, the meteorological data contain local estimates of the wind speed and the wind direction. In addition to importing the DTM into the modeling tools (ISC AEROMOD View Terrain Processor, FHWA-TNM and SLAB), the simplified shapes of buildings are used for US EPA BPIP Model, in order to effectively and quickly complete the building downwash analysis in the framework of running the ISCST3. The Computer Aided Design (CAD) files include the DTM complemented by surface objects. Traffic related data describe traffic

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Modeling of Traffic-Related Environmental Pollution in the GIS

127

intensity and vehicle types for setting of the modeling tools based on the MOBILE6 and the FHWA-TNM. The next phase includes pre-processing of the input data into the data formats for inputs of the selected models (ISCST3, MOBILE6, FHWA_TNM, and SLAB). In case of the ISCST3, the PC-RAMMET is used for pre-processing of meteorological data and the BPIP for adopting of spatial data (DTM and buildings). While the described data processing can be carried out in the GIS (with exception of the PC-RAMMET, and the BPIP), the models solved by ISCST3, MOBILE6, FHWA-TNM and SLAB are implemented in the standalone software tools. In spite of that they also include display functions, the simulation outputs are imported backward into the GIS project that manages a wide range of spatial analyses, and provide more complex outputs for risk assessment. In addition to the spatial data management, the spatial interpolations by deterministic techniques (Inverse Distance Weighted-IDW, and polynomial interpolations) and geostatistical methods (kriging and cokriging) assist for estimations of air pollution concentrations in the neighborhood of the sites predicted by models, and existing monitoring stations [27, 28]. The GIS also contains many spatial exploratory statistical tools such as histograms for checking normal distribution, quantile plots, trend analysis and semivariograms. The normal distribution of input data is required for geostatistical interpolation methods that produce not only prediction surfaces but also uncertainty surfaces, give an indication of prediction quality and can generate probability and quantile output maps depending on spatial modeling needs. Unsuitable sites can be located by combining layers focused on air pollution and noise assessment. After the deriving the air pollution layers and the noise level layer to a common scale, the cell based modeling can assist in calculation the output map for complex risk assessment.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4.1. Mapping of the Emission Sources As an example, the urban site in Prague, the Czech Republic, characterized by high population density is selected for modeling of traffic-related environmental pollution, Figure 2. The accumulation of traffic in the last decade has indicated an increasing pressure in the air quality and noise. The more detailed view of the main crossroad from the aerial image is in Figure 3. The image was not taken during the rush hours, which causes a small percentage of the vehicles on the roads. The significant spatial and temporal variability of traffic during the day also results in variability of traffic-related environmental pollution. A number of environmental studies have raised the question of how representative the site and time period of air quality actually can be in comparison with other sites and different time periods [29]. In order to provide validation of simulation outputs, data from surface monitoring stations and local mobile DIAL-LIDAR measurements can be used for the selected site [30]. In case of the presented case study, estimates of emissions from the motor vehicles are based on the mobile sources emission factor model implemented in the U.S. EPA MOBILE6. The setting of the parameters reflects the rush hours, in order to explore scenarios focused on expected higher levels of air pollution and noise. The traffic volumes, average vehicle speeds, humidity, and meteorological parameters are based on terrain observations and regular reports.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

128

Lubos Matejicek and Zbynek Janour

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 2. The selected site for modeling of traffic-related environmental pollution with a main crossroad in the middle part and arterial roads (Google Earth 2008, Image © GEODIS, Brno).

Figure 3. Simulation results of the ammonia horizontal jet release caused by a traffic accident at the main crossroad displayed together with the aerial images in the GIS environment.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Modeling of Traffic-Related Environmental Pollution in the GIS

129

4.2. Simulation of Traffic-Related Environmental Pollution

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

As an example, the ammonia horizontal jet release caused by the traffic accident at the main crossroad is simulated and displayed together with other spatial data in the GIS environment, Figure 3. The SLAB model simulates the atmospheric dispersion in accordance to the spill source type, source properties, spill parameters, local field parameters and meteorological conditions. The spread of a plume caused by an accidental release of a chemical is drifted above the neighbor residential zone. Line contours give an approximate concentration levels. Simulation of atmospheric dispersion originated from vehicles is demonstrated by the ISCST3 in the framework of the AEROMOD View. As an example, the simulation results based on a steady-state Gaussian plume algorithm are used for the approximate spatial prediction of NOx in Figure 4.

Figure 4. Simulation results of the ammonia horizontal jet release together with the traffic related air pollution (NOx) displayed in the GIS environment.

But, many other traffic-related compounds (CO, VOCs, and dust emissions caused by heavy transport) can be included into the simulation. The spatial distribution of the pollutant concentration shows the higher concentration in the neighborhood of the crossroads and the arterial roads that can be shifted in dependence on the wind speed and the wind direction. In order to provide more complex risk assessment, the traffic noise model (FHWATNM2.5) complements used models. The simulation results based on prediction of hourly Aweighted equivalent sound level (LAeq1h) are illustrated together with air pollution in Figure 5. Again, the simulation outputs are integrated in the framework of the GIS project. The higher noise levels are at the main crossroads and arterial roads, but optionally, can be decreased by

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

130

Lubos Matejicek and Zbynek Janour

barriers in their neighborhood. Considering to sharing other map layers, the GIS approach offers a more efficient way in design of the barriers.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Simulation results of the ammonia horizontal jet release together with the traffic related air pollution (NOx), and noise assessment displayed in the GIS environment.

In spite of that the displayed map layers are based on the dense network of individual receptors or grids of receptors, the spatial interpolations (IDW or the ordinary kriging, for normal input data distributions) are needed to produce continuous surfaces of pollutant concentrations or noise levels. In this case, the GIS provides a comprehensive set of tools for creating surfaces that are used for visualization and analysis.

Conclusion In spite of that many environmental studies focused on air pollution, noise assessment and other environmental phenomena have been carried out [31, 32, 33], there is still need to develop research tools for complex risk assessment. Different techniques of terrain measurements, different methodology of data processing, and finally, various decisionmaking processes have created a wide range of individual procedures in dependency on spatial and temporal resolution of the areas of interest. In order to unify some common procedures of data processing, the GISs can be used as a framework for running and development of powerful, efficient and complex tools in air pollution exploration and modeling [34, 35, 36]. In addition to spatial database management, many extensions focused on network analysis, spatial analysis, 3D spatial modeling and geostatistical methods can support modeling of the traffic-related environmental pollution. The used case oriented

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Modeling of Traffic-Related Environmental Pollution in the GIS

131

modeling tools for the accidental releases, the pollutants dispersion, and the noise assessment have shown that new ways of integration are needed. In case of environmental data, the unified data storage complemented by spatial information from GPS or remote sensing can be realized by the spatial database in the GIS environment. It is designed to offer best available knowledge to bear on environmental planning and policy making, reach a broad audience, be easy to use and understand and help to explore a wide range of options. Thus, the close integration of environmental models into the GIS will help to explore the traffic-related environmental pollution in a more sophisticated way, and will reach a broad audience.

Acknowledgements The input spatial data included into the case study are originally provided by the Institute of Municipal Informatics of Prague. The map layers of emission sources and the temporal measurements of the surface air pollution transferred from the Internet and from the annual reports are administrated by the IT Department of the Prague City Hall. The spatio-temporal data were processed by ESRI-ArcGIS and Leica GeosystemsERDAS Imagine in the GIS Laboratory, Faculty of Natural Science, Charles University in Prague in the framework of the project AVCR 1ET400760405 supported by the Information Society, Academy of Sciences in the Czech Republic.

References

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[1] [2] [3] [4] [5]

[6]

[7]

Seika, M., Metz, N. and Harrison, R.M. (1996). Characteristics of urban and state emission inventories-comparison from Europe and the United States. The Science of the Total Environment 189/190, 221-234. Salt, J.E. and Moran, A. (1997). International greenhouse gas inventory systems: A comparison between CORINAIR and IPCC methodologies in the EU. Global Environmental Change 7(4), 317-336. Maguire, D.J., Batty, M. and Goodchild, M.F. (2005). GIS, Spatial Analysis and Modeling. Redlands, California: ESRI Press. Mitchell, A. (2005). The ESRI Guide to GIS Analysis. Volume 2: Spatial Measurements and Statistics. Redlands, California: ESRI Press. Brigs, D.J., de Hoogh, C., Gulliver, J., Wills, J., Elliott, P., Kingham, S. and Smallbone, K. (2000). A regression-based method for mapping traffic-related air pollution: application and testing in four contrasting urban environments. The Science of the Total Environment 253, 151-167. Hochadel, M., Heinrich, J., Gehring, U., Morgenstern, V., Kuhlbusch, T., Link E., Wichmann, H.E. and Krämer, U. (2006). Predicting long-term average concentration of traffic-related air pollutants using GIS-based information. Atmospheric Environment 40, 542-553. Baxter, L.K., Clougherty, J.E., Paciorek, J., Wright, R.J. and Levy, J.I. (2007). Predicting residential indoor concentrations of nitrogen dioxide, fine particulate matter, and elemental carbon using questionnaire and geographic information system based data. Atmospheric Environment 41, 6561-6571.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

132 [8] [9] [10] [11] [12]

[13] [14] [15] [16]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[17] [18] [19] [20] [21]

[22] [23] [24]

Lubos Matejicek and Zbynek Janour Venkatram, A., Isakov, V., Thoma, E. and Baldauf, R. (2007). Analysis of air quality data near roadways using a dispersion model. Atmospheric Environment 41, 9481-9497. Skote, M., Sandberg, M., Westerberg, U., Claesson, L. and Johansson A.V. (2005). Numerical and experimental studies of wind environment in an urban morphology. Atmpospheric Environment 39, 6147-6158. Maidment, D.R. (1996). Environmental Modeling within GIS. In M.F. Goodchild, L.T. Steyaert, B.O. Parks (Ed.), GIS and Environmental Modeling: Progress and Research Issues (pp. 315-323). Fort Collins: GIS World, Inc. Steele, C. (2001). A critical review of some traffic noise prediction models. Applied Acoustics 62, 271-287. Jaeger, J.A.G., Bowman, J., Brennan, J., Fahrig, L., Bert, D., Bouchard, J., Charbonneau, N., Frank, K., Gruber, B. and von Toschanowitz, K.T. (2005). Predicting when animal populations are at risk from roads: an interactive model of road avoidance behavior. Ecological Modelling 185, 329-348. Holman, C. (1999). Sources of Air Pollution. In S. Holgate, H. Koren, J. Samet, R. Maynard (Ed.), Air Pollution and Health (pp. 115-148). Amsterdam: Elsevier. Sopasakis, A. (2004). Stochastic noise approach to traffic flow modeling. Physica A 342, 741-754. El-Fadel, M., Shazbak, S., Baaj, M.H. and Saliby, E. (2002). Parametric sensitivity analysis of noise impact of multihighways in urban areas. Environmental Impact Assessment Review 22, 145-162. Pamanikabud, P. and Vivitjinda, P. (2002). Noise prediction for highways in Thailand. Transportation Research Part D 7, 441-449. Wang, L., Parker, D.B., Parnell, C.B., Lacey, R.E. and Shaw, B.W. (2006). Comparison of CALPUFF and ISCST3 models for predicting downwind odor and source emission rates. Atmospheric Environment 40, 4663-4669. Lorber, M., Eschenroeder, A. and Robinson, R. (2000). Testing the USA EPA’s ISCSTVersion 3 model on dioxins: a comparison of predicted and observed air and soil concentrations. Atmospheric Environment 34, 3995-4010. Hao, J., He, D., Wu, Y., Fu, L. and He, K. (2000). A study of the emission and concentration distribution of vehicular pollutants in the urban area of Beijing. Atmospheric Environment 34, 453-465. Pokharel, S.S., Bishop, G.A. and Stedman, D.H. (2002). An on-road motor vehicle emissions inventory for Denver: an efficient alternative to modeling. Atmospheric Environment 36, 5177-5184. Kuhns, H.D., Mazzoleni, C., Moosmüller, H., Nikolic, D., Keislar, R.E., Barber, P.W., Li, Z., Etyemezian, V. and Watson, J.G. (2004). Remote sensing of PM, NO, CO and HC emission factors for on-road gasoline and diesel engine vehicles in Las Vegas, NV. Science of the Total Environment 322, 123-137. Lin, J. and Niemeier, D.A. (2003). Regional driving characteristics, regional driving cycles. Transportation Research Part D 8, 361-381. El-Fadel, M., Shazbak, S., Baaj, M.H. and Saliby E. (2002). Parametric sensitivity analysis of noise impact of multihighways in urban areas. Environmental Impact Assessment Review 22, 145-162. Li, B., Tao, S., Dawson, R.W., Cao, J. and Lam, K. (2002). A GIS based road traffic noise prediction model. Applied Acoustics 63, 679-691.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Modeling of Traffic-Related Environmental Pollution in the GIS

133

[25] Thé, J.L., Thé, C.L. and Johnson, M.A. (2004). Slab View. User’s Guide. Ontario, Canada: Lakes Environmental. [26] Johnston, C.A. (1998). Geographic Information Systems in Ecology. London: Blackwell. [27] Cressie, N.A.C. (1993). Statistics for Spatial Data. New York: Wiley. [28] Johnston, K., Ver Hoef, J.M., Krivoruchko, K. and Lucas, N. (2001). Using ArcGIS Geostatistical Analyst. Redlands, California: ESRI Press. [29] Vardoulakis, S., Gonzalez-Flesca, N., Fisher, B.E.A. and Pericleous, K. (2005). Spatial variability of air pollution in the vicinity of a permanent monitoring station in central Paris. Atmospheric Environment 39, 2725-2736. [30] Matejicek, L., Janour, Z. and Strizik, M. (2008). Spatial Modeling of Air Pollution Based on Traffic Emissions in Urban Areas. In S.E. Paterson, L.K. Allan (Ed.), Road Traffic: Safety, Modeling, and Impacts (in press). New York: Nova Science Publisher, Inc. [31] Mower, B. (1998). A multiple source approach to acute human health risk assessment. Waste Management 18, 377-384. [32] Borrego, C., Tchepel, O., Barros, N. and Miranda, A.I. (2000). Impact of road traffic emissions on air quality of the Lisbon region. Atmospheric Environment 34, 4683-4690. [33] Borrego, C., Tchepel, O., Costa, A.M., Martins, H., Ferreira, J. and Miranda, A.I. (2006). Traffic-related particulate air pollution exposure in urban areas. Atmospheric Environment 40, 7205-7214. [34] Matejicek, L. (2005). Spatial modelling of air pollution in urban areas with GIS: a case study on integrated database development. Advances in Geosciences 4, 63-68. [35] Matejicek, L., Engst, P. and Janour, Z. (2006). A GIS-based approach to spatiotemporal analysis of environmental pollution in urban areas: A case study of Prague’s environment extended by LIDAR data. Ecological Modelling 199, 261-277. [36] Grünfeld, K. (2005). Integration spatio-temporal information in environmental monitoring data-a visualization approach applied to moss data. Science of the Total Environment 347, 1-20.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 135-147

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 6

USING MONITORING DATA TO EVALUATE THE VARIATIONS OF TRAFFIC-RELATED AIR POLLUTION IN TAIWAN FROM 1994 TO 2006 Tzu-Yi Pai1, 2, Keisuke Hanaki2, Horng-Guang Leu3 and Shuenn-Chin Chang3 1

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Department of Environmental Engineering and Management, Chaoyang University of Technology, Wufeng, Taichung, 41349, Taiwan, R.O.C. 2 Department of Urban Engineering, School of Engineering, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan 3 Environmental Protection Administration, Taipei, 10042, Taiwan, R.O.C.

Abstract The Air Pollution Control Act (APCA) of Taiwan signed in 1975 prescribes the maximum permissible limits of motor vehicle exhausts as well as the monitoring of air pollution, etc. In this study, the concentrations of air pollutants including SO2, CO, O3, PM10, NO2, and nonmethane hydrocarbons (NMHC) from background air pollution monitoring stations (BAQMSs) and traffic area air pollution monitoring stations (TAQMSs) were evaluated to comprehend the variations of traffic-related air pollution in Taiwan from 1994 to 2006. The results indicated that the background concentrations of SO2 increased from 1994, peaked at 1997 and decreased to 4.31 ppb at 2006. The concentrations of traffic area SO2 decreased from 16.09 ppb to 7.19 ppb during this period. The background concentrations of CO varied between 0.35 ppm and 0.51 ppm, while the concentrations of traffic area CO decreased from 5.20 ppm to 1.17 ppm from 1994 to 2006. The background concentrations of O3 increased from 1994, peaked at 2004 (34.10 ppb) and maintained at 33.51 ppb at 2006. The concentrations of traffic area O3 varied between 18.70 ppb and 25.44 ppb during this period. From 1994 to 2006, the background concentrations of PM10 varied between 41.55 μ g/m3 and 60.73 μ g/m3, while the concentrations of traffic area PM10 decreased from 119.39 μ g/m3 to 69.62 μ g/m3. The background concentrations of NO2 varied between 13.93 ppb and 16.54 ppb, while the concentrations of traffic area NO2 decreased from 55.85 ppb to 31.68 ppb from 1994 to 2006. The background concentrations of NMHC decreased from 0.61 ppm to 0.11 ppm. The concentrations of traffic area NMHC decreased from 2.52 ppb to 0.72 ppb from 1994 to 2006. From 1994 to 2006, SO2, CO, PM10, NO2 and NMHC from TAQMSs decreased

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

136

Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu et al. by 55.3 %, 77.5 %, 41.7 %, 43.3 % and 71.4 %, respectively. Contrarily, the concentrations of O3 increased by 17.2 %. The traffic area air quality was improved as the transportation loading increased from 1994. Actually, many efforts including vehicle exhaust emission regulation, fuel and mechanical improvement for cars have been made after the signature of APCA in 1975. But according to the analysis, transportation was the major source of SO2, CO, PM10, NO2 and NMHC. For further improving the air quality in Taiwan, TEPA shall adopt stricter regulations for traffic area.

Keywords: traffic area air pollution monitoring stations, SO2, CO, O3, PM10, NO2, nonmethane hydrocarbons (NMHC).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Introduction In the past two decades, air pollution has improved in most cities in Western Europe, North American, Japan as well as Taiwan. Air pollution reductions have resulted mainly from greater efficiency and pollution-control technologies in factories, power plants, and other facilities (Cunningham and Cunningham, 2006). Although improvements are also achieved in transportation, the regulation efficiencies of mobile pollution sources are not as significant as those of stationary pollution sources because their mobile, emitted characteristics (Faiz et al., 1995; Fischer et al., 2000; Kingham et al., 2000; Lipfert et al., 2006; Pai et al., 2007). Air pollution from traffic is one of the main factors considered in the environmental assessment of road establishment. The emission and roadside concentration of those regulated pollutants are potentially harmful to the health or well-being of human, animal or plant life, or to ecological systems. Meanwhile, vehicle emissions may also cause a number of health and nuisance problems (Williams and McCrae, 1995; Künzli et al., 2000; Nicolai, 2002; Vargas, 2003; Monzón and Guerrero, 2004; Wu et al., 2005; Hyder et al., 2006; Pénard-Morand et al., 2006; Finkelstein and Jerrett, 2007). For examples, carbon monoxide (CO) is one of the important components of air pollution caused by traffic exhaust fumes. CO can cause chronic poisoning which reveals its first symptoms as headaches, blurry vision, difficulty in concentration, and confusion (Atimtay, 2002). The results of recent epidemiologic studies also indicated that ground level ozone (O3) can exacerbate asthma symptoms even at concentrations lower than 80 ppb (8-h average) (Delfino et al., 1998). The evidences from various studies have shown that micron particle is associated with morbidity and mortality rates particularly due to cardiovascular and respiratory illness (Pekkanen et al., 1997; Zhao et al., 2004; de Kok et al., 2006; Meng and Lu, 2007). In order to control air pollution, maintain public health and the living environment and improve the quality of life, the Air Pollution Control Act (APCA) of Taiwan was signed in 1975. For controlling the transportation air pollution, Article 16 of Chapter 2 prescribes that the competent authorities at all levels may collect air pollution control fees from stationary and mobile pollution sources that emit air pollutants. For mobile pollution sources, fees shall be collected from the vendor or user based on the types and quantity of air pollutants emitted, or from the vendor or importer based on the type, composition and quantity of fuel. Article 34 of Chapter 3 prescribes that air pollutants emitted by transportation vehicles shall meet the Vehicular Air Pollutant Emission Standards (VAPES). Article 35 prescribes that the owners of transportation vehicles shall maintain the effective operation of the air

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution… 137 pollution control equipment of their vehicle and may not remove or modify this equipment. The type, specifications and labeling of the equipment shall meet the regulations. Article 36 prescribes that the manufacture, import, sale or use of fuel supplied for use in transportation vehicles shall comply with the composition standards and property standards for fuel types. In Article 37, for those in-use motor vehicles that are determined to be unable to meet VAPES for transportation vehicles due to poor design or assembly, the central competent authority shall order the manufacturer or importer to recall for repair within a limited time period. Article 38 prescribes that the manufactured, marketed or imported motor vehicles shall obtain central competent authority issued vehicle model exhaust testing compliance verification before they may apply for license plates. Article 39 prescribes that the central competent authority in conjunction with the Ministry of Transportation and Communications shall determine regulations for the issuance, revocation and cancellation of the vehicle model exhaust testing compliance verification for motor vehicles, and testing and treatment regulations for the air pollutants emissions of transportation vehicles. Article 40 prescribes that the in-use motor vehicles shall undergo regular air pollutant emissions testing. Article 41 prescribes that the competent authorities may perform irregular air pollutant emissions testing or inspections of in-use transportation vehicles at car parks, at airports, at stations, on roadways, in port zones, on water bodies or at other appropriate locations. In this study, the concentrations of air pollutants including sulfur dioxide (SO2), CO, O3, suspended particulates (particles with a diameter of less than 10 micron, PM10), nitrogen dioxide (NO2) and non-methane hydrocarbons (NMHC) from various air pollution monitoring stations (AQMSs) were evaluated to comprehend the variations of traffic-related air pollution in Taiwan from 1994 to 2006.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Methods Area of Study and Measurements In APCA, Article 8 of Chapter 2 prescribes that the central competent authority may, based on topographical and meteorological conditions, designate single or multiple special municipalities, counties or cities between which it is possible for air pollutants to circulate as total quantity control zones, determine total quantity control plans, and officially announce and implement total quantity controls. In addition, Article 13 of Chapter 2 prescribes that competent authorities at all levels shall select appropriate locations for the installation of air quality monitoring stations and officially publish air quality conditions at regular intervals. From 1980, the Environmental Protection Administration in the Central Government of Taiwan (TEPA) began installing automatic air quality monitoring stations and by 2006, there were a total of 76 monitoring stations in the entire nation. TEPA concluded that the data collected by these monitoring stations are representational of each individual region and later divided Taiwan into 7 air quality regions (AQRs) according to their geographical and climatic characteristics. These 7 AQRs are Northern, Chu-Miao, Central, Yun-Chia-Nan, Kao-Ping, Ilan and Hwa-Tung AQRs as shown in Figure 1.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

138

Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu et al.

Figure 1. Air quality regions in Taiwan.

Monitoring Stations and Monitoring Items According to the Article 11 in Air Pollution Control Act Enforcement Rules (APCAER), AQMS designated in Article 13 of APCA shall include five types including general air quality monitoring stations (GAQMS), traffic area air quality monitoring stations (TAQMS), industrial area air quality monitoring stations (IAQMS), national park air quality monitoring stations (PAQMS), and background air quality monitoring stations (BAQMS). The established locations and purposes of different AQMSs are listed in Table 1.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution… 139 Table 1. The established locations and purposes of different AQMSs AQMS GAQMS TAQMS IAQMS PAQMS

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

BAQMS

Established locations and purposes established in areas that are densely populated or that may be subject to high pollution or that are able to reflect the air quality distribution in a larger region established in areas of heavy traffic established in windy downwind areas industrial parks established at appropriate sites in national parks established in areas where there is relatively little human pollution or in windy downwind areas in total quantity control zones

In Article 12 of APCAER, the selection of sites for the air quality monitoring stations in Article 13 of this Act shall take into consideration the following factors including: (1) the type of air quality monitoring stations to be established, (2) the distribution and types of pollution sources and pollutant concentration and distribution, (3) topography, terrain, and meteorological conditions, (4) population distribution and traffic conditions, (5) determination of benefit to control strategy effectiveness, (6) urban plan, regional plan, or other land utilization plan. The number of air quality monitoring stations established shall conform to the following principles: (1) In accordance with population and habitable area (buildings, paddies, upland fields), one general air quality monitoring station shall be established for every 300,000 persons in areas with a population density exceeding 15,000 persons per square kilometer; one general air quality monitoring station shall be established for every 350,000 persons in areas with a population density below 15,000 persons per square kilometer. The number of air quality monitoring stations may be increased in special municipalities. (2) The number of other types of air quality monitoring stations shall depend on actual needs. The central competent authority may establish a monitoring center connected with monitoring stations in light of actual needs. The establishment of air quality monitoring station sampling orifices shall conform to the following principles. First, sampling orifices may not be in locations directly affected by pollution from flues or exhaust outlets, etc. Second, avoid disturbance of air flow and pollutant concentration by nearby obstacles. Third, avoid nearby buildings or obstructing surfaces that may affect pollutant concentration. Forth, determine the height of the sampling orifice above the ground in accordance with the vertical concentration distribution of pollutants near the monitoring station. Article 13 of APCAER prescribes the pollutant items which shall be tested in different AQMSs. The required test items include SO2, CO, O3, PM10, NO2 and NMHC. The analytical methods for SO2, CO, O3, PM10, NO2 and NMHC are ultraviolet fluorescence method, nondispersive infrared method, ultraviolet absorption method, β -ray attenuation method (or tapered element oscillating microbalance technology), chemiluminescence method and flame ionization detector, respectively.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

140

Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu et al.

Results and Discussion Different Pollutant Variations in TAQMSs from 1994 to 2006 The data were analyzed individually over a long period of time. Figure 2 shows the average values of different pollutants in TAQMSs from 1994 to 2006. According to Figure 2, all pollutant average concentrations declined year by year from 1994 excepting O3. The concentrations of SO2, CO, PM10, NO2 and NMHC decreased to the lowest values of 6.63 ppb, 1.16 ppm, 61.36 μ g/m3, 31.3 ppb and 0.69 ppm, respectively in 2002, 2004, 1998, 2005

120 SO2 O3

16 14 12

NMHC NO2

100 80 60

10 8 6 4

40

2 0 1994

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

CO PM10

3

20 18

20

1995

1996

1997

1998

1999

2000 Year

2001

2002

2003

O3 , NO2 (ppb), PM 10 (ug/m )

SO2 (ppb), CO, NMHC (ppm)

and 2004. From 1994 to 2006, SO2, CO, PM10, NO2 and NMHC decreased by 55.3 %, 77.5 %, 41.7 %, 43.3 % and 71.4 %, respectively. Contrarily, the concentrations of O3 increased from 19.49 ppb to 22.85 ppb and by 17.2 % from 2000 to 2006 (TEPA, 2006).

2004

2005

0 2006

Figure 2. The average values of different pollutants in TAQMSs from 1994 to 2006.

Comparisons of SO2 Variations between TAQMSs and Other AQMSs Figure 3 (a) depicts the declining trends of SO2 between TAQMSs and other AQMSs from 1994 to 2006. According to Figure 3 (a), the concentrations of SO2 emitted from PAQMSs were the lowest and those from GAQMSs and BAQMSs were the second low. From 1994 to 2000, industries contributed the majority of SO2, but the SO2 from transportation was higher than those from industries. In 2000, SO2 of TAQMSs has exceeded by 1.49 ppb. From 1994 to 2006, the reduction percentages of SO2 for GAQMSs, TAQMSs, IAQMSs, PAQMSs and BAQMSs were 43.0 %, 55.3 %, 68.3 %, 40.4 % and 27.3 %, respectively. Although a high level reduction of 55.3 % for TAQMSs was reached, the reduction of 68.3 % from IAQMSs was even higher. Table 2 lists the correlation coefficients of SO2 between the TAQMSs and other AQMSs. The SO2 from TAQMSs were highly correlated with those from other AQMSs (R > 0.7). It revealed that the behaviors of SO2 from different sections would affect each others.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution… 141

Concentration (ppb)

(a)

20 15

GAQMS

TAQMS

PAQMS

BAQMS

IAQMS

10 5 0 1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Year

Concentration (ppm)

(b)

6 5 4

GAQMS

TAQMS

PAQMS

BAQMS

IAQMS

3 2 1 0 1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2001

2002

2003

2004

2005

2006

(c)

45 40

Concentration (ppb)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Year

35 30 25 20 15 10 5 0 1994

GAQMS

TAQMS

PAQMS

BAQMS

1995

1996

1997

1998

IAQMS

1999

2000

2004

2005

Year

Figure 3. Continued on next page.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

2006

142

Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu et al.

3

Concentration (ug/m )

(d)

130 120 110 100 90 80 70 60 50 40 30 20 10 0 1994

1995

1996

1997

1998

1999

2000

GAQMS

TAQMS

PAQMS

BAQMS

2001

2002

2003

IAQMS

2004

2005

2006

Year

Concentration (ppb)

(e)

60 50 40

GAQMS

TAQMS

PAQMS

BAQMS

IAQMS

30 20 10

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Year

(f) Concentration (ppb)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

0 1994

3 2.5 2

GAQMS

TAQMS

IAQMS

BAQMS

1.5 1 0.5 0 1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

Year

Figure 3. The variations of different pollutants between TAQMSs and other AQMSs from 1994 to 2006. (a) SO2, (b) CO, (c) O3, (d) PM10, (e) NO2 and (f) NMHC.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution… 143 Table 2. The correlation coefficients between the TAQMSs and other AQMSs. Station Pollutants SO2 CO O3 PM10 NO2 NMHC

GAQMS

IAQMS

PAQMS

BAQMS

0.94 0.90 0.58 0.71 0.87 0.78

0.83 -0.18 0.60 0.24 0.65 0.42

0.87 0.67 -0.01 0.08 -0.63 -

0.70 0.61 0.51 0.44 0.03 0.75

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Comparisons of CO Variations between TAQMSs and Other AQMSs The variations of CO between TAQMSs and other AQMSs from 1994 to 2006 are shown in Figure 3 (b). In Figure 3 (b), all CO concentrations were lower than 1.00 ppm excepting those from TAQMSs. Even in 2000, the concentration of CO was still higher than 1.00 ppm. From 1994 to 2006, the reduction percentages of CO for GAQMSs, TAQMSs, PAQMSs and BAQMSs were 40.2 %, 77.5 %, 44.1 % and 28.6 %, respectively. But the CO concentrations from IAQMSs were contrary to the declining trends of other AQMSs, the values increased by 10.0 % from 2000 to 2006. When analyzing the correlation coefficients of CO between the TAQMSs and other AQMSs, the CO from TAQMSs were highly correlated with those from GAQMSs, i.e. R = 0.90, but the CO from TAQMSs were moderately correlated with those from PAQMSs and BAQMSs (0.4 < R < 0.7). The relationship between the CO from TAQMSs and those of IAQMSs revealed a low negative correlation, R = -0.18. It revealed that the reductions of CO from different sections were not completely consistent, additionally, the concentrations of CO from TAQMSs would affect those from GAQMSs.

Comparisons of O3 Variations between TAQMSs and Other AQMSs When comparing with other pollutants, O3 revealed an opposite trends. The O3 concentrations from all AQMSs increased excepting those from IAQMSs. According to Figure 3 (c), the concentrations of O3 emitted from TAQMSs were the lowest. From 1994 to 2006, the O3 concentrations for GAQMSs, PAQMSs and BAQMSs increased by 36.9 %, 5.9 % and 19.4 %, respectively. The O3 concentrations from TAQMSs increased by 17.2 % from 2000 to 2006, contrarily, those from IAQMSs decreased by 2.4 %. The O3 from TAQMSs were moderately correlated with those from GAQMSs, PAQMSs and BAQMSs (0.4 < R < 0.7). But the relationship between the O3 of TAQMSs and those of IAQMSs revealed a very low negative correlation, i.e. R = -0.01. It revealed that the increasing behaviors of O3 from different sections were not completely consistent.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

144

Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu et al.

Comparisons of PM10 Variations between TAQMSs and Other AQMSs According to Figure 3 (d), the concentrations of PM10 from PAQMSs were the lowest and those from BAQMSs were the second low. From 1994 to 2000, transportation section contributed the majority of PM10, the industries contributed the second quantity. From 1994 to 2006, the reduction percentages of PM10 for GAQMSs, TAQMSs and BAQMSs were 16.9 %, 41.7 % and 6.3 %, respectively. The values from IAQMSs varied between 58.34 ppb and 68.22 ppb, without significant reduction. The reduction was not high from PAQMSs, their values varied between 18.83 ppb and 28.13 ppb. The PM10 from TAQMSs were highly correlated with those from GAQMSs (R > 0.7), but the R values between TAQMSs and other three AQMSs (IAQMSs, PAQMSs and BAQMSs) were low. It indicated that the concentrations of PM10 from TAQMSs would affect those from GAQMS.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Comparisons of NO2 Variations between TAQMSs and Other AQMSs The variations of NO2 between TAQMSs and other AQMSs from 1994 to 2006 are shown in Figure 3 (e). In Figure 3 (e), all NO2 concentrations were lower than 30.00 ppb excepting those from TAQMSs. Even in 2000, the concentration of NO2 was still higher than 30.00 ppb. From 1994 to 2006, the reduction percentages of NO2 for GAQMSs, TAQMSs and IAQMSs were 25.7 %, 43.3 % and 27.9 %, respectively. But the NO2 concentrations from PAQMSs and BAQMSs varied between the ranges of 0.92 – 3.02 ppb and 13.93 - 16.54 ppb, respectively. When analyzing the correlation coefficients of NO2 between the TAQMSs and other AQMSs, the NO2 from TAQMSs were highly correlated with those from GAQMSs, i.e. R = 0.87, but the NO2 from TAQMSs were moderately and lowly correlated with those from IAQMSs and BAQMSs. The relationship between the NO2 from TAQMSs and those of PAQMSs revealed a moderately negative correlation, R = -0.63. It suggested that the concentrations of NO2 from GAQMSs and TAQMSs affected each others.

Comparisons of NMHC Variations between TAQMSs and Other AQMSs Figure 3 (f) reveals the variations of NMHC between TAQMSs and other AQMSs. It was noticed that the data from PAQMSs were not available since no detector was installed. All NMHC concentrations were lower than 1.00 ppb excepting those from TAQMSs. Until 2000, the concentration of NMHC began below 1.00 ppb. From 1994 to 2006, the reduction percentages of NMHC for GAQMSs, TAQMSs, IAQMSs and BAQMSs were 44.0 %, 71.4 %, 100.0 % and 82.0 %, respectively. The NMHC from TAQMSs were highly correlated with those from GAQMSs and BAQMSs, i.e. R = 0.78 and 0.75, but the NMHC from TAQMSs were moderately correlated

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution… 145 with those from IAQMSs (0.4 < R < 0.7). It revealed that the concentrations of NMHC from TAQMSs would affect those from GAQMSs and BAQMSs.

Regulations to Control the Transportation Air Pollution

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

According to the analysis, the transportation was the major source of SO2, CO, PM10, NO2 and NMHC. According to the correlation analysis, the transportation would affect the concentrations of SO2, CO, PM10, NO2 and NMHC in GAQMSs. The concentrations of SO2 from IAQMSs and PAQMSs would be affected by transportation. The transportation would affect the concentrations of SO2 and NMHC in BAQMSs. To control the transportation air pollution, TEPA determines many regulations including Air Pollution Control Fee Collection Regulations, Gas Station Gasoline Vapor Recovery Facility Management Regulations, In-Use Motor Vehicle Recall and Correction Regulations, Regulations Governing Issuance, Revocation, and Cancellation of Compliance Certification for Diesel and Alternative Clean Fuel Engine Vehicle Emissions Inspections, Regulations Governing Issuance, Revocation, and Cancellation of Compliance Certification for Gasoline and Alternative Clean Fuel Engine Vehicle Emissions Inspections, Regulations Governing Verification Issuance and Cancellation of Certification Compliance for Motorized Bicycle Configuration Emissions, Regulations Governing the Certification and Authorization of Imported Motor Vehicle Air Pollutants and VAPES. Because these regulations, SO2, CO, PM10, NO2 and NMHC decreased by 55.3 %, 77.5 %, 41.7 %, 43.3 % and 71.4 %, respectively from 1994 to 2006. But the transportation was still the major source of SO2, CO, PM10, NO2 and NMHC. For further improving the air quality in Taiwan, TEPA shall adopt stricter regulations.

Conclusion In order to control air pollution, maintain public health and the living environment and improve the quality of life, the APCA of Taiwan was signed in 1975. After the signature of APCA, all air pollutant average concentrations declined year by year from 1994 excepting O3. The concentrations of SO2, CO, PM10, NO2 and NMHC decreased to the lowest values of 6.63 ppb, 1.16 ppm, 61.36 μ g/m3, 31.3 ppb and 0.69 ppm, respectively in 2002, 2004, 1998, 2005 and 2004. From 1994 to 2006, SO2, CO, PM10, NO2 and NMHC decreased by 55.3 %, 77.5 %, 41.7 %, 43.3 % and 71.4 %, respectively. Contrarily, the concentrations of O3 increased from 19.49 ppb to 22.85 ppb and by 17.2 % from 2000 to 2006. But according to the analysis, the transportation was still the major source of SO2, CO, PM10, NO2 and NMHC. The correlation analysis showed that the transportation would affect the concentrations of SO2, CO, PM10, NO2 and NMHC in GAQMSs. The concentrations of SO2 from IAQMSs and PAQMSs would be affected by transportation. The transportation would affect the concentrations of SO2 and NMHC in BAQMSs. For further improving the air quality in Taiwan, TEPA shall adopt stricter regulations for traffic area.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

146

Tzu-Yi Pai, Keisuke Hanaki, Horng-Guang Leu et al.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

References Atimtay A.T., Emri S., Bagci T., Demir A.U. (2000). Urban CO exposure and its health effects on traffic policemen in Ankara. Environmental Research, 82 (3), 222-230. Cunningham W.P., Cunningham M.A. (2006). Principles of Environmental Science: Inquiry and Applications. New York: McGraw-Hill Company. De Kok, T.M.C.M., Driece H.A.L., Hogervorst J.G..F., Briedé, J.J. (2006). Toxicological assessment of ambient and traffic-related particulate matter: A review of recent studies. Mutation Research-Reviews in Mutation Research, 613 (2-3), 103-122. Delfino R.J., Murphy-Moulton A.M., Becklake M.R. (1998). Emergency room visits for respiratory illnesses among the elderly in Montreal: association with low level ozone exposure. Environmental Research 76, 67–77. Faiz A., Gautam S., Burki E. (1995). Air pollution from motor vehicles: issues and options for Latin American countries. The Science of the Total Environment, 169 (1-3), 303-310. Finkelstein M.M., Jerrett M. (2007). A study of the relationships between Parkinson's disease and markers of traffic-derived and environmental manganese air pollution in two Canadian cities. Environmental Research, 104 (3), 420-432. Fischer P.H., Hoek G., van Reeuwijk H., Briggs D.J., Lebret E., van Wijnen, J.H., Kingham S. (2000). Traffic-related differences in outdoor and indoor concentrations of particles and volatile organic compounds in Amsterdam. Atmospheric Environment, 34 (22), 37133722. Hyder A.A., Ghaffar A.A., Sugerman D.E., Masood T.I., Ali L. (2006). Health and road transport in Pakistan. Public Health, 120 (2), 132-141. Kingham S., Briggs D., Elliott P., Fischer P., Erik L. (2000). Spatial variations in the concentrations of traffic-related pollutants in indoor and outdoor air in Huddersfield, England. Atmospheric Environment, 34 (6), 905-916. Künzli N., Kaiser R., Medina S., Studnicka M., Chanel O., Filliger P., Herry M., Horak F., Puybonnieux-Texier V. (2000). Public-health impact of outdoor and traffic-related air pollution: a European assessment. The Lancet, 356 (9232), 795-801. Lipfert F.W., Wyzga R.E., Baty J.D., Miller J.P. (2006). Traffic density as a surrogate measure of environmental exposures in studies of air pollution health effects: long-term mortality in a cohort of US veterans. Atmospheric Environment, 40 (1), 154-169. Meng Z., Lu B. (2007). Dust events as a risk factor for daily hospitalization for respiratory and cardiovascular diseases in Minqin, China. Atmospheric Environment, 41 (33), 70487058. Monzón A., Guerrero M.J. (2004). Valuation of social and health effects of transport-related air pollution in Madrid (Spain). Science of the Total Environment, 334-335, 427-434. Nicolai T. (2002). Pollution, environmental factors and childhood respiratory allergic disease. Toxicology, 181-182, 317 – 321. Pai T.Y., Hanaki K., Ho H.H., Hsieh C.M. (2007). Using grey system theory to evaluate transportation on air quality trends in Japan, Transportation Research Part D: Transport and Environment, 12 (3), 158-166. Pekkanen J., Timonen K.L., Ruuskanen J., Reponen A., Mirme A. (1997). Effects of Ultrafine and Fine Particles in Urban Air on Peak Expiratory Flow among Children with Asthmatic Symptoms. Environmental Research, 74 (1), 24-33.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Using Monitoring Data to Evaluate the Variations of Traffic-Related Air Pollution… 147

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Pénard-Morand C., Schillinger C., Armengaud A., Debotte G. (2006). Assessment of schoolchildren's exposure to traffic-related air pollution in the French Six Cities Study using a dispersion model. Atmospheric Environment, 40 (13), 2274-2287. Taiwan Environmental Protection Administration (2008). The Annual Assessment Report of the Air Pollution Control in Taiwan for 2006. Taipei: Taiwan Environmental Protection Administration. (In Chinese). Vargas V.M.F. (2003). Mutagenic activity as a parameter to assess ambient air quality for protection of the environment and human health. Mutation Research/Reviews in Mutation Research, 544 (2-3), 313-319. Williams I.D., McCrae I.S. (1995). Road traffic nuisance in residential and commercial areas. The Science of the Total Environment, 169 (1-3), 75-82. Wu J., Lurmann F., Winer A., Lu R., Turco R., Funk T. (2005). Development of an individual exposure model for application to the Southern California children's health study. Atmospheric Environment, 39 (2), 259-273. Zhao L., Wang X., He Q., Wang H., Sheng G., Chan L.Y., Fu J., Blake D.R. (2004). Exposure to hazardous volatile organic compounds, PM10 and CO while walking along streets in urban Guangzhou, China. Atmospheric Environment, 38 (36), 6177-6184.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 149-160

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 7

MOBILE LABORATORIES FOR PARTICLE AND GASEOUS POLLUTANTS U. Wa Tang∗1, Ni Sheng2 and Zhishi Wang1 1

Department of Civil and Environmental Engineering, University of Macau, Macau, China 2 General Studies, Macau University of Science and Technology, Macau, China

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Due to limited budget, installation space and labour resource, permanent monitoring sites are very scattered and thus complex distributions of particulates in various streets and urban environments are neglected. Recently, a number of mobile laboratories have been assembled for a variety of purposes such as monitoring the spatial and temporal distribution, investigating the emission and dispersion characteristics of tailpipe exhaust and aggregating fleet emission characteristics. This paper reviews their assembling of mobile laboratories, experimental designs and measurements. The authors’ work on the determination of gaseous emission factors for individual on-road vehicles is also introduced.

Keywords: review; spatial resolution; temporal resolution; traffic; air quality; vehicle chasing.

1. Introduction Spatial resolution is an important research topic in disciplines which underlying researches involving geographic information. Researches in environmental and urban management have shown that the outcomes can be altered according to the selection of scale in the studies (João, 2002; O’Neill et al., 1996; Osterkamp, 1995). Spatial distribution of air quality in a city can be obtained from monitoring networks or monitoring campaigns. Monitoring networks are established permanently to provide daily air quality information and

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

150

U. Wa Tang, Ni Sheng and Zhishi Wang

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

forecasts. Monitoring campaigns are conducted by installing monitoring stations temporarily in specific sites for particular research objectives. Due to limited budget, installation space, and labour resource, permanent or temporary monitoring sites are very scattered. Air quality assessment of a city based on scattered monitoring sites may be incorrect because nonhomogeneous distribution of air quality is neglected (Dockery et al., 1996; AckermannLiebrich et al., 1997). The error will be particularly significant in some urban areas with complex traffic conditions and urban geometries. By modelling approaches, the authors have shown that the distribution of air quality in a spatial resolution down to address level (see Figure 1a) (Tang and Wang, 2007 and 2008) can better reflect the non-homogeneous distributions than that a spatial resolution of 300 m x 300 m (see Figure 1b) (Wu et al., 2002a,b) even in a small urban area of 8.8 km2 (the Macau Peninsula). Monitoring studies co-conducted by the authors for spatial distributions of particulate pollutions in the Macau Peninsula also showed the significant spatial variation in an urban, neighbourhood or street scale (Wu et al., 2002a, 2003; Qi et al., 2001a,b; Chen et al., 2000). In an urban scale of 8.8 km2, the roadside polycyclic aromatic hydrocarbons levels of dustfall samples at eleven sites varied from 2.72 to 24.83 μg/g (Qi et al., 2001a,b). In an neighbourhood scale of 670m × 200m, toxic volatile organic compounds BTEX at 18 sampling sites varied to a great extent, i.e., the highest total BTEX was 136 times higher than that of the lowest (Chen et al., 2000).

Figure 1. Continued on next page. ∗

E-mail address: [email protected]

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Mobile Laboratories for Particle and Gaseous Pollutants

151

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Spatial distribution of CO concentrations by a spatial resolution at (a) address level (Tang and Wang, 2007) and (b) 0.3×0.3 km grids (Wu et al., 2002a,b) in the Macau Peninsula. Columns represent CO concentrations: blue 0–1.5 ppm, red 1.5–2.5 ppm, and black 2.5–5.6 ppm.

In a street scale, the mean concentrations of PM10, PM2.5 and PM1 on the leeward side of two major roads were all higher than those on the windward side during the sampling periods, which revealed that the effect of street canyon exactly occurred (Wu et al., 2002b, 2003). The vertical profiles and chemical analysis suggested that particle concentrations were affected significantly by those sources at ground level from traffic district, e.g., resuspended road dust and tailpipe exhaust from motor vehicles. To measure realistic distributions of particulates and gaseous pollution, a mobile laboratory can be assembled to measure the air pollution continuously on the road or at roadsides in urban areas with various urban forms (urban land use and urban geometries) and traffic conditions. A number of mobile laboratories have been assembled for a variety of purposes, such as monitoring the spatial/temporal distribution of air pollutions (Bukowiecki et al., 2002; Pirjola et al., 2004), investigating the emission and dispersion characteristics of tailpipe exhaust (Kittelson et al., 2000; Vogt et al., 2003; Pirjola et al., 2004; Yli-Tuomi et al., 2005; Giechaskiel et al., 2005), and aggregating fleet emission characteristics (Fruin et al., 2004; Canagaratna et al., 2004). This paper reviews their special research objectives which could affect the assembling of mobile laboratories, experimental designs and measurements. In addition, the paper will introduce a mobile laboratory designed by the authors to determine gaseous emission factors for individual on-road vehicles in the Macau Peninsula with extremely high traffic congestions and complex street configuration.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

152

U. Wa Tang, Ni Sheng and Zhishi Wang

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2. Spatial/Temporal Distribution Bukowiecki et al. (2002) was assembled a mobile laboratory at the Paul Scheerer Institute, Switzerland for a one-year project YOGAM. The main objectives and innovations of the project were the spatially and temporally resolved mapping of aerosol parameters for the Zurich area in detail and the investigation of the indicators for NOx or VOC sensitivity of the ground-near ozone formation in the same area without the need of a dense network of stationary measuring sites. To collect representative air quality samples, a route which covers around 25 km x 35 km was designed for the mobile measurement in downtown Zurich, suburban area (small towns, airport and industry) and rural regions. The mobile laboratory followed the route on a regular base to monitor the seasonal and spatial variation of selected ambient aerosol parameters and trace gases. In addition, short time variations between daytime/nighttime, weekday/Sunday were investigated and short distance variations regarding the proximity to heavy traffic roadways were investigated. To achieve the goal to investigate aerosol parameters and indicators for photochemical reaction, a variety of analysers were equipped in the mobile laboratory to measure particle size distribution (i.e., SMPS from TSI and OPC from Grimm), particle number concentration (i.e., CPC from TSI), active surface area (DC from Matter engineering), black carbon (Aethalometer from Magee), particulate matter < 2.5µm (Betameter from Ebreline) and ambient gases such as CO, CO2, NOx, O3, HNO3, PAN, HCHO and H2O2. Other instruments include global position system GPS, a video camera to record the traffic situation, and analysers for meteorological parameters temperature, pressure, relative humidity, wind direction and global radiation. Considering the loading (1000 kg) and power consumption (1.6±0.2kW) of the large amount of analysers, a diesel van which could deliver 2.8 kW power on the road was selected. A larger truck such as that used by Kittelson et al. (2000) was not considered because the van size is suitable for driving on small roads. Controlling of self-contamination during the measurements was an important issue. According to the objective of project, the main inlet was located above the van front end at a height of 2.35 m so that the ambient air can be caught before it is affected by the own van exhaust at the end and by the turbulence caused by the van front end. The detection limit of the particle and gaseous analysers are in ambient levels. Due to the different sampling system demands for gases and aerosol, two inlet systems were constructed. To better correlate data from the two systems, two individual CO2 monitors were connected so that CO2 concentrations could be used as an indicator for the successful catching of the exhaust plume from the vehicle being followed. Bukowiecki et al. (2002) reported a real case that a small exhaust plume from a truck only hit one inlet. Time resolution of the analyser varies significantly from 1s to ½ hour. Fast response analysers with time resolution < 2s include particle analysers CPC, DC, gaseous analysers for CO, CO2, NOx, HNO, PAN, O3, and analyser for wind direction and GPS; while those of other meteorological analysers are < 1 min. The SPMS, Aethalometer, and gaseous analysers H2O2 and HCHO have time resolution of 1 min to 3 min. The one with the slowest response is the PM2.5 analyser Betameter with a time resolution of ½ hour. In considering the main objective of the project YOGAM is to

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Mobile Laboratories for Particle and Gaseous Pollutants

153

measure the seasonal and spatial variation of a large urban area, time resolutions < ½ hour were in an acceptable range. Nevertheless, according to Kittelson et al. (2000), the average time behind a single vehicle in urban traffic conditions was 15–45 sec and a truck-following experiment showed that the duration in which the exhaust plume of the vehicle being followed was constantly captured and did not exceed 10–15 sec in most cases. According to Tang and Wang (2006), gaseous and particle emissions fluctuated significantly in a few seconds during rapid exchange of rapid exchange of idle, acceleration, cruise and deceleration modes in extremely high traffic congestions. Therefore, potential mobile measurement researchers should consider the time resolution of analysers according to the specific research objectives. To prevent instrument damage from vibration shocks while driving, shock absorbers were installed between the rack shelves and the individual instruments. This is particularly important for the particle counter CPC which is sensitive to shocks. (Note: According to a conversation between the first author and Magee’s president Dr. Anthony Hansen, the black carbon Aethalometer is suitable for on-road measurement under normal driving conditions). Bukowiecki et al. (2002) showed that daytime urban ambient air is dominated by high number concentrations of ultrafine particles (nanoparticles) with diameters 100 km/h). Nucleation mode formation potential measured during the chasing tests was also reproduced in the laboratory tests. With a similar experimental design, Vogt et al. (2003) recorded exhaust particle size distribution data together with exhaust gas concentrations (CO, CO2, and NOx) and compared to data obtained from a chassis dynamometer. Good agreement was found for the soot mode particles which occurred at a geometric mean diameter of approximately 50 nm and a total particle emission rate of 1014 particles /km.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4. Fleet Emission Characteristics Rodes et al. (1998) designed one of the earliest mobile laboratories in University of California, Los Angles in 1997 for vehicle following experiments in a variety of streets such as collector, major and minor arterial, highway, interstate, and carpool in Sacramento and Los Angeles. The mobile laboratory was a sedan which outfitted with instrumentation to continuously record 60-s averages of black carbon (BC) concentrations (Aethalometer from Magee), fine particle counts (LAS-X from Particle Measuring Systems) and CO concentrations (60-s average). Windows were closed; fan was set to either "high" or "low"; and air conditioning was set as needed for comfort. The speed of the mobile laboratory was measured with a digital transducer and the following distance was measured by a laser range finder. The driver’s forward view was captured and recorded via video camera. The video observations were used to assign labels to each BC concentration based on the type of vehicle followed, number of axles, location of exhaust, visibility of exhaust, and freeway or street number. Fruin et al. (2004) used the mobile measurement from Rodes et al. (1998) to calculate mean in-vehicle BC concentrations associated with following vehicle types: gasoline passenger car, diesel passenger car, 3-axle diesel vehicle, 5-axle tractor trailer, transit bus (high exhaust), deliver truck (low exhaust), transit bus (low exhaust) and situation when no vehicle followed. Categorised BC exposures were used to estimate in-vehicle contributions to overall particulate matter exposures. Results showed that the approximately 6% of time spent following diesel vehicles led to 23% of the in-vehicle BC exposure, while the remaining exposure was due to elevated roadway BC concentrations. In-vehicle BC exposures averaged

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

156

U. Wa Tang, Ni Sheng and Zhishi Wang

6 mg/m3 in Los Angeles and the Bay Area, the regions with the highest congestion and the majority of the state’s VMT. The statewide average in-vehicle BC exposure was 4 mg/m3, corresponding to DPM concentrations of 7–23 mg/m3, depending on the Aethalometer response to elemental carbon (EC) and the EC fraction of the DPM. In-vehicle contributions to overall DPM exposures ranged from approximately 30% to 55% of total DPM exposure on a statewide population basis. Thus, although time spent in vehicles was only 1.5 h/day on average, vehicles may be the most important microenvironment for overall DPM exposure. Canagaratna et al. (2004) used a mobile laboratory to follow a representative fraction of the New York Metropolitan Transit Authority bus fleet. The mobile laboratory was equipped with an aerosol mass spectrometer (AMS) to measure the nonrefractory particulate matter (NRPM1) and mass spectra of exhaust gas from the bus. Particle counters and gas analysers for CO2, CO, NO, NO2, N2O, CH4, SO2, and H2CO were also installed. In the experiment, the mobile laboratory followed a selected bus at a distance of approximately 3–15 m. The bus drove through city traffic or through quiet neighbourhoods, making stops to pick up or discharge passengers. The start and end points of the route together with notes on the drive conditions and pertinent vehicle-following details were written to text files and automatically tagged with the appropriate time for further analysis. A database provided by the MTA was used to categorize buses in terms of engine type, age and fuel type so that tailpipe emission indices for different types of buses could be obtained. The emission indices were obtained by referencing the measured NRPM1 mass loading to the instantaneous CO2 measured simultaneously in the plume.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

5. Emission Factors of Individual Vehicles Vehicle emission factors are essential input parameters in air quality modelling. Laboratory tests and real-world tests are two major approaches to measuring vehicle emissions. The types of laboratory tests include chassis dynamometer tests and idle tests. However, the operating conditions of the vehicle in the laboratory tests may not fully represent the complex real-world conditions. Moreover, many high-emitting vehicles would not volunteer for government emission testing, so emission inventories may be underestimated. The types of real-world tests include remote sensing tests and on-board emission measurements. A remote sensing test typically uses a stationary analyzer to measure emissions of moving real-world vehicles. However, it can capture only a snapshot (0.5 sec) emission from each moving vehicle, which cannot reflect continuous emissions of individual vehicles in different operating modes. For the on-board emission measurements, an on-street vehicle carries an analyzer to measure emissions from the vehicle itself. However, similar to the laboratory tests, only a few vehicles are available for measurement because cooperation from the vehicle’s owners is necessary. Based on the measured on-road emissions by Pirjola et al. (2004) (see Section 2), YliTuomi et al. (2005) calculated the overall fuel-based emission factors of the fleet for sizeresolved particle numbers, CO, NO and NOx. Nevertheless, due to the lack of vehicle samples and the low time resolution, the fleet emission factors were calculated and results only reflected general emission characteristics for all vehicles being followed. Recently, the authors successfully obtained the emission factors for each category of vehicle (motorcycle, passenger car, taxi, truck and bus) under traffic congestion conditions at

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Mobile Laboratories for Particle and Gaseous Pollutants

157

the urban hot spots in the Macau Peninsula (Tang and Wang, 2006). In comparison with the traditional emission tests such as chassis dynamometer and idle-emission tests, vehicle chase tests reflect real-world situations and more vehicle samples are available because it is not necessary to obtain the cooperation of the vehicle owners. To measure the CO, HC, NO, CO2 and O2 concentrations from the preceding vehicles, a five-gas analyser licensed by the Physikalisch-Technische Bundesanstalt (the national metrology institute) in Germany (Model DiCom 4000 from AVL, Austria) was assembled in a sedan mobile laboratory with the inlet mounted on the sedan’s front bumper. The detection ranges (ppmv) were as follows: CO 0–100,000; CO2 0–200,000; NO 0–4,000; O2 40,000– 220,000; and HC 0–200,000. With irradiation (IR) measurement, the response time for CO, CO2 and HC is ~100 ms. With electrochemical measurement, the response time for O2 and NO may take more than a second. A laptop computer logged data from the five-gas analyser via a standard RS232 serial port every two seconds and simultaneously recorded the driver’s forward view captured by video camera. In urban hot spots, the local background concentrations of traffic air pollutants are expected to be high, and the measurements from vehicle-following experiments may not represent the emissions of a single vehicle. Therefore, the lowest fifth percentile of values over 1 min was selected to estimate the background concentration (Bukowiecki et al., 2002; Pirjola et al., 2004). This low-percentile method aimed to avoid the influence of signal noise that could be induced by use of the minima method. The selection of 1-min duration was based on the estimation that the average time behind a single vehicle in urban traffic conditions was ~15–45 s. After elimination of the background concentrations by the 1-min fifth percentile method, the CO, HC, and NO emission factors in 2-s intervals were calculated by the fuel-based method (Singer and Harley, 2000). In the experiment, as the effective plume path length and amount of plume measured depend on turbulence and wind, it is only possible to determine molar ratios of CO/CO2, HC/CO2, and NO/CO2. These molar ratios are constants for a given exhaust plume. By using the fuel-based method, the amount of pollutants per kilogram of fuel burned can be expressed in terms of the molar ratios. For the purpose of roadside air quality modelling, fuel-based emission factors (grams per kilogram of fuel) were converted to distance-based emission factors in grams per kilometre based on the fuel consumption data. A total of 178 vehicles were individually followed and labelled and the measurements were re-sampled according to the licence plates, vehicle category, vehicle model and registration year. The CO, HC and NO emission factors for each vehicle category were obtained and compared with those estimated from the traditional MOBILE5 emission model (Hao et al., 2003). It can be seen that the CO and HC emissions of petrol passenger cars obtained from the vehicle chase tests are close to those from MOBILE5. The other emission factors from MOBILE5, however, are significantly higher than those from the vehicle chase tests. To evaluate the emission factors obtained from the vehicle chase tests and MOBILE5, these emission factors were input into an operational air quality model (OSPM) to predict the roadside air quality and results were compared with measured values. Evaluation results showed that the emission factors obtained from the vehicle-following measurement techniques are more realistic than those from MOBILE5 in Macau.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

158

U. Wa Tang, Ni Sheng and Zhishi Wang

Conclusion Assembling mobile laboratory to measure on-road gaseous and particle concentrations has been started since the late 1990s. The mobile laboratory can be a cargo container mounted on a diesel container truck, a diesel van or a petrol sedan. Cargo containers have the advantage of more working and storage space reserved for researchers and a variety of analysers but a container truck is too clumsy to follow a target vehicle on the road. Diesel vans are the most popular choice as mobile laboratories because they provide proper space and electricity supply with a size suitable for driving on narrow roads. While petrol sedans are most suitable to chase vehicles during traffic congestions with dynamic driving conditions. Investigations of real-world particle emission characteristics are the major objectives in the previous mobile laboratory measurement experiments. Nevertheless, with the improvement of time resolution and detection limit for gaseous analysers, the vehicle-following measurement technique has potential to be an important approach to develop on-road gaseous emission factor database because cooperation from the vehicle’s owners is not necessary.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

References Ackermann-Liebrich, U, Leuenberger, P., Schwartz, J., Schindler, C., Monn, C., Bolognini, G., Bongard, J.P., Brandli, O., Domenighetti, G., Elsasser, S., Grize, L., Karrer, W., Keller, R., Keller-Wossidlo, H., Kunzli, N., Martin, B.W., Medici, T.C., Perruchoud, A.P., Schoni, M.H., Tschopp, J.M., Villiger, B., Wuthrich, B., Zellweger, J.P., Zemp, E. (1997) ‘Lung function and long – term exposure to air pollutants in Switzerland’, Am Crit Care Med, Vol. 155, pp. 122–129. Bukowiecki, N., Dommen, J., Prévôt, A.S.H., Richter, R., Weingartner, E., Baltensperger, U. (2002) ‘A mobile pollutant measurement laboratory – measuring gas phase and aerosol ambient concentrations with high spatial and temporal resolution’, Atmospheric Environment, Vol. 36, pp. 5569–5579. Canagaratna, M.R., Jayne, J.T., Ghertner, D.A., Herndon, S., Shi, Q., Jimenez, J.L., Silva, P.J., Williams, P., Lanni, T., Drewnick, F., Demerjian, K.L., Kolb, C.E., Worsnop, D.R. (2004) ‘Chase Studies of Particulate Emissions from In-use New York City Vehicles’, Aerosol Science and Technology, Vol. 38, pp. 555–573. Dockery, D.W., Cunningham, J., Damokosh, A.I., Neas, L.M., Spengler JD, Koutrakis P, Ware JH, Speizer FE (1996) ‘Health effects of acid aerosols on North American children: respiratory symptoms’, Environ Health Perspect, Vol. 104, pp. 500–505. Fruin, S.A., Winer, A.M., Rodes, C.E. (2004) ‘Black carbon concentrations in California vehicles and estimation of in-vehicle diesel exhaust particulate matter exposures’, Atmospheric Environment, Vol. 38, pp.4123–4133. Gäggeler, H. W., Baltensperger, U., Emmenegger, M., Jost, D.T., Schmidt-Ott, A., Haller, P., Hoffmann, M. (1989) ‘The Epiphaniometer, a new device for continuous aerosol monitoring’, J. Aerosol Sci. Vol. 20, pp. 557-564. Giechaskiel, B., Ntziachristos, L., Samaras, Z., Scheer, V., Casati, R., Vogt, R. (2005) ‘Formation potential of vehicle exhaust nucleation mode particles on-road and in the laboratory’ Atmospheric Environment, Vol. 39, pp.3191–3198.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Mobile Laboratories for Particle and Gaseous Pollutants

159

João, E. (2002) ‘How scale affects environmental impact assessment’, Environmental Impact Assessment Review, Vol. 22, pp.287-306. Kittelson D, Johnson J, Watts W, Qiang W, Drayton M, Paulsen D, Bukowiecki N (2000) ‘Diesel aerosol sampling in the atmosphere’, SAE Technical Paper Series, No. 2000–01– 2212. O’Neill, R., Hunsaker, C., Timmins, S., Jackson, B., Jones, K., Riitters, K., Wickham, J. (1996) ‘Scale problems in reporting landscape pattern at the regional scale’ Landsc Ecol Vol. 11(3), pp. 169–180. Osterkamp, W. (1995) Effects of scale on interpretation and management of sediment and water quality, IAHS Publication, vol. 226. Wallingford, UK: International Association of Hydrological Sciences. Pirjola, L, Parviainen, H., Hussein, T., Valli A, Hämeri K, Aaalto P, Virtanen A, Keskinen J, Pakkanen TA, Mäkelä T, Hillamo RE (2004) ‘“Sniffer’’ – a novel tool for chasing vehicles and measuring traffic pollutants’, Atmospheric Environment, Vol. 38, pp.3625– 3635. Rodes, C., Sheldon, L, Whitaker, D, Clayton, A, Fitzgerald, K, Flanagan, J, DiGenova F, Hering S, Frazier C (1998) Measuring concentrations of selected air pollutants inside California vehicles. Final Report, Contract No. 95–339. California Air Resources Board, Sacramento, CA. Singer, B.C., Harley, R.A. (2000) ‘A fuel-based inventory of motor vehicle exhaust emissions in the Los Angeles Area during summer 1997’, Atmospheric Environment Vol. 34, pp.1783–1795. Tang, U.W. and Wang, Z.S. (2006) ‘Determining gaseous emission factors and driver’s particle exposures during traffic congestion by vehicle-following measurement techniques’, Journal of Air and Waste Management Association, Vol. 56, pp. 1532–1539. Tang, U.W. and Wang, Z.S. (2007) ‘Influences of urban forms on traffic-induced noise and air pollution: results from a modelling system’, Environmental Modelling and Software, Vol.22, pp.1750–1764. Tang, U.W. and Wang, Z.S. (2008) ‘A model system to determine roadside oxides of nitrogen in highly compact urban forms’, invited submission to Transportmetrica. Vogt, R., Scheer, V., Casati, R., Benter, T. (2003). ‘On-road measurement of particle emission in the exhaust plume of a diesel passenger car’, Environmental Science and Technology, Vol. 37, pp. 4070–4076. Wu, Y., Hao, J.M., Fu, L.X., Hu, J.N., Wang, Z.S., Tang, U.W. (2002a). ‘Emission inventory for mobile sources in Macau, China’, J Tsinghua Univ (Sci and Tech), Vol. 42(12), pp. 1601–1604 (in Chinese with abstract in English). Wu, Y., Hao, J.M., Fu, L.X. (2002b). ‘Study on Spatial Distribution of Air Pollutant Concentrations from Sources in Urban Area of Macao’, Shanghai Environmental Sciences, Vol. 21(6), pp. 338-341. Wu, Y., Hao, J.M., Fu, L.X., Wang, Z.S., Tang, U.W. (2002c). ‘Vertical and horizontal profiles of airborne particulate matter near major roads in Macao, China’, Atmospheric Environment, Vol. 36, pp. 4907–4918. Wu, Y., Hao, J.M., Fu, L.X., Hu, J.N., Wang, Z.S., Tang, U.W. (2003). ‘Chemical characteristics of airborne particulate matter near major roads and at background locations in Macao, China’, Science of the Total Environment, Vol. 317, pp. 159–172.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

160

U. Wa Tang, Ni Sheng and Zhishi Wang

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Yli-Tuomi, T., Aarnio, P., Pirjola, L., Mäkelä, T., Hillamo, R., Jantunen, M. (2005). ‘Emissions of fine particles, NOx, and co from on-road vehicles in Finland’, Atmospheric Environment, Vol. 39, pp. 6696–6706.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 161-178

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 8

URBAN TREES AND AIR AMELIORATION CAPABILITY Loretta Gratani, Laura Varone and Maria Fiore Crescente Department of Plant Biology, Sapienza University of Rome, P.le A. Moro 5, 00185 Rome – Italy

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Urbanisation processes have increased pollution levels in urban areas, and vehicular traffic is one of the major sources of air pollution. Nevertheless, understanding the urban ecosystem pollution dynamics requires long-term researches because of several factors involved (i.e. climate, pollution sources, urban characteristics, trees species, traffic density, vehicle type). In the last years, investigators are themselves concentrating on the possibility to develop models that can quantify the role of urban trees in removing pollutants from atmosphere. Nevertheless, these models need many data on the species composition, age, structure, and health. We analysed the main factors affecting atmospheric CO2 and heavy metal concentration in Rome in the long term, and the trees air amelioration capability, considering the most important deciduous and evergreen species widely distributed in the city. These species have a different role in carbon dioxide sequestration and in bioaccumulation of heavy metals. Such a role depends on many factors, including trees species, plant size and leaf longevity. In particular, information on air pollution could be deduced from heavy metal concentration in plant tissues, offering a low-cost information about urban environment quality. Moreover, crown volume is a discriminant factor for carbon sequestration, and leaf longevity is the most important leaf trait changing in response to heavy metal pollution. Nevertheless, incorrect pruning practices can reduce or undo the trees air amelioration role. Moreover, trees contribute to air temperature mitigation by shading and transpiration, thus lowering the energy consumption for air conditioning during summer. It is important to identify the tree species mostly contributing to air amelioration, and the most discriminant plant traits helping long-time monitoring and city management. The results might be exported into other urban areas to improve air quality enhancing social benefits.

