Accelerated Pavement Testing to Transport Infrastructure Innovation: Proceedings of 6th APT Conference [1st ed.] 9783030552350, 9783030552367

Table of contents :
Front Matter ....Pages i-xviii
Front Matter ....Pages 1-1
Promoting Implementation of Significant Findings from the NCAT Pavement Test Track (R. Buzz Powell)....Pages 3-11
The Design, Construction and Operation of the APT Facility at the University of Texas at Arlington (Stefan A. Romanoschi, Constantin Popescu, Ana-Maria Coca, Mohsen Talebsafa)....Pages 12-20
Research Progress of RIOHTRACK in China (Xu Dong Wang, Lei Zhang, Xing Ye Zhou, Qian Xiao, Wei Guan, Ling Yan Shan)....Pages 21-31
Two Years of APT Program on the New Test Site duraBASt (Bastian Wacker, Dirk Jansen)....Pages 32-39
Guidance for the Next Generation Accelerated Pavement Testing Facilities (Benjamin Worel, Michael Vrtis, R. Buzz Powell)....Pages 40-48
Front Matter ....Pages 49-49
Laboratory Investigation of Cracked Asphalt Pavement Structure Using Accelerated Loading System (Youness Berraha, Daniel Perraton, Guy Doré, Michel Vaillancourt, Jean-Pascal Bilodeau)....Pages 51-60
Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement Using Accelerated Pavement Test (Rongji Cao, Chunying Wu, Bingfeng Zheng, Zhenglong Lv, Yi Huang, Ying Gao)....Pages 61-69
Performance of Unbound Pavement Materials in Changing Moisture Conditions (Marit Fladvad, Sigurdur Erlingsson)....Pages 70-79
Performance Evaluation of Asphalt Pavement with Semi-rigid Base and Fine-sand Subgrade by Indoor Large-Scale Accelerated Pavement Testing (J. T. Wu, Y. T. Wu)....Pages 80-89
Accelerated Fatigue Damage Profile of Asphalt Concrete Placed on Semi-rigid Layer (Yi Li, Jiahao Li, Liping Liu, Lijun Sun)....Pages 90-99
Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic Simulations to Mitigate the Damage Effects on Road Structures (Paul Marsac, Juliette Blanc, Olivier Chupin, Thomas Gabet, Ferhat Hammoum, Navneet Garg et al.)....Pages 100-107
Impact of Mix Design Optimization on HMA Rutting Performance Under Accelerated Pavement Testing (Fabrizio Meroni, Wenjing Xue, Gerardo W. Flintsch, Brian K. Diefenderfer)....Pages 108-117
Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams from Accelerated Pavement Testing Trial (Sameera Pitawala, Arooran Sounthararajah, James Grenfell, Didier Bodin, Jayantha Kodikara)....Pages 118-126
Effect of the Incremental Loading Conditions in the Permanent Deformation in Heavy Vehicle Simulator Tests (Erdrick Pérez-González, Jean-Pascal Bilodeau, Guy Doré)....Pages 127-136
Comparison of In-situ Response from Companion Asphalt Pavement Sections in Different Climates (Michael Vrtis, Benjamin Worel, David H. Timm)....Pages 137-146
Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate via Full-Scale Accelerated Testing (Xingdong Wu, Mustaque Hossain, Greg Schieber)....Pages 147-156
Experimental Methods for Material Selection, Quality Control, and Forensic Investigations of Asphalt Paving Materials (Tianhao Yan, Jhenyffer Matias de Oliveira, Mugurel Turos, Mihai Marasteanu, Michael Vrtis, Dave van Deusen)....Pages 157-166
Front Matter ....Pages 167-167
Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced with Metal Fibers (Alfonso Pérez, Federico Castro, Miguel Ángel Franco, Paul Garnica)....Pages 169-177
The Challenges of the Accelerated Testing of Jointed Concrete Pavements (Angel Mateos, Rongzong Wu, John Harvey, Julio Paniagua, Fabian Paniagua, Robel Ayalew)....Pages 178-185
Accelerated Pavement Testing for the Evaluations of Structural Design and Safety Performance of an Innovative Road Coating (Mai Lan Nguyen, Pierre Hornych, Minh Tan Do, Thierry Sedran, Duc Tung Dao)....Pages 186-195
Experimental Investigation of Wheel-Load Induced Strain Responses in Roller Compacted Concrete Pavements (Moinul Mahdi, Yilong Liu, Zhong Wu, Tyson Rupnow)....Pages 196-205
Front Matter ....Pages 207-207
Failure Modes of Rapid-Setting Concrete Repairs Under Accelerated Aircraft Traffic (Lulu Edwards, Haley P. Bell, Jeb S. Tingle)....Pages 209-217
Evaluation of Warm Mix Asphalt (WMA) Technologies for Use in Airport Pavements (Navneet Garg, Hasan Kazmee, Lia Ricalde)....Pages 218-227
Impact of Heavy Airplanes on Asphalt Surface Behaviour, a Need for a Different Material Design (Laurent Porot, David Bell, Erik Scholten, Robert Kluttz)....Pages 228-237
The Evolution of Accelerated Pavement Testing for U. S. Military Airfield Pavements at the Waterways Experiment Station (Timothy W. Rushing)....Pages 238-247
Front Matter ....Pages 249-249
Heavy Vehicle Simulator (HVS) Testing of Innovative Pavement Structures in South Africa (I. Akhalwaya, M. A. Smit, F. C. Rust, L. Du Plessis)....Pages 251-260
From Laboratory Mixes Evaluation to Full Scale Test: Fatigue Behavior of Bio-Materials Recycled Asphalt Mixtures (Juliette Blanc, Emmanuel Chailleux, Pierre Hornych, Chris Williams, Zahra Sotoodeh-Nia, Laurent Porot et al.)....Pages 261-269
Evaluation of Cold Central-Plant Recycling (CCPR) Technique Using Full-Scale Accelerated Pavement Testing (Gerardo Flintsch, Wenjing Xue, Brian Diefenderfer, Fabrizio Meroni)....Pages 270-279
Structural Performance Evaluation of Block Pavements Using Heavy Vehicle Simulator (Shafiqur Rahman, Erik Simonsen, Fredrik Hellman, Abubeker Ahmed, Sigurdur Erlingsson)....Pages 280-288
Front Matter ....Pages 289-289
Downscaled Accelerated Trafficking of Novel Asphalt Joints Based on the Induction Heating Technology (Martin Arraigada, M. Bueno, M. N. Partl)....Pages 291-299
Deformation Performance of Foamed Bitumen Stabilised Pavements Under ALF Full-Scale Accelerated Loading (Didier Bodin, James Grenfell, Geoff Jameson)....Pages 300-308
Lessons Learnt from Accelerated Pavement Testing of Full-Depth Recycled Material Stabilized with Portland Cement (Stefan Louw, David Jones, Rongzong Wu)....Pages 309-318
Evaluating the Feasibility of Using an Engineered Cementitious Composite in the Rehabilitation of Pavements by Means of APT and Laboratory Tests (Camilo Andrés Múñoz Rodriguez, Washington Peres Núñez, Jorge Augusto Pereira Ceratti, Lélio Antônio Brito, Ângela Gaio Graeff, Luiz Carlos Pinto da Silva Filho)....Pages 319-328
Design of Reinforced Pavements with Glass Fiber Grids: From Laboratory Evaluation of the Fatigue Life to Accelerated Full-Scale Test (Mai Lan Nguyen, Cyrille Chazallon, Mehdi Sahli, Georg Koval, Pierre Hornych, Daniel Doligez et al.)....Pages 329-338
Full-Scale Evaluation of Balanced Cold in-Place Recycling (CIR) Asphalt Mixtures Using a Heavy Vehicle Simulator (Ahmed Saidi, Ayman Ali, Yusuf Mehta, Ben C. Cox, Wade Lein)....Pages 339-347
Micro Surfacing Performance Under Accelerated Pavement Testing (Adriana Vargas-Nordcbeck, Jason Nelson, Buzz Powell)....Pages 348-356
Large-Scale Test to Determine the Crack Retarding Ability of Asphalt Reinforcements (Jens Wetekam, Konrad Mollenhauer, Michael Wistuba, Stephan Büchler, Michael Schmalz, Thomas Ziegler)....Pages 357-366
Front Matter ....Pages 367-367
Development and Calibration of Performance Models Based on APT Data (Tania Ávila-Esquivel, José P. Aguiar-Moya, Edgar Camacho-Garita, Luis Guillermo Loria-Salazar)....Pages 369-378
Back-Calculation of the Moduli of Asphalt Pavement Layer Using Accelerated Pavement Testing Data (Huailei Cheng, Yuhong Wang, Liping Liu, Lijun Sun, Yue Hu, Yi Li)....Pages 379-388
Simulation of Damage Scenarios in a Bituminous Pavement Tested Under FABAC ALT Using M4-5n (Olivier Chupin, Jean-Michel Piau, Armelle Chabot, Hanan Nasser, Mai Lan Nguyen, Yann Lefeuvre)....Pages 389-398
Evaluation of Permanent Deformation Models for Flexible Pavements Using Accelerated Pavement Testing (Yared Dinegdae, Sigurdur Erlingsson)....Pages 399-408
Application of a General Methodology to Analyze Strain Data Obtained from Accelerated Pavement Testing (Andrae Francois, Ayman Ali, Yusuf Mehta)....Pages 409-417
Bayesian Calibration of ILLI-THERM for Temperature Prediction Within Airfield Concrete Pavement (Osman Erman Gungor, Imad L. Al-Qadi, Navneet Garg)....Pages 418-427
Development of a Distress Prediction Model Based on Deflection Parameters (Piero Laurent-Matamoros, Tania Avila-Esquivel, Edgar Camacho-Garita, Jose Aguiar-Moya, Luis Guillermo Loria-Salazar)....Pages 428-437
Modelling Asphalt Pavement Responses Based on Field and Laboratory Data (Andreas Loizos, Konstantinos Gkyrtis, Christina Plati)....Pages 438-447
Data Gathering and Evaluation of Tensile Strains Measured in APT with Mathematical Computation Method (Lubos Remek, Veronika Valaskova)....Pages 448-457
Use of Heavy Vehicle Simulator-Aircraft in the Development of the CBR-Beta Procedure (Jeremiah M. Stache, Carlos R. Gonzalez)....Pages 458-466
Full Scale Testing with the Mobile Load Simulator: Advanced Measurements Related to Pavement Behavior and Surface Layer Damage (Yamina Oubahdou, Philippe Reynaud, Christophe Petit, Anne Millien, Jérome Dopeux, Mickael Metrope et al.)....Pages 467-475
Front Matter ....Pages 477-477
Water Pressures in Porous Asphalt (David Alabaster, Frank Greenslade, Sabine Leischner)....Pages 479-487
Procedure for Temperature Correction of Strains Measured in a Road Pavement (Maria Barriera, Juliette Blanc, Emmanuel Chailleux, Simon Pouget, Julien Van Rompu, Bérengère Lebental)....Pages 488-496
Relationship Between Pavement Damage and the Combined Action of Frost Penetration and Axle Load (Jean-Pascal Bilodeau, Erdrick Perez Gonzalez)....Pages 497-506
APT Instrumentation – Where Do I Begin? (Harold T. Carr, W. Jeremy Robinson)....Pages 507-515
Towards an Adapted Ovalization System for Flexible Airfield Pavement Interface Characterization Using Rolling-Wheel or HWD Loads (Maissa Gharbi, Michael Broutin, Thomas Schneider, Maindroult Stephane, Armelle Chabot)....Pages 516-525
Evaluation of the Use of Geophones and Accelerometers for Monitoring Pavement Deflections, Using Accelerated Pavement Tests (Natasha Bahrani, Juliette Blanc, Pierre Hornych, Fabien Menant)....Pages 526-535
Dynamic Response Monitoring and Analysis of In-Service Asphalt Pavement Based on FBG Measuring Technology (Xianyong Ma, Tongxu Wang, Zejiao Dong)....Pages 536-544
Monitoring Road Pavement Performance Through a Novel Data Processing Approach, Accelerated Pavement Test Results (Mario Manosalvas-Paredes, Nizar Lajnef, Karim Chatti, Juliette Blanc, Nick Thom, Gordon Airey et al.)....Pages 545-554
Pavement Instrumentation and WIM Data from a Test Track on the BR-101 Highway in South of Brazil (Gustavo Garcia Otto, Amir Mattar Valente, Bruno de Melo Gevaerd, Rafael Aleixo de Souza, Adosindro Joaquim de Almeida, Keyla Junko Chaves Shinohara)....Pages 555-563
Instrumentation Response of Full-Scale Multi-axial Geogrid Stabilized Flexible Pavements (W. Jeremy Robinson, Jeb S. Tingle, Mark H. Wayne, Jayhyun Kwon, Gregory Norwood)....Pages 564-573
Temperature Normalization of Flexible Pavement Response Measurements (David Timm)....Pages 574-582
Capacitance Type Moisture Sensors in Unbound Granular Material of Pavement Structures (Michael Vrtis, Leonard Palek, Benjamin Worel)....Pages 583-591
Inverse Analysis of Pavement Layer Moduli Based on Data Collected by Buried Accelerometers and Geophones (Natasha Bahrani, Eyal Levenberg, Juliette Blanc, Pierre Hornych)....Pages 592-601
Temperature Correction Method of Deflection Basin and Stress/Strain Response of Asphalt Pavement (Qian Xiao, Xu Dong Wang, Xing Ye Zhou, Lei Zhang, Wei Guan)....Pages 602-611
Front Matter ....Pages 613-613
Thin-Bed Data Model for the Processing of GPR Data over Debonded Pavement Structures (V. Baltazart, S. Todkar, X. Dérobert, J.-M. Simonin)....Pages 615-622
Remote Monitoring System for Road Constructions (Dmitry Chirva, Vitaly Solodov, Sergey Mironchuk, Eugeny Isaev)....Pages 623-631
Radar Database Collected over Artificial Debonding Pavement Structures During APT at the IFSTTAR’s Fatigue Carrousel (X. Dérobert, V. Baltazart, J.-M. Simonin, C. Norgeot, S. Doué, O. Durand et al.)....Pages 632-639
Improvements to Backcalculation Procedure by Means of Structural Analysis Based on Deflection Parameters (Piero Laurent-Matamoros, Edgar Camacho-Garita, Tania Avila-Esquivel, Jose Aguiar-Moya, Luis Guillermo Loria-Salazar)....Pages 640-648
Rapid and Continuous Imaging for Crack Monitoring During APT Experiments (Mai Lan Nguyen, Juliette Blanc, Stephane Trichet, Thierry Gouy, Gilles Coirier, Yvan Baudru et al.)....Pages 649-657
Spectral Element Simulation of Heavy Weight Deflectometer Test Including Layer Interface Conditions and Linear Viscoelastic Behaviour of Bituminous Materials (Jean-Marie Roussel, Hervé Di Benedetto, Cédric Sauzéat, Michaël Broutin)....Pages 658-665
Backcalculation of Airfield Pavement Layer Moduli Under HWD Testing (Hao Wang, Pengyu Xie, Richard Ji)....Pages 666-675
SMS Structure Measurement System to Optimize Accelerated Pavement Testing APT (Bastian Wacker, Dirk Jansen)....Pages 676-685
Front Matter ....Pages 687-687
Evaluation of a Solution for Electric Supply of Vehicles by the Road, at Laboratory and Full Scale (Pierre Hornych, Thomas Gabet, Mai Lan Nguyen, Fabienne Anfosso Lédée, Patrick Duprat)....Pages 689-698
Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes (Bertrand Pouteau, Kamal Berrada, Sandrine Vergne, Mai Lan Nguyen, Stéphane Trichet, Thierry Gouy et al.)....Pages 699-707
Future APT – Thoughts on Future Evolution of APT (Wynand J. vdM. Steyn)....Pages 708-717
Back Matter ....Pages 719-722

Citation preview

Lecture Notes in Civil Engineering

Armelle Chabot Pierre Hornych John Harvey Luis Guillermo Loria-Salazar   Editors

Accelerated Pavement Testing to Transport Infrastructure Innovation Proceedings of 6th APT Conference

Lecture Notes in Civil Engineering Volume 96

Series Editors Marco di Prisco, Politecnico di Milano, Milano, Italy Sheng-Hong Chen, School of Water Resources and Hydropower Engineering, Wuhan University, Wuhan, China Ioannis Vayas, Institute of Steel Structures, National Technical University of Athens, Athens, Greece Sanjay Kumar Shukla, School of Engineering, Edith Cowan University, Joondalup, WA, Australia Anuj Sharma, Iowa State University, Ames, IA, USA Nagesh Kumar, Department of Civil Engineering, Indian Institute of Science Bangalore, Bengaluru, Karnataka, India Chien Ming Wang, School of Civil Engineering, The University of Queensland, Brisbane, QLD, Australia

Lecture Notes in Civil Engineering (LNCE) publishes the latest developments in Civil Engineering - quickly, informally and in top quality. Though original research reported in proceedings and post-proceedings represents the core of LNCE, edited volumes of exceptionally high quality and interest may also be considered for publication. Volumes published in LNCE embrace all aspects and subfields of, as well as new challenges in, Civil Engineering. Topics in the series include: • • • • • • • • • • • • • • •

Construction and Structural Mechanics Building Materials Concrete, Steel and Timber Structures Geotechnical Engineering Earthquake Engineering Coastal Engineering Ocean and Offshore Engineering; Ships and Floating Structures Hydraulics, Hydrology and Water Resources Engineering Environmental Engineering and Sustainability Structural Health and Monitoring Surveying and Geographical Information Systems Indoor Environments Transportation and Traffic Risk Analysis Safety and Security

To submit a proposal or request further information, please contact the appropriate Springer Editor: – Mr. Pierpaolo Riva at [email protected] (Europe and Americas); – Ms. Swati Meherishi at [email protected] (Asia - except China, and Australia, New Zealand); – Dr. Mengchu Huang at [email protected] (China). All books in the series now indexed by Scopus and EI Compendex database!