Introduction In the last years, much attention has been focused on the release of pollutants in the atmosphere (Gratani et al., 2000; 2008; Brack, 2002; Moreno et al., 2003; Petaloti et al., 2006; Maher et al., 2008). Vehicular traffic is one of the most significant source of air

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

162

Loretta Gratani, Laura Varone and Maria Fiore Crescente

pollution emission in urban areas; the main factors influencing vehicle emissions are vehicle type, technology and fuel used, and the operating mode of the vehicle, that is the speed, acceleration and engine temperature (Saedler et al., 1996). Pollutants are released at ground level and their upward movement is restricted because of tall buildings (Capannesi et al., 1981; Gratani and Crescente, 1999; Gratani et al., 2000; 2008; Pal et al., 2002). Moreover, much attention is actually focused on the atmospheric carbon dioxide increase, which is the dominant greenhouse gas–emission (Nowak and Crane, 2002). The atmospheric concentration of CO2 has increased from 280 ppm to 360 ppm since 1860 (Abdollahi et al., 2000), and its concentration in urban areas is greater up to 50% than the surrounding nonurban areas (Koerner and Klopatek, 2002; Gratani and Varone, 2005). Human and automobile activities produce more than 80% input of CO2 into the urban environment (Koerner and Klopatek, 2002). Moreover, CO2 is site and time dependent (Nasrallah et al., 2003; Salmond et al., 2005), and it is related to weather conditions and urbanistic characteristics (Gratani and Varone, 2005). The greenhouse gases are thought to be contributing to the increase of atmospheric temperature by trapping certain wavelengths of radiation in the atmosphere (Nowak and Crane, 2002). A current estimate of the expected rise in average air temperature, globally due to greenhouse gas concentration increase, is between 1.4 °C and 4.0 °C by the year 2100 (IPPC, 2007). Climatic differences in urban areas compared with rural environments are due in part to artificial surfaces, high levels of fossil fuel combustion, and traffic volume, which increase air temperature through an “urban heat island” effect (Idso et al., 2001; Koerner and Klopatek, 2002; Gratani and Varone 2005). Plants in the city not only have an ornamental role but they may have also a role in regulating environmental functions. In particular, trees may improve the quality of urban life (Akbari, 2002; Brack, 2002; Gratani and Varone 2006), because they can uptake and accumulate pollutants through their roots and leaf surfaces (Sawidis et al., 2001). Leaves can act as biological absorbers of pollutants (Gratani et al., 2000; Pal et al., 2002), and evergreen species are better traps for heavy metal pollution than deciduous ones, because of their longer leaf longevity, which can accumulate pollutants throughout the year. Particles can be kept as internal leaf accumulation and/or retention on the leaf surface (Gratani et al., 2000; Gratani et al., 2008). A great deal of researches have been focused on the chemical composition of the atmospheric particulate matter having detrimental effects on human health (Pearson et al., 2000; Röösli et al., 2001; Petaloti et al., 2006). Thus, information on atmospheric pollution may be deduced from the concentration of specific substances in plant tissues (Wolterbeek and Peters, 2000; Kapusta et al., 2006), offering low-cost information about the environmental quality (Çelik et al., 2005; Rossini Oliva and Espinosa, 2007). Moreover, trees act as sink for CO2 by fixing carbon during photosynthesis and storing the excess as biomass (Nowak and Crane, 2002). Crown volume is a discriminant factor for carbon sequestration, as leaf size and crown leaf density. The partitioning of stored CO2 for a typical forest tree is about 50% in trunk, 30% in branches and stems, and 3% in foliage (Birdsey, 1992). Larger trees extract and store more carbon dioxide from the atmosphere having a greater leaf area to trap air borne pollutants (Brack, 2002; Gratani and Varone 2006; 2007). Trees represent also excellent regulators of air temperature, heat and dampness in urban surroundings (Hobert et al., 1982); they provide shade, and their transpiration cools air beneath canopies, which can mitigate the urban “heat island” effect and lower energy

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

163

consumption for air conditioning, by altering the heat balance of the entire city (Akbari, 2002). The effects of humans and urbanisation on ecosystem processes can be better understood using the urban areas as a study tool in comparison with “control” sites (Groffman et al., 1995); moreover, cities may provide effective “natural laboratories” for global change impact studies (Idso et al., 2002). In the last years, the investigators are themselves concentrating on the possibility to develop models that can quantify the role of urban trees in removing pollutants from atmosphere (Nowak and Crane, 2000; Brack, 2002). These models need many data on the species composition, habitus, age, structure, and health. Increasing the number of trees in urban areas may potentially slow the accumulation of CO2 atmospheric concentration (Moulton and Richards, 1990) and heavy metal pollutants (Gratani et al. 2000; 2008). The overall objective of this research was to analyse the atmospheric pollution and CO2 concentration changing in the long term in Rome, and the air amelioration capability by the presence of tree species. To this aim, we have also used data collected by Capannesi et al. (1981), Gratani and Crescente (1999), Gratani et al. (2000; 2008), and Gratani and Varone (2005; 2006; 2007). The city of Rome may represent an ideal system to study the dynamic of pollution in relationship with the main affecting factors, and of which results may be exported in other urban areas. Understanding the relationship among urban trees, people, and the environment can facilitate future urban designs that might enhance the environmental and social benefits from trees (Dwyer et al., 1992).

Materials and Methods

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Site Description The study was carried out in the city of Rome (41°53’ N; 12°29’ E), characterised by a low altitude (the highest point is Mount Mario, 139 m on the sea level) (Zapparoli, 1997). From a geologic point of view, the physical environment of the city has been developed during the Pleistocene. The roman territory is comprised between two distinguished volcanic districts: to south-east the Albani Hills and to north-west the Sabatini Mountains, characterised mostly by explosive activity with magmas of alkaline - potassic type (Funiciello et al., 1995). Rome is under a Mediterranean type of climate. The average total year rainfall is 676 mm; the average minimum air temperature in the coldest month (February) is 5.5±1.8 °C and the average maximum air temperature in the hottest months (July and August) is 30.8±0.2 °C. Most of the total rainfall is distributed in autumn and winter (data provided by the Meteorological Station of the Collegio Romano, for the period 1995–2006). The natural vegetation of the city is constituted from strips of persistent meadows of the suburban areas, trampled down environments, shrubs, ruderal or nitrophilous vegetation, and fragments of deciduous and evergreen woods (Pignatti, 1995). Most of the roads are characterised by the presence of Quercus ilex L., Platanus hybrida Brot. and Pinus pinea L.; Quercus pubescens Willd. is present in the urban parks and in some roads. The city of Rome is characterised by a high extension of the urbanised area and movement of public and private means of transport (6.1 millions daily movements; data from ATAC S.p.A – Mobility Agency, 2004). The urbanisation process in Rome has been

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

164

Loretta Gratani, Laura Varone and Maria Fiore Crescente

increasing during the last years, and many new sub-urban areas have been built by scaling down free areas surrounding the city, which is changing into a mega city (129.000 ha of urbanised area, 2.810.931 inhabitants, with 32.569 in the historical centre). Rome is characterised by a considerable value of the archaeological presences, a centre of high historical value, and a high volume of “green” (15% of the total area). The urban sites were selected on the basis of the pollution data monitored by Capannesi et al. (1981), Gratani and Crescente (1999), and Gratani et al. (2000; 2008). It was possible to differentiate roads exhibiting high traffic level, (HT), low traffic level (LT ), an urban park (UP), and a surrounding zone (S) outside the city. We would verify the atmospheric heavy metal pollution levels changing in the long time in the historical centre in response to the limited traffic zone (LTZ), introduced by the City Council in the 1989, and implemented in the year 2001 by the introduction of electronic gates to control the access. The LTZ was active Monday to Friday, from 6.30 to 18.00, and Saturday from 2.00 to 18.00. At the end, data concerning heavy metal concentration in Q. ilex leaves, and collected in the year 2006, were compared with the data previously collected at the same sites, either in the year 1979, before the green fuel and catalytic exhaust introduction (Capannesi et al., 1981) and in the year 1996 (Gratani et al., 2000), when LTZ was active but without electronic gates.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Traffic Density The traffic density was monitored from 8:00 to 11:00 (the peak hours) at each HT, LT and S site. The cars frequency (number of cars per minute) at the investigation sites was recorded in winter and summer (mean of the first 10 days of November and December, and July and August, respectively). For this section we used published data by Gratani and Varone (2005), and unpublished data collected in the years 2006–2007.

Carbon Dioxide Concentration and Urban Climate Atmospheric carbon dioxide concentration (CO2, ppm) was monitored, monthly, using a CO2 gas analyser EGM–1 (PP Systems, UK), at a height of 2 m above the ground at each HT, LT, UP and S sites, according to Idso et al. (2001). Measurements were carried out during the first 10 days of each month from 7.00 a.m. to 19.00. At the same sites and hours of CO2 measurements, air temperature (T, °C) and relative air humidity (RH, %) were measured using a portable thermohygrometer (HD 8901, Delta Ohm, I). Moreover, related “below-canopy” microclimate (air temperature, TBC, °C, and air humidity RHBC, %) during summer was also analysed with the same instruments to determine the degree to which tree cover influences the climate surrounding people and houses, according to Heisler et al. (1994). For this section we used published data by Gratani and Varone (2005; 2006), and unpublished data collected in the year 2007.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

165

Tree Structure Tree structure was analyzed at HT, LT and UP sites, by individual trees of the selected species, comparable per size. Tree diameter at breast height (dbh, cm) was measured by a wheel. Plant (H, m) and crown height were measured by a clinometer; the height of the crown was calculated as the difference between plant height and the insertion point of the first coppers. Crown volume (Vc, m3) was calculated by approximating each crown to a simple geometric solid, according to Karlik and Winer (2001). Leaf Area Index (LAI) was measured by the “LAI 2000 Plant Canopy Analyzer” (LICOR Inc., Lincoln, USA), and Leaf Area Density (LAD, m-1) was calculated by the ratio of LAI and crown height (Küppers, 2003). Total photosynthetic leaf surface area (TPS, m2) was determined multiplying each LAI value by the projected crown area (Gratani and Varone, 2006). The projected crown area was measured by projecting the edges of the crown to the ground and measuring the length along an axis from edge to edge through the crown centre. TPS resulting from not correct pruning practice, comparing both pruned and unpruned trees, was analysed for each species. For this section we used published data by Gratani and Varone (2006), and unpublished data collected in the year 2007 for P. pinea and P. hybrida.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

CO2 Sequestration The total yearly CO2 sequestration capacity was calculated multiplying the photosynthetic rate of each crown by the total photosynthetic activity hours in the year. Total photosynthetic activity hours in the year was obtained by summing the hours in the year with positive net photosynthesis; photosynthetic rate of each crown was obtained multiplying TPS by the mean yearly photosynthetic rate per species, according to Gratani and Varone (2006). Net photosynthesis was monitored in the morning (10:30 – 12:30), by an infrared CO2 gas analyser ADC-LCA3 (Hoddesdon, U.K.) open system, equipped with a leaf chamber (type PLC3). For this section we used published data by Gratani and Varone (2006), and unpublished data collected in the year 2007 for P. pinea and P. hybrida.

Chemical Analysis Three Q. ilex trees, of comparable size (on an average, plant height was 16±1 m and diameter at breast height 50±4 cm, mean value ± S.E.), were selected within the limited traffic zone (LTZ) in the centre of the city. These trees were the same ones investigated in the years 1979 and 1996. Three leaf samples (per tree), of 250 g each, were collected from the south-eastern side of the lower crown portion, a week after the last rainfall, according to Capannesi et al. (1981) and Gratani et al. (2000). One year old sun leaves were collected.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

166

Loretta Gratani, Laura Varone and Maria Fiore Crescente

Leaf samples were stored in polyethylene bags and transferred immediately to the laboratory. Chemical analyses were carried out following the same analytical procedure used by Capannesi et al. (1981) and Gratani et al. (2000). Two leaf sub–samples, without petioles, were selected and one of them was washed with deionised water and an aqueous solution of 5% Triton–x100 (Ratcliffe and Beeby, 1980) to remove superficial leaf deposit. Both washed and unwashed leaf samples were oven dried at 90 °C to constant mass, and pulverised by a Retsch PMA pulveriser. The concentration of Al, Fe, Cu, Zn, and Pb was analysed in triplicate by flame atomic absorption spectrometry (AAS; maximum versatility 82547), after hashing them in a muffle furnace (Haeraeus). The hashed samples were digested by a hotplate at 80–90 °C for 3 h in 20 ml of 1:1 (w/w) (constant boiling) HCl pure water solution, according to Wheeler and Rolfe (1979). The leaf internal metal concentration was calculated as the difference between unwashed and washed leaf samples. For this section we used published data by Capannesi et al. (1981), and Gratani et al. (2000; 2008).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Statistic Analysis Statistical significances of the mean values of the considered variables were analyzed by a one-way analysis of variance (ANOVA), followed by the Tukey test for multiple comparisons. A multiple regression analysis was carried out using CO2 concentration as the dependent variable and air temperature, air humidity and traffic density as independent variables. Two Principal Component Analysis (PCA) were used to get associations of metal factors in order to identify their trend in 1979 and 2006. PCA was carried out on the basis of a matrix of the normalised data using unwashed leaf samples (Gratani et al., 2008). All statistical tests were performed using a statistical software package (Statistica, Statsoft, USA).

Results Traffic Density Traffic density changed significantly (p < 0.05) among the considered sites, the highest one being monitored at HT sites (on an average 95±10 cars min-1), decreasing at LT sites (51±12 cars min-1), and S site (< 1 car min-1). Within the city, weekday traffic density was significantly (p < 0.05) greater (22%) in winter than in summer (Figure 1). Weekend traffic density was significantly (p < 0.05) lower (73%) than weekday one (Figure 1).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

167

winter

summer

weekday

weekend 0

20

40

60

80

100

Number of cars min-1 Figure 1. Traffic density (number of cars min-1) in winter and summer, and during the weekday and weekend. Mean value (± S.E.) is shown. Differences between winter and summer, and between weekday and weekend, are significantly different (ANOVA, p < 0.05).

Urban Climate

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

T monitored in summer, in the city centre (HT and LT sites, mean value), was on an average 4.1 °C greater than in UP sites and S one (Table 1). RH was lower in the city centre (43.5±0.7 %) than in UP sites and S one (48.3±0.4 and 61.1±0.6 %, respectively). Table 1. Results of the microclimatic analysis at the city centre (mean value of high and low traffic level sites), urban park (UP) and the surrounding zone (S), in summer and winter during the study period. T (air temperature), RH (relative air humidity). The mean values with different letters are significantly different (ANOVA, p < 0.01). Mean value (± S.E.) is shown T (°C)

RH (%)

sites City centre Summer Winter

32.6±1.1 a 13.8±1.3 a

sites UP

S

29.0±1.3b 28.1±0.5c 13.5±0.3a 12.9±0.5b

City centre 43.5±0.7a 69.0±3.2a

UP 48.3±0.4b 72.0±4.1a

S 61.1±0.6c 87.3±0.4b

T and RH not significantly differed between the city centre (mean of HT and LT sites) and S site in winter (Table 1). Air temperature and air humidity in the city centre were affected by plant species during summer; TBC Q. pubescens monitored plants were on an average 3.0, 4.2 and 4.9 °C lower than those ones monitored around Q. ilex, P. hybrida and P. pinea plants, respectively (Table

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

168

Loretta Gratani, Laura Varone and Maria Fiore Crescente

2). Moreover, RHBC was higher below the two deciduous species canopy (on an average 52.0±0.5 %) than below the two evergreen ones (39.7±0.9 %) (Table 2). Plant structure did not significantly affect air temperature and air humidity in the UP site (Table 2). Table 2. Results of the microclimatic analysis “below the canopy” of Q. pubescens, Q. ilex, P. hybrida and P. pinea at the city centre (mean value of high and low traffic level sites) and the urban park (UP). TBC = air temperature below the canopy; RHBC = relative air humidity below canopy. In each row the means value with different letters are significantly different (ANOVA, p < 0.05). Mean value (± S.E.) is shown

Sites

Q. pubescens

Q. ilex

P. hybrida

P. pinea

32.7±0.4c 27.0±0.3a

33.4±0.2c 27.8±0.4a

51.4±1.2a 56.3±1.0a

38.7±1.0b 45.3±0.8a

TBC (°C) City centre UP

28.5±0.3a 27.3±0.1a

31.5±0.2b 27.5±0.3a RHBC (%)

City centre UP

52.5±2.1a 56.1±1.2a

40.6±2.1b 50.1±0.9a

The mean annual CO2 concentration has had an upward trend since 2001, with a 30% increase in the year 2007 as to the 1995 (Figure 2). The CO2 concentration varied during the year decreasing significantly (p < 0.05) 15% in summer compared to winter (Figure 3). 600

b

500

CO2 (ppm)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Carbon Dioxide Trend

400

a

b

b

b

b

b

b

a

300 200 100 0 1995 1998 2001 2002 2003 2004 2005 2006 2007 Study years

Figure 2. Annual atmospheric CO2 trend during the study years. Each bars is the mean (± S.E.) of 12 months. Mean values with the same letters are not significantly different (ANOVA, p > 0.05).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

169

600

CO2 (ppm)

500

*

400 300 200 100 0 Winter

Summer

Figure 3. Atmospheric CO2 concentration in winter and summer. The mean (± S.E.) of winter and summer months are shown. * p < 0.05. 550 A

b c

500 a

450 CO2 (ppm)

c

e C

C D

400 350

a

ab

af

df

cd

B

a

ab

c

D

c

E F a

a

300 City centre

250

UP S

200 7:00

8:00

9:00

11:00

13:00

15:00

17:00

19:00

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Hours of day

Figure 4. Mean daily trend of the atmospheric CO2 concentration at different sites: city centre (mean value of high and low traffic level sites), urban park (UP) and surrounding site (S). Each point is the mean hourly value (± S.E.). For each site hourly mean values with the same letters are not significantly different (ANOVA, p > 0.05).

The data revealed the existence of a strong urban CO2 dome, which exhibited a peak CO2 concentration (477±10 ppm) in the city centre (mean of HT and LT sites), 11 and 27% greater, respectively, than UP sites and S site. During the day the highest CO2 concentration was monitored in the first hours of the morning, from 7:00 to 9:00 (492±2 ppm, mean value), decreasing 10% at 3:00 p.m., and increasing 5% at 17:00 p.m. in the city centre (mean of HT and LT sites) (Figure 4). In the park, the highest values were monitored at 7:00 (499±4 ppm), decreasing 9% at 8:00, and 20% at 15:00 (Figure 4). The daily CO2 concentration in S site ranged 368±9 ppm to 385±5 ppm (Figure 4).

Tree Structure The structural crown traits of all the considered trees were significantly (p < 0.05) different between the city centre (HT and LT sites, mean value) and the UP site (Table 3). LAI

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

170

Loretta Gratani, Laura Varone and Maria Fiore Crescente

values were on an average 15, 12, 11 and 10% lower at the city centre than in UP for Q. pubescens, P. pinea, P. hybrida, and Q. ilex, respectively (Table 3). TPS and VC was 19, 27, 30 13 % and 21, 48, 31 and 17 %, respectively, lower at the city centre (HT and LT sites, mean value) than at UP site, for Q. pubescens, Q. ilex, P. hybrida and P. pinea, respectively (Table 3). LAD was 50, 15, 18 and 29% higher at UP site than at the city centre (HT and LT sites, mean value), for Q. pubescens, Q. ilex, P. hybrida, and P. pinea, respectively (Table 3). Incorrect pruning practices reduced, on an average, TPS 40% in Q. pubescens and Q. ilex, 70% in P. hybrida, and 10% in P. pinea. Table 3. Tree crown traits of the considered species at the city centre (mean value of high and low traffic level sites) and the urban park (UP). dbh (diameter at breast height), H (tree height), TPS (total photosynthetic leaf surface), LAI (leaf area index), LAD (leaf area density), Vc (crown volume). Mean value (± S. E.) and significant level (p) are shown City centre

UP

p

TPS (m2) LAI LAD (m-1) Vc (m3) Q. ilex (dbh = 61.0 ± 1.0 cm; H = 14.3 ± 0.8 m)

661±93 3.50±0.30 0.34±0.08 533 ± 52

821±67 4.10±0.10 0.51±0.07 672 ± 30

< 0.05 < 0.01 < 0.05 < 0.001

TPS (m2) LAI LAD (m-1) Vc (m3) P. hybrida (dbh = 65.2 ± 3.1 cm; H = 20.1 ± 2.3 m)

475±59 4.40±0.30 0.47±0.05 252 ± 19

650±28 4.90±0.20 0.54±0.04 485 ± 20

< 0.001 < 0.01 < 0.05 < 0.001

TPS (m2) LAI LAD (m-1) Vc (m3) P. pinea (dbh = 48.5 ± 5.3 cm; H = 19.1 ± 2.6 m)

340±12 2.30±0.1 0.11±0.01 200 ± 9

490±10 2.60±0. 1 0.13±0.01 290 ± 11

< 0.05 < 0.05 < 0.05 < 0.05

TPS (m2) LAI LAD (m-1) Vc (m3)

148±11 2.63±0.05 0.14±0.01 62 ± 6

170±8 3.00±0.050 0.18±0.01 75 ± 4

< 0.05 < 0.05 < 0.05 < 0.05

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Q. pubescens (dbh = 62.0 ± 3.0 cm; H = 15.0 ± 1.4 m)

CO2 Sequestration CO2 sequestration was the highest in Q. pubescens (185±7 Kg year-1), followed by Q. ilex, P. hybrida, and P. pinea (151±10, 100±7 and 51±6 Kg year-1, respectively).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

171

Q. Ilex leaf Metal Concentration The metal concentration of Q. ilex unwashed and washed leaf samples, monitored in the years 1979, 1996, and 2006 are shown in Figure 5. Significant differences (p < 0.001) in the metal concentration among the unwashed leaf samples collected in 1979, 1996 and 2006 were found (Figure 5). Comparing unwashed and washed leaf samples collected in the year 2006, the leaf internal metal concentration of Al, Fe, Cu, Zn, and Pb was 28, 22, 40, 77, and 37%, respectively, of the total concentration, and it was in the same range monitored in the year 1979 and in the year 1996. The total leaf metal concentration (superficial and internal leaf concentration) monitored in the year 2006 was, on an average, 92% lower than the one monitored in the year 1979. 5000

5000

Washed

4000

Fe

Al

3000

2000

1000

Metal concentration (ppm)

Metal concentration (ppm)

Unwashed

500

Al

3000

2000

1000

500 Unwashed

400

Pb

Cu

Zn

300

200

Metal concentration (ppm)

Metal concentration (ppm)

Fe

0

0

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4000

Washed 400

Pb

Cu

Zn

300

200

100

100

0

0 1979

1996

2006

1979

1996

2006

Figure 5. Metal concentration in unwashed and washed Quercus ilex leaf samples for the considered years (1979, 1996 and 2006). Mean value (± S.E.) is shown. Interannual differences for each considered metal are significantly different (ANOVA, p < 0.001).

Statistical Analysis Multiple regression analysis between the CO2 concentration and the urban climatic variables (T and RH) and the traffic density was significant (p < 0.05); RH and traffic density were the most significant variables explaining 68% of CO2 concentration variations (Table. 4).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

172

Loretta Gratani, Laura Varone and Maria Fiore Crescente

Table 4. Results of the multiple regression analysis using atmospheric CO2 concentration as the dependent variable and air humidity (RH), air temperature (T) and traffic density as independent variables. Multiple R value, intercept value, unstandardized (Beta coefficient) and standardized (B coefficient) regression coefficients and significance level of the coefficients (p-level) are shown Indipendent variable

RH

T

Traffic density

Beta regression coefficient

0.61

0.09

0.22

B regression coefficient

3.1

0.48

0.60

1 10-7

0.55321

0.0437

Multiple R value

0.68

Intercept

330

p -level

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The results of PCA, carried out using metal concentration data monitored in 1979 and 2006, showed two factors explaining the metal concentration variations between the considered years (Table 5). Factor 1 accounted for 49% of the total variance in 1979, and it was significantly correlated to Pb and Zn, while Factor 2, accounted for 25% of the total variance, and it was significantly correlated to Al and Fe. Table. 5. Factors loadings for PCA of the metal content in Q. ilex unwashed leaf samples, carried out using data collected in the years 1979 and 2006. In the year 1979 Factor 1 accounting for the most of the total variance (49 %) was significantly correlated to Pb and Zn. In the year 2006 Factor 1, explaining the most of the total variance (41 %), was significantly correlated to Al and Fe

1979 Metal Al

2006

Factor 1 Factor 2 Factor 1 Factor 2 -0.213

0.811*

0.963*

-0.135

Pb

0.915*

-0.165

0.179

0.731*

Fe

-0.269

0.850*

0.935*

0.253

Cu

-0.490

-0.463

0.076

-0.752*

Zn

0.895*

-0.263

-0.082

-0.756*

Eingenvalues % of explained variance

2.4 49

1.3 25

2.0 41

1.6 31

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

173

On the contrary, Factor 1, explained 41% of the total variance in the year 2006, and it was significantly correlated to Al and Fe, while Factor 2 accounted for 31% of the total variance and it was significantly correlated to Pb, Cu, and Zn.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Conclusions The results underline that the centre of Rome exhibits climatic differences compared with the surrounding zone outside of the city, and that urban air temperature may be positively influenced by the tree canopy cover. The evergreen and deciduous species both contribute to decreasing air temperature in the city, in particular during the hottest months. Nevertheless, air temperature around Q. pubescens trees is on an average 4°C lower than those measured around the other species; this is due to its largest crown volume compared to Q. ilex, P. pinea, and P. hybrida. The shade cast by trees also reduces glare and blocks the diffuse light reflected from the sky to surrounding surfaces, thereby altering the heat exchange between buildings and their surroundings (Akbari, 2002). Thus, a significant increase in the number of trees could moderate the intensity of the urban “heat island” effect, by altering the heat balance of the entire city. Moreover, the results show the existence of a strong urban carbon dioxide dome, with a peak CO2 concentration at city centre, 11% and 27% greater, respectively, than in the park and in the surrounding zone. The carbon dioxide concentration shows a strong diurnal trend with the highest values in the first hours of the day when the traffic is the highest, confirmed also by the significant correlation between CO2 concentration and traffic density. On the contrary, the CO2 trend in the parks shows a decrease from the first hours of the morning up to 11:00; this is due to the highest daily photosynthetic rates (Gratani et al., 1998). These data suggest the importance of tree presence in urban areas to sequestrate CO2, in particular when the traffic picks up. Gratani and Varone (2005) underline that freeway roads with a high traffic density have a CO2 concentration lower than the roads in the historical centre of the city, but with high buildings on both sides of the roads. The upward movement of CO2 released at the ground level is restricted because of tall buildings; therefore, high building up of pollutants frequently occur causing adverse effects on the environment quality of urban areas. The considered evergreen and deciduous tree species both contribute to decrease carbon dioxide concentration. Nevertheless, carbon sequestration, depends on plant traits and habitus; Q. pubescens by its greater TPS, and its higher yearly mean photosynthetic rate (Gratani and Crescente, 1999), contributes in major role to CO2 sequestration, followed by Q. ilex, P. hybrida, and P. pinea. Moreover, taking into account the leaf longevity (i.e. 12±3, 7±2, 6±1, 21±2 months for Q. ilex, Q. pubescens, P. hybrida and P. pinea, respectively (unpublished data), Q. ilex and P. pinea by their continuous photosynthetic activity, contribute to reduce CO2 throughout the year, and in particular in winter when the traffic picks up, while the deciduous species contributes to this same reduction only from spring to the beginning of autumn. Although metals are naturally present in soils, contamination comes from local sources, mostly industry, agriculture, sewage sludge, waste incineration, and road traffic (Çelik et al., 2005). Zn, Fe and Cu are natural constituents of the soils (Larcher, 2003); nevertheless the source of Zn and Cu in the street dust has been ascribed to corrosion of the metallic parts of

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

174

Loretta Gratani, Laura Varone and Maria Fiore Crescente

cars like engine wear, thrust bearing and brush wear (Divrikli et al., 2003; Al-Khashman and Shawabkeh, 2006). Pb is directly related to the emissions from motor vehicles (Koeppe, 1981; Al-Khlaifata and Al-Khashmanb, 2007). Our results concerning the bio-monitoring metal pollution trend in Rome, over a period of 27 years, confirm the effects of the Council provision regarding the introduction of LTZ and resulting in a significant traffic density lowering in the city centre. The principal component analysis, in fact, underlines that, in the year 1979, when there was a higher Pb concentration in gasoline, the Factor 1, describing most of the total variance, is dominated by Pb and Zn, which are directly correlated to motor vehicles (Capannesi et al., 1981). In the years following 1979, a significant increase of the traffic density has been observed (1.582.754 vehicles registered in 1985 to 2.476.179, in 2006). Since 1989 further measures have been taken in Italy to limit pollution levels increasing: “green” fuel and catalytic exhausts have become compulsory, and the Rome City Council introduced the LTZ in the centre of the city. The effects of these provisions resulted in a significant traffic density lowering, and it is underlined by the Factor 1 of PCA, explaining most of the total variance in the data monitored in 2006, is dominated by Fe and Al, which are typical soil constituents (Rossini Oliva and Mingorance, 2006); on the contrary, the concentration of Zn, Cu and Pb, which are directly related to the traffic (Çelik et al., 2005; Al-Khlaifata and Al-Khashmanb, 2007) are correlated to Factor 2, accounting for a lower variance. The results are in accordance to those of Urbat et al. (2004), Çelik et al. (2005) and AlAlawi and Mandiwana (2007), which underline the relationship between motor vehicle emissions and metal concentrations in leaves of Pinus nigra L., Robinia pseudoacacia L. and Pinus halepensis L. respectively. Gratani and Crescente (1999) and Gratani et al. (2000) show that metal pollution in urban area alters plant phenology and shoot production, reducing the leaf longevity. Moreover, air pollution influences plant and leaf traits: in particular Vc, LAI and LAD are significantly lower at polluted sites than in the park. On the whole, our results suggest that evergreen and deciduous trees play an important role in sequestering CO2, decreasing metal pollution level, and lowering air temperature in the city of Rome. Nevertheless, the incorrect pruning practices may reduce or undo the trees air amelioration role. The choice of plant species for urban areas might be set out favouring those species which characterise the natural environment, but taking into account their own air amelioration capability Plant traits of each species may be used for realise an urban inventory available for urban tree planting programs to ameliorate urban air. Moreover, our results suggest the use of Q. ilex for long-term monitoring of metal concentration in those urban areas where the species is naturally present and widely distributed. The obtained data of carbon dioxide concentration trend in the city of Rome, over a period of nine years and in relationship with traffic density and urbanistic characteristics can be used to forecast the urban carbon dioxide fluxes for a long term as a result of increasing urbanisation levels. These findings may be of relevance in international discussions related to the ongoing rise in the air CO2 concentration and its implications within the context of predicted future global change, according to Idso et al. (2001). Understanding the urban ecosystem dynamics require long-term researches because of the many factors involved. The city of Rome, characterised by the presence of a high volume

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Urban Trees and Air Amelioration Capability

175

of the “green”, may represent an ideal system to study the possibility to improving air quality by the selection of species according to their own amelioration capability.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

References Abdollahi, K. K., Ning, Z. H. and Appeanning, A. (2000). Global Climate Change and the Urban Forest. Baton Rouge, 1391 Highland Road, LA 70802: Franklin Press, Inc. Akbari, H. (2002). Shade trees reduce building energy use and CO2 emissions from power plants. Environmental Pollution, 116, 119-126. Al-Alawi, M. M. and Mandiwana, K. L. (2007). The use of Aleppo pine needles as a biomonitor of heavy metals in the atmosphere. Journal of Hazardous Materials, 148, 43-46. Al-Khashman, O. A. and Shawabkeh, R. A. (2006). Metals distribution in soils around the cement factory in southern Jordan. Environmental Pollution, 140, 387-394. Al-Khlaifata, A. L. and Al-Khashmanb, O. A. (2007). Atmospheric heavy metal pollution in Aqaba city, Jordan, using Phoenix dactylifera L. leaves. Atmospheric Environment, 41, 8891-8897. Birdsey, R. (1992). Carbon Storage and Accumulation in United States Forest Ecosystems. Gen. Tech. Rep. WO-GTR-59. Radnor, PA: Northeastern Forest Experiment Station, Forest Service, U.S. Department of Agriculture. Brack, C. L. (2002). Pollution mitigation and carbon sequestration by an urban forest. Environmental Pollution, 116, 195-200. Capannesi, G., Gratani, L., Amadori, M. and Bruno, F. (1981). Le foglie di Quercus ilex monitor del grado di inquinamento in area urbana. Notiziario Comitato Nazionale Energia Nucleare, 3, 33-56. Çelik, A., Kartal, A. A., Akdoğan, A. and Kaska, Y. (2005). Determining the heavy metal pollution in Denizli (Turkey) by using Robinia pseudo-acacia L. Environment International, 31, 105-112. Divrikli, V., Soylak, M., Elic, L. and Dogan, M. (2003). Trace heavy metal levels in street dust samples from Yazgat city center, Turkey. Journal of Trace and Microprobe Techniques, 21, 351-361. Dwyer, J. F., McPherson, E. G., Schroeder, H. W. and Rowntree, R. A. (1992). Assessing the benefits and costs of the urban forest. Journal of Arboriculture, 18, 227-234. Funiciello, R., Marra, F. and Rosa, C. (1995). I caratteri geologici – stratigrafici. In B. Cignini, G. Massari and S. Pignatti (Eds.), L’Ecosistema Roma: Ambiente e Territorio (pp. 29-39). Roma: Fratelli Palombi Editori. Gratani, L. and Crescente, M.F. (1999). Indici di valutazione del grado di inquinamento in area urbana. Archivio Geobotanico, 5, 43-50. Gratani, L. and Varone, L. (2005). Daily and seasonal variation of CO2 in the city of Rome in relationship with the traffic volume. Atmospheric Environment, 39, 2619-2624. Gratani, L. and Varone, L. (2006). Carbon sequestration by Quercus ilex L. and Quercus pubescens Willd. and their contribution to decreasing air temperature in Rome. Urban Ecosystems, 9, 27-37. Gratani, L. and Varone, L. (2007). Plant crown traits and carbon sequestration capability by Platanus hybrida Brot. in Rome. Landscape and Urban Planning, 81, 282-286.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

176

Loretta Gratani, Laura Varone and Maria Fiore Crescente

Gratani, L., Pesoli, P. and Crescente, M. F. (1998). Relationship between photosynthetic activity and chlorophyll content in an isolated Quercus ilex L. tree during the year. Photosynthetica, 35, 507-516. Gratani, L., Crescente, M. F. and Petruzzi, M. (2000). Relationship between leaf life–span and photosynthetic activity of Quercus ilex in polluted urban areas (Rome). Environmental Pollution, 110, 19-28. Gratani, L., Crescente, M. F. and Varone, L. (2008). Long-term monitoring of metal pollution by urban trees. Atmospheric Environment, doi:10.1016/j.atmosenv.2008.07.032. Groffman, R. M., Pouyat, R. V., McDonnell, M. J., Pickett, S. T. A. and Zipperer, W.C. (1995). Carbon pools and trace gas fluxes in urban forests. In R. Lai, J. Kimble, E. Levine and B. A. Stewart (Eds.), Soils Management and Greenhouse Effect: Advances in Soil Science (pp. 147-159). Boca Raton: CRC Lewis. Heisler, G. M., Grimmond, S., Grant, R. H. and Souch, C. (1994). Investigation of the influence of Chicago’s urban forests on wind and air temperature. In E. G. McPherson, D. J. Nowak and R. A. Rowntree (Eds.), Chicago’s Urban Forest Ecosystem: Results of the Chicago Urban Forest Climate Project (pp. 19-40). Radnor, PA: General Technical Report No. NE-186, U. S. Department of Agriculture, Forest Service, Northeastern Forest Experiment Station, Horbert, M., Blume, H.P., Elvers, H. and Sukopp, H. (1982). Ecological Contributions to Urban Planning. Oxford: Blackwell. Idso, C. D., Idso, S. B. and Balling, R. C. Jr. (2001). An intensive two–week study of an urban CO2 dome in Phoenix, Arizona, USA. Atmospheric Environment, 35, 995-1000. Idso, S. B., Idso, C. D., and Balling R. C. Jr. (2002). Seasonal and diurnal variations of near– surface atmospheric CO2 concentration within a residential sector of the urban CO2 dome of Phoenix, AZ, USA. Atmospheric Environment, 36, 1655-1660. IPPC (2007). Climate Change 2007: Mitigation. Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press. Kapusta, P., Szarek-Lukaszewska, G. and Godzik, B. (2006). Spatio-temporal variation of element accumulation by Moehringia trinervia in a polluted forest ecosystem (South Poland). Environmental Pollution, 143, 285-293. Karlik, J. F. and Winer, A. M. (2001). Plant species composition, calculated leaf masses and estimated biogenic emissions of urban landscape types from a field survey in Phoenix, Arizona. Landscape and Urban Planning, 53, 123-134. Koeppe, D.E., 1981. Lead: understanding the minimal toxicity of lead in plants. In N.W. Leep (Ed.), Effects of Trace Metals on Plant Function: Effects of Trace Metals on Plants (pp. 55–76). London: Applied Science Publishers. Koerner, B. and Klopatek, J. (2002). Anthropogenic and natural CO2 emission sources in an arid urban environment. Environmental Pollution, 116, 45- 51. Küppers, M. (2003). Space filling by foliage: the effect of growth form and architecture on the light environment. In W. Larcher (Ed.), Physiological Plant Ecology (4th edition, pp. 40-41). Berlin, Heidelberg, New York: Springer- Verlag. Larcher W. (2003). Physiological Plant Ecology (4th edition). Berlin, Heidelberg, New York: Springer-Verlag.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Urban Trees and Air Amelioration Capability

177

Maher, B. A., Moore, C. and Matzka, J. (2008). Spatial variation in vehicle-derived metal pollution identified by magnetic and elemental analysis of roadside tree leaves. Atmospheric Environment, 42, 364-373. Moreno, E., Sagnotti, L., Dinarès-Turell, J., Winkler, A. and Cascella, A. (2003). Biomonitoring of traffic air pollution in Rome using magnetic properties of tree leaves. Atmospheric Environment, 37, 2967-2977. Moulton, R. J. and Richards, K. R. (1990). Costs of Sequestering Carbon through Tree Planting and Forest management in the United States. Washington, DC: Gen. Tech. Rep. WO-58. USDA Forest Service. Nasrallah, H. A., Balling, R. C. Jr, Madi, S. M. and Al-Ansari, L. (2003). Temporal variations in atmospheric CO2 concentrations in Kuwait city, Kuwait with comparisons to Phoenix, Arizona, USA. Environmental Pollution, 121, 301-305. Nowak, D. J. and Crane, D. E. (2000). The Urban Forest Effects (UFORE) Model: quantifying urban forest structure and functions. In M. Hansen and T. Burk (Eds.), Integrated tools for natural resources inventories in the 21st century: proceedings of the IUFRO Conference (pp. 714-720). St. Paul, MN: Gen. Tech. Rep. NC-212, U.S. Department of Agriculture, Forest Service, North Central Research Station. Nowak, D. J. and Crane, D. E. (2002). Carbon storage and sequestration by urban trees in the USA. Environmental Pollution, 116, 381-389. Pal, A., Kulshreshtha, K., Ahmad, K. J. and Behl, H.M. (2002). Do leaf surface characters play a role in plant resistance to auto–exhaust pollution? Flora, 197, 47-55. Pearson, R. L., Wachtel, H. and Ebi, K. L. (2000). Distance–weighted traffic density in proximity to a home is a risk factor to leukemia and other childhood cancers. Journal of Air and waste Management Association, 50, 175-180. Petaloti, Ch., Triantafyllou, A., Kouimtzis, Th., and Samara, C. (2006). Trace elements in atmospheric particulate matter over a coal burning power production area of western Macedonia, Greece. Chemosphere, 65, 2233-2243. Pignatti, S. (1995) La vegetazione naturale. In B. Cignini, G. Massari and S. Pignatti (Eds.), L’Ecosistema Roma: ambiente e territorio (pp. 54-61). Roma, Fratelli Palombi Editori. Ratcliffe, D. and Beeby, A. (1980). Differential accumulation of lead in living and decaying grass on roadside verges. Environmental Pollution, 23, 279-286. Röösli, M., Theis, G., Künzli, N., Stoehelin, J., Mathys, P., Oglesby, L., Camenzind, M. and Braun-Fahrländer, Ch. (2001). Temporal and spatial variation of the chemical composition of PM10 at urban and rural sites in the Basel area, Switzerland. Atmospheric Environment, 35, 3701-3713. Rossini Oliva, S. and Mingorance, M. D. (2006). Assessment of airborne heavy metal pollution by aboveground plant parts. Chemosphere, 65, 177-182. Rossini Oliva, S. and Espinosa, A. H. F. (2007). Monitoring of heavy metals in topsoils, atmospheric particles and plant leaves to identify possible contamination sources. Microchemical Journal, 86, 131-139. Saedler, L., Jenkins, N., Legassick, W. and Sokhi, R.S. (1996). Remote sensing of vehicle emission on British urban roads. The Science of the Total Environment, 189/190, 155160. Salmond, J. A., Oke, T. R., Grimmond, C. S. B., Roberts, S. M. and Offerle, B. (2005). Venting of heat and carbon dioxide from street canyons at night. Journal of Applied Meteorology, 44, 1180-1194.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

178

Loretta Gratani, Laura Varone and Maria Fiore Crescente

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Sawidis, T., Chettri, M. K., Papaioannou, A., Zachariadis, G. and Stratis, J. (2001). A study of metal distribution from lignite fuels using trees as biological monitors. Ecotoxicology and Environmental Safety, 48, 27-35. Urbat, M., Lehndorff, E. and Schwark, L. (2004). Biomonitoring of air quality in the Cologne conurbation using pine needles as a passive sampler - Part I: magnetic properties. Atmospheric Environment, 38, 3781-3792. Wheeler, G. L. and Rolfe, G. L. (1979). The relationship between daily traffic volume and the distribution of lead in roadside soil and vegetation. Environmental Pollution, 18, 265274. Wolterbeek, H. E. and Peters, A. (2000). Biomonitoring of trace element air pollution: principles, possibilities and perspectives. Environmental Pollution, 120, 11-21. Zapparoli, M. (1997). Gli Insetti di Roma. Considerazioni Introduttive ad uno Studio di Entomologia Urbana. Roma: Fratelli Palombi Editori.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 179-210

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 9

DOWNSIZING DIRECT INJECTION SPARK IGNITION ENGINES: A TIMESCALE ANALYSIS John Shrimpton and Agissilaos Kourmatzis Energy Technology Research Group, School of Engineering Sciences University of Southampton, Highfield Campus Southampton, SO17 1BJ, England

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Downsizing direct injection spark ignition engines presents several challenges to the engine designer, but is a necessary requirement if significant savings in terms of fuel economy and CO2 emissions are to be realised. This challenge becomes more acute if we wish to employ a flexible fuel supply, for instance a range of biological fuel blends. These typically require more mechanical and thermal effort to provide good fuel vapour-air mixture preparation at ignition. In this chapter these issues are investigated by comparing engine timescales (a function of engine size, speed and injection timing) against droplet timescales (a function of drop diameter, liquid physical properties and local air thermodynamic conditions). The analysis is used to make predictions of the target drop diameter required for a given engine size, speed and injection timing (load). Finally, we briefly explore the possibility of employing electrostatic charging to reduce the mass transfer timescale, since this is the limiting timescale for small direct injection spark ignition operation.

Nomenclature a CL CR D, Do f2 f1 HT

Crank length (m) Liquid specific heat (Jkg-1K-1) Ratio of constant pressure gas phase heat capacity to liquid phase heat capacity Droplet diameter, Initial Droplet Diameter (μm) Energy conservation equation correction term for heat transfer component Drag coefficient Energy conservation equation non-equilibrium term

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

180 HM l Le,s LV mD P1,P2 Patm PG Pr Re R ScG Sh r s spiston T1,T2 TB TG TD t uG, uD,utf V1,V2 Wv xcl xs,eq YG, Ys,eq Greek γ εo θ θinj λ ρG ρD σ τmass τmom τheat τe,g ω

John Shrimpton and Agissilaos Kourmatzis Mass Transfer Differential Equation driving term Connecting rod length (m) Engine stroke restriction lengthscale (m) Latent heat of evaporation (Jkg-1) Droplet mass (kg) Pressure at states 1 and 2 (Nm-2) Atmospheric Pressure (Nm-2) Gas Phase pressure (Nm-2) Prandtl Number Reynolds number Ideal Gas constant (Jmol-1K-1) Gas Phase Schmidt Number Sherwood Number Compression ratio Stroke (m) Distance between piston crown and cylinder head (m) Temperature at states 1 and 2 (K) Boiling Point Temperature (K) Gas phase temperature (K) Droplet temperature (K) Time (ms) Gas velocity, Droplet Velocity, Turbulent fluctuating velocity (ms-1) Volume at states 1 and 2 (m3) Molar mass (gmol-1) Clearance height (m) Surface equilibrium mole fraction Gas vapour mass fraction and surface equilibrium mass fraction Ratio of specific heats Permittivity of continuum (C2N-1m-2) Instantaneous Crank angle degrees from Top dead centre (degrees) Instantaneous Crank angle degrees from Top dead centre at injection (degrees) Thermal Conductivity (Jm-1s-1K-1) Gas phase density (kgm-3) Droplet density (kgm-3) Surface tension (Nm-1) Droplet mass timescale (ms) Droplet momentum timescale (ms) Droplet heat-up timescale (ms) Global Engine timescale (ms) Engine Crank-shaft angular velocity (rads-1)

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

181

Abbreviations TDC- Top Dead Centre BDC- Bottom Dead Centre ODE- Ordinary Differential Equation CAD- Crank Angle Degrees

Introduction

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Background Over the last century the internal combustion engine has undergone many changes, especially after the introduction of electronically controlled fuel injection systems and an improved understanding of mixture preparation. In this chapter we explore the direct injection-spark ignition engine which has evolved from the original carburetor-based engine. This was superceded by the port fuel injection engine, due difficulties in reaching a stoichiometric air/fuel mixture [1]. Port fuel injection which involves fuel injection behind the intake valve is advantageous in the sense that the fuel has more time to atomize due to the promotion of turbulence before the entry of the fuel into the combustion chamber. However, as one can imagine, a significant proportion of fuel is poorly utilized due to the presence of a liquid film on the intake valve. Direct injection was therefore introduced which involves, as the words suggest, the injection of fuel directly into the cylinder, so that the intake valve is bypassed. This allows for stratified operation [1] and therefore for the utilization of a leaner fuel mixture and a system that consumes less fuel at part load. The development of electronic high pressure injection systems has led to improved fuel atomization and spray dispersal which in turn provides better fuel economy and improved emissions characteristics. Unfortunately, nothing comes without disadvantages, and in a direct injection system, fuel droplets colliding with the piston crown can lead to inefficient combustion. For late injection and for stratified charge operation there is a reduced amount of time between injection and ignition. Therefore the evaporation and momentum decay that the droplets undergo should complete before the piston reaches Top Dead Centre (TDC) and ignition occurs. This is difficult to achieve, especially for small engines. Furthermore, there is the issue of controlling the spray dispersal from pressure injection systems and of course the fact that we need to supply an injection pressure anywhere from 50 to 120 bars [2]. The above are issues that arise in both small and large engines and the severity of each issue may vary according to engine speed, engine size and fuel type. In particular, the issues become acute as the engine capacity is reduced, and due to the commercial interest in developing smaller internal combustion engines for hybrid engine/battery power plants, this is of relevance today. Therefore, interest exists in a novel system of fuel injection such as the use of electrostatic force in the injection process. This involves injecting an electric charge into the fuel prior to injection. The introduction of an electric charge effectively reduces the surface tension of the fuel droplets and thus promotes the improved atomization of the fuel and also promotes spray dispersal and improved mixture preparation [3]. The system proposed here has been explored through numerous simulations and experiments and is a fairly simple method of achieving droplet destabilization.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

182

John Shrimpton and Agissilaos Kourmatzis

Aims and Objectives In this chapter we want to explore a number of aspects of fuel injection which will outline important aspects of the engine, the fuel that we are to use, and the effects of electric charge and injection timing; the scope of this chapter is as follows: • •





The comparison of timescales of large and small engines for fast and slow engine speeds, as well as early and late injection. Analysis of the dynamics of a single fuel droplet in an engine cylinder. Droplet momentum, mass transfer and heat-up timescales play a very important role in understanding how different types of fuels can be used in various engine conditions. The comparison of droplet timescales with engine timescales. This is of fundamental importance as this will define what the limiting parameters of a particular engine are when interacting with a particular fuel. The investigation of electric charge on the modification of droplet timescales and how this technology allows us to relax engine parameters will be explored.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Plan The chapter will begin with the outlining of various engine parameters of interest. We will look at how engine size, speed and thermodynamic variations affect various restricting timescales of interest. Standard model equations describing droplet dynamics will be presented and explained. Results from a number of studies carried out by the authors relating to engine dynamics will then be presented followed by fuel droplet dynamics. The restricting nature of the various engine timescales will be discussed side by side with the droplet timescales of interest which will effectively bring the limiting engine timescales to the surface.

Definitions and Basic Theory Firstly, we analyze the geometry of a typical internal combustion engine and look at this directly affects important timescales. Figure 1 shows a very simplified version of piston motion in a cylinder found in an internal combustion engine, the various geometric parameters will be outlined in table 1. This is a simplified cross-sectional view where the crank rotates about the out-of page axis at a given angular velocity, ω. Table 1 mainly outlines the geometric engine parameters that are of interest to us in this chapter. The above values are enough to characterize the various engine timescales that we will be looking at and comparing with the droplet timescales shall be calculated.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis Injector TDC

xcl

183

Engine cylinder

s BDC l

a

θ

ω

Figure 1. Schematic of an internal combustion engine cylinder.

Table 1.Basic Engine Parameters of Interest

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Parameter ω l a θ xcl s BDC, TDC Early Injection Late Injection Ignition

Definition Angular Crank shaft velocity Connecting rod length Crank length (Stroke/2) Crank angle degrees (CAD) from TDC Clearance Height Engine Stroke Bottom and Top dead centre Injection@230CAD Injection@330CAD Ignition@360CAD

Units rad/s m m degrees m m

Engine Restrictions Having looked at relevant engine parameters at a very basic level we define exactly what is meant by engine restrictions. There are two restrictions of interest here; The global engine timescale, τe,g and the engine stroke restriction lengthscale Le,s. The global engine timescale is the main limiting time for the fuel droplet and is the time interval between injection and ignition. From an engine design point of view we must ensure that the droplet has lost most of its momentum, heated up and also lost most of its mass via evaporation before ignition. This will depend on the engine speed and will not depend on the engine geometry at all. As an example, say we have an engine spinning at 3000rpm (314.2 rad/s), If we inject at 230 degrees, and ignition occurs at 360 degrees, 130 degrees have elapsed. We know that

ω=

πθ 180t

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(1)

184

John Shrimpton and Agissilaos Kourmatzis

where t is time. Substituting the engine speed and elapsed degrees a time t=7.22ms is calculated. This is the global engine timescale τe,g and is a restricting parameter for the droplet timescales which we will define later on. Of course if we spin faster and inject later the global engine timescale will change so τe,g=f(ω,θinj). We also have to define our second engine restriction of interest, a length-scale we referred to earlier as Le,s. This is dependant on both of the parameters that τe,g is dependant on but it is also a function of the engine geometry. Referring to figure 1, we can derive an expression that will relate various engine geometric parameters conveniently:

s piston = xcl + [ a + l − (l 2 − a 2 sin 2 θ )1/ 2 − a cos θ ]

(2)

In table 1 we defined all the variables on the R.H.S of equation (2). As shown in figure 1, xcl is defined as the clearance height (the distance between the piston crown at TDC and the cylinder head), this can be directly related to the compression ratio of the engine, r which is defined by (3):

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

r=

s + xcl xcl

(3)

spiston is the instantaneous position of the piston which will depend on all of the parameters on the R.H.S of the equation, note that we can relate the current CAD (θ), to the angular velocity and time via equation (1). Having equations (1)-(3) will help us define Le,s; We consider this restriction because if we can define the piston locus at any given point in time (spiston) we can then compare the piston position to that of a droplet undergoing deceleration. This will inform us whether or not the droplet has hit the piston. Assuming they do not collide within the global engine timescale means that our system is acceptable as far as momentum is concerned, i.e. we need to make sure the droplet slows down, heats up and significantly evaporates before ignition (τe,g criterion) while at the same time not colliding with the piston crown (Le,s criterion).

Droplet Timescales-Momentum In the introduction, the term ‘droplet timescales’ was mentioned, in this section we define what we mean by this. The droplet momentum timescale is defined as the time elapsed from the point of injection to when the momentum has significantly decayed. Solving equation (4) yields equation (5) which we can use to define our momentum timescale, τmom.

du D uG − u D = τ mom dt Setting the fluid velocity equal to zero:

uG = 0 Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(4)

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

185

and integrating the ODE : u D ,t



u D ,o

t

duD dt =∫ τ uD 0 mom u (τ ) =

u (o ) e

(5)

So, the final velocity at which most of the momentum has decayed is the (initial velocity*1/e), so the time elapsed from t=0 to t at (initial velocity*1/e) is the momentum timescale, τmom.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Droplet Timescales-Heat We can carry out an essentially identical analysis as that above in order to define the droplet heat-up timescale, τheat. This timescale is the time elapsed from t=0 to a point where the droplet has heated up to a value close to its steady state temperature. The same method as for the momentum timescale shall be employed only that in this case, we reach a steady state maximum. We will treat a heat up from an initial temperature To to a final temperature Tf, as a decay from (Tf-To) to zero, and then employ equation (5) for the case of temperatures. This is easiest to explain using an example: Assume we begin from 290K and heat up to 400K, we treat this as a decay from 110K (400-290) to zero, i.e. we flip the plot over. Noting that 1/e*110=40.46K, i.e. 69.5K from the maximum (from 110K) and now adding that number to 290 (the previous initial temperature, To) equals 359K, this calculated temperature is a value which represents a point in time when the majority of heat up that is to occur in the droplet has already happened.

Droplet Timescales-Mass Our third timescale of interest is the mass transfer timescale, τmass, this is simply defined as the amount of time it takes for our droplet mass to reach 0.1mD,o [4] a point where the majority of the droplet has evaporated.

Engine Thermodynamics We have so far discussed engine parameters such as geometry and speed and have also defined our droplet timescales. What still needs to be looked at is some basic engine thermodynamics. We need to know what the density, pressure and temperature will be at any given point in order to define our in-cylinder conditions which we will use in other

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

186

John Shrimpton and Agissilaos Kourmatzis

calculations. In order to do this we will assume that on the compression stage the engine undergoes a reversible and adiabatic process. Using equation (2), we can calculate the piston position and therefore the volume of air available above the piston, and from this, the temperature (6): 1

⎛ V2 ⎞ ⎛ T2 ⎞ ⎜ ⎟=⎜ ⎟ ⎝ V3 ⎠ ⎝ T3 ⎠

1−γ

(6)

where γ is the ratio of specific heats (of air in this case). We can then also determine the density and pressure at any given Crank angle degree and thus time, this will be of vital importance as we will see later. For all the results presented in this chapter we assume the initial conditions to be ambient i.e initial TG=293K and PG=1atm. Density is calculated from the ideal gas equation:

PG = ρG RTG

(7)

where P is pressure, R is the gas constant for air, and ρG is the gas (air) density.

Fuel Droplet Equations and Physical Properties

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

As stated in the introduction we will be analyzing the dynamics of a fuel droplet within a combustion chamber. By dynamics one means momentum, mass transfer and heat-up timescales, τmom , τmass , and τheat respectively.

Momentum Equations The momentum conservation equation for a drop is given by equation (8) [4,5].

duD f = 1 (uG − uD ) dt τ D

(8)

where uD, is the droplet velocity, uG is the gas velocity, f1 is a drag coefficient, and τD is the particle time constant. Here we use equation (9) for f1 [6].

1 f1 = 1 + Re D 2 / 3 6

(9)

ReD is the droplet Reynold’s number and is given by (10):

Re D =

ρ G uG − u D D μG

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(10)

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

187

where ρG is the gas density, D is the diameter and μg is the gas dynamic viscosity. The particle time constant τD is given by equation (11). [4]

τD =

ρD D2 18μ G

(11)

So by using equations (8) to (11) we can define all momentum characteristics, i.e. we can define how the velocity of the droplet will temporally evolve according to fuel type and incylinder conditions. We can also easily calculate the position of the droplet as we know how its velocity varies over time.

Droplet Heat-Up Equations The energy conservation equation for the drop is equation (12) [4,5].

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

dTd f Nu ⎛ C = 2 ⎜ r 3PrG ⎝ τ D dt

⎞ ⎛ LV ⎞ m d − HΤ ⎟ (TG − Td ) + ⎜ ⎟ ⎠ ⎝ CL ⎠ md

(12)

Equation (12) is a general equation where the terms of the RHS relate to heat transfer, Latent heat and non-equilibrium convection. Here, because the aim of this chapter is to study timescale issues relatively broadly we use a simple rapid-mixing model. No corrections for Stefan flow will be made and we assume that our internal droplet temperature will be uniform thus making HΤ=0 and f2=1. The rest of the terms are those which define the physical properties of the droplet; Nu is the Nusselt number which we derive from the widely used Ranz-Marshall correlation (see Appendices), PrG is the gas phase Prandtl number which we also derive from an empirical correlation (see Appendices). Cr is the ratio of the constant pressure gas phase heat capacity to the liquid phase heat capacity (both of which are derived from correlations). TG is the bulk gas temperature (derived from the adiabatic process equations described in section 2.3), Td is the droplet temperature, Lv is the latent heat of evaporation (derived from correlation), CL is the liquid phase heat capacity, md is the mass of

m

the droplet and d is the mass transfer rate defined by equation (13). So, using the physical property correlations (see Appendices) and equation (12) we can define the temporal evolution of temperature for a given droplet.

Droplet Mass-Transfer Equations The final droplet timescale that we are interested in is that one which will characterize our mass transfer. Equation (13) is the governing ODE for the droplet mass transfer. [4]

dmd Sh =− dt 3ScG

⎛ md ⎜ ⎝ τD

⎞ ⎟ HM ⎠

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(13)

188

John Shrimpton and Agissilaos Kourmatzis

We introduce some more terms here. Sh is the Sherwood number which we can calculate using a Ranz-Marshall correlation (see Appendices), ScG is the gas phase Schmidt number which is equal to μG/ρGΓG where ΓG is the binary mass diffusion coefficient. HΜ in equation (13) is effectively the mass transfer driving term and for the rapid mixing model this is equal to ln[1+BM,eq] where BM,eq is the equilibrium Spalding mass transfer number [4] which depends on the equilibrium vapour mass fraction (YS,eq) and the free stream vapour mass fraction away from the droplet surface (YG) (see equation 14). [4]

BM ,eq =

Ys ,eq − YG 1 − Ys ,eq

(14)

YS,eq is calculated from the equilibrium assumption and the well known ClausiusClapeyron equation (equations 15 and 16 respectively). [4]

Ys ,eq =

xs ,eq =

xs ,eq xs ,eq + (1 − xs ,eq )θ 2

(15)

⎡ L ⎛ 1 1 ⎞⎤ Psat Patm exp ⎢ V ⎜ − ⎟ ⎥ = PG PG ⎣⎢ R / Wv ⎝ TB Td ⎠ ⎦⎥

(16)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

where xs,eq is the equilibrium mole fraction of the vapour, Patm is atmospheric pressure, Psat is the saturation pressure, PG is the bulk gas pressure (derived from the adiabatic process equations described in section 2.3), R is the universal gas constant, WV is the fuel molecular weight and TB is the fuel liquid boiling point temperature. What we can see here is that the mass transfer potential will directly depend on the droplet temperature and must therefore be solved coupled with equation (12). Furthermore, the droplet diameter, which is used in the calculation of momentum timescale, will temporally evolve as the droplet is evaporating, so the momentum ODE must be coupled to the mass and heat transfer ODEs.

Simulations All ODE’s were solved using a 4th order Explicit Runge-Kutta Scheme along with a 1st order implicit Euler scheme for validation purposes. In order to validate the code, simulations were run with parameters used by Miller et.al [4] and compared and the results were identical. The results to be displayed have been acquired from the following simulations. Here follows a short clarification of table 2: 1. Two separate simulations were done for dodecane, one which compared the droplet timescales to the global engine timescales for an early and late injection, and the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

189

other which compared the droplet dynamics to the engine stroke restriction for 4 different engine sizes and for an early and late injection. 2. Only one simulation was done for iso-octane, that which compares the droplet timescales to the global engine timescale. Engine size was not investigated, as the effect of engine downsizing on droplet dynamics is made very apparent from a single fuel, there is no need to investigate two. Table 2. Simulation Parameters Fuel Type Dodecane (Fuel properties in Appendices)

Iso-octane (Fuel properties in Appendices)

Injection Regime Early injection@230CAD

Late injection@330 CAD Early injection@230CAD Late injection@330 CAD

Engine Size Small and Slow (s1), B=39mm,S=28mm,ω=2000rpm Small and Fast (s2), B=39mm,S=28mm,ω=8000rpm Medium (m1), B=84mm,S=89mm,ω=4000rpm Large and Slow (L1), B=S=136mm,ω=2000rpm Same as for Early Injection -No simulation run which depends on engine size-No simulation run which depends on engine size-

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

This section will be structured in the following manner: 1. Initially, we shall explore how engine speed and injection regime affect the global engine timescale, τe,g and how engine size affects the piston locus. 2. Secondly, we shall explore how the momentum, heat-up and mass transfer timescales compare to the global engine timescales for dodecane and iso-octane, for early and late injection. 3. Finally, we shall explore how droplet distances traveled (calculated from the momentum timescale) compare with the engine stroke restriction (engine size dependant) for dodecane, for early and late injection and for four different engines. This study will outline how fuel properties, initial droplet diameter and engine characteristics are related, for an uncharged liquid spray. In the subsequent section, the effect of electric charge on enhancing engine performance will be explored.

Global Engine Timescale and Piston Locus Results We shall firstly look at how τe,g is related to ω and θinj. This is shown in figure 2. It is clear that for late injection the timescale constraint is severe regardless of engine speed, where as for early injection, the timescale constraint is only relevant for higher engine speeds. Figure 2 gives us an excellent indication of where our droplet timescales have to stay within given a particular engine speed and injection regime.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

190

John Shrimpton and Agissilaos Kourmatzis

The variation of piston position according to time for two different cases is shown in figures 3 and 4.

Figure 2. τe,g vs ω for an early and late injection.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3 plots the piston position vs time, the piston’s location at t=0 is at the start of injection. So the y-axis of figure 3 is the distance the piston is from the injector (i.e. the distance includes the clearance height, xcl.) when it is at θ=230CAD from TDC. Similarly, figure 4 plots the variation but for the late injection strategy.

Figure 3. Piston position vs time after start of injection for four different engine sizes of table 2 for early injection. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

191

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 4. Piston position vs time after start of injection for four different engine sizes of table 2 for late injection.

What we can see here is that the x-axes of figures 3 and 4 are in the same range as the yaxis of figure 2, and the maximum values of taSOI (time after start of injection) for each curve of figures 3 and 4 represent τe,g for that particular engine speed and injection regime (defined in table 2). Each individual curve is calculated from the respective engine geometry parameters of table 2. In order to acquire realistic engine sizes the authors referred to Kashdan et.al [7] and Fuji IMVAC, Inc. The figures very conveniently show us where the piston will be after a given amount of time (calculated using equation (2)). Later on, we will plot droplet positions and compare them to figures 3 and 4 in order to determine whether or not we have violated the engine stroke restriction criterion suggesting that droplet impact on the piston will occur.

Droplet Timescale Results Here we will look at how in-cylinder conditions affect droplet momentum, heat-up and mass transfer timescales and we will also compare these timescales to the global engine timescales described earlier. In the examples we will look at here, the in-cylinder conditions in table 3 apply and the fuel types and injection regimes mentioned in table 2 are valid; All fuel properties/correlations are provided in the Appendices. Table 3. Cylinder Condition parameters CASE Injection@230 (early)

Variable Tg Pg ρg

Value 313K 1.26atm 1.4kg/m^3

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

192

John Shrimpton and Agissilaos Kourmatzis Table 3. Continued CASE Injection@330 (late)

Both Cases

Variable Tg Pg ρg r uD,o

Value 604K 12.6atm 7.25kg/m^3 10 70m/s

Note: Maximum pressure at TDC is (Patmrγ) where γ is the ratio of specific heats of air.

Droplet Momentum Timescale Results

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5 shows four cases; an early injection (CAD@230) and a late injection (CAD@330) for dodecane and the same for iso-octane. Here we are plotting droplet diameter vs the time it takes for a droplet to go from the initial velocity to 1/e*initial velocity. Notice how as we inject later, the timescale for a given droplet diameter decreases, this is owed to the fact that the drag coefficient of the droplet will significantly increase due to an increase in air density. Furthermore, note that isooctane loses its momentum slightly more rapidly than dodecane (fractions of a millisecond faster), due to the higher liquid density of dodecane and thus altered particle time constant. Furthermore, the timescales observed here are much lower than even the higher rpm area of the global engine timescale plot, from a design point of view this means that if we are interested in the majority of our droplet momentum decay we are free to use both the fuels displayed and at a range of diameters and engine speeds; however, if we were completely realistic and captured the complete droplet deceleration larger timescales would be acquired.

Figure 5.Droplet Diameter vs τmom (defined by equation 5) for iso-octane and dodecane for two different injection strategies. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

193

Figures 6-7 plot Engine speed vs. the ratio of droplet timescale to global engine timescale for a range of droplet diameters, for an early and late injection with isooctane (figure 6) and for an early and late injection with dodecane (figure 7).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 6a. Engine speed vs. τmom/τe,g for late injection as defined by table 2, for D=5-50μm (5 to 50 from left to right) for iso-octane.

Figure 6b. Engine speed vs. τmom/τe,g for early injection as defined by table 2, for D=5-50μm (5 to 50 from left to right) for iso-octane.

Looking at figure 6 first, it can be seen that for an early injection the droplet timescale is a smaller proportion of the global engine timescale when compared to the late injection (as Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

194

John Shrimpton and Agissilaos Kourmatzis

expected) and we observe that at larger diameters the droplet timescale gets closer to the global engine timescale. The fuels may be compared by choosing a curve from figure 6 and comparing to figure 7 and going to a particular engine speed, this tells us how close we are to the global engine timescale for a given fuel and speed. From a design point of view we can see the flexibility existent here through the range of diameters and speeds available.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 7a. Engine speed vs. τmom/τe,g for late injection as defined by table 2, for D=5-50μm (5 to 50 from left to right) for dodecane.

Figure 7b. Engine speed vs. τmom/τe,g for early injection as defined by table 2, for D=5-50μm (5 to 50 from left to right) for dodecane. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

195

Droplet Heat-Up Timescale Results

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Now we can have a look at the heat up timescales for early and late injection for the two fuels. All thermodynamic initial conditions are set to ambient as defined in section 2.5. We observe similarities between figure 8 and figure 5, again we see that at later injections our timescale decreases, this is expected; as we inject later our average in-cylinder temperature (bulk gas temperature Tg) has increased, thus accelerating the heat transfer from the gas phase to the liquid phase. Comparing the two fuels, we now see that the iso-octane timescale is lower than that of dodecane for any given diameter, an easy way to explain that is simply by noticing that iso-octane’s boiling point is significantly lower (more than 100 degrees difference) than that of dodecane therefore explaining the accelerated heat transfer rate of iso-octane. We can also produce non-dimensional figures again, only now for τheat/τe,g. We observe the same general trends as with the momentum cases and now the dodecane curves clearly show larger timescales than those of iso-octane (heat-up in this case), the difference is significant. By comparing the x-axes of the two fuels, for early injection cases the iso-octane timescales are approximately 3.5 times less than the dodecane for all cases and about 2 times less for late injection. Furthermore, for dodecane at large diameter droplets and high engine speeds we get close to the global engine timescale suggesting that we are more restricted by the droplet heat-up results than the droplet momentum decay results indicating that from an engine design point of view larger droplets cannot be employed at higher engine speeds.

Figure 8. Droplet Diameter vs τheat for iso-octane and dodecane for two different injection strategies.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

196

John Shrimpton and Agissilaos Kourmatzis

Figure 9. Engine speed vs. τheat/τe,g for early injection (bottom) and for late injection (top) as defined by table 2 for D=5-50μm (5 to 50 from left to right) for iso-octane.

Figure 10. Continued on next page. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

197

Figure 10. Engine speed vs. τheat/τe,g for early injection (bottom) and for late injection (top) as defined by table 2 for D=5-50μm (5 to 50 from left to right) for dodecane.

Droplet Mass-Transfer Timescale Results

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Here follows the timescale results for the evaporation of the droplet.

Figure 11.Droplet Diameter vs τmass for iso-octane and dodecane for two different injection strategies.

It is suffice to say that τmass is the governing timescale, we see values surpassing those of the global engine timescale values and now the restricting nature of the engine is made apparent. As with the heat-up case, the timescale for dodecane is larger than that of isooctane, and this is for the same reasons as the heat-up case. All of the other trends we observe

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

198

John Shrimpton and Agissilaos Kourmatzis

are the same as with the previous results only that now we see a much greater difference between the early and late injection strategies, i.e. injecting late increases the mass transfer timescale much more than it does the momentum or the heat-up timescales. This is particularly visible with dodecane. Again, let us directly compare these results to τe,g nondimensionally. Observing figures 12 and 13 we see the restriction when considering the evaporation of the droplet, if we want 90% of it to vaporize before ignition the above plots suggest very small diameters (difficult to achieve) or average sized droplets (30μm) but at low engine revolutions. Preheating the fuel is necessary if we want our droplet to vaporize otherwise we would have small amounts of un-vaporized fuel that we would have to burn (which in reality happens).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

a

b

Figure 12. Engine speed vs. τmasst/τe,g for early injection (b) and for late injection (a) for D=5-50μm (5 to 50 from left to right) for iso-octane.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

199

Figure 13. Engine speed vs. τmasst/τe,g for early injection (top) and for late injection (bottom) for D=550μm (5 to 50 from left to right) for dodecane

Droplet Position Compared to Piston Position In section 3.1 we briefly discussed some results regarding the position of the piston which was regarded as something of importance in our analysis. The results we have displayed up until now regarding the droplets do not take into account possible contact between the droplet and the piston crown and they do not take engine sizes into account. So far we have only looked at how we can compensate for different engine speeds and injection regimes for two different fuels. Now we will examine the effect of engine downsizing on the droplet in as a realistic way as possible. In table 2 we saw 4 different engines, each one with a different geometry. For each of the engines, a different angular velocity was provided in order to realistically capture what speeds those engines will be rotating at and we had a look graphically at where the piston will be after a given point in time, given those values.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

200

John Shrimpton and Agissilaos Kourmatzis

In order to calculate the droplet position after a given time we use our momentum calculation procedure, only now, we will not use the timescale measure, τmom. Here, we will look at the distance the droplet has traversed from its initial velocity of 70 m/s at t=0 to a final velocity which will depend on the engine geometry and speed, its turbulent fluctuating velocity. We will calculate this as 10% of the mean piston speed, i.e. equation (17).

utf = .1(2ω S )

(17)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

For example for engine s1, utf=2*(2000*2π/60)*(.028)*0.2=1.17m/s; We use this procedure instead of τmom as it will capture more of the droplet’s lifetime therefore giving us a more conservative measure of where the droplet will be after a given point in time and as it includes engine dependant parameters. We will do this analysis for dodecane only (for an early and late injection) as we are now interested in the effects of downsizing rather than fuel type. The results are particularly interesting, let us go through what they mean by observing figure 14 (s1). The curve represents where the piston of engine s1 will be from the cylinder head (y-axis) vs time (x-axis). The individual points represent different diameters (labeled; 550μm from the bottom-up in increments of 5μm) whose y-axis represents the distance they traveled until there momentum decayed (distance traveled until utf was reached) vs the amount of time it took for that to happen.

Figure 14. Continued on next page.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

201

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 14. Piston distance (s1 and s2) and droplet decay distance vs time after start of injection for early injection for dodecane.

The meaning of the plots is easiest to explain using examples, we observe figure 14 (s1) once more. Here we can see that the 35 micron droplet is below the engine curve, this means that it is closer to the injector than the piston is when its momentum has decayed which suggests that collision will not occur. If we wished to use the same engine size but with a speed of 8000rpm’s (figure 14 s2), we will see that the 35 micron droplet is now further from the injector than the piston head is at 1.2ms; this suggests that collision has already occurred (the limiting diameter for engine s2 at the early injection regime is at about 32microns, where the droplet data line intersects the engine curve). For larger engines, one can see that we are not very restricted at all if we want to inject early; this is of course using the assumption that the momentum decay occurs at the engine turbulent fluctuating velocity. We can now compare figures 14 and 15 to figures 16 and 17, the latter plots are with the same engine sizes and speeds, only now we inject later and are therefore much more restricted. Comparing any of the plots of figure 16 and 17 to those of 14 and 15, it can be seen that we are now severely limited as to what droplet size we can use. Figure 16 (s1 and s2) tells us that anything above 15microns will collide with the piston head, and for the medium engine (figure 17 m1), anything above 35 microns is unacceptable in contrast to the early injection regime for that engine where the restriction would have been at much larger diameters.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

202

John Shrimpton and Agissilaos Kourmatzis

Figure 15.Piston distance (m1 and L1) and droplet decay distance vs time after start of injection for early injection for dodecane

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

203

Figure 16. Piston distance ( s1 and s2) and droplet decay distance vs time after start of injection for late injection for dodecane.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

204

John Shrimpton and Agissilaos Kourmatzis

Figure 17. Piston distance ( m1 and L1) and droplet decay distance vs time after start of injection for late injection for dodecane.

At this point another important observation must be made that concerns figures 14-17. The very right end of all of the engine curves represents the point in time when that piston has reached TDC i.e. ignition (as we discussed earlier). So, for example, in figure 16 (s2) we can Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

205

see that from 35microns and onwards all of the droplets are to the right of the engine curve, this means that they decay at a time which has surpassed the global engine timescale τe,g. In some situations (none presented here) it may be that a droplet is below the engine curve (suggesting no collision), but to the right of it. Even though that would suggest that the droplet has not met the piston crown, it does suggest that the droplet has not decayed before τe,g which is unacceptable.

Summary This part of the chapter has covered how an engine timescale can restrict the choice of fuel, injection regime and engine speed via comparing the global engine timescale to the three droplet timescales of interest (τmom, τmass, and τheat). Throughout that study we saw that the main restricting droplet timescale was the evaporation scale which in many cases surpasses τe,g. This restriction can be relaxed via the use of electrostatic charge as we will see in the following section. In addition to droplet timescales we also explored how the engine stroke length-scale affects the choice of engine geometry, injection regime and fuel type and how we must also consider this in addition to the restricting value of τe,g.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Electrostatic Charging We will now explore the effects of electrostatic charging on the droplet mass timescale which will show us how charge can prove to be advantageous in engine design by reducing the droplet mass timescale. The effect that we will consider within this chapter shall be the effect of electrostatic charging on the evaporation of a droplet; we shall ignore any effects on the droplet momentum which are negligible in our analyses as we are only considering the movement of the droplet along the axial direction (electrostatic charging mainly has an effect on the radial dynamics of the droplet [5]). We also ignore effects on the heat up rate as we are interested in investigating the effect of charge on droplet evaporation which is the most restricting timescale. The temperatures that we shall use in these simulations are those at the early and late injection conditions, defined in table 3.

Background Theory As explained in the introduction of this chapter, when we inject a current into a fuel droplet the surface tension is disturbed thus promoting the improved atomization of the droplet. There is a limit to how much charge a droplet can hold before it breaks up and this is known as the Rayleigh limit [3,5].

Qray = π (ε oσ )1/ 2 (2 D)3/ 2

(18)

where εo is the permittivity of the surrounding gas, σ is the surface tension of the fuel and D is the droplet diameter. The charge density of the droplet is defined as the charge it holds per Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

206

John Shrimpton and Agissilaos Kourmatzis

unit volume. So we can change the Rayleigh charge limit to a Rayleigh charge density limit by dividing Qray by the droplet volume.

Simulation Parameters Table 4 shows the parameters used and the various studies carried out within these simulations: Table 4. Charge Simulation parameters Fuel Type

Injection Regime

Charge Condition 1

Charge Condition 2

Dodecane (Fuel properties in Appendices)

Early injection@230CAD

Qinitial=0.7Qray Break-up@ Q=0.8Qray At each break up 80% of mass remains and 30% of charge remains. [8]

Qinitial=0.4Qray Break-up@ Q=0.8Qray At each break up 80% of mass remains and 30% of charge remains. [8]

Late injection@330 CAD

Same as for early

Same as for early

Early injection@230CAD

Same as above

Same as above

Late injection@330 CAD

Same as above

Same as above

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Iso-octane (Fuel properties in Appendices)

The reader will notice that the break up conditions used are at 80% of the Rayleigh charge, this is because it has been shown that in practice, break-up occurs before the limiting Rayleigh value [9]. In order to illustrate the effect of electrostatic charging we shall present a result for a single 20 micron droplet of dodecane; at early and late injection conditions (with charge) plotted alongside a curve with identical injection conditions but without charge. Those plots will show the ratio of final diameter to initial diameter all squared, vs time. Following this we shall present curves of τmass (with charge) vs Diameter for the cases of table 4 which will be compared with figure 11, τmass (no charge) vs. Diameter.

Results As we can see both from figures 16 and 17 charge has a significant effect on the evaporation of the droplet. The electric charge decreases the timescale taken for evaporation and can thus be used to enhance engine performance. Each discontinuity in the charged curve represents the droplet reaching the Rayleigh limit and thus losing 20% of its mass and 70% of its charge.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

207

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 16. (D2/Do2) vs time for charge condition 1 of table 4 for early injection.

Figure 17. (D2/Do2) vs time for charge condition 1 of table 4 for late injection.

Figures 18 and 19 show the effect of charge for the two different fuels and injection regimes looked at throughout this chapter. The differences present are more easily noticed by observing the very slow evaporation rate of dodecane for the early injection case. Directly comparing figure 18 with figure 11 we can see that for all of the diameters, we have lowered the mass timescale thus enhancing the performance of our engine.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

208

John Shrimpton and Agissilaos Kourmatzis

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 18. Droplet Diameter vs τmass for iso-octane and dodecane for two different injection strategies with charge condition 1 of table 4.

Figure 19. Droplet Diameter vs τmass for iso-octane and dodecane for two different injection strategies with charge condition 2 of table 4.

Of course the enhancement that we will get will also depend on the level of charge that we choose to inject into our fuel; we can see this by observing figure 19 where we inject a lower initial charge value into the fuel. We see a 10ms difference between the two charge Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Downsizing Direct Injection Spark Ignition Engines: A Timescale Analysis

209

conditions and a difference of approximately 20ms between the low charge regime and the zero charge regime of figure 11 for the early injection case for dodecane. The reader should note that these timescale improvements are extremely conservative. In reality, liquid droplets shatter due to aerodynamics forces [10] and since electric charge can be viewed as a mechanism to suppress surface tension forces, we can expect enhanced rates of secondary atomization. This should, as suggested elsewhere [5] significantly improve the overall fuel evaporation rate.

Conclusion This chapter has explored how different injection regimes, engine speeds, fuel types, engine geometry and charge conditions affect the momentum, heating up and evaporation timescales of a single fuel droplet within an engine cylinder. The aim was to show to the reader what parameters need to be considered in the design of an engine when concerned with fuel efficiency and generally what the restricting engine parameters are in such a design. The technology of electrostatic charging was briefly discussed and introduced as an option that can be used to enhance engine performance.

Appendix 1. Fuel Properties and Correlations Fuel Correlations acquired from Miller et al. [4] and from work done at Imperial College by J.S. Shrimpton. Various other properties were obtained from the American National Institute of Standards and Technology.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Air μG= 6.109e-6 + 4.604e-8(T) - 1.051e-11(T2) kgm-1s-1 PrG=.815 - 4.958e-4(T) + 4.514e-7(T2) λG= 3.227e-3 + 8.3894e-5(T) - 1.9858e-8(T2) Wv=28.97 kgkmol-1 R=8314.5 kJkmol-1K-1 Dodecane ρD=750; kgm-3 CL=2207; Jkg-1K-1 Wv=170.3348; gmol-1

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

210

John Shrimpton and Agissilaos Kourmatzis TB=489; K Lv=112100(7.08(1-T/658).354+10.95*.576*(1-T/658).456); Jkg-1 Iso-octane ρD =690; kgm-3 CL =2098; Jkg-1KWv =114.22; gmol-1 TB =372.4; K Lv =62012(7.08(1-T/543)^.354+10.95*.304(1-T/543)^.456); Jkg-1 Ranz-Marshall Correlations Nu= 2 + 0.552(Re^1/2)(PrG^1/3) Sh=2 + 0.552(Re^1/2)(ScG^1/3)

References

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[1] [2]

John L. Lumley. Engines an Introduction, Cambridge University Press 1999. H. Lienemann, J.S. Shrimpton, 2002. Timescale Considerations for Internal Combustion Engine Sprays. ILASS-Europe, Zaragoza 9-11 September. [3] J.S. Shrimpton. Pulsed Charged Sprays: application to DISI engines during early injection. Internation Journal for Numerical Methods in Engineering 2003. 58:513-536. [4] R.S. Miller, K. Harstad, J. Bellan. Evaluation of equilibrium and non-equilibrium evaporation models for many-droplet gas-liquid flow simulations. International Journal of Multiphase Flow 1998. 1025-1055. [5] J. S. Shrimpton, Y. Laoonual. Dynamics of electrically charged transient evaporating sprays. International Journal for Numerical Methods in Engineering 2006. Published online in Wiley InterScience. [6] Putnam A. Integratable form of the drop drag coefficient. Journal of the American Rocket Society 1961. 31:1467-1468. [7] J.T. Kashdan, J.S. Shrimpton. Dynamic Structure of Direct-Injection Gasoline Engine Sprays: Air Flow and Density effects. Atomization and Sprays 2002. vol.12 pp. 539-557. [8] J.S. Shrimpton. Modeling Dielectric Charged Drop Break up Using an Energy Conservation Method. School of Engineering Sciences, Southampton University, UK. Manuscript in final form and accepted 10 June 2008. [9] A Gomez, K Tang. Charge and Fission of Droplets in Electrostatic Sprays. Physics of Fluids 6, pp 404-414, 1994. [10] R.D Reitz, R Diwacker. Effect of Drop Breakup on Fuel Sprays. SAE 860469, 1986.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution... Editors: S. Demidov and J. Bonnet, pp. 211-249

ISBN 978-1-60741-145-1 c 2009 Nova Science Publishers, Inc.

Chapter 10

R ECENT P ROGRESS IN H YDROGEN -F UELED I NTERNAL C OMBUSTION E NGINES Sebastian Verhelst∗ and Roger Sierens Department of Flow, Heat and Combustion Mechanics Ghent University, Belgium

Abstract

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Greenhouse gas emission by the transport sector is a hot topic these days. There is a strong drive towards legislation limiting the fleet average CO2 emissions. The use of hydrogen as an energy carrier is one option with the potential of lowering CO2 emissions investigated by the vehicle manufacturers. Governments from regions that are pollutant hot spots are also looking into hydrogen as a potential solution. Hydrogen is mostly associated with fuel cells. However, affordable fuel cell vehicles seem to be a long way off, both for technological as for economic reasons. An interesting alternative is using hydrogen in internal combustion engines (ICEs). Next to being less expensive, hydrogen-fueled ICEs offer a number of other benefits of which the most practical one is the ability to run in bi-fuel or flex-fuel operation. Research on hydrogen-fueled ICEs has been reported from the 1930’s onwards. In the past decade, the research has shifted from general laboratory-type ”proof of concept” work to efforts focused on achieving practical vehicles through thorough optimization. This includes modern concepts such as direct injection, supercharging, exhaust gas recirculation, variable valve timing etc. This work reports the pros and cons of hydrogen-fueled ICEs, the properties of hydrogen relevant to its application as a fuel in engines and the resulting engine hardware and software features. The current state of the art is discussed, for achieving maximum power output with minimal emissions and minimal fuel consumption. Results from numerous engine test bench experiments in the authors’ department are shown, and results obtained by others are reviewed. ∗

E-mail address: [email protected]

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

212

1.

Sebastian Verhelst and Roger Sierens

Introduction

1.1.

Contents

In the following, first some thoughts on hydrogen as an energy carrier are given, and whether hydrogen should be used in internal combustion engines (ICEs) or fuel cells (FCs). The remainder of the text lists the properties of hydrogen relevant to ICEs, how hydrogen compares with conventional fuels, what past research has accomplished and discusses the current state of hydrogen ICE research. Part of this has been published previously [1, 2]. However, the number of recent new publications as well as new work at the authors’ department warrants an update. The use of hydrogen in dual-fuel applications (hydrogen-gasoline fueling, hydrogendiesel, hydrogen-natural gas etc.), or as an additive, is not discussed here.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1.2.

Why Hydrogen?

The incentives for using hydrogen as an energy carrier are well known, with the most important ones being the clean emissions in its use and the variety as well as long term viability of methods to produce it. One could argue that the potential of clean emissions has become less of a drive, as modern gasoline engines equipped with three-way catalysts and sophisticated engine management systems have been demonstrated to produce extremely low emission levels, with some ultra-clean vehicles already on the market1 . Nevertheless, emissions remain a concern as the number of vehicles as well as the vehicle miles driven yearly keep rising, offsetting the advances in emission reduction technology. A second interest in hydrogen relates to the source of fossil fuels. As fossil fuels are hydrocarbons, their combustion will produce carbon dioxide, which has been designated as the most important contributor to radiative forcing in the Earth’s atmosphere, resulting in a global warming or the so-called greenhouse effect. As all parties that signed the Kyoto Protocol [3] have committed themselves to reducing their anthropogenic carbon dioxide emissions to below the 1990 levels by the period 2008-2012, the energy conversion efficiency will have to increase, or alternatives to carbon containing energy sources must be promoted. The European Commission is setting up a legislative framework to bring down the CO2 emissions from road transport (accounting for roughly 25% of total CO2 emissions) to 130 g/km by 2012 (fleet average). Further decreases will have to come from taking carbon out of the fuel. Hydrogen is also viewed as a means to enhance energy security, as the global fossil fuels reserves are geographically concentrated, with the largest concentration in politically unstable regions. Furthermore, these reserves are finite and there is increasing evidence that the peak in oil production has already happened or will happen in the very near future [4]. Finally, fossil fuels are a valuable raw material for the polymer industry. Hence, although different regions have different policies, there is an unmistakable trend away from fossil fuels. Biofuels such as bio-ethanol, biodiesel etc. are one option with the potential to reduce well to wheel CO2 emissions. However, if we take Europe as an 1

For examples see for instance the California Air Resources Board list of California certified vehicles, http://www.arb.ca.gov Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

213

example, it is clear that the potential for biomass derived fuels, even for second generation biofuels, is limited to about 15-20% of the transport need due to limited land availability for energy crops (avoiding competition with food crops) [5]. A recent study compared the yield of final fuel per hectare of land for different biomass derived fuels, to hydrogen from photovoltaics (PV) and hydrogen from wind power [6]. The results show that the energy yield of land area is much higher when it is used to capture wind or solar energy. This is easily explained when comparing the efficiency of PV panels (about 15% for commercially available PV) or wind turbines to the efficiency of the photosynthesis process for capturing solar energy (in theory maximum 9 or 10%, in practice ranging from 0.2% for rapeseed oil to about 2%). Hydrogen is then an obvious choice for storing the wind or solar energy so it can be used to power transport. While hydrogen has compelling advantages, it has equally dissuading disadvantages. The biggest challenge is its low density. As a consequence, the energy density is low even when compressed to 700 bar or liquefied, which both incur substantial energy losses. Thus, distribution, storage and on-board vehicle storage are heavily compromised. Even when advocating decentralized hydrogen production and taking into account the typically higher end-user efficiencies when using hydrogen (see below), the vehicle driving range will suffer and the fuel storage system will be (a lot) more expensive compared to the liquid hydrocarbon fuels we are currently used to. For some applications, such as captive fleets, the advantages of hydrogen can outweigh the disadvantages, e.g. where true (tailpipe) zero emission is required. This justifies research into hydrogen-fueled powertrains. The main powertrain for transport, the internal combustion engine, is one of the options and the subject of this work.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1.3.

Why Hydrogen-Fueled Internal Combustion Engines?

As mentioned above, there are a number of hydrogen utilization technologies. Hydrogen fueled internal combustion engines and hydrogen fueled fuel cells are the two that qualify for transportation purposes. The present work focuses on hydrogen fueled ICEs for the following reasons. The internal combustion engine has benefited from a continuous development during more than a century and is still showing potential for further optimization. Fuel cell technology on the other hand is still in its infancy. This also reflects in the price, with a prohibitive cost for fuel cells. Advocates of the fuel cell claim the price will drop orders of magnitude (the current price scale difference between FC and ICE) through further development and the economics of scale but one has to keep in mind similar claims for electric vehicle battery prices which have failed to come true. Naturally, the conversion of an ICE to hydrogen increases its cost but this cost is very limited2 . Using ICEs allows bi-fuel operation (e.g. the engine can run on gasoline as well as on hydrogen), alleviating fuel station density and autonomy requirements. This could facilitate the start-up of a hydrogen infrastructure, where the experience gained with transport, fueling and storage directly translates to fuel cell vehicles. 2

Not counting the cost of the hydrogen storage and safety devices, as these are needed regardless of the propulsion unit, the only additional cost of hydrogen fueled ICEs is the cost of H2 injectors, a modified engine control unit and some changes to the ignition and crankcase ventilation system Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

214

Sebastian Verhelst and Roger Sierens

Fuel cells are currently still handicapped by cold-start problems (freezing of the fuel cell stack) and the necessity of very pure hydrogen to avoid poisoning of the FC [7, 8], which also involves a cost penalty. The hydrogen fueled ICE does not suffer from these problems. The most frequently hailed advantage of fuel cells is its high theoretical efficiency. However, not only do practical fuel cells not (yet) reach these high efficiencies, the fuel cell stack (of which the efficiency is mostly cited) is also part of a fuel cell system and the overall efficiency is thus lower. Furthermore, the efficiency decreases as the load increases (the cell ohmic losses increase with the square of the cell current). This is not an important disadvantage for light-duty applications as these are in part load most of the time, but could become important for heavy duty. The large difference between the theoretical efficiency of the fuel cell stack and the effective efficiency of an ICE thus mostly exists on paper and is much smaller in practice. Furthermore, hydrogen fueled ICEs also have the potential for an increased engine efficiency (see later), with a demonstrated indicated efficiency of 52% for a hydrogen fueled spark-ignition engine [9] and a power generation efficiency of 49% for a hydrogen fueled compression-ignition engine [10]. In summary: the hydrogen fueled ICE and FC both have their advantages and both merit research to show their full potential. The hydrogen fueled ICE can function as a transition technology to fuel cells or might take up its own share of the market next to fuel cells (and other technologies).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1.4.

Hydrogen Properties Relevant to ICEs

Hydrogen has significantly different properties compared to the more traditional fuels. Table 1 summarizes the most important physical and combustion-related properties as a reference for the discussion in the following. All temperature and pressure dependent data is given at normal temperature and pressure (300 K and 1 atm), combustion-related properties cited are for stoichiometric combustion. Most data is either rounded off or is given approximately, for ease of comparison. The exact data can be found in the source references [1, 11, 12, 13]. The auto-ignition temperature of hydrogen can be seen to exceed the values for methane and gasoline. This makes hydrogen particularly suited for spark ignition operation and unsuited for compression ignition. The remainder of this work therefore deals exclusively with hydrogen spark-ignition engines unless otherwise stated. Note that some works state a higher adiabatic flame temperature for gasoline compared to hydrogen [13], which should be the other way round (as in Table 1). This is important when comparing N Ox emissions. No octane rating is given in Table 1, for reasons discussed below. It is noteworthy that for the determination of the ‘methane number’ of a gaseous fuel, hydrogen is taken as having a methane number of zero [14], giving the impression that is very prone to knock (more about this in Section 2.1.1.).

2.

Literature Review

The literature on hydrogen fueled internal combustion engines is surprisingly extensive and papers have been published continuously from the 1930’s up to the present day, alDemidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

215

Table 1. Hydrogen properties compared with methane and gasoline properties Property Molecular weight (g/mol) Density (kg/m3 ) Mass diffusivity in air (cm2 /s) Kinematic viscosity (mm2 /s) Stoichiometric volume fraction (in air) Minimum ignition energy (mJ) Auto-ignition temperature (K) Adiabatic flame temperature (K) Normalized flame emissivity (200 K, 1 atm) Flammability limits in air (vol%) Laminar burning velocity (m/s) Quenching distance (mm) Lower heating value (M J/kg) Higher heating value (M J/kg)

Hydrogen 2.016 0.08 0.61 110 29.5 0.02 858 2390 1 4-75 2.1 0.64 120 142

Methane 16.043 0.65 0.16 17.2 9.5 0.28 813 2225 1.7 5-15 0.36 2.03 50 55

Gasoline ∼107 ∼750 0.05 1.18 1.65 0.25 ∼500-750 ∼2275 1.7 1.0-7.6 ∼0.40 ∼2.0 45 48

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

though most of them are concentrated around a few points in time (e.g. during and in the years following the oil crises). A summary of the most important findings of these works is given in the following sections and is an update from the overview given in [1]. This will serve as the background for elucidating the more recent research reported in the later sections.

2.1.

Proof of Concept Reports

This section summarizes the findings of the initial ‘conversion project’ works. Initial experimentation on H2 ICEs was to a large extent ‘proof of concept’ work: converting a base, mostly gasoline engine to hydrogen operation with relatively minor changes. Most of these works focused on the suppression of abnormal combustion, and to lesser extent on the suppression of N Ox emissions. In the following, abnormal combustion and mixture formation in H2 ICEs are treated, after which some design ‘guidelines’ are given for dedicated hydrogen engines.

2.1.1.

Abnormal Combustion

The suppression of abnormal combustion in hydrogen engines has proven to be quite a challenge and measures taken to avoid abnormal combustion have important implications for engine design, mixture formation and load control. For spark-ignition engines, three regimes of abnormal combustion exist: auto-ignition of the end gas region (for gasoline engines mostly referred to as knock, as this is how it manifests itself), surface ignition (uncontrolled ignition induced by a hot spot, referred to as pre-ignition if it occurs premature

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

216

Sebastian Verhelst and Roger Sierens

to the spark ignition [15]) and backfire3 (premature ignition during the intake stroke, which could be seen as an early form of pre-ignition). The effects auto-ignition and glow-ignition can have are well known: in the best case increased noise and vibration, in the worst case major engine damage. The effects of backfire are a loud ‘bang’ in the best case, an engine stop as the fuel is consumed before it can enter the cylinders and deliver work, or a destruction of the intake manifold in the worst case. Backfire has been a particularly tenacious obstacle to the development of hydrogen engines. Most, if not all, of the early literature mentions causes of backfire and countermeasures as it so frequently occurs in hydrogen engines with external mixture formation (backfire can only occur when a combustible charge is present in the intake port). The causes cited for backfire are: • Hot spots in the combustion chamber: deposits and particulates [16, 17], the spark plug [18, 19], residual gas [12, 19, 20, 21], exhaust valves [20, 22, 23, 24], etc. These hot spots are cited to easily cause a backfire ‘because of the low ignition energy of hydrogen’, which is an order of magnitude smaller than for typical hydrocarbons, and the wide flammability limits (see Table 1). Deposits and particulates originate from the (partial) combustion of lubricating oil and/or rust formation during an extended standstill (corrosion can be a problem, due to the large amounts of water and N Ox formed during stoichiometric combustion [25]), or from the hydrocarbon fuel in case of a bi-fuel engine.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• Residual energy in the ignition circuit: due to the lower ion concentration of a hydrogen/air flame compared to a hydrocarbon/air flame, it is possible that the ignition energy is not completely deposited in the flame and remains in the ignition circuit until the cylinder conditions are such that a second, unwanted, ignition can occur, namely during the expansion or the intake stroke, when the pressure is low [19, 26]. • Induction in the ignition cable: with multi-cylinder engines, the (controlled) ignition in one cylinder can cause an induced ignition in another cylinder when the individual ignition cables are placed close to each other [17]. • Combustion in the piston top land persisting up to inlet valve opening time and igniting the fresh charge [19, 27, 28, 29]. This is caused by the smaller quenching gap of hydrogen mixtures compared to typical hydrocarbons, which enables a hydrogen flame to propagate into the top land. • Pre-ignition: pre-ignition is often encountered in hydrogen engines because of the low ignition energy and wide flammability limits of hydrogen. As a premature ignition causes the mixture to burn mostly during the compression stroke, the temperature in the combustion chamber rises, which causes the hot spot that led to the pre-ignition to increase in temperature, resulting in another, earlier, pre-ignition in the next cycle. This advancement of the pre-ignition continues until it occurs during the intake stroke and causes backfire [9, 23, 28, 30, 31]. The mechanism is termed a runaway pre-ignition and can also result from a knocking cycle, increasing the chamber temperature and creating a hot spot [17]. 3

Also referred to as backflash, flashback and induction ignition

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

217

In summary, there is still much to learn about the precise causes for backfire and how to avoid them. In engines, it is often hard to single out one parameter, e.g. decreasing the valve overlap has been shown by Huynh et al. [32] to extend the backfire-free operation range, but as the authors point out this simultaneously decreases the backflow of (hot) exhaust gas to the intake, as well as the power output (and combustion chamber temperature) due to a lower volumetric efficiency, so it is difficult to pinpoint the exact cause for the suppression of backfire. In the present authors’ opinion, the low ignition energy of hydrogen is often too easily pointed at as the main cause of backfire. The ignition energy is defined by the minimal spark energy needed to ignite the mixture [15], whereas the ignition by thermal masses such as the valves and residual gases is perhaps more related to the auto-ignition temperature of the mixture, the temperature at which the mixture will spontaneously ignite. As the auto-ignition temperature of hydrogen is quite high (higher than for methane and gasoline, see Table 1), it seems unlikely that these thermal masses would induce backfire as they normally do not reach the auto-ignition temperature if the engine is properly set up. As hydrogen compression ignition engines have been demonstrated to require very high compression ratios in order to ensure self-ignition [10], it is improbable that e.g. residual gases could initiate ignition thermally (again: assuming normal working conditions, with optimal spark timing etc.). Furthermore, this cannot explain the occurrence of backfire at lean conditions (low temperatures). Also, deposits and particulates are frequently cited although (assuming an engine in good condition) the concentration of these are extremely low for hydrogen engines. The ‘inert dust in air’ has even been cited [12], probably forgetting the function of the engine air filter. Experiments have been conducted where all hot spots were eliminated (careful cleaning of the engine, enhanced oil control or even non-lubricated operation, scavenging of the residual gases, cold spark plugs, cooled exhaust valves, ...), as well as any uncontrolled spark-induced ignition, and backfire still occurred [19, 28]. This suggests that the small quenching distance of hydrogen (together with the wide flammability limits), allowing combustion in the piston top land4 , is a parameter that has been overlooked by many workers. Hydrogen engines have been demonstrated, running on stoichiometric mixtures without any occurrence of backfire, by careful selection of piston rings and crevice volumes, without any need for timed injection (see later) or cooled exhaust valves [27]. Workers that have paid attention to increased cooling, enhanced ‘oil control’ by mounting different piston rings, increased scavenging etc., attribute the resulting wider backfire-free operation region to a reduction of hot spots but have simultaneously (sometimes possibly without realizing it) taken measures to suppress crevice combustion. There is some ambiguity in the literature on the relation backfire–compression ratio. Some authors mention a decrease of the compression ratio to increase the resistance to backfire [24, 33] by lowering the combustion chamber temperature; others state that an increase in compression ratio is advised, resulting in an increased combustion chamber area to volume ratio, thus increasing the heat transfer and cooling residual gases [12, 23, 34]. An increased compression ratio also lowers the amount of residuals. Both are valid mechanisms and indicate the existence of an optimum compression ratio: increasing it results in higher 4

When the piston rings and crevice geometries are those as used for the traditional fuels

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

218

Sebastian Verhelst and Roger Sierens

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

power due to increased efficiency up to a certain point, where the mixture has to be set leaner to avoid pre-ignition and a power decrease is noted [20, 35]. To conclude the discussion of the backfire phenomenon: all causes itemized above can indeed result in backfire and the design of a hydrogen engine should try to avoid them, as engine conditions different from normal operation are always a possibility. Even though backfire that is stated to be the result of hot spots might have other causes, hot spots still have to be avoided as they can lead to surface ignition which increases the engine’s thermal loading and can have detrimental effects even without leading to backfire (e.g. in an internal mixture formation engine, see later). The knocking behavior of hydrogen engines, in the sense of auto-ignition of the endgas, has been misunderstood more than the backfire phenomenon. With backfire, some causes have been mixed up or their effects overestimated, but concerning ‘knock’ there are completely contradictory claims in the literature [36]. To begin with, most papers fail to point out that the ‘knock resistance’ is a property of the fuel/air mixture, stating octane numbers without the corresponding equivalence ratio. Some claim the octane number to be very low [37, 38], others claim it to be very high [9, 12, 23]. One paper was even found that stated both ‘hydrogen has a high effective octane number’ and ‘the equivalent octane number of hydrogen is rather low’ [13]! Only very few papers mention octane numbers as a function of the mixture richness [34, 39]. Experiments have been reported that show hydrogen to act as an anti-knock agent when added to unleaded iso-octane [12]. There is some evidence that the causes of hydrogen engine knock could be different from gasoline knock, being caused by excessive flame speeds rather than an end-gas reaction [12, 23]. Thus, reducing the rate of pressure rise might be more effective to control knock than limiting the combustion period (e.g. using a quiescent combustion chamber, see later). Reviewing the experimental literature on hydrogen SI engines, surface ignition seems to be the limiting factor concerning compression ratios, spark timings and mixture equivalence ratios, rather than auto-ignition. Measurements with a compression ratio of 11:1 and a supercharging pressure of 0.85 bar (gauge) on stoichiometric mixtures have been reported [20], as well as measurements on lean mixtures using compression ratios of 14:1 and higher [9, 40], all without any appearance of auto-ignition. It thus seems safe to say that hydrogen has a higher effective octane number than regular gasoline, although it would be interesting to have quantitative data (studies by Ford and Mobil Oil showed an effective octane rating of about 140 for a hydrogen-air mixture with an air-to-fuel equivalence ratio λ of 2.5 [2]). It is noteworthy that the experimental and theoretical work of Karim and co-workers [38, 41] reports very wide ‘knocking regions’, where stoichiometric mixtures are claimed to knock even at compression ratios as low as 6:1. As these results disagree with every other experiment reported in the literature, they seem highly unlikely and are probably affected by causes unknown to the authors. 2.1.2.

Mixture Formation

A range of mixture formation methods has been tested for hydrogen engines, mostly in the pursuit of backfire-free operation: Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

219

• external mixture formation with a gas carburetor [19, 34, 42, 43], sometimes with water injection [24, 37], sometimes with additional exhaust gas recirculation (EGR) [44] • external mixture formation with ‘parallel induction’, that is: some means of delaying the introduction of hydrogen, e.g. a fuel line closed by a separate valve on top of the intake valve that only opens when the intake valve has lifted enough [45] • external mixture formation with timed manifold or port fuel injection (PFI) [9, 17, 20, 27, 30, 35, 46, 47, 48], sometimes also with some means of ‘parallel induction’ [39], sometimes with water injection [49] • internal mixture formation through direct injection (DI) [50, 51, 52, 53], sometimes with water injection [54, 55]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• external and internal mixture formation combined (PFI+DI) [56] In these works, the injection of water, delayed introduction of hydrogen and direct injection are all primarily intended to delay or prevent backfire either by additional cooling (also affecting N Ox formation, which is discussed further below) or by avoiding a combustible mixture during the intake phase. During the last decade, only timed port injection and direct injection (during the compression stroke or later), or a combination, have been used, as the other methods are less flexible and controllable. External mixture formation by means of port fuel injection has been demonstrated to result in higher engine efficiencies, extended lean operation, lower cyclic variation and lower N Ox production (for lean mixtures) compared to direct injection [40, 57]. This is the consequence of the higher mixture homogeneity due to longer mixing times for PFI as well as decreased mixing for DI as the intake generated turbulence contributes less to the mixing. Additionally, the cost and complexity are significantly lower for PFI than for DI [22] and retrofitting an existing engine is possible. On the other hand, the power output of an external mixture formation hydrogen engine is limited because of the decrease in volumetric efficiency: due to the low density of hydrogen and small air requirement for stoichiometric mixtures, the cylinder volume taken up by the hydrogen in a stoichiometric mixture amounts to 29.5%, see Table 1. This results in a volumetric energy content decrease of some 18% for hydrogen compared to gasoline (however, the volumetric efficiency of the engine can also be affected by the higher local acoustic velocity due to the presence of hydrogen [58, 59]). If direct injection is used to introduce the hydrogen after the intake valve has closed, the maximum power output can be 17% higher compared to gasoline. An important advantage of DI over PFI is the impossibility of backfire. This too increases the maximum power output of DI compared to PFI as richer mixtures can be used without fear of backfire. Pre-ignition can still occur though, unless very late injection is used. External mixture formation provides a greater degree of freedom concerning storage methods: direct injection during the compression stroke needs high pressure hydrogen and thus effectively requires liquid hydrogen storage5 . 5

Metal hydrides can only provide low pressure hydrogen, compressed hydrogen could be used but this limits the effective tank contents as the tank can only be emptied down to the fuel injection pressure. Compressing gaseous hydrogen on board would mean an extra compressor and a substantial energy demand. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

220

Sebastian Verhelst and Roger Sierens

Thus, both external and internal mixture formation have their advantages and disadvantages. DI is better for full load performance (maximum power output), PFI is better at part load (maximum engine efficiency). Engine designs have been proposed using both mixture formation techniques [56, 57, 60, 61]. Contemporary reviews of mixture formation techniques for hydrogen engines can be found in refs. [23, 60]. 2.1.3.

Hydrogen Engine Design Features

The initial findings discussed in the above, and other relevant literature, can be summarized in some H2 ICE ‘guidelines’, given in the following:

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• Spark plugs: use cold rated spark plugs to avoid spark plug electrode temperatures exceeding the auto-ignition limit and causing backfire [18, 26]. Cold rated spark plugs can be used since there are hardly any spark plug deposits to burn off. Do not use spark plugs with platinum electrodes as this can be a catalyst to hydrogen oxidation [12, 22] (platinum has been used in the exhaust to oxidize unburned hydrogen [45]). • Ignition system: avoid uncontrolled ignition due to residual ignition energy by properly grounding the ignition system or changing the ignition cable’s electrical resistance [24, 26]; avoid induction ignition in an adjacent ignition cable, for instance by using a coil-on-plug system; provide a high voltage output ignition system as the ignition of hydrogen mixtures asks for an increased secondary ignition voltage [24, 26, 62, 63] (probably because of the lower ion concentration of a hydrogen flame, see above), coil-on-plug systems also satisfy this condition. Alternatively, the spark plug gap can be decreased to lower the ignition voltage, this is no problem for hydrogen engines as there will be almost no deposit formation. Spark plug gaps as small as 0.25 mm have been used [19] (although the gap was subsequently increased to 0.5 mm because of cold start difficulties due to water condensation at the spark plug tip). • Injection system: provide a timed injection, either using port injection and programming the injection timing such that an air cooling period is created in the initial phase of the intake stroke and the end of injection is such that all hydrogen is inducted, leaving no hydrogen in the manifold when the intake valve closes; or using direct injection during the compression stroke. High flow rate injectors are needed in both cases, multiple injectors per cylinder can alleviate this requirement for PFI engines. The timing described here might not be necessary as work has been reported where no relation between injection timing and backfire or surface ignition limited equivalence ratio was found [9]. Timed injection also decreases the amount of unburned fuel in the intake manifold at any given time, limiting the severity of a backfire should it occur. • Hot spots: avoid hot spots in the combustion chamber that could initiate surface ignition or backfire, use cooled exhaust valves; use multi-valve engine heads to further lower the exhaust valve temperature [22, 23, 24]; ensure proper oil control; provide additional engine coolant passages around valves and other areas with high thermal Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

221

loads [27] (if possible); delay fuel introduction to create a period of air cooling (using timed manifold or direct injection); ensure adequate scavenging (e.g. using variable valve timing [9, 20]) to decrease residual gas temperatures. • Piston rings and crevice volumes: decrease the piston top land clearance to prevent hydrogen flames from propagating into the top land, Swain et al. [27] use a clearance of 0.152 mm to quench the hydrogen flame. Change the crevice volumes and/or piston rings with the aim of reducing the reflow of unburned mixture from the second land to the top land [27, 28, 29] (preventing ‘fueling’ of a top land flame during exhaust and intake). The smaller quenching distance of a hydrogen flame also implies an increased thermal load for the piston top land, Berger et al. [64] report changes (a special coating) to the top piston ring groove area to account for this.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• Valve seats and injectors: the very low lubricity of hydrogen has to be accounted for, suitable valve seat materials have to be chosen [22, 24] and the design of the injectors should take this into account. This is the case with any dry gaseous fuel (such as natural gas) but can be more critical for hydrogen (compressed natural gas contains small amounts of oil originating from the oil mist in the compressor whereas hydrogen compressors normally have tighter clearances to limit the leak rate). • Lubrication: an engine lubrication oil compatible with increased water concentration in the crankcase has to be chosen [25], the report on the hydrogen drive test in Ger¨ [24] cites two options, a demulsifying oil and a synthetic oil which many by TUV forms a solution with water. DeLuchi [65] claims a longer oil lifetime as the oil is not diluted by hydrogen and there is less formation of acids (perhaps doubtful given the large quantities of water and N Ox that can be formed during stoichiometric combustion). An ashless oil is recommended to avoid deposit formation (hot spots) [18, 59]. Measurements at the authors’ department [66], of the composition of the gases in the crankcase, showed a very high percentage of hydrogen (+ 5 vol%, out of range of the testing equipment), arising from the blowby. Blowby can be expected to be quite high because of the rapid pressure rise and the low density of hydrogen gas. The composition of the lubricating oil (semi synthetic ’universal’ oil, viscosity class 15W50) was investigated and compared to that of the unused oil. The properties of the oil had severely changed with a strong decrease of the lubricating qualities. The concentration of various additives (both lubricating and wearresisting, e.g. zincdialkyldithiophosphate) had strongly decreased, esters appearing in the unused oil had almost completely disappeared in the used oil. These conclusions were drawn from the difference in absorption of the various elements in an infrared spectrum. This is understandable when one knows that hydrogen is used in the industry to harden oils to fats (breaking open the double carbon to carbon bonds). The viscosity of the oil in atmospheric conditions had increased (causing more friction during starting) and decreased more quickly when the temperature rose (causing poor lubrication when the engine is at operating temperature). The kinematic viscosity at 40o C of the used oil was 141.9 mm2 /s, as compared to the value for the unused oil of 111.8 mm2 /s. At 100o C these values were respectively 14.33 mm2 /s versus

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

222

Sebastian Verhelst and Roger Sierens 17.25 mm2 /s. The viscosity index of the used oil thus amounted to 99, substantially lower than that of the unused oil of 163. An X-ray fluorescence spectrometry showed no substantial engine part wear, which is normal considering the limited amount of testing time of the engine. This means that all changes of the oil characteristics are to be ascribed to the influence of the blowby gases. An engine oil specifically developed for hydrogen engines is probably the best solution but currently unavailable. For safety reasons, a forced crankcase ventilation system (see also below) was mounted on the engine to keep the hydrogen concentration well below the lower flammability limit. Air is fed to the crankcase from the lab compressed air net, set to a small overpressure using a pressure regulating valve. A vacuum pump is used to evacuate the crankcase gases, which pass an oil separator first. The crankcase pressure is controlled to a slight underpressure by a balance between the compressed air pressure and a bypass valve on the vacuum pump inlet. The resulting hydrogen concentration in the crankcase with the ventilation system was measured to be below 1 vol%.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• Crankcase ventilation: positive crankcase ventilation is generally recommended due to unthrottled operation (high manifold air pressures) and to decrease hydrogen concentrations (from blowby) in the crankcase [22, 67]. • Compression ratio: the choice of the optimal compression ratio is similar to that for any fuel, it should be chosen as high as possible to increase engine efficiency, with the limit given by increased heat losses or appearance of abnormal combustion (in the case of hydrogen primarily surface ignition). The choice may depend on the application, as the optimum compression ratio for highest engine efficiency might be different from the optimum for highest power output [35]. In general, the compression ratio of a hydrogen engine can be chosen higher than that for a gasoline engine. • In-cylinder turbulence: because of the high flame speeds of hydrogen, low turbulence combustion chambers (pancake or disk chamber and axially aligned symmetric intake port) can be used which are beneficial for the engine efficiency [23, 27, 68]. They might even be necessary to avoid abnormal combustion at stoichiometric operation [27]. • Electronic throttle: as discussed in more detail below, hydrogen engines should be operated at wide open throttle wherever possible, but throttling is needed at very low loads to maintain combustion stability and limit unburned hydrogen emissions. At medium to high loads, throttling might be necessary to limit N Ox emissions. This can only be realized with a drive-by-wire system.

2.2.

2nd Generation H2 ICEs

Based on the knowledge gained from the proof-of-concept research, the ‘1st generation’ of H2 ICEs, a number of demonstration projects have been set up. The resulting demonstration vehicles used H2 ICEs optimized for reliable operation, which can be seen as ‘2nd Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

223

generation’ H2 ICEs. These vehicles have resulted in additional interesting information on the optimal set-up of H2 ICEs as well as useful data on power output, efficiency and emissions in real-world operation. Without going into detail, a few of these vehicles are listed here, grouped by manufacturer. More information can be found in the original references.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• BMW: the 750hL [69, 70, 71] had a 5.4 liter bi-fuel V12 engine using PFI gasoline and a separate (multipoint sequential) PFI system for hydrogen (with a dedicated electronic pressure regulator). This involved changes to the intake manifold as well as changes to the ignition system. The hydrogen operation strategy is the qualitative, lean burn, wide open throttle approach (see below). 15 of these vehicles were used, covering over 140 000 km. The BMW 745h based on the subsequent 7 series used a bi-fuel 4.4 liter V8 with essentially the same technology. Later, the new 6 liter V12 was also converted to bi-fuel operation. This engine was used in the series production run of 100 BMW ‘Hydrogen 7’ vehicles and uses a hydrogen PFI system next to the standard gasoline DI system [59, 64]. The load control strategy is more complex, using both qualitative and quantitative control strategies based on the power demand (this is further discussed in the next section). Recently, a mono-fuel derivative of the BMW Hydrogen 7 was demonstrated, with an optimized catalyst loading. This variant achieves very impressive emission results [72]. • Ford: the P2000 [9, 22, 73] used a 2 liter 4 cylinder engine optimized for hydrogen (mono-fuel), with the original references containing a wealth of data on the mapping parameters etc. The strategy chosen for the demo vehicle was throttled at a fixed lean equivalence ratio, but data was also taken for a qualitative approach. Later, the Ford model U show vehicle used a supercharged 2.3 liter variant of the engine, in a hybrid set-up. This powertrain was also used in the H2 RV prototype [74]. Next to these passenger cars, E-450 shuttle buses were also converted to hydrogen, using a 6.8 liter supercharged V10 [75]. • MAN: the first MAN hydrogen-fueled city buses were equipped with 6 cylinder inline bi-fuel engines, with separate PFI systems for gasoline and hydrogen [76]. Subsequent buses were mono-fuel [25]. The MAN buses were (and are) operated for public transport in Berlin and Munich and have accumulated over 400 000 km. • Mazda: the properties of the rotary engined RX-8 were put to good use by converting the engine to hydrogen (for a rotary engine the intake is in a different place than the combustion, so backfire can be eliminated). The vehicle is bi-fuel, with the original PFI system for gasoline and a separate DI+PFI system for hydrogen [56, 77]. The load control strategy switches between lean operation and stoichiometric operation (with EGR), depending on the power demand and engine speed. The same powertrain was later used in a hybrid set-up in the Premacy H2 RE vehicle. • Quantum: a Toyota Prius converted to hydrogen operation is offered by Quantum Technologies. The original hybrid powertrain is left untouched, but the engine is turbocharged, using a throttled strategy at a fixed lean equivalence ratio [78]. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

224

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2.3.

Sebastian Verhelst and Roger Sierens

3rd Generation: Increasing the Power Output

Hydrogen is a very versatile fuel when it comes to load control. The high flame speeds of hydrogen mixtures and its wide flammability limits permit very lean operation and substantial dilution. The engine efficiency and the emission of N Ox are the two main parameters used to decide the load control strategy. Constant equivalence ratio throttled operation has been used but mainly for demonstration purposes [44, 45] (also see previous section), as it is fairly easy to run a lean burn throttled hydrogen engine6 . Where possible, wide open throttle (WOT) operation is used to take advantage of the associated increase in engine efficiency [39, 40], regulating load with mixture richness (qualitative control) instead of volumetric efficiency (quantitative control) and thus avoiding pumping losses. Limitations to WOT operation are due to misfires, unburned hydrogen and decreased stability at very low load (e.g. idling) and N Ox emissions at medium to full load. Thus, throttling is used at very low loads to increase combustion stability and decrease unburned hydrogen emissions [9, 20, 24, 34, 52]. Moreover, this increases the efficiency at these conditions: the efficiency gain through decrease of unburned hydrogen emissions offsets the efficiency loss by throttling. The engine efficiency using throttled or WOT operation is compared in refs. [9, 35], the lean limit at which throttling is introduced is engine dependent and ranges from λ = 3 [24] to λ = 4 [9, 20]. For higher loads, flame temperatures quickly exceed the N Ox generation limit. This results in a N Ox limit to WOT operation. One could restrict the mixture richness and use sufficiently lean mixtures to stay below a 10 or 100 ppm N Ox limit, but this implies a large decrease in maximum power output. Alternatively, the engine can be throttled above this limit, using stoichiometric mixtures and thus enabling the use of a conventional three way catalyst for N Ox reduction7 [20], with an associated decrease in engine efficiency. Yet another strategy is the use of EGR to control load: using stoichiometric mixtures but instead of throttling, recycling exhaust gas in a proportion dependent on the power demand [35, 79]. This gives a better efficiency compared to throttling. EGR is also a means to allow backfire-free operation at stoichiometric mixtures, enabling a higher power output if N Ox emissions are a boundary condition [54, 79]. Water injection can also be used to decrease N Ox emissions from the richer mixtures [24, 49], and is more effective than EGR [44] but is mostly considered impractical. Work has been reported using a ‘dual fluid injector’ for DI [54], which injects hydrogen and liquid water directly in the combustion chamber, for decreased N Ox . As stated above, a PFI hydrogen engine operating stoichiometric at WOT, has a theoretical power deficit of about 18% compared to a gasoline engine, due to the lower volumetric energy density. In practice, the power deficit can be even higher if the equivalence ratio has to be limited to avoid abnormal combustion. Variable valve timing has been used to enable hydrogen engines to run stoichiometric without backfire [20], through better scavenging of hot exhaust gases. There are a number of options to increase the power output to levels higher than a comparable atmospheric gasoline engine. Supercharging is one means. Berckm¨uller et 6

Accepting the severe power output penalty The mixture richness is then set slightly rich of stoichiometric so that some unburned H2 is present in the exhaust which is a very effective reducing agent for N Ox 7

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

225

al. [20] report work on a supercharged stoichiometric PFI H2 engine, reaching peak power outputs of about one third above that of an atmospheric gasoline engine. Compared to the naturally aspirated case, they had to lower the compression ratio from 12 to 11:1 when charging to 0.85 barg to enable stoichiometric operation without the occurrence of surface ignition. The cylinder head coolant flow and the valve timing were optimized to minimize surface ignition tendencies. Supercharging lean burn PFI engines is also reported [35, 80], although this requires special measures if gasoline-level power outputs are to be reached. White et al. [81] present a review of H2 ICEs in which they discuss work on boosted H2 ICEs, concluding that such engines are a compelling alternative to fuel cells when used in a series hybrid vehicle, at lean equivalence ratios so that very low emissions are reached without any after treatment (more about this option can be found at the end of this section).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

When liquid hydrogen is used as the on-board vehicle storage, one can take advantage of the extremely low temperatures to increase the volumetric energy density of a PFI H2 engine. Heller and Ellgas [82] report measurements using extremely cold (gaseous) H2 injection in the intake manifold, resulting in mixture intake temperatures of down to 210K. This increases the bmep with about 25% when compared to ‘warm’ PFI and results in power output similar to or higher than gasoline.

A third option for higher power is the use of direct injection. As mentioned above, based on volumetric energy density, a DI hydrogen engine can have up to 17% higher power output compared to gasoline [61]. The possibility of postponing the introduction of hydrogen can eliminate backfire, so in practice the power increase over PFI H2 can be even higher than what would be expected based on the difference in volumetric energy density. The higher power output, backfire avoidance as well as the extra freedom in mixture formation (homogeneous, stratified, mixing controlled combustion, . . . ) have led to lots of H2 DI development efforts recently [36, 54, 83, 84, 85, 86]. Initial results on power output, N Ox emissions and efficiency look very promising. However, the engine hardware poses some challenges. First of all, the DI injectors need further development [83], to reach the high flow rates needed at full load and most of all, for durability. The injectors are subjected to a high thermal load (high flame temperature, high flame speeds, short quenching distance) and receive no cooling from a vaporizing fuel as gaseous hydrogen is injected. For the same reason, overfueling cannot be used to decrease exhaust gas temperatures at high loads and engine speeds to protect the catalyst. For mono-fuel applications, the catalyst can be placed further downstream as there are no cold-start HC emissions [83].

If a hydrogen engine is designed for mostly single speed/power operation, e.g. for stationary power generation or for a series hybrid vehicle, very clean and highly efficient operation is possible without any after treatment (of which the effectiveness could deteriorate with time). N Ox emissions below 10 ppm or even 1 ppm, with indicated efficiencies of perhaps 50% are possible, effectively approaching fuel cell efficiencies [40, 68, 80, 87, 88]. Hydrogen is the only fuel with which this is possible (with hydrocarbons, decreasing N Ox emissions with lean burn implies increased unburned hydrocarbon emissions).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

226

3. 3.1.

Sebastian Verhelst and Roger Sierens

Optimizing PFI Engines Introduction

As stated above, H2 PFI engines offer a number of advantages. However, their theoretical power output is lower compared to gasoline, and in practice, if abnormal combustion phenomena make stoichiometric combustion impossible, the power deficit can be even higher. This section discusses work at Ghent University focussed on increasing the power output of H2 PFI engines while limiting N Ox emissions, and maximizing engine efficiency. First, the often reported efficiency benefit of H2 enabled control strategies compared to gasoline is quantified. Next the potential of variable valve timing is investigated, looking at an extension of the wide open throttle strategy and the effects on the possibility of stoichiometric combustion. Finally, supercharging experiments are reported, where the use of exhaust gas recirculation (EGR) enabled stoichiometric operation with a resulting high catalyst efficiency.

3.2.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3.2.1.

Efficiency Comparison Hydrogen-Gasoline Experimental Set-up

A Volvo four cylinder 16 valve gasoline engine with a total swept volume of 1783cc and a compression ratio of 10.3:1 was converted to bi-fuel operation by mounting an additional fuel rail supplying gaseous fuel to 8 (2 per cylinder) Teleflex GSI gas injectors, mounted on the intake manifold. The intake manifold was modified to avoid any damage if backfire would occur during the hydrogen measurements: a T-type branch pipe was mounted on the intake manifold with the ’straight ahead’ branch closed by a foam plug and the other branch leading to the air filter - mass air flow (MAF) sensor - throttle valve assembly, see Fig. 1. Any pressure rise in the intake manifold due to the occurrence of backfire results in the foam plug being blown out instead of damaging other components such as the MAF sensor. The engine has continuously variable valve timing (CVVT) on the intake camshaft, allowing up to 40 degrees crank angle (ca) advance of the intake valve opening and closing time. A MoTeC M800 engine control unit is used to control ignition timing, start of injection, injection duration and intake valve timing. The exhaust gas components O2 , CO, CO2 , N O, N Ox and H2 are measured. A direct reading of the air to fuel equivalence ratio λ is given by a Bosch wide band sensor and digital air/fuel ratio meter with calibrations for hydrogen and gasoline. The engine test bench has a tapered roof with a Buveco Bucom ST600EX hydrogen sensor located at the highest point. 3.2.2.

Results

For reasons discussed above, the brake thermal efficiency (BTE) of a hydrogen ICE is expected to be substantially higher than that of a gasoline ICE. The bi-fuel Volvo engine is the suitable experimental set-up to examine this statement. The efficiencies while using gasoline and hydrogen will be compared at different torque settings (20, 40 and 80 Nm equivalent to 1.41, 2.82 and 5.64 bar bmep) and different engine speeds since the influence of the gas dynamics increases with an increasing engine speed. For hydrogen, high torque

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

227

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Modified intake of four-cylinder engine. outputs were limited by the appearance of backfire. At each point, MBT-timing and a fixed Intake Valve Opening (IVO) angle of 4 o ca BTDC was used. The influence of the IVO advance on the brake torque is fairly limited (up to 3% torque rise), as described in the next section. This justifies the fixed IVO advance. Most hydrogen measurements were carried out at WOT. The highest BTE is expected for this strategy. This is mainly caused by the lower pressure losses in the intake system as a consequence of the absence of throttle losses. The remaining measurements were taken for a throttle position of 50% to be able to evaluate the influence of the increasing pressure losses. As a consequence of throttling, the mixture had to be set richer for the same power output. For TP=50% it was not possible to run the engine at 80 Nm since the mixture became too rich with backfire as a consequence. Figures 2 and 3 show the brake thermal efficiencies BTE as a function of engine speed, for fixed torque outputs of 20 and 40 Nm respectively. Three BTE curves are shown, one is for gasoline (throttled, stoichiometric) operation for which the corresponding throttle position is shown in the middle part of the graph. The other two are for hydrogen with wide open throttle and with a TP=50% respectively, for these two the corresponding equivalence ratios are given in the top part of the graph. From Figs. 2 and 3 it is clear that at these low loads, the brake thermal efficiency on hydrogen is (much) higher than on gasoline, the hydrogen BTEs are 40 to 60% higher relative to the gasoline BTEs. This difference is due to the absence of throttling losses (or much lower throttle losses in the H2 TP=50% case) and the lean mixtures for hydrogen. The higher burning velocity of hydrogen is also a contributing factor, as this leads to a more isochoric combustion. The influence of this factor can be seen directly in Fig. 3 for the 4500 rpm point: the throttle position for gasoline is about 50% there, so the efficiency can be compared to the H2 TP=50% case. The BTE on hydrogen for this condition is about 18% higher relative to gasoline. This difference is not entirely down to a difference in burning velocity however: as hydrogen displaces more air due to its low density, throttling losses are lower even though the throttle position is identical, as the lower air flow results

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

228

Sebastian Verhelst and Roger Sierens

5

λ (-)

4 3 2

50

H2 WOT H2 TP=50%

45

gasoline

40 35

TP (%)

1

30 25 25

20

BTE (%)

15 20

15

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

10

1000

2000

3000

4000

5000

rpm Figure 2. Brake thermal efficiency and respective throttle position or equivalence ratio as a function of engine speed, for a fixed brake torque of 20 Nm.

in lower flow losses. The gasoline BTE can be seen to be relatively insensitive to the engine speed. On the one hand, the increasing air flow with engine speed cause higher flow losses but on the other hand, as seen from the TP curve, the throttle opening has to be increased with engine speed to keep the torque output constant, which decreases the throttling losses. Both effects seem to cancel each other out. For hydrogen, the BTE decreases with engine speed, although the decrease is less pronounced for the WOT case. Two effects explain this behavior: first, due to the lean burn operation and large throttle openings, the air flow is much higher in the hydrogen case than for gasoline. This leads to higher flow losses in the intake manifold. To keep the torque output fixed, this means more hydrogen needs to be injected at higher engine speeds to com-

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

229

4

λ (-)

3

2

1

55

H2 WOT

50

gasoline

45 40

TP (%)

H2 TP=50%

35 30 25

BTE (%)

33

20

30 27 24 21

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

18 1000

2000

3000

4000

5000

rpm Figure 3. Brake thermal efficiency and respective throttle position or equivalence ratio as a function of engine speed, for a fixed brake torque of 40 Nm.

pensate for these higher losses, with a reduced efficiency as a consequence. The increasing hydrogen flow can be seen in the λ curves in Figs. 2 and 3, which show a decreasing air-tofuel equivalence ratio with increasing engine speed. This increasing hydrogen flow results in a second effect, of an increasing air displacement and thus lower flow losses. However, the net air flow increases with engine speed so the intake flow losses increase with engine speed. For the TP=50% case, there are also throttle losses so the efficiency is lower than for the WOT case, and decreases more strongly with engine speed. For the lowest engine speeds at the lowest load (Fig. 2) the BTE for the H2 TP=50% case can be seen to be slightly higher than for the WOT case. Here, the increased turbulence due to throttling is beneficial for the combustion stability of the ultra lean mixtures (see the λ curves).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

230

Sebastian Verhelst and Roger Sierens

In Figs. 2 and 3, the air-to-fuel equivalence ratio for the WOT case is always higher than 2, as a result N Ox emissions are very low. All these points can thus be part of a practical load control strategy. For the TP=50% points, the equivalence ratios do drop under 2, with corresponding high N Ox emissions, in an oxygen-rich exhaust. Thus, these cannot be used in practice. However, as the BTE of this strategy is lower than in the WOT case this is not an issue. With a throttle position at 50%, it was not possible to reach 80 Nm with hydrogen. Thus Fig. 4, which shows the brake thermal efficiencies BTE as a function of engine speed, for a fixed torque outputs of 80 Nm, does not contain a TP=50% curve. Instead, for the throttled hydrogen case, the TP was set so that a stoichiometric mixture was obtained. In that case, N Ox emissions can be treated in a TWC (three way catalyst) with high conversion efficiencies.

2.0 H2 WOT 1.8

H2 λ=1

λ (-)

gasoline 1.6 1.4 65

1.2

60

50 45 40 35

35 30

BTE (%)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

55

TP (%)

1.0

30

25

20

1000

2000

3000

4000

5000

rpm Figure 4. Brake thermal efficiency and respective throttle position or equivalence ratio as a function of engine speed, for a fixed brake torque of 80 Nm.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

231

First, comparing Figs. 2, 3 and 4, the efficiencies of both gasoline and hydrogen can be seen to increase as the delivered torque increases. The explanation differs slightly for gasoline and hydrogen. In the two cases, as a result of the increasing torque, the mechanical efficiency increases strongly. For gasoline, the flow losses across the throttle valve decrease because of a larger TP. In the case of hydrogen, the flow losses decrease because of a smaller air flow since more air is displaced by hydrogen as a result of the richer mixture. This also leads to a decreased influence of engine speed on the hydrogen BTEs: from the figures it can be seen that as the load increases, the BTE decreases less strongly with engine speed. Secondly, Fig. 4 shows that hydrogen WOT measurements at 80 Nm (for all engine speeds) have an air-to-fuel equivalence ratio between 1 and 2. This is below the ‘threshold’ equivalence ratio, taking a N Ox emission of 100 ppm as the threshold. As explained above, for an equivalence ratio between 1 and 2 it is impossible to reduce the N Ox emissions with sufficient efficiency using a TWC since the exhaust oxygen concentration is too high. As a consequence, these points are useless for an automotive application. The efficiency penalty caused by the N Ox boundary condition can be seen by comparing the H2 WOT BTE curve to the H2 λ = 1 BTE curve, showing a relative decrease in BTE of 5 to 10%. Finally, for the 80 Nm case, the BTE on hydrogen is still higher than on gasoline, but the difference is lower than for the lower loads (between 10 and 30% relatively, comparing gasoline to the H2 λ = 1 case). In Fig. 4 there are no values given for the H2 λ = 1 case at 4500 rpm. This condition could not be set because of backfire occurrence.

3.3.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3.3.1.