More information about this series at http://www.springer.com/series/15087

Armelle Chabot Pierre Hornych John Harvey Luis Guillermo Loria-Salazar •

•

•

Editors

Accelerated Pavement Testing to Transport Infrastructure Innovation Proceedings of 6th APT Conference

123

Editors Armelle Chabot MAST-LAMES Université Gustave Eiffel, IFSTTAR Nantes, France

Pierre Hornych MAST-LAMES Université Gustave Eiffel, IFSTTAR Nantes, France

John Harvey University of California Pavement Research Center Davis, CA, USA

Luis Guillermo Loria-Salazar Civil Engineering School University of Costa Rica San José, Costa Rica

ISSN 2366-2557 ISSN 2366-2565 (electronic) Lecture Notes in Civil Engineering ISBN 978-3-030-55235-0 ISBN 978-3-030-55236-7 (eBook) https://doi.org/10.1007/978-3-030-55236-7 © Springer Nature Switzerland AG 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

The road and airfield pavements that are a critical part of transport infrastructure have been the subject of accelerating change and innovation in recent years. This has included the development of new materials and maintenance techniques, the growth of traffic, the development of new types of vehicles, and a broadening of the scope of pavement performance requirements to include increased consideration of environmental impacts. Even as pavement design methods have progressed, modeling the long-term performance of pavement structures remains a challenge. For this reason, accelerated pavement testing (APT) facilities continue to be essential tools for the validation of transport infrastructure innovations, due to their capacity to perform full-scale tests in shorter testing times. This is confirmed by the continuous growth of the number of APT facilities in service, including facilities in more countries and increased APT capacity within countries. APT experiments play a major role in the validation of models and design methods, due to their capacity of applying controlled loading and environmental conditions. APT is also used for initial screening of new concepts for pavement structures, materials, and treatments. The increasing pace of innovation in pavements has led to innovation in APT as well, including new instrumentation and monitoring equipment, and methods of handling and using increasing amounts of data. Movement is also considered toward standardization of APT databases to improve the sharing of pavement knowledge and to facilitate use with numerical tools. This new volume of Lecture Notes for Civil Engineering presents the latest developments in full-scale accelerated pavement testing, with papers coming from nineteen countries. All the original papers published in this book have been reviewed by at least two international experts from the Scientific Committee. These papers have been finally selected for presentation at the 6th International Conference on Accelerated Pavement Testing organized by Nantes Campus of Université Gustave Eiffel, France, which has been active in APT since the 1980s, when its unique circular test facility, known as the fatigue carrousel, was developed.

v

vi

Preface

Co-sponsored by the Transportation Research Board (TRB) Committee AFD40 and following the success of the five previous conferences held in Nevada (1999), Minnesota (2004), Spain (2008), California (2012), and Costa Rica (2016), the purpose of this 6th edition is to share the knowledge on the “accelerated pavement testing to transport infrastructure innovation” among international students, scientists, and experts from industry and academic institutions. Organized in ten different parts, the contents of this proceedings reflect recent international research developments related to: • Establishing new APT facilities in different countries • Evaluation of the performance and sustainability of traditional and innovative pavement materials for road and airfield pavements • Instrumentation and monitoring in APT experiments, data analysis, and modeling • Testing of the latest maintenance and rehabilitation solutions and of new pavement concepts. Originally planned to be held in September 2020, this 6th APT Conference edition has been postponed, due to the global pandemic of COVID-19, to September 27–29, 2021. The editors of this volume would like to very sincerely acknowledge James Green, Chairman of TRB Committee AFD40, for his constant support, all members of the International Scientific Committee as well as all the authors of papers, who have contributed to the success of this conference. Armelle Chabot Pierre Hornych John Harvey Luis Guillermo Loria-Salazar

Organization

International Scientific Committee José P. Aguiar-Moya Imad Al Qadi Juliette Blanc Didier Bodin Michael Broutin Armelle Chabot Karim Chatti Bouzid Choubane Olivier Chupin Erdem Coleri Guy Doré Sigurdur Erlingsson Jeffrey Gagnon Navneet Garg Antonio Gomes Correia John Harvey Pierre Hornych Mustaque Hossain Lambert Houben Dirk Jansen David Jones Issam Khoury Jeremy D. Lea Yann-Lefeuvre Sabine Leischner Andreas Loizos Luis Guillermo Loria Salazar

University of Costa Rica, Costa Rica University of Illinois at Urbana-Champaign, USA Université Gustave Eiffel, IFSTTAR, France ARRB Group, Australia STAC, France Université Gustave Eiffel, IFSTTAR, France Michigan State University, USA Florida Department of Transportation, USA Université Gustave Eiffel, IFSTTAR, France Oregon State University, USA Université Laval, Canada VTI, Sweden FAA, USA FAA, USA University of Minho, Portugal University of California, Davis, USA Université Gustave Eiffel, IFSTTAR, France Kansas State University, USA TU Delft, Netherlands BASt, Germany UC Davis, USA Ohio University, USA University of California, Davis, USA Colas (e-mobility), France TU Dresden, Germany NTU Athens, Greece University of Costa Rica, Costa Rica

vii

viii

Angel Mateos Rebecca S. McDaniel Mai Lan Nguyen Hugues Odéon Manfred N. Partl Christophe Petit Jean-Michel Piau Christina Plati Laurent Porot Simon Poujet Bertrand Pouteau Raymond-Buzz Powell Christiane Raab Stefan A. Romanoschi Thierry Sedran Jean-Michel Simonin Wynand JvdM Steyn Nick Thom Michael C. Vrtis Bastian Wacker Rongzong Wu Zhong Wu

Organization

University of California Pavement Research Center, USA Purdue University, USA Université Gustave Eiffel, IFSTTAR, France CEREMA, France EMPA, Switzerland Université de Limoges, France Université Gustave Eiffel, IFSTTAR, France NTU Athens, Greece Kraton, Netherlands Eiffage Infrastructures, France Eurovia Research Center, France National Center for Asphalt Technology (NCAT), USA EMPA, Switzerland The University of Texas at Arlington, USA Université Gustave Eiffel, IFSTTAR (as AIPCR member), France Université Gustave Eiffel, IFSTTAR, France University of Pretoria, South Africa University of Nottingham, UK Minnesota Department of Transportation, USA BASt, Germany University of California, Davis, USA Louisiana Transportation Research Center, USA

Conference Committee of the 6th International APT Conference Chairman Pierre Hornych

MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France

Co-chairs Armelle Chabot John Harvey Luis Guillermo Loría-Salazar

MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France University of California Pavement Research Center, Davis, USA Civil Engineering School, Universidad de Costa Rica, Costa Rica

Organization

ix

Local Organizing Committee of the 6th International APT Conference President Juliette Blanc

MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France

Members Yvan Baudru Daniel Bourbotte Philippe Caquelard Armelle Chabot Olivier Chupin Gilles Coirier Agnès Gaudicheau Fadila Ghedjati Thierry Gouy Nathalie Juignet Jacques Kerveillant Isabelle Larrue Angel Mateos Fabien Menant Mai Lan Nguyen Jean-Michel Piau Jean-Michel Simonin Stéphane Trichet Franck Tartrou Benoît Velier

MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France Université Gustave Eiffel, IFSTTAR, France Université Gustave Eiffel, IFSTTAR, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France Université Gustave Eiffel, IFSTTAR, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France Université Gustave Eiffel, IFSTTAR, Nantes, France University of California Pavement Research Center, USA MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France MAST-LAMES, Université Gustave Eiffel, IFSTTAR, Nantes, France Université Gustave Eiffel, IFSTTAR, France

Contents

Establishing APT Facilities Promoting Implementation of Significant Findings from the NCAT Pavement Test Track . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . R. Buzz Powell The Design, Construction and Operation of the APT Facility at the University of Texas at Arlington . . . . . . . . . . . . . . . . . . . . . . . . . Stefan A. Romanoschi, Constantin Popescu, Ana-Maria Coca, and Mohsen Talebsafa

3

12

Research Progress of RIOHTRACK in China . . . . . . . . . . . . . . . . . . . . Xu Dong Wang, Lei Zhang, Xing Ye Zhou, Qian Xiao, Wei Guan, and Ling Yan Shan

21

Two Years of APT Program on the New Test Site duraBASt . . . . . . . . Bastian Wacker and Dirk Jansen

32

Guidance for the Next Generation Accelerated Pavement Testing Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Benjamin Worel, Michael Vrtis, and R. Buzz Powell

40

APT of Asphalt Concrete Laboratory Investigation of Cracked Asphalt Pavement Structure Using Accelerated Loading System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Youness Berraha, Daniel Perraton, Guy Doré, Michel Vaillancourt, and Jean-Pascal Bilodeau Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement Using Accelerated Pavement Test . . . . . . . . . . . . . . . . . . . . . Rongji Cao, Chunying Wu, Bingfeng Zheng, Zhenglong Lv, Yi Huang, and Ying Gao

51

61

xi

xii

Contents

Performance of Unbound Pavement Materials in Changing Moisture Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Marit Fladvad and Sigurdur Erlingsson Performance Evaluation of Asphalt Pavement with Semi-rigid Base and Fine-sand Subgrade by Indoor Large-Scale Accelerated Pavement Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . J. T. Wu and Y. T. Wu Accelerated Fatigue Damage Profile of Asphalt Concrete Placed on Semi-rigid Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yi Li, Jiahao Li, Liping Liu, and Lijun Sun

70

80

90

Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic Simulations to Mitigate the Damage Effects on Road Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100 Paul Marsac, Juliette Blanc, Olivier Chupin, Thomas Gabet, Ferhat Hammoum, Navneet Garg, and Mai Lan Nguyen Impact of Mix Design Optimization on HMA Rutting Performance Under Accelerated Pavement Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Fabrizio Meroni, Wenjing Xue, Gerardo W. Flintsch, and Brian K. Diefenderfer Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams from Accelerated Pavement Testing Trial . . . . . . . . . . 118 Sameera Pitawala, Arooran Sounthararajah, James Grenfell, Didier Bodin, and Jayantha Kodikara Effect of the Incremental Loading Conditions in the Permanent Deformation in Heavy Vehicle Simulator Tests . . . . . . . . . . . . . . . . . . . 127 Erdrick Pérez-González, Jean-Pascal Bilodeau, and Guy Doré Comparison of In-situ Response from Companion Asphalt Pavement Sections in Different Climates . . . . . . . . . . . . . . . . . . . . . . . . 137 Michael Vrtis, Benjamin Worel, and David H. Timm Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate via Full-Scale Accelerated Testing . . . . . . . . . . . . . . . . 147 Xingdong Wu, Mustaque Hossain, and Greg Schieber Experimental Methods for Material Selection, Quality Control, and Forensic Investigations of Asphalt Paving Materials . . . . . . . . . . . . 157 Tianhao Yan, Jhenyffer Matias de Oliveira, Mugurel Turos, Mihai Marasteanu, Michael Vrtis, and Dave van Deusen

Contents

xiii

APT of Portland Cement Concrete Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced with Metal Fibers . . . . . . . . . . . . . . . . . . . . . . . 169 Alfonso Pérez, Federico Castro, Miguel Ángel Franco, and Paul Garnica The Challenges of the Accelerated Testing of Jointed Concrete Pavements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178 Angel Mateos, Rongzong Wu, John Harvey, Julio Paniagua, Fabian Paniagua, and Robel Ayalew Accelerated Pavement Testing for the Evaluations of Structural Design and Safety Performance of an Innovative Road Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186 Mai Lan Nguyen, Pierre Hornych, Minh Tan Do, Thierry Sedran, and Duc Tung Dao Experimental Investigation of Wheel-Load Induced Strain Responses in Roller Compacted Concrete Pavements . . . . . . . . . . . . . . . 196 Moinul Mahdi, Yilong Liu, Zhong Wu, and Tyson Rupnow APT for Airfield Pavements Failure Modes of Rapid-Setting Concrete Repairs Under Accelerated Aircraft Traffic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Lulu Edwards, Haley P. Bell, and Jeb S. Tingle Evaluation of Warm Mix Asphalt (WMA) Technologies for Use in Airport Pavements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218 Navneet Garg, Hasan Kazmee, and Lia Ricalde Impact of Heavy Airplanes on Asphalt Surface Behaviour, a Need for a Different Material Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228 Laurent Porot, David Bell, Erik Scholten, and Robert Kluttz The Evolution of Accelerated Pavement Testing for U. S. Military Airfield Pavements at the Waterways Experiment Station . . . . . . . . . . . 238 Timothy W. Rushing APT of Sustainable and Innovative Materials Heavy Vehicle Simulator (HVS) Testing of Innovative Pavement Structures in South Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251 I. Akhalwaya, M. A. Smit, F. C. Rust, and L. Du Plessis

xiv

Contents

From Laboratory Mixes Evaluation to Full Scale Test: Fatigue Behavior of Bio-Materials Recycled Asphalt Mixtures . . . . . . . . 261 Juliette Blanc, Emmanuel Chailleux, Pierre Hornych, Chris Williams, Zahra Sotoodeh-Nia, Laurent Porot, Simon Pouget, François Olard, Jean-Pascal Planche, Davide Lo Presti, and Ana Jimenez del Barco Evaluation of Cold Central-Plant Recycling (CCPR) Technique Using Full-Scale Accelerated Pavement Testing . . . . . . . . . . . . . . . . . . . 270 Gerardo Flintsch, Wenjing Xue, Brian Diefenderfer, and Fabrizio Meroni Structural Performance Evaluation of Block Pavements Using Heavy Vehicle Simulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 280 Shafiqur Rahman, Erik Simonsen, Fredrik Hellman, Abubeker Ahmed, and Sigurdur Erlingsson Testing of Maintenance and Rehabilitation Solutions Downscaled Accelerated Trafficking of Novel Asphalt Joints Based on the Induction Heating Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . 291 Martin Arraigada, M. Bueno, and M. N. Partl Deformation Performance of Foamed Bitumen Stabilised Pavements Under ALF Full-Scale Accelerated Loading . . . . . . . . . . . . . . . . . . . . . . 300 Didier Bodin, James Grenfell, and Geoff Jameson Lessons Learnt from Accelerated Pavement Testing of Full-Depth Recycled Material Stabilized with Portland Cement . . . . . . . . . . . . . . . . 309 Stefan Louw, David Jones, and Rongzong Wu Evaluating the Feasibility of Using an Engineered Cementitious Composite in the Rehabilitation of Pavements by Means of APT and Laboratory Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319 Camilo Andrés Múñoz Rodriguez, Washington Peres Núñez, Jorge Augusto Pereira Ceratti, Lélio Antônio Brito, Ângela Gaio Graeff, and Luiz Carlos Pinto da Silva Filho Design of Reinforced Pavements with Glass Fiber Grids: From Laboratory Evaluation of the Fatigue Life to Accelerated Full-Scale Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 329 Mai Lan Nguyen, Cyrille Chazallon, Mehdi Sahli, Georg Koval, Pierre Hornych, Daniel Doligez, Armelle Chabot, Yves Le Gal, Laurent Brissaud, and Eric Godard Full-Scale Evaluation of Balanced Cold in-Place Recycling (CIR) Asphalt Mixtures Using a Heavy Vehicle Simulator . . . . . . . . . . . . . . . . 339 Ahmed Saidi, Ayman Ali, Yusuf Mehta, Ben C. Cox, and Wade Lein

Contents

xv

Micro Surfacing Performance Under Accelerated Pavement Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 348 Adriana Vargas-Nordcbeck, Jason Nelson, and Buzz Powell Large-Scale Test to Determine the Crack Retarding Ability of Asphalt Reinforcements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 357 Jens Wetekam, Konrad Mollenhauer, Michael Wistuba, Stephan Büchler, Michael Schmalz, and Thomas Ziegler Data Analysis and Modelling Development and Calibration of Performance Models Based on APT Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369 Tania Ávila-Esquivel, José P. Aguiar-Moya, Edgar Camacho-Garita, and Luis Guillermo Loria-Salazar Back-Calculation of the Moduli of Asphalt Pavement Layer Using Accelerated Pavement Testing Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 379 Huailei Cheng, Yuhong Wang, Liping Liu, Lijun Sun, Yue Hu, and Yi Li Simulation of Damage Scenarios in a Bituminous Pavement Tested Under FABAC ALT Using M4-5n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 389 Olivier Chupin, Jean-Michel Piau, Armelle Chabot, Hanan Nasser, Mai Lan Nguyen, and Yann Lefeuvre Evaluation of Permanent Deformation Models for Flexible Pavements Using Accelerated Pavement Testing . . . . . . . . . . . . . . . . . . 399 Yared Dinegdae and Sigurdur Erlingsson Application of a General Methodology to Analyze Strain Data Obtained from Accelerated Pavement Testing . . . . . . . . . . . . . . . . . . . . 409 Andrae Francois, Ayman Ali, and Yusuf Mehta Bayesian Calibration of ILLI-THERM for Temperature Prediction Within Airfield Concrete Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . 418 Osman Erman Gungor, Imad L. Al-Qadi, and Navneet Garg Development of a Distress Prediction Model Based on Deflection Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 428 Piero Laurent-Matamoros, Tania Avila-Esquivel, Edgar Camacho-Garita, Jose Aguiar-Moya, and Luis Guillermo Loria-Salazar Modelling Asphalt Pavement Responses Based on Field and Laboratory Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 438 Andreas Loizos, Konstantinos Gkyrtis, and Christina Plati

xvi

Contents

Data Gathering and Evaluation of Tensile Strains Measured in APT with Mathematical Computation Method . . . . . . . . . . . . . . . . . . . . . . . . 448 Lubos Remek and Veronika Valaskova Use of Heavy Vehicle Simulator-Aircraft in the Development of the CBR-Beta Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 458 Jeremiah M. Stache and Carlos R. Gonzalez Full Scale Testing with the Mobile Load Simulator: Advanced Measurements Related to Pavement Behavior and Surface Layer Damage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 467 Yamina Oubahdou, Philippe Reynaud, Christophe Petit, Anne Millien, Jérome Dopeux, Mickael Metrope, Benoît Picoux, Charlotte Gerbaud, and Rémi Tautou Instrumentation and Data Processing Water Pressures in Porous Asphalt . . . . . . . . . . . . . . . . . . . . . . . . . . . . 479 David Alabaster, Frank Greenslade, and Sabine Leischner Procedure for Temperature Correction of Strains Measured in a Road Pavement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 488 Maria Barriera, Juliette Blanc, Emmanuel Chailleux, Simon Pouget, Julien Van Rompu, and Bérengère Lebental Relationship Between Pavement Damage and the Combined Action of Frost Penetration and Axle Load . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 Jean-Pascal Bilodeau and Erdrick Perez Gonzalez APT Instrumentation – Where Do I Begin? . . . . . . . . . . . . . . . . . . . . . . 507 Harold T. Carr and W. Jeremy Robinson Towards an Adapted Ovalization System for Flexible Airfield Pavement Interface Characterization Using Rolling-Wheel or HWD Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 516 Maissa Gharbi, Michael Broutin, Thomas Schneider, Maindroult Stephane, and Armelle Chabot Evaluation of the Use of Geophones and Accelerometers for Monitoring Pavement Deflections, Using Accelerated Pavement Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526 Natasha Bahrani, Juliette Blanc, Pierre Hornych, and Fabien Menant Dynamic Response Monitoring and Analysis of In-Service Asphalt Pavement Based on FBG Measuring Technology . . . . . . . . . . . . . . . . . . 536 Xianyong Ma, Tongxu Wang, and Zejiao Dong

Contents

xvii

Monitoring Road Pavement Performance Through a Novel Data Processing Approach, Accelerated Pavement Test Results . . . . . . . 545 Mario Manosalvas-Paredes, Nizar Lajnef, Karim Chatti, Juliette Blanc, Nick Thom, Gordon Airey, and Davide Lo Presti Pavement Instrumentation and WIM Data from a Test Track on the BR-101 Highway in South of Brazil . . . . . . . . . . . . . . . . . . . . . . 555 Gustavo Garcia Otto, Amir Mattar Valente, Bruno de Melo Gevaerd, Rafael Aleixo de Souza, Adosindro Joaquim de Almeida, and Keyla Junko Chaves Shinohara Instrumentation Response of Full-Scale Multi-axial Geogrid Stabilized Flexible Pavements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564 W. Jeremy Robinson, Jeb S. Tingle, Mark H. Wayne, Jayhyun Kwon, and Gregory Norwood Temperature Normalization of Flexible Pavement Response Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 574 David Timm Capacitance Type Moisture Sensors in Unbound Granular Material of Pavement Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583 Michael Vrtis, Leonard Palek, and Benjamin Worel Inverse Analysis of Pavement Layer Moduli Based on Data Collected by Buried Accelerometers and Geophones . . . . . . . . . . . . . . . 592 Natasha Bahrani, Eyal Levenberg, Juliette Blanc, and Pierre Hornych Temperature Correction Method of Deflection Basin and Stress/Strain Response of Asphalt Pavement . . . . . . . . . . . . . . . . . . 602 Qian Xiao, Xu Dong Wang, Xing Ye Zhou, Lei Zhang, and Wei Guan Monitoring and non Destructive Testing Thin-Bed Data Model for the Processing of GPR Data over Debonded Pavement Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 615 V. Baltazart, S. Todkar, X. Dérobert, and J.-M. Simonin Remote Monitoring System for Road Constructions . . . . . . . . . . . . . . . 623 Dmitry Chirva, Vitaly Solodov, Sergey Mironchuk, and Eugeny Isaev Radar Database Collected over Artificial Debonding Pavement Structures During APT at the IFSTTAR’s Fatigue Carrousel . . . . . . . . 632 X. Dérobert, V. Baltazart, J.-M. Simonin, C. Norgeot, S. Doué, O. Durand, and S. S. Todkar Improvements to Backcalculation Procedure by Means of Structural Analysis Based on Deflection Parameters . . . . . . . . . . . . . . . . . . . . . . . . 640 Piero Laurent-Matamoros, Edgar Camacho-Garita, Tania Avila-Esquivel, Jose Aguiar-Moya, and Luis Guillermo Loria-Salazar

xviii

Contents

Rapid and Continuous Imaging for Crack Monitoring During APT Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 649 Mai Lan Nguyen, Juliette Blanc, Stephane Trichet, Thierry Gouy, Gilles Coirier, Yvan Baudru, Xuan Quy Le, Minh Duc Nguyen, Rodrigo Shigueiro Siroma, Pierre Hornych, and Fabrice Blaineau Spectral Element Simulation of Heavy Weight Deflectometer Test Including Layer Interface Conditions and Linear Viscoelastic Behaviour of Bituminous Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . 658 Jean-Marie Roussel, Hervé Di Benedetto, Cédric Sauzéat, and Michaël Broutin Backcalculation of Airfield Pavement Layer Moduli Under HWD Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 666 Hao Wang, Pengyu Xie, and Richard Ji SMS Structure Measurement System to Optimize Accelerated Pavement Testing APT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 676 Bastian Wacker and Dirk Jansen New Pavement Concepts and APT Challenges Evaluation of a Solution for Electric Supply of Vehicles by the Road, at Laboratory and Full Scale . . . . . . . . . . . . . . . . . . . . . . 689 Pierre Hornych, Thomas Gabet, Mai Lan Nguyen, Fabienne Anfosso Lédée, and Patrick Duprat Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 699 Bertrand Pouteau, Kamal Berrada, Sandrine Vergne, Mai Lan Nguyen, Stéphane Trichet, Thierry Gouy, and Pierre Hornych Future APT – Thoughts on Future Evolution of APT . . . . . . . . . . . . . . 708 Wynand J. vdM. Steyn Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 719

Establishing APT Facilities

Promoting Implementation of Significant Findings from the NCAT Pavement Test Track R. Buzz Powell(&) National Center for Asphalt Technology, 277 Technology Parkway, Auburn, AL 36830, USA [email protected]

Abstract. The National Center for Asphalt Technology’s (NCAT) Pavement Test Track is a unique accelerated pavement testing facility that has been operational for over 20 years. It is a cooperatively funded project with individual test sections sponsored by state and federal highway agencies as well as commercial interest groups. The focus of research within each section can include mix/materials, structural pavement design, and/or pavement preservation. A fleet of triple trailer trains with legally loaded axles is used to compress a design lifetime of pavement damage within each 3-year project cycle, with fleet operations suspended weekly to facilitate detailed pavement condition assessments. Off-Track pavement preservation sections were added to the program in 2012. A pavement research partnership with the Minnesota Road Research Project (MnROAD) was added in 2015 that included both pavement preservation and asphalt mix performance testing. This paper highlights a new initiative by the project team in the 2018 research cycle to promote implementation of significant findings that have the greatest relevance to sponsoring state agencies. The effort consisted of reviewing experimental outcomes from all seven research cycles, determining an implementation readiness level for each finding, seeking feedback from industry stakeholders, and working one-on-one with state agency representatives to identify the best deployment candidates to pursue successful demonstration projects. Keywords: Accelerated pavement testing (APT) Implementation
Deployment


Asphalt pavement


1 Background The National Center for Asphalt Technology’s (NCAT) Pavement Test Track is a unique accelerated pavement testing facility that has been operational for over 20 years. It is a cooperatively funded project with individual test sections sponsored by state and federal highway agencies as well as commercial interest groups. The focus of research within each section can include mix/materials, structural pavement design, and/or pavement preservation. A fleet of triple trailer trains with legally loaded axles is used to compress a design lifetime of pavement damage with-in each 3-year project cycle, with fleet operations suspended weekly to facilitate detailed pavement condition assessments. The 2018 NCAT Pavement Test Track (referred to hereafter as the Track) represents the seventh 3-year research cycle. © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 3–11, 2020. https://doi.org/10.1007/978-3-030-55236-7_1

4

R. Buzz Powell

Off-Track pavement preservation sections were added to the program in 2012. A pavement research partnership with the Minnesota Road Research Project (MnROAD) was added in 2016 that included both pavement preservation and asphalt mix performance testing. As a result of the MnROAD partnership, life extending and condition improving benefits of pavement preservation are being quantified on low and high traffic roadways in both hot and cold climates. Additionally, laboratory results from asphalt mix performance testing are being compared to actual cracking performance at both facilities (southern fatigue and northern thermal). This paper highlights a new initiative by the project team in the 2018 research cycle to promote implementation of significant findings that have the greatest potential value for state agencies. This effort was necessary because products from Track experiments are sometimes only implemented by the sponsoring agency and are not widely embraced by all the other agencies in the research cooperative.