Influence of Variable Valve Timing Introduction

The Volvo bi-fuel engine used for the efficiency comparison discussed above, was also used to check the impact of a variable intake valve event on torque output and emissions using hydrogen. Initially, measurements were limited to 1500 rpm due to limitations of the hydrogen storage in the lab at that time. The results at 1500 rpm will be discussed first, after which the work is extended to include higher engine speeds. All measurements reported here are for MBT ignition timing. Both qualitative load control (varying the equivalence ratio at WOT) and quantitative load control (throttling at stoichiometric) have been investigated. 3.3.2.

Wide Open Throttle Operation

Figure 5 shows results at wide open throttle. Here, the air-to-fuel equivalence ratio is used to regulate load, with the brake torque increasing as more hydrogen is injected. Brake torque and N Ox emissions are shown as a function of the air-to-fuel equivalence ratio, for different intake cam phasing. Advancing IVO at a fixed equivalence ratio leads to a higher brake torque. As injection durations had to increase to keep the equivalence ratio constant, this means the volumetric efficiency increases with advancing IVO. This is caused by the low momentum of the fresh charge at low engine speed, meaning it is best to close the intake valves early8 so no fresh charge is pushed back in the intake manifold at the end of 8

As the valve opening time is fixed, advancing the IVO means an earlier intake valve closing time

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

232

Sebastian Verhelst and Roger Sierens

120

5000

100

4000

80 3000

brake torque (Nm)

raw NOx emissions (ppm)

the intake phase.

60

2000

raw NOx at 4°ca IVO advance

40

raw NOx at 16°ca IVO advance raw NOx at 28°ca IVO advance raw NOx at 40°ca IVO advance

1000

20

Brake torque at 4°ca IVO advance Brake torque at 16°ca IVO advance Brake torque at 28°ca IVO advance

0

0 1

1.5

2

2.5

3

air to fuel equivalence ratio

Brake torque at 40°ca IVO advance

3.5

(-)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Brake torque and N Ox emissions on hydrogen, at 1500 rpm and wide open throttle, for different intake cam phasing, as a function of the equivalence ratio. N Ox emissions have no clear trend with the intake cam phasing. As the throttle valve is fully opened, there is no depression in the intake, no backflow of exhaust gas to the intake during the valve overlap, i.e. there is no effect on the amount of internal EGR and thus, N Ox emissions are unaffected by the intake valve timing. The threshold equivalence ratio for this engine can be seen to be around λ = 2. One would expect the N Ox emissions to peak around λ = 1.2. This peak is not visible in Fig. 5 as the measurements were done at 0.25 increments in air-to-fuel equivalence ratio. An important result to note from Fig. 5 is that it was not possible to run at stoichiometric for the most advanced intake cam timings. At these timings, backfire occurred. It is not clear what causes this. There is an increased contact between fresh charge and exhaust gas due to the longer valve overlap, but the injection duration of hydrogen for these low speeds is such that there is more than enough time to delay the start of injection until after the exhaust valve closes and still have an end of injection before BDC. Contrary to our findings, Tang et al. [9] did not find an influence of the intake valve timing on the richest mixture possible for WOT on their 2.0 liter Zetec engine. In their case, the mixture richness was limited (to λ ∼ 1.35) due to the occurrence of surface ignition events. The gain in torque output with the richer mixtures that are possible with the small intake cam advance, far outweighs the effects of the intake cam phasing. For example, at λ = 1.5, which is the richest point possible using the maximum intake cam advance of 40o ca, the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

233

brake torque is 83.2 Nm, compared to 70 Nm at 4o ca IVO advance. However, up until 16o ca IVO advance, stoichiometric operation is possible resulting in a torque output of 97 Nm (compared to about 120 Nm on gasoline). 3.3.3.

Throttled Operation

6000

100 90

5500 80 5000

70 60

4500

brake torque (Nm)

raw NOx emissions (ppm)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

From Fig. 5 it is clear that the region between stoichiometric mixtures and the threshold equivalence ratio is to be avoided due to excessive N Ox emissions and (oxygen-rich) conditions that prohibit efficient after treatment. When the load demand is higher than that reachable (from the viewpoint of acceptable N Ox ) with the WOT, qualitative control strategy, one has to switch to a throttled strategy at stoichiometric mixtures. Figure 6 shows results for stoichiometric mixtures as a function of the throttle position. Again, an advanced IVO increases the brake torque. This effect is strongest for the larger throttle openings, as the momentum of the gas (fresh charge or exhaust gas) in the manifolds is higher. The mechanism by which torque increases is that less fresh charge is being pushed back in the intake when closing the intake valve shortly after BDC. For the smaller throttle openings, an advanced IVO leads to higher internal EGR due to the depression in the intake. At these throttle openings, the conflicting trends between internal EGR and volumetric efficiency result in a negligible effect of intake cam phasing on torque output.

50 4000

40 30

3500

raw NOx at 4°ca IVO advance raw NOx at 16°ca IVO advance raw NOx at 28°ca IVO advance raw NOx at 40°ca IVO advance

20 3000

Brake torque at 4°ca IVO advance Brake torque at 16°ca IVO advance

10 2500

0 20

40

60

80

Brake torque at 28°ca IVO advance Brake torque at 40°ca IVO advance

100

throttle position (opening %)

Figure 6. Brake torque and N Ox emissions on hydrogen, at 1500 rpm, λ = 1, for different intake cam phasing, as a function of throttle opening. The effects on N Ox emissions for throttled hydrogen operation are expected to be a deDemidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

234

Sebastian Verhelst and Roger Sierens

crease in N Ox with advancing IVO at the smaller throttle openings (internal EGR effects), and a negligible effect for the larger throttle openings (no internal EGR). This is partly confirmed from the results in Fig. 6, although the results at 4o ca IVO advance do not follow the expected trend. It can be seen in Fig. 6 that no results are shown for advanced IVO phasing at the larger throttle openings. This is because the maximum throttle opening for these IVO timings is limited due to the occurrence of backfire. This is similar to the limitations on the richest equivalence ratio for advanced IVO in Fig. 5. However, for the throttled case the benefit of IVO advance on torque is greater than going to a less advanced IVO and WOT (the highest torque output of 100 Nm is reached at an IVO advance of 28o ca and 80% open throttle, compared to 97 Nm at 16o ca and WOT).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3.3.4.

Extension to Higher Engine Speeds

The experiments on the effects of a variable intake valve timing were extended to higher engine speeds: measurements have been taken at WOT for 2500, 3500 and 4500 rpm. For these measurements, the amount of hydrogen was held constant at 0,23 Ndm3 /cycle, for the different speeds. This corresponds with an air-to-fuel equivalence ratio of around 2. This way it is possible to evaluate the influence of IVO advance on N Ox emissions for a fixed energy input. The comparison has been limited to λ = 2, because backfire occurred at some conditions using stoichiometric mixtures. At λ = 2, WOT operation and medium to high engine speeds, the inertia of the exhaust gases is relatively high. During the valve overlap, the inertia of the exhaust gases creates a vacuum that sucks fresh charge into the cylinder. Thus, a longer valve overlap period leads to a better cylinder filling. This is confirmed by Fig. 7, showing how the brake torque varies with IVO advance for 2500, 3500 and 4500 rpm. The torque outputs rise slightly (up to 3% gain) if the IVO advance increases, for all engine speeds. The higher the engine speed, the higher the inertia of the exhaust gases, so cylinder filling and torque output vary more strongly with IVO advance as the engine speed is increased. The torque output can be seen to vary with engine speed, this results from variations in volumetric efficiency, increasing flow losses and decreasing heat losses with an increasing engine speed. Figure 8 shows the raw N Ox emissions and the corresponding air-to-fuel equivalence ratios at the different engine speeds, as a function of the IVO advance. Due to the increasing amount of air per cycle, caused by the suction effect, the air-to-fuel equivalence ratio increases with an increasing IVO advance. Also, the greater amount of air has a cooling effect in the cylinder, causing a lower intake mixture temperature. Both effects lead to less N Ox formation. The relative position of the N Ox emission curves for the three engine speeds in Fig. 8 is a consequence of various influences: the time available for N Ox formation is higher for the low engine speeds and results in higher emissions; at high engine speeds there is less heat loss to the walls resulting in higher emissions; and as shown in Fig. 7 the brake torque varies with engine speed which also changes in-cylinder temperatures. For the conditions shown in Figs. ?? and 8 it can be concluded that the brake torque is maximized and N Ox emissions minimized, with the most advanced IVO of 40o ca BTDC.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

235

71 IVO=4°ca

IVO=16°ca

IVO=28°ca

IVO=40°ca

69

brake torque (Nm)

67 65 63 61 59 57 55 2500

3500

4500

engine speed (rpm)

Figure 7. Brake torque for WOT λ ∼ 2 operation, influence of intake valve opening time shown for different engine speeds. 2.4 2.3

2.1

2500 rpm 3500 rpm 4500 rpm

NOx (ppm)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

200 160

λ (-)

2.2

2.0 1.9 1.8

120 80 40 0 0

10

20

30

IVO (°ca BTDC)

40

Figure 8. Engine-out N Ox and actual air-to-fuel equivalence ratios for WOT λ ∼ 2 operation, influence of intake valve opening time shown for different engine speeds.

3.4. 3.4.1.

Supercharging and EGR Introduction

With a naturally aspirated, PFI H2 ICE, the N Ox emissions are a trade-off against the power AtAirlow loads, WOT operation possible. As long as λ remains above Demidov, Sergey,engine and Jacques Bonnet. output. Traffic Related Pollution and Internal Combustion Engines,is Nova Science Publishers, Incorporated, 2009.

236

Sebastian Verhelst and Roger Sierens

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

a value of around 2 (specific values depend on the type of engine), N Ox emissions are nearzero. For lower air to fuel ratios the N Ox emissions rise very steeply. For higher loads, it is required to throttle and to operate stoichiometric. With stoichiometric mixtures, after treatment of the exhaust gases is possible, using a TWC. When using this strategy, higher power outputs are possible, while losing in efficiency. Instead of throttling in order to have a stoichiometric mixture, EGR can be used. With EGR, a part of the exhaust gases is brought back into the cylinder, diluting the fresh charge. Therefore, the peak cylinder temperatures are lower, so N Ox formation is reduced [49]. After treatment with a TWC is also possible. The power output is determined by the amount of EGR. Because of the lower pumping losses, due to the WOT operation, efficiencies reach higher values. For hydrogen PFI engines to achieve a similar power output as with gasoline ICEs, supercharging can be used [20]. There are two options: either the air-fuel mixture has to be kept lean enough, so that N Ox formation stays limited. The air to fuel ratio at which the N Ox emissions reach a value of 100 ppm (the threshold value) depends on the supercharging pressure. By using this method, power output is increased by increasing the charging pressure. In ref. [20] engine boosting experiments with boosted manifold air pressure until 1 barg have been performed. The other option is a strategy that combines EGR and supercharging, so that the engine can run stoichiometric. Because of the diluting effect and the possibility to use a TWC, N Ox emissions do not exceed the limit of 100 ppm. The EGR also diminishes the risk of abnormal combustion. Therefore, high output power and high efficiencies are achieved. In the following, supercharging and a combination of supercharging and EGR is studied with the aim of increasing the power output with N Ox emissions as a restriction. 3.4.2.

Experimental Set-up

The single cylinder 2 valve engine used in this study is based on an Audi-NSU research engine. Figure 9 shows the layout of the test bench. The engine has a swept volume of 400 cc. The compression ratio was set at 11:1 by machining a DI diesel piston. The engine is coupled to a DC motor and is operated between 1500 and 4500 rpm. Two Teleflex GSI gas injectors are used for port fuel injection of gaseous fuels. Fuel (hydrogen) is supplied at 2 barg. Fuel mass flow is measured using a Bronkhorst Hi-Tec sensor. The test bench is equipped with a Busch MM1102BP claw compressor. Both rotors are fitted with a waterproof coating to stand up to the condensation water coming from the EGRsystem. A volumetric compressor was chosen in order to be able to supercharge the engine up to 1 barg, at low as well as at high engine speeds. The compressor is driven by a 7 kW electric motor which is fed by an inverter, allowing varying the inlet pressure by adjusting the compressor speed. A damper vessel (200 liter) is used to allow intake air mass flow measurements (strongly pulsating flow) with a second Bronkhorst Hi-Tec sensor. The second purpose of this relatively large damper vessel is making the inlet pressure less sensitive to little variations of the compressor speed. On the other hand, a large damper vessel is to the detriment of the time necessary to reach a stable gas composition in case of EGR operation. In order to control the gas temperature before the gas enters the engine, an adjustable

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

237

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 9. Single cylinder engine bench layout. intercooler is placed between the compressor and the damper vessel. Exhaust gas can be recirculated from the exhaust to the intake of the compressor. The layout as shown in Fig. 9 was chosen to avoid the complexity of a high pressure loop EGR (supercharging + EGR combination). It turned out to be necessary to install a choke valve to create vacuum at the inlet of the compressor. This vacuum overcomes the pressure losses in the relatively long EGR-pipe, and causes a flow of exhaust gases. The surplus of energy to create this vacuum is provided by the compressor. In order to protect the compressor against excessively high inlet temperatures, an EGR-cooler was installed to keep the inlet temperature around 50o C. This is a compromise because it is necessary to keep the temperature high enough to prevent the production of condensation water in the compressor as much as possible. At the end of the exhaust system, a TWC was installed. A TWC makes it possible to reduce the N Ox emissions at stoichiometric operation. For a more efficient reduction, mixtures slightly rich of stoichiometric should be used so that the small amount of excess hydrogen can act as reducing agent. The recirculation of unburned hydrogen is a safety concern as in a worst case scenario hydrogen could accumulate in the damper vessel until a flammable mixture is formed. The engine test bench has a tapered roof with a Buveco Bucom ST600EX hydrogen sensor located at the highest point. As an additional safety measure for the supercharging experiments, a small controlled flow is taken from the damper vessel (less than 0,5% of the intake flow) through a flow controller and pressure reduction valve to another ST600EX hydrogen sensor. This set-up is necessary as most commercial hydrogen sensors cannot handle strong pressure fluctuations or overpressure. The monitoring of hydrogen concentration in the (large) damper vessel precludes any build-up of a combustible mixture in the supercharged vessel. A MoTeC M4Pro engine control unit is used to control ignition timing, start of injection and injection duration. A cold rated spark plug with a silver central electrode was used to minimize spark plug hot spots and catalytic reactions. Again, the exhaust gas components O2 , CO, CO2 , N O, N Ox and H2 are measured. A direct reading of the air to fuel equivalence ratio λ is given by a Bosch wide band sensor and digital air/fuel ratio meter with calibration for hydrogen.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

238 3.4.3.

Sebastian Verhelst and Roger Sierens Results

Supercharging is an effective way to increase the power output of hydrogen ICEs. By supercharging, the density and the mass of the fresh charge increase. However, a higher charging pressure leads to a higher peak cylinder pressure and temperature. As a result, the likelihood of hot spot formation and the occurrence of backfire increases. For this engine, running at stoichiometric (allowing for after treatment with a TWC) was impossible. Consequently, in order not to exceed the N Ox limit of 100 ppm, the air-fuel mixture has to be lean enough (lean burn). Experiments have been conducted at an engine speed of 2000 rpm and WOT at charging pressures from 0 barg (atmospheric) to 1 barg. Measurements have been taken at these and the intervening charging pressures, with an interval of 0.2 barg. At these measuring points, the richness of the mixture at which the N Ox emissions remained just below the N Ox limit of 100 ppm was determined. For every charging pressure, the torque that corresponds with this λ, is the maximum torque that could be reached, with the N Ox limit as a restriction. The results of the experiments are shown in Fig. 10. All brake torque and efficiency numbers given in this Section are net values, accounting for the energy needed for the supercharging. Concerning the BTE values, the low mechanical efficiency of a single cylinder research engine should be kept in mind. As the main interest is a comparison of BTE values between strategies, this is no problem.

2.6

2.0

28

Me (Nm)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2.2

λ (-)

2.4

1.8

24 20 16 12 0.0

0.2

0.4

0.6

0.8

1.0

charging pressure (barg)

Figure 10. Supercharging at the N Ox threshold: brake torque and air-to-fuel equivalence ratio. As can be seen in Fig. 10, the minimal richness of the mixture, to keep N Ox emissions below the threshold, decreases (λ increases) as the charging pressure increases. If the airto-fuel equivalence ratio would not have been increased, the cylinder temperature would increase, accelerating N Ox formation, resulting in N Ox emissions higher then 100 ppm. Although the mixture richness decreases, brake torque increases with increasing charging

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

239

pressure. A maximum brake torque of 23.9 Nm (bmep = 7.47 bar, if the research engine were to run on gasoline a maximum bmep of about 8.5 bar is estimated) is reached at the maximum charging pressure of 1 barg with λ = 2.5. For atmospheric operation the torque output is 13.3 Nm (bmep = 4.15 bar) with λ = 1.9. At lean operation, supercharging up to 1 barg thus results in a net torque increase of 80%. Table 1 shows the backfire limit (air-tofuel equivalence ratio at which backfire occurred) and the corresponding brake torque for charging pressures from 0 to 1 barg. This table shows that the backfire limit is well below the air-to-fuel equivalence ratio at which the N Ox limit of 100 ppm is reached, so there is no risk of backfire at these points. Table 2. Backfire limit as a function of supercharging pressure 0 0.95 20.7

0.2 1.09 23.6

0.4 1.23 24.5

0.6 1.36 26.6

0.8 1.50 32.4

1.0 1.64 35.4

When combining supercharging with EGR, stoichiometric operation is possible. This is due to the diluting effect of the exhaust gases, as a result of which N Ox formation is reduced. Stoichiometric operation allows for the use of a TWC. As with lean operation, measurements were taken at an engine speed of 2000 rpm with charging pressures varying from 0 to 1 barg. For this range of charging pressure, the amount of EGR was set so that N Ox emissions after conversion (i.e. after the TWC) remained just below the N Ox limit. Thus, with N Ox emissions as a restriction, the corresponding brake torque is maximal at this EGR percentage for a particular charging pressure. Figure 11 shows the brake torque and EGR percentage as a function of the charging pressure.

48 44 40 36

24

Me (Nm)

EGR %

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Pressure (barg) Backfire limit (λ) Brake torque (Nm)

21 18 15 12 0.0

0.2

0.4

0.6

0.8

1.0

charging pressure (barg)

Figure 11. Supercharging at the N Ox threshold: brake torque and EGR percentage (λ = 1). Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

240

Sebastian Verhelst and Roger Sierens

In order to keep the cylinder peak temperature and N Ox formation limited, the EGR percentage increases with increasing charging pressure, with a maximum of 45.9% at a charging pressure of 1 barg. Similar to the lean operation, the brake torque increases when increasing the charging pressure. At atmospheric operation, the brake torque is 12.3 Nm (bmep = 3.85 bar). At 1 barg a maximum torque of 22.6 Nm (bmep = 7.08 bar) is reached. Supercharging in combination with EGR, results in a power increase over atmospheric operation (λ = 1) of 84.7%. Figure 12 shows the brake torque (see also Figs. 10 and 11) and efficiency for both methods. For the whole pressure range, the efficiency of the lean burn strategy exceeds the efficiency of the λ = 1+EGR operation. This is due to the higher power needed for supercharging of the mixture of EGR and fresh charge at EGR operation, as a result of the higher temperature at the inlet of the compressor. The brake torque that could be achieved is similar for both methods. At 1 barg, where the maximum torque for both methods is reached, the torque at lean operation exceeds the torque at EGR operation. 34 32 30

26

Me (Nm)

28

lean burn λ=1+EGR

24

24

BTE (%)

28

22

20 16

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

12 0.0

0.2

0.4

0.6

0.8

1.0

charging pressure (barg)

Figure 12. Comparison of brake torque and efficiency as a function of supercharging, between lean burn operation and stoichiometric+EGR operation, at the N Ox threshold. Similar experiments have been conducted at an engine speed of 3000 rpm. Table 3 summarizes the results for both engine speeds at a charging pressure of 1 barg. As with 2000 rpm, the torque and efficiency at 3000 rpm are optimal when operating lean. However, with this engine configuration, it was not possible to run at exactly λ = 1. Because of the safety reasons discussed above (buffer vessel and low pressure EGR loop), the air-fuel mixture was set slightly lean, with λ ∼ 1, 07. When deviating from stoichiometric operation, the conversion efficiency of the TWC decreases drastically. Therefore, when taking the 100 ppm limit into account, the raw N Ox emissions (before conversion) had to be limited. This restricts the maximum torque. Taking into account the non-optimized TWC location, low conversion efficiencies were to be expected. When assuming a more realistic (conservative) conversion efficiency of 95% for the TWC, the restriction on the engine-out N Ox emissions are less severe. As a consequence, the necessary EGR percent-

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

241

Table 3. Comparison of maximum brake torque and corresponding efficiency when supercharging, between lean burn operation and stoichiometric+EGR operation, at the N Ox threshold, for 2000 rpm and 3000 rpm

Engine speed 2000 rpm 3000 rpm

Lean burn Me (Nm) ηe (%) 23.9 28.6 25.6 27.7

EGR Me (Nm) ηe (%) 22.6 24.7 21.2 22.5

age decreases to 23%, while brake torque and efficiency increase. At a charging pressure of 1 barg, the brake torque reaches a maximum value of 34.8 Nm. This means an increase of 46% in comparison with the maximum torque at lean operation. This brake torque of 34.8 Nm exceeds the torque of a comparable gasoline engine with almost 30%. The brake thermal efficiency is 27.9%, which is similar to the efficiency at lean burn. These results emphasize the importance of a correct λ-control.

4.

Conclusion

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

This text presented an overview of past and present experimental research on hydrogenfueled internal combustion engines. The most important conclusions from this overview are: • there remains a lot to be learned about the abnormal combustion phenomena when operating on hydrogen: what are the exact causes, how can they be avoided in practice, how can they be modeled for development work • there are various demonstration vehicles now, accumulating many kilometers while emitting very few emissions. However, all of these vehicles are still restricted in power output compared to gasoline • power outputs greater than on gasoline have been demonstrated on research engines, as well as efficiency numbers exceeding diesel efficiencies, all with extremely low emissions. These results now need to be translated to actual (demonstration) vehicles. • further research is needed into the fundamentals: abnormal combustion, as stated above; injection and mixture formation; laminar and turbulent combustion of hydrogen in engines; so that modeling can be improved. In the second half, experiments on hydrogen-fueled PFI engines were discussed. The most important results are: • the brake thermal efficiency of a hydrogen PFI engine exceeds that of a gasoline engine over the entire speed and load range. N Ox emissions control incurs an efficiency penalty at the higher loads, although hydrogen keeps its efficiency advantage over gasoline. At low loads, important in drive cycles as well as for everyday driving, Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

242

Sebastian Verhelst and Roger Sierens the efficiency gains of hydrogen compared to gasoline are particularly high (up to 60% relative increase of the BTE)

• variable valve timing was shown to influence brake torque and N Ox emissions, with brake torque increasing and N Ox emissions decreasing for an advanced intake valve opening time. This was the case both for wide open throttle (lean burn) operation and throttled (stoichiometric) operation. Backfire was a limiting factor for the IVO advance in some cases. VVT allows an extension of the qualitative control range to slightly higher engine loads, which is beneficial for the brake thermal efficiency • supercharging is a means to increase the power output. When supercharging at equivalence ratios lean of the N Ox threshold, the maximum power output at the currently used maximum supercharging pressure of 1 barg exceeded that of stoichiometric atmospheric operation but was still lower than what would be obtained on gasoline. Supercharging while using stoichiometric mixtures was possible by introducing EGR. That way, power outputs of up to 30% higher compared to gasoline could be reached (assuming a 95% TWC conversion efficiency), albeit at lower efficiencies compared to the lean burn strategy

Acknowledgements The authors would like to thank technicians Rene Janssens and Patrick De Pue for their efforts, and the numerous master thesis students choosing the H2 ICE as their graduation thesis topic, for conducting many experiments throughout the years.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Nomenclature BDC barg bmep BTDC BTE CVVT DI EGR FC ICE IVO MAF MBT PFI PV rpm TDC TP TWC

bottom dead center bar gauge (pressure relative to atmospheric) brake mean effective pressure before top dead center brake thermal efficiency continuously variable valve timing direct injection exhaust gas recirculation fuel cell internal combustion engine inlet valve opening time mass air flow minimum spark advance for best torque port fuel injection photovoltaic revolutions per minute top dead center throttle position three way catalyst

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

243

WOT wide open throttle λ air-to-fuel equivalence ratio

References [1] Verhelst S. A Study of the Combustion in Hydrogen-Fuelled Internal Combustion Engines. PhD thesis, Ghent University, URL http://hdl.handle.net/1854/3378, 2005. [2] Verhelst S., Sierens R., and Verstraeten S. A critical review of experimental research on hydrogen fueled SI engines. SAE, paper nr 2006-01-0430, also in SAE 2006 Transactions, Journal of Engines, pp264-274, 2006. [3] United Nations Framework Convention on Climate Change, http://unfccc.int/. [4] Rifkin J. The hydrogen economy. Polity Press, UK edition, 2002. [5] Grube T. and Peters R. Evaluation of biofuel production routes. International Conference Automotive Biofuels, Berlin, Germany, 2008. [6] Weindorf W. and Altmann M. Yield of biofuels versus hydrogen from photovoltaics and wind power. Ludwig-B¨olkow-Systemtechnik GmbH, 2007.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[7] Gosselink J.W. Pathways to a more sustainable production of energy: sustainable hydrogen – a research objective for Shell. Int. J. Hydrogen Energy, 27:1125–1129, 2002. [8] Kazuyuki N., Morisama H., Odaka M., Kamiya Y., Daisho Y., and Murooka K. Study on fuel cell poisoning resulting from hydrogen fuel containing impurities. Fisita World Automotive Congress, paper nr F2004F397, 2004. [9] Tang X., Kabat D.M., Natkin R.J., Stockhausen W.F., and Heffel J. Ford P2000 hydrogen engine dynamometer development. SAE, paper nr 2002-01-0242, 2002. [10] Akagawa H., Ishida H., Osafune S., Egashira H., Kuma Y., Furutani H., and Iwasaki W. Development of hydrogen injection clean engine. 15th World Hydrogen Energy Conference, paper nr 28J-05, Yokohama, Japan, July 2004. [11] Das L.M. Hydrogen engines: a view of the past and a look into the future. Int. J. Hydrogen Energy, 15:425–443, 1990. [12] Das L.M. Hydrogen-oxygen reaction mechanism and its implication to hydrogen engine combustion. Int. J. Hydrogen Energy, 21:703–715, 1996. [13] Karim G.A. Hydrogen as a spark ignition engine fuel. Int. J. Hydrogen Energy, 28:569–577, 2003. [14] Herdin G., Gruber F., Klausner J., Robitschko R., and Chvatal D. Hydrogen and hydrogen mixtures as fuel in stationary gas engines. SAE, paper nr 2007-01-0012, 2007. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

244

Sebastian Verhelst and Roger Sierens

[15] Heywood J.B. Internal Combustion Engine Fundamentals. McGraw-Hill, 1988. [16] Bardon M.F. and Haycock R.G. The hydrogen research of R.O. King. 14th World Hydrogen Energy Conference, invited paper, Montreal, Canada, 2002. [17] MacCarley C.A. A study of factors influencing thermally induced backfiring in hydrogen fuelled engines, and methods for backfire control. 16th IECEC conference, Atlanta, USA, 1981. [18] Das L.M. Near-term introduction of hydrogen engines for automotive and agricultural application. Int. J. Hydrogen Energy, 27:479–487, 2002. [19] Lucas G.G. and Morris L.E. The backfire problem of the hydrogen engine. Symposium organized by the university’s internal combustion engine group, King’s College, London, UK, 1980. [20] Berckm¨uller M., Rottengruber H., Eder A., Brehm N., Els¨asser G., M¨uller-Alander G., and Schwarz C. Potentials of a charged SI-hydrogen engine. SAE, paper nr 200301-3210, 2003. [21] Ciatti S.A., Wallner T., Ng H., Stockhausen W.F., and Boyer B. Study of combustion anomalies of H2 -ICE with external mixture formation. ASME Spring Technical Conference, paper nr ICES 2006-1398, 2006.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[22] Stockhausen W.F., Stockhausen W.F., Natkin R.J., Kabat D.M., Reams L., Tang X., Hashemi S., Szwabowski S.J., and Zanardelli V.P. Ford P2000 hydrogen engine design and vehicle development program. SAE, paper nr 2002-01-0240, 2002. [23] Swain M.R., Swain M.N., and Adt R.R. Consideration in the design of an inexpensive hydrogen-fueled engine. SAE, paper nr 881630, 1988. ¨ Rheinland e.V. for the [24] Project Coordinator Motor Vehicles and Road Transport, TUV Federal Ministry for Research and Technology. Alternative energy sources for road ¨ Rheinland, 1990. transport - hydrogen drive test. Technical report, TUV [25] Pr¨umm W. Hydrogen engines for city buses. 1st (Int.) Symp. on Hydrogen Internal Combustion Engines, pages 1–11, 2006. [26] Kondo T., Iio S., and Hiruma M. A study on the mechanism of backfire in external mixture formation hydrogen engines –about backfire occurred by the cause of the spark plug–. SAE, paper nr 971704, 1997. [27] Swain M.R., Schade G.J., and Swain M.N. Design and testing of a dedicated hydrogen-fueled engine. SAE, paper nr 961077, 1996. [28] Koyanagi K., Hiruma M., and Furuhama S. Study on mechanism of backfire in hydrogen engines. SAE, paper nr 942035, 1994. [29] Lee J.T., Kim Y.Y., Lee C.W., and Caton J.A. An investigation of a cause of backfire and its control due to crevice volumes in a hydrogen fueled engine. ASME Spring Technical Conference, paper nr 2000-ICE-284, San Antonio, USA, 2000. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

245

[30] Lee S.J., Yi H.S., and Kim E.S. Combustion characteristics of intake port injection type hydrogen fueled engine. Int. J. Hydrogen Energy, 20:317–322, 1995. [31] Sierens R. and Rosseel E. Backfire mechanism in a carburetted hydrogen fuelled engine. 12th World Hydrogen Energy Conference, pages 1537–1546, Buenos Aires, 1998. [32] Huynh T.C., Kang J.K., Noh K.C., Lee J.T., and Caton J.A. Controlling backfire using changes of the valve overlap period for a hydrogen-fueled engine using an external mixture. ASME Fall Technical Conference, paper nr ICEF 2007-1702, 2007. [33] Jorach R.W. Low nitrogen oxide emission combined with high power density by using the hydrogen combustion method with advanced internal mixture formation. MTZ Worldwide, 58-4:5–8, 1997. [34] Jing-Ding L., Ying-Qing L., and Tian-Shen D. Improvement on the combustion of a hydrogen fueled engine. Int. J. Hydrogen Energy, 11:661–668, 1986. [35] Natkin R.J., Tang X., Boyer B., Oltmans B., Denlinger A., and Heffel J.W. Hydrogen IC engine boosting performance and N Ox study. SAE, paper nr 2003-01-0631, 2003. [36] Wimmer A., Wallner T., Ringler J., and Gerbig F. H2 -direct injection - a highly promising combustion concept. SAE, paper nr 2005-01-0108, 2005.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[37] Binder K. and Withalm G. Mixture formation and combustion in hydrogen engine using hydrogen storage technology. Int. J. Hydrogen Energy, 7:651–659, 1982. [38] Li H. and Karim G.A. Knock in spark ignition hydrogen engines. Int. J. Hydrogen Energy, 29:859–865, 2004. [39] Heffel J.W., Johnson D.C., and Shelby C. Hydrogen powered Shelby Cobra: vehicle conversion. SAE, paper nr 2001-01-2530, 2001. [40] Smith J.R., Aceves S., and Van Blarigan P. Series hybrid vehicle and optimized hydrogen engine design. SAE, paper nr 951955, 1995. [41] Li H. and Karim G.A. Hydrogen fuelled spark ignition engines: predictive and experimental performance. ASME Spring Technical Conference, paper nr ICES2003-548, Salzburg, Austria, 2003. [42] Sierens R. Installation and first experimental results of a hydrogen fuelled engine. 9th World Hydrogen Energy Conference, pages 31–40 (Addendum), 1992. [43] Sierens R. and Verhelst S. Comparison between a carburetted and a port injected hydrogen fuelled single cylinder engine. 8th EAEC European Automotive Congress, pages paper nr. SAITS 01009,, Bratislava, 2001. [44] Davidson D., Fairlie M., and Stuart A.E. Development of a hydrogen-fuelled farm tractor. Int. J. Hydrogen Energy, 11:39–42, 1986. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

246

Sebastian Verhelst and Roger Sierens

[45] Olavson L.G., Baker N.R., Lynch F.E., and Meija L.C. Hydrogen fuel for underground mining machinery. SAE, paper nr 840233, 1984. [46] Heffel J.W., McClanahan M.N., and Norbeck J.M. Electronic fuel injection for hydrogen fueled internal combustion engines. SAE, paper nr 981924, 1998. [47] Sierens R. and Rosseel E. Sequential injection of gaseous fuels. 5th EAEC European Automotive Congress, paper nr SIA 9506A03, 1995. [48] Sierens R. and Verhelst S. Influence of the injection parameters on the efficiency and power output of a hydrogen fueled engine. Transactions of the ASME: J. Eng. Gas Turbines and Power, 125:444–449, 2003. [49] Subramanian V., Mallikanjuna J.M., and Ramesh A. Effect of water injection and spark timing on the nitric oxide emission and combustion parameters of a hydrogen fuelled spark ignition engine. Int. J. Hydrogen Energy, 32:1159–1173, 2007. [50] Meier F., K¨ohler J., Stolz W., Bloss W.H., and Al-Garni M. Cycle-resolved hydrogen flame speed measurement with high-speed Schlieren technique in a hydrogen direct injection SI engine. SAE, paper nr 942036, 1994. [51] Furuhama S. Problems of forecasting the future of advanced engines and engine characteristics of the hydrogen injection with LH2 tank and pump. Journal of Engineering for Gas Turbines and Power, 119:227–242, 1997.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[52] Guo L.S., Lu H.B., and Li J.D. A hydrogen injection system with solenoid valves for a four-cylinder hydrogen-fuelled engine. Int. J. Hydrogen Energy, 24:377–382, 1999. [53] Kim J.M., Kim Y.T., Lee J.T., and Lee S.Y. Performance characteristics of hydrogen fueled engine with the direct injection and spark ignition system. SAE, paper nr 952498, 1995. [54] Bleechmore C. and Brewster S. Dilution strategies for load and N Ox management in a hydrogen fuelled direct injection engine. SAE, paper nr 2007-01-4097, 2007. [55] Nande A.M., Wallner T., and Naber J. Influence of water injection on performance and emissions of a direct-injection hydrogen research engine. SAE, paper nr 2008-012377, 2008. [56] Wakayama N., Morimoto K., Kashiwagi A., and Saito T. Development of hydrogen rotary engine vehicle. 16th World Hydrogen Energy Conference, 2006. [57] Yi H.S., Min K., and Kim E.S. The optimised mixture formation for hydrogen fuelled engines. Int. J. Hydrogen Energy, 25:685–690, 2000. [58] Drew A.N., Timoney D.J., and Smith W.J. A simulation and design tool for hydrogen SI engine systems - validation of the intake hydrogen flow model. Int. J. Hydrogen Energy, 32:3084–3092, 2007. [59] Kiesgen G., Kl¨uting M., Bock C., and Fischer H. The new 12-cylinder hydrogen engine in the 7 series: the H2 ICE age has begun. SAE, paper nr 2006-01-0431, 2006. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

247

[60] Peschka W. and Escher W.J.D. Germany’s contribution to the demonstrated technical feasibility of the liquid-hydrogen fueled passenger automobile. SAE, paper nr 931812, 1993. [61] Rottengruber H., Berckm¨uller M., Els¨asser G., Brehm N., and Schwarz C. A highefficient combustion concept for direct injection hydrogen internal combustion engine. 15th World Hydrogen Energy Conference, paper nr 28J-01, Yokohama, Japan, 2004. [62] Gerbig F., Strobl W., Eichlseder H., and Wimmer A. Potentials of the hydrogen combustion engine with innovative hydrogen-specific combustion proces. Fisita World Automotive Congress, paper nr F2004V113, Barcelona, Spain, 2004. [63] Verhelst S. and Sierens R. Hydrogen engine - specific properties. Int. J. Hydrogen Energy, 26:987–990, 2001. [64] Berger E., Bock C., Fischer H., Gruber M., Kiesgen G., and Rottengruber H. The new BMW 12-cylinder hydrogen engine as clean efficient and powerful vehicle powertrain. Fisita World Automotive Congress, paper nr F2006P114, Yokohama, Japan, 2006. [65] DeLuchi M.A. Hydrogen vehicles: an evaluation of fuel storage, performance, safety, environmental impacts, and cost. Int. J. Hydrogen Energy, 14:81–130, 1989. [66] Sierens R. and Verhelst S. Experimental study of a hydrogen fuelled engine. Transactions of the ASME: J. Eng. Gas Turbines and Power, 123:211–216, 2001.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[67] Strebig K.C. and Waytulonis R.W. The bureau of mines’ hydrogen powered mine vehicle. SAE, paper nr 871678, 1987. [68] Van Blarigan P. Development of a hydrogen fueled internal combustion engine designed for single speed/power operation. SAE, paper nr 961690, 1996. [69] Wolf J.F. and Nordheimer R.K. BMW’s energy strategy – promoting the technical and political implementation. SAE, paper nr 2000-01-1324, 2000. [70] Pehr K., Burckhardt S., Koppi J., Korn T., and Partsch P. Mit Wasserstoff in die Zukunft - der BMW 750hL. Automobiltechnische Zeitschrift - ATZ, 104:120–131, 2002. [71] Sch¨uers A., Abel A., Fickel H.C., Preis M., and Artmann R. Der Zw¨olfzylinderWasserstoffmotor im BMW 750hL. Motortechnische Zeitschrift - MTZ, 63:98–105, 2002. [72] Wallner T., Lohse-Busch H., Gurski S., Duoba M., and Thiel W. Fuel economy and emissions evaluation of BMW Hydrogen 7 mono-fuel demonstration vehicles. Int. J. Hydrogen Energy, in press, 2008. [73] Szwabowski S.J., Hashemi S., Stockhausen W.F., Natkin R.J., Reams L., Kabat D.M., and Potts C. Ford hydrogen engine powered P2000 vehicle. SAE, paper nr 2002-010243, 2002. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

248

Sebastian Verhelst and Roger Sierens

[74] Jaura A.K., Ortmann W., Stunz R., Natkin B., and Grabowski T. Ford’s H2 RV: an industry first HEV propelled with a H2 fueled engine - a fuel efficient and clean solution for sustainable mobility. SAE, paper nr 2004-01-0058, 2004. [75] Lapetz J., Natkin R., and Zanardelli V. The design, development, validation and delivery of the Ford H2 ICE E-450 shuttle bus. 1st (Int.) Symp. on Hydrogen Internal Combustion Engines, pages 20–33, 2006. [76] Knorr H., Held W., Pr¨umm W., and R¨udiger H. The MAN hydrogen propulsion system for city buses. Proceedings 11th World Hydrogen Energy Conference, pp1611-1620, Stuttgart, Germany, June 1996. [77] Tomoaki S., Masanori M., Hiroaki M., and Takayuki V. Development of hydrogen rotary engine with dual-fuel system. Fisita World Automotive Congress, paper nr F2006P200, Yokohama, Japan, 2006. [78] Geiss R., Webster B., Ovshinsky S.R., Stempel R., Chiang Young R., Li Y., Myasnikov V., Falls B., and Lutz A. Hydrogen-fueled hybrid: pathway to a hydrogen economy. SAE, paper nr 2004-01-0060, 2004. [79] Heffel J.W. N Ox emission reduction in a hydrogen fuelled internal combustion engine at 3000 rpm using exhaust gas recirculation. Int. J. Hydrogen Energy, 28:1285–1292, 2003.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[80] Lohse-Busch H., Wallner T., and Shidore N. Efficiency-optimized operating strategy of a supercharged hydrogen-powered 4-cylinder engine for hybrid environment. SAE, paper nr 2007-01-2046, 2007. [81] White C.M., Steeper R.R, and Lutz A.E. The hydrogen-fueled internal combustion engine: a technical review. Int. J. Hydrogen Energy, 31:1292–1305, 2006. [82] Heller K. and Ellgas S. Optimisation of a hydrogen internal combustion engine with cryogenic mixture formation. 1st (Int.) Symp. on Hydrogen Internal Combustion Engines, pages 49–58, 2006. [83] Rottengruber H., Berckm¨uller M., Els¨asser G., Brehm N., and Schwarz C. Direct injection hydrogen SI engine - operation strategy and power density potentials. SAE, paper nr 2004-01-2927, 2004. [84] Eichlseder H., Wallner T., Freyman R., and Ringler J. The potential of hydrogen internal combustion engines in a future mobility scenario. SAE, paper nr 2003-012267, 2003. [85] Wallner T., Nande A.M., and Naber J. Evaluation of injector location and nozzle design in a direct-injection hydrogen research engine. SAE, paper nr 2008-01-1785, 2005. [86] Mohammadi A., Shioji M., Nakai Y., Ishikura W., and Tabo E. Performance and combustion characteristics of a direct injection SI hydrogen engine. Int. J. Hydrogen Energy, 32:296–304, 2007. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Recent Progress in Hydrogen-Fueled Internal Combustion Engines

249

[87] Aceves S.M. and Smith J.R. Hybrid and conventional hydrogen engine vehicles that meet EZEV emissions. SAE, paper nr 970290, 1997.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[88] Keller J. and Lutz A. Hydrogen fueled engines in hybrid vehicles. SAE, paper nr 2001-01-0546, 2001.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution... Editors: S. Demidov and J. Bonnet, pp. 251-338

ISBN 978-1-60741-145-1 c 2009 Nova Science Publishers, Inc.

Chapter 11

I N S EARCH OF I MPROVEMENTS FOR THE C OMPUTATIONAL S IMULATION OF I NTERNAL C OMBUSTION E NGINES Ezequiel J. L´opez∗, Norberto M. Nigro† and Mario A. Storti‡ Centro Internacional de M´etodos Computacionales en Ingenier´ıa (CIMEC), INTEC-CONICET, Universidad Nacional del Litoral, G¨uemes 3450, 3000 Santa Fe, Argentina§

Abstract

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The modeling of internal combustion engines is a multidisciplinary subject that involves chemical thermodynamics, fluid mechanics, turbulence, heat transfer, combustion, and numerical methods. In this chapter, we focus on some aspects of the computational resolution of the fluid dynamic problem. We present strategies designed in order to improve the simulation tools available today. A Computational Mesh Dynamics (CMD) to solve the movement of the mesh is presented. This kind of techniques are useful when an Arbitrary Lagrangian Eulerian (ALE) method is applied in the resolution of flows on moving domains. For in-cylinder flows in internal combustion engines, the domain has a very high relative deformation and even changes on its topology. This demands a CMD strategy with great robustness to avoid the deterioration of the grid quality and to reduce at minimum possible the number of remeshing needed in the whole simulation. The CMD strategy proposed is based on an optimization problem solved in a global way. The strategy can handle meshes with inverted elements, being a simultaneous mesh untangling and smoothing method. The flow inside of an internal combustion engine is characterized by a low Mach number, except in the early moments in which the exhaust valve (or port) is opened. The numerical methods for compressible flow based on the density fail when they are applied to flows with low Mach numbers, which is due to the ill-conditioning of the system of equations. For this reason, it is necessary to apply a technique that allows ∗

E-mail address: [email protected] E-mail address: [email protected] ‡ E-mail address: [email protected] § http://www.cimec.org.ar †

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

252

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti the resolution of compressible flows in all the range of Mach numbers, especially in the low Mach limit. Here, we apply the method of preconditioning of the equations in conjunction with the dual time stepping technique. The preconditioning matrix used was originally designed by Choi and Merkle to solve steady compressible flows with the Finite Volume Method. We adapt this matrix to the unsteady flows case via an eigenvalues analysis of the system. A stabilized Finite Element formulation for the preconditioned system of equations is presented. The dynamic boundary conditions on inlet/outlet of the 3D problem are obtained by using a code with thermodynamic (0D) and gas-dynamic (1D) models. Therefore, the need to couple appropriately the solutions obtained in the computational 1D and 3D domains arises. We propose a coupling strategy of 1D/multi-D domains for compressible flows based on constraints of the state at the coupling section. Finally, these tools are applied to solve the fluid flow in the novel rotative engine MRCVC.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1.

Introduction

The increase in the capability of computing in conjunction with the development of new mathematical models and numerical methods, allow to deal with the resolution of complex problems of importance for both science and engineering. Among these, the CFD (Computational Fluid Dynamics) problems in moving domains, such as Fluid-Structure Interaction (FSI) problems, are a topic of particular interest for researchers because of the difficulty that they present and the large number of applications in which these kind of problems are present. One of such problems is the computation of in-cylinder flows in internal combustion (IC) engines. The modeling of IC engines is a multidisciplinary subject that involves chemical thermodynamics, fluid mechanics, turbulence, heat transfer, combustion, and numerical methods. In this chapter, we focus on some aspects of the computational resolution of the fluid dynamics problem. In particular, the topics addressed are the mesh dynamics problem, the ill-conditioning of the system of flow equations at low Mach numbers, and the coupling of 1D/multi-D domains for compressible flows. When an Arbitrary Lagrangian Eulerian (ALE) strategy is applied to solve problems with deformable domains, it is necessary to have a Computational Mesh Dynamics (CMD) technique to resolve the dynamics of the mesh. While the movement of the mesh is an artificial field in a FSI problem, its significance is relevant because it affects considerably the efficiency and accuracy of the computation. For in-cylinder flows in IC engines the movement of the boundary domain is known a priori. In these cases the domain has a very high relative deformation and even changes on its topology. This demands great robustness from the CMD strategy to avoid an excessive deterioration of the grid quality and to reduce the number of remeshing needed in the whole simulation. The flow inside of an IC engine is characterized by a low Mach number, except in the early moments in which the exhaust valve (or port) is opened. The numerical methods for compressible flow based on the density fail when are applied to flows with low Mach numbers, which is due to the bad conditioning of the system of equations. For this reason, it is necessary to apply a technique that allows the resolution of compressible flows in all the range of Mach numbers, especially in the low Mach limit.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

253

Then, to perform a simulation in an IC engine is necessary to have a CFD code able of compute compressible turbulent flows with low (and also relatively high) Mach numbers in deformable 3D domains. Given the highly complex geometry of the IC engines and the physical processes that occur within them, it is at present only possible to solve one part of such machines with a 3D model. In this way, and because of its dynamic behavior, another difficulty that appears is related to the boundary conditions to impose to the model. Usually, these problems are addressed by the simulation of the rest of the engine through 0D/1D models, which is achieved in one hand, modeling the entire machine simultaneously (but the level of detail varies depending on the model) and, on the other hand, providing appropriate conditions to the 3D code. Applying the above approximation, the need to couple appropriately the solutions obtained in the computing domains arises, which can be calculated by different codes. The large spread in length and time scales of in-cylinder flows in IC engines requires a high degree of refinement in the finite element mesh and, then requires very large computational resources. Thus, a parallel code is needed in order to achieve accurate results in that problems. In addition, due to explicit and semi-implicit schemes have demonstrated to be inefficient when they are applied to IC engines [25], a full implicit scheme might be used. The base code utilized here is PETSc-FEM [55]. PETSc-FEM is a general purpose, parallel, multi-physics FEM program for CFD applications based on PETSc [6]. PETSc-FEM comprises both a library that allows the user to develop FEM (or FEM-like, i.e. non-structured mesh oriented) programs, and a suite of application programs.

2.

Governing Equations and Numerical Approximation

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2.1.

Governing Equations

The flow field in an internal combustion engine model is governed by the instantaneous time-dependent three-dimensional conservation equations of mass, momentum and energy. These equations can be simplified by neglecting radiation [49] and, in the particular case of this chapter, a single component is considered. Let Ω ⊂ Rnd the spatial domain and (0, tf ) the temporal domain, where nd is the number of space dimensions, and let Γ the boundary of Ω. The spatial and temporal coordinates are denoted by x and t, respectively. The Navier-Stokes equations governing the fluid flow, in conservation form, are ∂ρ + ∇ · (ρu) = 0 on Ω × (0, tf ) ∂t ∂(ρu) + ∇ · (ρuu) + ∇p − ∇ · T = ρfe ∂t

on Ω × (0, tf )

∂(ρE) + ∇ · (ρEu) + ∇ · (pu) − ∇ · (Tu) + ∇ · q = ρfe · u ∂t

(1)

on Ω × (0, tf )

where ρ, u, p, T, E and q are the density, velocity, pressure, viscous stress tensor, total energy per unit mass, and heat flux vector, respectively, and fe are external forces. Generally, these forces are null in IC engine simulation. It is assumed a perfect gas constitutive relation and a Newtonian fluid defined by the two viscosity coefficients λ and µ. Thus, the Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

254

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

viscous stress tensor is defined as T = µ((∇u) + (∇u)T ) + λ(∇ · u)I = 2µǫ(u) + λ(∇ · u)I

(2)

I being the second order identity tensor, ǫ(u) = 1/2((∇u) + (∇u)T ) is the deformation rate tensor and superscript T denotes transpose. In addition, it is considered the range of fluid behavior within local thermodynamic equilibrium for which the Stokes relation 3λ + 2µ = 0 is valid. Pressure is related to the other variables via the equation of state. For ideal cases, this equation have the form p = (γ − 1)ρe (3) where γ is the ratio of specific heats, and e is the internal energy per unit mass which is related to the total energy per unit mass and kinetic energy as 1 e = E − kuk2 2

(4)

q = −κ∇T

(5)

The heat flux vector is defined as

where κ is the heat conductivity and T is the temperature. In the particular case of ideal gases, the following relations hold e cv R cv = γ−1 γR cp = γ−1

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

T =

(6)

where cv is the specific heat of the fluid at constant volume, cp is the specific heat of the fluid at constant pressure, and R is the ideal gas constant. Prandtl number (P r) relates the heat conductivity of the fluid to its viscosity according to the following relation κ=

µcp Pr

(7)

The governing equations (1) can be written in compact form as [26] ∂Fdi ∂U ∂Fai + = +S ∂t ∂xi ∂xi

on Ω × (0, tf )

(8)

where U = [ρ, ρu, ρE]T is the vector of conservative variables, S = [0, ρfe , ρfe · u]T is the source vector, Fa and Fd are the advective (or inviscid) and viscous flux vectors respectively, defined as   ρui  ρu1 ui + δi1 p    a  Fi =  (9)  ρu2 ui + δi2 p   ρu3 ui + δi3 p  (ρE + p)ui

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...   0   Ti1   d   Ti2 Fi =     Ti3 Tik uk − qi

255

(10)

Here, ui and qi are the components of the velocity and heat flux vectors, respectively, Tik are the components of the viscous stress tensor, and δij is the Kronecker delta.. In the quasi-linear form, equation (8) is written as [26]   ∂U ∂ ∂U ∂U + Ai = Kij + S on Ω × (0, tf ) (11) ∂t ∂xi ∂xi ∂xj where

∂Fai ∂U is the advective jacobian matrix, and Kij is the diffusivity matrix satisfying Ai =

Kij

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2.1.1.

∂U = Fdi ∂xj

(12)

(13)

Turbulence Modeling

The flow field in an internal combustion engine is turbulent and comprises many time and length scales. The ratio between the time (length) scales of the Kolmogorov and energy-containing eddies is of the order of Re−1/2 (Re−3/4 ), where Re is the Reynolds number [56]. Thus, it is practically impossible with available processors obtain a numerical solution of equations (1) that accounts for all the turbulent time and length scales and, therefore, models need to be introduced. There are two main approaches to model turbulence, the so-called RANS (Reynolds-Average Navier-Stokes) approach, and the LES (Large Eddy Simulation) models. In RANS methods, a set of partial differential equations describing suitable averaged quantities are used everywhere in the flow. For periodic engine flows, ensemble or phase average is replaced instead typical time averaging [25, 49]. Since the flow during the engine cycle is compressed and expanded, mass-weighted averaging (called Favre averaging) is commonly applied in conjunction with ensemble averaging [25]. LES is an approach in which the large-scale three-dimensional time-dependent turbulence structure is calculated in a single realization of the flow. Thus, only the small-scale turbulence need to be modeled, which is more isotropic than the large-scale structure [67]. Although the RANS methods involve a greater number of equations than LES strategies, they are cheaper since they can work with coarser meshes and time steps. In addition, the LES methods could need a posteriori statistical analysis in order to compute turbulent variables. Nevertheless, RANS methods have demonstrated their inability to produce solutions for the Navier-Stokes equations. This is obviously a very serious shortcoming of any turbulence modeling procedure, and although it has been recognized for a long time by theorists, especially mathematicians, it has had little, if any, impact on engineering analyzes of turbulence. In contrast to RANS, it can be shown that LES procedures generally converge to DNS (Direct Numerical Simulation) as discretization step sizes are refined. Hence, the use of a LES flow solver is desirable.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

256

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

In the problems solved here, the simplest Smagorinsky model [67, 52] is applied, which takes the Smagorinsky coefficient as constant, in contrast with the dynamic counterpart proposed by Germano [22]. This is one of the most popular choices into the LES family of turbulence models. In this eddy viscosity model, the turbulent dynamic viscosity is defined as p (14) µt = ρ(Cs h)2 ∆ 2ǫ(u) : ǫ(u)

where Cs = 0.1 − 0.2 is the Smagorinsky constant, ∆ is a damping function to reduce the amount of turbulent viscosity in the vicinity of solid objects immersed in the fluid flow, and h is the grid size (a parameter that divides the size of vortexes being resolved by the sizes of p vortexes being modeled). Finally, ǫ(u) : ǫ(u) represents the trace of the strain velocity tensor making the eddy viscosity a local parameter. 2.1.2.

Boundary Conditions

At solid walls, the ‘classical’ approach to impose boundary conditions to the compressible Navier-Stokes equations is the non-slip condition [26]. For the velocity this condition is expressed by u = uwall (15) where uwall is the velocity of the wall in the considered reference system. For the temperature, either the wall temperature (Twall ) is fixed T = Twall

(16)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

or the heat flux is determined by the physical conditions, that is −κ

∂T = qwall ∂n

(17)

where qwall is the wall heat flux, and n refers to the normal direction to the wall. The last boundary condition at the wall could be obtained projecting the momentum equation on the normal direction ∂p = (∇ · T)n (18) ∂n For thin shear layers at high Reynolds numbers, this might be replaced by the boundary layer approximation [26] ∂p =0 (19) ∂n which was used for numerical simulations of internal combustion engines [49] by several researchers. For turbulent flows, non-slip boundary condition could be a mistake if the mesh is not enough refined at boundary layers. In addition, most of the turbulence models that have been used to calculate the flow field in reciprocating and rotary engines do not account for low-Reynolds number effects, preferential dissipation, and streamline curvature. Therefore, these models cannot be applied up to the solid walls, and the boundary conditions are applied close to the wall but not on it. These boundary conditions are analogous to those

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

257

used in statistically stationary, incompressible, turbulent flows along flat plates in the absence of pressure gradients. The flows in reciprocating and rotary engines are neither steady nor incompressible, but involve recirculation zones and pressure gradients. Nevertheless, boundary conditions based on solutions to flat plates are frequently used. For the velocity, the boundary conditions near a solid wall can be written as follows [49] (u − uwall ) · n = 0  2 u d uf d   if 0 ≤ ≤ 11.63  f ν ν  u·t= uf d uf d    uf 2.5 ln + 5.5 if > 11.63 ν ν

(20)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

where n and t are the unit vectors normal and tangential to the solid wall, respectively; uf = (τw /ρ)1/2 is the friction velocity, with τw the shear stress at the wall; d is the distance from the solid wall to the point closest to the wall; and ν the gas kinematic viscosity. The boundary condition for the temperature profile near the wall in a turbulent flow is frequently imposed by means of the Reynolds analogy between linear momentum and energy [50]. The Reynolds analogy is strictly applicable when the Reynolds fluxes of linear momentum and energy are equal, and the profiles of the mean velocity and enthalpy are similar, that is, the laminar and turbulent Prandtl numbers are close to 1. This boundary condition is expressed as [49]  P rqwall d uf d   if 0 ≤ ≤ 11.63  Twall − ρc ν ν p   T =  P rt qwall uf d uf d   Twall − + 5.5 if > 11.63 2.5 ln ρcp uf ν ν

(21)

νt is the turbulent Prandtl number, νt = µt /ρ and αt being, respectively, the αt turbulent kinematic viscosity and the turbulent thermal diffusivity. The analogy between linear momentum and energy may not be applicable to reciprocating and rotary engine flows. Thus, equation (21) should be used carefully. Conditions for inlet and outlet boundaries are applied following the approach proposed by Storti et al. [54]. Considering a point on an inlet/outlet boundary, it is possible to do a simplified 1D analysis in the normal direction to the local boundary. The projection matrices onto the right/left-going characteristics modes are defined as where P rt =

± −1 Π± n = Sn ΠV n Sn

(22)

where Sn is the matrix of eigenvectors diagonalizing the projected system, being Λn = diag [(λn )j ] their respective eigenvalues; and (Π− V n )jk

=

(

1 if j = k and (λn )j < 0 0 otherwise

+ Π− V n + ΠV n = I Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(23)

258

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Then, the boundary condition is applied as a constraint to the system of governing equations as follows ˆ ˆ Π− (24) n (U)(U − U) = 0 ˆ is defined depending on whether the boundary is either an inlet or an outlet. At where U ˆ = Uref , with Uref a reference state. At outlet regions, Storti et al. [54] inlet regions U ˆ as the state of fluid in the previous time step if the external conditions are propose to take U unknown. They named this strategy ULSAR (Use Last State As Reference) and show that Riemann invariants are preserved in the limit ∆t → 0 and h → 0, if such invariants exist. Note that in equation (24) the projection matrix, which is a non linear function of the fluid ˆ This is true if it is assumed that the flow is composed of state, is evaluated at the state U. ˆ However, as long as the fluid state departs from the small perturbations around the state U. ˆ value, the condition becomes less and less absorbing. U

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2.1.3.

Arbitrary Lagrangian Eulerian Description of Governing Equations

The Arbitrary Lagrangian Eulerian (ALE) strategy applied here for the resolution of moving domain problems is the proposed by Donea et al. [16]. It is well known that there are two viewpoints mostly used in the description of the flow motion equations: one is called the Lagrangian approach, where the observer moves with the fluid velocity, and the other option is the Eulerian approach, in which the observer is fixed. The ALE description is a generalization of these two approaches, where the observer moves with an arbitrary velocity w(x, t) in the laboratory system. Using this ALE strategy, it can be shown (see, for example, the works by Donea et al. [16], Nomura [44], etc.) that the system of Navier-Stokes in its quasi-linear form can be written as   ∂U ∂ ∂U ∂U + (Ai − wi I) = Kij + S on Ωt × (0, tf ) (25) ∂t ∂xi ∂xi ∂xj where wi are the components of the arbitrary velocity vector w. Note the subscript in Ωt denoting the fact that the domain could change in time.

2.2. 2.2.1.

Numerical Implementation Finite Element Formulation

In this section, the variational formulation of the Navier-Stokes equations for compressible flows is presented. The Finite Element Method stabilized by means of the Streamline Upwind/Petrov-Galerkin (SUPG) strategy and with the addition of a shock capturing operator is used. Consider a finite element discretization of the domain Ω into nel sub-domains Ωe , e = 1, 2, . . . , nel . Based on this discretization, the finite element function spaces for the trial solutions and for the weighting functions, S h and V h respectively, can be defined (see equation (27)). Then, the finite element formulation of the problem (11) using SUPG is written as follows [42]:

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... Find Uh ∈ S h such that ∀ Wh ∈ V h   Z Z h ∂Uh ∂Wh ∂Uh h h ∂U dΩ + W · + Ai · Khij dΩ ∂t ∂xi ∂xj Ω Ω ∂xi     nel Z h h h X ∂Uh ∂ h ∂U h ∂U h T ∂W · + Ai − Kij − S dΩe + τ (Ak ) ∂x ∂t ∂x ∂x ∂x e i i j k e=1 Ω Z Z nel Z X ∂Wh ∂Uh e + · dΩ = δsc Wh · SdΩ + Wh · f dΓ ∂x ∂x e h i i Ω Ω Γ

259

(26)

e=1

where S h = {Uh |Uh ∈ [H1h (Ω)]ndof , Uh |Ωe ∈ [P 1 (Ωe )]ndof , Uh = g

on Γg }

V h = {Wh |Wh ∈ [H1h (Ω)]ndof , Wh |Ωe ∈ [P 1 (Ωe )]ndof , Wh = 0 on Γg }

(27)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

H1h (Ω) being the finite dimensional Sobolev functional space over Ω, and with f and g representing the natural and Dirichlet boundary conditions vectors, respectively. Γg and Γh are the portion of the boundary with Dirichlet and Neumann (or Robin) conditions, respectively. The first series of element level integrals in equation (26) are added to the variational formulation to stabilize the computations against numerical instabilities. In the advectiondominated range, these terms prevent the node-to-node oscillations of the flow variables, where τ is known as the intrinsic time tensor. The second series of element level integrals in equation (26) are the shock capturing terms that stabilize the computations in the presence of sharp gradients, δsc being the coefficient of shock capturing. Matrix τ is defined by Aliabadi et al. [1] in the following way τ = max [0, τa − τd − τδ ] where τa = τd =

h I 2(c + kuk) Pnd 2 j=1 βj diag(Kjj )

(c + kuk)2 δsc τδ = I (c + kuk)2

I

(28)

(29)

√ Here, c = γRT is the sonic speed, h is the element size computed as the element length in the direction of the streamline 2 a=1 ks · ∇Na k

h = Pnen

(30)

Na being the trial function associated with the node a, nen the number of nodes in the element, s a unit normalized velocity vector, and β = ∇kUk2 /k∇kUk2 k. Regarding the shock capturing term, an isotropic operator proposed by Tezduyar and Senga [57] is presented here. Let j = ∇ρh /k∇ρh k a unit vector oriented with the density Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

260

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti P en gradient and the characteristic length hJGN = 2( na=1 kj · ∇Na k)−1 . The isotropic shock capturing factor included in equation (26) is then defined as hJGN δsc = uchar 2



k∇ρh khJGN ρref

β ∗

(31)

where uchar = kuk+c is the characteristic velocity defined as the addition of the flow velocity magnitude and the sonic speed, ρref is the density interpolated at gaussian point, and β ∗ is a parameter that could be taken as 1 or 2 according to the sharpness of the discontinuity to be captured [57]. Comparing (11) with (25), in the ALE formulation of Navier-Stokes equations only the advective jacobian are modified. Thus, the finite element formulation stabilized with SUPG of equation (25) is obtained by replacing Ai by Ai − wi I in (26), and in the definition of stabilization coefficients (29) kuk by ku − wk. In the case of moving domains, care should be taken with integrals containing time derivatives due to integration domain is a function of t. 2.2.2.

Time Discretization

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Derivatives with respect to time are discretized using the trapezoidal difference scheme, expressed as Uϑ − Un ∂U ≈ (32) ∂t ∆t where Uϑ = ϑUn+1 + (1 − ϑ)Un , ϑ is the implicitness parameter, ∆t is the time step and the superscript n indicates the level time t. Furthermore an implicit scheme is used, in such a way that all variables in equation (26) are evaluated at n + ϑ level time. 2.2.3.

Dynamic Boundary Conditions Using Lagrange Multipliers

Boundary conditions at inlet and outlet (24) are imposed via Lagrange multipliers, as proposed by Storti et al. [54]. Let i a node lying on the inlet (or outlet) boundary. Then, the equations for this node are modified in the following way + ˆ ˆ ˆ Π− n (U)(Ui − U) + Πn (U)Ulm = 0 ˆ lm = 0 Ri + Π− (U)U

(33)

n

where Ulm is the vector of Lagrange multipliers and Ri is the FEM residue for node i. Some internal combustion engines utilize ports for the gas-exchange process, such as two-stroke and rotative (Wankel [5], MRCVC [62], etc.) engines. Generally, the ports are placed on fixed walls of the engine (the cylinder or the housing) and, thus, have a relative motion respect to the flow domain. For example, figure 1 shows a scheme of a two-stroke engine with intake and exhaust ports located on cylinder wall. In this case, an observer placed on the centroid of the flow domain sees the area of ports moving away in the bottom direction .

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... flow domain

intake port

flow domain

exhaust port

intake port

261

flow domain

exhaust port

intake port

exhaust port

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Two-stroke engine scheme.

A port could be modeled as a ‘hole’ in relative motion with respect to the boundary domain. This hole changes its passage area as the boundary moves, from open position to the closed one, and viceversa. Due to the nodal displacement produced by the deformation of the flow domain, mesh nodes lying on a boundary with a hole could change their position between the wall and the port (an inlet/outlet for the flow problem). Therefore, the boundary condition applied on each of such nodes must be changed appropriately in order to account for the node position. This is sketched in figure 2, where the mesh boundary nodes are filled according to the type of boundary condition. At time t, nodes 1 to 5 have absorbing boundary conditions since they lie on the passage ports area. At time t + ∆t and due to the mesh movement, nodes 1 to 4 are located on the cylinder walls and their boundary condition must change to the wall type. The strategy used consists in switching from an absorbing boundary condition (equation (33)) when the node is placed on the port region to a wall boundary condition when the node moves on the solid wall. The wall boundary condition is applied by means of constraints using Lagrange multipliers in order to keep the total number of degree of freedom constant. For instance, in a 3D problem, using a no-slip boundary condition, and considering as null the velocity of the solid wall; the system of equations to solve for the node i is written as MUi + (I − M)Ulm = 0 Ri + MUlm = 0

where M = diag [0, 1, 1, 1, 0]. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(34)

262

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

t+ ∆t

t

intake port

1 2

3 4 5

exhaust port

1 2

intake port

3 4 5

exhaust port

wall boundary condition absorbing boundary condition

Figure 2. Change in the type of boundary condition due to the mesh movement.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3.

Mesh Dynamics

Several scientific and industrial applications of Computational Mechanics problems involve domains with moving boundaries. Examples of this kind of problems include free surfaces, two-fluid interfaces, fluid-object interaction, fluid-structure interaction, and moving mechanical components. In computation of fluid problems with moving boundaries and interfaces, either an interface-tracking or an interface-capturing technique could be used, depending on the complexity of the interface as well as other aspects of the problem. An interface-tracking technique requires meshes that ‘track’ the interfaces, then, they need to be updated as the flow evolves. Besides, in an interface-capturing technique the computations are based on fixed spatial domains, where an interface function sets the location of the interface. This function needs to be computed in order to ‘capture’ the interface within the finite element mesh covering the area where the interface is located [58, 59]. In fluid-structure interaction problems, one of the most popular interface-tracking techniques is the ALE formulation [29, 8, 16], as described in section 2.. In such applications, a body-conforming mesh has to be regenerated at each time step, or the existing grid has to be allowed to deform in order to follow the computational domain geometries. The former option is rather cumbersome and computationally expensive, especially for 3D problems, and could introduce an additional degradation of the numerical solution due to the projection of solutions from a mesh to another one. The latter option introduces the concept of a moving and deforming grid known as ‘dynamic’ mesh. In this case, the motion of the grid could cause the deterioration of the mesh quality and, in some situations, generate an invalid mesh where any of the grid elements is inverted. It is well known that poor quality elements have strong influence on stability, convergence and accuracy of the numerical method used. Furthermore, when implicit schemes are applied in an environment of parallel computing, the matrix profile must be calculated at each remeshing stage. More precisely, a total or a partial change in the topology of the mesh involves changes in the matrix profile. Thus, this additional computational cost introduced by the remeshing could become very important if the frequency of remeshing stages increases. Although a FSI problem may have instants

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

263

at which a new mesh must be introduced, the goal is to develop a Computational Mesh Dynamics (CMD) strategy that allows to reduce the total number of remeshing stages. For relatively small domain deformations, there are many techniques which can solve the dynamics of the mesh. However, when the boundary displacements are relatively high most of these methods could fail to give a valid mesh. The domain deformation in internal combustion engines is very high, with topological changes and contact between different boundaries. Therefore, the CMD strategy utilized to solve the mesh dynamics in these problems should be as robust as possible. An advantage in the particular case of internal combustion engines, is the fact that the movement of the boundaries is known a priori and has a periodic behavior. Thus, the complete sequence of meshes could be generated before the resolution of the CFD problem.

3.1.

Mesh Quality

Following the paper by Knupp [31], an element quality metric is defined as Definition. An element quality metric is a scalar function of node positions that measures some geometric property of the element. If a 3D element has J nodes with coordinates xj ∈ R3 , j = 0, 1, . . . , J − 1, then a mesh quality metric is denoted by fˆ : R3J → R. Some examples of element quality metrics are

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• Volume (V ). • Aspect ratio, defined as the ratio between the radius of the sphere circumscribed to the element and the radius of the sphere inscribed in the element. • Minimal dihedral angle. • Size skewness (η), defined as η=

V − Vref Vref

where Vref is the volume of the equilateral element with the same (radius of the) sphere circumscribed as the actual element. Definition. For a given element quality metric fˆ, the mesh quality (fˆmesh ) is the minimum of fˆ over all elements in the mesh, i.e. fˆmesh = mine fˆe .

3.2.

The Mesh Dynamics Strategy

The Computational Mesh Dynamics strategy developed could be classified as a mesh smoothing method. The strategy is based on an optimization problem, where the functional is defined in terms of some appropriate element quality indicator. The mesh topology (element connectivity) is assumed to remain constant, and the nodal coordinates are updated Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

264

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

at each time step minimizing a given functional. In their most general form, the proposed functional is written as F = F (x) = F ({xαj }) (35) where x = {xαj } represents the whole set of nodal coordinates, and xαj is the α coordinate of the node j. The widely used pseudo-elastic CMD strategies can be expressed in that form, being F the functional of elastic energy. At the time step n, the problem to solve is written as min F (xn ) n x

(36)

s. to xn · ntn = 0 on Γtn where Γt is the boundary domain and nt is the outward normal. As it is observed, the nodes on the boundary domain are free to slide in the tangent direction to the boundary. This possibility represents a non-linear restriction for the optimization problem (36) in the general case. Note that the constraint in equation (36) reduces in 1 the number of degree of freedom for each boundary node. Thus, nodes lying on the intersection of two boundaries have two degrees of freedom less, and so on. In the particular case of vertexes, for instance, the number of constraints equalize the number of spatial dimensions (and the number of degree of freedom) and, hence, the node is attached to the vertex. A most simple type of boundary condition consists in fixing the nodal displacement in a predetermined value, but it is more restrictive from the point of view of the optimization problem.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3.2.1.

Functional Design

Being the dynamics of the mesh an artificial field in the problem, there is enough freedom to design the functional in order to obtain meshes having good quality. Some design conditions for F are • F should be computed from element contributions (as in an usual finite element assembly process). • The minimum of F should give the best mesh quality. • F should be well behaved enough in order to solve the minimization problem with Newton-like methods. In general, it will require F to have continuous first derivatives. • F should be convex in order to guarantee unicity of the minimum and positivity of the stiffness matrices (the Hessian matrices of the functional). The first item requires that the functional may be computed as F = g(F1 ({xαj }1 ), F2 ({xαj }2 ), . . .)

(37)

where Fe ({xαj }e ) is the functional for element e, which is a function of the coordinates of its nodes and g(·) is some associative function that preserves convexity, for instance the sum or the maximum g(F1 , F2 , ...) = F1 + F2 + ... (38) g(F1 , F2 , ...) = max(F1 , F2 , ...)

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

265

The second requirement is somewhat in conflict with the third and fourth ones. In the ideal case, Fe would be some indicator of the element badness (or distortion), and g(·) would be the maximum of its arguments, so the minimization of the mesh functional is equivalent to search for the mesh whose badness (i.e. the badness of the worst element) is minimum. However, using the maximum for g(·) it leads to functionals with non-differentiable first derivatives. Therefore functionals of the form X F = |Fe ({xαj }e )|p , (39) e

will be consider, preserving regularity while for p → ∞ the maximum (L∞ ) criterion is recovered. Regarding for the design of the element functional Fe itself, at first sight it should be a function of its deformation only, in order to be invariant under dilatation, translation or rotation. However, the corresponding functional would be non-convex. Consider, for instance, a 1D problem covering the interval [0, 1] with two linear elements, as in figure 3. There are three nodes, for which the position of nodes 1 and 3 are fixed by the boundary

x1=0

x2

x3=1

x

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Invariance under dilatation. conditions and the only unknown variable is the position of node 2. But if the functional is invariant under dilatation, then the functional for all 1D elements would be the same, and the position of the node 2 would be undetermined. In order to regularize the problem a term depending on the volume should be added to the functional. The convexity of the functional is perhaps the most difficult restriction to accomplish. Consider for instance the case of two triangles T = ABC and T ′ = A′ B ′ C ′ in figure 4. Both of them are well oriented (counterclockwise) and they are not too far from an equilateral one, therefore they should have a relatively low functional value (badness). Consider now the linear path that connects both of them, i.e. the family of triangles T (α) that are formed by linear interpolation of the coordinates of T and T ′ . For instance, for the A vertex xA (α) = (1 − α) xA + αxA′

(40)

The triangles for α = 0, 0.25, 0.5, 0.75 and 1 are shown in figure 4. If the functional is convex, it should satisfy the inequality F (T (α)) ≤ (1 − α)F (T ) + αF (T ′ )

(41)

in such a way the badness of the interpolated triangles for 0 < α < 1 should be lower than the extreme ones, i.e. they should be ‘nicer’. But as could be seen in the figure, for α = 0.5 the triangles collapse in a line, so it not seems to be an appropriate criterion of badness that would be convex. Then the convexity requirement on the element functional is drooped. Considering the items discussed above, the following expression for the element functional is proposed m  Ve (42) + Cq qen Fe = Cv e −1 Vref

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

266

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

B’

A

C

C’

α=0

α=1 α=0.5 B

A’

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 4. Convexity preservation. e is the target element volume, q is any elementwhere Ve is the element volume, Vref e wise quality indicator, Cq and Cv are weight coefficients for shape and size terms in the functional respectively. The exponent m must be even and n ∈ Z− in order to set the optimization problem as a minimization one. Then, the element distortion is defined as the reciprocal of the element quality. Normally, Cv should be kept as small as possible but preserving the well-posedness of the problem. Exponents m and n allow to use different discrete norms to measure the element distortion and the element size change, e.g. n = −2 means Euclidean norm, and n → −∞ means maximum norm. This last case is equivalent to maximize the worst element quality. On the other hand, it is possible to see that the element functional chosen is convex in a neighborhood of the equilateral element. Depending on how the system is solved, there are two possibilities: local methods and global methods. The global methods update the nodal position simultaneously for the whole set of nodes, while the local algorithms apply their methodologies over each subset of nodes until the whole set of nodes is updated, i.e., the free nodes are relocated one by one iteratively, keeping the remainder fixed until the convergence is reached. Local methods are to global ones as explicit schemes are to implicit schemes for the resolution of differential equations with preponderant diffusive character. Furthermore, there is no guarantee that a solution of a global strategy may always be reached for a local method. Hence, the optimization problem is solved with a global scheme in order to avoid the drawbacks that local methods present. The way in which the functional was written allows its application for any type of element if the quality indicator is properly defined. Here, it is proposed to use the following geometric quality indicator based on the subdivision of the element in simplexes (triangles in 2D and tetrahedra in 3D) #1/n "N X n (43) q=C (qS,i ) i=1

where C is a normalization constant such that 0 < q ≤ 1, N is the total number of simplexes in the all possible subdivisions of the element in simplicial ones, and qS,i is computed for Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

267

the simplex element i in the subdivision and it is given by qS =

V Se

(44)

P with Se = j ljnd , being lj and V the length of the j−edge and the volume of the simplex, respectively. As can be shown, with an appropriate normalization constant qS is an algebraic quality metric for simplicial elements. Thus, the quality indicator for non-simplicial elements is based on those defined for simplicial ones. Due to this fact, without loss of generality, only the simplex element case is considered in the analysis that follows. The element functional (42) is continuous if qe 6= 0 for all mesh elements, but Fe (x) tends to infinity when qe tends to zero. With the quality metric (44), this last situation could happen when there is at least one element in the mesh with Ve → 0, due to Se is bounded below if the element do not shrinks to a single point. Therefore, the application of this technique is restricted only to valid meshes, since infinite barriers arise when the element volume tends to zero, making it impossible to recover a valid mesh starting from an invalid one. 3.2.2.

Differential Predictor

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The optimization strategy presented in section §3.2. means that at each time step the unknown node positions are obtained by solving the minimization problem (36). The mesh coordinates vector (x) is composed by nodes at the boundary (xb ) and the internal nodes (xint )   xb x= (45) xint A Newton-like strategy is used to solve this optimization problem and, for the sake of clarity, it is assumed that boundary nodes have their displacement prescribed. At each time step, the minimization problem consists in finding the vector x that minimizes the functional F (x). Due to the fact that some components of x (those in xb ) are fixed by the boundary conditions, then  n  xb n (46) xint = argmin F ˆ x int ˆ int x The recurrence formula from the Newton-Raphson strategy is

where

k −1 k xn,k+1 = xn,k int int − (K ) R

(47)

∂F ∂xint ∂R K= ∂xint

(48)

R=

This generates a sequence xn,k int that, if it converges, will give the solution for the optimization problem n lim xn,k (49) int = xint k→∞ Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

268

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

The simplest choice for the initial value xn,0 int is taking the unknown vector at the previous time step, i.e. n−1,∞ xn,0 (50) int = xint However, this has the drawback that, if the elements near the moving boundary are small, then the initial combination [xn−1,∞ , xnb ] may lead to invalid elements, even for small time int steps. In fact, the time step is limited by the element size at the wall, and the limit time step of the moving mesh problem decreases with mesh refinement. To avoid this, a linear predictor for the initial mesh is performed. If the solution xint (t) is considered for each t in the range tn−1 ≤ t ≤ tn , then R(xint (t), xb (t)) = 0 Taking derivatives with respect to time and making an evaluation at t = tn−1     ∂R ∂R n−1 x˙ int (t )+ x˙ b (tn−1 ) = 0 ∂xint tn−1 ∂xb tn−1

(51)

(52)

then the Newton-Raphson sequence can be initialized with the extrapolation

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

n−1,∞ xn,0 + ∆t x˙ int (tn−1 ) int = xint

(53)

where the dot designates time differentiation. For instance, consider a 1D problem with a homogeneous mesh of N linear elements in the interval [0, 1]. The right boundary is fixed and the left boundary moves to the right with velocity 1. With the standard initialization strategy, the limit time step is initially ∆tCMD = h = 1/N , since a larger time step will cause the left boundary to pass over the position of the first internal node (initially at x = h). Besides, with the differential predictor, the limit time step is ∆tCMD = 1, since, in fact, the differential predictor gives the optimal solution, and the subsequent Newton-Raphson iteration is not needed. It has been verified through numerical experiments that with the differential predictor the limiting time step ∆tCMD is independent of the mesh refinement. 3.2.3.

Avoiding the Relaxation of the Initial Mesh

Pseudo-elastic CMD strategies have the property that they do not move the initial mesh unless the domain boundary is deformed. This may be justified by means of elastic energy minimization arguments. This property is not shared by the proposed functional because sometimes the initial mesh introduced by the user is not the optimal mesh with respect to this functional. Consider, for instance, the structured mesh M1 shown in figure 5. The mesh is composed by 200 triangular elements. Even if the mesh has a good quality, the optimization strategy tends to bring each element to a regular (equilateral) shape, then after a relaxation stage the mesh M3 is obtained. In this case, during the relaxation process the nodes on sides AB, CD are fixed, whereas those on BC, AD are left to slide on the horizontal direction. As a consequence of the optimization problem, the elements near vertexes A and C tend to shrink, whereas those near B and D tend to grow. This effect is caused by the particular way in which the squares have been split up into triangles. Note how the elements tend to

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

269

reach the equilateral shape in the relaxed mesh. After the mesh has relaxed, subsequent displacements of the boundary nodes produce displacement of the internal nodes, as described before. This initial ‘relaxation’ stage may or may not be desirable. If the initial mesh has bad quality, then this stage should produce a better new mesh. However, if the initial mesh has some ad-hoc refinement, then it is possible that the relaxation stage will revert this refinement. Consider for instance mesh M2 in figure 5, which has a refinement towards side AB in such a way that the horizontal spacing near CD is 3.5 times larger than the same at AB. As a result of the relaxation process, the relaxed mesh M3 is reached. The resulting relaxed mesh depends only on the topology of the mesh and on the constraints on the boundary nodes, but not on the initial position of the internal nodes. In fact, both meshes M2 and M1 (with and without refinement) produce the same final mesh M3 after relaxation. M1 = initial homogeneous mesh A

D

rel

B

axa

tio

n

M3 = relaxed mesh A

D

B

C

C

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

M2 = initial mesh with refinement A

D

ion xat a l re

B

C

Figure 5. Relaxation of meshes. The functional can be easily modified in order to keep the initial refinement. First, note that for simplicial elements there is a unique linear transformation (x0 , T) that transforms the coordinates {xreg,j } of the regular element (i.e. equilateral triangle in 2D, regular tetrahedron in 3D) to the actual element coordinates {xj } xj = x0 + T xreg,j

(54)

It is easy to see that the functional can be expressed as a function of the transformation Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

270

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti Xreg 111111 000000 111111 000000 regular 111111 000000 element 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 T’

reference element

X

T

T T’ −1

actual 1111111 0000000 1111111 0000000 element 1111111 0000000 1111111 0000000 1111111 0000000 1111111 0000000 1111111 0000000 1111111 0000000 1111111 0000000 1111111 0000000

111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 111111 000000 X ref

Figure 6. Compensation for initial deformation in reference mesh. matrix T

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Fe = g(T)

(55)

This fact can be seen because the functional can be computed by taking the nodal coordinates of the regular elements, applying the transformation and finally computing the edge lengths, volume, and the functional. All these computations are encapsulated in the function g( · ). Of course, the functional does not depend on the translation x0 . In fact, it does only depend on the metric of the transformation TT T, because it is independent of rotations. However, for the analysis that follows, it is needed to accept that it only depends on the transformation matrix, as reflected in (55). By construction, g has a minimum when T = c O, with c ∈ R a scaling factor and O an orthogonal matrix, since in this case the actual element is similar to the regular one. The purpose is to modify the functional in order to the optimal element shape for Fe is not longer the regular element shape but it could be the shape of some reference element with coordinates {xref,j } (see figure 6). It is easy to see that this can be done by considering the transformation from the reference element to the actual element as Fe = g(T T′

−1

)

(56)

where T′ transforms the regular element to the reference element. For instance, as mentioned above, a minimum is reached when T T′ −1 = c O, i.e. when the current element is similar in shape to the reference one. Note that this modification can be simply introduced by computing the transformations T, T′ and then computing the functional with the coordinates x′j = T T′ −1 xref,j . An example can be seen in figure 7. The original mesh on the right has a refinement ratio of 1:10 near the AB side. Then, it is deformed on the side AB with a ramp of amplitude 0.2 (resulting in the mesh shown on the left of that figure). Note that if no initial relaxation is produced, the final mesh still has the refinement towards the AB side. Computations of the analytical jacobians are also straightforward. The jacobians with respect to x′j are computed in the standard way, and then they are composed with the jaco-

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... initial mesh with refinement

271

deformed mesh

A

D

B

C

A

B

D

C

Figure 7. Mesh deformation with surface refinement. bian

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3.2.4.

∂x′j −1 = T T′ T−1 ∂xj

(57)

Results

The proposed CMD strategy was applied to several test problems, in 2D and 3D. A test is considered successful if it achieves a valid mesh at every time step. Values for Cq = 1, Cv = 0 and n = −1 were adopted in the tests solved. The boundary was moved in such a way to produce an initial valid mesh for the nonlinear solver at each time step. The imposed law of movement for a boundary moving node i was of the form φi (t) = ai φ(t), where ai is the spatial amplitude and   t φ(t) = tanh (58) τ˜ is the temporal part of the dependency, τ˜ being a coefficient that represents the duration of the change from the starting value to the end value and t ∈ (0, 1). Regarding for the nonlinear system resolution, the computational cost depends on the percentage of the domain deformation. Using a tolerance of 1 × 10−5 in the residue, for small deformation domain it is enough with one or two Newton iterations, while for harder problems, between five and seven Newton iterations were employed to guarantee the residual convergence. Step 2D Figure 8 sketches the domain and the deformation sequence for this problem. The mesh used has 200 triangular elements and 121 nodes, and the time step adopted was ∆t = 0.005. On the nodes of the boundary domain the displacements have been imposed, being null for those nodes on fixed boundaries. In figure 9 several meshes obtained during the deformation sequence are shown. In figure 10 the minimum and mean values of the element quality indicator q as a function of the mesh deformation is plotted. As can be noted, the application of the strategy to this problem makes it possible to reach deformations larger than 99 %. Of course, the meshes obtained with the highest percentages of relative deformation may not be useful

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

272

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

x2 C t B A O

x1 Figure 8. Domain of step 2D test.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

for a computation. Such deformations are reached in order to show the robustness of the method.

Figure 9. Mesh deformations for step 2D test: 50 % (left), 90 % (center) and 99 % (right).

Step 3D This test is the 3D extension of the test presented in the last section. Figure 11 shows the domain for different deformations. The top face of the cube is moved in vertical direction and, during the deformation, this face is transformed into two planes of different heights joined by a truncated cone with upper radius r1 and lower radius r2 . The mesh used has 1080 elements and 343 nodes. As boundary conditions, imposed displacements were used. Only the component of displacement in direction x3 is nonzero for nodes on moving boundaries. Figure 12 shows the surface mesh at three different time steps allowing to see how the moving mesh is transformed. The mean and minimum values of element quality indicator q as a function of mesh deformation is shown in figure 13.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

273

1

0.8

quality

mean 0.6

0.4 minimum 0.2

0

0

20

40 60 mesh deformation [%]

80

100

Figure 10. Step 2D minimum and mean element quality as a function of mesh deformation.

C

x3

r1 B

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A

x2 F

t

C’

r2

O

B’

A’

E

D

x1 Figure 11. Problem definition for step 3D test.

3.3.

Simultaneous Mesh Untangling and Smoothing

The CMD strategy presented requires valid meshes at the begin of each time step, as discussed in section §3.2.. In FSI problems, this limitation is sometimes by-passed decreasing the time step size and, thus, avoiding the tangling of the grid caused by the motion of Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

274

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Figure 12. Mesh deformations for step 3D test: 50 % (left), 80 % (center) and 87 % (right).

1

0.8

quality

mean 0.6

0.4 minimum

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

0.2

0

0

10

20

30 40 50 60 mesh deformation [%]

70

80

90

Figure 13. Step 3D minimum and mean element quality as a function of mesh deformation.

the boundary. However, the computational cost suffers large increments, specially if some clustering of nodes is used near the moving boundary to capture fluid dynamics details like boundary layers. When the mesh turns invalid, an untangling methodology should be used to recover a valid mesh and then to be able to apply the smoothing technique. Untangling methods are commonly based on the element volume [30] and, in general, both procedures smoothing and untangling, are treated separately. Thinking of FSI problems, the CMD method should have the capability of solving the mesh motion even though inverted elements were found, guaranteeing a smooth mesh at each time step. Therefore, a simultaneous procedure of smoothing and untangling is preferable [19, 14, 17, 39]. It is in this sense that the CMD strategy, enhanced with simultaneous untangling and smoothing, is useful, providing a way to recover a valid mesh, if it exists, despite starting from an invalid one.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... 3.3.1.

275

Functional Regularization

In order to circumvent the drawbacks produced by the singularities of the proposed functional (42) when the element volume tends to zero, it was modified the element quality indicator following an idea proposed by Escobar et al. [17]. The modification consists in replacing the volume in equation (44) by the function p 1 h(V ) = (V + V 2 + 4δ 2 ) (59) 2 This is a strictly increasing function of the volume and it is also a positive function for all V (see figure 14). The parameter δ represents the value of the function for a null volume.

h(V)

δ

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

O

V

Figure 14. Function h(V ). The modified functional is written as  m Ve Fe∗ (x) = Cv + Cq qe∗n − 1 e Vref

(60)

where, for simplicial elements, Ch(V ) (61) Se Due to V > 0 ∀V , the regularized functional is continuous in the whole space of nodal coordinates. The dependence of h(V ) with the parameter δ is such that  V if V ≥ 0 lim h(V ) = 0 if V < 0 δ→0 q∗ =

Therefore, when the parameter δ tends to zero, the modified functional tends to the original one (for V ≥ 0), and also, the modified optimal solution tends to the original one. Particularly, in the region of valid meshes, as δ → 0, the function F ∗ (x) converges pointwise to F (x). Besides, by considering that ∀V > 0, lim h′ (V ) = 1 δ→0

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(62)

276

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

and lim h(r) (V ) = 0,

for r ≥ 2

(63)

δ→0

it is easy to show that the derivatives of the modified functional verify the same property of convergence. The purpose is to find a solution close enough to the optimal solution of the original functional, assuming that a valid mesh exists for the given topology and boundary position. Thus, it should be defined a decreasing sequence {δ k } such that δ k → 0 as k → ∞. Then, a simultaneous mesh untangling and smoothing strategy is obtained. This strategy could be used as a CMD technique with the property of no conditioning the time step in FSI problems. 3.3.2.

Solution Strategy

According to numerical examples, the lower the parameter δ, the slower the convergence rate of the optimization algorithm, without guarantee of final convergence. Moreover, if δ is not small enough the ‘optimal’ mesh finally obtained may be invalid, being even worse for high relative domain deformations. Therefore, two main problems arise • finding the decreasing sequence {δ k } to ensure the final convergence to a valid mesh, if exists.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• finding a suitable initial value for the sequence (δ 0 ). To determine an equation that allows the decreasing of δ, this parameter is assumed as a new global variable only for theoretical purposes, i.e., considering F ∗ = F ∗ (x, δ). If the problem is posed in terms of the variables (x, δ) and using a Newton-like solver, the problem is written as follows     2 ∗ ∂2F ∗ " ∂F ∗ ∂ F #    ∂x2 ∂x∂δ  ∆x = −  ∂x∗    2 ∗ ∂F ∂2F ∗ ∂ F ∆δ ∂δ ∂δ∂x ∂δ 2 Writing the last equation in the following way ∂2F ∗ ∆x + ∂x2 ∂2F ∗ ∆x + ∂δ∂x

∂2F ∗ ∂F ∗ ∆δ = − ∂x∂δ ∂x ∂2F ∗ ∂F ∗ ∆δ = − ∂δ 2 ∂δ

(64)

it is observed that this system may be solved in an uncoupled way if the parameter δ is kept fixed for the first equation (∆δ = 0). This is equivalent to solve the system (64) using the block Gauss-Seidel method. Thus, the variable increments ∆x and ∆δ are obtained  2 ∗ −1 ∂ F ∂F ∗ ∆x = − ∂x2 ∂x   ∗ 2 ∂ F∗ ∂F (65) + ∆x ∂δ ∂δ∂x ∆δ = − ∂2F ∗ ∂δ 2

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

277

The expression for ∆δ in (65) is adopted as the maximum value to reduce δ. Therefore, the updated δ in the iteration k is defined in the following way ˜ k−1 ) δ k = max(δ k−1 − α ˜ |∆δ k−1 |, βδ

(66)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

where α ˜ and β˜ are constants lower than one. In addition, it was found that the off-diagonal terms in the element-wise matrix have a strong influence on the convergence of this optimization method. In the untangling stage, it is advisable to relax these off-diagonal terms to make the matrix more diagonal-dominant. However, in the smoothing stage these terms should be restored to take advantage of the convergence rate of full Newton schemes. Here, the relaxation parameter for these offdiagonal terms (˜ γ ≤ 1) may be constant or a function of the iterations. For example, for a 2D case the element-wise matrix is modified as  2 ∗     2 ∗ ∂ F ∂ F γ˜ 2  ∂x1 ∂x2 e  ∂x1 e      Ke =  2 ∗   ∂ F ∂2F ∗ γ˜ 2 ∂x2 ∂x1 e ∂x2 e

The problem is solved by the Newton-Raphson method with Armijo inexact line search [45]. At each iteration δ is diminished only if the line search strategy gives a unit step as result. If the mesh is initially invalid, the initial value δ 0 is chosen according to the following criterion based on the minimum volume (Vmin = mine Ve ). Due to the fact that h(V ) is a strictly increasing function, then   q 1 2 2 Vmin + Vmin + 4δ hmin = h(Vmin ) = (67) 2 Defining h∗min = hmin /δ as an user-defined parameter and getting δ from the last equation, the following criterion to initialize δ arises  ∗  hmin Vmin + ǫ if V ≤ ǫ δ δ 0 δ = h∗ 2 − 1  min 0 if V > ǫδ

where ǫδ > 0 is the minimum value given to the initial value of δ such that δ 0 > 0 when Vmin = 0.

3.3.3.

Results

In this section, the numerical results for some test examples are presented. These examples are CMD problems in 2D and 3D with different deformations of the boundary. In the whole set of test cases, the following convergence criteria had been applied • Valid mesh. • For the iteration k,

k−1 k |qmin − qmin | < ǫq , being qmin = mine qe and ǫq > 0 a prefixed k qmin

tolerance. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

278

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

The relaxation coefficient for the Hessian matrix was chosen as γ˜ = 0.5 for the untangling stage, and γ˜ = 1 for the smoothing stage. In the whole set of numerical examples, the following set of parameter values were used: Cv = 0, Cq = 1, n = −1, α ˜ = 1, β˜ = 0.1, ǫq = 0.01, h∗min = 0.75 and ǫδ = 1 × 10−6 . In these tests, the reference element used was the regular element.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Step 2D This test was defined in section §3.2.4. for a relative domain deformation of 50 %, 90 % and 99 %, and here is repeated using the simultaneous untangling and smoothing strategy. The total mesh deformation is carried out in one time step in order to show the robustness of the proposed algorithm. For each case, the initial tangled mesh and the resulting valid grid are presented in figures 15 to 17. In addition, the evolution of the mesh quality with iterations is included in figure 18.

Figure 15. Mesh deformation of 50 % for step 2D test, initial (left) and final (right) meshes.

Figure 16. Mesh deformation of 90 % for step 2D test, initial (left) and final (right) meshes. Figure 19 shows the comparison between the smoothing (S) and the simultaneous untangling and smoothing (U-S) strategies. It is observed that the elapsed time in the computation with the U-S technique is approximately independent of the deformation. It does

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

279

Figure 17. Mesh deformation of 99 % for step 2D test, initial (left) and final (right) meshes. 1 50 % def. 90 % def. 99 % def.

q mesh

0.5

0

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

−0.5

−1

0

2

4

6

8 10 iteration number

12

14

16

18

Figure 18. Mesh quality as a function of iterations for step 2D test. depend on the number of time steps used, and tends towards the elapsed time of the S method as the number of time steps are increased. In order to verify the utility of the differential predictor (DP) explained in section §3.2.2., the test was solved for a relative domain deformation of 90 % varying the amount of time steps used. In table 1 the total number of iterations in each case (with and without DP) is presented. As it is observed, the use of the DP makes it possible to decrease in a noticeable way the number of iterations (and therefore the cost), mainly when the time step used diminishes. Step 3D As presented in section §3.2.4., this test is the 3D extension of the step 2D test. The test was solved for 50 %, 80 % and 87 % of relative domain deformation, and using one time step for each case. Figures 20 to 22 show the initial and final meshes obtained. Figure 23 shows the mesh quality as a function of iterations. For 87 % of mesh defor-

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

280

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti 60 S − 200 time steps elapsed comp. time [sec]

50

U−S − 1 time step U−S − 2 time steps U−S − 4 time steps

40

U−S − 10 time steps 30

20

10

0

0

10

20

30

40 50 60 mesh deformation [%]

70

80

90

100

Figure 19. Computational time in terms of the relative domain deformation for step 2D test. Table 1. Total number of iterations to reach a relative domain deformation of 90 % for step 2D test.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

number of time steps used 1 2 4 8

iterations with DP 15 19 34 39

iterations without DP 13 27 47 69

Figure 20. Mesh deformation of 50 % for step 3D test, initial (left) and final (right) meshes. mation, the number of iterations is high since the boundary position is close to the limit in which a valid mesh exists for the conditions of the test. A comparison of computational cost

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

281

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 21. Mesh deformation of 80 % for step 3D test, initial (left) and final (right) meshes.

Figure 22. Mesh deformation of 87 % for step 3D test, initial (left) and final (right) meshes. with mesh deformation between S and U-S strategies was done, and the results achieved are shown in figure 24. Again, as in the previous test, the advantage of applying the DP can be observed in table 2. In this case, the problem was solved for a relative domain deformation of 70 % varying the size of the time step. Axisymmetrical flowmeter Finally, the methodology is applied to IC engine simulations. One of the most important device to measure the mass flow rate of different cylinder heads is the flowmeter bench. This example consists in solving the mesh dynamics of an axisymmetrical flowmeter, whose geometry is shown in figure 25. The head of the valve moves as a rigid solid and its stem stretches and shortens depending on the displacement of the valve. The other boundary that stretches and shortens is the axis of symmetry. Initially, the valve lift is 5 mm, and the valve is moved with a linear law in the x2 -direction until reaching the minimum valve lift of 0.5 mm. The mesh has 12K triangular elements and 6.4K nodes, with h ≃ 0.2 mm in the region

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

282

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti 1 50 % def. 80 % def. 87 % def.

q mesh

0.5

0

−0.5

−1

0

5

10 iteration number

15

20

Figure 23. Mesh quality as a function of iterations for step 3D test. 140 S − 45 time steps

elapsed comp. time [sec]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

120

U−S − 1 time step U−S − 2 time steps

100

U−S − 4 time steps U−S − 10 time steps

80 60 40 20 0

0

10

20

30 40 50 60 mesh deformation [%]

70

80

90

Figure 24. Computational time in terms of the relative domain deformation for step 3D test.

between the valve and its seat. Nodes lying on the vertical walls of the cylinder and those lying on the valve stem are left to slide in the x2 -direction, while the remaining boundary nodes have their displacements prescribed. In figure 26(a) it is shown a close-up of the mesh utilized. Figure 26(b) presents the mesh from which the CMD strategy starts, and where invalid elements were filled in green. The final valid mesh is shown in figure 26(c). In this case, the total domain deformation was applied in one time step. In table 3 it is summarized the elapsed computational time for obtain a valid mesh for the minimum valve lift varying the number of time steps. The problem was solved with and without the differential predictor for comparison purposes. The coefficients used for the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

283

Table 2. Total number of iterations to reach a relative domain deformation of 70 % for step 3D test. number of time steps used 1 2 4 8

iterations with DP 8 9 13 20

iterations without DP 7 12 24 48

U-S technique were chosen in order to minimize the total number of iterations when one time step is utilized. As it is observed in the table, when the differential predictor is applied the elapsed computational time remains approximately unchanged. Table 3. Total elapsed computational time for the axisymmetrical flowmeter test.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

number of time steps 1 2 5 10

elapsed time without DP 39.486 s. 113.897 s. 121.616 s. 204.734 s.

elapsed time with DP 77.721 s. 79.217 s. 82.272 s. 78.549 s.

Diesel engine with three valves The last case presented is the mesh dynamics for the 3D geometry of a diesel engine. The engine has three valves and bowl-in piston. The purpose is to solve the movement of a unique mesh along the whole cycle in order to show how the CMD strategy works with a real IC engine geometry. Hence, the valve closure is approximated with a small enough minimum valve lift greater than zero. Some views of the chamber geometry are included in figure 27. The cylinder bore is 93.0 mm, the stroke is 103.0 mm, and the geometric compression ratio is 17.8:1. The maximum intake valve lift is 9.0 mm and the maximum exhaust valve lift is 11.0 mm. The valve timing values are the following • Intake Valves Opening (IVO): 5◦ • Intake Valves Closing (IVC): 212◦ • Exhaust Valve Opening (EVO): 485◦ • Exhaust Valve Closing (EVC): 715◦ The (topology of the) mesh was generated with the valves placed at a half of the maximum lift from their seats and the crown piston at a half of the stroke from the head. The mesh is coarse in the cylinder, but refined in the region of valves, ducts and piston bowl. The mesh has 925700 tetrahedrons and 157630 nodes. The mesh dynamics was solved with the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

284

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti x2

moving valve

h

rv

H R

O

x1

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 25. Axisymmetrical flowmeter. simultaneous mesh untangling and smoothing technique and using the differential predictor presented in section §3.2.2.. The time step used corresponds to 1 crank angle degree. Due to the initial mesh was generated with a configuration of the boundary that not corresponds to any instant in the cycle, firstly the valves and piston are moved to the reference crank angle. In this case, such angle was adopted as the TDC when the intake stroke starts. The mesh quality as a function of the crank angle is plotted in figure 28, where the element quality metric applied was q (see equation (43)). Figure 29 shows the minimum mesh dihedral angle as a function of the crank angle. As could be observed, the mesh quality is approximately constant along the whole cycle. In order to compare, the initial mesh has a quality q = 0.0211 and a minimum dihedral angle of 1.4812◦ . Figures 30 to 32 show the distribution of the element quality in the domain for three crank angles. These angles correspond to the TDC with the three valves closed (0◦ ), the maximum intake valve lift (108.5◦ ), and maximum exhaust valve lift (600◦ ). The elements with worse quality in the mesh are located between a valve and its seat when the valve approaches the closed position.

4.

Resolution of Compressible Flows in the Low Mach Number Limit

Low Mach number (M ) flows represent a limit situation in the solution of compressible flows. When the Mach number approaches to zero, the strategies based on density to Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

(a) Initial position

(b) 100 % def. - Initial mesh

285

(c) 100 % def. - Final mesh

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 26. Mesh zoom for the axisymmetrical flowmeter.

Figure 27. Geometry of the combustion chamber of a diesel engine. solve the flow equations suffer severe deficiencies, both in efficiency and accuracy. Turkel et al. [64] and Guillard and Viozat [24] have identified that, in the low Mach number limit, the discretized solution of the compressible fluid flow equations may fail to provide an accurate approximation to the incompressible equations. In the subsonic regime, when the magnitude of the flow velocity is small in comparison with the acoustic wave speed, dominance of convective terms within the time-dependent equation system renders the system stiff and the solvers could converge slowly [11]. Time-marching procedures may suffer severe stability and accuracy restrictions and become inefficient for low Mach number flow regimes. Here, for explicit schemes, the time step must satisfy the Courant-Friedrichs-Levy (CFL) conditions, where numerical stability considerations lead to small time steps. On

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

286

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti IVO

BDC IVC

TDC

EVO BDC

EVC

qmesh

0.1

0.01

0.001

0

100

200

300 400 crank angle [deg]

500

600

700

Figure 28. Mesh quality as a function of crank angle for the diesel engine. IVO

BDC IVC

TDC

EVO BDC

EVC

minimum dihedral angle [deg]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1

0.8

0.6

0.4

0.2

0

0

100

200

300 400 crank angle [deg]

500

600

700

Figure 29. Minimum mesh dihedral angle as a function of crank angle for the diesel engine.

the other hand, implicit methods suffer from stiffness due to large disparity in the eigenvalues of the system. The condition number of the system of equations is O(1/M ) in the low Mach number limit [11]. There are two main approaches to circumvent this drawback: firstly, the modification of compressible solvers (density-based) downward to low Mach numbers [11, 65, 66, 35, 68]; secondly, extending incompressible solvers (pressure-based) towards this regime [26]. Also there are unified formulations, as the method proposed by Mittal and Tezduyar [38].

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

287

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 30. Mesh quality field at 0 crank angle degree (TDC) for the diesel engine.

Figure 31. Mesh quality field at 108.5 crank angle degree (maximum intake valve lift) for the diesel engine. The in-cylinder flow in an IC engine is characterized by low Mach numbers, except in the exhaust blowdown phase [25]. Thus, this application limits the strategies for low Mach number flows to whose, under an unified formulation, work correctly for (at least) all

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

288

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 32. Mesh quality field at 600 crank angle degree (maximum exhaust valve lift) for the diesel engine. subsonic Mach numbers. For density-based methods, two distinct techniques have been proposed to capture solution convergence for low-Mach number regimes: preconditioning and asymptotic methods. Both techniques achieve rescaling of the system condition number. The asymptotic method introduces a perturbed form of the equations discarding specific terms, so that the physical acoustic waves are replaced by pseudo-acoustic modes. The magnitude of the propagation speeds of this pseudo-acoustic modes is similar to the fluid velocity [11]. Although perturbation procedures are highly robust and applicable for both viscous and inviscid flows, the nature of the perturbation limits their usage, particularly with respect to mixed compressible-incompressible flows. Preconditioning schemes consist in premultiplying time derivatives by a suitable preconditioning matrix. This scales the eigenvalues of the system to similar orders of magnitude and removes the disparity in wave-speeds, leading to a well-conditioned system [64]. The modified equations have only steady-state solutions in common with the original system (hence, are devoid of true transients). For the application of these methods to unsteady problems, the ‘dual-time-stepping’ technique has emerged, where the physical time derivative terms are treated as source and/or reactive terms. During each physical time step, the system of pseudo-temporal equations is advanced in artificial time to reach a pseudosteady-state [65, 66, 35]. Several local preconditioning matrices have been designed for

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

289

steady state problems, with very good results [11, 41]. However, recent works have shown that these preconditioners may not be appropriate for unsteady flows [66]. In IC engines applications, one procedure widely used is a semi-implicit method named acoustic subcycling [25]. In this method, all terms in the governing equations that are not associated with sound waves are explicitly advanced with a larger time step ∆t similar to the used with implicit methods. The terms associated with acoustic waves (the compression terms in the continuity and energy equations and the pressure gradient in the momentum equation) are explicitly advanced using a smaller time step δt that satisfies the CFL stability criterion, and of which the main time step is an integral multiple. While this method works well in many internal combustion engines applications where the Mach number is not too low, it is unsuitable for very low Mach number flows since the number of subcycles (∆t/δt) tends to infinity as the Mach number tends to zero. Pressure gradient scaling can be used to extend the method to lower Mach numbers [25]. The Mach number is artificially increased to a larger value (but still small in an absolute sense) by multiplying the pressure gradient in the momentum equation by a time-dependent scaling factor 1/α2 (t), where α(t) > 1. This reduces the effective sound speed by the factor α. Coupling pressure gradient scaling with acoustic subcycling reduces the number of subcycles by α [4]. We propose to use the method of preconditioning due to its ability to work in an wide range of Mach and Reynolds numbers [43]. The preconditioning matrix applied is the proposed by Choi and Merkle [11] to solve steady compressible flows and which has been adapted by Nigro et al. [40, 41] to the finite element method. Via an eigenvalue analysis of the system of equations, some parameters involved in this matrix are redefined in order to solve transient problems.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4.1.

Problem Definition and Eigenvalues Analysis

The eigenvalue analysis of the Navier-Stokes equations for compressible flows is done by using viscous variables, defined as Q = [p, u, T ]T

(68)

In the low Mach number limit, it is convenient to do the analysis using the viscous variables instead the conservative ones because, at a fixed time, the density tends to be constant in the space domain. The preconditioning of equations consists in premultiply the time derivatives by a properly defined matrix. The purpose is to modify the eigenvalues of the system of equations in order to decrease the condition number. Due to this fact, it is only applicable to steady state simulations. In order to apply the preconditioning strategy in unsteady problems, the dual time technique has emerged [34]. In this technique, two times must be considered: the physical time (t) and the pseudo-time (τ ). The solution is obtained by means of preconditioned fully implicit pseudo-transient iterations adding a pseudo-time derivative to equation (11). At each physical time step, the system is solved until a pseudo-steady state is reached when τ → ∞. If Γ denotes the preconditioning matrix, the system of equations modified by the dual time strategy is written as [65]   ∂U ∂U ∂U ∂U ∂ Γ + + Ai Kij +S (69) = ∂τ ∂t ∂xi ∂xi ∂xj

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

290

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

∂U When the system approaches to the pseudo-steady state, the term tends to zero. Thus, ∂τ the solution of equation (69) at fixed time t tends to the corresponding solution of equation (11). In the viscous variables basis, equation (69) is expressed as   ∂U ∂Q ∂ ∂U ∂Q ∂U ∂Q ∂U ∂Q + + Ai = Kij +S (70) Γ ∂Q ∂τ ∂Q ∂t ∂Q ∂xi ∂xi ∂Q ∂xj Let Γv = Γ

∂U ∂Q

(71)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

the preconditioning matrix in the viscous variables basis. Hence, after premultiplying equation (70) by the inverse matrix Γ−1 v , the following expression is obtained   ∂Q ∂U ∂Q ∂U ∂Q −1 ∂U ∂Q −1 −1 ∂ + Γv + Γv Ai = Γv Kij + Γ−1 (72) v S ∂τ ∂Q ∂t ∂Q ∂xi ∂xi ∂Q ∂xj The analysis is done by using the preconditioning matrix proposed by Choi and Merkle [11] for the resolution of steady state problems at the low Mach limit. The preconditioning matrix takes the form   1 0 0 0 0   βMr2   u1   ρ 0 0 0   βMr2   u2    0 ρ 0 0  Γv =  (73) 2  βM r   u 3  0 0 ρ 0    2 βM   r  ρe + p γρR  − δ ρu ρu ρu 1 2 3 ρβMr2 γ−1

where Mr is a reference Mach number, δ is a constant that plays the role of a coefficient of the time derivative of pressure, and β = zc2 , being z = max(zinv , zvis ), zinv = Mr2 αvis (αvis − 1) αvis − 1 + c2 /(u · s)2 CF L = σReh

zvis = αvis

In the last equation, σ is the Fourier number, Reh is the cell Reynolds number based on the characteristic element length h, and s is the unit vector aligned with the flow velocity. The reference Mach number Mr replaces the Mach number because the term 1/M 2 in the preconditioning matrix becomes singular in regions where the Mach number tends to zero (e.g., stagnation points). In the next subsection a discussion about the election of Mr is presented (see equation (86) for the definition given by Choi and Merkle [11]).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

291

The preconditioned equations are nearly identical to the equations obtained when the method of artificial compressibility is applied, with the addition of the energy conservation equation [11]. Considering null the source vector, the individual equations are 1 ∂p + ∇ · (ρu) = 0 βMr2 ∂t   ∂u + u · ∇u + ∇p = ∇ · T ρ ∂t   ∂T ∂p ρcp + u · ∇T = δ + u · ∇p + ∇ · (κ∇T ) + ∇ · (Tu) ∂t ∂t

(74)

In order to study the eigenvalues of the system of equations, a dispersion analysis on the equation (72) is done. Let ˜ v, i = Γ−1 Ai ∂U A v ∂Q ˜ v, ij = Γ−1 Kij ∂U K v ∂Q −1 ˜ Sv = Γv S

(75)

The discretization of physical time derivative is done using backward differences

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

∂Q ct Qn+1 − E(Qn , Qn−1 , . . .) ≈ ∂t ∆t

(76)

where ct is a constant that depends on the temporal order of the scheme. Then, the system of equations (72) can be written as   ∂Q ct −1 ∂U ˜ v, i ∂Q = ∂ ˜ v, ij ∂Q +S ˜ v + 1 Γ−1 ∂U E(Qn , Qn−1 , . . .) + Γv Q+A K ∂τ ∆t ∂Q ∂xi ∂xi ∂xj ∆t v ∂Q (77) In order to simplify the notation, the index n + 1 for the variables evaluated at the current time is dropped from equation (77). If the source term is assumed to have no effect on the dispersion equation and neglecting the diffusive terms (Euler equations), the following equation is reached ct −1 ∂U ∂Q ˜ v, i ∂Q = 0 + Γv Q+A ∂τ ∆t ∂Q ∂xi

(78)

By introducing a Fourier mode   Q = Q0 exp i(kT x − ωτ )

into the equation (78), the following equation of dispersion for ω is obtained   ct −1 ∂U ˜ v, i Q = 0 Γ + iki A −iωI + ∆t v ∂Q

(79)

(80)

Due to the finite number of mesh nodes, the wavelengths are limited by the grid spacing. To take this filtering into account, it could set kkk = φ/h, where h is a measure of the grid Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

292

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

spacing, and φ ∈ [0, π]. Let λ = ω/kkk the wave speed and uk = uT k/kkk. Writing equation (80) as a system of equations GQ = 0, and looking for solutions for Q 6= 0 achieves G as ct ki ˜ ∂U G = −i Γ−1 + Av, i − λI (81) v kkk∆t ∂Q kkk after dividing by ikkk. The equation GQ = 0 have a non-trivial solution if det G = 0. This condition is equivalent to compute the eigenvalues of the matrix ˆ = −i G Let the CFL numbers

ct ∂U ki ˜ Γ−1 + Av, i v kkk∆t ∂Q kkk

(82)

uk ∆t h c∆t CF Lc = h

(83)

ˆ 1,2,3 = uk (1 − ict CF L−1 ) λ(G) u uk ˆ (1 − ict CF L−1 λ(G)4,5 = u )T± 2

(84)

CF Lu =

ˆ are Then, the eigenvalues of G

where

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

T± = (1 +

Mr2 χ)

±

s

(1 −

Mr2 χ)2



4Mr2



 1 +1−χ 2 (iM + ct CF L−1 c )

(85)

In equation (85), χ = γ − (γ − 1)δ and the definition of Mr allows to choose among different preconditioning matrices. The eigenvalues of the preconditioned system using the preconditioning matrix designed for steady state solutions can be obtained by means of the definition of Mr proposed by Choi and Merkle [11]    Mǫ if M < Mǫ M if Mǫ ≤ M < 1 Mr = (86)   1 if M ≥ 1

or, equivalently, Mr = min(1, max(M, Mǫ )), where Mǫ is a cut-off of the Mach number in a neighborhood of stagnation points. When the ALE strategy is applied (see section §2.1.3.), the eigenvalues for the preconditioned system of Euler equations using the dual-time formulation are written as ˆ ALE )1,2,3 = uk (1 − ict CF L−1 ) − wk λ(G u ˆ ALE )4,5 = 1 [uk (1 − ict CF L−1 ) − wk ]T ALE λ(G u ± 2

(87)

being T±ALE

= (1 +

Mr2 χ)

±

s

(1 −

Mr2 χ)2



4Mr2



 1 +1−χ , 2 ˜ + ct CF L−1 (iM c )

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(88)

In Search of Improvements for the Computational Simulation...

293

˜ = (uk − wk )/c, and wk = wT k/kkk. M For viscous flows, it is not longer easy to compute the eigenvalues of the system. Therefore, many authors use approximations that depend on the Reynolds number. In the following subsection, a numerical analysis of the viscous flow case is presented. 4.1.1.

Preconditioning Strategies

The condition number of the system is defined as CN =

max(|λi |) max(1, |T+ |, |T− |) = min(|λi |) min(1, |T+ |, |T− |)

(89)

Figure 33 shows the condition number of the system of Euler equations as a function of Courant number CF Lc , where ct = 1, δ = 1 (χ = 1), Mǫ = 1 × 10−6 and M = 1 × 10−3 . The preconditioning matrix for steady state solutions (SP, Steady Preconditioner) corresponds to the definition of Mr given by equation (86). For unsteady problems, Vigneron et al. [66] suggest q Mr = min(1, max(

M 2 + CF L−2 c , Mǫ ))

(90)

This definition is named UP (Unsteady Preconditioner) on figure 33. In addition, the condition number of the non-preconditioned system (NP) is included, which is obtained when Mr = χ−1/2 in equation (85). 1e+06 UP SP NP

condition number

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

100000 10000 1000 100 10 1 0.1 0.1

1

10

100

1000

10000

CFLc

Figure 33. Condition number as a function of Courant number CF Lc with M = 1 × 10−3 and Mǫ = 1 × 10−6 for the inviscid case. As it is shown in the figure, when the ‘unsteady’ preconditioning is applied the condition number of the system is O(1) for all CF Lc numbers in the inviscid case. For viscous flows, the eigenvalues of the system were computed numerically adopting z = zinv for the definition of β in equation (73). Figure 34 shows the condition number of the system for several Reynolds numbers using the inviscid reference Mach number (Mr |inv ) defined in equation (90). For Reynolds numbers lower than 1 the condition number of the system

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

294

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti 100000

condition number

10000

inviscid Re = 1 Re = 0.1 Re = 0.01

1000

100

10

1 0.1

1

10

100

1000

10000

CFLc

Figure 34. Condition number as a function of Courant number CF Lc for several Reynolds numbers with the reference Mach number for the inviscid case and M = 1 × 10−3 .

increase significantly, specially for high CF Lc numbers. Merkle [37] proposed to use the following approximation

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Mr =

(

Mr |inv

if

Re > 1

M/Re

if

Re < 1

(91)

In figure 35 the condition number of the system as a function of Reynolds number for CF Lc = 1 × 104 , 1 × 102 is plotted. Two pairs of curves are shown in the figure, one of them corresponds to the reference Mach number for inviscid flows and the other one was obtained by using the approximation (91). The approximation could not be good enough for low CF Lc numbers, as it is shown in figure 35. We propose to use no correction for the viscous flow case since the original definition of the preconditioning matrix includes some control of the time step in the viscous regions via the β parameter.

4.2. 4.2.1.

Numerical Implementation Variational Formulation

In order to simplify the notation, the mesh velocity is considered null since when the ALE technique is used, only the advective jacobians are modified in the system of equations (see equation (25)). The variational formulation of the problem is written as follows: Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

295

1e+07 CFLc = 1e4

condition number

1e+06

CFLc = 1e2

100000 10000 Mr inviscid 1000 100 10 Mr corrected 1 0.001 0.01

0.1 Re

1

10

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 35. Condition number as a function of Reynolds number with CF Lc = 1×104 , 1× 102 and M = 1 × 10−3 . Find Uh ∈ S h such that ∀ Wh ∈ V h   Z Z h h ∂Uh ∂Wh ∂Uh h h ∂U h ∂U W · Γ + + Ai · Khij dΩ dΩ + ∂τ ∂t ∂xi ∂xj Ω Ω ∂xi     nel Z h h h h X ∂Uh ∂ ′ h T ∂W h ∂U h ∂U h ∂U + τ (Ak ) + + Ai − · Γ Kij − S dΩe ∂x ∂τ ∂t ∂x ∂x ∂x e i i j k e=1 Ω Z Z nel Z h ∂Uh X ′ ∂W · dΩe = Wh · SdΩ + Wh · f d∂Ω + δsc ∂x ∂x h e i i Ω Γ e=1 Ω (92) ′ indicate where spaces S h and V h are defined by equation (27). The primes in τ ′ and δsc that a redefinition of these stabilization parameters must be done due to the preconditioning. The derivative with respect to τ is discretized using the backward Euler difference scheme ∂U Un+1,m+1 − Un+1,m ≈ (93) ∂τ ∆τ Notice the indexes used to indicate each time level: n + 1 is the current physical time step and m + 1 is the current pseudo-time step. In addition, an implicit formulation is proposed in both, t and τ . The definition of the matrix of intrinsic time scale (τ ′ ) is very important in order to stabilize the numerical scheme correctly. For this formulation, it is proposed to apply the SUPG strategy, i.e. to stabilize the numerical scheme considering the advective part of the system only. From equation (72) and using the definitions given in (75), the system of Navier-Stokes equations expressed in the viscous variables basis is written as ∂Q ∂U ∂Q ˜ v, i ∂Q = ∂ + Γ−1 +A v ∂τ ∂Q ∂t ∂xi ∂xi

  ˜v ˜ v, ij ∂Q + S K ∂xj

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(94)

296

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti The numerical diffusivity introduced by the SUPG method in the inviscid case is [28] ˜ num = A ˜ v τ˜v A ˜v K v

(95)

where τ˜v is the matrix of intrinsic time scale in the viscous variables basis. In the conservative variables basis this matrix is expressed as τ′ =

∂U ∂Q −1 ∂U τ˜v Γ = τ˜v Γ−1 v ∂Q ∂U ∂Q

(96)

There are several approaches to compute the matrix τ˜v . One of them is the definition given by Hughes and Mallet [28] adapted to the preconditioned system ˜ v k−1 τ˜v = kB

(97)

˜ v, j ˜ v = ∂ξi A B ∂xj

(98)

being

the preconditioned advective jacobians transformed to the master element, in which ξi represents the master element coordinates. Another option is the proposed by Le Beau et al. [32]. For the preconditioned system, this proposal is expressed as h τ˜v = (99) ˜ v )| 2 max |λ(A In this chapter, the definition given by equation (99) is used to stabilize the numerical scheme plus a correction due to viscous effects (see, for instance, the work by Mittal and ˜ v are Tezduyar [38]). The eigenvalues of the advective jacobian matrix A Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

˜ v )1,2,3 = u λ(A (100) u(1 + βMr2 χ/c2 ± c˜) 2     2 βMr2 χ u2 2 where c˜2 = u2 1 + + 4βM . 1 − r c2 c2 Considering the non preconditioned system, for the wave with propagation speed λ(A)j the numerical diffusion introduced by the stabilization strategy given by equation (99) is proportional to hλ2 (A)j 2 maxi |λ(A)i | Thus, at the low Mach number limit, there are sub-stabilized modes due to the disparity in the wavespeeds. This sub-stabilization could leads to spurious numerical oscillations in the solution. Regarding for the shock capturing operator, it is used a modification of the ‘standard’ definition given by (31). The characteristic velocity is computed as uchar = kuk + c˜, and the operator is affected by a coefficient (˜ a)  β ∗ hJGN k∇ρh khJGN ′ δsc = a ˜ (kuk + c˜) (101) 2 ρref ˜ v )4,5 = λ(A

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... 4.2.2.

297

Dynamic Boundary Conditions

As with the stabilization in the finite element method, the dynamic boundary conditions must be reformulated for the preconditioned system of equations. The idea here is to follow the proposal by Storti et al. [54] but applied to the equation (77) expressed in viscous variable basis. In multidimensional problems a simplified 1D analysis in the normal direction to the boundary is done by considering the projection of the advective jacobians onto this direction, as follows ˜ v, n = A ˜ v, i ni A (102) where ni are the components of the unit vector normal to the local boundary. After diagonalization of the projected jacobian ˜ v, n = M ˜ v, n Λ ˜ v, n M ˜ −1 A v, n

(103)

˜ v, n = diag[λ(A ˜ v, n )], the projection matrices onto the right/left-going characteriswith Λ tics modes in the diagonal basis are obtained by ( ˜ v, n ) < 0  1 if i = j and λi (A − ΠV n ij = 0 otherwise (104) + Π− V n + ΠV n = I

The projection matrices in the viscous variables basis are computed changing the basis ± ˜ −1 ˜ Π± Qn = Mv, n ΠV n Mv, n

(105)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Finally, coming back to the U basis, the projection matrices expressed in this basis are Π± Un =

∂U ± ∂Q Π ∂Q Qn ∂U

(106)

As explained in section §2.2.3., the dynamic boundary conditions are applied using Lagrange multipliers. Therefore, for a node i on the boundary, the system of equations to be solved is + ˆ ˆ ˆ Π− U n (U)(Ui − U) + ΠU n (U)Ulm = 0 (107) ˆ lm = 0 Ri (U) + ΓΠ− (U)U Un

ˆ is adopted according to the direction of normal flow respect to boundary, as in The state U section §2.2.3..

4.3.

Results

The preconditioning strategy presented in this section to solve compressible flows at the low Mach number limit was applied to several problems in order to test the potentiality of the technique. Among the solved problems, there are steady and unsteady incompressible flows with moving domains. The purpose of these problems was to compare the preconditioned-system solution and the solution obtained by using a standard incompressible Navier-Stokes code. Furthermore, a test inherently compressible, which is the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

.

298

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

simulation of the in-cylinder flow in an opposed-piston IC engine under cold conditions, was solved. The pseudo-time step is increased during the pseudo-transient following the rule ∆τ m+1 = ∆τ 0

kRn,0 k kRn,m k

(108)

where Rn,m is the global residue and ∆τ 0 is an initial pseudo-time step defined by the user. 4.3.1.

Flow in a Lid Driven Cavity

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

This test has served as a benchmark for the Navier-Stokes equations for incompressible flow for decades. The problem consists in a fluid into a square cavity whose top wall moves with a uniform velocity. Two Reynolds numbers cases 100 and 1000 are considered and a 60×60 quadrangles mesh is employed. A uniform grid is used for Re = 100, and a stretched grid is used for Re = 1000, with a ratio of 1:10 between elements near the wall and elements in the central region of the domain. The Mach number of the moving lid is 4.5×10−4 . Wall boundary conditions are no slip and constant temperature. Initially, the fluid is at rest and its pressure and temperature are constants. The test was solved with the preconditioning strategy presented in this section and also with a standard incompressible Navier-Stokes (NSI) code. The incompressible Navier-Stokes equations were solved using a stabilized finite element SUPG-PSPG [10] [60] (Pressure-Stabilizing/Petrov-Galerkin) method. Figures 36 to 38 show the density, the velocity module and the pressure perturbation, respectively. The pressure perturbation is computed as p − p¯, with p¯ = 1 × 105 Pa. Notice that the solution obtained is smooth and with no numerical oscillations.

(a) Re = 100 - UP

(b) Re = 1000 - UP

Figure 36. Density field ([kg/m3 ]) for flow in a lid driven cavity. In order to verify the accuracy of the presented method, the velocity profiles at vertical and horizontal centerlines of the cavity (x1 = 0.5 and x2 = 0.5, respectively) are compared with a numerical solution of incompressible Navier-Stokes equations by Ghia et al. [23]. The u1 velocity is compared at vertical centerline of the cavity, and the u2 velocity is compared at horizontal centerline of the cavity. Figures 39 and 40 show the results, where good agreement can be observed.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

(a) Re = 100 - UP

(b) Re = 1000 - UP

299

(c) Re = 1000 - NSI

Figure 37. Velocity module field ([m/s]) for flow in a lid driven cavity.

(a) Re = 100 - UP

(b) Re = 1000 - UP

(c) Re = 1000 - NSI

Figure 38. Pressure perturbation field ([Pa]) for flow in a lid driven cavity. 1

0.8

Re = 1000

x2

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Re = 100 0.6

0.4 UP 0.2

N−S incompressible ref. sol. Ghia et al. (1982)

0 −0.4

−0.2

0

0.2

0.4 u 1 velocity

0.6

0.8

1

1.2

Figure 39. Comparison of u1 velocity component at vertical centerline of the cavity with numerical solution by Ghia et al. [23].

4.3.2.

Flow in a Channel with a Moving Indentation

This test case consists in a flow through a 2D channel with a moving indentation, which has been studied experimentally by Pedley and Stephanoff [46], and numerically by Ralph and Pedley [48] and by Demird˘zi´c and Peri´c [15]. Figure 41 shows a scheme of the channel geometry. The shape of the indentation was taken from Pedley and Stephanoff [46], whose specified the following analytic function which approximately fit the real shape used in the

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

300

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti 0.4 0.3 0.2

u 2 velocity

0.1

Re = 100

0 −0.1 −0.2 −0.3 −0.4

UP N−S incompressible

−0.5

ref. sol. Ghia et al. (1982)

−0.6

0

0.2

Re = 1000

0.6

0.4

0.8

1

x1

Figure 40. Comparison of u2 velocity component at horizontal centerline of the cavity with numerical solution by Ghia et al. [23].

experiment  for 0 < x1 < c1  H 0.5H{1 − tanh [a(x1 − c2 )]} for c1 < x1 < c3 x2 (x1 ) =  0 for x1 > c3

where a = 4.14, c1 = 4b, c3 = 6.5b, c2 = 0.5(c1 + c3 ), and

H = 0.5Hmax [1 − cos(2πt∗ )] t − t0 . Υ x2

b

c1 c2 c3

H

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

being t∗ =

(109)

l1

x1

l2

Figure 41. Geometry of the channel (not to scale): b = 1 cm, l1 = 9.85 cm, l2 = 18.0 cm. Here b is the channel height, Υ is the oscillation period and Hmax = 0.38b specifies the maximum blockage of the channel cross-section at t∗ = 0.5. The geometry is symmetric around x1 = 0. The Strouhal number based on the channel height, bulk velocity U = 2/3u1, max and the oscillation period, St =

b , UΥ

(110)

is 0.037. The Reynolds number based on the same reference quantities is 507. At the initial time t = t0 the flow is assumed to be fully developed (Poiseuille flow). The maximum Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

301

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

velocity is u1, max = 1.5 m/s. The velocity profile at the inlet cross-section is taken to remain constant throughout the cycle. Also a unit density is imposed at the inlet crosssection. At the other channel end, dynamic boundary conditions with the ULSAR strategy were imposed, as presented in section §4.2.2.. Walls are assumed isothermic and no slip boundary condition is imposed on them. Mesh dynamics was solved applying the method described in section §3.. The mesh used has 12.4K triangular elements and 6.8K nodes. The (physical) time step adopted in the simulation was ∆t = Υ/200. Figures 42, 43 and 44 show the density field, the velocity module, and the pressure perturbation (with p¯ = 1 × 105 Pa) at times t∗ = 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1. These results were obtained by using the preconditioning strategy presented above.

Figure 42. Density field ([kg/m3 ]) for the flow in a channel with a moving indentation computed by using the UP strategy. From top to bottom, times t∗ = 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1. Table 4 compares the solution obtained by means of the UP strategy and two solutions of the Navier-Stokes equations for incompressible flow, one obtained using the PETSC-FEM code and the other extracted from the bibliography. The comparison is done taking the maximum velocity differences and the time in which appear the three first vortexes. Figure 45 shows the positions of the first three eddies center as a function of time. In that figure, the experimental data reported by Pedley and Stephanoff [46], and the numerical solutions of the UP technique and the Navier-Stokes equations for incompressible flow [36]

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

302

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Figure 43. Velocity module field ([m/s]) for the flow in a channel with a moving indentation computed by using the UP strategy. From top to bottom, times t∗ = 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1. Table 4. Comparison of results obtained by using the UP strategy and two solutions of the incompressible Navier-Stokes equations for the channel with a moving indentation test. UP Ref. [36] Ref. [15] Max. vel. [m/s] 2.916 at t∗ = 0.37 2.931 at t∗ = 0.38 2.645 at t∗ = 0.4 1st vortex t∗ = 0.22 t∗ = 0.23 t∗ = 0.2-0.25 nd ∗ ∗ 2 vortex t = 0.345 t = 0.35 t∗ = 0.35-0.4 3rd vortex t∗ = 0.425 t∗ = 0.42 t∗ = 0.45 are included. According to Pedley and Stephanoff [46], the abscissa is defined as x∗1 =

(x1 − c1 )(10St)1/3 b

(111)

Also, this test was solved using the non-preconditioned system of equations, with the same conditions as presented above. In figures 46, 47 and 48 the density field, the velocity magnitude field and the pressure perturbation field are shown at t∗ = 0.2, 0.4, 0.6, 0.8 and 1 respectively. The non-preconditioned solution does not represent the behavior of the flow Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

In Search of Improvements for the Computational Simulation...

303

Figure 44. Pressure perturbation field ([Pa]) for the flow in a channel with a moving indentation computed by using the UP strategy. From top to bottom, times t∗ = 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1. since it does not produce the different vortexes experimentally observed. The pressure field has numerical oscillations, which can not be observed in figure 48 due to the scale used. Thus, the pressure perturbation field at t∗ = 0.5 is shown in figure 49 with an appropriate color scale. 4.3.3.

Opposed-Piston Engine

The case consists in the resolution of the fluid flow inside the cylinder of an opposedpiston engine under cold conditions, i.e. without combustion. This test was selected in order to apply the preconditioning strategy to an inherently compressible case similar to what it is found in real engine geometries and also due to its simplicity. The engine geometry was taken from the KIVA-3 [3] tutorial. The cylinder bore is 100 mm, the stroke of each piston is 85 mm, and the geometric compression ratio is 9.5:1. The cylinder has 8 exhaust ports equally distributed in the circumferential direction and 12 intake ports uniformly separated also. Assuming the reference angle as the EDC (External Dead Center), the timing of the ports are the following • Intake Port Opening (IPO) = 295.13◦

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

304

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti 1

0.8

t*

0.6

0.4

UP 0.2

0

NSI Experiment 0

1

2

3

4

5

6

7

x*1

Figure 45. Comparison of predicted and experimentally observed positions of first three vortexes center. • Intake Port Closing (IPC) = 64.87◦ • Exhaust Port Opening (EPO) = 280.2◦

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

• Exhaust Port Closing (EPC) = 79.8◦ In order to simplify the problem, the flow domain is reduced to a two-dimensional one by means of the intersection between the 3D cylinder and a plane containing the axis of the cylinder. This plane must be oriented in such a way to ‘cut’ two intake and exhaust ports simultaneously. The resultant geometry is shown in figure 50 for pistons located at EDC. The mesh was generated with the pistons at EDC (ports totally opened) and has 76K triangular elements (structured) and 38K nodes. The mean element size is h = 0.5 mm. Due to the simplicity of the geometry and the boundary movement, the mesh dynamics is solved by using an algebraic law following a linear distribution with respect to the position of pistons at IDC (Internal Dead Center). No-slip condition is imposed at solid walls. In addition, solid walls are assumed insulated. Mixed absorbing/wall boundary conditions are used to model the ports, as explained in section §2.2.3.. For absorbing boundary conditions, the reference state used for intake ports is Uiref = [1.2195 kg/m3 , 0 m/s, 0 m/s, 105 kPa]T , and for the exhaust ports is Ueref = [0.662 kg/m3 , 0 m/s, 0 m/s, 95 kPa]T . The engine speed is 3000 rpm. The time step used in the simulation corresponds to 0.5 crank angle degree (CAD) and, for the conditions of the flow in the test, a CF Lu O(1). The stationary cyclic state is reached with approximately four cycles. The following results correspond to the last cycle simulated. For some instants in the cycle, the density and pressure fields into the chamber are depicted in the following figures. The purpose is to show that the unsteady preconditioning strategy presented in this chapter produces smooth solutions (without numerical oscillations) when it is applied to computations of in-cylinder flows problems. Figures 51 and 52

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

305

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 46. Density field ([kg/m3 ]) for the flow in a channel with a moving indentation computed by using the NP strategy. From top to bottom, times t∗ = 0.2, 0.4, 0.6, 0.8 and 1.