2 Methodology The project began with a review of all seven past 3-year research cycles at the Track. The research team compiled a list of fourteen implementable findings that were considered to offer a high potential value to adopting agencies. Next, the list was sorted into high, medium, and low readiness levels based on the amount of available documentation. In addition to supporting the current effort, this methodology also helped the research team identify topics that need more documentation that will enable them to move up to higher levels of readiness in the future. The next step in the process was to determine how industry viewed each topic. A presentation was made to the summer meeting of the state asphalt pavement association (SAPA) executives in which each topic was explained in significant detail. Following the presentation, each SAPA executive director was asked to participate in an online poll to indicate whether they would either enthusiastically support, be neutral on, or oppose each of the fourteen topics within their state jurisdiction. This information was compiled and summarized so that it could be shared with agencies at a later time in onsite workshops. In this manner, agencies could make fully informed implementation decisions on each topic.

3 Significant Findings 3.1

Highest Readiness Level

3.1.1 Recalibrated Layer Coefficient for Modern Materials/Methods The HMA layer coefficient developed from the AASHO Road Test in Ottawa, Illinois in the late 1950s has not been updated in approximately 60 years. Data from numerous test sections over several Track research cycles were used to recalibrate the asphalt layer coefficient for modern methods and materials. Using the 1993 AASHTO Design Guide flexible pavement design equation, predicted ESALs were compared to actual traffic on the sections, and a new asphalt layer coefficient was conservatively determined. Increasing the coefficient from 0.44 to 0.54 results in approximately 18%

Promoting Implementation of Significant Findings

5

thinner, more cost effective HMA cross-sections using the 1993 AASHTO Design Guide (Peter-Davis and Timm 2009). Other sections on the Track were used to determine a layer coefficient of 0.36 for 100% RAP cold recycle mix (which is more than twice the value typically used for crushed aggregate base) and 0.15 for spray reducing open graded friction course (OGFC) surfaces that are used to prevent wet weather accidents (Diaz-Sanchez et al. 2017; Timm and Vargas-Nordcbeck 2012). 3.1.2 Maximum Thickness for Perpetual Pavement Long-life or perpetual pavements are designed and built to last longer than 50 years, requiring only a periodic mill and inlay of the surface layer. The pavement structure is designed using appropriate materials and layer thicknesses to prevent structural distresses that begin at the bottom of the pavement structure, such as bottom-up fatigue cracking and subgrade rutting. Instrumented sections over numerous Track research cycles have been used to develop conservative tables that relate maximum asphalt thicknesses to varying subgrades and bases. Perpetual asphalt thicknesses using conventional materials range from a minimum of 6½ in. (on thicker, stiffer base over a good subgrade) to a maximum of 15½ in. (on thinner, softer base over a poor subgrade) (Tran et al. 2015). 3.1.3 Strategic Use of Polymer Modification for Lower Life Cycle Cost Targeted polymer modification is an important tool in satisfying long-life pavement layer design objectives (e.g., strain tolerant lower layers and rut resistant surfaces). Conventional polymer modification increases fatigue resistance in lower layers and improves both rutting and cracking resistance in upper layers, which can reduce life cycle cost when used in thinner pavement structures. More value is added when the stiffness of the intermediate layer can be increased through the targeted use of reclaimed and recycled materials (West et al. 2016). Likewise, high polymer modification can be used to build even thinner pavements that take advantage of greater improvement in both cracking and rutting performance (Timm et al. 2013). High polymer modification can also add value in maintenance and rehabilitation. An intentionally under designed pavement on a soft subgrade that was exhibiting complete structural failure on the Track was converted into a perpetual pavement by a highly polymer modified inlay (Powell 2012). 3.1.4 Arizona Rubber Mix for Durable and Crack Resistant Pavements The State of Arizona (AZ) Department of Transportation has used crumb rubber modified surface mixes for decades to improve the life cycle performance of high traffic pavements, delay reflective cracking in jointed concrete pavement overlays, and reduce urban road noise (Scofield 1989). One fifth of the asphalt binder in these mixes is coarse recycled ground tire rubber, which increases the elasticity of the mix. The aggregate structure is gap graded to provide room for the coarsely ground (minus 16 to 20 mesh) rubber particles. Arizona rubber mixes have exhibited good performance and high life cycle value at the Track over three 3-year research cycles using local materials in bottom, intermediate, and surface layer applications. No special precautions were taken at the plant to prevent maintenance problems during the Track’s relatively short production runs. In order to simplify the production process on high tonnage projects, it

6

R. Buzz Powell

is now possible to purchase pre-reacted rubber. This material is added to the plant via RAP feed or mineral filler system at a rate that produces the necessary rubber and total asphalt binder contents. 3.2

Medium Readiness Level

3.2.1 Balanced Mix Design (BMD) Testing for Construction and Design Recipe based volumetric mix design practices originally developed for virgin mixes and neat asphalt binders can hinder innovation with new materials (e.g., highly polymer modified binder, aramid fibers, etc.) and limit sustainability (i.e., higher percentages of reclaimed and recycled materials) in modern asphalt pavements. Practical balanced mix design (BMD) tests are now available to verify the performance potential of new mix designs (using critical aging) and ensure that produced mix will not fail prematurely (using only plant aging). The turnaround time for testing during construction is only a few hours, which enables a contractor to make plant changes before a significant tonnage of deficient mix has been produced. Research is ongoing at the Track to establish threshold criteria for both rutting and fatigue cracking performance, while research at MnROAD will establish threshold criteria for rutting and thermal cracking performance. 3.2.2 Implementation of Pavement Preservation Benefit Curves The second Strategic Highway Research Program (SHRP 2) only provided US agencies with general preservation treatment selection guidance (Smith et al. 2011); however, life extending and condition improving benefit curves generated by the partnership between NCAT and MnROAD can now be used to quantify the life expectancy of standalone and combination treatments using MAP-21 based pretreatment pavement condition data. The condition improving benefit at any point in time is available for all treatment options by comparing treated to untreated performance curves that are separated by MAP-21 based good, fair, and poor pretreatment condition (Vargas-Nordcbeck and Powell 2018). Additionally, the entirety of the treated and untreated banded performance curves can be hard coded into a pavement preservation decision tree to automate the agency selection process using local bid prices. 3.2.3 Use of Smaller/Finer “Screenings” Mixes for Pavement Preservation Thin asphalt overlays are a common preservation treatment option. Currently, about half of US states utilize 4.75 mm Nominal Maximum Aggregate Size (NMAS) mixtures in thin overlay applications. An advantage of these mixtures is that they can be placed as thin as 12.5 mm, allowing more miles to be surfaced. A 4.75 NMAS thinlay made from limestone screenings and native sand was placed on the Track at the beginning of the second research cycle in the summer of 2003. The section has been in place for seventeen years and has supported approximately 60 million equivalent single axle loadings (ESALs) with virtually no rutting and only minor cracking. This section is proof that well-designed 4.75 mm NMAS mixes are a tough and durable option for pavement preservation (Powell and Buchanan 2012).

Promoting Implementation of Significant Findings

7

3.2.4 Better Design and Construction of OGFC Drainable Surface Mixes Drainable surface mixes are utilized by many agencies to improve road spray and reduce wet weather accidents on high speed roadways. These mixes are typically designed with 20 to 25% air voids; therefore, they must exhibit good durability while repeatedly exposed to water. Agencies would use more OGFC surfaces if they can be constructed to last as long as relatively impermeable surfaces. Smaller NMAS designs, liquid antistrip, ground tire rubber, and aramid fibers have all been strategically used at the Track to enhance OGFC durability (West et al. 2016). The performance benefits of different tack materials (conventional and reduced track emulsions versus hot applied), equipment (conventional versus spray paver), and rates (conventional versus thick) have been documented (Jones 2013). 3.2.5 Expanded Use of Locally Available Virgin and Recycled Aggregates Many agencies must set safe limits on the use of local aggregates in order to protect the motoring public from low friction roadways. Similar limits necessarily apply to RAP when it is conservatively assumed low friction aggregates may be the dominant raw material. Technology developed at NCAT can be used to compare mix with higher levels of local aggregates and/or RAP to well established control mixes. The process involves making mix slabs in the laboratory, polishing the surface under water with a mechanical polishing device, and making periodic measurements with a dynamic friction tester (DFT) to characterize friction loss as a function of wheel passes (Heitzman et al. 2015). Removing recipe limits on mix designs for both virgin and recycle materials incentivizes innovation and improves sustainability without compromising the safety of the motoring public. 3.3

Lower Readiness Level

3.3.1 100% Utilization of All RAP Generated In many urban markets, RAP piles have become mountains that continue to get larger by the workday. This is because 100% asphalt surfaces are removed and replaced with mix that is limited by specification to a relatively low RAP percentage (typically 20%). Findings from numerous experiments on the Track can be used to stop the growth of giant urban market RAP piles in a manner that creates value for taxpayers and extends pavement life. For example, 100% RAP cold recycle mix can be used in place of virgin aggregates to cut base thickness in half and expedite construction (Diaz-Sanchez et al. 2017). In some cases, cold recycle mix can even replace intermediate layers in the pavement structure. Surface and base mixes can be designed using BMD principles at the highest possible RAP content that does not negatively impact performance (e.g., rich bases to prevent bottom-up cracking and surfaces that are rut resistant but not susceptible to top-down cracking) (West et al. 2016). Depending on foundation quality, it may be possible to design mixes for intermediate layers with very high RAP contents to minimize overall deflections in the pavement structure, which also reduces the strains that cause bottom-up fatigue cracking.

8

R. Buzz Powell

3.3.2 Asphalt Based Enhanced Friction Surfaces Although the standard of excellence for high friction surfaces is epoxy bauxite, several other surfaces have been proven to provide long term enhanced friction on the Track (Heitzman et al. 2015). Some examples of asphalt based enhanced friction surfaces placed on the Track include chip seals, micro surface, and thinlays. Track chip seals were placed using kiln fired lightweight clay, while calcined bauxite was blended with conventional virgin aggregates in a micro surface as well as an asphalt pavement thinlay. In all cases, surface friction measured with a full-size ribbed tire in wet conditions compared well with epoxy bauxite without the liability of the low friction epoxy surface that results from chip loss. 3.3.3 Quality Tack Practices, Products, and Equipment The life of asphalt pavements is dramatically shortened when layers debond (Jones 2013). Interlayer stresses increase significantly when adjacent layers exhibit big differences in stiffness (such as when a lower RAP layer is placed adjacent to a higher RAP layer), and/or when significant deflections are expected due to a soft foundation. Reduced interlocking of adjacent layers is also likely with stiffer mixes. In consideration of all these factors, it has never been more important to ensure that proven tack products are applied at a suitable residual rate, that the quality and consistency of the application is good, and that excessive tracking does not occur. Best practices, products, and equipment that improve the chances of successfully bonding adjacent asphalt pavement layers have been proven on the Track in seven construction cycles and over 20 years of practical experience. 3.3.4 Producing 100% RAP Recycle Mix from a Cold Hot-Mix Plant One hundred percent RAP mix produced with a dedicated portable cold central plant has exhibited excellent performance on and off the Track through numerous research cycles. As a proof of concept, an Astec Double Barrel Green hot-mix plant with the burner off was used to produce foamed 100% RAP cold recycle mix during the 2015 Track build. No significant plant modifications were required to run this mix at approximately 125 tons per hour, which is about 40% of the rated hot-mix capacity. This mix was used as the base layer in a section that can be directly compared to another (control) section that was built with a hot-mix asphalt base layer at a thickness to provide theoretically equivalent performance. 3.3.5 Using Small NMAS “Screenings” Mix for Crack Relief Under OGFC Although 4.75 NMAS surface mixes are some of the most crack resistant mixes tested on the Track, low macrotexture can exclude their use on high speed roadways. In order to overcome this limitation, versatile 4.75 NMAS mixes can be combined with OGFC surfaces in a very cost-effective way. Many agencies have adopted a standard practice of micro milling and inlaying only the deteriorated OGFC surface layer to minimize the cost of rehabilitation. This works well when distresses do not extend beneath the old surface; however, it sets the stage for poor OGFC performance when surface cracks extend down into the underlying layer. When this is the case, milling an additional

Promoting Implementation of Significant Findings

9

¾ in. and inlaying with a 4.75 NMAS “surface” mix can provide a solid foundation for the new OGFC surface while providing a crack inhibiting interlayer that prevents reflective cracking.

4 Industry Feedback After this list of fourteen high value implementation topics were banded into high, medium, and low readiness levels, an overview presentation was made at a national meeting of SAPA executive directors. Following the presentation, the SAPAs were asked to complete an online poll to indicate how industry would respond to each implementation topic in their unique jurisdiction (i.e., supportive, neutral, or opposed). Twenty-five SAPA representatives initially completed the online poll, representing half the states in the US. As a result of this process, it was clear that industry was highly supportive of the following four findings (with the highest positive rankings shown first in the list, and NCAT’s readiness level assessment for each topic shown in parentheses): 1) Strategic Use of Polymer (high); 2) Innovation via Balanced Mix Design (medium); 3) Maximum Pavement Thickness Table (high); 4) High Performance Tack (low).

5 Implementation Workshops Armed with a banded findings list and knowledge of industry’s view of each, researchers set out to develop an onsite workshop that could be presented to Track sponsor agencies to promote implementation. A sales funnel approach was applied because it is analogous to bringing about institutional change in a transportation agency where the mission of public protection and stewardship can make practitioners change averse. Within this implementation effort, the top of the funnel represents the list of fourteen high value implementation findings. The bottom of the funnel represents the finding or findings from the list that are chosen by the agency for use on actual highway projects. NCAT expected to emerge from each workshop with an action item list to support the implementation effort(s) that may consist of reference documents, best practices, training, project support, etc. for selected topics. The goal was for each workshop to be interactive and informative for both agency and NCAT personnel. A continuous improvement process would be used to ensure that each workshop was improved upon based on the experience gained from all previous workshops. 5.1

AM Commitment by Decision Makers with Technical Experts

Onsite implementation workshops were designed to last a full day. In the morning, the audience would ideally consist of both subject matter experts and decision makers. Subject matter experts would engage in a technical exchange with NCAT personnel to inform the decision makers of potential agency value as each of the fourteen high value research topics was reviewed. The goal of the morning session would be to get a

10

R. Buzz Powell

commitment from decision makers on which topic or topics should be selected for implementation. With this commitment in place, decision makers could be dismissed for the rest of the day to take care of other pressing agency business, as necessary. An overview of the morning session is provided below: 0800 0930 1000 1030 Noon

5.2

Overview of Fourteen NCAT Selected High Value Findings Brainstorming Other High Value Findings Not Selected by NCAT Agency Selection of Findings for Further Discussion Detailed Review and Discussion of Selected Findings Final Selection of High Value Findings for Implementation.

PM Roadmap and Timeline with Technical Experts

In the afternoon, subject matter experts would work with NCAT personnel to identify what is needed for implementation to be successful. In order to support this effort, NCAT personnel would invest time during lunch to pull together specific technical information to support the topics selected. The products of the meeting are an implementation roadmap for selected topics, a projected timeline for trial projects, and an action list of support needed from NCAT to facilitate the process. An overview of the afternoon session is provided below: 1330 1400 1500 1630

Additional Discussion of Final High Value Finding Selections Identification of Needs to Support Implementation Development of Implementation Roadmaps/Timelines Adjourn.

6 Summary and Takeaways Workshops have been presented in two pooled fund sponsor states as of the end of 2019. The first state chosen just joined the project for the 3-year research cycle that began in 2018. Because they did not have an extensive history with the Track, the morning time allocated for brainstorming other findings was essentially a review session for past research cycles. At the end of the first workshop, the following topics were chosen for implementation: 1) Arizona rubber surface mix to prevent reflective cracking; 2) Strategic use of polymer modification with focus on high polymer; 3) Recalibrated (increased) layer coefficient; 4) Maximum pavement thickness table; 5) Active support for ongoing BMD implementation. The second state has been a Track sponsor since the inaugural research cycle that began in 2000, which meant they already had extensive experience with research findings generated for other sponsor states. At the end of the day in the second workshop, the following three topics were chosen for implementation: 1) Quality tack practices/products/equipment; 2) Active support for BMD implementation with focus on aging; 3) Curves quantifying the benefit of pavement preservation.

Promoting Implementation of Significant Findings

11

There are many positive takeaways from the initial experience with onsite implementation workshops. It was encouraging to find that the agenda for the first two workshops was generally followed and the meetings flowed as expected. It was also encouraging that attendees for the first two workshops seemed to be an ideal mix of decision makers and subject matter experts. All attendees stayed for the entire day, which was not expected. In the afternoon, decision makers were fully engaged in the technical discussion with subject matter experts. There was diversity in selected topics. and it is expected that future selections will also vary significantly with each agency. The next step is to schedule workshops in all the other states that support the Track and/or the MnROAD pavement research partnership. Acknowledgements. The activity presented in the paper is a product of the national transportation pooled fund (TPF) hosted by the Alabama Department of Transportation and funded in a prorated manner by twenty-four state departments of transportation (DOTs), the Federal Highway Administration (FHWA), and four private sector partners.

References Diaz-Sanchez, M.A., Timm, D., Diefenderfer, B.: Structural coefficients of cold central-plant mixtures. J. Transp. Eng. 143(6) (2017) Heitzman, M., Turner, P., Greer, M.: High Friction Surface Treatment Alternative Aggregates Study. NCAT 15-04. Auburn, Alabama (2015) Jones, C.: Debonding of asphalt pavement layers. Asphalt Technol. News 25(2) (2013) Peter-Davis, K., Timm, D.: Recalibration of the Asphalt Layer Coefficient. NCAT 09-03. Auburn, Alabama (2009) Powell, B.: Accelerated performance of a failed pavement on a soft clay subgrade after rehabilitation with high polymer mix at the NCAT track. In: Advances in Pavement Design through Full-scale APT, Davis, California (2012) Scofield, L.: The History, Development, and Performance of Asphalt Rubber at ADOT. Technical report AZ-SP-8902. Arizona Transportation Research Center, Phoenix, Arizona (1989) Smith, K.L., Peshkin, D., Wolters, A., Krstulovich, J., Moulthrop, J., Alvarado, C.: Guidelines for the Preservation of High-Traffic-Volume Roadways, SHRP 2 Report S2-R26-RR-2, Transportation Research Board. Washington, DC (2011) Timm, D., Robbins, M., Willis, R., Tran, N., Taylor, A.: Field and Laboratory Study of High Polymer Mixtures at the NCAT Test Track: Final Report. NCAT 13-03. Auburn, Alabama (2013) Timm, D., Vargas-Nordcbeck, A.: Structural coefficient of open-graded friction course. Transp. Res. Rec. 2305(1), 102–110 (2012) Tran, N., Robbins, M., Timm, D., Willis, J.R., Rodezno, C.: Refined Limiting Strain Criteria and Approximate Ranges of Maximum Thicknesses for Designing Long-Life Asphalt Pavements. NCAT Report 15-05R. Auburn, Alabama (2015) Vargas-Nordcbeck, A., Powell, B.: Developing Life-Extending Benefit Curves for Asphalt Pavements with Preservation Treatments. AAPT. Jacksonville, Florida (2018) West, R., Timm, D., Powell, B., Heitzman, M., Tran, N., Rodezno, C., Watson, D., Leiva, F., Vargas, A., Willis, R., Vrtis, M., Sanchez, M.: Phase V (2012–2014) NCAT Test Track Findings. NCAT Report 16-04. Auburn, Alabama (2016)

The Design, Construction and Operation of the APT Facility at the University of Texas at Arlington Stefan A. Romanoschi(&), Constantin Popescu, Ana-Maria Coca, and Mohsen Talebsafa Department of Civil Engineering, University of Texas at Arlington, Arlington, USA [email protected]

Abstract. The University of Texas at Arlington has decided to own and operate an Accelerated Pavement Testing (APT) facility in 2012 to expand its research capabilities. The facility is located in the Dallas-Fort Worth urban area on university owned land. The loading machine is a mobile linear device capable of loading two pavements simultaneously at a rate of about 100,000 passes per week. The maximum axle load that can be applied is 162 kN (36 kip). Unidirectional and bi-directional loading can be accommodated. Controlled heating and cooling of the air above the tested pavement section is also possible. The facility and the loading machine were designed and built entirely with internal funds over an eighteen-month period. Two research projects have been conducted in close collaboration with the Texas A&M Transportation Institute (TTI) to study the optimization of asphalt mixes containing various percentages of Recycled Asphalt Pavement (RAP) and Recycled Asphalt Shingles (RAS). The paper presents the development of the APT facility and the main features of the loading machine and discusses the experience with the construction and the four-year long operation. The reasoning behind the design choices, the financial decision making, the challenges as well as the successes are highlighted. Keywords: Facility design Instrumentation


Pavement testing machine
Operation


1 Introduction The University of Texas at Arlington started to develop an Accelerated Pavement Testing facility (APT) in 2012, reflecting the desire to expand the pavement research capabilities within the University. The facility is unique in the United States not only because it is entirely owned and operated by a university but also because it was developed with internal funds. Most APT facilities in the United States are either owned and operated by state or federal transportation agencies or have been built with funding from these agencies due to their high initial costs. Having an APT facility at the University brings national and international prominence and visibility to the University. It enhances the research capabilities in the pavement engineering field, it © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 12–20, 2020. https://doi.org/10.1007/978-3-030-55236-7_2

The Design, Construction and Operation of the APT Facility

13

fosters increased collaboration with other universities and it increases the graduate student enrollment. It is a cost effective investment and effort which allows the University to serve better its constituents.

2 Test Site The Accelerated Pavement Testing Facility (APTF) was constructed on University owned land near The University of Texas at Arlington’s Research Institute (UTARI), near State Highway SH-820, about one mile north of Interstate Highway I-30, on the east side of Fort Worth, TX. The site is located less than a mile away from an asphalt plant owned by a local paving contractor, which has produced and placed the hot mix asphalt for the first two experiments. The site is larger than the area occupied by the APTF, allowing further expansions of the facility if needed. The facility was designed such that a large number of test sections can be built; it is desirable to have sufficient space to allow construction of new test sections as the existing ones are subjected to accelerated testing. No building was planned for the facility to reduce the overall cost. Outdoor test sections were considered as better than indoor sections since they can be built with conventional highway construction equipment. Figure 1 shows a schematic of the APRF. The main components of the APRF are: a) The Experimental Test Area is a 45 m by 45 m (150 ft by 150 ft) elevated area with 0.9 m (3 ft) of imported subgrade soil. A total of 30 experimental pavement sections, each 23 m (75 ft) long and 3.0 m (10 ft) wide can be constructed on top of the imported subgrade soil. Twelve pavement sections were built for the first experiment while eight sections were built for the second experiment. The Test Pad was built by removing 0.6 m (2 ft) of the existing soil and replacing it with imported soil. This was done to ensure a good uniformity of the subgrade soil support underneath the experimental pavement sections. b) Parking and Access Areas around the experimental area provide sufficient space not only for parking but also for maneuvering the large construction equipment used in removing the existing sections and building new ones. c) Entry Gate (sliding) on the east side and a wider, Equipment Access Gate on the north side of the site. d) Personnel Trailer. A large office trailer was purchased, brought to the site and modified to satisfy the requirements of the work to be conducted at the APRF. The trailer contains an office room, a small conference room and a storage room for tools, materials and equipment. The trailer is equipped with a bathroom and a kitchenette. e) An electrical transformer with several disconnect switches placed near the trailer provides electricity to the trailer and to the Pavement Testing Machine (PTM). The entire site is bordered by a chain-link fence and illuminated by two light poles for security purposes during nighttime. The site is provided with all needed utilities: electricity; water, sewer lines and internet connection.

14

S. A. Romanoschi et al.

Fig. 1. The layout of the APRF Site

3 Pavement Testing Equipment The testing equipment, named the Pavement Testing Machine (PTM), was designed and built in-house. The PTM is a linear APT testing device that can be transported easily to remote locations, if needed. Figure 2 shows the PTM while it is positioned on top of test sections. The design philosophy was to build a portable machine capable of loading two pavement sections at a time, with as many off-the shelf components as possible and at a low cost. The conceptual design is similar to that at the APT facility of Kansas State University (AFD40 2019). The use of a conventional truck axle ensures that concrete slabs are loaded more realistically relative to loading them with only one wheel.

Fig. 2. Photo of the Pavement Testing Machine (PTM)

The Design, Construction and Operation of the APT Facility

15

As shown in Fig. 3, the main components of the PTM are: – A steel frame made of two 15 m (45 ft) long beams bolted to four pillars. The beams are stabilized with transverse and diagonal bars at the top. The frame is heavy and it acts as a reaction frame. One railroad rail is fixed underneath the bottom flange of each beam. The steel frame was fabricated by a steel building construction company. – A moving bogie that travels at speeds of up to 8 km/h (5 mph). The bogie has four rail wheels mounted at the top that push up into the rails underneath the beams when loads are applied to the test pavement. Four additional steel wheels travel on top of the inner part of the bottom flanges of the beams to support the bogie when the loading axle is lifted and it doesn’t touch the pavement. A reinforced single axle of a truck with four bus tires is mounted at the bottom of the bogie. – A hydraulic pump mounted on top of the bogie provides power to two hydraulic pistons that push down or lift the axle; the maximum single axle load that can be applied is 162 kN (36,000 lbf), which is twice the legal load limit for single axles in most U.S. states. The pistons can lift the axle when testing is done in uni-directional loading mode or when the PTM is transported. The pistons push down continually when testing is done in bi-directional mode. The force applied by each piston is measured using a load cell. The load applied by each wheel is periodically calibrated using static scales. – A motor mounted between the two front columns pulls the bogie back and forth using a cable. A belt, sprockets and drum system transmits the power from the motor to the cable. The maximum speed the bogie travels is 8 km/h (5 mph). In bidirectional loading mode, the PTM applies about 600 passes in one hour or about 100,000 passes in a week. – Two transverse support footings are mounted underneath the two front and the two rear columns. Each footing is fixed to the ground and attached with a screw jack system to the columns. The two screw jacks move in synchronized fashion the entire PTM sideways to provide the lateral wheel wander. The maximum lateral position the PTM can move on either side of the central position is 0.6 m (24 in). – A front and a rear platforms. The front platform is fixed. A king pin mounted underneath it is used to connect to a truck when the machine is towed. The rear platform mounted on top of a tandem truck axle is not fixed; it can slide vertically relative to the rear posts of the machine using two hydraulic pistons connected to a small hydraulic pump. To change the PTM in transport mode, the cylinders push up the two steel columns when the tandem axle sits on the ground. In pavement-testing mode, the rear platform is lifted off the ground by the two pistons. – The central portion of the PTM is encased in a temperature control chamber equipped with heating and cooling units and fans. The temperature of the air inside the chamber can be control between 0 °C and 50 °C (32 °F to 122 °F) by a thermostat connected to thermocouple sensors glued to the surface of the pavement. – Electrical and electronic equipment. It controls the movement of the main motor, the two screw jacks and the pressure in the hydraulic system. The software in the industrial PC controls all the components on the PTM except the heating and

16

S. A. Romanoschi et al.

cooling system. The electric power needed is 460 V, 100 Amps for the PTM machine only or 200 Amps when the heating and cooling units are also used. In the first two experiments conducted so far on flexible pavement structures, the following loading conditions were used: – Bi-directional trafficking for the “fatigue cracking” and “rutting” pavement sections and uni-directional trafficking for the “reflection cracking” pavement sections. In the uni-directional traffic, the axle is loading the pavement sections only when it travels from south toward north. – A single axle load of 81.6 kN (18,000 lb), equal to the legal limit for single axles in Texas, and a tire inflation pressure of 690 kPa (100 psi). The tire inflation pressure has been checked every four weeks and it has been found to be very stable. – A lateral wander with a maximum lateral position of 0.3 m (12 in.) with a standard deviation of 0.2 m (8 in.). In this way, the lateral wonder follows a normal distribution truncated at the central 87th percentile (6.6% on each tail) (Romanoschi et al. 2015). – The target temperature for the air inside the chamber of 20 °C (68 °F) when loading the “fatigue cracking” and “reflection cracking” sections and 40 °C (104 °F) for the “rutting” sections. To ease the utilization of the heating and cooling units, the cracking sections were tested in the Fall and Spring while the rutting sections were tested in the Summer since in this way the target temperature are close to the ambient temperatures.

Fig. 3. Schematic configuration of the PTM

Accelerated loading was applied until at least one of the following distress levels is reached: – 19 mm (0.75 in.) rut depth at the pavement surface; – 25% of each lane area is cracked (equivalent to 50% of the trafficked area cracked). The operation of the PTM has been without any major issue; the machine proved to be quite reliable. Most incidents have been related to water intrusion in the electrical

The Design, Construction and Operation of the APT Facility

17

circuits, failure of the main computer due to exposure to heat and wearing of gaskets in the hydraulic circuits. An update of the control software is scheduled for Winter 2020. Two research projects have been conducted in close collaboration with the Texas A&M Transportation Institute (TTI) to study the optimization of asphalt mixes containing various percentages of Recycled Asphalt Pavement (RAP) and Recycled Asphalt Shingles (RAS). More than four million passes have been applied in total so far to twenty pavement sections. All sections were built and tested at the APTF in Fort Worth, Texas; no testing at remote locations has been done so far. The pavements have been constructed with conventional highway construction equipment. The typical quality control tests have been performed according to AASHTO and TxDOT (2014) specifications. In addition, the control of the temperature of the asphalt mat during paving was done using an infrared based PAVE-IR device developed by the Texas A&M Transportation Institute (Romanoschi and Scullion 2014).

4 Instrumentation The accelerated loading is usually applied for one week. Then the response and condition of the tested pavement sections is measured. The instrumentation embedded in the wheel path of each pavement sections during construction included strain gages (Tokyo Sokki models KM-100HAS and PMFLS-60-50-2LTSC) and pressure cells (Geokon). The survival rate of these sensors has been approximately ninety percent. The pavement response measurements are done with the PTM machine centered and at the lateral positions of 76 mm (3 in.) and 150 mm (6 in.). Thermocouples inserted between the tested pavement sections at the surface, at 25 mm (1.0 in.) depth and at mid-depth of the asphalt surface layer are connected to an Omega logging system; they record the temperature every 15 min. TDR moisture sensors were inserted in the soil subgrade during construction but have not provided reasonable results. A Light Falling Weight Deflectometer (LFWD) equipped with three geophones has been used to measure the surface deflection at a drop load of around 2.0 kN (450 lbf) in five locations in the wheel path of each section. Akbariyeh (2015) developed a simple, regression based procedure to estimate the pavement layer moduli from the LFWD deflection data. The procedure can be used efficiently for conventional flexible pavement structures having the thickness of the asphalt concrete layer less than 100 mm (4 in.). Figure 4 shows the LFWD tests being done inside the PTM. Transverse profile measurements are performed using a specially constructed profiler (Fig. 4). The components of the profiler are: – A 2.4 m (8 ft) long aluminum beam that is very rigid so it doesn’t bend during the measurements. – A carriage that moves along the beam. Two sensors are mounted on the carriage to measure the vertical elevation of the pavement surface and the lateral position of the carriage. – A data acquisition system connected to a laptop computer. – A power supply to provide power to the sensors and the acquisition system.

18

S. A. Romanoschi et al.

Fig. 4. Measurement of the LFWD deflections and transverse profiles

When a transverse profile is measured, the sensor recording the vertical position is first positioned on a nail mounted in the centerline of the PTM machine which does not move when APT loading is applied. Then the recording is started and the carriage is moved transversely away from the nail. The recorded data is used to obtain transverse profiles with elevation data at 1.0 in. interval. Then the elevation data for the transverse profiles is then assembled in a database in Excel spreadsheet format. The Permanent Deformation (PD) at the pavement surface is calculated first in each point of the transverse profile by subtracting the measured elevation after a given number of APT passes from the initial elevation data. The permanent deformation in one point is positive when the current elevation of the point is lower than the initial elevation of that point (the point moves downward). Then, for each transverse profile, the permanent deformation (PD) is computed as the maximum value obtained from the 84 points. Then, the average of the permanent deformation for the five transverse profiles is computed. After surface cracks are first observed, crack mapping is performed at the same time with the profile measurements on the 5 m (16.6 ft) central portion of the section where the axle travels at constant speed. The cracking extent and severity are determined from the mapped data. The calculation of the percentage of area with fatigue cracking is done for a grid with the size of the squared openings of 6 in. A rectangular wood frame was built to record the location and extent of surface cracks. The frame has longitudinal and transverse strings at 150 mm (6 in.) spacing to form a grid with 150 mm (6 in.) square openings. The wood frame is positioned on each pavement section in the north half first and then in the south half of the trafficked area. While the wood frame is on the pavement, each observed crack is recorded manually on a paper template using the string grid as the reference. The extent of cracking is determined by counting the number of squares in which the cracks are present. Photos of pavement surface are also taken. Figure 5 illustrates the mapping of surface cracks.

The Design, Construction and Operation of the APT Facility

19

Fig. 5. Measuring, recording, and calculating of fatigue cracking sections

5 Summary The University of Texas at Arlington has owned and operated an Accelerated Pavement Testing (APT) facility in the Dallas-Fort Worth urban area since 2012. The loading machine is a mobile linear device designed and built in-house to reduce cost. It is capable of loading two pavements simultaneously at a rate of about 100,000 passes per week in bi-directional loading mode; uni-directional loading mode is also possible. The maximum single axle load that can be applied is 162 kN (36 kip). Controlled heating and cooling of the air above the tested pavement sections is also possible. The facility and the loading machine were designed and built entirely with internal funds over an eighteen month period. This proved to be a cost effective method. Two research projects have been conducted in close collaboration with the Texas A&M Transportation Institute (TTI) to study the optimization of asphalt mixes containing various percentages of Recycled Asphalt Pavement (RAP) and Recycled Asphalt Shingles (RAS). More than four million passes have been applied so far with the machine. The four-year long operation proved that the investment in an APT facility has been beneficial to the University. The reduced cost for the design, construction and operation of the machine have been the key to the success of the investment.

20

S. A. Romanoschi et al.

References Akbariyeh, N.: A new technique for the estimation of the elastic moduli of pavement layers from light weight deflectometer data. Master thesis. The University of Texas at Arlington, USA (2015) APT Facilities in the USA. TRB Committee AFD40 website. https://sites.google.com/site/ afd40web/apt-facilities-in-the-usa. Accessed Nov 2019 Romanoschi, S.A., Scullion, T.: Validation of the Maximum Allowable Amounts of Recycled Binder, RAP, & RAS Using Accelerated Pavement Testing. Interim Report. Report No. FHWA/TX-14/0-6682-1, The University of Texas at Arlington, USA (2014) Romanoschi, S.A., Zhou, F., Scullion, T., Saeedzadeh, R.: Testing and Evaluation of RAP and RAS Experimental Pavement Sections. Final Report. Report No. FHWA/TX-14/0-6823-1, The University of Texas at Arlington, USA (2015) TxDOT: Standard Specifications for Construction and Maintenance of Highways, Streets and Bridges (2014). ftp://ftp.dot.state.tx.us/pub/txdot-info/des/spec-book-1114.pdf

Research Progress of RIOHTRACK in China Xu Dong Wang, Lei Zhang(&), Xing Ye Zhou, Qian Xiao, Wei Guan, and Ling Yan Shan Research Institute of Highway Ministry of Transport (RIOH), Beijing, China {xd.wang,xy.zhou,w.guan,ly.shan}@rioh.cn, [email protected], [email protected]

Abstract. For the purposes of developing long-life pavements in China, in 2015. The Research Institute of Highway Ministry of Transport built a 2.038 km full-scale field pavement testing road named RIOHTRACK. The RIOHTRACK consisted of 38 pavement testing sections, in all of them some potential long-life asphalt pavement structure and material sections were set and tested, especially some semi-rigid base long-life pavement structures. The testing road was planed to finish 50 million ESALs (10 tons standard axle loading) conducted by the real trucks and trails accelerated loading before the end of 2022. Then based on the pavement performance changing rule and the structural response characteristics of different pavement, especially of the long-life pavement candidates, to study the long-life pavement design method, prediction model, distress criterion and so on. This paper introduces the basic research details of RIOHTRACK, which includes the design scheme, the loading situation, the detecting method, And the 19 main structural section pavement performance detecting results and the research findings from December 2016 to November 2019. Keywords: Test track
Long-life pavement
Performance
Asphalt pavement

1 Introduction The accelerated loading test in the field is a worldwide accepted way for research on the pavement structure and material, which can effectively reflect the service behavior of the pavement under the condition near the real-world. The earliest full-scale testing road was built by the Dutch East India Highway Association at 1920 (Metcalf 2016). By 1950s, AASHO (USA) had carried out the research of the AASHO road, the achievements of which supported the establishment of the evaluation system of pavement performance. This project is the core of AASHO empirical design method which laid the design foundation for the construction of highway network in USA (AASHO 1961). Until now, there are still many researches about full-scale testing road by different organizations in globe scope. For example, the NCAT testing track (West et al. 2019) and the MnRoad in USA (Worel 2015), the IFSTTAR accelerated pavement testing facility in France (Kerzrého et al. 2012) and the RIOHTRACK in China. Long-life or perpetual pavement is an expected goal because it has lower life-cycle cost, lower user delay costs, lower consumption of resources and lower energy consumption (Timm and Newcomb 2016). But the long-life pavement design theory and method, the threshold level of damage, the performance prediction model are need to © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 21–31, 2020. https://doi.org/10.1007/978-3-030-55236-7_3

22

X. D. Wang et al.

be confirmed firstly. These need large number of studies especially long-term field performance observation and data collection. It is difficult to realize because it needs to spend so many years. The accelerated loading facility can be used to help simulating the lifecycle in shortened time and under similar loading and environment condition with the real-world conditions. For the purposes of developing long-life pavements in china, in 2015, the Research Institute of Highway Ministry of Transport built a 2.038 km field full-scale pavement testing road named RIOHTRACK. Some long-life pavement potential structure and material were selected and be set on the testing sections. The structure includes not only some pavement structure type which had been designed as long-life pavement goal and used in china in the past 10 years, but also the full depth flexible pavement in USA. The semi-rigid base layer asphalt pavement is the main pavement type in china, so this type of long-life pavement is the main study objective of RIOHTRACK. The design loading capacity of the RIOHTRACK is 10 million ESALs (10 ton standard axle loading) per year and it is conducted by real trucks and trails, The total target ESALs of long-life pavement is 50 million, so this project is a five years plan. By detecting and analyzing the pavement performance changing rule and the structural response characteristics of different pavement especially the long-life pavement candidates, to achieve the long-life pavement design method, prediction model, distress criterion and so on. This paper introduces the basic research details of RIOHTRACK, which includes the design scheme, the loading situation, the detecting method, And the 19 main structural section pavement performance detecting results from December 2016 to November 2019. 1.1

Design Scheme of Testing Road

The RIOHTRACK is located in Beijing (39.9° north latitude), the annual average, maximum and minimum temperature are 12 °C, 35–40 °C and −20 °C respectively. In addition, the average relative humidity of Beijing is 55% with the annual rainfall of 600–700 mm. It is 2.038 km long oval with two lanes. The inner lane is the carriageway and the outside lane is the overtaking. The design driving speed is 50 km/h. The radius of the circular curve on the north and south sides are 130.5 m with the super elevation slope of 7%. The total length of straight line and easement curve sections on the east and west sides is 1428 m. The RIOHTRACK includes 25 types of asphalt pavement sections. Among them, there are 19 structural sections set on the straight line and easement curve named STR1–STR19, and 6 anti-rutting material sections set on the circular curve which are named YB-STR1–YB-STR6. The total pavement thickness is from 68 cm to 100 cm, including 12 cm to 52 cm asphalt layer. because some of the pavement structures are designed aim to the goal of long-life asphalt pavement, so with relatively thick pavement structure to bear the high-volume traffic. These 19 structural sections (shown as Table 1) can be divided into 6 categories according to the structure combination and the thickness of asphalt layer as follow: Type I: STR1–STR3 (semi-rigid base and thin asphalt layer pavement); Type II: STR4 and STR5 (composite pavement);

Research Progress of RIOHTRACK in China

23

Type III: STR6–STR9 (traditional semi-rigid base pavement popular used in china); Type IV: STR10 and STR12 (inverted); Type V: STR11–STR17 (thick asphalt layer asphalt pavement); Type VI: STR18 and STR19 (full depth asphalt pavement); Aim to achieve more balanced performance of asphalt mixture. All the types of asphalt mixtures are designed by the method of dense-skeleton, which contain the optimization design of gradation and asphalt aggregate ratio (Wang and Zhang 2014).