Figure 47. Velocity module field ([m/s]) for the flow in a channel with a moving indentation computed by using the NP strategy. From top to bottom, times t∗ = 0.2, 0.4, 0.6, 0.8 and 1.

show the distribution of density and pressure into the cylinder at EDC and IDC, respectively. In figure 53 these fields are plotted at an intermediate position of the pistons, corresponding to 270 CAD. In figures 54 and 55, the mean density and pressure in the cylinder during a whole cycle are presented, respectively. This brief overview of the results obtained is completed with some pictures of the magnitude of flow velocity at several instants in the cycle. Figure 56 shows these results.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

306

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 48. Pressure perturbation field ([Pa]) for the flow in a channel with a moving indentation computed by using the NP strategy. From top to bottom, times t∗ = 0.2, 0.4, 0.6, 0.8 and 1.

Figure 49. Pressure perturbation field ([Pa]) for the flow in a channel with a moving indentation computed by using the NP strategy at t∗ = 0.5, where the scale was modified.

5.

Coupling of 1D/Multi-D Domains for Compressible Flows

Generally, when CFD-3D models are used to simulate the fluid flow in IC engines, due to computational resources availability reasons only a few components of the engine are studied at each time step. Usually, the boundary conditions for these 3D models are dynamic and, hence, are not easy to impose. A typical approach is to use 0D/1D codes as boundary condition generators for the 3D problem. When dimensionally heterogeneous models are applied to solve a given problem, the need to perform the coupling between sub-domain arises. For IC engine simulation, the coupling between multi-D and 1D domains is the most useful. Thus, a method focused on such coupling type is presented in this section. The Domain Decomposition theory provides the framework to develop a coupling domain algorithm. Several research work was done in this sense, specially considering elliptic operators and the advection-diffusion-reaction equation [9, 47, 20, 12, 63, 21, 2]. Given a boundary value problem defined on a domain Ω, a partition of that domain in ns subdomains (Ωi , i = 1, . . . , ns ) is built. These sub-domains could be disjoint or overlapping. The original boundary value problem is reformulated in a split form on the sub-domains,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... e

U ref

i

U ref

1111111111111 0000000000000 0000000000000 1111111111111 exhaust ports

307

e

U ref

0 1 1010 0 1 0 1 1010 0 1 0 1 1010 0 1 0 1 100 0 1 110 0 1 0 1 100 0 1 110 0 1 0 1 1010 0 1 0 1 100 0 1 110 0 1 0 1 100 0 1 1 i intake ports 0000000000000 U ref 1111111111111 0000000000000 1111111111111

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 50. Geometry of the simplified 2D model for the opposed-piston engine case (pistons at EDC).

Figure 51. Density ([kg/m3 ], left) and pressure ([Pa], right) fields at EDC (0◦ ).

and the sub-domain solutions satisfy suitable matching conditions at sub-domain interfaces. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

308

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 52. Density ([kg/m3 ], left) and pressure ([Pa], right) fields at IDC (180◦ ).

Figure 53. Density ([kg/m3 ], left) and pressure ([Pa], right) fields at 270◦ . These boundary conditions are named transmission conditions [47], and could involve appropriate combinations of the following types • Dirichlet type: condition on the unknown of the problem. • Neumann type: condition on the first derivative of the unknown.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... IPC EPC

IDC

309

EPO IPO

9 8

density [kg/m 3]

7 6 5 4 3 2 1

0

50

100

150 200 crank angle [deg]

250

300

350

Figure 54. Cylinder mean density through a cycle of the opposed-piston engine. IPC EPC

IDC

EPO IPO

2000

pressure [kPa]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

1500

1000

500

0

0

50

100

150 200 crank angle [deg]

250

300

350

Figure 55. Cylinder mean pressure through a cycle of the opposed-piston engine. • Robin type: linear combination of Dirichlet and Neumann conditions. For second order differential equations problems, some combination of the above cited boundary conditions could fail to give the original solution on the corresponding sub-domain [27]. In particular, Dirichlet/Dirichlet (D/D) coupling as well as the Neumann/Neumann (N/N) coupling are not possible. For advection-dominated advectiondiffusion equations it is important that the boundary conditions accounts for the flow direction. Some techniques, such adaptive strategies, introduce iterative methods splitting the above interface conditions in a way which is adapted to the local flow direction [12, 63]. Other methods, such as the proposed by Alonso et al. [2], do not care about the local direc-

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

310

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

(a) 0 CAD

(b) 90 CAD

(c) 180 CAD

(d) 300 CAD

Figure 56. Magnitude of flow velocity field ([m/s]) at several instants during a cycle.

tion of the advective field, but only need that the boundary value problems defined on the sub-domains are associated for a suitable coercive bilinear form. As mentioned above, the purpose is to link dimensionally heterogeneous models for numerical simulation of IC engines. Thus, the focus is placed on the governing equations of compressible fluid flows, namely, the Navier-Stokes and Euler equations. For these equations, the different boundary condition types are interpreted here as follows

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

311

• Dirichlet: condition on the vector state (U) or the advective flux (Fa ). This is true due to Fa is a function of the vector state only, and not of their derivatives. Moreover, in the case of a perfect gas the advective flux is a homogeneous function of degree 1 of the conservative variable vector [7, 53], this is Fa (λU) = λFa (U)

for any λ ∈ R

and taking derivatives with respect to λ and setting λ = 1, it is obtained the relation [26] ∂Fa U = AU Fa (U) = ∂U • Neumann: condition on the diffusive flux (Fd ), due to it contains first order derivatives (see equation (10)). • Robin: some linear combination of the other two types, for instance, a condition on the total flux (F = Fa + Fd ). For viscous flows, the system of equations contains second order derivatives and, hence, conditions on state and its first derivatives are necessary to impose on the coupling boundary [26]. In addition, it is considering a non-overlapping sub-domain coupling method. It will begin discussing a 1D domain case divided in two intervals. Let Ω a bounded, open interval of R discretized by a grid with N elements and N + 1 nodes numbered from 1 to N + 1. After standard discretization, for instance with the SUPG stabilized Finite Element Method, a system of equations of the following form is obtained

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

E1 (U1 , U2 ) = 0 E2 (U1 , U2 , U3 ) = 0 .. .

(112)

EN +1 (UN , UN +1 ) = 0 It has been assumed that the equation at node i involves only the nodal states at nodes i−1, i and i+1, as is true for first order Finite Difference Method and FEM discretization methods. Also a steady system of equations is assumed. The system (112) represents (N + 1) × ndof equations in the (N + 1) × ndof unknowns {U1 , U2 , . . . , UN +1 }. It is assumed that this non-linear system of equations has a unique solution. Equations at the boundary nodes may include some mixture of Dirichlet or Robin and Neumann boundary conditions. This system of equations is split up at a certain internal node i, so that the domain Ω = [x1, 1 , x1, N +1 ] is split in the sub-domains Ω1 = [x1, 1 , x1, i ] and Ω2 = [x1, i , x1, N +1 ] (see figure 57). Now, node i splits in i1 and i2 for the left and right sub-domains, respectively. Appropriate boundary conditions at i1 (i2) for Ω1 (Ω2 ) must be provided. These conditions must ensure that each sub-domain problem can be solved independently, and that the problem is well posed. An iterative approach to solve the problem must guarantee that the solutions at each boundary converge to the solution of the coupled system, i.e., Uki1 , Uki2 → Ui Ukj → Uj

∀j, j 6= i

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

(113)

312

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

for k → ∞, k being the iteration number. In addition, it is assumed that the equation at node i can be separated in its right and left contributions Ei (Ui−1 , Ui , Ui+1 ) = Ei1 (Ui−1 , Ui ) + Ei2 (Ui , Ui+1 ) = 0

(114)

P x 1,i x 1,1

x 1,N+1 x 1,i1

x 1,i2

x 1,1

x 1,N+1 P2

P1

Figure 57. Sketch of 1D domain splitting.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

5.1.

Coupling for Implicit Schemes ‘Monolithically’ Solved

If an implicit scheme for time integration is used and assuming that the resolution could be performed as a ‘monolithic’ system, the coupling strategy reduces to a constraint between the states at interface nodes. This strategy could be useful when the codes that perform the computation on each sub-domain are not ‘black boxes’, but the contributions to the global residue (and perhaps the global jacobian matrix) are available. For the problem (112), if the equality of states at the coupling node is imposed, then it results in a linear constraint. With the notation of previous section, the system of equations is expressed as E1 (U1 , U2 ) = 0 E2 (U1 , U2 , U3 ) = 0 .. . Ei1 (Ui−1 , Ui1 ) = 0 Ui1 = Ui2

(115)

Ei2 (Ui2 , Ui+1 ) = 0 .. . EN +1 (UN , UN +1 ) = 0 where the constraint Ui1 = Ui2 imposes the continuity of the solution and forces the continuity of fluxes through the coupling interface. By elimination and using (114), it is easy to see that the problem (115) reduces to (112).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... 5.1.1.

313

Coupling of 1D/Multi-D Domains

If the purpose is to link a multi-D domain with a 1D domain, the constraints are not straightforward as in the 1D/1D coupling. In this case, the conditions at the coupling interface for the multi-D domain are defective [18, 33] and a simplification must be done at the coupling interface. One possibility is to impose each node at coupling surface of multi-D domain to have the same state as the corresponding node in the 1D domain. Let M the number of nodes lying on the coupling interface of the multi-D domain, and let j the node identifying the end of the 1D domain. Then, the M ndof constraints are Uj = Ui

i = 1, . . . , M

(116)

This kind of coupling could be useful with an uniform (‘piston’-like) flow through the multiD domain boundary, limiting the shape of variables profile in the coupling interface. For instance, this limitation do not allow to apply a non-slip boundary condition at walls, even whether the flow is parallel to the wall. Another option is to equalize the mean value of the state on the whole coupling surface and the corresponding state in the 1D domain, i.e. Z Z UM D dS = U1D dS = U1D S (117) S

S

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

where S is the coupling surface. The constraint (117) could cause the code failure since the problem is not restricted enough, without guarantee of a unique solution. For example, it is possible to obtain a profile of density with non physical negative values on the coupling surface and whose mean value is a feasible value.

5.2.

Results

Several tests cases were solved using the linear constraints presented above. The following sections show some results from 1D/1D, 2D/1D and 3D/1D couplings. In the whole set of test cases, the ‘standard’ Navier-Stokes equations for compressible flow are solved, i.e. the equations (8) with the variational formulation given by (26). 5.2.1.

1D/1D Coupling

In this particular case, strategies (116) and (117) are equivalent. Using two or more subdomains, the solution obtained is equal to the solution computed from an unique domain. 5.2.2.

2D/1D Coupling

The 2D/1D coupling test proposed consists of a reservoir connected to the atmosphere by a pipe with length L = 7 m. The pressure and the temperature of the gas inside the reservoir are constant, with values of p0 = 1.2 × 105 Pa and T0 = 278.75 K, respectively. The atmospheric pressure is pa = 1 × 105 Pa. Initially, the gas inside the pipe is at rest, and at the same pressure and density as the gas in the reservoir. At t = 0, the pipe end connected to the atmosphere is suddenly open. The pipe is modeled using three sub-domains, 1D at

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

314

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

the pipe ends and 2D in its middle region as shown in figure 58 (1D-2D-1D). Also, in order to compare the solutions, the problem is solved using three 1D domains (1D-1D-1D). The duct is adiabatic and without friction, and the fluid is inviscid.

p

1D

1D

0

2D

p a

T0

Figure 58. Gas discharge from a reservoir to the atmosphere.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The 2D domain is discretized by using an unstructured mesh containing 592 triangular elements. For 1D domains uniform meshes are used, with element size h = 0.05 m. The time step in the simulations is ∆t = 1 × 10−4 s, which gives a CFL number O(1). The left coupling section is located at x1 = 2 m, and the right coupling section at x1 = 5 m. Solving the problem with the constraints given by equation (116) for the density, the pressure and the mean axial velocity, the test is successful and the whole time interval can be solved. Figures 59(a) to 59(c) show the results achieved at coupling sections. As expected, there are no differences between the cases 1D-2D-1D and 1D-1D-1D. 5.2.3.

3D/1D Coupling

This case consists of the exhaust manifold of a six-cylinder spark-ignition four-stroke engine1 . The purpose is to solve a junction of the manifold with a CFD-3D code and to solve the rest of the engine with 0D/1D models. The exhaust manifold is composed by three junctions, as it is shown in figure 60. Two of these junctions connect three header pipes with one of the two intermediate pipes (junctions 3-to-1). These pipes converge to the third junction that connects them with the exhaust tailpipe (junction 2-to-1). The diameters and lengths of the pipes in the manifold are the following • Header pipes: D = 51.7 mm, L = 570 mm. • Intermediate pipes: D = 61 mm, L = 650 mm. • Tailpipe: D = 70 − 85 mm, L = 600 mm. One of the junctions 3-to-1 is modeled as a 3D domain, and coupled to 1D domains representing the pipes connected with it. The geometric model is shown in figure 61, where 1

We want to thank to Juan Pablo Alianak for provide me the computational 3D geometric model of the six-cylinder exhaust manifold. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... 300

1.5 left section

1.48

250 axial velocity [m/s]

density [kg/m 3]

1.46 1.44 1.42 1.4 right section 1.38 1.36

200

150

100

1D−2D−1D 1D−2D−1D 1D−1D−1D

1.34 1.32 1.3

315

0

0.02

0.04

0.06

0.08

0.1 t [sec]

0.12

0.14

0.16

0.18

left section (1D−2D−1D) right section (1D−2D−1D) 1D−1D−1D

50

0

0.2

0

0.05

(a) Density

0.1 t [sec]

0.15

0.2

(b) Axial velocity

120 118

left section

116 pressure [kPa]

114 112 110 108

right section

106 1D−2D−1D

104

1D−2D−1D 102 100

1D−1D−1D 0

0.02

0.04

0.06

0.08

0.1 t [sec]

0.12

0.14

0.16

0.18

0.2

(c) Pressure

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 59. Time evolution of solution at coupling sections.

Figure 60. Exhaust manifold geometry.

a (relatively short) stretch of the pipes were added to the 3D model of the junction in order to avoid the failure of the 1D approximation at the coupling section. With the aim of simplifying the resolution, only the branch of the exhaust manifold containing the junction 3-to-1 is solved. In figure 62 a sketch of the computational model of the branch is presented. This figure shows two models, one of them is composed by four 1D pipes and the 3D junction, while in the other one the junction is represented as a 0D component. In this last case, the model by Corberan (see appendix 7.) is applied to solve the junction pipes. Thus, a comparison of the solution in the 1D pipes between both computational models is done and, a verification of the hypothesis in the junction model by Corberan is made also.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

316

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The boundary conditions for these models are obtained simulating the whole engine by using a 0D/1D [51] code until a stationary (cyclic) state is reached. The engine speed is 8000 rpm and, therefore, the period of the cycle is Υ = 0.015 s. For the 3D junction, insulated walls are considered and slip condition is imposed on them. At coupling interfaces, the density and pressure over the whole section are equalized to the corresponding value of the 1D domains. In addition, the component of the velocity in the normal direction to the surface is equalized to the axial velocity at end pipes connected to the 3D domain. The tangential components of velocity on inlet/outlet sections are constrained to be null. The mesh of the 3D model has 278K tetrahedra and 67K nodes, and the 1D domains were discretized by means of uniform meshes with element size of h = 5 mm. The time step of the simulation is ∆t = 5 × 10−5 s. Several periods were simulated in order to reach (approximately) a stationary state in the solution. For the 1D-1D/0D-1D model, the discretizations in time and space are the same as in the 1D-3D-1D model.

Figure 61. Geometry of the junction 3-to-1. Some instantaneous distributions of the pressure over the junction skin are shown in figure 63, where t − t0 t∗ = (118) Υ is the non-dimensional time, being t0 the start time of the cycle. Solutions from 1D-3D-1D and 1D-1D/0D-1D models are compared at coupling 1D/3D interfaces, named from P1 to P4 as it is indicated in figure 62. These comparisons are presented in figures 66 to 68. In figure 67 the velocity is referred as axial velocity which, for the 3D model, should be interpreted as the component of the velocity vector in the normal direction to the given surface. In the coupling sections of the header pipes, the waves in both models tend to have similar behavior, especially when the pulse of the velocity wave reaches the interface. The solution of the 1D-1D/0D-1D model is more oscillatory than the solution of the other model,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... 1D

3D P3 P2 P1

317

1D

J3 J2

J4

P4

J1

junction (0D)

1D

1D / 0D

1D

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 62. Computational models to solve a branch of the exhaust manifold.

Figure 63. Pressure field over the junction 3-to-1 skin at t∗ = 0.

which is because the pipes junction 0D model has no inertia. In general, the amplitude values are near each other on both solutions, being more attenuated for the 1D-3D-1D solution. The largest differences are found when the flow is established from the junction to the pipe. The influence of both the finite volume of the pipe junction and its geometric shape is more clearly evident in the solutions obtained in coupling section P4. In this interface, the pressure peaks on the 1D-3D-1D solution are higher than on the 1D-1D/0D-1D model. The CFD 3D could be useful to validate or to improve a 0D model. For the junction 3-to-1 solved, the hypothesis of the pipes junction 0D model by Corberan will be checked. This model is composed by conservation equations, incoming path/Mach lines equations,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

318

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 64. Pressure field over the junction 3-to-1 skin at t∗ = 0.4.

Figure 65. Pressure field over the junction 3-to-1 skin at t∗ = 0.8. the equality of the pressure in all pipes end, and the equality of the enthalpy in all outgoing branches in the junction, as presented in appendix 7.. With the exception of the incoming path/Mach lines equations, the remaining equations can be verified.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation... 0.26

0.26 1D−3D−1D 1D−1D/0D−1D

0.25

0.24 density [kg/m 3]

density [kg/m 3]

1D−3D−1D 1D−1D/0D−1D

0.25

0.24 0.23 0.22 0.21

0.23 0.22 0.21

0.2

0.2

0.19

0.19

0.18

0

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

0.18

1

0

0.1

0.2

(a) Section P1 0.26

0.24

0.25

0.4

0.5 t*

0.6

0.7

0.8

0.9

1

0.7

0.8

0.9

1

1D−3D−1D 1D−1D/0D−1D

0.24 density [kg/m 3]

0.23 density [kg/m 3]

0.3

(b) Section P2

0.25

0.22 0.21 0.2

0.23 0.22 0.21 0.2

1D−3D−1D 1D−1D/0D−1D

0.19 0.18

319

0

0.1

0.2

0.3

0.19 0.4

0.5 t*

0.6

0.7

0.8

0.9

0.18

1

0

0.1

0.2

(c) Section P3

0.3

0.4

0.5 t*

0.6

(d) Section P4

Figure 66. Time evolution of density at coupling sections through a cycle.

500

500 1D−3D−1D 1D−1D/0D−1D

400

axial velocity [m/s]

axial velocity [m/s]

300

200 100 0 −100

200 100 0 −100

−200

−200 0

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

−300

1

0

0.1

0.2

(a) Section P1

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

1

0.7

0.8

0.9

1

(b) Section P2 260

500 1D−3D−1D 1D−1D/0D−1D

400

240 220 axial velocity [m/s]

300 axial velocity [m/s]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

300

−300

1D−3D−1D 1D−1D/0D−1D

400

200 100 0

200 180 160 140 120

−100 100 −200 −300

1D−3D−1D 1D−1D/0D−1D

80 0

0.1

0.2

0.3

0.4

0.5 t*

0.6

(c) Section P3

0.7

0.8

0.9

1

60

0

0.1

0.2

0.3

0.4

0.5 t*

0.6

(d) Section P4

Figure 67. Time evolution of axial velocity at coupling sections through a cycle.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

320

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti 140

140 1D−3D−1D 1D−1D/0D−1D

120 110 100 90 80

1D−3D−1D 1D−1D/0D−1D

130

pressure [kPa]

pressure [kPa]

130

120 110 100 90

0

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

80

1

0

0.1

0.2

(a) Section P1

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

1

0.7

0.8

0.9

1

(b) Section P2

140

135 1D−3D−1D 1D−1D/0D−1D

130

1D−3D−1D 1D−1D/0D−1D

130

pressure [kPa]

pressure [kPa]

125 120 110 100

120 115 110 105 100

90 95 80

0

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

1

90

0

0.1

(c) Section P3

0.2

0.3

0.4

0.5 t*

0.6

(d) Section P4

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 68. Time evolution of pressure at coupling sections through a cycle. The mass conservation equation is evaluated by computing the following quantity as a function of time P4 ˙i i=1 m (119) m ˙ ref P where m ˙ ref = 4i=1 |m ˙ i |. For the last cycle simulated, the curve for expression (119) is plotted in figure 69(a). Important deviations from the zero value (as assumed in the 0D model) could be observed in the figure. In a similar manner, for the energy conservation it is calculated the following expression P4 ˙ i=1 hi (120) ˙href P with h˙ ref = 4i=1 |h˙ i |. The results obtained are shown in figure 69(b) where, again, the difference between the 0D assumption and the 3D solution is relatively high. In figure 69(c) the relative difference of pressure is presented, which is computed as p4 − pi , p4

i = 1, 2, 3

(121)

where p4 is the mean pressure over the surface corresponding to the interface between the junction and the intermediate pipe (J4, see figure 62), and pi the mean pressure over the respective surfaces for header pipes (named Ji, i = 1, 2, 3 in figure 62). With p4 as reference pressure the maximum error is approximately 18 % and, as could be inferred from the figure, it is expected errors of the same order whether other pressures were adopted as reference.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

321

The final hypothesis in the model by Corberan is the equality of the enthalpy in all outgoing pipes. From the viewpoint of the intermediate pipe the interface J4 is an inlet for all times, then, the enthalpy over such a section (h4 ) is adopted as reference. The relative difference of outgoing enthalpy is computed as h4 − hi , h4

i = 1, 2, 3

(122)

hi being the mean enthalpy over the section Ji. Figure 69(d) shows this relative difference, where the maximum error is about 10 %. As could be noted, the curves are discontinuous due to the change in the sense of flow through the cycle. 0.4

0.3

0.3

0.2

0.2

Σ h i / href

Σ m i / mref

0.4

0.1 0

−0.1

−0.2

−0.2

0

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

−0.3

1

(a) Equilibrium of relative mass fluxes at junction

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

1

15 relative dif. of outgoing enthalpy [%]

head pipe 1 head pipe 2 head pipe 3

15 10 5 0 −5 −10

0

(b) Equilibrium of relative enthalpy fluxes at junction

20 relative difference of pressure [%]

0

−0.1

−0.3

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

0.1

0

0.1

0.2

0.3

0.4

0.5 t*

0.6

0.7

0.8

0.9

1

(c) Pressure difference with respect to pressure at J4

head pipe 1 head pipe 2 head pipe 3 10

5

0

−5

0

0.2

0.4

0.6

0.8

1

t*

(d) Outgoing enthalpy difference at junction with respect to enthalpy at J4

Figure 69. Relative errors in the hypothesis of the pipes junction 0D model by Corberan. Summarizing, for the example solved the major deviations in the assumptions of the Corberan model were found in the conservation of mass and energy. This could be because the volume of the junction is neglected in this 0D model. The hypothesis of equality of pressure in all branches and the equality of the enthalpy in all outgoing pipes present a moderate relative error.

6.

Numerical Simulation of the MRCVC Engine

In this section, the computational tools developed in previous sections are applied to simulate the Constant-Volume Combustion Rotative Engine (MRCVC, for Motor Rotativo de Combusti´on a Volumen Constante [62]). Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

322

Ezequiel J. L´opez, Norberto M. Nigro and Mario A. Storti

The main feature of the MRCVC engine is that the combustion could be performed at constant volume (effectively). To be more precise, when the chamber reaches its minimum volume and during a finite angular interval, the chamber changes its shape but not its volume. Furthermore, the combustion chamber has a ratio surface area/volume similar to those found in reciprocating engines [61]. Thus, a net increment on the engine thermodynamic efficiency could be reached when it is compared with both, the rotative Wankel engine and the classical engine with reciprocating pistons. The MRCVC engine has a perfect static and dynamic balance of its moving components, and hence, allowing to achieve high smoothness and low engine vibration. Also, the contact between apex seals and walls is harmonic, which should permits to reduce wear and noise.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

6.1.

Operation and Geometry of MRCVC

The MRCVC was invented and patented by Jorge Toth [62], and it is under development at the Applied Mechanical Department of National University of Comahue (Neuqu´en, Argentina). In order to test the kinematic of the mechanism designed, at present, only a compressor with the geometry of the MRCVC was built [61]. This engine is composed by a rotor and two or more vanes inside a cylindrical housing. Figure 70 illustrates the geometry of a MRCVC with four vanes. The central region of the housing and the vanes have oval shape, with apex seals to avoid gas leakage. The rotor is a ring with cylindrical hollows allowing the relative rotation movement of the vanes. Each vane must to keep parallel its centerline with respect to the other vanes while their centers revolve around the output shaft. In this form, the vanes have translational movement only. This kinematic constraint is accomplished by means of a rim, which also links the engine shaft with the rotor and vanes. Breathing could be through ports in the side housings and/or through lateral ports in the center housing. However details of the gas-exchange system is undefined yet and it is under development. Considering as null the radii of apex seals, the geometry of MRCVC engine is completely defined by specifying the number of vanes (n), the radius of the trajectory center of the vanes (R), the half length of vane centerline (r), and the height of the chamber (h) [61]. Figure 71 shows a sketch of the top view of the MRCVC indicating the main geometric parameters. This geometric simplification is applied in this analysis to model the flow domain. Although a MRCVC with the simplest geometry is technologically unfeasible, the simplification turns simpler the analysis but retaining the main characteristics of the machine. The rotation angle of engine shaft (θ) is measured clockwise relative to the trailing vane position when the chamber has its maximum volume at the start of compression stroke. Due to the symmetry of the MRCVC geometry, it is enough to analyze the angular interval n+2 π (123) 2n which corresponds to a variation of the chamber volume from its maximum value to the minimum one. The geometry chamber for the remain part of the cycle is obtained by flipping the domain around the geometric symmetry axis. In order to get analytical formulae for describing the chamber geometry, the interval (123) is split in five sub-intervals due to topological changes and boundary redefinition. These intervals are defined as 0≤θ≤

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In Search of Improvements for the Computational Simulation...

323

Figure 70. Cutaway drawing of four-vanes MRCVC engine. rotor

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

vane R r r

r

stator θ

Figure 71. Basic geometry of a three-vane MRCVC and definition of its main geometric parameters. • Interval 1: 0≤θ
1) and contraction (s < 1) of the mother wavelet, whereas the translation parameter indicates the location of the mother wavelet in time. In Eq. (3), the expression 1 / s is introduced as a normalization factor to ensure that the wavelet has the same energy at every scale. The amount of signal energy contained in a specific scale s and location τ is given by the squared modulus of the CWT:

P ( s,τ ) = W ( s ,τ ) 2 .

(4)

This energy density function is defined as the wavelet power spectrum (WPS). The integral formulation given by Eq. (2) applies to a signal x(t) that is a continuous function of t. In order to use it for a time series, this integral representation must be discretized in an appropriate fashion. Consider a time series {xn } with n =1, 2, 3, …, N. For such a time series, Eq. (2) may be discretized as [23]:

Wn ( s ) =

N



n '= 1

⎛ δτ ⎞ ⎜ ⎟ ⎝ s ⎠

1/ 2

⎡ (n' − n)δτ ⎤ xn' ψ * ⎢ ⎥. s ⎣ ⎦

(5)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Here n is the time index, and δτ is the sampling interval. In order to calculate the CWT from Eq. (5), the convolution procedure given by this equations should be performed N times for each scale. However, it is possible to carry out all N convolutions simultaneously in Fourier space using a discrete Fourier transform (DFT). Using Eq. (5), the WPS can be evaluated as:

| Wn ( s ) | 2 . The WPS provides a good description of the fluctuation of the variance at different scales. The WPS which depends on both scale and time is represented by a surface. By taking contours of this surface and plotting them on a plane, a time-scale representation of the wavelet power spectrum may be derived. The time-scale representation is also referred to as a scalogram. From a scalogram the various periodicities and intermittency in the time series can be identified by visual inspection. In our analysis we used a Morlet wavelet as the mother wavelet. A Morlet wavelet consists of a plane wave modulated by a Gaussian function and is described by:

ψ (η ) = π −1 / 4 e

iω0η

e −η

2

/2

,

(6)

where ω 0 is the center frequency, also referred to as the order of the wavelet. Morlet wavelets have been used successfully in a wide variety of applications for feature extraction in time series data. We have used a Morlet wavelet with ω 0 = 6. This value of ω 0 provides a good balance between time and frequency localizations. For ω 0 = 6, the scale is approximately equal to the Fourier period. As a consequence, the scale and the period can be used

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

354

Asok K. Sen and Grzegorz Litak

interchangeably for interpreting the results. Note that in Eq. (6), the coefficient π −1 / 4 is used as a normalization factor so that the wavelet has unit energy. For a time series of finite length, computation of CWT using DFT requires that the time series is cyclic. To satisfy this requirement, the time series is often padded with zeros at the ends. Zero padding introduces discontinuities at the end points, especially at larger scales, and decreases the amplitude near the edges. In a scalogram, the region where the edge effects become important is referred to as the cone of influence (COI). Within the COI, the wavelet results may be unreliable and should be used with caution [23]. In addition to the information provided by a scalogram, useful information about the behavior of a time series can be obtained by calculating the scale-averaged wavelet power (SAWP). The scale-averaged wavelet power represents the average variance in a certain range of scales (or a band). The SAWP over a specific band represents the average variance in that band; in other words, it describes the fluctuations in power in that band. In particular, by plotting the SAWP, the presence of intermittency, if any, can be clearly identified. The SAWP is defined as the weighted sum of the wavelet power spectrum (WPS) over scales corresponding to j1 and j 2 :

Wn

2

=

δ j δτ Cδ

j = j2

Wn ( s j )

j = j1

sj



2

.

(7)

Here δτ is the sampling interval as indicated earlier, and the factor δj determines the scale resolution. The constant Cδ is a reconstruction factor whose value depends on the choice of

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

the specific mother wavelet; for a Morlet wavelet with ω 0 = 6, Cδ = 0.776 (see [23] for details).

3. Experimental Facilities and Measurement Procedure 3.1. Pressure Measurements in a Spark-Ignition Engine The SI engine data analyzed in this paper were collected from a four-stroke, singlecylinder Aprilia/Rotax engine (see Figures 1a,b). The experiments were performed in the Engine Laboratory at the University of Trieste, Italy. The in-cylinder pressure was measured using a piezoelectric pressure sensor. The output of the pressure sensor was passed through a charge amplifier and was subsequently digitized and stored in a personal computer. Details of the experimental design can be found in our earlier paper [10]. For the purpose of acquiring data, the torque on the engine was varied and the internal pressure in the engine cylinder was measured over 1000 consecutive cycles. From these measurements the maximum value of pressure ( p max ) in each cycle is obtained. The time series of cycle-to-cycle variations of p max for the first 960 cycles are plotted in Figure 2, for torques F = 0, 10, 20, 28, 40 and 43 Nm. Using a continuous wavelet transform (CWT) with a

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Dynamics of Pressure Fluctuations in Internal Combustion Engines

355

Morlet wavelet of order 6 as the mother wavelet, we performed wavelet analysis of these time series to determine the various periodicities of maximum pressure fluctuations.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

(a)

(b)

(c) Figure 1. (a) Experimental design of a four-stroke, single-cylinder spark-ignition engine in the Engine Laboratory at the University of Trieste, Italy, and (b) a cut-out view of the engine. (c) Experimental design of the diesel engine at the Engine Laboratory in Radom University of Technology, Poland.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

356

Asok K. Sen and Grzegorz Litak

3.2. Pressure Measurements in a Diesel Engine

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The experiments on the diesel engine were carried out in the Engine Laboratory at the Radom University of Technology, Poland (Figure 1c). A three-cylinder diesel engine was

Number of Engine Cycles Figure 2. Time series of cycle-to-cycle variations of maximum pressure measured in the sparkignition engine for loads: F = 0, 10, 20, 28, 40, 43 Nm . Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Dynamics of Pressure Fluctuations in Internal Combustion Engines

357

used and the in-cylinder pressure was measured in one of the three cylinders. The engine was fueled by standard diesel fuel. Analogous to the pressure measurement procedure used in the SI engine, the in-cylinder pressure was measured, using a piezoelectric pressure sensor. From the sensor the signal was transferred to a charge amplifier, and then digitized and stored in a computer with a data acquisition card. The loading of the engine was controlled by an eddy current brake coupled to the crankshaft. Pressure data were collected over 978 cycles for six different rotational speeds of the crankshaft: N = 1000, 1200, 1400, 1600, 1800 and 2000 rpm [11,12]. Measurements were made with a sampling frequency of 1024 times per combustion cycle. The time series of the maximum pressure variations are depicted in Figure 5. We analyzed these time series using a continuous wavelet transform (CWT) with a Morlet wavelet of order 6 as a mother wavelet.

4. Wavelet Results for Maximum Pressure Variations

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4.1. The Spark-Ignition Engine Figure 3a shows a time-period representation of the wavelet power spectrum (WPS) of the pressure signal in the SI engine for zero loading. In this and all other figures depicting WPS, the thick contour lines enclose regions with greater than 95% confidence with respect to a red-noise spectrum, whereas the region below the thin U-shaped curve denotes the cone of influence (COI). As indicated before, the results inside the COI may be unreliable and should be used with caution. The colorbar indicating the power levels is shown on the right side of each figure. The various periodicities and the number of cycles over which they persist can be readily observed in Figure 3a. For example, there are periodic bands bordering the cone of influence (COI). Among them, the 165-265 cycle band has the strongest power spanning over 450 engine cycles from approximately 225 to 675 cycles. A weaker periodicity is observed around the 58-cycle period; this band persists from approximately 370 to 460 cycles. Another periodic band is seen around the 11-cycle period and it persists between 420 to 440 cycles. Note however that these periodicities persist with less than two full oscillations. In addition, there are short-term periodicities in the 2-8 cycle band that appear intermittently. A time-period representation of the WPS for a load of F = 10 Nm is shown in Figure 3b. Similar to the scenario in Figure 3a (with no load), the strongest periodicity is now confined to the 185-265 cycle band. This band persists between approximately 300 and 650 engine cycles. Regions with weaker power extending from the COI are also seen in this figure; these regions indicate the presence of frequency modulation. Furthermore, the pressure variations exhibit short-term intermittent oscillations in the 2-6 cycle band. The results for F = 20 Nm are illustrated in Figure 3c. A distinct 75-120 cycle band spanning approximately from 140 to 375 cycles is visible in this figure. In addition, lowfrequency bands extending from the COI are present. We also observe several short-term periodicities and high-frequency intermittent patterns. The scalogram for a torque of 28 Nm is depicted in Figure 3d. This figure reveals that as the torque is increased from 20 Nm to 28 Nm, the periodicities extending from the COI have almost completely disappeared. We now observe a 28-40 cycle band over the span of 815 to

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

358

Asok K. Sen and Grzegorz Litak

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

895 cycles. In addition, there are a few short-term periodicities in the 8-16 cycle band and several even shorter-term periodicities that are intermittent. Figure 3e presents the results for F = 40 Nm. The two prominent periodicities in this figure are: (a) the 95-175 cycle band occurring between 250 and 640 cycles, and (b) the 60-80 cycle band persisting from 135 to 300 cycles. In addition, a few short-term periodicities as well as high-frequency intermittent oscillations are seen in this fugure. Finally, the results for F = 43 Nm are given in Figure 3f. The following periodicities are apparent in this figure. Bordering the COI, there are two periodic bands: (a) the 120-210 cycle band, extending approximately between 410 and 750 cycles, and (b) the 37-70 cycle band spanning 790 to 890 cycles. In addition, there are: (a) the 45-70 cycle band from 570 to 690 cycles, (b) the 13-25 cycle band between 815 and 880 cycles, and (c) the 15-24 cycle band spanning 50 to 90 cycles. There are also several short-term periodicities appearing in an intermittent fashion.

(a)

F = 0 Nm

(b)

F = 10 Nm

Figure 3. Continued on next page. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Dynamics of Pressure Fluctuations in Internal Combustion Engines

(c)

F = 20 Nm

(d)

F = 28 Nm

(e)

F = 40 Nm

Figure 3. Continued on next page. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

359

360

Asok K. Sen and Grzegorz Litak

(f)

F = 43 Nm

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. Wavelet power spectra of the time series of the peak pressure signal in the SI engine for different loads: (a) F = 0 Nm, (b) F = 10 Nm, (c) F = 20 Nm, (d) F = 28 Nm, (e) F = 40 Nm, and (f) F = 43 Nm. In these figures, the thick contour lines represent the 95% confidence level and the region below the thin U-shaped curve denotes the cone of influence.

Figure 4. Scale-averaged wavelet power (SAWP) of the time series of the pressure signal for F = 20 Nm shown in Figure 2: (a) averaged over the 2-8 cycle band, and (b) averaged over the 70-120 cycle band .

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Dynamics of Pressure Fluctuations in Internal Combustion Engines

361

It was mentioned in Sec. 2 that in addition to a wavelet power spectrum (WPS), useful information about the behavior of a time series can be obtained from the scale-averaged wavelet power (SAWP). To illustrate the use of SAWP, we consider the time series of maximum pressure variations with a torque of 20 Nm (see Figure 2). For this time series, Figure 4a shows the variation of SAWP in the 2-8 cycle band revealing the intermittent nature of the pressure fluctuations. To put things in perspective, we have also plotted the scaleaveraged wavelet power for the 75-120 cycle band. This is displayed in Figure 4b. It is clear from this figure that in this periodic band, the SAWP has a peak around the cycle 275, and the average variance is very low between approximately 400 and 700 cycles, consistent with the scalogram shown in Figure 3c. It is apparent from the results obtained above that depending on the load, the dynamics of cycle-to-cycle pressure fluctuations in an SI engine may evolve on multiple timescales.

4.2. The Diesel Engine

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

As mentioned in the Introduction, we have also examined the cyclic peak pressure fluctuations in a diesel engine. Figure 5 depicts the time series of the pressure signal for six different speeds of the crankshaft: N = 1000, 1200, 1400, 1600, 1800 and 2000 rpm. The wavelet power spectra (WPS) of these time series are shown in Figures 6a–f. They are obtained by applying a continuous wavelet transform to the time series with a Morlet wavlet of order 6 as the mother wavelet. Several characteristic features about the dynamics of pressure fluctuations can be discerned from these figures.

Figure 5. Time series of cycle-to-cycle variations of cyclic peak pressure measured in the diesel engine for six different speeds of the crankshaft: N = 1000, 1200, 1400, 1600, 1800 and 2000 rpm (starting from the bottom) .

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

362

Asok K. Sen and Grzegorz Litak

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

(a) N = 1000 rpm

(b) N = 1200 rpm

(c) N = 1400 rpm Figure 6. Continued on next page.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Dynamics of Pressure Fluctuations in Internal Combustion Engines

363

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

(d) N = 1600 rpm

(e) N = 1800 rpm

(f) N = 2000 rpm Figure 6. Wavelet power spectra of the peak pressure time series in the diesel engine for crankshaft speeds: N = 1000, 1200, 1400, 1600, 1800 and 2000 rpm.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

364

Asok K. Sen and Grzegorz Litak

For instance, at low speeds (N = 1000, 1200, 1400 and 1600 rpm), no true periodicities are detected at the 95% confidence level. When the crankshaft speed is sufficiently high (i.e., N = 1800 and 2000 rpm), strong periodic bands appear and they persist continuously over many engine cycles. In particular, for N = 1800 rpm, it is seen from Figure 6e that the strong periodic band exhibits signatures of frequency modulation. This band persists from approximately 250 to 850 cycles. At N = 2000 rpm (see Figure 6f), there is a periodic band around the 90-cycle period spanning approximately 100 to 850 cycles. In addition, the various wavelet power spectra reveal weak oscillations of much shorter periods that appear intermittently.

5. Concluding Remarks We have described the use of wavelet analysis for detecting the periodicities in cycle-tocycle pressure fluctuations in spark-ignition and diesel engines. It is found that depending on the load, the cyclic pressure fluctuations can occur on multiple timescales. Generally, the fluctuations are better developed in spark-ignition engines than in diesel engines (see Figures 2-6). This is partially due to the misfire phenomenon present in spark-ignition engines but absent in diesel engines. An advantage of wavelet analysis is that it can delineate the dominant spectral modes as well as the number of engine cycles over which each mode persists. Knowledge of these periodicities can be useful for developing new effective control strategies [14,15] for efficient combustion, which may reduce the amount of harmful exhaust gas emissions into the atmosphere and contribute to a cleaner environment.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Acknowledgements We would like to thank R. Taccani, K. Gorski, and R. Longwic for providing us with the photos of the experimental designs (Figure 1).

References [1] [2] [3] [4] [5]

Heywood, J.B. Internal Combustion Engine Fundamentals. McGraw-Hill, New York, 1988. Leonhardt, S., Muller, N., Isermann, R. Methods for engine supervision based on cylinder pressure information. IEEE/ASME Transactions on Mechatronics 4 (1999) 425-435. Daw, C.S., Finney, C.E.A., Kennel, M.B., Connelly F.T. Observing and modeling nonlinear dynamics in an internal combustion engine. Physical Review E 57 (1998) 2811-2819. Daily, J.W. Cycle-to-cycle variations: a chaotic process? Combustion Science and Technology 57 (1988) 149-162. Foakes, A.P., Pollard, D.C. Investigation of a chaotic mechanism for cycle-to-cycle variations. Combustion Science and Technology 90 (1993) 281-287.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Dynamics of Pressure Fluctuations in Internal Combustion Engines [6] [7] [8] [9] [10] [11] [12] [13] [14]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[15]

[16] [17] [18] [19] [20] [21] [22] [23]

365

Wendeker, M., Czarnigowski, J., Litak, G., Szabelski, K. Chaotic Combustion in spark ignition engines. Chaos, Solitons and Fractals 18 (2004) 805-808. Wendeker, M., Litak, G., Czarnigowski, J., Szabelski, K. Nonperiodic oscillations of pressure in a spark ignition engine. International Journal of Bifurcation and. Chaos 14 (2004) 1801-1806. Roberts, J.B., Peyton-Jones, J.C., Landsborough, K.J. Cylinder pressure variations as a stochastic process. SAE Paper 970059 (1997). Kaminski, J., Wendeker, M., Urbanowicz, K., Litak, G. Combustion process in a spark ignition engine: dynamics and noise level estimation. Chaos 14 (2004) 401-406. Sen, A.K., Litak, G., Taccani, R., Radu, R. Wavelet analysis of cycle-to-cycle pressure variations in an internal combustion engine. Chaos, Solitons and Fractals, 38 (2008) 886-893. Sen, A.K., Longwic, R., Litak, G., Gorski, K. Analysis of cycle-to-cycle pressure oscillations in a diesel engine. Mechanical Systems and Signal Processing 22 (2008) 362–373. Longwic, R., Litak, G. Sen, A.K. Recurrence plots for diesel engine variability tests. Zeitschrift fuer Naturforschung A 64 (2009) 96-102. Orianto, M., Latorre, R., Charnews, D. Experimental study of diesel engine cycle-tocycle variability. Journal of Ship Research 37 (1992) 273-279. Matsumoto, K., Tsuda, I., Hosoi, Y. Controlling engine system: a low–dimensional dynamics in a spark ignition engine of a motorcycle. Zeitschrift fuer Naturforschung A 62 (2007) 587-595. Kaul, B.C., Vance, J.B., Drallmeier, J.A., Sarangapani, J. A method for predicting performance improvements with effective cycle-to-cycle control for highly dilute sparkignition engine combustion. Proceedings of the Instiution of Mechanical Engineers Part D: Journal of Automobile Engineering 223 (2009) 423-432. Kumar, P., Foufoula-Georgiou, E. Wavelet analysis for geophysical applications. Reviews in Geophysics 35 (1997) 385-412. Gencay, R., Selcuk, F., Whitcher, B. An Introduction to Wavelets and Other Filtering Methods in Finance and Economics. Academic Press, London, 2001. Ozguc, A., Atac, T., Rybak, J. Temporal variability of the flare index (1966- 2001) Solar Physics 214 (2003) 375-397. Chainais, P., Abry, P., Pinton, J.-F. Intermittency and coherent structures in a swirling flow: A wavelet analysis of pressure and velocity measurements Physics of Fluids 11 (1999) 3524-3539. Sen, A.K., Dostrovsky, J.O. Evidence of intermittency in the local field potentials recorded from patients with Parkinson’s disease: A wavelet-based approach. Computational and Mathematical Methods in Medicine 8 (2007) 165-171. Sen, A.K. Spectral-temporal characterization of riverflow variability in England and Wales for the period 1865-2002. Hydrological Processes 23 (2009) 1147-1157. Sen, A.K., Filippelli, G., Flores, J.A. An application of wavelet analysis to paleoproductivity records in the Southern Ocean. Computers and Geosciences 35 (2009) 1445-1450. Torrence, C., Compo, G.P. A practical guide to wavelet analysis. Bulletin of the American Meteorological Society 79 (1998) 61-78.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 367-378

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 14

SYNGAS PRODUCTION BY PARTIAL OXIDATION USING A COMPRESSION IGNITION ENGINE Young Nam Chun∗ Department of Environmental Engineering, Chosun University, Seosuk-dong, Gwangju, 501-759, R. Korea

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract It is essential to develop the environment-friendly alternative energies urgently considering the limited fossil fuel and the global warming caused by environmental destruction. In this study the new technology was studied to produce syngas from methane with a compression ignition engine. This experiment was conducted on syngas production according to the variations of oxygen/methane ratio, total flow rate, air intake temperature and oxygen enrichment, with a commercial diesel engine without modification as a reformer with partial oxidation. Results showed that the concentration of hydrogen and carbon monoxide was 20.84% and 13.36%, respectively, under the optimal standard condition of oxygen/methane ratio 0.26, total flow rate 106.5 L/min and intake preheating temperature 355 . Under the same condition, the concentration of hydrogen became 20.31% when the oxygen enrichment ratio was 55.6%, while that of carbon monoxide became 20.85% when the oxygen enrichment ratio was 50.33%.

Keywords: Synthesis Gas, Hydrogen Production, Reforming, Biogas, Compression Ignition Engine.

Introduction Interest in low-pollution alternative energies have been increasing gradually, as the environmental problems, such as the global warming caused by the use of fossil fuels, have been getting more and more serious. One of the technologies that are deemed realistic at present is the production of synthesis gas using methane, which is the principal ingredient of ∗

E-mail address: [email protected]

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

368

Young Nam Chun

natural gas and biogas. Methane, in particular, constitutes some 50-70% the biogas that comes from the landfills and the anaerobic digester of the wastewater treatment plants. If it could be utilized as energy source for the production of synthesis gas, it will be an eco-friendly technology that enables not only the treatment of methane gas that has 20 times higher Global Warming Potential (GWP) than carbon dioxide but also the production of pollution-free energy. So far, two methods have largely been applied to the production of synthesis gas based on methane: the steam reforming and the partial oxidation reforming. Currently, the steam reforming[1,2] is used most widely thanks to its strong points such as the high throughput of gas processing and the high ratio of synthesis gas production. But the method also has some shortcomings: it requires exterior energy source such as high temperature and high pressure and it should be huge in size. On the other hand, the partial oxidation reforming[3,4] is a method using the incomplete combustion of fuels and, therefore, the whole reaction is operated through exothermic reaction. But the method also has technological difficulty since it requires low oxygen/fuel ratio to maintain the high yield of synthesis gas and concentration. While the partial oxidation reforming shows a high efficiency compared with the other existing methods and has benefits in terms of energy costs because it uses its own generation of heat, the method also faces difficulty due to narrow combustibility range in case of the biogas which has low caloric value. In order to make up for the weak points, the superadiabatic combustion reforming[5] was suggested for further researches. Still, in this case, the regenerator should also be used for the method. It is also considered inappropriate for mass production. In this study, therefore, a new form of compression ignition reforming reactor was devised with the utilization of a super-adiabatic compression reforming so as to resolve the abovementioned problems of the existing methods. It was proved through experiments that the proposed reformer is an efficient reactor, which does not use catalyst, enables partial oxidation and guarantee high yield of synthesis gas and high throughput. Beside, this study also presented the optimal condition for the operation of the proposed reformer through experiment by variables that affect the production of synthesis gas.

Experimental Experimental Apparatus Figure 1 shows the experimental apparatus used in this study. It is made up of the gas supply line, the compression ignition reformer and the measurement and analysis system. The gas supply line is composed of the methane supply line, oxygen supply line and air supply line. Methane is supplied through the Compressed Natural Gas (CNG) cylinder filled up with a high pressure of 22MPa (224.3kgf/cm2) and is injected through the mixer via a regulator, a flowmeter (Dwyer, USA) and a surge tank (7.5ℓ) for mixture with air. The oxygen supply line is composed of an oxygen cylinder, a regulator and a flowmeter. The air supply line is made up of an orifice flowmeter (KFE, Korea), a surge tank (19ℓ), a diaframe with a 10mm orifice diameter, a safety valve and 6kW electric heater. A mixer from an LPG motor vehicle was used for the mixture of methane.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Syngas Production by Partial Oxidation Using a Compression Ignition Engine

369

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A commercial diesel engine was used for the compression ignition reformer and its technical data is shown in Table 1. The measuring system is composed of the measurement of temperature and the measurement of the engine’s rotational frequency. For the measurement of temperature, a thermocouple (K-type, OD:6mm) was installed onto the intake manifold and exhaust manifold and a data logger (FLUKE, USA) was used for temperature monitoring.

Figure 1. Experimental apparatus of compression ignition reformer.

Table 1. Technical data of compression ignition reforming reactor Item Model Type Fuel injection Cylinder number Bore(mm) and Stroke(mm) Compression ratio Displacement(cc) Max Power (PS/rpm) Rated

Specification Daedong, ND130DIE Horizontal water-cooled 4 cycle diesel engine Direct injection 1 95 × 95 18:1 673 13/2400 10/2200

A magnetic-type temperature control device was also made and used to control the temperature of the heater. For the measurement of the rotational frequency, a tachometer (HIOKI, Japan) was installed onto the engine poly. As for the exhaust gas analysis system, a sampling probe was inserted into the exhaust line at the point of 390mm away from the exhaust valve. Exhaust gas was inhaled through a vacuum pump (Gast Manufacturing Inc.,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

370

Young Nam Chun

USA) and then passed through the impinger so that soot and moisture could be got rid of. After that the waste gas was analyzed by the gas chromatograph (SHIMADZU, Japan).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Experimental Method In order to secure the recurrence of the experiment, the refiner was first test-run for 15 minutes with gasoline and, only after the temperature inside the engine got stabilized, the refining fuels of methane and air were injected. In order to minimize the connection between factors that affect the yield of fuel-rich syngas, the optimal condition was worked out as a standard condition through repeated experiments. And then experiments were carried out in accordance with each of the variables, including an oxygen/methane ratio, a total flow rate, an intake preheating temperature and an oxygen-enrichment ratio. Experiments were carried out in two ways, the basic experiment and the oxygen-enrichment experiment. In the basic experiments, two variables were fixed so that the influence other variable has on the production of synthesis gas could be shown. In the oxygen-enrichment experiments, the rate of oxygen enrichment was first set at 21% and then the amount of oxygen was gradually added up. Oxygen was flowed into the flowmeter, with the pressure of the exit of the pure oxygen cylinder decompressed to 0.1MPa (1kgf/cm2). In the initial stage, the standard condition of oxygen/methane ratio was set at 0.26. But the ratio changed later in accordance with the addition of oxygen. As aforementioned, methane has a high self-ignition temperature and, therefore, a high compression ratio and the preheating of the intake temperature is required in order to operate the engine through methane[6]. Addition of oxygen promotes a faster reaction with methane and results in an increased thermal efficiency and an improved output of engine[7]. Experiments were carried out with an increasing amount of oxygen from the lowest temperature condition of the engine, which was first operated under the standard condition. Oxygen enrichment ratio was calculated by the following Eq. (1)[8].

OEC (%) =

0.21A + O2 add ×100 Qtotal

(1)

In the equation, OEC is the oxygen enrichment concentration, A is the amount of the air inflow, O2 add is the amount of added oxygen and Qtotal is the sum of the air inflow and the added oxygen. Table 2. Experimental conditions by variables and standard condition

Conditions

O2/CH4 ratio

Total gas flow rate(L/min)

Intake Oxygen temperature(℃) enrichment (%)

Experiment Range

0.22~ 0.63

59.0~ 171.4

280~ 355

21.0~ 60.53

Reference

0.26

106.5

330

55.6

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Syngas Production by Partial Oxidation Using a Compression Ignition Engine

371

In order to analyze the concentration of the reformed gas, collected by the sampling probe, a gas chromatograph was used. A TCD detector was used for the analysis, with Molecular Sieve 5A (80/100 mesh) used for hydrogen, Molecular Sieve 13X (80/100 mesh) for carbon monoxide and HayeSep R (100/120 mesh) for carbon dioxide and CmHn.

Results and Discussion Limit of Combustibility Since this study is aimed at the production of synthesis gas through the partial oxidation in the fuel-rich state, the characteristics of the engine and the range of combustibility were examined first, which are shown in Figure 2. Experimental limit zone means the area where the experiments could not be carried out due to such problems as the capacity of the heater and the durability of the engine. Fuel-rich knocking region is the area where knocking, an incomplete combustion, took place and the operation of the engine had difficulties due to excessive inflow of fuel compared with the air inflow. Knocking in this region is caused by the increased pressure due to the sudden combustion, the result of delayed ignition which was caused by the lack of oxygen because of the fuel-rich state. 10

Fuel-rich knocking region

Equivalence ratio

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

8

6

Operating region

Experimental limit zone

4

2

0 280

Lean misfire region

300

320

340

Preheating temperature (oC)

360

Figure 2. Combustibility range of compression ignition reformer.

Lean misfire region, on the other hand, is the area where normal operation of the engine was difficult due to weak generation power. It is the area where the fuel is relatively rare compared with the combustibility-operating region. A misfire was caused, as the total caloric Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

372

Young Nam Chun

value of the mixed gas decreased below the level of ignition temperature. In the area where the temperature was high, the range of combustibility of the lean mixer was somewhat expanded in tandem with the increased heat content.

Results of the Parametric Studies 1. Effect of Oxygen/Methane Ratio Figure 3 shows the concentration of the hydrogen and carbon monoxide when the oxygen/methane ratio was changed between 0.22 and 0.63 with the total flow of the mixed gas set at 106.5 L/min and the intake preheating temperature at 330 . As the oxygen/methane ratio increase, so did the concentration of hydrogen. When the oxygen/methane ratio became 0.26, it reached the peak of 15.39% and then began to fall again. A similar pattern was found for the carbon monoxide. As shown above, the concentration of hydrogen and carbon monoxide increases as the oxygen/methane ratio increases. It is because the combustibility gets enhanced thanks to the sufficient supply of oxygen, which enables the partial oxidation promoted by the decreased equivalent ratio. On the contrary, the concentration decreases after it reached the maximum point, despite the continual increase in the oxygen/methane ratio. And this is because of the excessive supply of air which causes the loss of waste heat, thereby leading to the reduced partial oxidation reaction and dilution of air.

Concentration (%)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

20

15

H2 10

CO

5

0

9.15

8.15

7.15

6.15

5.15

4.15

3.15

Equivalence ratio 0.2

0.3

0.4

0.5

0.6

0.7

O2/CH 4 ratio Figure 3. Concentrations of oxygen/methane ratio.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Syngas Production by Partial Oxidation Using a Compression Ignition Engine

373

An increase in the oxygen/methane ratio means the decrease of fuel in a state where air is fixed. When the oxygen/methane ratio was below 0.22, a normal operation became difficult due to the pre-ignition and knocking, or an abnormal combustion, caused in the fuel-rich state. In the area where the ratio got above 0.63, the lack of fuel and the exhaust loss caused misfire.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2. Effect of Total Flow Rate Figure 4 shows the results of the measurement of the concentration of reformed gas, the temperature of exhaust and the engine frequency, when the total flow was changed between 59 L/min and 171 L/min, with the oxygen/methane ratio fixed at 0.26 and the intake preheating temperature is set at 330 . Figure 4(a) indicates the concentration of hydrogen and carbon monoxide among the reformed gas. When the total gas flow stood at 106.5 L/min and 117.3 L/min, the concentration of hydrogen and carbon monoxide reached the maximum of 18.76% and 14.39%, respectively. As the total flow increases, the amount of hydrogen and carbon monoxide out of the total synthesis gas also increased, due to the increase of the amount of oxygen in tandem with the amount of injected methane. But the maximum state of concentration, although there are some differences in the flow of hydrogen and carbon monoxide, began to fall down after it reached the peak. It is because the increased amount of mixed gas causes quenching effect inside the cylinder during the intake process, forming a generally low combustion temperature and decreasing the concentration of synthesis gas. Figure 4(b) is the result of the analysis on the concentration of methane and oxygen included in the reformed gas. As the total amount of gas increased, that of methane and oxygen also increased. It is because the increased flow lowers the combustibility inside the cylinder and the reactivity of methane and oxygen within the engine. In this situation, the gas does not react well and just gets out of the engine. In particular, if the total flow is increased in a large quantity, the conversion into synthesis gas is not carried out well and the methane just comes out as it originally was, which was already mentioned in Figure 4(a). Figure 4(c) shows the relationship between the temperature of exhaust gas and the number of rotations of the engine in accordance with the change in the total gas flow. As the total flow increased, the temperature and the frequency increased to have the maximum value and then began to drop. It could be explained in the similar way of Figure 4(a), which dealt with the amount of synthesis gas production. As aforementioned, the temperature inside the combustion chamber increases in tandem with the increased flow, raising the engine frequency. But the temperature and the engine frequency decrease after reaching the peak, because the supply of a large amount of mixed gas causes quenching effect inside the cylinder. In this experiment, the reaction of methane was most active and the engine frequency was the highest in the area where the hydrogen yield was high. As seen in the experimental results, the temperature of exhaust gas and the engine frequency have a very close relationship. And this means that the combustion state of an engine could be checked out by simply checking the number of rotation of the engine.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

374

Young Nam Chun 20

15

15

10

10

5

50

100

150

CO concentration (%)

H 2 concentration (%)

20

5

Total gas flow rate (L/min) (a) Concentrations of hydrogen and carbon monoxide 30 8

20

6

15 4

10

O2 concentration (%)

CH 4 concentraion (%)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

25

2 5

0

50

100

150

0

Total gas flow rate (L/min) (b) Concentrations of oxygen and methane Figure 4. Continued on next page.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

350

1600

340

1500

330

1400

320

1300

310

1200

300

1100

290

375

rpm

Exhaust temperature (oC)

Syngas Production by Partial Oxidation Using a Compression Ignition Engine

50

100

1000 200

150

Total gas flow rate (L/min)

(c) Temperature of exhaust and engine revolution Figure 4. Results of total flow rate.

3. Effect of Intake Preheating Temperature

25

20

Concentration (%)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5 shows the concentration of hydrogen and carbon monoxide when the intake temperature is heightened to 355℃, with the oxygen/methane ratio is fixed at 0.26 and the total mixed gas flow at 106.5 L/min.

15

H2

10

5

CO

280

300

320

340

360

o

Preheating temperature ( C) Figure 5. Influence of intake preheating temperature. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

376

Young Nam Chun

This compression ignition reformer was operated normally with the preheating of the mixed gas maintained above 280℃ and the self-ignition temperature. As the preheating temperature increased, the concentration of hydrogen and carbon monoxide also increased continuously. At the maximum temperature of 355℃, the concentration of hydrogen and carbon monoxide reached the maximum of 20.84% and 14.99%, respectively. By increasing the intake preheating temperature, the concentration of the mixed gas is also increasing. It is because the total calorie increases thanks to the increase in entropy and the inside of the reactor is brought to a state of high pressure, a sufficient condition for partial oxidation.

4. Effect of Oxygen-enrichment Figure 6 shows the results of the experiment to check the yield of synthesis gas in accordance with the addition of oxygen in the reformer. Under the standard condition, the oxygen/methane ratio was fixed at 0.26, the total gas flow at 106.5 L/min and the intake temperature at 330℃. And the oxygen enrichment was raised from 21% to 60.53%. Exhaust gas fluctuation got relatively serious in accordance with the oxygen enrichment. As seen in Figure 6, there was some fluctuation in concentration of synthesis gas. But the concentration of hydrogen and oxygen increased to the maximum of 20.31% and 20.85%, respectively, in tandem with the general increase in the oxygen enrichment ratio. It is because the discharge loss decreases in accordance with the decreased amount of nitrogen, which raises temperature and improves combustibility. Temperature and the concentration of nitrogen could be checked in Figure 6 (b).

20

Concentration (%)

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

25

H2 15

CO 10

5

20

30

40

50

60

Oxygen enriched (%) (a) Concentration of hydrogen and carbon monoxide Figure 6. Continued on next page.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Syngas Production by Partial Oxidation Using a Compression Ignition Engine

377

62

60

400

58

56 350

N2concentration (%)

Exhaust temperature(oC)

450

54 300 20

30

40

50

60

52

Oxygen enriched (%) (b) Temperature and concentration of nitrogen Figure 6. Influence of oxygen enrichment.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Conclusion This study partially altered a commercial diesel combustion engine to set up a compression ignition reforming system to carry out experiments about the production of synthesis gas. Through repeated experiments, the standard condition, under which the amount of hydrogen out of the total synthesis gas reaches the maximum, was worked out. According to the experimental results, the concentration of hydrogen and carbon monoxide became 20.84% and 13.36%, respectively, when the oxygen/methane ratio was set at 0.26, the total gas flow was 106.5 L/min and the intake preheating temperature was 355 . Also, under the same condition, the concentration of hydrogen and carbon monoxide became 20.31% and 20.85%, respectively, when the oxygen enrichment ratio was about 53%. Beside, the feature of the reformer’s movement was also examined through experiments by variables that affect the yield of hydrogen. As the oxygen/methane ratio and the total gas flow increased, the concentration of the synthesis gas also increased to get the maximum value and then began to fall down. Therefore, it would be an important factor to find out the optimal condition for the abovementioned variables to get the maximum yield. Also, the synthesis gas yield increased in proportion to the increase in the intake preheating temperature and the oxygen enrichment ratio. It means that, if only the economic efficiency is secured, higher temperature and enrichment ratio would be favorable to the yield.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

378

Young Nam Chun

References [1] [2] [3] [4] [5]

[6] [7]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[8]

Beckhaus P, Heinzel A, Mathiak, and J, Roes, J of Power Sources, 127, 294(2004). Kajornsak F, Ryuji K, and Koichi E., J of Power Sources, 161, 87(2006). Lutz AE, RW Bradshaw, L Bromberg, and A Rabinovich, Int. J of Hydrogen Energy, 29, 809(2004). Barrio VL, Schaub G, Rohde M, Rabe S, Vogel F, and Cambra JF, Int. J of Hydrogen Energy, 32, 1421(2007). Yu.M. Dmitrenko, R.A. Kelvan, V.G. Mimkina, and S.A. Zhdanok, Minsk International Colloquium on Physics of Shock Waves, Detonation and Non-Equilibrium Processes, Nov. 12-17(2005). Apoorva A, and Dennis N.A., SAE International congress and exposition, Feb. 2326(1998). Assanis DN, Poola RB, Sekar R, and Catal GR, Journal of Engineering for Gas Turbines and Power, 123, 157(2001). Kwark JH, Jeon CH, and Chang YJ., The Korean Society of Mechanical Engineers, 28, 160(2004).

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 379-384

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

Chapter 15

A POSSIBLE EFFECT OF EXHAUST EMISSIONS FROM VEHICLES USING UNLEADED PETROL ON HOUSE SPARROW POPULATIONS J.D. Summers-Smith

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

United Kingdom

The House Sparrow Passer domesticus is one of the most familiar wild birds. It is common, with an immense natural range that stretches from the west coast of Ireland to the east coast of Siberia, north to the Arctic Circle and south to North Africa, the Middle East, India and Sri Lanka. In addition, it has also been introduced to America, both North and South, South Africa, Australia, New Zealand and many points between, making it one of the most widespread of all birds; furthermore it is currently expanding its range by colonizing new areas in the north of South America, West Africa and the Far East. Its success, as its name implies, is through its association with man, both in farmland and built-up areas, even extending into the centers of large towns. This association began about 10,000 years ago when man developed from a hunter/gatherer to a sedentary agriculturist, the bird taking advantage of the cereals that man developed from the large-seeded grasses, which had previously provided its diet, supplemented by the food put out for domesticated animals. Yet all is not well with this apparently most successful of birds. In 1962 The British Trust for Ornithology (BTO) initiated a project to keep a check on the state of the nation’s birds in the wider countryside. A decline in a suite of farmland birds, including the House Sparrow, was noted in 1980; it continued until the mid-1990s, when it stabilized after a decrease of about 60% and it has broadly remained at this level to the present day. This decline was investigated and it is now accepted that it was the result of the intensification of farming practices that reduced the amount of food available to sparrows and other farmland species (Chamberlain et al 2000). The much more significant population of House Sparrows living in towns and villages was not monitored by the BTO project and it was not until the 1990s that it was realized that a serious decline of the bird was occurring in the centers of London and other large towns in

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

380

J.D. Summers-Smith

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Britain. Interestingly, the evidence for this came, not from ornithologists and bird-watchers, but from the ordinary town-dwellers who in Britain hold their town sparrows in great affection. This led to reports in the press and even questions in the UK Parliament about what was going on and what the Government was going to do about it! Figure 1 gives a plot of House Sparrow densities as a function of date for a range of towns in northwest Europe (Summers-Smith 2003). The plot has to be viewed with some caution as, apart from the Kensington Garden counts, there are almost no repeats and the counts have been made by different observers using different protocols, but even with this caveat it appeared that the urban decline was taking place, not only in Britain, but also in Ireland, Germany, the Netherlands and Belgium, and that the decline in all these countries began about 1990-91, suggesting a common environmental factor. Further investigation showed that the urban decline was quite separate from the earlier farmland one – the observation of color-ringed (banded) birds and recoveries of ringed birds in any case suggest that there is little interchange between the farmland birds and those occurring in urban areas. Moreover, the urban populations were not stabilizing at a lower level, but were inexorably proceeding to complete extinction. A number of possible causes have been proposed: predation by cats and birds of prey, lack of availability of nesting sites in new buildings and improved maintenance of older properties, competition with other birds occupying the urban habitat (gulls, crows, pigeons), atmospheric pollution, increasing use of herbicides and pesticides reducing the availability of food for sparrows and, more recently, electromagnetic radiation from GSM masts and WiFi transmitters, and light pollution (see below). A notable feature of the urban decline is its apparent coincidence in timing over a large area of northwest Europe. This decline, however, has been by no means uniform; for example it has not occurred in Berlin and Paris. Again, populations in villages, small towns and the outer suburbs of large towns were not affected, or at least not to the extent of that in the centers of large towns.