Table 1. The structures and materials of all the 19 structural asphalt testing sections Structural code/thickness STR1/92 cm STR2/72 cm STR3/72 cm STR4/72 cm STR5/72 cm STR6/76 cm STR7/76 cm STR8/76 cm STR9/76 cm STR10/88 cm

STR11/88 cm STR12/84 cm STR13/84 cm STR14/84 cm STR15/100 cm STR16/76 cm STR17/76 cm

Structural layers 4 cm SBSI-AC13I + 8 cm A30-AC20I + 40 cm CBG-A + 40 cm CS 4 cm SBSI-AC13I + 8 cm A30-AC20I + 40 cm CBG-A + 20 cm CS 4 cm SBSI-AC13I + 8 cm A30-AC20I + 40 cm CBG-A + 20 cm GB 4 cm SBSI-AC13II + 6 cm A30-AC20I + 2 cm SBSI-AC10 + 20 cm LCC + 20 cm CBG-A + 20 cm CS 4 cm SBSI-AC13II + 6 cm A30-AC20I + 2 cm SBSI-AC10 + 20 cm CC + 20 cm CBG-A + 20 cm CS 4 cm SBSI-AC13II + 10 cm A30-AC25I + 2 cm SBSIAC10 + 38 cm CBG-A + 20 cm CS 4 cm SBSI-AC13II + 6 cm SBSI-AC20I + 8 cm A70-AC25I + 38 cm CBG-A + 20 cm CS 4 cm SBSI-AC13II + 6 cm SBSI-AC20I + 8 cm A70-AC25I + 38 cm CBG-B + 20 cm CS 4 cm SBSIII-PAC13 + 6 cm SBSI-AC20I + 8 cm A70AC25I + 38 cm CBG-B + 20 cm CS 4 cm SBSI-AC13I + 6 cm SBSI-AC20I + 8 cm A70-AC25I + 8 cm A70-AC25I + 2 cm SBSI-AC10 + 20 cm GB + 20 cm CBGB + 20 cm CS 4 cm SBSI-AC13I + 6 cm SBSI-AC20I + 8 cm A70-AC25I + 8 cm A70-AC25I + 40 cm CBG-A + 20 cm CS 4 cm SBSI-AC13I + 8 cm SBSI-AC20I + 12 cm A70AC25I + 20 cm GB + 20 cm CBG-B + 20 cm CS 4 cm SBSI-AC13I + 8 cm SBSI-AC20I + 12 cm A70AC25I + 40 cm CBG-A + 20 cm CS 4 cm SBSI-AC13I +8 cm SBSI-AC20I + 12 cm RAPAC25II + 40 cm CBG-B + 20 cm CS 4 cm SBSI-AC13I + 8 cm A50-AC20I + 12 cm A50-AC25I + 12 cm A50-AC25I + 20 cm CBG-A + 44 cm GB 4 cm SBSI-SMA13 + 8 cm SBSI-AC20I + 12 cm A70AC25I + 12 cm A70-AC25I + 20c CBG-A + 20 cm CS 4 cm SBSI-SMA13 + 8 cm A30-AC20I + 8 cm A30-AC25I + 8 cm A30-AC25I + 20 cm CBG-A + 20 cm CS (continued)

24

X. D. Wang et al. Table 1. (continued)

Structural code/thickness STR18/100 cm STR19/68 cm

Structural layers 4 cm SBSI-SMA13 + 8 cm A50-AC20I + 8 cm A50-AC25I + 8 cm A50-AC25I + 8 cm A50-AC25I + 2 cm SBSI-AC10 + 48 cm GB 4 cm SBSI-SMA13 + 8 cm A30-AC20I + 8 cmA30AC25I + 8 cmA30-AC25I + 8 cm A30-AC25I + 20 cm CBG-B

Note: CBG-A: Cement Bonded Graded Aggregate material, which has 6% cement dosage and  6 MPa 7 day Unconfined compression strength. CBG-B: Cement Bonded Graded Aggregate material, which has 4% cement dosage, and  4.5 MPa 7 day Unconfined compression strength. CS: Cement soil which has 9.5% cement dosage and  2 MPa 7 day Unconfined compression strength. LCC: Lean Cement Concrete which has 8% cement dosage and  8 MPa 7 day Unconfined compression strength. GS: Graded Stone. SBSI, SBSII, SBSIII: SBS modified asphalt with different SBS dosage and the technical requirements. A30, A50, A70: Raw bitumen with the penetration 20–30, 40–60 and 60–80 respectively. A+70: Raw bitumen added certain dosage anti-rut agent. AC13I, AC13II, AC20I, AC20II, AC25I, AC25II: the asphalt mixture gradations with corresponding maximum particle size and passing rate. PAC13: Porous asphalt concrete.

1.2

Loading

The RIOHTRACK started to test at December 2016, It was loaded by real trucks and trails. Each vehicle drove 12 h per day and about 500 km. Until November 2019, Two types of loading mode had been used which was mode A from December 2016– December 2018 and mode B from January 2019 to now, shown as Fig. 1. The cumulative driving mileage was 1.4 million kilometers, and it was equal to approximately 20 million ESALs as 10 ton standard axle load calculate by fatigue equivalent formula in Specifications for Design of Highway Asphalt Pavement (JTG D50-2006).

Fig. 1. The axle type and axle weight of loading trucks and trails

1.3

Performance Detection

There are two types of performance detection. The one is the real-time monitoring including axle weight, dynamic stress, strain, temperature and humidity captured by the

Research Progress of RIOHTRACK in China

25

sensors. Another type is the periodical testing including FWD deflection, Rutting Depth (RD), Texture Depth (TD), Sideway Force Coefficient (SFC), International Roughness Index (IRI) and cracking amount. The periodical detecting is conducted after each 20–30 thousand kilometers. Until December 2019, there had collected about 0.43 million detecting data during the finished 70 cycles.

2 Performance and Research Findings Until December 2019, the overall road performance of RIOHTRACK is good. There were no distresses such as potholes and raveling appeared and the main typical distresses appeared were rutting and cracking. 2.1

Carrying Capacity of Structures

The deflections were detected by FWD at 10 meters interval after each 20000 km loading, calculated the average values of center point deflection D0 for each section. Figure 2(a) show the results at each detecting cycle, the horizontal ordinate was the corresponding ESALs. The average values of D0 had the same periodical changing with the temperature changing during one year. The detecting values at low temperature in January would be only half compared with the values detected at high temperature in Jane to July.

Fig. 2. (a) The detecting D0 deflection average values of each section corresponding to different ESALs. (b) The D0 deflection average values at 20 °C of each section corresponding to different ESALs

26

X. D. Wang et al.

2 0

45% 35% 25% 15% 5% -5% -15% -25% -35%

Changing rate

4

15.04%

21.63%

11.55%

26.00%

15.22%

6.77%

8.23%

9.18% -0.73%

12.20%

7.42%

17.44%

0.71%

-1.89%

32.12%

6

10.46%

8

changing rate

overtaking lane

carriage lane

-1.61%

10

15.46%

12

1.82%

Deflection of center point(0.01mm)

In order to eliminating the effect of temperature, convert all the results to the values at 20 °C standard temperature (Xiao and Wang 2019), shown as the Fig. 2(b), the values of the corrected deflection values of all the structures (except the two inverted structures of STR10 and STR12) were generally less than 10 (0.01 mm). These illustrate the stiffness of the sections were high and had no obviously decline compared with the initial situation after nearly 20 million ESALs. Moreover, in the lasted detecting cycle at November 2019, comparing the each section average D0 value of carriageway lane and overtaking lane, the average D0 value had a small increasing, the average changing rage was 10.9%. The bearing capacity was still in good situation, shown as Fig. 3.

Fig. 3. The comparison of the D0 average value of carriageway lane and overtaking lane

Further, the strain sensors in the 19 structural sections were used to capture the strain response. The layout position of the sensors was shown as Fig. 4. In each detecting cycle, utilized the FWD to load at the appointed position 1 (shown at Fig. 4), the loading level is 5 ton, at same time the strain responses at each direction were captured and analyzed.

Fig. 4. The layout position of sensors at testing section

Figure 5 gives the transverse and longitudinal strain detecting value at the 12 cm depth of pavement of STR19 during 2017–2019. The horizontal ordinate was the temperature values captured by the thermal gauges at 12 cm position at the same time of strain detecting. According to the strain data in three years, the results illustrate that the strains were sensitive with temperature, the values would increase with the

Research Progress of RIOHTRACK in China

27

increasing of temperature, but compared the strain values at the similar temperature in different year, the strain values had no obviously increasing trend but were cross or overlap.

με Strain value

-5 2017 2018 2019

-10

Longitudinal strain εy

0

Transverse strain εx

Strain value

με

0

-5

-10

2017 2018 2019

-15

-15 -10 -5

0

5

10

15

20

25

30

Temperature at 12cm below surface

35

40

-10 -5 0 5 10 15 20 25 30 35 40 Temperature at 12cm below surface

Fig. 5. The results of longitudinal and transverse strain value of STR19 from 2017 to 2019

2.2

Rutting

Figure 6 shows the rutting depth detecting results of each section with the increasing of ESALs. The rutting depth increased sharply at the first phase of loading and at the high temperature season. The rutting depths did not show a continuous increasing, but would decrease at cold season. It presented a periodical wave with temperature changing during one year. The variation curve was similar with the half sine curve. And these results were different with the traditional laboratory test results and the rutting prediction model.

Fig. 6. The results of rutting depth with ESALs

Based on the data from RIOHTRACK laboratory rutting resistance testing results of asphalt mixture, used the rutting prediction model refer to the specifications for design of highway asphalt pavement (JTG D50-2017), the predicting results were obviously different with the detection results, shown as Fig. 7.

X. D. Wang et al. 10 prediction

8

STR1 Rutting depth (mm)

Rutting depth (mm)

10

measured by autodetecting device 3m rule measured

6 4 2

prediction

8

10

STR10 Rutting depth (mm)

28

measured by autodetecting device 3m rule measured

6 4 2 0

0 4

4.5

5

5.5 6 lg(Ne)

6.5

7

7.5

STR18

prediction

8

measured by autodetecting device 3m rule measured

6 4 2 0

4

4.5

5

5.5 6 6.5 lg(Ne)

7

7.5

4

4.5

5

5.5 6 6.5 lg(Ne)

7

7.5

Fig. 7. Comparison of the RIOHTRACK detecting results and the predicting rutting depth

These could be concluded: • The elastic deformation delayed recovery property of asphalt material in high temperature conditions were not to be ignored, but these characteristics were not considered in the prediction model adequately. • The high frequency of detects help to discover and verify this phenomenon. • The prediction error depends on the structure type, the asphalt binder and so on. But because of the limitation of layered-elastic theory calculation method, the prediction model now can’t reflect the large different of different structure composition. In addition, the A30 binder engineering application data is not enough, so the prediction model has not been sufficient verified. 2.3

Crack

So far, all the cracks found in the RIOHTRACK were transverse crack, there were two main types of crack. The first type is usual transverse crack occurred in part of carriageway or throughout carriageway and overtake lane. The other type is “watermark” crack refers to the cracks which usually occur at the humid hours in the morning and gradually disappear in the afternoon. The statistics cracks in the whole road are presented in Table 2.

Table 2. Statistic of cracks in all the sections Type

Statistic results Only in carriageway Longitudinal crack 0 Transverse crack Ordinary transverse crack 15 Watermark crack 97

Whole two lane Total 0 0 4 19 0 97

This “watermark” cracks were found at May 2019 firstly, which developed fast and occupied 83.6% of all the cracks. In addition, as shown in Fig. 8. The “watermark” crack only found under the wheel track belt, which were not grow to the whole carriageway, and they were disappeared after the middle of July, see the statistics in

Research Progress of RIOHTRACK in China

29

Fig. 8. “Watermark” crack

2.4

STR19

STR17

STR18

27-May 8-Jul

STR16

21-May 23-Jun

STR14

STR12

STR13

STR11

STR9

STR10

STR7

STR8

STR5

STR6

STR4

STR3

STR1

14-May 7-Jun

STR15

18 16 14 12 10 8 6 4 2 0 STR2

Cracking amount

Fig. 9. The observations of this type of crack were going on to find the developing progress, and certain study to discover the mechanism were in progress.

Fig. 9. The occur and development time of the watermark crack

Performance of Skid Resistance

There were four types of asphalt mixture used in the surface layer of RIOHTRACK. They were SBSI-AC13I, SBSI-AC13II, SBSI-SMA13 and SBSIII-PAC13 respectively, the materials parameters given in Table 1, The same basalt aggregate were used. Collecting the texture depth detecting results of each section shown in Fig. 10. As prediction the texture depth decreased gradually with the increasing of loading repeat. And the largest decline rate occurred at the initial traffic opening period. After about 130 thousand repeated loading (equivalent to 2 million ESALs), the decreasing value reached to 73% of the total, then it would be gradually stable. Moreover, it had a stable order of texture depth as SBSIII-PAC13 > SBSI-AC13II > SBSI-AC13I  SBSISMA13. 1.60

SBSIII-PAC13

SBSI-AC13II

SBSI-AC13I

SBSI-SMA13

TD mm

1.40 1.20 1.00 0.80 0.60 0.00E+00

5.00E+06

1.00E+07

ESALs

1.50E+07

2.00E+07

Fig. 10. The detecting results of the texture depth

On the other hand, during April to November of each year, The Side-way Force Coefficient (SFC) were tested monthly by the SFC vehicle, the results shown in Fig. 11. The SFC value presented a decreasing trend with the increasing of repeated loading. But because of the Instabilities of the detecting equipment, the testing results exit a larger fluctuation regretfully.

X. D. Wang et al.

SFC

30

70 65 60 55 50 45 40 35 30 0.00E+00

SBSIII-PAC13

5.00E+06

SBSI-AC13II

1.00E+07 ESALs

SBSI-AC13I

1.50E+07

SBSI-SMA13

2.00E+07

Fig. 11. The detecting results of SFC

2.5

International Roughness Index (IRI)

The IRI detecting results of the 19 structural sections were a remarkable difference, the IRI results were in the scope of 1–2.5 m/km. But it was stable in each section during the whole 70 detecting cycles, because there were no remarkable potholes and raveling occurred, the IRI values are depending on and close to their own initial levels.

3 Conclusions RIOHTRACK has multi-type pavement structural testing sections, and its coupling condition of the loading and environment is closely real-world. In addition it conduct the life-cycle and high frequency performance monitoring and data analyzed. These all make it to be an effective approach to improve the pavement design method and prediction models. This paper only gives the basic information and detecting results in the past three years, the mechanism of certain performance, the relationship between the structure composition and material characteristic and so on are not involved in. Acknowledgement. The research is funded by the Fundamental Research Funds for the Central Research Institute (Grant No. 2019-0027).

References Metcalf, J.B.: A brief history of full-scale accelerated pavement testing facilities to 1962. Paper Presented at the International Conference on Accelerated Pavement Testing, University of Costa Rica, Costa Rica, 19–21 September 2016 (2016) AASHO: The AASHO road test, history and description of the project. National academy of sciences-national research council, Special Report 61A, Highway Research Board, Washington, DC. (1961) West, R.: Phase VI (2015–2017) NCAT test track findings. National center for asphalt technology, July 2019 (2019) Worel: Benefits of MnROAD Phase - II Research, Research Project Final Report 2015–19, May 2015 (2015)

Research Progress of RIOHTRACK in China

31

Kerzrého, J.P., Hornych, P., Chabot, A., Deloffre, L., Trichet, S., Coirier, G., Gouy, T.: Evaluation of the aggressiveness of different multi-axle loads using APT tests. In: Proceedings of the 4th International Conference on Accelerated Pavement Testing, Davis, California, USA, 19–21 September (2012) Timm, D.H., Newcomb, D.E.: Perpetual pavement design for flexible pavements in the US. Int. J. Pavement Eng. 7(2), 111–119 (2016) Wang, X., Zhang, L.: Multi-performance Design of HMA Based on Stone Inter-lock Structure Theory. Beijing (2014) Xiao, Q., Wang, X.: Study on feature and fitting of RIOHTRACK pavement temperature curve. J. Highway Transp. Res. Dev. 2019(3), 36:1–36:6 (2019) Liu, B.-Y., et al.: Specifications for Design of Highway Asphalt Pavement. China Communications Press, Beijing (2017)

Two Years of APT Program on the New Test Site duraBASt Bastian Wacker(&) and Dirk Jansen Federal Highway Research Institute (BASt), Bergisch Gladbach, Germany [email protected]

Abstract. In addition to laboratory investigation, innovations in road construction require the opportunity to be analyzed and tested before they are used in the sensitive road network. This minimizes the risk of failure of a new development. For this purpose, the German Federal Ministry of Transport and Digital Infrastructure (BMVI) has developed a research program in cooperation with the Federal Highway Research Institute (BASt) to promote innovations. One objective of the research framework program “Roads in the 21st century Innovative road construction in Germany” (BMVBS 2012) is to bring innovations faster towards real-world applications. A test area was planned and built for this purpose in order to realize large-scale projects. The demonstration, investigation and reference areal of BASt (duraBASt) was officially opened in October 2017. The realization of eight different scientific projects has been part of the first construction cycle. The APT program of BASt was used at five out of these projects. Among other things, self-healing asphalt, constructions for energy harvesting from roads, but also various joint connections between precast concrete elements were tested. Due to the variety of projects in this first construction cycle, valuable and comprehensive information could be collected in order to plan and implement future APT projects in a more project-specific way. In addition, the complexity of the tests under real weather conditions was continuously investigated. With duraBASt and the existing test facilities, the national and international research community has a test infrastructure at its disposal that has grown to meet the challenges of future road constructions. Keywords: duraBASt
Accelerated Pavement Testing
Innovation
Research infrastructure

1 Introduction The current Federal Transport Infrastructure Plan 2030 of the Federal Republic of Germany (BMVI 2016) deals primarily with the maintenance and renewal of existing infrastructure. This means rethinking of strategies at some points for an industry which mainly focused on new constructions for decades. This has been identified as a task by the Federal Ministry of Transport and Digital Infrastructure (BMVI) in cooperation with the Federal Highway Research Institute (BASt) within the framework “Roads in the 21st Century” (BMVBS 2012) and, among other things, promoted with the construction of a large-scale test facility. © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 32–39, 2020. https://doi.org/10.1007/978-3-030-55236-7_4

Two Years of APT Program on the New Test Site duraBASt

33

In addition to the necessary laboratory scale evaluation, BASt’s demonstration, investigation and reference areal (duraBASt) offers researchers from external institutes and companies, as well as BASt’s own researchers, the opportunity to conduct investigations under realistic and full-scale conditions. These investigations are crucial in order to check and further improve the development prior to use in the sensitive transport network. The main aim is to optimize maintenance and renewal actions in the existing infrastructure. Parallel to a rapid repair systems for concrete pavements, based on precast concrete slabs, new methods for maintaining asphalt pavements were tested. These included a project for the further development of self-healing asphalt (HEALROAD), but also a project aiming on the optimization of asphalt layer design on top of a relaxed concrete pavement (relaxed hybrid) and the investigation of the potential of cold recycling with foamed bitumen. Before, technologies for optimized asphalt paving, durable drill core sealing techniques and energy harvesting from road construction (SEDa) were also investigated. Research projects with a low Technology Readiness Level (TRL) up to a high TRL can be implemented, because the areal is separated from the public traffic. This paper is intended to summarize the BASt experiences, especially in the field of Accelerated Pavement Testing (APT).

2 Demonstration, Investigation and Reference Areal of BASt In addition to BASt, the Federal Ministry of Transport and Digital Infrastructure (BMVI) and the Road Administration of North Rhine-Westphalia (Straßen.NRW) were also involved in the implementation of the demonstration, investigation and reference areal (duraBASt). In addition to the sufficiently large area, the existing infrastructure was also decisive for the choice of location. duraBASt is located east of Cologne on the A3 motorway parallel to the northbound direction of travel at the A4 motorway intersection. It consists of an open area, a bridge structure and a tunnel-like situation. In the course of the construction process, the entire site was fenced in and secured by surveillance technology. As a result, public road traffic is not possible and the necessary equipment, such as data logging and measuring systems, but also the machine for the APT is protected. The construction of the 25,000 m2 (length approx. 1,000 m) test facility took place between the years 2015 and 2017. The reference tracks (R areal) are required for the approval of measuring vehicles for the regular condition assessment and evaluation of surface characteristics (ZEB). There defined surface characteristics and damages have been constructed. These are also helpful for further developments in measurement technology. For the demonstration and investigation of constructional matters (D/U areal), a total of nine sections are available on the entire area. Six of these (width 3.9 m to 5.5 m and lengths up to 100 m) are located in the central area and are used for APT. In addition to sufficient space for paving with standard paving equipment, the required drainage and power supply are also provided. This makes it possible to respond to the individual project-specific requirements. The D/U area can therefore also be used for industrial issues and can thus support the development work of road construction and mechanical engineering companies.

34

B. Wacker and D. Jansen

A modern office building is available for project meetings and events as well as for the employees. This also includes large enough parking spaces to allow larger groups to drive onto the site. Since 2017 the following projects have been realized: self-healing road (HEALROAD), polyurethane liners (INNOPAVE), Process Optimization in Asphalt Road Construction (OBAS), open-porous asphalt (OPA), open-porous concrete (OPB), optimization of core closures, precast concrete elements for preservation and new construction, use of reclaimed asphalt together with rejuvenators, energy harvesting in roads (SEDa). Of the nine projects, five were conducted with an APT program, the others with conventional measurement methods. For the years 2019 and 2020 several projects are planned, for example experiments in the field of cold recycling, design of road constructions and analysis of innovative materials.

3 Accelerated Pavement Testing Accelerated Pavement Testing (APT) is used to simulate the load of heavy freight vehicles in a shorter time (Steyn 2012). With the standardized BASt APT program (Wacker et al. 2018a; Wacker 2019) it is possible to work on the projects described above in a structured way and to generate measurement data that can also be used in other projects or for comparison purposes. Three key functions (load facility, research infrastructure and investigation program) were detected for performing sufficient APT projects. The BASt uses the Mobile Load Simulator MLS30 and uses both indoor (test halls) and outdoor (duraBASt) research infrastructures. The investigation program is partially modified or slightly adapted for the necessary project objectives, but consists mainly on non destructive testing, embedded sensors and laboratory tests at the end of the project. 3.1

HEALROAD

The HEALROAD project (Induction HEating AsphaLt mixes to increase ROAD durability and reduce costs, pollution and disruptions) was sponsored by INFRAVATION and was carried out by the project partners University of Cantabria, University of Nottingham, HEIJMANS, SGS Intron, ERF and BASt. After completion of the laboratory tests, a test track was realized on duraBASt and loaded with the MLS30 (Wacker et al. 2018b). The asphalt mix could be produced and paved according to current standards and a homogeneous test track was created. Further investigations focused on producing small cracks within the bitumen film between the aggregates and in the worst case stone loss from the surface course. This was to be prevented by healing the pavement with the help of induction energy at different locations and at different times. For the implementation, the APT program was supplemented by additional settings on the MLS30 and enhanced documentation during the measurements. The loading device was set as follows: high wheel loads (75 kN), slow speeds (2,000 mm/s) and the use of a lateral moving offset were to lead to an increase in shear forces. For documentation purposes, all dissolved particles were collected after each loading day and the area as well as particles was documented by surface images and photos (Fig. 1).

Two Years of APT Program on the New Test Site duraBASt

35

Fig. 1. Stone loss during APT program

During the loading process it was noticed that stones were loosened and damaged under the load. Unfortunately, the quantity was not sufficient to be able to make a statistically verified analysis for evaluating the positive influence of induction energy. Further laboratory tests on samples from the test track showed that the samples where induction energy was used at an early stage showed less stone loss than the other samples. From this project, valuable experience was gained with regard to sensor installation and documentation, as well as uniform data labeling. In addition, it has to be said that the MLS30 is not the right tool for this type of testing, because the lateral movement seems to take place to slowly (4,0 mm/s) and therefore less shear forces could be applied. Since the speed of the lateral movement cannot be adjusted, the above mentioned settings are the maximum that can be achieved. Also first experiences with the standard measurements (evenness, bearing capacity, etc.) under not controllable climatic conditions with use of reference points were collected. 3.2

SEDa

The SEDa project (Höller 2019) dealt with energy harvesting from roads. For this purpose a 100 m long test track was realized and two different collector systems were installed (Fig. 2). These collector systems are necessary to dissipate the energy introduced by the sun by means of a liquid and to convert it into energy. For this purpose, these collector systems were arranged below the surface layer in a so-called bedding layer. Two different materials (mastic asphalt and thin-layer cold asphalt) were also used in the bedding layer. A total of approx. 200 temperature sensors were installed at different depths and different constructions as well as sensors to monitor the mechanical status of the collector systems. The sensors recorded the current situation every 15 min.

36

B. Wacker and D. Jansen

Fig. 2. Two different collector systems (left: steel, right: plastic)

Three load points were provided for the Accelerated Pavement Testing (all with mastic asphalt bedding, one collector system each and one point without collector systems as a reference). The responsible research group decided to load one of the load points with 1.2 million and the others with 0.2 million loading cycles. Unfortunately, the strain and compressive sensors failed. Therefore, only data by external measurement techniques are available. From these measurement data it can be deduced that there has been little change within the construction and that only the surface has drawn the tire profile due to the true to track load. In this project, moisture sensors (M) were also installed for the first time in order to monitor the conditions. These experiences should be incorporated into the further development of the sensor equipment. The measured values of the sensors were classified as not plausible, because the difference to each other was clearly too large and values of 100% were indicated. This led to the development of a new sensor system for the following test sites, in which other sensor types are used. In addition, this shows the need for regular exchange with other institutes regarding the instrumentation of test sites. 3.3

Relaxed Hybrid

The “relaxed hybrid” project is being implemented as part of cooperation between the company TPA GmbH and BASt. The aim is to reduce the asphalt layer thickness by not removing an existing, old concrete pavement, but simply relaxing it with an Impactor (Fig. 3) and then pave the Asphalt over the cracked concrete surface. On the one hand, this is intended to create long-lasting road structures, but above all to save resources, optimize construction site logistics and save costs. Since this is not a new construction principle, the project aims on deriving input parameters for the further development of pavement design methods in these cases.

Two Years of APT Program on the New Test Site duraBASt

37

Fig. 3. Use of an Impactor to relax the concrete pavement

The installation of the test section was completed by mid 2019 and has since been subject to an APT program to determine initial parameters. These parameters are to reflect the changes during the construction process. In the further course of the project, the existing design methods will be further developed. Of particular interest here are the parameters with which the relaxed concrete road surface can be taken into account as subbase. For this purpose, extensive bearing capacity measurements were carried out after the individual construction phases (on the unbound layer, on the paved concrete (40 days after paving), on the relaxed concrete (68 days after paving) and on the following asphalt layers. In addition, the behavior of an old concrete pavement has been simulated using a very stiff concrete and an artificially installed weak spot. First evaluations show that the previous assumptions for the pavement design, at which relaxed concrete pavements have been treated like unbound subbase layers, are clearly on the safe side and that there is therefore a high degree of safety in design. This is to be confirmed by further investigations and calculations after completion of the APT program in November 2019. Figure 4 shows very clearly the effect of the relaxation on the concrete roadway and that the artificially installed weak spot can be clearly seen at all three measurement intervals (approx. station 13 on the x-axis). This means that, in case of real constructions sites, this weak spot could have been also be detected during the construction process with the help of bearing capacity measurements and that measures for improvement could be initiated during the ongoing construction process. The entire structure was equipped with various sensors in order to obtain strain and acceleration values as well as temperatures. During this installation, the experience gained previously was put into practice with success. 31 acceleration and two strain sensors survived the installation process and since then have regularly delivered values at the respective measuring points. In addition to the sensor measurements, various surface investigations are carried out. Transverse evenness measurements and highresolution surface image recording as well as skid resistance and texture measurements are carried out regularly. After completion of the load tests at the end of 2019, further

38

B. Wacker and D. Jansen

Fig. 4. Results of FWD bearing capacity measurements - before and after relaxing and on top of the asphalt pavement

findings are to be generated by extensive blending of the measured values determined and further laboratory tests. These will be fed into the development of regulations and other research projects.

4 Conclusion With the demonstration, investigation and reference areal of the BASt (duraBASt) there will be wide-ranging possibilities in national and international road construction research in the future. Within the first project and construction cycle, various projects were realized and investigated by means of Accelerated Pavement Testing (APT). In

Two Years of APT Program on the New Test Site duraBASt

39

addition to the detection of problems in the instrumentation, the processes have been continuously improved and standardized. This enables BASt to tackle research tasks with the APT program and also to assist industry with further developments. One example is the relaxed hybrid project, in which parameters for designing the asphalt structure on a relaxed concrete pavement can be determined. These extensive tests can be carried out very well and above all safely on such a test site. The previous research projects HEALROAD and SEDa have shown the respective research groups on the basis of the investigations and APT programs in which areas further research is required, but also that large-scale installations are possible with innovative materials and designs.

References BMVBS: Straße im 21. Jahrhundert – Innovativer Straßenbau in Deutschland (Road in the 21st century - Innovative road construction in Germany) (2012). https://www.bast.de/BASt_2017/ DE/Publikationen/Medien/Dokumente. Accessed 18 Oct 2019 BMVI: Bundesverkehrswegeplan 2030 (Federal Transport Infrastructure Plan 2030). Federal Ministry of Transport and Digital Infrastructure, Berlin (2016) Höller, S., Golkowski, G.: Untersuchung multifunktionaler Straßenbaumaterialien und Verbundwerkstoffe zur Nutzung solarer Energie und Verbesserung der Dauerhaftigkeit Großmaßstäbliche Analyse der mechanischen Belastbarkeit unterschiedlicher Asphaltkollektorsysteme (Investigation of multifunctional road construction materials and composites for the use of solar energy and improvement of durability - Large-scale analysis of the mechanical load capacity of different asphalt collector systems), BMBF (2019) Steyn, W.: NCHRP Synthesis 433: Significant Findings from Full-Scale Accelerated Pavement Testing. Transport Research Board, Washington, D.C. (2012) Wacker, B., Jansen, D.: Toolbox for APT program monitoring at BASt and optimized data handling on new test facility duraBASt, TRB Annual Meeting (2018a) Wacker, B., Kalantari, M., Beatens, B., van Bochove, G.: HEALROAD: induction heating asphalt mixes to increase road durability and reduce maintenance costs and disruptions Deliverable 5.2: evaluation of accelerated pavement test and analysis of structural changes, research program Infravation (2018b) Wacker, B.: Zeitraffende Belastungsversuche mit integriertem Einsatz zerstörungsfreier Messverfahren (Accelerated Pavement Testing with integrated use of non-destructive measuring methods), PHD Manuskript April 2019 (2019)

Guidance for the Next Generation Accelerated Pavement Testing Facilities Benjamin Worel1(&), Michael Vrtis1, and R. Buzz Powell2(&) 1

2

Minnesota Department of Transportation, Office of Materials and Road Research, 1400 Gervais Avenue, Maplewood, MN 55109, USA {ben.worel,michael.vrtis}@state.mn.us National Center for Asphalt Technology, Auburn University, 2700 Technology Parkway, Auburn, AL 36830, USA [email protected]

Abstract. Full-scale Accelerated Pavement Testing (APT) facilities are constantly evolving since their inception in terms of their contributions to research/engineering, funding, and facility operations. This paper is intended to serve as a guidance to new facilities that are being initiated, as well as a longterm vision for longstanding facilities such as the Minnesota Department of Transportation’s Road Research Facility (MnROAD) and the National Center for Asphalt Technology’s (NCAT) Pavement Test Track. The recommendations provided are based on decades of experience at both facilities. Both facilities have recognized the importance of collaboration and partnership in improving data quality, research impacts, and fiscal responsibility. The benefits of various types of collaborations and partnerships from both facilities are presented. Each facility has learned to cater their research objectives to different customer types or funding sources. MnROAD and NCAT have found it beneficial to utilize their facilities for non-pavement research to offset operating costs and improve relationships with parent agencies and sponsors. Keywords: Accelerated Pavement Testing
Facility operations
Funding and partnerships

1 Introduction and Background Longstanding full-scale accelerated pavement (APT) testing facilities have been forced to adapt to changes in local research priorities and funding availability. These changes have forced some APT facilities out of existence over the past 50 years. While the benefits of APT facilities may appear to be obvious to those closely involved, the facilities are often challenged with justifying their existence. The Minnesota Road Research Facility (MnROAD) and the National Center for Asphalt Technology (NCAT) have been fortunate enough to experience and learn from these challenges over the past 20 years. MnROAD was conceived in the early 1990’s and was initially funded primarily by the Minnesota Department of Transportation (MnDOT) and the Minnesota Local Road Research Board (LRRB). MnDOT continues to own and operate the facility however © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 40–48, 2020. https://doi.org/10.1007/978-3-030-55236-7_5

Guidance for the Next Generation Accelerated Pavement Testing Facilities

41

MnROAD has evolved into more of a self-sufficient funding model over the years. At the present time, MnROAD has a number of states, industry, and academia working on common research needs including its current partnerships. NCAT’s Pavement Test Track is a full-scale APT facility that has been operational for over 20 years. It is a cooperatively funded project with individual test sections sponsored by state and federal highway agencies as well as commercial interest groups. The focus of research within each section can include mix/materials, structural pavement design, and/or pavement preservation. The NCAT Pavement Test Track is a nonprofit research project in which all of the cost to execute each 3-year research cycle must be recovered via test section sponsor fees that vary by scope of work (Brown et al. 2002; Timm et al. 2006; Willis et al. 2009; West et al. 2012; West et al. 2018). The pavement research partnership with MnROAD was added in the 2015 research cycle that included both pavement preservation and asphalt mix performance testing (West et al. 2019).

2 Objective and Scope The objective of this paper is to serve as a guidance to new facilities that are being initiated, as well as a long-term vision for longstanding facilities. The paper highlights the limitations encountered when APT facilities operate independently, importance of partnerships, understanding customer needs, and how non-pavement research can utilized to help diversify funding and participation.

3 Individual Acting Test Track Limitations Historically, pavement test tracks in the United State have provided critical insight into pavement performance, design, and construction and have been very beneficial to its direct and indirect customers. While providing this benefit, test tracks have also learned that there are limitations that need to be understood if the results are to be utilized by a larger group of customers. The current MnROAD and NCAT efforts are working to limit these limitations from our past experience over the years through the development of partnership. It should be noted that the limitations discussed are generalized and are not assumed or based on an individual facility. The singularity of an APT’s location and climate have been a limitation from the onset for any study or test track. Research is typically developed and funded on a local level; therefore, only the local group can readily implement the findings. Other groups (states) are often skeptical of the results noting there are unique materials, specifications, climate, traffic, etc. APT facilities build test sections based on the funding that can be obtained. This is typically from local states interested in local topics or interpretation of the national trends but typically related back to the local groups. This limits a regional and national input in the development of the study, test section construction, and input on results that are found including the implementation of the results. Limiting ideas, creativity,

42

B. Worel et al.

and larger perspective diminishes the larger success of any work done at any facility because other states do not have influence in or even awareness of the study. There is an art to designing test sections that demonstrate long-term pavement performance in a short time period for the sponsors. Pavements typically take many years to develop either roughness/cracking (or withstand distress) to demonstrate a design or study objective. Many times test studies develop too many variables than can be researched. Influence from numerous perspectives is important for getting input however, test sections that isolate individual variables need to be developed. Control sections are important for comparison of the experiment variables back to existing network conditions. Examples include MnROAD’s experiences with several thermal cracking investigations. The initial phase incorporated too many variables that confounded the results. The subsequent phase utilized consistent structure and aggregates with only a varying asphalt binder grade. This allowed a direct comparison of the three binder grades without other variable impacting results (Marasteanu et al. 2007). Diversified funding is imperative for APT facilities as funding is inconsistent and subject to political changes or other changes that are outside of the facility’s control. Have various funding sources improves the facility resilience to changes that undoubtedly will occur. The need for diversified funding provides the impetus for understanding customer needs and developing partnerships.

4 Types of Customer Needs A key aspect for success of APT facilities is the ability to meet the needs of different customers. Although there is some overlap in customer type, most research sponsors fall into the following categories of: government transportation agencies (local, state, federal), private industry (associations, contractors, products, consultants) and academia. Each of these customer groups has different types of research objectives, data collection and sharing requirements, and reporting focus and needs. APTs need to understand partner’s needs and develop balance in their focus between applied and basic research needs. Figure 1 shows a schematic of where different APT customer types typical fit with their research focus ranging from applied research to basic research. On the left side of the scale is applied research and to the right is basic research. Lawrence Berkeley National Laboratory defines applied research as “… research designed to solve practical problems of the modern world.” Basic research is also known as pure or fundamental research and is research with no obvious immediate commercial value that advance the scientific understanding (Lawrence Berkeley National Laboratory 2017). The chart shows the primary and secondary focus of each group. For example local agencies typically need products that explain how to do things now and have little time to wait for future research efforts. States and Federal agencies typically are also looking for applied research but have a little more resources and time to wait on the development of basic research for the future. Associations and contractors typically lean more towards applied research other than industry developing products do push out more innovation and new ideas to help sell highway products. Academia goal is to promote basic research that will solve todays problems with new technology. while

Guidance for the Next Generation Accelerated Pavement Testing Facilities

43

they are interested in the future. Test tracks need to be able to balance both applied and basic research to effectively utilize their facilities to the benefit of all customers.

Fig. 1. Schematic of typical research emphasis for APT customer types

4.1

Government Transportation Agency Research

Both MnROAD and NCAT were started from large investments from their state transportation department (MnDOT and Alabama Department of Transportation (ALDOT), respectively). These agencies can be considered the parent agencies for the facilities. MnROAD is owned and operated by MnDOT. Addressing the needs of the parent agency is imperative to the survival of the facility, especially with the changes that can occur in top agency management with election cycles. The research for parent agencies typically have shorter timeframes and require more immediately implementable results mostly focusing on state and local research needs. Agency APT research is often aimed at smaller adjustments to materials or design that could lead to large cost savings when applied to the entire state network. For example, ALDOT sponsored NCAT research that adjusted the layer coefficient for asphalt concrete design in Alabama. The project resulted in a thinner recommended design thickness for ALDOT’s asphalt pavements and was estimated to save ALDOT between $25 and $50 million a year, depending upon the size of the annual paving program (Peters-Davis and Timm 2009). MnROAD has also seen this type of benefits to their local government agencies as noted in the Phase-I and Phase-II benefits (REF). One of the major results listed in the benefits report, were the placement of spring load restrictions which were developed from MnROAD research. Similar to NCAT’s work for ALDOT, small changes were generated from APT research that led to large benefits when applied to the state and local network. 4.2

Private Industry Research

APT research for private industry is equally important to advance the state-of-practice in pavements. Private research often is willing to take risks that agencies cannot and

44

B. Worel et al.

thus the results can bring large benefits to the pavement community. This type of research often presents unique challenges with proprietary information and data sharing, especially since MnROAD and NCAT are primarily publically funded facilities. Overcoming these challenges can seem daunting but MnROAD has been able to find ethical ways to accommodate this type of customer and facilitate this vital source of research progress. 4.3

Academic Research

Research projects for academia generally tend to be more complex, have longer timeline, and have less of a focus on immediate network implementation. As shown in Fig. 1, academia is able to develop advances on basic understanding research. This type of research is extremely valuable to the pavement community and should be encouraged however it is often underappreciated from local and state funding sources. Again, each institution and their ability to cover applied and basic research are unique.