Figure. 1. House Sparrow densities in built-up areas. The lines are the linear regressions of the Kensington Gardens counts for 1945-1975 and 1995-2002. The large town center results (London, Edinburgh, Dublin, Hamburg) show a reasonable fit to the regression lines, but those for the small rural towns do not show a dramatic decline after 1990.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A Possible Effect of Exhaust Emissions from Vehicles Using Unleaded Petrol…

381

The reason for the urban center decline has not yet been resolved despite increasing research. In a study in Hamburg, Germany, in 1997, Bower (1999) found that none of the broods that were started in April were successful. House Sparrows depend on animal food for the rearing of their nestlings and Bower suggested that a lack of insects at that time was responsible. This was supported by a PhD study by Kate Vincent in Leicester, England (Vincent 2005). In the breeding seasons of 2002 and 2003, Vincent found complete broods of young had died of starvation and there was a good correlation between the availability of aphids, an important food for young nestlings, and fledging success. Moreover, a shortage of invertebrates could put the parents at hazard through increased exposure to predators while foraging and the need to forage in more risky situations. Earlier studies have shown that, in contrast to most species, the adult mortality of House Sparrows was already at a maximum during the breeding season (Summers-Smith 1959, Will 1973, Heij 1985). Vincent also found that those nestlings that did survive tended to be underweight at fledging and thus were more likely to suffer a higher rate of mortality in the crucial post-fledging period and hence a reduced chance of successfully recruiting into the population. In 1922, the American Thomas Midgely discovered that the addition of tetraethyl lead (TEL) to petrol (gasoline) enhanced the anti-knock properties giving improved combustion and hence reduced fuel consumption in engines. It was adopted universally, but in the 1980s, following the finding that lead was responsible for brain damage in growing children, it was suggested that the exhaust emissions from cars were a major source of lead and the petroleum companies were pressurized into finding an alternative. They came up with an ether – methyl tertiary-butyl ether (MTBE). MTBE is a carcinogen and was progressively phased out as an additive to petrol in the EU in the early years of this century. However, as it is not easy to produce petrol with an RON95 (Research Octane Number), the standard for normal unleaded petrol, by simple fractionation of petroleum, it is likely that MTBE has been replaced by a similar additive that could have the same effect. A hypothesis that exhaust emissions from the engines of vehicles running on unleaded petrol might be responsible for the decline of urban sparrows was put forward in 2000 (Summers-Smith quoted by McCarthy 2000b). This is based on the following circumstantial evidence: •



The temporal coincidence between the introduction of unleaded petrol in the EU in 1989 and the onset shortly afterwards of the collapse of urban House Sparrow populations in northwest Europe. Work by Peter Joseph, University of Pennsylvania, USA, where he was investigating the increased incidence of asthma in children and cats, showed the presence of highly toxic vapors (methyl nitrite, CH3ONO, and a range of compounds with peroxy radicals, -OOH) in the exhaust emissions from engines running on unleaded petrol (Joseph 1999); CH3ONO is highly toxic, having a lethal concentration, LC50, in tests with rats Rattus of 170 ppm, making it 100 times more toxic than benzene. (The LC50 value is the concentration of the vapor in air that will kill 50% of the test animals after an exposure of 4 hours.) The compounds involved are unstable, being quickly decomposed by exposure to UV light, but nevertheless they could persist long enough in the hours of darkness or in conditions of heavy overcast to reduce the populations of invertebrates on which House Sparrows depend to rear their young.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

382

J.D. Summers-Smith





Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.





Moreover, under conditions of heavy traffic the toxic materials will be continuously replenished day and night in large towns. Although House Sparrows have declined almost to extinction in the centers of London, Edinburgh, Dublin, Hamburg, Amsterdam, Rotterdam, Antwerp, Ghent and Brussels (De Laet and Summers-Smith 2007), they are still common in Paris (McCarthy 2000a). This could be explained by the fact that France has the highest proportion of diesel-engined cars in the EU; for example, in 1998, only 12% of cars in Britain had diesel engines, compared with 33% in France (European Environment Agency 2003). A high incidence of animal deaths (with reports of more than 17,000 bird deaths, including House Sparrows) occurred in New York, USA, in the summer of 1999. The NY State authorities attributed the deaths to an outbreak of West Nile Disease (WND), a mosquito-borne flavivirus that had not previously been reported in the USA. However, an independent eco-consultant, Jim West (2002), showed that the pattern of deaths correlated strongly with traffic density and he suggested that WND had affected organisms already weakened by exposure to vehicle exhaust emissions. The finding some time later that House Sparrows, although carriers of the WND virus, are not susceptible to the disease (Wingerson 2000), adds some support to West’s suggestion. Moreover, although WND was not detected in the USA until this occurrence, there is no evidence that it had not been present prior to the incident in NY in 1999. In 1998, I carried out a breeding season census in 120 ha in the center of the small town of Guisborough, NE England (population 18,000, area 541 ha). The density of House Sparrows was 5.3 individuals/ha. A repeat census in 2006 gave a similar density, except for an area of 5 ha centered on the only traffic lights in the town center, which were erected in 1998 after I had carried out my first census. Three major roads meet at this controlled junction, resulting in queues of cars with idling engines in two of the roads. In 1998, there were 11 breeding pairs in the area round the lights, in 2006 there was only one. Moreover, a traditional winter social singing site, 12 m from the lights, was abandoned by the sparrows. Peach et al. (2008) in a further analysis of Vincent’s data (Vincent 2005) found that atmospheric pollution (as indexed by NO2 content, largely the product of vehicle exhaust emissions) was the only significant predictor of House Sparrow reproductive failure among a range of variables monitored (low temperature, extremes of rainfall, lack of deciduous vegetation and grass, low aphid densities – an important component of chick diet, level of vegetable material in the diet)

The hypothesis has been criticized on the grounds that other small bird species living in the urban environment have not been affected. This difference, however, may be more apparent than real. House Sparrows are extremely sedentary, the vast majority living out their lives within a compass of 1-2 km. The young of other birds that breed in towns tend to disperse and would thus be able to top up the urban ‘sink’ from ‘sources’ outside the town (Summers-Smith 2005). Town Pigeons Columba livia are still common in towns, but they rear their young on vegetable matter, not on invertebrates.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

A Possible Effect of Exhaust Emissions from Vehicles Using Unleaded Petrol…

383

It has to be emphasized that the above observations are purely circumstantial and not evidence of a causal relationship. The chemistry of combustion is very complex and the situation is confounded by the varying use of MTBE in different countries and at differing times of the year. Further, as mentioned above, it has now been replaced by a noncarcinogenic material, but it is likely that such an additive may well produce similar toxic combustion products. This is not to suggest that unleaded petrol is the only factor in the decline of urban invertebrates. Other possible causes, in addition to increased usage of pesticides and herbicides, that could have an adverse impact on invertebrate populations have been suggested: electromagnetic radiation (Balmori and Hallberg 2007, Evereart and Bauwens 2007), light pollution (Henshaw, 1994), but while these may well be contributory factors, they do not have the same temporal coincidence with the onset of the urban center House Sparrow decline as does the introduction of unleaded petrol. A research project looking at the effect of atmospheric pollution, including exhaust from unleaded petrol, was begun by the Royal Society for the Protection of Birds in England in 2008. Most of us now live in towns. We might ask if the House Sparrow is today’s equivalent of the miners’ canary giving us warning of some impending hazard in our towns. Birds are one of the UK ‘Quality of Life Indicators’ (DETR 1999). What does the sudden decline of House Sparrows tell us about the quality of life in our cities?

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

References Balmori, A. and Hallberg, Ö. 2007. The urban decline of the House Sparrow (Passer domesticus): a possible link with electromagnetic radiation. Electromagn. Biol. Med. 26: 141-151. Bower, S. 1999. Fortpflanzungsaktivität, Habitatnutzung und Populationsstruktur eines Schwarms von Haussperlingen (Passer d. domesticus) in Hamburger Stadtgebeit. Hamburger avifaun. Beitr. 30: 90-129. Chamberlain, D. E., Fuller, R. J. Bunce, R. G. H., Duckworth, J. C. and Shrubb, M. 2000. Changes in abundance of farmland birds in relation to the timing of agricultural intensification in England and Wales. J. Appl. Ecol. 37: 771-788. De Laet, J. and Summers-Smith, J. D. 2007. The status of the urban house sparrow Passer domesticus in north-western Europe: a review. J. Ornithol. 148 (Suppl. 2): 257-278. DETR 1999. UK Government Department of the Environment, Transport and the Regions White Paper ‘A Better Quality of Life’: strategy for sustainable development for the United Kingdom. The Stationary Office: London. European Environment Agency. 2003. Term 2002 32 EU – Size and composition of the vehicle fleet, 1st June 2003. http://themes.eea/europa.eu. Evereart, J. and Bauwens, D. 2007. A possible effect of electromagnetic radiation from mobile phone base stations on the number of breeding House Sparrows Passer domesticus. Electromagn. Biol. Med. 26: 63-72. Heij, C. J. 1985. Comparable ecology of the House Sparrow Passer domesticus in rural, suburban and urban situations. PhD Thesis, University of Amsterdam.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

384

J.D. Summers-Smith

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Henshaw, C. 1994. The environmental effects of light pollution. J. Brit. Astronomical Ass. 104: 3. Joseph, P. M. 1999. New hypotheses for MTBE combustion products. Univ. Pennsylvania, USA. www.osti.gov/em52/eprints/99885.pdf. McCarthy, M. 2000a. Where have all the sparrows gone. The Independent 18.03 00. McCarthy, M. 2000b. Lead-free petrol may be the villain in the mystery of demise of the world’s most familiar bird. The Independent 11.09.00. Peach, W., Vincent, K., Fowler, J. and Grice, P. (2008). Reproductive success in urban House Sparrows along an urban gradient. Animal Conservation 11: 493-503. Summers-Smith, J. D. 1959. The House Sparrow Passer domesticus: population problems. Ibis 101: 449-454. Summers-Smith, J. D. 2003. The decline of the House Sparrow: a review. Brit. Birds 96: 439446. Summers-Smith, J. D. 2005. Changes in the House Sparrow population in Britain. Int. Studies on Sparrows 30: 23-27. Vincent, K. E. 2005. Investigating the causes of the decline of the urban House Sparrow Passer domesticus population in Britain. PhD Thesis, De Montfort University, Leicester. West, J. 2002. West Nile Virus positives and MTBE analysis of NYSDEC Wildlife Pathology Unit’s West Nile Disease database 1999. http://findarticles.com/p/articles/ mi_mOISW/is_2002_July/ai_87719992. Will, R, L. Breeding success, numbers and movements of House Sparrows at McLeansboro, Illinois. In: Hardy, J W. and Morton, M. (eds.). A symposium on the House Sparrow (Passer domesticus) in North America. Orn. Monograph No. 14 AOU. Wingerson, L. 2000. Report on Meeting of the Soc. Trop. Med. and Hygiene. BioMedNat, 1st November 2000.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

SHORT COMMUNICATIONS

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 387-394

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

BENZENE EXPOSURE AND THE STORY OF CARCINOGENESIS: EXPERIENCE OF TRAFFIC POLICEMEN IN BANGKOK Viroj Wiwanitkit Wiwanitkit House, Bangkhae, Bangkok Thailand 10160

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Benzene is of particular concern because recent research indicates that benzene exposure can result in chronic toxicity including induction of hematological cancer. Benzene exposure and the story of carcinogenesis is a present focus in occupational health. In this article, the author will present and comment on experience involving traffic policemen in Bangkok. Data on chromosome and flow cytometry study with correlation to biomarker monitoring will be reported and presented in this article.

Introduction Benzene is of particular concern because recent research indicates that benzene exposure can result in chronic toxicity including induction of hematological cancer [1]. At present, work with benzene is subject to the Control of Substances Hazardous to Health (COSHH) Regulations 1999. Apart from the industrial workers, there are other occupations with high risk for benzene exposure. Police belong to another occupation at risk for benzene exposure. Crebelli et al. studied the police in the city of Rome in Italy and found that the exposure to traffic fumes during working activities might give a relatively greater contribution to general personal exposure to benzene than indoor sources [2]. In this article, the author will present and comment on experience involving traffic policemen in Bangkok. Data on a chromosome and flow cytometry study with correlation to biomarker monitoring will be reported and presented in this article.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

388

Viroj Wiwanitkit

Biomonitoring of Exposure to Traffic Benzene Vapor: Analysis of Sister Chromatid Exchange and Urine Biomarker, Trans, Trans - Muconic Acid of Traffic Policemen 1. Background According to a recent study of Wiwanitkit et al., the Thai police were highly exposed to traffic benzene vapor [3]. Wiwanitkit et al. also noted the possible leukemogenesis from exposure to the benzene vapor for the police [4]. To fulfill the previous report, this study was set aiming at determination for sister chromatid exchange (SCE), a marker for genotoxicity, and its correlation to the novel urine biomarker for benzene exposure urine trans, trans muconic acid (tt-MA ).

2. Materials and Methods

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

2.1. Subjects Thirty-nine healthy volunteer male police were included in this study. These police who had to work daily as traffic police in an urban area, namely Pathumwon District, Bangkok. All subjects gave informed consent. The Faculty of Medicine and Faculty of Allied Health Science, Chulalongkorn University, approved the study. All subjects were healthy and had worked for at least 2 years. All have the same lifestyles, as traffic policemen, and have similar diets. We also performed an additional investigation in 20 healthy male subjects, and used a group of students studying at Chulalongkorn University to serve as the control. These students lived in the same area, Pathumwan District, but in a dormitory about 0.5 kilometers away from roads. All subjects in this study presented the same eating and drinking habits. All subjects were healthy adults. Before the study, all were interviewed for possible exposure to other sources of non-traffic vapors, especially for smoking and volatile substances. The exclusion was set in the case of a history of possible exposure to those vapors.

2.2. Laboratory Analysis 2.2.1. Specimen Collection Each subject gave a urine specimen and a blood sample for further laboratory analysis. The work schedule of all subjects is 8.00–17.00 o’clock (official time of Thailand), 5 days per week for every month of the year. The time of sampling is in the evening, 17.00 o’clock, after the subjects finished their daily work. All samplings were performed at a same period on the same day in August 2002. 2.2.2. Laboratory Analysis for the Urine Ttma Level Each subject provided a urine sample and a blood sample for laboratory analysis. These subjects were included for a previous study on the benzene contact and confirmed for the high

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Benzene Exposure and the Story of Carcinogenesis

389

exposure of benzene, having high urine ttMA, during their daily work [3]. Concerning the laboratory analysis for urine ttMA, high-performance liquid chromatography method as described in our recent previous study was used [5]. 2.2.3. Determination for SCE A heparinized blood sample from each subject was collected for further determination for SCE. The laboratory analysis was performed using the modified Woff method [6]. For each case, 1,000 cells were scored for SCE. All laboratory analysis was performed at the Department of Clinical Microscopy, Faculty of Allied Health Science under standard laboratory quality control process.

2.3. Statistical Analysis The statistical analysis of the results was carried by SPSS 7.0 for Windows Program. For assessment of the correlation between SCE and urine ttMA, the regression analysis was used. Statistical significant difference was accepted at P value < 0.05 similar to the other medical studies.

3. Results

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A total of 39 police (all males) were included in this study. The reported average (mean + SD) urine tt-MA level in these police was 0.79 + 1.43 mg/gCr while the average level in the controls was under the detection limit. The average (mean + SD) SCE in these police was 4.21 + 0.86 (/cell).

Figure 1. The correlation between urine ttMA and SCE among the traffic police. Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

390

Viroj Wiwanitkit

The average SCE among the police is higher than the average SCE of Thai controls (0.24 + 0.12 /cell). Concerning the regression analysis, the least square equation plot SCE (Y) versus urine ttMA (X) is 0.30 X + 0.18 (r = 0.93, p = 0.02).

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4. Discussion Benzene is a common toxic volatile substance, found in many industrial processes in the present day [1]. International organizations such as Agency for Toxic Substances and Disease Registry (ATSDR) [7] have documented benzene toxicity and recommend the monitoring of benzene exposure for groups at risk. The toxicity includes genotoxicity, neurotoxicity and hematotoxicity. It is classified as a carcinogen and can cause serious health problems. In exposure- and risk- evaluation, the monitoring of benzene by peripheral biomarker has several advantages over technical assessment of exposure. For this purpose, several biomarkers for benzene have been proposed. Of those biomarkers, the urine trans, trans– muconic acid (ttMA) level is accepted as a useful monitoring tool for early diagnosis of dangerous exposure [8-9]. The urine ttMA level is a novel peripheral biomarker for benzene exposure. It is more specific than other biomarkers and has no interference from ingested food, better than a standard biomarker, urine phenol [9]. Ong et al. also noted that ttMA correlated best with environmental benzene concentration, better than other biomarkers [10]. At present, urine ttMA level has been validated as an effective test in monitoring for benzene exposure [5, 9, 11]. High benzene exposure among the police working in the road is noted by Leong et al. [12]. The study of Verma et al. among Indian police also presented similar results [13]. Therefore, working in the air pollution in the urban area can be a health hazard for the police. Exposure to the benzene from automobile exhaust can be an important occupational problem for these police [14], and this exposure can lead to the consequent carcinoma. Concerning the genotoxicity of benzene, Golding and Watson noted that benzene caused oxidative stress, which could be detected as oxidative damage to DNA [15]. They concluded that mechanisms other than DNA damage might play a role in benzene-related toxicity, such as reactions of benzene metabolites with essential enzymes such as topoisomerase II [15]. Here, the author assessed the correlation between SCE and urine ttMA among a sample of Thai traffic police from the same setting (with the daily exposure of benzene level equal to 150 μg /m3 [3]). A high urine biomarker among these police compared to controls can be seen [3]. Considering the SCE, the average SCE among the police is significantly higher than the average SCE of Thai controls (0.24 + 0.12 /cell). This finding can imply high exposure of benzene as well as high chromosomal aberration among these police. Indeed, there was a previous report on the SCE in the police in Italy [16]. In that study [16], Carere et al. noted that the SCE among the police is not significantly different from the general population in Rome. However, in that study, the authors did not control other confounding factors such as smoking. Here, the author performed a more controlled study and could show that the SCE among the police is significantly high. The correlation study between urine ttMA and SCE in this study can show a significant correlation. It can say that high exposure, determined from high urine ttMA, can imply the high chromosome aberration or mutagen. Of interest, the monitoring for benzene exposure among the at-risk workers is recommended [3]. In cases with high urine ttMA from screening,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Benzene Exposure and the Story of Carcinogenesis

391

the special consideration on the risk for cancer should be set. However, it should be noted that although it may be able to relate SCE to DNA damage, SCE definitely is not a reliable indicator of potential human health effects. The use of chromosome aberrations, micronuclei or even the Comet assay is much preferred as a biomarker of human health effects and should be assessed in the further studies. The author also recommends that an annual check up and monitoring for benzene exposure among the traffic police should be set as a primary prevention of occupational–related cancer for them.

Effects on Blood Cells: A Flow Cytometry Study on Urine Biomarker, Trans, Trans - Muconic Acid of Traffic Policemen and Blood Cell Parameters 1. Red Blood Cell

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Symptoms of benzene intoxication include headache, dizziness, fatigue, muscular weakness, drowsiness, and poor coordination with staggering gait, skin paresthesia, collapse, and coma [17]. The hematotoxic effect of benzene is mentioned [18]. Anemia is a common hematological manifestation due to benzene intoxication [18]. However, there is little knowledge about the correlation between the biomarker of benzene exposure and changes in red blood cell parameters. In a recent study [19], the author determined the correlation between urine ttMA level among a sample of exposed subjects to the red blood cell parameters. The regression analysis shows no significant correlation between urine ttMA and Hb, MCV or MCH [19]. Based on our hematologic data, the author proposed that only investigation for the urine ttMA might not be sufficient in detection of alteration in red blood cell parameters in the exposed population [19].

2. White Blood Cell Workers occupationally exposed to benzene exhibit increased frequencies of both structural and numerical chromosomal aberrations in their peripheral blood lymphocytes [20]. According to the recent study of Hristeva-Mirtcheva, leukocytosis with lymphocytosis was found in 20% of subjects occupationally exposed to benzene [21]. Rothman et al., noted a dose response relationship with various measures of current benzene exposure by personal air monitoring for the total white blood count and absolute lymphocyte count [22]. They supported the use of the lymphocyte count as the most sensitive indicator of benzene-induced hematoimmunotoxicity [22]. Here, the author investigated the correlation between ttMA, benzene urine biomarker, and lymphocyte count in occupationally exposed traffic policemen. An EDTA blood sample was collected for lymphocyte count analysis by the automated hematology flow cytometry analyzer, Technicon-H*3. The level of ttMA was measured according to the previous description in the previous section. The averages (mean + SD) of urine ttMA and lymphocyte count of the volunteer subjects were 0.79 + 1.43 mg/gCreatinine (range 0.05–1.86 mg/gCreatinine) and 2.52 + 0.68 x 103/μL (range 1.29–3.96 x 103/μL), respectively. The

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

392

Viroj Wiwanitkit

average urine ttMA in our subjects is high, however, the average lymphocyte count is not. No case of lymphocytosis can be detected. In addition, we found no significant correlation between ttMA and lymphocyte count. Classified by the reported cutoff high level for urine ttMA (0.13 mg/gCreatinine) [23], the lymphocyte count in the subjects with high urine ttMA level (n = 12, 0.67 + 0.19 mg/gCreatinine) is not significantly different from those with low urine ttMA (n = 13, 0.70 + 0.19 mg/gCreatinine) (independent sample T test, P = 0.41). In addition to this work, the author also studied the correlation between myeloperoxidation index and ttMA and detected a significant correlation implying the risk for leukemogenesis [24]. Based on these data, the author proposed that only investigation for the urine metabolite, ttMA, for benzene exposure is not sufficient in detection of lymphocyte change in the exposed population. A combination between biomaker and hematological test are recommended. Further studies with larger sample size focusing on the details of pathogenesis of lymphocyte disorders are also recommended.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

3. Platelets The effects of benzene on platelets have not been well studied. The correlation between the level of trans, trans muconic acid (ttMA), a urine metabolite of benzene, and the platelet count of subjects occupationally exposed to benzene was studied in a recent report by Wiwanitkit [25]. Traffic policemen in Bangkok who were occupationally exposed to benzene were investigated for both urine ttMA levels and platelet parameters [25]. According to the regression analysis, no significant correlation existed between urine ttMA level and any platelet parameters (p > 0.05) [25]. However, although there is no statistical significance, the platelet count and PCT decreased while the urine ttMA increased [24]. In addition, using the upper normal limit ttMA level as the cutoff level, the statistically significant lower platelet count and PCT was observed in the subjects with urine ttMA higher than the upper normal limit [25].

Acknowledgements This study was supported by the Rajchadapisakesompote Fund, Chulalongkorn University. The author would like to thank all subjects participating in this study and all health care workers, especially for Suphan Soogarun and Jamsai Suwansaksri, who helped perform the research and laboratory analysis.

References [1] [2]

Chocheo V. Polluting agents and sources of urban air pollution. Ann. Ist. Super Sanita 2000; 36:267-74. Crebelli R, Tomei F, Zijno A, Ghittori S, Imbriani M, Gamberale D, Martini A, Carere A. Exposure to benzene in urban workers: environmental and biological monitoring of traffic police in Rome. Occup. Environ. Med. 2001; 58:165-71.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Benzene Exposure and the Story of Carcinogenesis [3] [4] [5]

[6] [7] [8] [9] [10] [11]

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[12] [13] [14] [15] [16]

[17] [18] [19]

393

Wiwanitkit V, Suwansaksri J, Soogarun S. A note on urine trans, trans muconic acid level among a sample of Thai police: implication for an occupational health issue. Yale J. Biol. Med. 2003;76:103-8. Wiwanitkit V, Soogarun S, Suwansaksri J. Urine phenol and myeloperoxidase index: an observation in benzene exposed subjects. Leuk Lymphoma 2004;45:1643-5. Wiwanitkit, V., Suwansaksri, J.and Nasuan, P. Feasibility of using trans, trans muconic acid determination using high performance liquid chromatography for biological monitoring of benzene exposure. J. Med. Assoc. Thai. 84(Suppl 1): S263 – 8, 2001. Sripanidkulchal B, Tattawasart U. Chromosome aberration assay. In: Khon Kaen University. The 3rd Southeast Asian workshop on environmental mutagens and carcinogen, 1994. Khon Kaen, Khon Kaen University Press, 1994: 66 – 76. Agency for Toxic Substances and Disease Registry (ATSDR). Toxicological Profile for Benzene. Atlanta, G.A.: U.S. Department of Health and Human Services, Public Health Service, 1997. Lauwerys RR, Buchet JP. Biological monitoring of exposure to benzene, toluene and xylene. IARC Sci. Publ. 1988; 42: 205 – 222. Wiwanitkit V. Use of a novel peripheral biomarker, urine trans, trans, muconic acid, for benzene toxicity monitoring. Toxin Rev. 2004; 23: 467-475. Ong CN, Kok PW, Lee BL, Shi CY, Ong HY, Chia KS, Lee CS, Luo XW. Evaluation of biomarkers for occupational exposure to benzene. Occup. Environ. Med. 1995;52:528-33. Ong CN, Kok PW, Ong HY, Shi CY, Lee BL, Phoon WH, Tan KT. Biomarkers of exposure to low concentrations of benzene: a field assessment. Occup. Environ. Med. 1996;53:328-33. Leong ST. and Laortanakul, P. Indicators of benzene emissions and exposure in Bangkok street. Environ. Res. 2003; 92:173-81. Verma Y, Kumar A, Rana SV. Biological monitoring of exposure to benzene in traffic policemen of north India. Ind Health 2003; 41:260-4. Priante E, Schiavon I., Boschi G, Gori G, Bartolucci GB, Soave C, Brugnone F, Clonfero E. Urban air pollutant exposure among traffic policemen. Med. Lav. 87: 31422, 1996. Golding BT, Watson WP. Possible mechanisms of carcinogenesis after exposure to benzene. IARC Sci. Publ. 1999;(150):75-88. Carere A, Andreoli C, Galati R, Leopardi P, Marcon F, Rosati MV, Rossi S, Tomei F, Verdina A, Zijno A, Crebelli R. Biomonitoring of exposure to urban air pollutants: analysis of sister chromatid exchanges and DNA lesions in peripheral lymphocytes of traffic policemen. Mutat. Res. 2002;518:215-24. Ross D. Metabolic basis of benzene toxicity. Eur. J. Haematol. Suppl. 1996; 60:111-8. Kalf GF. Recent advances in the metabolism and toxicity of benzene. Crit. Rev. Toxicol. 1987;18:141-59. Wiwanitkit V, Soogarun S, Suwansaksri J. A correlative study on red blood cell parameters and urine trans, trans-muconic acid in subjects with occupational benzene exposure. Toxicol. Pathol. 2007;35(2):268-9.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

394

Viroj Wiwanitkit

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[20] Yager JW, Eastmond DA, Robertson ML, Paradisin WM, Smith MT. Characterization of micronuclei induced in human lymphocytes by benzene metabolites. Cancer Res. 1990;50:393-9. [21] Hristeva-Mirtcheva V. Changes in the peripheral blood of workers with occupational exposure to aromatic hydrocarbons. Int. Arch. Occup. Environ. Health 1998;71 Suppl: S81-3. [22] Rothman N, Li GL, Dosemeci M, Bechtold WE, Marti GE, Wang YZ, Linet M, Xi LQ, Lu W, Smith MT, Titenko-Holland N, Zhang LP, Blot W, Yin SN, Hayes RB. Hematotoxocity among Chinese workers heavily exposed to benzene. Am. J. Ind. Med. 1996;29:236-46. [23] Suwansaksri J, Wiwanitkit V. Urine trans,trans-muconic acid determination for monitoring of benzene exposure in mechanics. Southeast Asian J. Trop. Med. Public Health 2000;31:587-9. [24] Wiwanitkit V, Soogarun S, Suwansaksri J. A note on myeloperoxidation index and its correlation to the biomarker, urine trans, trans-muconic acid level, in the subjects occupationally exposed to benzene. Leukemia. 2004 Feb;18(2):369. [25] Wiwanitkit V, Suwansaksri J, Soogarun S. The urine trans, trans muconic acid biomarker and platelet count in a sample of subjects with benzene exposure. Clin. Appl. Thromb. Hemost. 2004 Jan;10(1):73-6.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

In: Traffic Related Air Pollution… Editors: S. Demidov and J. Bonnet, pp. 395-400

ISBN: 978-1-60741-145-1 © 2009 Nova Science Publishers, Inc.

TRAFFIC BENZENE POLLUTION IN BANGKOK: AIR BENZENE LEVEL AND IMPLICATION Viroj Wiwanitkit Wiwanitkit House, Bangkhae, Bangkok Thailand 10160

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Benzene exposure is of particular concern because of recent research indicating that benzene exposure can result in several chronic toxicities including carcinogenesis. Long-term benzene exposure is hematotoxic, genotoxic and immunotoxic. Exposure to benzene from automobile exhaust can be a significant occupational problem for urban populations. Here, the author assesses the seasonal pattern of air benzene levels in an urban area of Bangkok. In this report, the author also calculated the cancer risk for different occupations exposed to benzene vapor based on risk classification.

Introduction Benzene is the focus of particular concern because benzene exposure can result in several chronic toxicities, including leukemogenesis [1]. At present, work with benzene is subject to the Control of Substances Hazardous to Health (COSHH) Regulations of 1999. Recent epidemiological studies have described epidemiological evidence of lung cancer and childhood leukemia in relation to traffic-related air pollution, with particular reference to benzene [2]. Exposure to benzene from automobile exhaust can be an important occupational problem for urban populations. According to a recent study by Wiwanitkit et al., the Thai police were highly exposed to traffic benzene vapor [3]. Wiwanitkit et al. also noted the possible carcinogenesis from exposure to the benzene vapor for this occupation [4]. In addition, people living in urban areas are also reported to be subjected to high levels of exposure to benzene vapor [5]. Exposure to benzene from automobile exhaust can be a significant occupational problem for urban populations. Here, the author assesses the seasonal pattern of air benzene levels in an urban area of Bangkok, and calculates the cancer risk for different occupations exposed to benzene vapor based on risk classification.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

396

Viroj Wiwanitkit

Seasonal Variation of Air Benzene Level in Bangkok 1. Data on Environmental Benzene Levels in Bangkok The present study was conducted to assess the seasonal variation of air benzene levels in an urban area of Bangkok. The data on environmental benzene levels in Bangkok is derived from the Pollution Control Department, Bangkok Thailand. Measurement of environmental benzene was performed daily by the Pollution Control Department using standard collection and analytical methods. The data on the monthly environmental benzene levels in the Payathai area were used for further study on the seasonal pattern.

2. Seasonal Pattern of Environmental Benzene Level in Bangkok In this study, the distribution of monthly environmental benzene level was tested for normal distribution by the Kolmogorov-Smirnov test. In addition, the correlation between the environmental benzene level and basic important climate parameters, including monthly rainfall and humidity, was also analyzed.

3. Result

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The average concentration of environmental benzene during each month is presented in Table 1. The distribution of this data is not normal. The peak of environmental benzene level can be detected in October and November. The data on monthly rainfall and humidity are shown in Table 2. There is no significant correlation between environmental benzene level and studied climate parameters. Table 1. Environmental benzene level by month Month January February March April May June July August September October November December

Environmental benzene level (μg/m3) minimum maximum average 0 0.06 0.02 0 0.07 0.02 0.01 0.05 0.02 0 0.04 0.02 0 0.04 0.02 0 0.05 0.02 0.01 0.06 0.02 0.01 0.05 0.02 0.01 0.07 0.02 0.01 0.05 0.03 0.01 0.07 0.03 0.01 0.05 0.02

*Data from annual report of Pollution Control Department, Bangkok, Thailand.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Traffic Benzene Pollution in Bangkok: Air Benzene Level and Implication

397

Table 2. Environmental benzene level, rainfall and humidity by month Month January February March April May June July August September October November December

Average environmental benzene level (μg/m3) 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.02

Climate parameters Rainfall* Humidity* (mm) (%) 26.4 69.3 0 66.2 6.3 64.4 38.8 62.6 174.4 76.0 165.4 77.8 181.0 74.7 216.4 78.2 306.8 85.5 187.1 81.6 13.6 79.0 50.0 78.1

*Data from annual report of Thai Meteorological Department.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

4. Discussion Benzene is a common toxic volatile substance found in many present-day industrial processes [1]. It is classified as a carcinogen and can cause serious health problems. Epidemiological evidence indicates a relationship between exposure to benzene and the occurrence of acute non-lymphocytic leukemia in humans [7]. Exposure to benzene from automobile exhaust can be a significant occupational problem for urban populations. Therefore, working and living in an urban area in the presence of air pollution can be health hazard for the people living in urban area. Monitoring of air benzene level is useful. Although there are some studies on the seasonal change of air benzene levels, those studies usually focus on industrial areas. Rao et al. studied the seasonal variation of air benzene level around an industrial factory in India and reported no seasonal variation [8]. However, there is limited data on the seasonal change of air benzene level in areas of traffic. Here, the author assesses the seasonal pattern of air benzene level in an urban area of Bangkok. Of interest, there is a peak increase in the air benzene level in the interim period between the rainy season and winter. This observation is similar to a recent report conducted in Hong Kong [9]. However, the author cannot detect the correlation between air benzene level and the studied climate parameters, rainfall and humidity. Therefore, there might be other factors corresponding to the observed seasonal pattern. This might be the changing of the wind direction in that period of the year.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

398

Viroj Wiwanitkit

Calculated Cancer Risk for Different Occupations Exposed to Benzene Vapor Based on Risk Classification A. Introduction Benzene is a colorless poisonous liquid with a sweet odor [1]. Inhaling an extremely high level of benzene can result in death, while exposure to a high level can cause drowsiness, dizziness, rapid heart rate, headaches, tremors, confusion, and unconsciousness. Long-term benzene exposure is hematotoxic, genotoxic and immunotoxic. Although the mechanisms underlying benzene-induced toxicity and leukemogenicity are not yet fully understood, they are likely to be complicated by various pathways, especially those of metabolism and programmed cell death [10]. There are many occupations with high risk for benzene exposure. Wiwanitkit et al. [4] recently noted the possible leukemogenesis from exposure to the benzene vapor. In this report, the author calculated cancer risk for different occupations exposed to benzene vapor based on risk classification.

B. Materials and Methods

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

The author makes use of the data from the previously-published paper on the risk classification report on the environmental benzene exposure of different occupations (traffic policeman, oil refinery worker, press worker, fisherman, gas station attendant and mechanics) [11–12]. The author also used a reported estimated cancer incidence for the police [11] as primary data for further calculation of other at-risk occupations in the list.

C. Results The quoted classified risks for environmentally-exposed occupations are presented in Table 3. The primary references predicted cancer incidence for traffic police is 0.13% over a 70-year period. The calculated risks for other occupations are shown in Table 3. Table 3. Exposure risk ratio in previous studies concerning urine ttMA determination in at-risk occupations Predicted cancer incidence (% over a 70-year period) 2 Traffic policeman 15.8 0.13 Oil refinery worker 19.0 0.05 Press worker 7.0 0.06 Fisherman 9.0 0.07 Gas station attendant 33.3 0.27 Mechanic 2.3 0.02 1 This exposure risk ratio according the previous report by Wiwanitkit [12]. 2 Primary data on predicted cancer incidence for traffic police (0.13 % over a 70-year period) from the previous report by Wiwanitkit et al. [11] is used for further calculation. For example, the predicted cancer incidence for oil refinery work is equal to “(0.13 x 19.0)/15.8.” Occupation

Exposure risk ratio1

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Traffic Benzene Pollution in Bangkok: Air Benzene Level and Implication

399

D. Discussion Epidemiological evidence shows a significant relationship between exposure to benzene and the occurrence of acute non-lymphocytic leukaemia in humans [13]. Golding and Watson said that the significance of DNA adduct formation with respect to human leukaemia was uncertain [13]. They noted that mechanisms other than DNA damage might play a role in benzene-related toxicity, such as reactions of benzene metabolites with essential enzymes such as topoisomerase II [13]. Several occupations are mentioned as at risk for benzene exposure. Biomonitoring for risk of these workers is recommended. There are some reports on the high level of urine biomarker among many occupational workers [14–20]. Recently, Wiwanitkit estimated the incidence for cancer among traffic policeman and reported 0.05 cases from 39 studied cases in a 70-year following up [12]. Here, the author used this information for further calculation by simulating the reported risk classification [11]. According to this study, it can be shown that the estimated cancer incidence varies from 0.02% to 0.27%. Of interest, this rate is considerably important and should be a cause for concern. Annual checkups and monitoring for benzene exposure among workers in occupations at high risk for exposure should be established as primary prevention of occupation-related cancer.

References

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[1]

Chocheo V. Polluting agents and sources of urban air pollution. Ann. Ist Super Sanita. 2000;36(3):267-74. [2] Lagorio S, Forastiere F, Lipsett M, Menichini E. Air pollution from traffic and the risk of tumors. Ann. Ist Super Sanita. 2000;36(3):311-29. [3] Wiwanitkit V, Suwansaksri J, Soogarun S. A note on urine trans, trans muconic acid level among a sample of Thai police: implication for an occupational health issue. Yale J. Biol. Med. 2003 May 1;76(3):103-8. [4] Wiwanitkit V, Soogarun S, Suwansaksri J. Urine phenol and myeloperoxidase index: an observation in benzene exposed subjects. Leuk Lymphoma. 2004 Aug;45(8):1643-5. [5] Suwansaksri J, Wiwanitkit V, Neramitraram P, Praneesrisawasdi P. Urine trans,transmuconic acid levels in residents of a business area of Bangkok, Thailand. Clin. Chem. Lab. Med. 2002 Nov;40(11):1174-5. [6] Pollution Clearing House (1993), Airport noise - Estimation of Cancer Risk. Available online: http://www.nonoise.org/resource/trans/air/cancer/cancer.htm. [7] Golding BT, Watson WP. Possible mechanisms of carcinogenesis after exposure to benzene. IARC Sci. Publ. 1999;(150):75-88. [8] Rao PS, Ansari MF, Gavane AG, Pandit VI, Nema P, Devotta S. Seasonal variation of toxic benzene emissions in petroleum refinery. Environ. Monit Assess. 2007 May;128(1-3):323-8. [9] Lee SC, Chiu MY, Ho KF, Zou SC, Wang X. Volatile organic compounds (VOCs) in urban atmosphere of Hong Kong. Chemosphere. 2002 Jul;48(3):375-82. [10] Crebelli R, Tomei F, Zijno A, Ghittori S, Imbriani M, Gamberale D, Martini A, Carere A. Exposure to benzene in urban workers: environmental and biological monitoring of traffic police in Rome. Occup. Environ. Med. 2001; 58:165-71.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

400

Viroj Wiwanitkit

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

[11] Wiwanitkit V, Suwansaksri J, Soogarun S. Cancer risk for Thai traffic police exposed to traffic benzene vapor. Asian Pac. J. Cancer Prev. 2005;6:219-20. [12] Wiwanitkit V. Classification of risk occupation for benzene exposure by urine trans, trans-munconic acid level. Asian Pac. J. Cancer Prev. 2006;7:149-50. [13] Golding BT, Watson WP. Possible mechanisms of carcinogenesis after exposure to benzene. IARC Sci. Publ. 1999;(150):75-88. [14] de Paula FC, Silveira JN, Junqueira RG, Leite EM. Assessment of urinary trans, transmuconic acid as a biomarker of exposure to benzene. Rev. Saude Publica 2003; 37: 780-5. [15] Suwansaksri J, Wiwanitkit V. Urine trans,trans-muconic acid determination for monitoring of benzene exposure in mechanics. Southeast Asian J. Trop. Med. Public Health 2000;31:587-9. [16] Thummachinda S, Kaewpongsri S, Wiwanitkit V, Suwansaksri J. High urine ttMA levels among fishermen from a Thai rural village. Southeast Asian J. Trop. Med. Public Health 2002;33:878-80. [17] Wiwanitkit V, Suwansaksri J, Nasuan P. Urine trans,trans-muconic acid as a biomarker for benzene exposure in gas station attendants in Bangkok, Thailand. Ann. Clin. Lab. Sci. 2001;31:399-401. [18] Wiwanitkit V, Soogarun S, Suwansaksri J. Urine phenol and myeloperoxidase index: an observation in benzene exposed subjects. Leuk Lymphoma 2004;45:1643-5. [19] Wiwanitkit V, Suwansaksri J, Neramitraram P, Praneesrisawasdi P. A note on urinary trans,trans-muconic acid level among Thai press workers. Biomarkers 2003a;8:339-42. [20] Wiwanitkit V, Suwansaksri J, Soogarun S. A note on urine trans, trans muconic acid level among a sample of Thai police: implication for an occupational health issue. Yale J. Biol. Med. 2003b;76:103-8.

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

INDEX

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

A AAS, 166 ABC, 265 abnormalities, 15 absorption, 10, 12, 17, 23, 85, 93, 94, 139, 166, 221 access, 18, 70, 110, 164 accidental, viii, 121, 123, 125, 129, 131 accidents, 14, 56, 106, 107, 122, 123 accounting, 85, 172, 174, 212, 238 accuracy, vii, 29, 33, 40, 41, 42, 46, 51, 53, 54, 55, 56, 57, 58, 59, 63, 70, 122, 252, 262, 285, 298 acid, 6, 8, 24, 158, 388, 390, 392, 393, 394, 399, 400 acidity, 9 acoustic, 219, 285, 288, 289 acoustic waves, 288, 289 activation, 3 actual output, 46 acute, x, 5, 10, 13, 15, 17, 18, 21, 133, 179, 181, 397, 399 acute lung injury, 21 adaptive control, 64 additives, 12, 221 adducts, 3, 19, 25 adiabatic, 186, 187, 188, 214, 314, 368 adolescents, 23 adult, 13, 23, 381 adulthood, 13 adults, 3, 11, 12, 13, 25, 388 advection-diffusion, 333, 334, 335, 338 aerosol, 23, 97, 100, 152, 154, 155, 156, 158, 159 aerosols, 6, 9, 12, 14, 25, 158 Africa, vii, xii, 1, 379 afternoon, 113 age, x, 1, 10, 12, 13, 35, 70, 71, 122, 156, 161, 163, 241, 246 ageing, 30, 53 agent, 6, 24, 218, 224, 237 agents, 7, 12, 20, 23, 392, 399 aggregates, 112, 115 agricultural, 9, 244, 383 agriculture, 173

air, vii, viii, ix, x, xii, xiii, 1, 2, 3, 4, 5, 6, 9, 10, 12, 14, 15, 16, 18, 19, 20, 21, 22, 23, 24, 25, 28, 29, 30, 31, 32, 42, 43, 45, 46, 47, 50, 51, 56, 57, 58, 59, 61, 62, 63, 64, 65, 66, 69, 80, 90, 100, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 118, 119, 121, 122, 123, 124, 126, 127, 129, 130, 131, 132, 133, 135, 136, 137, 138, 139, 145, 146, 147, 149, 150, 151, 152, 153, 154, 155, 156, 157, 158, 159, 161, 162, 163, 164, 166, 167, 168, 172, 173, 174, 175, 176, 177, 178, 179, 181, 186, 192, 215, 216, 217, 218, 219, 220, 221, 222, 226, 227, 228, 229, 231, 232, 234, 236, 237, 238, 242, 339, 340, 341, 342, 343, 344, 345, 347, 348, 367, 368, 370, 371, 372, 373, 381, 390, 391, 392, 393, 395, 396, 397, 399 air emissions, 12, 65 air pollutant, vii, ix, 2, 3, 5, 10, 20, 29, 30, 31, 56, 58, 65, 104, 117, 118, 119, 123, 131, 135, 136, 137, 145, 157, 158, 159, 393 air pollutants, ix, 2, 3, 5, 10, 20, 56, 104, 118, 119, 123, 131, 135, 136, 137, 157, 158, 159, 393 air pollution, vii, ix, x, 1, 2, 3, 4, 5, 10, 14, 19, 20, 22, 23, 24, 25, 28, 29, 30, 31, 32, 56, 59, 63, 64, 66, 121, 122, 123, 127, 129, 130, 131, 133, 135, 136, 137, 145, 146, 147, 151, 159, 161, 174, 177, 178, 390, 392, 395, 397, 399 air quality, vii, ix, x, 28, 31, 43, 45, 46, 47, 50, 51, 57, 61, 62, 64, 69, 90, 100, 106, 113, 118, 127, 132, 133, 136, 137, 138, 139, 145, 146, 149, 150, 152, 156, 157, 161, 175, 178 air quality model, 156, 157 air toxics, viii, 103, 104, 105, 106, 107, 111, 112, 113, 114, 115, 116, 117, 119, 124 airborne particles, 7, 10 airflow obstruction, 23 airports, 137 airway epithelial cells, 11, 21 airways, 10 Alaska, 105, 116 algorithm, 124, 129, 276, 278, 306, 334 alkaline, 7, 163 alkane, 110 ALL, 76, 80, 81 allergens, 8, 11

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

402

Index

alloys, 18 alternative, ix, xi, xii, 16, 65, 121, 132, 211, 225, 339, 367, 381 alternatives, 64, 117, 212 alters, 20, 174 aluminium, 154 alveolar macrophage, 5, 20, 22 alveolar macrophages, 5, 20, 22 alveoli, 10 alveolitis, 27 Amadori, 175 ambient air, viii, 6, 10, 21, 22, 28, 80, 103, 104, 105, 109, 112, 116, 147, 152, 153, 154 ambient air temperature, 112 ambient pressure, 111 ambiguity, 217 amelioration, x, 161, 163, 174, 175 amendments, 118 amine, 13 aminoaciduria, 13 ammonia, 104, 128, 129, 130 amplitude, 270, 271, 317, 354 AMS, 156 Amsterdam, 132, 146, 382, 383 anaerobic, 368 analysis of variance, 166 anemia, 13, 14, 391 anger, 3 angular velocity, 180, 182, 184, 199 animal models, 10, 20 animal studies, 15, 16 animals, xiii, 14, 17, 18, 106, 122, 379, 381 ANOVA, 72, 166, 167, 168, 169, 171 anthracene, 3 anthropogenic, 4, 8, 9, 16, 29, 212 aorta, 3 aplasia, 4 appendix, 315, 318, 332 application, vii, ix, xi, 29, 38, 42, 47, 48, 52, 54, 55, 56, 57, 59, 61, 67, 71, 85, 89, 121, 131, 147, 210, 211, 222, 231, 244, 253, 266, 267, 271, 287, 288, 334, 365 aqueous solution, 166 ARB, 95 Arctic, xii, 379 Argentina, 251, 322 arid, 176 Arizona, 176, 177 aromatic compounds, 4 aromatic hydrocarbons, 394 arrhythmia, 24, 25 arrhythmias, 2, 4 arsenic, 27 ash, 21, 22, 23 Asia, vii, 1, 16, 38 Asian, 393, 394, 400 asphalt, 125 asphyxia, 6

assessment, viii, 31, 46, 58, 64, 103, 105, 107, 119, 121, 126, 127, 130, 131, 136, 146, 150, 159, 389, 390, 393 assignment, 67 assumptions, 52, 54, 56, 72, 75, 104, 112, 113, 123, 321 asthma, 5, 6, 8, 10, 19, 22, 26, 27, 28, 136, 381 asthmatic children, 28 asthmatic inflammation, 17 asymptotic, 288 ataxia, 13 Athens, 5, 22 atmosphere, x, 4, 7, 8, 9, 14, 15, 16, 18, 31, 101, 104, 106, 110, 123, 159, 161, 162, 163, 175, 212, 313, 314, 364, 399 atmospheric particles, 177 atmospheric pressure, 188, 313 attacks, 6, 10 Australasia, 64 Australia, xii, 28, 29, 48, 59, 60, 61, 64, 65, 66, 67, 68, 69, 85, 86, 93, 96, 97, 98, 379 Austria, 28, 65, 68, 93, 95, 99, 124, 157, 245 authority, 137, 139 automobiles, 14, 125 automotive application, 231 autonomy, 213 autopsy, 3 availability, 46, 47, 54, 55, 57, 59, 213, 306, 380, 381 averaging, 50, 255 avoidance, 132, 225

B backfire, 216, 217, 218, 219, 220, 223, 224, 225, 226, 227, 231, 232, 234, 238, 239, 244, 245 bacterial, 5, 7, 8 bacteriophage, 7, 27 barrier, 7 barriers, 122, 125, 130, 267 battery, 181, 213 Bayesian, 122 behavior, 21, 132, 218, 228, 253, 254, 263, 302, 316, 335, 354, 361 behavioral problems, 122 Beijing, 62, 65, 132, 337 Belgium, 62, 93, 95, 380 benchmark, 298 benefits, x, xi, 5, 48, 161, 163, 175, 211, 368 benzene, viii, xiii, 4, 103, 106, 107, 110, 113, 114, 115, 116, 117, 381, 387, 388, 389, 390, 391, 392, 393, 394, 395, 396, 397, 398, 399, 400 benzo(a)pyrene, 6 bioaccumulation, x, 161 biochemistry, 25 biodegradation, 22 biodiesel, 38, 212 biofuel, 243

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index biofuels, 37, 213, 243 biogas, 367, 368 biological systems, 12 biomarker, xiii, 387, 388, 390, 391, 393, 394, 399, 400 biomarkers, 27, 390, 393 biomass, 9, 104, 162, 213 biomass materials, 9 biomolecules, 8 biosphere, 25 birds, xii, 379, 380, 382, 383 birth, 4, 104 birth weight, 4 black carbon, 152, 155 blends, x, 179 blocks, 173 blood, 3, 7, 13, 15, 21, 24, 388, 389, 391, 393, 394 blood lead levels, 13 blood plasma, 13, 21 bloodstream, 10 boiling, 166, 188, 195 bonds, 221 bone marrow, 7, 21, 24 Boston, 3, 20 bottom-up, 57, 200 boundary conditions, xi, 252, 253, 256, 257, 261, 272, 297, 298, 301, 304, 306, 308, 309, 311, 316, 328, 332, 335, 337 boundary value problem, 306, 310 brain, 7, 12, 26, 381 brain damage, 381 branching, 153 brane, 16 breakdown, 35 breathing, 10, 14, 104 breeding, 381, 382, 383 Britain, 98, 379, 380, 382, 384 Brno, 128 bronchial airways, 17 bronchitis, 6, 10, 15, 16, 20 bronchoalveolar lavage, 8, 20 Brussels, 62, 96, 382 Buenos Aires, 245 buffer, 22, 240 building blocks, 37 buildings, 57, 125, 126, 127, 139, 162, 173, 380 burn, 3, 198, 216, 220, 223, 224, 225, 228, 238, 240, 241, 242 burning, viii, xii, 3, 6, 14, 103, 104, 177, 215, 227, 339, 340, 343, 348 burns, 15 buses, 34, 75, 79, 81, 82, 83, 87, 90, 94, 98, 99, 123, 125, 155, 156, 223, 244, 248 bypass, 222

C cables, 216

403

CAD, 126, 181, 183, 184, 189, 192, 206, 304, 305, 310, 328 cadmium, 12, 14, 20, 21, 26, 27 calcification, 13 calcium, 12, 13, 20 calibration, 237 calorie, 376 campaigns, 83, 149, 150 Canada, viii, 38, 66, 93, 103, 104, 105, 106, 107, 108, 112, 113, 115, 116, 117, 133, 244 cancer, xiii, 3, 4, 10, 14, 17, 23, 104, 387, 391, 395, 398, 399 capacity, 5, 41, 46, 51, 58, 67, 165, 179, 181, 187, 371 caps, 56 car accidents, 123 carbon, viii, x, xii, 3, 8, 9, 10, 64, 99, 104, 121, 122, 123, 124, 131, 136, 152, 153, 155, 156, 158, 161, 162, 164, 168, 173, 174, 175, 176, 177, 212, 221, 367, 368, 371, 372, 373, 374, 375, 376, 377 carbon dioxide, viii, x, 104, 121, 123, 161, 162, 164, 173, 174, 177, 212, 368, 371 carbon monoxide, viii, xii, 3, 64, 104, 121, 123, 124, 136, 367, 371, 372, 373, 374, 375, 376, 377 carboxyl, 13 carboxyl groups, 13 carcinogen, 14, 15, 16, 106, 381, 390, 393, 397 carcinogenesis, xiii, 27, 387, 393, 395, 399, 400 carcinogenic, 7, 14, 16 carcinogenicity, 6, 16 carcinogens, 3, 107 carcinoma, 390 cardiac arrhythmia, 2, 25 cardiopulmonary, 11, 20 cardiovascular disease, 3, 28, 146 cardiovascular system, 3 cargo, 154, 158 carrier, x, 211, 212 case study, ix, 121, 124, 127, 131, 133 cast, 173 casting, 56 catalase, 26 catalyst, 27, 38, 42, 53, 83, 99, 123, 155, 220, 223, 224, 225, 226, 230, 242, 368 cation, 137 cats, 380, 381 causal relationship, 383 CBS, 63 cell, ix, xi, 5, 10, 11, 12, 17, 18, 19, 42, 121, 127, 211, 213, 214, 225, 242, 243, 290, 389, 390, 391, 393, 398 cell death, 17, 398 cell division, 18 cement, 14, 125, 175 Centers for Disease Control, 106, 117 central nervous system, 106 cereals, xii, 379 CFD, 252, 253, 263, 317, 325, 335, 336, 338 CH4, 156, 370, 372

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

404

Index

charge density, 205, 206 chelating agents, 21 chemical composition, 162 chemical reactions, 123 chemicals, 29 chemiluminescence, 139 chemokine, 23 childbirth, 15 childhood, 13, 146, 177, 395 children, 5, 10, 11, 13, 19, 22, 23, 25, 27, 28, 147, 158, 381 Chile, 61, 338 China, 62, 65, 146, 147, 149, 159, 337 chlorophyll, 176 chromatography, 111 chromium, 12, 14, 15, 16, 18 chromosome, xiii, 387, 390, 391 chromosomes, 28 chronic disease, 13, 18 chronic diseases, 18 chronic obstructive pulmonary disease, 27 cigarette smoking, 108 cilia, 18 circulation, 337 classes, 9, 34, 35, 38, 39, 40, 41, 46, 104 classical, 256, 322, 325 classification, vii, xiii, 29, 32, 34, 35, 36, 37, 46, 55, 58, 123, 395, 398, 399 clean air, 118 Clean Air Act, 107 cleaning, 104, 217 closure, 283 clustering, 274 CO2, x, 42, 51, 64, 104, 123, 124, 152, 153, 154, 155, 156, 157, 161, 162, 163, 164, 165, 166, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 179, 211, 212, 226, 237 coaches, 123 coal, 9, 14, 16, 177 codes, 51, 253, 306, 312 cohort, 10, 11, 146 coke, 104 coma, 13, 391 combined effect, 31 combustibility, 368, 371, 372, 373, 376 combustion, vii, xi, xii, 3, 4, 5, 8, 9, 14, 15, 16, 21, 23, 80, 108, 162, 181, 182, 183, 186, 211, 212, 213, 214, 215, 216, 217, 218, 220, 221, 222, 223, 224, 225, 226, 227, 229, 236, 241, 242, 243, 244, 245, 246, 247, 248, 251, 252, 253, 255, 256, 260, 263, 285, 289, 303, 322, 325, 328, 332, 337, 339, 340, 341, 342, 343, 344, 345, 347, 348, 349, 351, 357, 364, 365, 368, 371, 373, 377, 381, 383, 384 combustion chamber, 181, 186, 216, 217, 218, 220, 222, 224, 285, 322, 340, 342, 343, 345, 373 combustion characteristics, 248, 339, 341 combustion processes, 3 combustion stability, 222, 224, 229 communication, 57

communities, viii, 104, 105, 106, 107, 110, 116, 117 community, viii, 16, 103, 105, 110, 113 competition, 213, 380 compilation, 107 complement, ix, 121, 125 complexity, vii, 29, 31, 36, 38, 43, 46, 47, 55, 57, 219, 237, 262 compliance, 43, 45, 137 complications, 15, 17 components, 3, 9, 11, 18, 24, 30, 50, 136, 226, 237, 255, 258, 262, 267, 297, 306, 316, 322, 342, 343 composition, x, 11, 25, 30, 34, 35, 37, 55, 105, 112, 113, 136, 137, 161, 162, 163, 176, 177, 221, 236, 342, 383 compounds, 3, 4, 8, 12, 14, 16, 17, 18, 104, 105, 106, 110, 113, 114, 116, 117, 118, 129, 381 compressibility, 291 computation, 32, 40, 41, 58, 252, 262, 272, 278, 312, 327, 328, 333, 334, 336, 354 Computational Fluid Dynamics, 252, 337, 338 computer simulations, 123 computing, 57, 62, 252, 253, 262, 270, 320 concentration, x, xii, 5, 6, 31, 50, 64, 68, 70, 99, 101, 107, 123, 129, 131, 132, 136, 139, 143, 144, 152, 153, 154, 155, 157, 161, 162, 163, 164, 166, 168, 169, 171, 172, 173, 174, 176, 212, 216, 217, 220, 221, 222, 231, 237, 367, 368, 370, 371, 372, 373, 375, 376, 377, 381, 390, 396 concrete, 125 condensation, 9, 110, 220, 236, 237 conditioning, x, 30, 42, 122, 155, 161, 163, 252, 276 conductivity, 254, 339 confidence, viii, 41, 58, 69, 72, 75, 76, 77, 79, 80, 81, 82, 83, 84, 88, 357, 360, 364 confidence interval, viii, 41, 69, 72, 75, 76, 77, 79, 80, 81, 83 confidence intervals, viii, 41, 69, 72, 76, 77, 80, 81, 83 configuration, 58, 123, 151, 240, 284, 325 conflict, 265 confusion, 136, 398 Congestion, 66 congestive heart failure, 24 Congress, 62, 68, 243, 245, 246, 247, 248, 378 Connecticut, 97 connectivity, 263 consent, 388 conservation, 179, 186, 187, 253, 291, 317, 320, 321, 332, 333, 336 constraints, xi, 32, 110, 252, 261, 264, 269, 313, 314, 332 construction, 8, 82, 104, 270 consumption, vii, x, xi, 29, 30, 32, 34, 37, 38, 40, 42, 43, 48, 51, 56, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 123, 152, 157, 161, 163, 211, 381 contamination, 16, 22, 173, 177 continuity, 289, 312 control, vii, xii, 11, 12, 29, 30, 42, 46, 51, 53, 56, 57, 59, 61, 64, 117, 124, 136, 137, 139, 145, 163, 164,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index 213, 215, 217, 218, 220, 223, 224, 226, 230, 231, 233, 236, 237, 241, 242, 244, 294, 351, 364, 365, 369, 388, 390 convection, 187, 334, 340 convective, 285 convergence, 262, 266, 271, 276, 277, 288 convergence criteria, 277 conversion, 6, 9, 212, 213, 215, 230, 239, 240, 242, 245, 373 convex, 264, 265, 266 cooking, 108 cooling, xii, 217, 219, 220, 221, 225, 234, 339, 340, 341, 342, 343, 345, 347, 348, 349 coordination, 43, 391 Copenhagen, 28, 65, 98, 99, 100 copper, 12, 27 correlation, xiii, 38, 81, 140, 143, 144, 145, 173, 187, 188, 341, 381, 387, 388, 389, 390, 391, 392, 394, 396, 397 correlation analysis, 145 correlation coefficient, 140, 143, 144 correlations, 187, 191 corrosion, 173, 216 corrosive, 6 cost-effective, 33, 51, 55, 57, 59 costs, 31, 33, 48, 55, 57, 98, 175, 368 cough, 17 coughing, 6, 10, 15 countermeasures, 216 coupling, xi, 252, 306, 309, 311, 312, 313, 314, 315, 316, 317, 319, 320, 329, 332 coverage, 57 covering, viii, 48, 69, 90, 223, 262, 265, 324 CPC, 75, 76, 77, 79, 84, 85, 87, 88, 93, 152, 153, 154 CRC, 95, 176 crops, 213 cross-sectional, 182 crude oil, 14 crust, 9 cryogenic, 248 culture, 26 cycles, xii, 32, 33, 37, 40, 43, 48, 50, 51, 55, 56, 58, 60, 63, 67, 71, 96, 98, 132, 241, 304, 325, 329, 351, 352, 354, 357, 358, 361, 364 cycling, 11 cytokine, 17 cytokines, 8, 11, 17 cytometry, xiii, 387, 391 cytosine, 7 cytotoxicity, 12 Czech Republic, 121, 127, 131

D damping, 256 data availability, 46, 52, 55, 59 data collection, 57 data distribution, 130

405

data processing, ix, 121, 126, 127, 130 data set, ix, 121 database, 43, 56, 59, 80, 123, 130, 131, 133, 156, 158, 384 database management, 130 death, 5, 10, 13, 14, 18, 23, 398 deaths, 3, 382 decay, 181, 185, 192, 195, 201, 202, 203, 204, 205 deciduous, x, 161, 162, 163, 168, 173, 174, 382 decision-making process, ix decisions, 60 decomposition, 333, 334, 335 defects, 104 defense, 5 deficiency, 13 deficit, 224, 226 definition, 48, 58, 59, 260, 273, 290, 292, 293, 294, 295, 296, 323, 324, 325 deformation, xi, 251, 252, 254, 261, 263, 265, 270, 271, 272, 273, 274, 278, 279, 280, 281, 282, 283 degradation, 18, 262 degrees of freedom, 264 dehydrogenase, 26 delivery, 248 demand, 31, 47, 56, 59, 65, 219, 223, 224, 233 dendritic spines, 17 Denmark, 28, 93, 95 density, x, xi, 27, 30, 39, 146, 161, 162, 164, 166, 167, 170, 171, 172, 173, 174, 177, 180, 185, 186, 187, 192, 213, 219, 221, 227, 238, 245, 248, 251, 252, 253, 259, 260, 284, 289, 298, 301, 302, 304, 305, 309, 313, 314, 315, 316, 319, 382 Department of Agriculture, 175, 176, 177 Department of Energy, 125 Department of Health and Human Services, 19, 393 Department of Transportation, 61 dependent variable, 72, 166, 172 deposition, 20, 31, 124 deposits, 216, 217, 220 depression, 106, 232, 233 derivatives, 7, 24, 260, 264, 265, 268, 276, 288, 289, 311 dermatitis, 15 destruction, xii, 8, 216, 367 detachment, 18 detection, 33, 62, 152, 157, 158, 389, 391, 392 developed countries, viii, 69, 71, 75, 85, 89 developing countries, vii, 1, 70 developing nations, 33 DFT, 353, 354 diabetes, 1 diarrhea, 14 diesel, xii, 3, 4, 9, 11, 34, 40, 43, 53, 60, 80, 81, 84, 85, 86, 94, 96, 98, 99, 100, 107, 123, 132, 152, 153, 154, 155, 158, 159, 212, 236, 241, 283, 285, 286, 287, 288, 351, 352, 355, 356, 357, 361, 363, 364, 365, 367, 369, 377, 382 diesel engines, xii, 99, 351, 352, 364, 382 diesel fuel, 94, 98, 357

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

406

Index

diet, xiii, 379, 382 diets, 388 differential equations, 309 differentiation, 268 diffusion, 5, 154, 188, 296, 309 diffusivity, 215, 255, 257, 296 dilation, 353 dioxins, 132 direct measure, 112 Dirichlet boundary conditions, 259 discontinuity, 206, 260 Discovery, 28 discretization, 255, 258, 291, 311 disease rate, 27 diseases, 1, 2, 146 dispersion, viii, x, 31, 50, 57, 65, 99, 106, 121, 122, 124, 125, 126, 129, 131, 132, 147, 149, 151, 153, 154, 155, 291 displacement, 113, 229, 261, 264, 267, 269, 272, 281, 325 distribution, x, 12, 16, 19, 26, 31, 38, 62, 70, 79, 89, 95, 97, 99, 101, 127, 129, 132, 139, 149, 150, 151, 152, 153, 154, 155, 175, 178, 213, 284, 304, 305, 326, 396 diversification, 37, 57, 59 division, 18 dizziness, 106, 391, 398 DNA, 3, 7, 18, 19, 24, 25, 26, 390, 391, 393, 399 DNA damage, 7, 18, 26, 390, 391, 399 DNA lesions, 393 DNA repair, 25 dominance, 285 downsizing, 189, 199, 200 drinking, 15, 388 drinking water, 15 drowsiness, 391, 398 durability, 225, 371 duration, 22, 125, 153, 157, 226, 232, 237, 271, 325, 343 dust, 8, 9, 14, 16, 17, 129, 173, 175, 217 dusts, 16 dynamic viscosity, 187, 256

E earth, 9, 14 eating, 15, 388 ecological, 136 ecological systems, 136 ecology, 383 economic activity, 31, 124 economic efficiency, 377 economics, 213 ecosystem, x, 161, 163, 174 eddies, 255, 301 edema, 5 Education, 66 elderly, 11, 146

election, 290 electric charge, 181, 182, 189, 206, 209 electric utilities, 5 electrical resistance, 220 electricity, 40, 158 electrodes, 220 electromagnetic, 380, 383 electrostatic force, 181 emission, vii, viii, ix, x, 29, 30, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 93, 95, 96, 97, 98, 99, 100, 101, 104, 106, 111, 112, 114, 115, 116, 117, 122, 123, 124, 125, 127, 131, 132, 136, 149, 151, 154, 155, 156, 157, 158, 159, 162, 176, 177, 211, 212, 213, 223, 224, 231, 234, 245, 246, 248 emission source, 122, 124, 131, 176 emitters, 32, 33, 40 employees, 107, 114 encapsulated, 270 encephalopathy, 13 endocrine, 27 endotoxins, 8 energy, x, 4, 25, 40, 161, 162, 175, 187, 211, 212, 213, 216, 217, 219, 224, 225, 234, 237, 238, 243, 244, 247, 253, 254, 257, 264, 268, 289, 291, 320, 321, 352, 353, 354, 368 energy consumption, x, 40, 161 energy density, 213, 224, 225, 353 engines, vii, x, xi, xii, 15, 30, 32, 34, 99, 106, 111, 114, 115, 119, 123, 179, 181, 182, 189, 199, 201, 210, 211, 212, 213, 214, 215, 216, 217, 218, 220, 222, 223, 224, 225, 226, 236, 241, 243, 244, 245, 246, 248, 249, 251, 252, 253, 256, 257, 260, 263, 289, 306, 310, 322, 325, 329, 330, 332, 339, 340, 341, 342, 347, 348, 349, 351, 352, 364, 365, 381, 382 England, 146, 179, 365, 381, 382, 383 entropy, 333, 336, 376 environment, ix, x, 14, 18, 25, 28, 31, 51, 57, 71, 121, 122, 125, 126, 128, 129, 130, 131, 132, 133, 147, 161, 162, 163, 173, 174, 176, 248, 262, 364, 382 environmental conditions, 111 environmental effects, 104, 384 environmental factors, 25, 146 environmental impact, 48, 60, 62, 64, 159, 247 Environmental Protection Agency, viii, ix, 2, 9, 12, 13, 16, 21, 33, 38, 40, 62, 65, 67, 100, 101, 104, 116, 118, 119, 121, 122, 124, 125, 126, 127, 132 enzymes, 390, 399 EPC, 304, 309 ependymal, 18 epidemiologic studies, 136 epithelial cell, 8, 11, 21 epithelial cells, 8, 11, 21 epithelial lining fluid, 8