5 MnROAD and NCAT Partnership Partnerships have helped MnROAD and NCAT improve their data quality, expand the impact area of their research, and reduce the fiscal burden of operating an APT facility. MnROAD and NCAT developed a formal partnership in 2015. At the time, both facilities were interested in similar investigations and decided that a collaboration was the best way to pursue their mutual objectives. Two national pooled-fund experiments were initiated which includes a cracking group experiment and a pavement preservation group experiment. The objective of the Cracking Group (CG) CG experiment was to investigate asphalt concrete laboratory cracking tests. Seven different laboratory cracking tests were evaluated on mixtures designed to have differing cracking susceptibilities. The test were evaluated based on repeatability, ease of testing/sample preparation, differentiation of results, and correlation to field performance. To facilitate the field performance comparison, seven test sections were constructed on the NCAT in 2015 and eight sections were constructed at MnROAD in 2016 with a range of expected cracking performance. The NCAT sections were designed to investigate top-down cracking and the MnROAD sections focused on low-temperature cracking. Instrumentation, performance monitoring, data collection and processing were carefully coordinated between the two facilities. The coordination has allowed for comparison between the sections and the sites. The Pavement Preservation (PP) experiment was initiated to investigate numerous pavement preservation treatments at different traffic levels in warm and cold climates. The study began at NCAT in 2012 when 30 unique test sections were created on a lowvolume county road in Auburn, AL. In 2015, companion test sections were built on a high-volume highway in Opelika, AL (US 280). The sections included chip-seals, micro-surfacing, thin-asphalt overlays, and recycling/reclamation treatments (full-depth reclamation, cold in-place recycling, and cold central plant recycling). Similar sections were constructed in Minnesota in 2016 on both low and high-volume roadways in

Guidance for the Next Generation Accelerated Pavement Testing Facilities

45

Pease, MN. The final phase of construction was completed in summer 2019 when different recycling technique sections were constructed on a local road immediately outside of MnROAD. The companion recycled sections were part of the high-volume construction in Opelika, AL in 2015. As described in the project descriptions from the MnROAD/NCAT partnership, the partnership allowed the impact of climate to be compared. MnROAD and NCAT have very different climates especially in the winter months. The average high and low temperatures in January are −5.6 C and −15 C in Minnesota and 14 C and 2 C in Alabama (rSSWeather.com). Alabama has a wetter climate than Minnesota and averages 75% more annual rainfall (US Climate Data). The MnROAD/NCAT Partnership has helped overcome the regional limitations of APT facilites. Figure 2 shows the current sponsors of MnROAD/NCAT partnership experiments. It can be seen that the combination has led to new sponsors outside the geographical region of each facility.

Fig. 2. United States map with MnROAD/NCAT Partnership sponsors highlighted

6 MnROAD’s Individual Partnerships MnROAD has evolved into a long history of utilizing partnerships to improve research including partnerships with the LRRB and a number of other pooled-funds leading up to the current National Road Research Alliance (NRRA) efforts that has greatly expanded the influence of partnerships for MnROAD. The NRRA was started in 2016 and currently has eight state agency members and over 55 associate members from industry and academia. One of the key benefits of the NRRA is bringing agency, industry, and academia together to help coordinate and design research that is beneficial to all stakeholders. The NRRA has funded 37 research projects (over $4.4 million) in concrete, asphalt, geotech, intelligent construction, and pavement maintenance, primarily at the MnROAD facility REF to website. NRRA contracts have been award to multiple universities, consultants and associations. MnROAD has found it beneficial to include non-pavement research. It is of utmost importance that the non-pavement research cannot impact or jeopardize any of the

46

B. Worel et al.

pavement research that the facility was built for. The non-pavement research activities have helped with operating expenses and also have helped expanded the perceived value of the facility to those outside the traditional pavement audience. MnROAD has allowed numerous internal and external users to utilize the facility. Some of the internal (state agency) users include: Minnesota Pollution Control Agency, MnDOT Office of Connected and Autonomous Vehicles (CAV-X), and Minnesota State Patrol. External users of MnROAD include 3M, Honeywell, Intercomp, National Cooperative Highway Research Program, and many other consultants and universities.

7 NCAT’s Individual Partnerships NCAT also has a history of pursuing partnerships that are complementary to the pavement research program. The transportation pooled fund that supported the first research cycle was funded completely by eight state agencies. Private sector partners began to fund test sections as full partners with state agencies in the second (2003) research cycle. In all subsequent research cycles, test sections were funded by private sector partners. Additionally, several private companies always support Track construction and maintenance activities through the donation of materials, equipment, services, and technical support. Access to the NCAT Pavement Test Track is available to researchers inside and outside Auburn University (AU) for nonprofit cost recovery through what college accountants call a Service Center. Some examples of past on-Track research/activities include precision and bias determination of published test procedures, validation of other APT technologies, certification of inertial profilers, transit bus testing, and pavement striping and marking. Some examples of off-Track research/activities include erosion control products/processes, sinkhole studies, herbicides, aerial vehicle testing, and emergency response exercises. Numerous on- and off-Track autonomous vehicle projects have been run that could be categorized either way. The entire 309 acre property is designated as a National Geotechnical Experimentation Site (NGES), which makes it attractive to geotechnical researchers at many other schools. NCAT has benefited from other research in several ways. Most directly, the Track Service Center is setup to reduce the cost of test sections by shifting cost to other research clients. Autonomous vehicle research has great value in informing Track personnel on the best options for future fleet automation. Erosion control research is beneficial to the best management practices that are utilized every 3 years when the Track is rebuilt, and it’s a great example of research that is perfectly complementary to the pavement research program.

8 Future Needs and Support Future operations of test tracks including MnROAD and NCAT are expensive to run and operate and are dependent on consistent funding support. APT facilities need to have a range of customers and develop partnerships to expand beyond their regional limitations. In the United States, there is space for the Federal Highway Administration

Guidance for the Next Generation Accelerated Pavement Testing Facilities

47

to increase collaboration with and among APT facilities. An example is the recent inclusion of MnROAD data into the infopave.com website. Infopave hosts Long-term Pavement Performance (LTPP) data. By including MnROAD data with the LTPP platform, there is potential to increase usage of MnROAD data by non-traditional APT audience. TRB’s AFD40 committee has been critical to the development of partnerships and collaborations both formally and informally. Continued meetings and discussions are necessary to improve all facilities around the world and preserve the benefits of APT facilities for the pavement community.

9 Summary and Conclusions APT is a vital tool to the pavement community that has provide numerous advances. Successful APT facilities need to be able to adapt and learn from the lessons of previous facilities. MnROAD and NCAT have made their share of missteps over their decades of operation but have been able to continue operating by adapting to new customers and research objectives. • Limitations that arise from independent APT facilities can be overcome through collaboration and partnerships. • Partnerships are an essential part of successfully operating an APT facility. Collaboration and partnerships have allowed MnROAD and NCAT to improve data quality, expand the impact area of their research, and improve funding constraints. • APT facilities have different customers with different research demands. Being able to meet the needs of each customer is vital to the longevity of the APT facility. • Facilities need a balance of basic and applied research. • Non-pavement research can be incorporated into APT facilities to offset operation costs and improve relationships with other agencies/departments.

References Brown, E.R., et al.: NCAT Test Track Design, Construction, and Performance. National Center for Asphalt Technology, Auburn. US Climate Data (2002). https://www.usclimatedata.com/ climate/auburn/alabama/united-states/usal0035. Accessed 24 Nov 2019 Lawrence Berkeley National Laboratory: Basic vs. Applied Research (2017). http://www.sjsu. edu/people/fred.prochaska/courses/ScWk170/s0/Basic-vs.-Applied-Research.pdf. Accessed 24 Nov 2019 Marasteanu, M., et al.: Investigation of Low Temperature Cracking. Minnesota Department of Transportation, St. Paul (2007) Peters-Davis, K., Timm, D.: Recalibiration of the Aspalt Layer Coefficient. National Center for Asphalt Technology, Auburn. rSSWeather.com, n.d. Climate for Minneapolis-St.Paul, Minnesota (2009). http://www.rssweather.com/climate/Minnesota/Minneapolis-St.Paul/. Accessed 24 Nov 2019 Timm, D., et al.: Phase II NCAT Test Track Results. National Center for Asphalt Technology, Auburn (2006)

48

B. Worel et al.

West, R., et al.: Phase V (2012–2014) NCAT Test Track Findings. National Center for Asphalt Technology. Auburn (2018) West, R., et al.: Phase VI (2015–2017) NCAT Test Track Findings. National Center for Asphalt Technology, Auburn (2019) West, R., et al.: Phase IV NCAT Pavement Test Track Findings. National Center for Asphalt Technology, Auburn (2012) Willis, R., et al.: Phase III NCAT Test Track Findings. National Center for Asphalt Technology, Auburn (2009) Worel, B.J., Jensen, M., Clyne, T.R.: Economic benefits resulting from road research performed at MnROAD. In: Third International Conference on Accelerated Pavement Testing, Madrid (2008) Worel, B., Van Deusen, D.: Benefits of MnROAD Phase-II Research. Minnesota Department of Transportation, St. Paul (2015)

APT of Asphalt Concrete

Laboratory Investigation of Cracked Asphalt Pavement Structure Using Accelerated Loading System Youness Berraha1,2(&), Daniel Perraton1, Guy Doré2, Michel Vaillancourt1, and Jean-Pascal Bilodeau2 1

2

Department of Civil Engineering, Laval University, Quebec, Canada [email protected] Construction Engineering Department, École de Technologie Supérieure, Montréal, Canada

Abstract. During winter season, cold temperatures will promote the development of tension stresses in the hot-mix asphalt (HMA) layers which can, in turn, lead to the occurrence of low-temperature cracking. Although the mechanisms and factors fostering the appearance of low-temperature cracks are already well understood, the consequences of this degradation on the structural capacity and freezing behavior of the pavement structure near the crack are still poorly documented. The purpose of the present project is to improve the state of knowledge regarding the effects of cracking on pavement performance. This paper documents the methodology undertook to investigate the mechanical behavior of a cracked asphalt pavement structure subject to varying thermal and dynamic load cycles. The pavement structure was built in the indoor test track pit of Université Laval in which an idealized transverse crack has been created in the HMA layer. Cyclic loading of the structure was accomplished using the Accelerated Transportation Loading System (ATLAS). The mechanical response of the pavement structure is monitored using different types of strain gauges (H-gauges, optic fiber gauges) and pressure cells embedded at precise locations in the different pavement layers with a particular interest for the area around the crack. Strain and stress levels measured around the idealized transverse crack are compared with those of a crack-free area. Keywords: Hot mix asphalt Instrumentation


Thermal cracking
Accelerated testing


1 Introduction Low temperature cracking is one of the most serious pavement distress in cold regions, as it creates a discontinuity in the pavement structure and fosters the manifestation of other types of deteriorations (spalling, raveling, edge cracking, crack deflection, potholes, increase in pavement roughness, frost heave, …), resulting in progressive deterioration of the hot mix asphalt (HMA) layer. This project is a collaboration between the industrial research chair on the interaction of heavy loads/climate/pavement of Université Laval (Québec, Canada) and the pavements and © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 51–60, 2020. https://doi.org/10.1007/978-3-030-55236-7_6

52

Y. Berraha et al.

bituminous materials laboratory of École de Technologie Supérieure (ÉTS) (Montréal, Canada) where the main objective is to quantify the pavement damage evolution near a transverse crack compared with a reference section, and under several loading and temperature conditions. The main focus of this paper is on the methodology undertook during this project.

2 Experiment Description 2.1

ATLAS

Tests were conducted at Université Laval’s accelerated pavement facility. The indoor test track is a 6  2  2 m3 (length by depth by width) pit and the loading system can be placed above the track as shown in Fig. 1-b. The pavement structure was built in November 2018 and instrumented in December 2018. The Accelerated Transportation Loading System (ATLAS) allows thermal insulation of the pavement structure using reinforced polyester panels placed on all sides of its frame. This also allows control of the pavement structure’s temperature via a cooling system that regulates the temperature of the confined air above the pavement surface, thus allowing to test the pavement structure at any desired temperature between −10 °C and +30 °C.

Fig. 1. ATLAS a) mounted on trailer, b) standing on the test track with insulation panels on its sides

ATLAS is mounted with a loading system that can be equipped with various load configurations. In our case, a half-axle was equipped which is tracked back and forth in a straight line along the track at a constant speed of 2.0 m/s. It is also allowed to wander laterally ±10 cm around the central axis of the track, following a normal distribution. The half-axle is composed of a dual tire 12.00R20 with a contact surface of 0.045 m2 per wheel. The applied load and inflation pressure are respectively 50 kN and 700 kPa. During this experimental campaign, the cyclic loading is only applied to the pavement in a single direction. On its way back the wheels are not in contact with the pavement surface, meaning a unidirectional cyclic loading. In North America, pavement engineers generally use the concept of equivalent single-axle load (ESAL) to measure the effects of axle loads on the pavement. ESAL gives a measure of pavement damage caused by the passage of a standard axle load of

Laboratory Investigation of Cracked Asphalt Pavement Structure

53

80 kN (18000 lbs or 8165 kg) carried by a single axle with dual tires (Hutchison et al. 1987). The ESAL values for other axle configurations and load levels are expressed relative to the standard 80 kN via a load equivalency factor. The American Association of State Highway and Transportation Officials (AASHTO) provides load-equivalency factors or ESAL factors, which vary sharply with the applied load, following roughly a fourth-power law (AASHTO 1986). Thus, the load-equivalence factor for a 50 kN halfaxle is about 2.44 ((250)4/804). 2.2

Methodology

2.2.1 Pavement Structure and Material Properties The pavement structure is the same all along the track and consists of a 100 mm of HMA (GB20), a 250 mm granular base course 0/20 mm (MG20), a 450 mm granular subbase course 0/112 mm (MG112) and 950 mm of compacted subgrade soil (Fig. 2). Under testing, the ground water level was maintained at 0.5 m from the bottom of the pit (z = 1.5 m). During the compaction phase, each structural layer (subgrade, MG20, and MG112) was subdivided into sublayers of about 150 mm for homogeneous compaction. An evaluation of the compaction was performed after the placement of each of these layers using a lightweight deflectometer (LWD). Nuclear gauge density tests were also carried out, giving a measure of the dry (qd ) and wet (qh ) density as well as moisture content of each layer. The LWD modulus (ELWD ) obtained and the results of the nuclear gauge tests are reported in Table 1.

Table 1. Measured values of LWD modulus, wet and dry density, and moisture content values measured after the compaction of each layer Layer ELWD (MPa) qh (g/cm3) MG20 122 2.210 MG112 86 1.940 Subgrade 139 2.050

qd (g/cm3) 2.160 1.900 1.970

w (%) 2.4 3.0 3.9

As shown on the plan view of Fig. 2, the 6 m long pit is divided into two main areas of concern. The first is the reference area which corresponds to the pavement structure as-built. The second is the cracked area where an idealized transverse crack was created using a circular saw. The saw blade used has a thickness of 5 mm which produced a straight and thin transverse crack on the entire HMA layer.

54

Y. Berraha et al.

Fig. 2. Plan and section views of the pavement structure.

2.2.2 Thermomechanical Properties and Fatigue Pavement Design To evaluate the fatigue resistance of the pavement structure, complex modulus and fatigue performance of the HMA were considered. The HMA layer is composed of a 0/20 mm aggregate mixes blended with 3.65% of a PG 58S-28 bitumen grade. The HMA was compacted at 95% ± 2%. Table 2 gives 2S2P1D parameters used to model the complex modulus of a typical HMA GB20 produced in Québec. Table 3 summarizes fatigue properties of the HMA GB20 used for pavement design. Based on the French design manual for pavement structures (SETRA-LCPC, 1997), the constructed test section has a fatigue resistance which corresponds to 655,000 ESAL at 15 °C when considering risk of 50% and a loading frequency of 1 Hz. Table 2. 2S2P1D parameters and WLD (Williams, Landel and Ferry) constants set at 4.1 °C in accordance with typical modulus data for a GB20 mix as produced in Quebec 2S2P1D parameters WLF constants h d sE b C1 C2 T0 (°C) E∞ (MPa) E0 (MPa) k 25 41000 0.17 0.56 2.25 1.5 400 25.89 178.84 4.1

Table 3. Fatigue law parameters to express GB20 mix as produced in Quebec Risk consideration (50%) Parameter of the Fatigue law(a) a3 E(htest, ftest) htest ftest kc ks kr SN c Sh Nf1(htest, ftest)(b) a2 (MPa) (cm−1) (cm) (°C) (Hz) 4.50 2.25 15,843 10 10 1.3 1.0 1.0 0.30 0.02 1 4.78  10−13 (a)

2 Based on Wöhler’s concept: Nf;h ¼ Nf1 ðhtest ; f test Þ
Eðhtest ; f test Þa3
Eðh; f Þa3
ea 0   1=a 2 e6 ðhtest ; f test Þ ¼ Nf1 ðhtest ; f test Þ
106 ¼ 85  106 ¼ 85 lm=m

(b)

Laboratory Investigation of Cracked Asphalt Pavement Structure

55

2.2.3 Sensors As shown in Fig. 2, vertical strains (ezz ) and vertical stress (rzz ) were collected in each unbound layer. For HMA layer, four H-beam strain gauges (TML-1, 2, 3 and 5) allowed the measurement of longitudinal strains at the bottom of the layer (z = 100 mm) under the wheel track (Fig. 3). TML-4 was embedded in-between the wheel track in the reference area.

Fig. 3. Plan view of the sensors’ layout at the bottom of HMA layer (z = 100 mm)

A retrofit technique was also used to measure longitudinal and transverse strains (resp. exx and eyy ) in the upper (z = 25 mm) and lower (z = 100 mm) parts of the HMA layer. It consists of an asphalt concrete core specially trimmed for the installation of optic fiber gauges (Doré et al. 2007). This method uses polymeric proof bodies with mechanical (E) and thermal (coefficient of thermal expansion) compatibility with asphalt concrete. Three asphalt concrete cores with optical strain gauges (Core A, B, and C) were made and sealed back in place using low viscosity epoxy (Fig. 4). Two of them are located on each side of the idealized crack (Core A and B) and the third (Core C) is in the reference area as shown in Fig. 3.

Fig. 4. Instrumented asphalt concrete core: a) before being placed back in the HMA layer b) in place before being sealed with low viscosity epoxy

56

Y. Berraha et al.

Another experimental setup was created in order to measure the vertical displacement (Lz ) of both crack lips under cyclic loading. The setup is placed in a hole drilled in between wheel tracks (Fig. 5b) and consists of two linear potentiometers mounted on a frame that is fixed at the base of the test pit and independent of the pavement structure. A steel angle is attached to each crack lip using two parts resin epoxy glue and each potentiometer lays on one of the steel angles (Fig. 5a). Under moving load, the vertical movement of a crack lip results in the displacement of the corresponding steel angle which in turn causes the linear potentiometer to extend or contract.

Fig. 5. Experimental setup for the measure of crack lips vertical displacement (Lz ): a) diagram of the general principle, b) plan view photo of the setup in place

Finally, a line of thermocouples runs from the bottom to the top of the pavement structure with temperatures recorded every 30 min at z = 20, 120, 320, 420, 520, 620, 720, 820, 920, 1020, 1120, 1320, 1520, 1720 and 1920 mm. For measuring volumetric water content in the unbound layers, 3 different holes allowed the use of a PR2 profile probe (Fig. 2) with measures periodically at different depths (z = 300, 400, 500, 600, 800 and 1100 mm). 2.2.4 Experimental Protocol The experimental protocol was created in order for the pavement structure to undergo a year of dynamic loading with a precise sequence of air temperature. Figure 6 gives a representation of the experimental protocol with the evolution of the number of ESAL and measured HMA temperature over 365 days. Figure 6 gives a representation of the evolution of the temperature in the HMA layer as well as the number of ESAL imposed on the pavement structure for the fictive year of solicitation. As shown in Fig. 6, three freeze-thaw cycles were achieved. During two of these cycles, a local saturation was performed through the idealized crack in order to emulate the infiltration of rainwater and evaluate the effect on the pavement’s behavior. This local infiltration is achieved using a flexible porous drip hose placed inside the crack. Knowing the porosity of the MG20 layer, the amount of water injected in the crack was estimated in order to reach a high degree of saturation in

Laboratory Investigation of Cracked Asphalt Pavement Structure

57

a delimited region (half-cylinder (radius = 150 mm)) of the MG20 layer below the crack. In total, 45 L and 30 L were poured all along the crack during respectively the second and the third thawing cycles (Fig. 2).