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index epithelium, 2, 8, 18, 19 equality, 312, 318, 321 equilibrium, 9, 180, 188, 210, 254 equipment, 30, 32, 53, 100, 104, 137, 221 erosion, vii, 1 erythrocytes, 13, 20 Escherichia coli, 7 esters, 221 estimating, 61, 65, 79, 89, 107, 115, 124 ethylbenzene, 106, 113, 114, 115, 116 Euler equations, 291, 292, 293, 310 Eulerian, xi, 251, 252, 258 Euro, 38, 52, 53, 155 Europe, 5, 16, 28, 32, 38, 60, 66, 131, 212, 380, 381, 383 European Commission, 5, 28, 61, 65, 100, 212 European Environment Agency, 65, 99, 100, 383 evaporation, 106, 180, 181, 183, 187, 197, 198, 205, 206, 207, 209, 210 evening, 388 evolution, 187, 278, 315, 319, 320, 325 exclusion, 388 excretion, 13, 16, 26 exercise, 1, 23, 31, 58 exhaust heat, 347, 348 experimental design, x, 149, 151, 155, 354, 364 explosions, 348 exposure, viii, xiii, 1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 22, 23, 24, 25, 26, 27, 28, 31, 62, 103, 104, 105, 106, 107, 113, 116, 117, 118, 133, 146, 147, 155, 156, 158, 381, 382, 387, 388, 389, 390, 391, 392, 393, 394, 395, 397, 398, 399, 400 extinction, 380, 382 extraction, 353 extrapolation, 268 eyes, 16

F failure, 313, 315, 382 family, 256, 265 Far East, xii, 379 farming, 379 farmland, xii, 379, 380, 383 fatigue, 3, 14, 391 fats, 221 fear, 219 February, 62, 66, 109, 113, 114, 163, 396, 397 feces, 12 feedback, 58 fees, 136 FEM, 253, 260, 311, 328, 337 fertility, 18 fetus, 16 fibrinogen, 3 fibronectin, 8 fibrosis, 5, 17

407

filters, 21, 81, 83 finite element method, 289, 297, 334, 335, 337 finite volume, 317, 331, 334, 338 finite volume method, 331 Finland, 25, 153, 154, 160 fires, 8, 14, 16 firms, 115 First Nations, 117 flame, 139, 166, 214, 215, 216, 218, 220, 221, 222, 224, 225, 246 flammability, 216, 217, 222, 224 flammability limit, 216, 217, 222, 224 flare, 365 flexibility, 194 flow, xi, xii, xiii, 8, 27, 51, 111, 114, 115, 116, 124, 126, 139, 187, 210, 220, 225, 226, 227, 228, 229, 231, 234, 236, 237, 242, 246, 251, 252, 253, 255, 256, 257, 258, 259, 260, 261, 262, 281, 285, 287, 290, 293, 294, 297, 298, 299, 300, 301, 302, 303, 304, 305, 306, 309, 310, 313, 317, 321, 322, 324, 325, 327, 329, 330, 334, 335, 338, 365, 367, 370, 372, 373, 375, 376, 377, 387, 391 flow field, 253, 255, 256, 325 flow rate, xii, 27, 111, 114, 115, 116, 220, 225, 281, 367, 370, 375 fluctuations, xii, 237, 351, 354, 355, 361, 364 flue gas, 20 fluid, xi, xii, 8, 14, 184, 224, 251, 252, 253, 254, 256, 258, 262, 274, 285, 288, 298, 303, 306, 310, 314, 325, 327, 328, 329, 333, 334, 337 fluid mechanics, xi, 251, 252 fluorescence, 139, 222 focusing, 352, 392 food, xiii, 15, 213, 379, 380, 381, 390 Ford, 110, 155, 218, 223, 243, 244, 247, 248 forecasting, 124, 246 forest ecosystem, 176 forest fire, 16 forest fires, 16 Forest Service, 175, 176, 177 forests, 176 forgetting, 217 fossil, xii, 3, 5, 6, 9, 14, 16, 162, 212, 339, 367 fossil fuel, xii, 3, 5, 6, 9, 14, 16, 162, 212, 339, 367 fossil fuels, 3, 6, 9, 14, 16, 212, 367 Fourier, 290, 291, 353 Fox, 117 fractionation, 381 France, 117, 382 free radical, 7 free radicals, 7 freedom, 219, 225, 261, 264 freezing, 214 frequency resolution, 352 friction, 221, 257, 314, 333 Friday, 164 fuel, vii, ix, x, xi, xii, 5, 8, 15, 16, 18, 22, 23, 29, 30, 32, 34, 37, 38, 42, 43, 48, 51, 53, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 70, 71, 83, 98, 99,

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

408

Index

106, 117, 123, 136, 137, 155, 156, 157, 162, 164, 174, 179, 181, 182, 183, 186, 187, 188, 189, 191, 194, 198, 200, 205, 208, 209, 211, 212, 213, 214, 216, 218, 219, 220, 221, 222, 224, 225, 226, 232, 236, 237, 242, 243, 246, 247, 248, 339, 340, 342, 347, 348, 367, 368, 371, 373, 381 fuel cell, xi, 42, 211, 212, 213, 214, 225, 242, 243 fuel efficiency, 209 fuel type, 34, 43, 57, 59, 71, 137, 156, 181, 187, 191, 205, 209, 342 funding, 89

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

G gait, 391 gas, x, xi, xii, 6, 8, 20, 81, 98, 111, 155, 156, 158, 164, 165, 176, 179, 186, 187, 188, 195, 205, 211, 212, 215, 216, 217, 218, 219, 221, 224, 225, 226, 232, 233, 236, 237, 242, 243, 248, 254, 257, 313, 322, 328, 339, 340, 341, 342, 343, 347, 348, 364, 367, 368, 369, 370, 371, 372, 373, 375, 376, 377, 398, 400 gas chromatograph, 370, 371 gas phase, 158, 179, 187, 188, 195 gases, viii, 11, 56, 98, 110, 111, 152, 217, 221, 222, 224, 234, 236, 237, 239, 254, 340, 348, 352 gasoline, 9, 12, 18, 96, 106, 117, 132, 155, 174, 212, 213, 214, 215, 217, 218, 219, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 233, 236, 239, 241, 242, 370, 381 gastrointestinal, 15, 16 gauge, 105, 116, 218, 242 Gaussian, 124, 129, 353 generalization, 258 generation, 22, 31, 58, 64, 65, 100, 214, 222, 223, 224, 225, 337, 368, 371 generators, 306 Geneva, 28, 62, 68 genotoxic, xiii, 7, 19, 20, 395, 398 Geographic Information System, vii, viii, ix, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 132, 133 geophysical, 365 Georgia, 27 Germany, 93, 95, 124, 157, 221, 243, 247, 248, 380, 381 GHG, 67 Global Positioning System, 46, 56, 57, 63, 111, 131, 152, 153 global warming, xii, 212, 367 glutathione, 8, 26 glutathione peroxidase, 26 GM-CSF, 8 government, 104, 124, 156 grains, 9 granulocyte, 8 graph, 227 grass, 177, 382

grasses, xiii, 379 Greece, 66, 177 greenhouse, x, 29, 30, 31, 32, 38, 45, 51, 56, 57, 59, 61, 63, 131, 162, 176, 211, 212 greenhouse gas, 29, 30, 31, 32, 38, 45, 51, 57, 59, 61, 131, 162 greenhouse gases, 31, 32, 38, 57, 59, 162 grids, 130, 151, 338 grounding, 220 groups, 13, 85, 86, 88, 97, 105, 390 growth, 11, 30, 35, 176 growth rate, 35 GSM, 380 Guangzhou, 147 guidelines, 122, 215, 220 gulls, 380 guns, 46 gut, 16

H habitat, 380 half-life, 13 handicapped, 214 harmful effects, 13 headache, 391 health, vii, viii, x, 1, 2, 5, 9, 10, 11, 12, 13, 18, 20, 22, 23, 24, 25, 26, 28, 31, 50, 69, 70, 71, 73, 74, 75, 76, 89, 90, 93, 103, 104, 105, 106, 107, 114, 116, 117, 122, 133, 136, 146, 147, 161, 162, 163, 390, 391, 392, 397 Health and Human Services, 19, 393 health care, vii, 1, 392 health care workers, 392 health effects, 5, 10, 11, 12, 13, 18, 20, 22, 24, 26, 31, 104, 106, 107, 114, 122, 146, 391 health problems, 1, 9, 10, 390, 397 hearing, 122 hearing loss, 122 heart, 2, 3, 7, 10, 11, 12, 14, 24, 398 heart attack, 10 heart disease, 2 heart rate, 3, 398 heartbeat, 10 heat, xi, xii, 162, 163, 173, 177, 179, 180, 185, 187, 188, 195, 205, 217, 222, 234, 251, 252, 253, 254, 255, 256, 333, 339, 340, 341, 342, 343, 345, 347, 348, 349, 368, 372 heat capacity, 179, 187 heat conductivity, 254 heat loss, 222, 234 heat release, 340, 341, 343, 348 heat transfer, xi, xii, 179, 187, 188, 195, 217, 251, 252, 333, 339, 340, 348, 349 heating, 209, 215 heavy metal, x, 9, 22, 161, 162, 163, 164, 175, 177 heavy metals, x, 9, 175, 177

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index height, 112, 114, 139, 152, 153, 164, 165, 170, 180, 184, 190, 300, 322 hematologic, 20, 391 hematological, xiii, 387, 391, 392 hematology, 391 hemolytic anemia, 13 hemorrhage, 15 hemorrhages, 17 hepatocytes, 17 herbicides, 380, 383 Hessian matrix, 278 heterogeneous, 306, 310 high power density, 245 high pressure, 181, 219, 237, 368, 376 high resolution, 40, 56, 57, 58 high risk, 398, 399 high temperature, 368 high-frequency, 352, 357, 358 high-level, 13 high-performance liquid chromatography, 389 high-speed, 155, 246 high-tech, 59 highways, 18, 98, 100, 123, 132, 154 hippocampus, 17 histamine, 26 histogram, 127 histological, 17, 18 homogeneity, 219 Hong Kong, 28, 397, 399 hospital, 5, 6, 8, 19, 24, 27, 28 hospitalization, 146 hospitals, ix, 121, 123 hot spots, 157, 211, 216, 217, 218, 220, 221, 237 House, xii, 379, 380, 381, 382, 383, 384, 387, 395, 399 household, 14 household waste, 14 housing, 260, 322 human, vii, 1, 3, 4, 5, 7, 10, 11, 12, 13, 14, 15, 16, 19, 21, 22, 24, 25, 26, 27, 28, 31, 104, 105, 106, 107, 116, 122, 133, 136, 139, 147, 162, 391, 394, 399 human exposure, 11, 15, 25, 105, 106 human subjects, 27 humans, 13, 14, 15, 16, 18, 26, 106, 163, 397, 399 humidity, 6, 30, 127, 152, 153, 164, 166, 167, 168, 172, 396, 397 hybrid, 42, 81, 181, 223, 225, 245, 248, 249 hybrids, 37, 38, 40 hydrides, 219 hydro, viii, ix, xii, 3, 4, 8, 104, 111, 121, 124, 135, 136, 137, 150, 212, 216, 225, 339, 394 hydrocarbon, 30, 66, 213, 216, 225, 339 hydrocarbon fuels, 213, 339 hydrocarbons, viii, ix, xii, 3, 4, 8, 104, 111, 121, 124, 135, 136, 137, 212, 216, 225, 339 hydrogen, vii, x, xi, xii, 3, 9, 12, 211, 212, 213, 214, 215, 216, 217, 218, 219, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230, 231, 232, 233, 234,

409

236, 237, 238, 241, 242, 243, 244, 245, 246, 247, 248, 249, 339, 340, 341, 342, 343, 347, 348, 349, 367, 371, 372, 373, 374, 375, 376, 377 hydrogen gas, 221 hydrogen peroxide, 12 hydrothermal, 113 hydroxyl, 6, 12 hygienic, 23 hyperbolic, 334 hyperplasia, 17, 25 hypertension, 13, 122 hyperthyroidism, 13 hypothesis, 315, 317, 321, 332, 381, 382

I ICE, 212, 213, 214, 220, 226, 235, 242, 246, 248 identification, 90 identity, 254 ignition energy, 215, 216, 217, 220 Illinois, 384 IMA, 335 images, 125, 126, 128 imaging, 57 immune response, 17, 25 immune system, 14, 15, 17 immunocompetence, 17 impact assessment, vii, viii, 31, 42, 69, 70, 71, 73, 75, 76, 89, 90, 93 implementation, ix, 43, 51, 112, 121, 247 importer, 136, 137 impurities, 243 in situ, 104 in vitro, 7, 10, 11, 23, 28 in vivo, 7, 10, 24 inactivation, 8 incentives, 212 incidence, 3, 106, 381, 382, 398, 399 incineration, 20, 21, 173 inclusion, 32, 54, 57, 59, 85 incomplete combustion, 3, 368, 371 incompressible, 257, 285, 286, 297, 298, 299, 300, 301, 302, 332, 334, 335, 336 independent variable, 72, 166, 172 India, xii, 379, 393, 397 Indian, 390 Indiana, 351 indication, 72, 127, 189 indicators, 10, 152 indices, 156 induction, xiii, 17, 46, 216, 219, 220, 387 industrial, 14, 15, 16, 122, 138, 139, 262, 387, 390, 397 industrial accidents, 14 industrial application, 262 industry, 14, 152, 173, 212, 221, 248 inequality, 265 inert, 13, 217

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

410

Index

inertia, 234, 317 infancy, 213 infants, 5, 26 infarction, 1 infection, 23 infections, 2, 5, 6, 8 infinite, 34, 35, 267 inflammation, 8, 10, 11, 17, 20, 22, 27 inflammatory, 2, 8, 10, 12, 17, 18 inflammatory cells, 17 inflammatory mediators, 12, 18 informed consent, 388 infrared, 125, 139, 165, 221 infrastructure, 31, 39, 213 ingestion, 18 inhalation, 6, 7, 8, 12, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 107 injection, x, xi, 34, 42, 53, 155, 179, 180, 181, 182, 183, 184, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206, 207, 208, 209, 210, 211, 217, 219, 220, 221, 224, 225, 226, 231, 232, 236, 237, 241, 242, 243, 245, 246, 247, 248, 340, 349, 369 injections, 195 injury, 5, 10, 11, 18, 20, 21 inorganic, 6, 21 insects, 381 insertion, 165 inspection, 353 inspections, 137 instabilities, 259 instability, 336 instruments, 33, 152, 153, 164 insults, 3 integration, 131, 260, 312 intensity, 31, 54, 127, 173 interaction, 30, 31, 58, 59, 262, 334, 337 interactions, 11, 334 interface, viii, 121, 262, 309, 312, 313, 316, 317, 320, 321, 332 interference, 390 Intergovernmental Panel on Climate Change, 162, 176 interleukin, 26 internal combustion, vii, xi, xii, 181, 182, 183, 211, 212, 213, 214, 241, 242, 244, 246, 247, 248, 251, 252, 253, 255, 256, 260, 289, 325, 339, 351, 364, 365 International Agency for Research on Cancer, 106, 117, 393, 399, 400 Internet, 117, 118, 119, 131 interpretation, 89, 159 interstate, 155 interstitial, 13 interstitial nephritis, 13 interval, 41, 183, 238, 265, 268, 311, 314, 322, 324, 325, 326, 327, 328, 353, 354 intoxication, 13, 391 intrinsic, 259, 295, 296

invariants, 258 inventories, viii, 32, 43, 45, 64, 66, 69, 70, 89, 90, 99, 104, 107, 108, 124, 156, 177 inversion, 330 invertebrates, 381, 382, 383 Investigations, 158 investment, 18 ionization, 139 ions, 8, 9, 21, 29, 30, 45, 51, 61, 67, 123, 137, 151, 153, 158, 285 IPCC, 37, 63, 131 IPO, 303, 309 Ireland, xii, 379, 380 iron, 6, 10, 12 irradiation, 157 irritability, 13 irritation, 2, 10, 16, 17, 106 ISC, 126 ischemic, 2, 24 island, 162, 173 isotope, 21 isotropic, 255, 259, 260 Italy, 22, 23, 99, 161, 174, 354, 355, 387, 390 iteration, 268, 277, 279, 282, 312 IVC, 283, 286

J JAMA, 117 January, 396, 397 Japan, 124, 135, 136, 146, 243, 247, 248, 339, 369, 370 Jordan, 175 justice, 117

K kappa, 11 kappa B, 11 kidney, 7, 13, 14, 15 kidneys, 12, 16, 18 kinetic energy, 40, 254 King, 244 Kolmogorov, 255 Korea, 367, 368 Korean, 378 Kuwait, 177

L labeling, 137 laboratory studies, 154 labor-intensive, 122 labour, ix, 33, 149, 150 lactation, 13 Lagrange multipliers, 260, 261, 297, 328

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index Lagrangian, xi, 251, 252, 258 Lagrangian approach, 258 lamina, 241, 257 laminar, 215, 241, 257 land, 31, 70, 122, 139, 151, 213, 216, 221 land use, 31, 70, 122, 151 landfills, 368 landscapes, viii, 103, 105, 106 laptop, 157 large-scale, 255, 352 laser, 155 Latin America, vii, 1, 146 Latin American countries, 146 law, 271, 281, 304, 332 leaching, 21 lead, 3, 5, 12, 13, 19, 20, 21, 22, 23, 26, 27, 33, 38, 40, 47, 48, 52, 57, 59, 106, 123, 176, 177, 178, 181, 218, 234, 268, 285, 381, 390 leakage, 322 leather, 14 legislation, x, 33, 51, 211 legislative, 212 lens, 352 lesions, 3, 16 leukaemia, 399 leukemia, 106, 177, 395, 397 leukocytes, 17 leukocytosis, 391 Leydig cells, 18, 23 life expectancy, 2, 10 life quality, 18 lifestyles, 388 lifetime, 6, 9, 200, 221 likelihood, 238 limitation, 17, 32, 122, 273, 313 limitations, 231, 234 Lincoln, 165 linear, 72, 124, 257, 258, 265, 268, 269, 281, 304, 309, 311, 312, 313, 338, 380 linear function, 258 linear law, 281 linear model, 72 linear regression, 72, 380 linkage, 125 links, 3, 39, 45, 46, 51, 63, 322 lipid, 7, 8, 26 lipid peroxidation, 8 lipids, 5 liquid chromatography, 389, 393 liquid film, 181 liquid hydrogen, 219, 225 liquid phase, 179, 187, 195 liquid water, 224 liver, 7, 12, 14, 15, 16, 17 liver damage, 14 living environment, 136, 145 LNG, 83, 88 local authorities, 122 localization, 27

411

location, 30, 31, 35, 75, 83, 110, 111, 123, 125, 155, 190, 240, 248, 262, 353 locus, 184, 189 London, 19, 22, 60, 133, 176, 244, 334, 365, 379, 380, 382, 383 long period, 16, 140 longevity, x, 161, 162, 173, 174 long-term, x, 5, 10, 11, 15, 107, 131, 146, 161, 174 Los Angeles, 26, 66, 97, 156, 159 losses, 30, 123, 213, 214, 222, 224, 227, 228, 229, 231, 234, 236, 237 low temperatures, 217, 225 low-level, 25 LPG, 34, 44, 368 LSD, 83, 88, 94 lubricating oil, 216, 221 lubrication, 221 lung, 2, 3, 5, 6, 7, 8, 10, 11, 12, 14, 15, 17, 19, 20, 21, 22, 23, 25, 27, 395 lung cancer, 3, 10, 14, 15, 19, 25, 395 lung disease, 10, 11, 27 lung function, 8, 10, 11, 22, 23, 27 lungs, 5, 8, 10, 11, 12, 14, 16, 17 lying, 260, 261, 264, 282, 313 lymphocyte, 10, 391, 392 lymphocytes, 7, 24, 391, 393, 394 lymphocytosis, 391, 392

M Macao, 159 Macau, 149, 150, 151, 157, 159 Macedonia, 177 machinery, 246 machines, viii, 103, 105, 106, 107, 109, 111, 112, 113, 114, 115, 116, 253 macrophage, 8, 17 macrophages, 17 magnetic, 177, 178 magnetic properties, 177, 178 maintenance, 30, 32, 82, 380 major cities, 66 males, 389 Malta, 67 mammalian cell, 7 mammalian cells, 7 management, x, 32, 43, 48, 51, 53, 122, 124, 127, 130, 149, 159, 161, 177, 212, 246 manganese, 6, 12, 146 Manhattan, 94 manifold, 42, 154, 216, 219, 220, 221, 222, 223, 225, 226, 228, 231, 233, 236, 314, 315, 317, 332, 369 manifolds, 233 manufacturer, 137, 223 manufacturing, 15, 106, 115 mapping, 123, 131, 152, 223 market, 214 marketing, 106

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

412

Index

marrow, 7 mass transfer, x, 179, 182, 185, 186, 187, 188, 189, 191, 198 Massachusetts, 3, 24 mast cell, 10 Mathematical Methods, 365 mathematicians, 255 matrix, xi, 166, 252, 255, 257, 258, 262, 270, 277, 278, 288, 289, 290, 292, 293, 294, 295, 296, 312, 331, 332 MCV, 391 measurement, 32, 33, 68, 70, 71, 74, 76, 83, 84, 89, 114, 152, 153, 154, 155, 156, 157, 158, 159, 246, 352, 357, 368, 369, 373 measures, 12, 43, 48, 51, 56, 59, 79, 85, 87, 88, 174, 215, 217, 225, 263, 391 meat, 108 median, 48, 113 mediators, 12, 18 medication, 11 medicine, 22 Mediterranean, 163 megakaryocytes, 21 membranes, 17 memory, 3, 13, 17, 19, 106 men, 1, 11, 19, 27 menopause, 13 mercury, 21 mesh node, 261, 291 metabolic, 3, 26 metabolism, 23, 26, 393, 398 metabolite, 392 metabolites, 390, 394, 399 metal content, 11, 12, 172 metals, 6, 8, 11, 12, 20, 21, 22, 26, 28, 161, 173 meteorological, 31, 109, 114, 126, 127, 129, 137, 139, 152, 153 methane, vii, xii, 135, 136, 214, 215, 217, 340, 341, 342, 348, 367, 368, 370, 372, 373, 374, 375, 376, 377 metric, 72, 95, 263, 267, 270, 284 metropolitan area, 101 Mexico, 1, 16, 21, 28, 101 Mexico City, 1, 16, 21, 28 mice, 7, 14, 15, 18, 19, 20, 22, 24, 26 microbial, 11 microclimate, 164 microenvironment, 156 micrograms, 15 microtubules, 18 Middle East, xii, 379 migraines, 3 mild asthma, 5 mines, 247 mining, 9, 14, 104, 246 Minnesota, 98, 100, 154, 336 missions, 66 MIT, 337

mixing, 33, 110, 111, 112, 113, 114, 116, 188, 219, 225 mobile phone, 383 mobility, 248 model system, 159 modeling, viii, ix, xi, 64, 104, 105, 107, 117, 121, 122, 123, 124, 125, 126, 127, 128, 130, 131, 132, 241, 251, 252, 253, 255, 332, 337, 364 models, vii, viii, ix, x, xi, 10, 20, 29, 32, 33, 35, 36, 37, 38, 39, 40, 41, 42, 43, 46, 47, 48, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 69, 70, 71, 72, 73, 75, 76, 82, 89, 95, 104, 106, 107, 112, 121, 122, 124, 125, 127, 129, 131, 132, 161, 163, 210, 252, 253, 255, 256, 306, 310, 314, 315, 316, 317, 329 modulation, 357, 364 modules, 42 modulus, 353 moisture, 370 molar ratio, 157 molar ratios, 157 mole, 180, 188 molecular weight, 112 momentum, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 191, 192, 195, 198, 200, 201, 205, 209, 231, 233, 253, 256, 257, 289 monkeys, 17 monolithic, 312, 332 Montana, viii, 103, 105, 110, 118 Montenegro, 335, 336 Moon, 41, 64 morbidity, 2, 6, 8, 10, 11, 24, 136 Morlet wavelets, 353 morning, 113, 165, 169, 173 morphology, 122, 132 mortality, 2, 5, 6, 10, 11, 19, 20, 22, 24, 106, 122, 136, 146, 381 mortality rate, 136 motion, 182, 258, 260, 261, 262, 273, 274 motor neurons, 21 motor vehicle emissions, 8, 65, 66 motorcycles, viii, 43, 106, 121, 123, 124, 125 mould spores, 9 mouse, 7, 24 movement, xi, 46, 162, 163, 173, 205, 251, 252, 261, 262, 263, 271, 283, 304, 322, 325, 377 MTBE, 381, 383, 384 mucosa, 6, 26 mucus, 16, 17 multidimensional, 297, 335, 338 multidisciplinary, xi, 31, 58, 251, 252 multiple regression, 166, 172 multiple regression analysis, 166, 172 multiplication, 34 multiplicity, 71 multivariate, 41, 42 municipal solid waste, 21, 22 muscle, 3 muscle cells, 3

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Index muscles, 12 mutagen, 390 mutagenesis, 27 mutagenic, 3, 9, 24 mutations, 3, 7 myeloperoxidase, 393, 399, 400 myocardial infarction, 1, 3, 25

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

N NaCl, 23 NAM, 1 nanoparticles, 90, 153, 154 nation, 137, 379 national, viii, 32, 37, 43, 45, 51, 57, 103, 104, 107, 119, 124, 138, 139, 157 National Academy of Sciences, 7, 19, 25 National Institute for Occupational Safety and Health, 107 National Park Service (NPS), 105, 106, 107, 110, 112, 118 national parks, 139 National Research Council, 19, 67 natural, xii, 8, 9, 15, 16, 81, 98, 163, 173, 174, 176, 177, 221, 259, 368, 379 natural environment, 174 natural gas, 81, 98, 221, 368 natural resources, 177 nausea, 14 Navier-Stokes, 253, 255, 256, 258, 260, 289, 295, 297, 298, 301, 302, 310, 313, 332, 335, 336 Navier-Stokes equation, 253, 255, 256, 258, 260, 289, 295, 298, 301, 302, 313, 332, 335, 336 necrosis, 8, 18 needles, 175, 178 neglect, 32 negligence, 18 neighbourhoods, 156 neoplasm, 6 neoplasms, 6, 17 nephritis, 13 nephropathy, 13 nervous system, 13, 14, 17 nesting, 380 Netherlands, 63, 66, 380 network, 34, 39, 42, 43, 45, 46, 47, 48, 51, 52, 54, 58, 61, 70, 119, 123, 130, 152 Neumann condition, 309 neural network, 42, 61 neural networks, 42 neuronal death, 18 neurotoxicity, 22, 390 neutrophil, 10 neutrophilia, 17 Nevada, 98 New York, 20, 23, 25, 94, 98, 99, 133, 146, 156, 158, 176, 337, 338, 364, 382 New Zealand, xii, 66, 67, 97, 379

413

Newton, 271, 277 Newton iteration, 271 Newtonian, 253 NFkB, 11 nickel, 12, 149, 150, 152, 154, 156, 158, 160 Nielsen, 97 Nile, 382, 384 nitrate, 110 nitrates, 9, 111 nitric oxide, 246 nitrogen, viii, 3, 4, 5, 21, 22, 23, 26, 28, 60, 61, 65, 104, 110, 121, 122, 123, 124, 131, 137, 159, 245, 376, 377 nitrogen dioxide, 23, 26, 28, 60, 65, 131 nitrogen oxides, viii, 3, 4, 5, 21, 22, 121, 123, 124 nodes, 17, 45, 259, 261, 263, 264, 265, 266, 267, 268, 269, 271, 272, 274, 281, 282, 283, 291, 301, 304, 311, 312, 313, 316, 325 noise, viii, ix, 1, 121, 122, 123, 124, 125, 126, 127, 129, 130, 131, 132, 157, 159, 216, 322, 365, 399 non diabetic, 1 non-human, 23 nonlinear, 42, 264, 271, 311, 352, 364 nonlinear dynamics, 352, 364 non-smokers, 11 normal, 6, 27, 32, 35, 72, 75, 111, 127, 130, 153, 214, 217, 218, 222, 256, 257, 264, 297, 316, 332, 371, 373, 381, 392, 396 normal distribution, 127, 396 normalization, 266, 267, 353, 354 normalization constant, 266, 267 norms, 266 North Africa, 379 North America, viii, 5, 103, 104, 105, 106, 107, 109, 111, 112, 113, 115, 116, 117, 119, 136, 158, 384 North Carolina, 62, 100 Northeast, 96 nuclear, 8, 11 nucleation, 9, 79, 87, 90, 154, 158 nuclei, 9, 16, 110, 154 numerical analysis, 293 Nusselt, 187

O OAL, 124 observations, 6, 43, 98, 113, 125, 127, 155, 383 obstruction, 6, 17 occupational, xiii, 15, 105, 106, 387, 390, 393, 394, 395, 397, 399, 400 occupational health, xiii, 387, 393, 399, 400 oceans, 14 octane, 214, 218 octane number, 218 oil, 9, 14, 16, 21, 22, 23, 42, 117, 212, 213, 215, 216, 217, 220, 221, 222, 398 oil production, 212 oil refining, 117

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

414

Index

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

oils, 16, 221 older adults, 11 online, 210, 399 on-ramp, 125 open heart surgery, 3 operator, 258, 259, 296, 335 optical, 61 optimization, xi, 211, 213, 251, 263, 264, 266, 267, 268, 276, 277, 334 optimization method, 277 oral, 15 organ, 13, 15 organic, 3, 4, 9, 11, 25, 67, 99, 104, 105, 118, 123, 146, 147, 150, 399 organic compounds, 3, 4, 9, 67, 104, 118, 123, 146, 147, 150, 399 organizations, 390 oscillation, 42, 300 oscillations, 259, 296, 298, 303, 304, 357, 358, 364, 365 oxidants, 6 oxidation, xii, 6, 7, 81, 83, 123, 155, 220, 367, 368, 371, 372, 376 oxidative, 11, 12, 24, 25, 390 oxidative damage, 24, 390 oxidative stress, 11, 12, 390 oxide, 4, 10, 14, 123, 245, 246 oxides, 4, 6, 7, 21, 26, 61, 65, 104, 110, 159 oxygen, xii, 27, 231, 367, 368, 370, 371, 372, 373, 374, 375, 376, 377 ozone, 4, 8, 11, 19, 26, 27, 28, 110, 123, 136, 146, 152

P P. pinea, 165, 167, 168, 170, 173 Pacific, 65, 114 PAHs, 11, 154 pain, 14, 15 Pakistan, 146 PAN, 152 Pap, 96 paper, 28, 60, 61, 67, 123, 149, 151, 159, 214, 218, 243, 244, 245, 246, 247, 248, 249, 263, 334, 339, 340, 352, 354, 365, 383, 398 parameter, 125, 147, 154, 184, 217, 256, 260, 275, 276, 277, 278, 294, 353 parents, 381 Paris, 133, 380, 382 Parkinson, 146, 365 Parliament, 380 partial differential equations, 122, 255 particle mass, 71, 73, 76, 82, 84, 87, 89, 93, 94, 95 particles, viii, 3, 5, 6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17, 20, 22, 23, 28, 69, 70, 80, 84, 88, 89, 90, 93, 96, 97, 98, 100, 101, 137, 146, 153, 154, 155, 158, 160, 177

particulate matter, vii, 1, 2, 8, 9, 10, 11, 12, 20, 24, 27, 45, 68, 69, 70, 90, 96, 97, 98, 99, 104, 107, 110, 121, 122, 124, 131, 146, 152, 155, 156, 158, 159, 162, 177 partition, 306 passenger, 34, 35, 39, 53, 61, 62, 63, 64, 68, 117, 154, 155, 156, 157, 159, 223, 247 passive, 178 pathogenesis, 392 pathophysiological, 17 pathways, 398 patients, 1, 2, 3, 365 PCA, 166, 172, 174 PCT, 392 peak concentration, 153 pedestrian, 62 pediatric, 27 penalty, 214, 224, 231, 241 Pennsylvania, 381, 384 percentile, 157 perception, 59 perfect gas, 253, 311 performance, 31, 46, 60, 70, 106, 123, 189, 206, 207, 209, 220, 245, 246, 247, 365, 393 periodic, 255, 263, 357, 358, 361, 364 periodicity, 19, 325, 357 peripheral blood, 7, 24, 391, 394 peripheral blood lymphocytes, 7, 391 permeability, 5, 8, 18 permit, 224 permittivity, 205 peroxidation, 5, 7, 26 personal, viii, 30, 43, 44, 54, 103, 104, 105, 107, 118, 122, 354, 387, 391 perturbation, 288, 298, 299, 301, 302, 303, 306 perturbations, 258 pesticides, 380, 383 petroleum, 125, 381, 399 pH, 7, 22 pH values, 7 pharmacology, 28 pharynx, 17 phenol, 390, 393, 399, 400 Philadelphia, 27 Phoenix, 175, 176, 177 phone, 383 phosphate, 13, 21 phosphorylation, 22 photochemical, 4, 6, 8, 45, 152 photosynthesis, 162, 165, 213 photosynthetic, 165, 170, 173, 176 photovoltaic, 242 photovoltaics (PV), 213, 243 physical environment, 163 physical properties, x, 179 piezoelectric, 354, 357 pigments, 15 Pinus halepensis, 174 planar, 112, 336

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index planning, vii, viii, 46, 55, 62, 63, 67, 69, 90, 103, 104, 105, 113, 118, 131 plants, 14, 15, 104, 167, 175, 176, 368 plasma, 13, 16, 20, 21 plastic, 154 platelet, 17, 392, 394 platelet count, 392, 394 platelets, 17, 392 platinum, 220 play, 17, 174, 177, 182, 390, 399 Pleistocene, 163 PMA, 166 pneumonia, 5, 6, 15 pneumonitis, 16 poisoning, 136, 214, 243 poisonous, 398 Poland, 176, 355, 356 police, 387, 388, 389, 390, 391, 392, 393, 395, 398, 399, 400 policy makers, 51, 56 policy making, 131 pollen, 8 pollutant, x, 4, 8, 16, 34, 41, 48, 56, 60, 61, 64, 117, 122, 123, 129, 130, 139, 140, 158, 211 pollutants, vii, viii, x, 1, 2, 3, 5, 9, 11, 18, 25, 31, 32, 37, 38, 40, 43, 45, 50, 56, 57, 58, 59, 62, 98, 104, 107, 121, 122, 123, 124, 126, 131, 132, 136, 139, 140, 142, 143, 146, 153, 157, 159, 161, 162, 163, 173 pollution, vii, viii, ix, x, 1, 2, 3, 4, 5, 8, 10, 11, 14, 16, 18, 19, 20, 22, 23, 24, 25, 27, 28, 29, 30, 31, 32, 56, 57, 59, 63, 64, 66, 69, 70, 90, 93, 106, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 131, 133, 135, 136, 137, 139, 145, 146, 147, 151, 159, 161, 162, 163, 164, 174, 175, 176, 177, 178, 380, 382, 383, 384, 390, 392, 395, 397, 399 polycyclic aromatic hydrocarbon, 150 polyethylene, 166 polymer, 212 polymer industry, 212 polymorphonuclear, 17 polynomial, 127 pools, 176 poor, 13, 51, 106, 137, 221, 262, 391 population, 5, 12, 15, 20, 31, 74, 76, 79, 80, 81, 83, 107, 122, 127, 139, 156, 379, 381, 382, 384, 390, 391, 392 population density, 122, 127, 139 ports, 260, 261, 303, 304, 307, 322, 327, 328 power, xi, 7, 8, 35, 36, 41, 42, 65, 90, 136, 152, 175, 177, 181, 211, 213, 214, 217, 218, 219, 220, 222, 223, 224, 225, 226, 227, 235, 236, 238, 240, 241, 242, 243, 245, 246, 247, 248, 325, 352, 353, 354, 357, 360, 361, 363, 364, 371 power generation, 214, 225 power plant, 7, 8, 136, 181 power plants, 7, 8, 136, 181 Prandtl, 180, 187, 254, 257

415

preconditioning, xi, 252, 288, 289, 290, 292, 293, 294, 295, 297, 298, 301, 303, 304, 328, 330, 331, 332, 334, 338 predators, 381 prediction, vii, ix, 29, 34, 35, 40, 41, 42, 45, 51, 52, 53, 55, 56, 57, 58, 59, 65, 121, 122, 124, 125, 127, 129, 132, 334 prediction models, 124, 132 predictive model, 63 pre-existing, 8, 11 pregnancy, 13, 15 pregnant, 16 premature death, 8, 10 president, 153 pressure, xii, 42, 93, 94, 111, 112, 124, 127, 152, 179, 180, 181, 185, 186, 187, 188, 192, 214, 216, 218, 219, 221, 222, 223, 226, 227, 236, 237, 238, 239, 240, 241, 242, 253, 254, 257, 289, 290, 298, 301, 302, 303, 304, 305, 307, 308, 309, 313, 314, 315, 316, 317, 318, 320, 321, 328, 329, 331, 333, 334, 340, 347, 351, 352, 354, 355, 356, 357, 360, 361, 363, 364, 365, 368, 370, 371, 376 Pressure-Stabilizing/Petrov-Galerkin, 298 prevention, 391, 399 prices, 213 primary data, 398 primate, 23 principal component analysis, 174 private, 105, 163 private property, 105 probability, ix, 15, 58, 121, 127 probe, 369, 371 production, xii, 8, 14, 15, 17, 25, 110, 115, 174, 177, 213, 219, 223, 237, 243, 367, 368, 370, 371, 373, 377 program, 100, 104, 107, 118, 244, 253, 333 programming, 220 proinflammatory, 11, 17 proliferation, 17 promote, vii, 1, 7, 19 propagation, 288, 296, 332 property, 137, 187, 218, 263, 268, 276 proposition, 329 propulsion, 213 prostate, 14, 26, 27 prostate cancer, 26, 27 protection, 147 proteins, 5, 13 protocols, 380 prototype, 223 proximal, 13 pruning, x, 161, 165, 170, 174 pseudo, 17, 288 PSI, 19 public, 1, 5, 12, 56, 63, 107, 110, 117, 136, 145, 163, 223 public health, 5, 12, 107, 117, 136, 145 Public Health Service, 18, 19 pulmonary edema, 6, 17

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

416

Index

pulse, 316 pulses, 325 pumping, 224, 236, 345, 347 pure water, 166 pyrene, 3, 6

Q quality control, 112, 389 quality of life, 136, 145, 383 quasi-linear, 255, 258 Quercus, 163, 171, 175, 176 Quercus ilex, 163, 171, 175, 176 questionnaire, 122, 131

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

R radar, 46 radiation, 8, 152, 162, 253, 380, 383 radical, 6 radius, 263, 272, 322 rail, 226 rain, 14, 30 rainfall, 163, 165, 382, 396, 397 random, 109 range, viii, x, xi, xii, 9, 16, 18, 33, 38, 67, 69, 70, 71, 72, 73, 74, 76, 77, 79, 80, 82, 85, 88, 89, 90, 115, 122, 125, 127, 130, 131, 153, 154, 155, 171, 179, 191, 192, 193, 194, 213, 217, 218, 221, 239, 240, 241, 242, 252, 254, 259, 268, 289, 330, 354, 368, 371, 372, 379, 380, 381, 382, 391 rat, 20, 26 rats, 6, 14, 16, 17, 20, 23, 26, 27, 381 raw material, 212 Rayleigh, 205, 206 reaction mechanism, 243 reactive oxygen, 11 reactive oxygen species (ROS), 11, 12 reactivity, 23, 373 reading, 226, 237 reality, 34, 38, 58, 198, 209 recall, 137 receptacle, 13 receptor sites, 26 receptors, 31, 130 reconstruction, 354 recreation, 106, 118 recreational, 104, 105, 108, 115, 119 recruiting, 381 recurrence, 267, 370 recycling, 224 red blood cell, 13, 391, 393 red blood cells, 13 redox, 11 reduction, 12, 34, 42, 48, 51, 64, 140, 143, 144, 173, 212, 217, 237, 248, 330, 340, 342, 345, 347, 348 reference system, 256

refining, 14, 117, 370 refractory, 14 regional, 16, 32, 43, 45, 46, 47, 51, 57, 124, 132, 139, 159 Registry, 18, 19, 107, 390, 393 regression, 41, 42, 50, 63, 67, 95, 99, 122, 171, 172, 380, 389, 390, 391, 392 regression analysis, 122, 171, 389, 390, 391, 392 regression line, 380 regular, 127, 137, 152, 218, 268, 269, 270, 278 regulation, vii, ix, 69, 90, 100, 107, 113, 114, 117, 123, 136 regulations, ix, 107, 117, 118, 124, 136, 137, 145 regulators, 162 relationship, viii, 6, 20, 28, 36, 70, 88, 101, 123, 143, 144, 163, 174, 175, 178, 373, 383, 391, 397, 399 relationships, 30, 31, 32, 42, 74, 82, 84, 146 relaxation, 268, 269, 270, 277, 278 relaxation coefficient, 278 relaxation process, 268, 269 relevance, 89, 95, 174, 181 reliability, 42, 63, 72 remote sensing, 32, 33, 43, 61, 68, 93, 94, 125, 131, 156 renal, 13, 18 renal function, 18 repair, 25, 137 research, viii, xi, xiii, 26, 33, 40, 50, 55, 56, 57, 58, 59, 63, 80, 103, 105, 122, 124, 130, 149, 150, 151, 153, 163, 211, 212, 213, 214, 215, 222, 236, 238, 239, 241, 243, 244, 246, 248, 306, 348, 352, 381, 383, 387, 392, 395 Research and Development, 21 researchers, 33, 153, 158, 252, 256 reserves, 212 reservoir, 313, 314 residential, ix, 5, 39, 46, 48, 51, 108, 121, 122, 123, 129, 131, 147, 176 residuals, 217 residues, 24 resistance, 5, 8, 10, 27, 177, 217, 218, 220 resolution, xi, 33, 34, 37, 40, 48, 49, 50, 51, 56, 57, 58, 59, 130, 149, 150, 151, 152, 153, 156, 158, 251, 252, 258, 263, 266, 271, 290, 303, 312, 315, 329, 354 resources, 47, 177, 253, 306 respiratory, 2, 3, 5, 6, 8, 9, 10, 11, 12, 14, 15, 17, 22, 23, 25, 26, 27, 106, 136, 146, 158 respiratory arrest, 106 respiratory problems, 5, 6 response time, 154, 157 responsiveness, 8, 10, 26 retardation, 348 retention, 16, 162 Reynolds, 38, 50, 55, 66, 180, 255, 256, 257, 289, 290, 293, 294, 295, 298, 300, 336, 337 Reynolds number, 180, 256, 289, 290, 293, 294, 295, 298, 300 rhinitis, 27

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Index rings, 3, 217, 221 risk, xiii, 1, 2, 3, 5, 15, 17, 19, 23, 25, 104, 107, 114, 117, 122, 124, 125, 127, 129, 130, 132, 133, 146, 177, 236, 239, 387, 390, 391, 392, 395, 398, 399, 400 risk assessment, 5, 107, 124, 125, 127, 129, 130, 133 risks, 26, 106, 107, 398 road dust, 71, 73, 75, 82, 84, 85, 89, 90, 95, 151 Roads, 67 robustness, xi, 42, 72, 251, 252, 272, 278 rodent, 20 Rome, x, 161, 163, 164, 173, 174, 175, 176, 177, 387, 390, 392, 399 rotation transformation, 329 rotations, 270, 373 Royal Society, 383 RTI, 112, 118 RTI International, 112, 118 runaway, 216 rural, 23, 39, 82, 94, 123, 152, 153, 154, 162, 177, 380, 383, 400 rural areas, 153 rust, 216

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

S SAE, 60, 61, 64, 67, 95, 96, 98, 99, 159, 210, 243, 244, 245, 246, 247, 248, 249, 334, 349, 365, 378 safety, viii, 103, 105, 213, 222, 237, 240, 247, 368 sales, 35, 106, 113, 115 sample, ix, 18, 33, 71, 72, 73, 74, 82, 87, 88, 95, 110, 113, 121, 154, 388, 389, 390, 391, 392, 393, 394, 399, 400 sampling, 32, 33, 107, 110, 111, 112, 116, 139, 150, 151, 152, 154, 159, 353, 354, 357, 369, 371, 388 satellite, 125, 126 saturation, 188 Saturday, 164 savings, x, 60, 179 scalar, 263 scaling, 115, 164, 270, 289 scarcity, 83 schema, 126 schizophrenia, 3, 25 Schmid, 63, 99 Schmidt number, 188 school, ix, 121, 123 sea level, 163 seals, 322 search, 251, 265, 277 seasonal pattern, xiii, 395, 396, 397 second generation, 213 security, 212 sedentary, xii, 379, 382 sediment, 159 segmentation, 51 selecting, 77, 79 seminiferous tubules, 15, 21

417

sensing, 32, 33, 43, 61, 62, 68, 93, 94, 125, 131, 132, 156, 177 sensitivity, 15, 52, 55, 125, 132, 152, 154 sensors, 57, 62, 237 SEPA, 104 septum, 15 series, xii, 4, 6, 154, 223, 225, 246, 259, 351, 352, 353, 354, 355, 356, 357, 360, 361, 363 Sertoli cells, 18 settlements, vii, 1 severity, 2, 181, 220 sewage, 173 shade, 162, 173 Shanghai, 159 shape, 154, 266, 268, 269, 270, 299, 313, 317, 322, 324, 325 sharing, 130 shear, 256, 257, 333 Shell, 243 Sherwood number, 188 shock, 153, 258, 259, 260, 296 shocks, 153 shoot, 174 shortage, 381 shortness of breath, 14, 17 short-term, xii, 8, 10, 15, 17, 19, 20, 351, 352, 357, 358 shrubs, 163 Siberia, xii, 379 signaling, 8, 20 signaling pathway, 8 signaling pathways, 8 signals, 43, 60, 61, 125, 352 significance level, 172 signs, 13, 125 silver, 237 similarity, 87 simple linear regression, 72 simulation, ix, xi, 37, 47, 48, 55, 58, 63, 96, 121, 124, 125, 127, 129, 189, 246, 251, 252, 253, 298, 301, 304, 306, 310, 316, 325, 332, 338 simulations, 42, 97, 123, 181, 188, 205, 206, 210, 256, 281, 289, 314, 332, 334, 338 singular, 290 singularities, 275 sister chromatid exchange, 7, 24, 388, 393 sites, ix, 4, 5, 15, 26, 107, 109, 110, 122, 123, 124, 127, 139, 149, 150, 152, 163, 164, 165, 166, 167, 168, 169, 170, 174, 177, 380 skewness, 263 skin, 15, 16, 316, 317, 318, 391 sleep, 3, 122 sleep disorders, 3 sleep disturbance, 122 sludge, 173 SMA, 105 smog, 4, 6, 8, 45, 62 smoke, 3, 14, 27, 123, 154 smokers, 11

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

418

Index

smoking, 108, 388, 390 smooth muscle, 3 smooth muscle cells, 3 smoothing, xi, 251, 263, 274, 276, 277, 278, 284, 329, 330, 335 smoothness, 322 snowmobiles, viii, 103, 104, 105, 106, 108, 109, 110, 111, 112, 113, 114, 115, 116, 117, 119 SO2, ix, 2, 6, 7, 9, 23, 44, 45, 67, 123, 135, 136, 137, 139, 140, 142, 143, 145, 156 social benefits, x, 161, 163 sodium, 7, 24 software, ix, xi, 41, 72, 121, 122, 124, 125, 127, 166, 211 soil, 9, 15, 132, 174, 178 soil particles, 15 soils, 173, 175 solar, 110, 213 solar energy, 213 solid waste, 20, 21, 22 solubility, 16 solutions, xi, 32, 252, 253, 255, 257, 258, 262, 288, 292, 293, 301, 302, 304, 307, 311, 314, 317, 332, 335 soot, 155, 370 sound speed, 289 South Africa, xii, 379 South America, xii, 379 Southampton, 64, 179, 210 Southeast Asia, 393, 394, 400 Spain, 146, 247 spatial, ix, x, 17, 19, 31, 33, 34, 37, 48, 49, 50, 57, 58, 59, 89, 108, 110, 121, 122, 123, 124, 125, 126, 127, 129, 130, 131, 149, 150, 151, 152, 153, 158, 177, 253, 262, 264, 271 spatial analysis, ix, 110, 121, 125, 130 spatial information, 131 spatial location, 123 spatial memory, 17 spatiotemporal, 105, 133 speciation, 21, 67 species, x, 7, 11, 26, 27, 161, 162, 163, 165, 167, 168, 170, 173, 174, 175, 176, 379, 381, 382 specific heat, 179, 180, 186, 192, 254, 340, 349 specificity, 24 spectrum, 221, 352, 353, 354, 357, 361 speech, 3 speed, x, 30, 36, 37, 38, 39, 40, 41, 42, 43, 46, 47, 48, 49, 50, 51, 52, 55, 56, 57, 58, 60, 61, 62, 64, 65, 66, 67, 70, 71, 75, 82, 84, 85, 90, 94, 111, 115, 122, 126, 129, 155, 162, 179, 181, 182, 183, 184, 185, 189, 191, 193, 194, 196, 197, 198, 199, 200, 201, 205, 223, 225, 226, 227, 228, 229, 230, 231, 234, 235, 236, 238, 239, 240, 241, 246, 247, 259, 260, 285, 292, 296, 304, 316, 328, 341, 342, 343, 344, 345, 346, 348, 364 speed limit, 30, 39, 46, 56, 65, 84, 94 sperm, 13, 15 spermatocytes, 18

spermatogenesis, 18 spermatogonium, 18 spin, 184 spines, 17 spleen, 7, 12, 16, 17 SPSS, 72, 389 Sri Lanka, xii, 379 stability, 109, 222, 224, 229, 262, 285, 289 stabilization, 21, 260, 295, 296, 297, 337 stabilize, 259, 295, 296 stages, 262, 263, 325 standard error, viii, 69, 72, 76, 77, 80, 81, 83 standards, 2, 30, 34, 43, 50, 53, 80, 106, 114, 136, 137, 155, 209 starvation, 381 statistical analysis, viii, 35, 62, 69, 71, 73, 74, 84, 89, 90, 93, 95, 121, 255, 389 statistics, 36, 37, 117 steady state, 185, 289, 290, 292, 293 stiffness, 264, 286 stochastic, 132, 365 stomach, 7 storage, 53, 131, 158, 177, 213, 219, 225, 231, 245, 247 strain, 256 strategic, 46, 55 strategic planning, 46, 55 strategies, vii, xi, xii, 48, 69, 117, 122, 124, 192, 195, 197, 198, 208, 223, 226, 238, 246, 251, 255, 264, 268, 278, 281, 284, 287, 309, 313, 330, 351, 364 strategy use, 261 streams, 54, 58, 111, 116 stress, 8, 11, 12, 13, 20, 41, 106, 122, 253, 254, 255, 257, 333, 390 stress level, 122 stroke, 24, 114, 180, 183, 189, 191, 205, 216, 219, 220, 283, 284, 303, 322, 324, 325, 326, 340 stroke volume, 340 strokes, 3 students, 242, 388 subacute, 22 subsistence, 105 subsonic, 285, 288 substances, 9, 22, 107, 154, 162 suburban, 152, 163, 383 suburbs, 380 suffering, 5 sulfur, 6, 19, 23, 24, 26, 27, 98, 104, 123, 137 sulfur dioxide, 6, 19, 23, 24, 26, 27, 123, 137 sulfur oxides, 26, 104 sulfuric acid, 6 sulphur, 6, 7, 23, 24, 27, 28, 94, 98, 155 summer, x, 8, 159, 161, 164, 166, 167, 168, 169, 382 Sunday, 152 sunlight, 6, 8 supercomputers, 336 superoxide, 26 superoxide dismutase, 26

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Index supervision, 364 SUPG, 258, 260, 295, 296, 311, 333, 336, 337 supply, x, 31, 158, 179, 181, 368, 372, 373 suppression, 215, 217 surface area, 38, 71, 90, 112, 152, 153, 154, 165, 322 surface tension, 181, 205, 209 surgery, 3 surplus, 237 susceptibility, 2, 5, 8, 10, 11 suspensions, 12 sustainable development, 383 Sweden, 93, 95, 97, 98 switching, 261 Switzerland, 63, 67, 84, 93, 95, 97, 100, 124, 152, 154, 158, 177 symbols, 353 symmetry, 281, 322, 325 symptoms, 5, 10, 11, 13, 18, 22, 28, 136, 158 syndrome, 13 synthesis, ix, 11, 121, 367, 368, 370, 371, 373, 376, 377 system analysis, 42 systems, 5, 12, 33, 46, 57, 63, 64, 124, 125, 131, 136, 152, 181, 212, 220, 223, 246, 333, 334, 335

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

T Taiwan, ix, 135, 136, 137, 138, 145, 147 targets, 21 technicians, 242 technology, vii, viii, xii, 29, 30, 33, 34, 35, 38, 42, 46, 53, 56, 57, 59, 64, 67, 103, 106, 107, 123, 124, 139, 155, 162, 182, 209, 212, 213, 214, 223, 245, 367, 368 teeth, 12, 13, 27 Teflon, 93, 94 temperature, x, xii, 4, 25, 30, 32, 42, 111, 114, 152, 154, 161, 162, 163, 164, 166, 167, 168, 172, 173, 174, 175, 176, 180, 185, 186, 187, 188, 195, 214, 215, 216, 217, 220, 221, 225, 234, 236, 237, 238, 240, 254, 256, 257, 298, 313, 328, 332, 340, 341, 343, 344, 347, 348, 367, 368, 369, 370, 372, 373, 375, 376, 377, 382 temporal, x, 31, 33, 37, 40, 50, 57, 58, 59, 108, 127, 130, 131, 149, 151, 158, 187, 253, 271, 291, 381, 383 temporal distribution, x, 31, 149, 151 tension, 180 teratogen, 14 territory, 163 test data, 32, 33, 35, 41, 42, 43, 50, 54, 55, 59 test items, 139 testes, 18, 24 testicle, 14 testosterone, 18 Thai, 388, 390, 393, 395, 397, 399, 400 Thailand, 132, 387, 388, 395, 396, 399, 400 theory, 146, 213, 306

419

Thermal Conductivity, 180 thermal efficiency, vii, xii, 226, 227, 228, 229, 230, 241, 242, 339, 340, 341, 342, 343, 344, 345, 346, 347, 348, 349, 370 thermal load, 221, 225 thermodynamic, x, xi, 179, 182, 195, 252, 254, 322 thermodynamics, xi, 185, 251, 252 thoracic, 3 threat, 12, 107 three-dimensional, 122, 253, 255 threshold, 26, 231, 232, 233, 236, 238, 239, 240, 241, 242 throat, 2, 16 thrombocytosis, 17 thromboembolic, 1, 17 thymus, 7 tics, 5 time, xi, xii, 6, 13, 14, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 46, 48, 50, 53, 54, 56, 57, 58, 73, 105, 106, 112, 113, 114, 125, 127, 137, 140, 152, 153, 155, 156, 157, 158, 162, 164, 181, 183, 184, 185, 186, 187, 190, 191, 192, 199, 200, 201, 202, 203, 204, 205, 206, 207, 214, 215, 216, 220, 222, 225, 226, 231, 232, 234, 235, 236, 242, 252, 253, 255, 258, 259, 260, 261, 262, 264, 267, 268, 271, 272, 273, 274, 276, 278, 279, 280, 281, 282, 283, 284, 285, 288, 289, 290, 291, 294, 295, 296, 300, 301, 304, 306, 312, 314, 316, 320, 328, 329, 330, 331, 332, 333, 351, 352, 353, 354, 355, 357, 360, 361, 363, 381, 382, 388 time periods, 112, 127 time resolution, 33, 50, 152, 153, 156, 158 time series, xii, 351, 353, 354, 357, 360, 361, 363 timing, x, xi, 42, 51, 58, 75, 83, 179, 182, 211, 217, 220, 221, 224, 225, 226, 231, 232, 234, 237, 242, 246, 283, 303, 339, 341, 342, 343, 344, 345, 346, 347, 348, 380, 383 tissue, 6, 10, 11, 12, 13, 17, 18, 21 TNF, 8 toddlers, 5 Tokyo, 135 tolerance, 271, 277 toluene, viii, 103, 106, 107, 110, 114, 115, 116, 393 top-down, 37 topological, 263, 322 topology, xi, 251, 252, 262, 263, 269, 276, 283, 325 torque, 42, 226, 227, 228, 229, 230, 231, 232, 233, 234, 235, 238, 239, 240, 241, 242, 354, 357, 361 total energy, 254 toxic, viii, 6, 10, 12, 13, 14, 15, 16, 98, 103, 104, 105, 106, 107, 111, 114, 116, 119, 121, 123, 150, 381, 382, 383, 390, 397, 399 toxic effect, 10 toxic gases, 121, 123 toxic metals, 12 toxic substances, 107 toxicities, xiii, 395 toxicity, xiii, 5, 6, 11, 12, 13, 15, 18, 20, 25, 26, 27, 176, 387, 390, 393, 398, 399

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

420

Index

toxicological, 7, 11, 106 toxicology, 19, 20, 23, 25 toxins, 8, 11 Toyota, 223 trabecular bone, 13 trachea, 6 trade-off, 47, 54, 235 traffic, vii, ix, x, xiii, 1, 2, 3, 4, 18, 25, 27, 29, 30, 31, 32, 33, 34, 35, 37, 38, 39, 40, 42, 43, 46, 47, 48, 49, 50, 51, 52, 54, 55, 56, 57, 58, 59, 61, 62, 63, 64, 65, 66, 67, 82, 96, 97, 98, 99, 100, 101, 106, 107, 110, 113, 115, 121, 122, 123, 124, 125, 126, 127, 128, 129, 130, 132, 133, 135, 136, 138, 139, 145, 146, 147, 149, 150, 151, 152, 153, 156, 157, 158, 159, 161, 162, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 177, 178, 382, 387, 388, 389, 390, 391, 392, 393, 395, 397, 398, 399, 400 traffic flow, 31, 48, 56, 66, 67, 132 traits, x, 161, 169, 170, 173, 174, 175 trajectory, 322 trans, 213, 388, 390, 392, 393, 394, 399, 400 transcription, 11 transcription factor, 11 transcription factors, 11 transducer, 155 transfer, xi, xii, 16, 119, 179, 187, 188, 195, 217, 251, 252, 333, 339, 340, 348, 349 transferrin, 16 transformation, 9, 25, 31, 269, 270, 329 transformation matrix, 270 transformation product, 9 transformations, 270 transition, 8, 11, 12, 20, 21, 22, 28, 214 transition metal, 8, 11, 12, 20, 21, 22, 28 translation, 265, 270, 353 translational, 322 transmission, 30, 34, 41, 308 transpiration, x, 161, 162 transport, vii, viii, x, 31, 32, 43, 46, 51, 63, 64, 67, 69, 70, 71, 73, 74, 75, 76, 90, 93, 99, 100, 110, 129, 146, 163, 211, 212, 213, 223, 244, 335 transportation, vii, ix, 1, 5, 62, 63, 64, 106, 136, 137, 140, 144, 145, 146, 213 transpose, 254 traps, 162 trauma, 106 travel, 9, 14, 35, 38, 46, 48, 55, 98 travel time, 38, 46, 48 tree cover, 164 trees, x, 161, 162, 163, 165, 169, 173, 174, 175, 176, 177, 178 tremor, 13 trend, 43, 54, 127, 166, 168, 169, 173, 174, 212, 232, 234, 341 trial, 258, 259 tribal, 104 TRIPS, 46 trucks, viii, 34, 37, 40, 94, 121, 124, 125

Tsuga, 349 tubular, 13 tumor, 8 tumor necrosis factor, 8 tumors, 15, 399 turbulence, xi, 152, 157, 181, 219, 222, 229, 251, 252, 255, 256 turbulent, 200, 201, 241, 253, 255, 256, 257, 332, 338 Turbulent, 180, 337 turbulent flows, 253, 256, 257 Turkey, 175 two-dimensional, 41, 304, 325, 329 tyrosine, 17 tyrosine hydroxylase, 17

U U.S. Department of Agriculture, 175, 177 ulceration, 15 ultra-fine, 8 ultraviolet, 4, 8, 139 ultraviolet light, 4 uncertainty, 52, 127 uniform, 187, 298, 313, 314, 316, 380 United Kingdom, 24, 62, 93, 95, 100, 124, 159, 164, 210, 243, 244, 380, 383 United Nations, 243, 311, 312 United States, viii, 38, 103, 104, 107, 118, 119, 131, 175, 177 univariate, 72 universal gas constant, 188 upload, 118 upper respiratory infection, 2, 5, 23 upper respiratory tract, 10, 106 urban areas, vii, x, 4, 14, 29, 50, 69, 70, 71, 90, 122, 123, 132, 133, 150, 151, 161, 162, 163, 173, 174, 176, 380, 395 urban population, xiii, 24, 380, 395, 397 urbanisation, 163, 174 uric acid, 8 urinary, 400 urine, 12, 16, 388, 389, 390, 391, 392, 393, 394, 398, 399, 400 US Department of Health and Human Services, 19 USEPA, viii, 44, 70, 101, 103, 104, 105, 106, 107, 113, 115, 116, 117, 118, 119 user-defined, 277 USSR, 16 Utah, 21 UV light, 381

V vacuum, 222, 234, 237, 369 validation, 33, 55, 56, 59, 62, 66, 127, 188, 248 validity, 55

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

Index values, 7, 21, 33, 38, 51, 72, 73, 74, 75, 76, 80, 81, 82, 84, 85, 86, 88, 89, 90, 112, 113, 114, 115, 116, 140, 143, 144, 145, 153, 157, 166, 167, 168, 169, 170, 173, 182, 191, 197, 199, 214, 221, 231, 236, 238, 271, 272, 278, 283, 313, 317, 325, 328, 330, 331 vanadium, 6, 12, 16, 17, 18, 19, 20, 21, 23, 24, 25, 26, 27 vanadium pentoxide, 16, 17, 18, 19, 20, 21, 23, 24, 26, 27 vapor, xiii, 388, 395, 398, 400 variability, 41, 50, 52, 53, 54, 57, 58, 60, 62, 72, 85, 108, 127, 133, 352, 365 variable, xi, 32, 37, 39, 40, 41, 47, 49, 50, 51, 55, 57, 58, 71, 72, 84, 88, 95, 166, 172, 211, 221, 226, 231, 234, 242, 265, 276, 297, 311, 351, 370 variables, viii, 30, 35, 36, 39, 40, 41, 42, 55, 58, 59, 62, 69, 70, 71, 72, 73, 74, 76, 77, 79, 80, 81, 82, 83, 84, 85, 88, 89, 93, 95, 114, 166, 171, 184, 254, 255, 259, 260, 276, 289, 290, 291, 295, 296, 297, 313, 333, 336, 351, 368, 370, 377, 382 variance, 53, 166, 172, 173, 174, 353, 354, 361 variation, 55, 72, 73, 75, 77, 79, 80, 81, 82, 84, 89, 90, 105, 118, 150, 152, 153, 175, 176, 177, 190, 219, 322, 348, 361, 396, 397, 399 vector, 253, 254, 258, 259, 260, 267, 268, 290, 291, 297, 311, 316, 337 vegetation, 163, 178, 382 vehicles, vii, viii, x, xi, 3, 5, 8, 30, 32, 33, 34, 35, 38, 39, 40, 42, 43, 44, 45, 46, 48, 50, 53, 54, 56, 57, 58, 59, 60, 61, 62, 67, 69, 70, 71, 73, 74, 75, 77, 80, 82, 85, 86, 89, 90, 93, 94, 95, 96, 97, 98, 99, 104, 105, 106, 109, 110, 111, 112, 116, 117, 118, 119, 123, 124, 125, 127, 129, 132, 136, 137, 146, 149, 151, 155, 156, 157, 158, 159, 160, 174, 211, 212, 213, 222, 223, 241, 247, 249, 381 vein, 17 velocity, xii, 111, 180, 182, 183, 184, 185, 186, 187, 192, 199, 200, 201, 215, 219, 227, 253, 255, 256, 257, 258, 259, 260, 261, 268, 285, 288, 290, 294, 296, 298, 299, 300, 301, 302, 305, 310, 314, 315, 316, 319, 329, 330, 339, 340, 343, 348, 365 ventilation, 122, 213, 222 versatility, 166 veterans, 146 vibration, 153, 216, 322 village, 400 virus, 382 viscosity, 215, 221, 222, 253, 254, 256, 257, 335 visible, 198, 232, 357 vision, 3, 136 visualization, 124, 125, 130, 133 volatile substances, 388 vomiting, 14, 15 vortex, 302

421

W Wales, 365, 383 walking, 147 wall temperature, 256, 340 war, 105 warrants, 212 waste disposal, 15 waste disposal sites, 15 waste incineration, 173 waste incinerator, 24 wastewater, 368 wastewater treatment, 368 water, 6, 14, 15, 107, 110, 137, 159, 166, 216, 219, 220, 221, 236, 237, 246 water quality, 159 water-soluble, 11 wavelengths, 162, 291 wavelet, xii, 351, 352, 353, 354, 355, 357, 360, 361, 364, 365 wavelet analysis, xii, 351, 352, 355, 364, 365 weakness, 391 wealth, 223 wear, 71, 90, 174, 221, 222, 322 weathering, 14 well-being, 136 West Africa, xii, 379 Western Europe, 136 wheezing, 15, 27 wildlife, 106, 111 wind, 15, 111, 122, 126, 129, 132, 152, 153, 157, 176, 213, 243, 397 wind turbines, 213 windows, 122 winter, viii, 103, 104, 105, 106, 107, 110, 115, 118, 163, 164, 166, 167, 168, 169, 173, 382, 397 wintertime, 110, 117 wireless, 57 women, 1, 11 wood, 9 woods, 163 workers, 7, 12, 13, 14, 15, 22, 23, 24, 25, 42, 217, 387, 390, 392, 394, 399, 400 working conditions, 217 World Health Organization, 2, 4, 5, 6, 8, 9, 10, 17, 18, 28, 50, 68, 117

X xylene, 393 xylenes, viii, 103, 106, 114, 115, 116

Y Yellowstone National Park, viii, 103, 104, 105, 109, 118, 119

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.

422 yield, 9, 213, 368, 370, 373, 376, 377

Index

Z

Copyright © 2009. Nova Science Publishers, Incorporated. All rights reserved.

zinc, 12, 27, 166, 171, 172, 173, 174

Demidov, Sergey, and Jacques Bonnet. Traffic Related Air Pollution and Internal Combustion Engines, Nova Science Publishers, Incorporated, 2009.