Fig. 6. Representation of the experimental protocol followed during the project

3 Preliminary Results 3.1

TML Response Before the Creation of the Idealized Transverse Crack

During the experimental protocol (Fig. 6), after the stabilization of the HMA layer’s temperature, a set of cyclic loading passes were applied to the pavement structure according to the predefined sequence. Following the cyclic loading, a dynamic response test is carried out which corresponds to very few additional passes with the acquisition of all sensors’ response to the loading. Figure 7 represents the longitudinal strain (exx ) at the base of HMA layer before the creation of the idealized crack. Figure 7a defines maximum tensile strain (et ), maximum compressive strain before and after the passage of the load over the strain gauge (resp. ecb and eca ), and strain pulse durations Tpk and Tload . Stain pulse frequencies are defined as Fpk ¼ 1=Tpk and Fload ¼ 1=Tload . Figure 7b shows the evolution of exx recorded by all strain gauges before the creation of the idealized transverse crack and for a temperature of the HMA layer of 16 °C. Before the creation of the idealized transverse crack, the amplitudes of exx are similar in both tested areas and the signal shows typical behavior.

58

Y. Berraha et al.

Fig. 7. Longitudinal strain (exx ) at the base of HMA layer before the creation of the idealized transverse crack, (a) typical response, (b) TML response

3.2

Tensile Strain at the Base of the HMA Layer After the Creation of the Idealized Transverse Crack

Figure 8a and 8b show the evolution of the longitudinal strain (exx ) recorded at the bottom of the HMA layer (z = 100 mm) by respectively the five TML gauges and the optic fiber gauges just after the creation of the idealize transverse crack for a temperature of HMA layer of 16 °C. Values of et , ecb , eca , Fpk and Fload measured from Fig. 7b and Fig. 8a are reported in Table 4. Before the creation of the idealized transverse crack, the amplitude of longitudinal strain is similar in both the cracked area (TML 1,2 and 3) and the reference area (TML 4 and 5). After the creation of the idealized transverse crack, we see a significant drop of et for TML-2 which is the closest to the crack. We also notice an increase in ecb for TML-2 and TML-3.

Fig. 8. Longitudinal strain (exx ) at the bottom of the HMA layer just after the creation of the idealized crack, (a) TML signal, (b) optic fiber signal

Laboratory Investigation of Cracked Asphalt Pavement Structure

59

Table 4. Values of et , ecb , eca and Fpk Fload before and after the creation of the idealized transverse crack AFTER (4,400 ESAL) THMA = 16 °C BEFORE (0 ESAL) ecb eca Fpk Fload et ecb eca Fpk Fload et µm/m µm/m µm/m Hz Hz µm/m µm/m µm/m Hz Hz TML-1 TML-2 TML-3 TML-5 TML-4

351 375 350 386 369

−93 −111 −105 −110 −133

−50 −44 −31 −57 −53

2.9 2.7 2.9 2.8 2.7

0.9 0.9 1.0 1.0 1.0

307 170 352 370 370

−75 −5 −168 32 −150 −15 −110 −51 −131 −53

4.2 1.9 2.7 2.7 2.7

1.4 2.9 1.5 1.0 1.0

Considering a multi-layer linear elastic system (MEL) (Burmister 1943), the theoretical et value calculated at 16 °C and 1 Hz (EHMA = 7,055 MPa) is nearby 222 µm/m for a load of 721 kPa apply on circular area of 340 cm2. The calculated et value is clearly lower than TML measurements. Moreover, for the TML-2, the typical shape of the contraction signal observed before and after the passage of the wheel over the strain gauge is no more visible. Furthermore, results in Fig. 8b show that the et values from optic fiber gauges located in the reference area (core C) is really close to theoretical value with a value of 200 lm/m. 3.3

Influence of the Temperature of the HMA Layer on the Tensile Strain

Figure 9a and 9b show the evolution of the longitudinal strain (exx ) after respectively 41 800 (THMA = −7.5 °C) and 132 000 ESAL (THMA = 20 °C). Figure 9 illustrates the influence of temperature of the HMA layer on the amplitude of longitudinal strain (0 °C and 20 °C). As seen in Fig. 9a, at −7.5 °C, exx in the HMA layer are non-existent due to the increase of HMA’s modulus at low temperatures. Figure 9b represents exx measured by the five TML after the first year of loading.

Fig. 9. exx at the bottom of the HMA layer after (a) 41 800 ESAL and (b) 132 000 ESAL

60

Y. Berraha et al.

4 Conclusion This paper focuses on the methodology followed during this accelerated pavement testing program. The paper presents the experimental protocol and different sensors placed in the pavement structure in order to investigate the pavement behavior around an idealized transverse crack. This procedure allowed to monitor deflection, strain and stress developed in a cracked asphalt pavement structure. Preliminary results of the longitudinal strains measured near the idealized transverse crack shows that the amplitude of et is lower that what is measured in the reference section. Further analysis of results from other sensors and tests carried out during the experiment (FWD, rutting, …) are required in order to completely quantify the evolution of damage around an idealized transverse crack.

References Burmister, D.M.: The theory of stresses and displacements in layered systems and applications of the design of airport runways. In: Proceedings of the Highway Research Board, vol. 23, pp. 126–148 (1943) AASHTO: Guide for Design of Pavement Structures, Washington, D.C (1986) Hutchison, B.G., Haas, R.C.G., Meyer, P., Hadipour, K., Papagiannakis, T.: Equivalencies of different axle load groups. In: Proceedings of the 2nd North American Conference on Managing Pavements, Toronto, Ontario, Canada (1987) Doré, G., Duplain, G., Pierre, P.: Monitoring mechanical response of in-service pavements using retrofitted fiber optic sensors. In: Proceedings of the International Conference on the Advanced Characterization of Pavement and Soil Engineering Materials, Athens, Greece, pp. 883–891 (2007)

Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement Using Accelerated Pavement Test Rongji Cao1, Chunying Wu1, Bingfeng Zheng1, Zhenglong Lv1(&), Yi Huang1, and Ying Gao2 1

2

JSTI Group, Nanjing 211112, China {crj,lzl31}@jsti.com School of Transportation, Southeast University, Nanjing 210096, China [email protected]

Abstract. A test section of asphalt pavement with typical semi-rigid base was constructed to analyze its dynamic response under accelerated pavement tester. The strain behavior inside asphalt layers was measured by installed pavement sensors. A tension-compression process was observed in vertical strain time history curves over one cycle, while longitudinal time history strain curves exhibited an alternating compression-tension-compression course. The characterization of strain responses under different speeds was studied in this paper and it was found that strain amplitude and the duration of high-strain zone increased at both vertical and longitudinal direction of asphalt layers when loading velocity reduced. Moreover, based on test results, the predicted fatigue life of Sup-25 course suggested that lower loading speed resulted in shorter fatigue life. All in all, the increase of vertical and longitudinal strain in circumstances of low velocity more easily caused pavement distresses, e.g., rutting and fatigue cracking. Keywords: Strain response
Accelerated Pavement Test
Low speed
Fatigue life

1 Introduction Researchers widely accept that Accelerated Pavement Testing (APT) is an effective method to study the dynamic response and performance of asphalt pavement, although the experimental cost is expensive. Mobile Load Simulator 66 (MLS66) is a new generation of full-scale APT device, which can simulate the standard axle and load 6000 times within an hour. Additionally, strain/stress behavior of pavement structure is capable of being accurately measured through MLS66 conducting on instrumented experimental section, but this is difficult for indoor laboratory or numerical simulation. Some studies related to dynamic response of semi-rigid base asphalt pavement have been conducted. The performance of semi-rigid base asphalt pavement over fine-sand subgrade was evaluated via APT, and additionally vertical strain at the top of subgrade was measured and proven that the designed pavement structure can meet the requirement of heavy traffic (Wu et al. 2015). Dong et al. (2009) installed three© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 61–69, 2020. https://doi.org/10.1007/978-3-030-55236-7_7

62

R. Cao et al.

direction strain sensors inside asphalt pavement and compared the characteristics of dynamic response under different loading conditions. Nevertheless, there are few investigations regarding the influence of low loading speed on dynamic response of semi-rigid base asphalt pavement. According to pavement surface distress survey, pavement distresses, e.g., rutting, fatigue cracking and swelling, are normally more severe in low-speed sections, like junctions. In this paper, a test section was constructed with buried pavement sensors, and then dynamic response of pavement was analyzed under different loading speeds using MLS66 Tester.

2 Test Section and Instrumentation The test section was 50 m in length and 6 m in width, as shown in Fig. 1(a). It was a typical asphalt pavement structure for heavy-load traffic road in China. The asphalt layers were composed of 4 cm SMA-13, 6 cm Sup-20 and 8 cm Sup-25, with asphalt content of 5.8%, 4.3% and 3.9% respectively. SBS modified asphalt was applied in SMA-13 and Sup-20, while 70# asphalt was applied in Sup-25. The base was 36 cm cement stabilized macadam with 4.2% cement content. The subbase consisted of lowdosage cement stabilized macadam with 3% cement content. The design gradations of asphalt mixtures were illustrated in Fig. 1(b), and design gradations of base and subbase were listed in Table 1.

4cm SMA-13SBS 6cm SUP-20 (SBS)

8cm SUP-25 (70#)

36cm cement stabilized macadam 20cm low-dosage cement stabilized macadam

100

Passing Percentage / %

(a)

80

b

60 40

SMA-13 SUP-20 SUP-25

20 0 0.075 0.15 0.3

0.6 1.18 2.36 4.75 9.5 13.2 16.0 19.0 26.5 31.5 Sieve Size / mm

Fig. 1. Experimental pavement: a) Pavement structure, b) Design gradation of asphalt mixtures

Table 1. Design aggregate gradation of base and subbase Sieve size/mm 31.5 26.5 19 9.5 4.75 2.36 0.6 0.075 Passing percentage/% Base 100.0 96.2 78.3 51.3 31.7 20.6 8.7 2.9 Subbase 100.0 99.0 79.5 50.7 33.6 21.8 10.8 4.8

The strain/stress at the bottom of asphalt layers is regarded as one of the most significant indicators required in the prediction of asphalt pavement performance. In this study, longitudinal strain sensors (KM-100HAS), and vertical strain sensors (KM50F) were selected. The layout of embedded sensors is presented in Fig. 2. Two

Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement

63

longitudinal strain sensors and two vertical strain sensors were installed directly under twin tires at the bottom of Sup-20 binder course and Sup25 base course. Nine thermocouples (PT100) were embedded along the pavement depth (i.e., 7, 10, 14, 18, 27, 36, 45, 54 and 64 cm respectively) to monitor inner temperature of pavement. Furthermore, two pressure cells (OL141D500) were installed at the top of base and subgrade to monitor compressive stress.

(b)

4cm SMA-13

(a) 6cm SUP-20 8cm SUP-25

wheel path

0.5m 0.5m at the bottom of middle asphalt layer

36cm cement treated macadam base

20cm low-dosage cement treated macadam subbase

wheel path

at the bottom of bottom asphalt layer

longitudinal strain sensor

vertical strain sensor

subgrade

temperature sensor

pressure cell

Fig. 2. Layout of sensors: a) Plan view at the bottom of Sup-20 course and Sup-25 course, b) Right view of experimental section

3 MLS66 Setup and Strain Analysis Method After the test section was paved, MLS66, a full-scale accelerated device, was used to run along the wheel path. The MLS66 was equipped with six 305/70R22.5 twin tires, and its load was set to 50 kN with 0.7 MPa inflation pressure, which corresponded to a standard single axle of 100 kN. The test section was loaded at 10 km/h, 15 km/h and 22 km/h respectively to study the influence of low speed on dynamic response of pavement structure. All sensors were connected to the data acquisition system DHDAS. Considering that there would be a large amount of data, the sampling frequency of strain/stress sensors was 200 Hz, while that of thermocouples was 2 Hz. All these data were stored automatically once test started. Previous studies revealed that temperature field of asphalt pavement significantly influenced dynamic response (Xiao 2019 and Arraigada et al. 2014). Temperature inside test pavement was thus monitored simultaneously. To quantify the strain curves, some critical points were defined in reference to documented literatures (Han et al. 2019 and Wu et al. 2015). Figure 3(a) gives an example of vertical strain curve within one cycle. As seen, vertical strain curve exhibits a minor tensile-strain peak closely followed by a larger compressive-strain peak and eventually returns to zero baseline. The absolute value of highest compressive strain (C) was approximately 4 times larger than that of the highest tensile strain (B). Typical

64

R. Cao et al.

longitudinal strain curve was plotted in Fig. 3(b). It indicates that longitudinal strain curve consists of a compressive strain state when wheel approached, followed by a tensile strain state when wheel is directly upon sensors, and lastly a compressive strain state when wheel leaves. Unlike vertical strain curve, values of tensile strain and compressive strain are in the same order of magnitude. Only high-strain phases (B-C and C-D) were discussed in paper. The definitions of critical points B, C and D of vertical and longitudinal strain curves were listed in Table 2. In order to analyze the strain response under low speed, the variations of strain (Δe) and time (Δt) need to be summarized in B-C and C-D phases.

20

15

tension

(a

B

(b) C

10

wheel approaching

0 E

D

-20

compression

-40

wheel leaving

wheel approaching

-60

wheel leaving

5 Strain / με

Strain / με

A

tension

0 E

A -5

compression D

C

-10 B

-80 0

0.1

0.2

0.3

0.4

0.5

0.6

-15 0

Time / s

0.1

0.2

0.3

0.4

0.5

0.6

Time / s

Fig. 3. Example of strain curves: a) Vertical strain, b) Longitudinal strain

Table 2. Definitions of critical points of vertical and longitudinal strain Critical points B

C D

Definition Vertical strain curve The maximum tensile strain in the stage of wheel approaching The maximum compressive strain within one cycle The first derivative of strain curve is positive, and meanwhile its second derivative is nearly zero when wheel moves off

Longitudinal strain curve The maximum compressive strain in the stage of wheel approaching The maximum tensile strain within one cycle The maximum compressive strain when wheel moves off

4 Results and Analysis The temperature filed was measured simultaneously since asphalt mixtures had strong temperature dependency. Results suggested that temperature remained approximately identical when MLS66 loaded at different velocities. Strain curves of different loading speeds at the bottom of Sup-20 course were presented in Fig. 4. It was noted that strain

Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement

65

curves exhibited obvious periodic variation. As expected, the duration of one cycle increased from 0.6 s to 1.32 s when loading velocity reduced from 22 km/h to 10 km/h.

40

vertical

longitudinal

40

vertical

(a)

longitudinal tension

tension Strain / με

Strain / με

0 compression duration=1.32s

-40

-80 0

0.5

40

1

1.5 Time / s

2

2.5

vertical

0 compression -40

duration =0.88s

-80

3

longitudinal

(b

0

0.5

1

1.5 Time / s

2

2.5

3

(c)

Strain / με

tension 0 compression -40

duration =0.6s

-80 0

0.5

1

1.5 Time / s

2

2.5

3

Fig. 4. Strain curves of different loading velocities at the bottom of Sup-20 course: a) Loading speed = 10 km/h, b) Loading speed = 15 km/h, c) Loading speed = 22 km/h

4.1

Vertical Strain

Twenty cycles of strain curves were chosen for each loading velocity, and their variations of vertical strain (ΔeB-C and ΔeC-D) versus time (ΔtB-C and ΔtC-D) were depicted in Fig. 5. Table 3 listed the average values of Δe and Δt of vertical strain. At the bottom of Sup-20 course, the absolute value of ΔeB-C of 22 km/h was 60.80 le, and that of 15 km/h and 10 km/h increased by 14% and 31% respectively. Besides, when wheel speed decreased from 22 km/h to 10 km/h, ΔtB-C increased from 0.051 s to 0.108 s. This indicated that low velocity influenced not only the amplitude of Δe but the corresponding duration Δt. Furthermore, the vertical response observed in Sup-25 course was more distinct than that of Sup-20 course. It seems that the accumulative vertical strain of slower speed more easily contributed to pavement rutting.

66

R. Cao et al.

Fig. 5. Variations of vertical strain (Δe) versus variations of time (Δt) over twenty cycles at the bottom of Sup-20 course: a) B-C phase, b) C-D phase Table 3. Average results of variations of vertical strain and time over twenty cycles Position Bottom of Sup20 course Bottom of Sup25 course

4.2

Speed/km/h |Δe|/le B-C 10 79.59 ± 15 69.31 ± 22 60.80 ± 10 83.59 ± 15 81.09 ± 22 78.70 ±

4.71 3.75 3.45 3.17 2.96 2.83

C-D 54.68 50.09 47.26 68.02 69.14 69.36

± ± ± ± ± ±

3.79 3.01 3.00 2.80 2.53 2.64

Δt/s B-C 0.108 0.076 0.051 0.157 0.113 0.084

± ± ± ± ± ±

0.005 0.005 0.004 0.007 0.007 0.005

C-D 0.125 0.100 0.088 0.164 0.135 0.113

± ± ± ± ± ±

0.017 0.013 0.014 0.015 0.014 0.013

Longitudinal Strain

Results obtained from longitudinal strain sensors were given in Fig. 6 and Table 4. For instance with the longitudinal strain at the bottom of Sup-20 course, the absolute value of ΔeB-C of 22 km/h was 19.59 le, and that of 15 km/h and 10 km/h increased by 10.3% and 21.7% respectively. In phase C-D, ΔeC-D of 10 km/h was 17.25 le, an increase of 14.3% in comparison with that of 22 km/h. Besides, when wheel speed decreased from 22 km/h to 10 km/h, ΔtB-C increased from 0.037 s to 0.081 s. Similarly, the strain variation at the bottom of Sup-25 course exhibited the same tendency.

Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement

67

Fig. 6. Variations of longitudinal strain (Δe) versus the variations of time (Δt) over twenty cycles at the bottom of Sup-20 course: a) B-C phase, b) C-D phase Table 4. Average results of variations of longitudinal strain and time over twenty cycles Position Bottom of Sup20 course Bottom of Sup25 course

Speed/km/h |Δe|/le B-C 10 23.85 ± 15 21.60 ± 22 19.59 ± 10 7.07 ± 15 7.05 ± 22 6.44 ±

0.67 0.59 0.59 0.38 0.36 0.30

C-D 17.25 16.22 15.09 5.13 5.31 4.95

± ± ± ± ± ±

0.48 0.43 0.42 0.35 0.34 0.29

Δt/s B-C 0.081 0.056 0.037 0.220 0.142 0.106

± ± ± ± ± ±

0.005 0.003 0.003 0.037 0.018 0.016

C-D 0.087 0.061 0.043 0.238 0.190 0.166

± ± ± ± ± ±

0.007 0.006 0.004 0.037 0.021 0.025

The alternating tensile/compressive strain in longitudinal direction easily causes fatigue cracking, which is one major pavement distress. Hence, the fatigue life is predicted herein using Eq. 1 (Assogba et al. 2019). Nf ¼ A  0:00432  104:84ðVFA0:69Þ  e3:291  jE j0:854 t

ð1Þ

where Nf is fatigue life, A = 18.4 is the adjustment from laboratory to field, VFA = 0.672 is void filled with asphalt of Sup-25, et is measured maximum tensile strain at the bottom of Sup-25 course, and |E*| is dynamic modulus. To account for the influence of different loading velocities on loading frequency and furthermore the modulus of asphalt mixture, dynamic modulus of asphalt mixture was measured under different temperatures (4.4, 21.1, 37.8 and 54.4 °C) and different frequency (from 0.1 to 25 Hz). Loading frequency can be converted from loading velocity according to Eq. 2 (Qian et al. 2015). The master curve of dynamic modulus of Sup-25 was constructed using Eq. 3 at 20 °C reference temperature (Rahman and Tarefder 2016). f ¼

v 12  d

ð2Þ

68

R. Cao et al.

where f is loading frequency, v is loading speed, and d = 0.213 m is equivalent circle diameter of tire touching ground. logðjE  jÞ ¼ d þ

a 1 þ eb þ c logðfrÞ

ð3Þ

where |E*| is dynamic modulus, fr is reduced frequency, d is minimum modulus, d+a is maximum modulus, and both b and c are shape parameters. 1E+5 R2=0.997 Se/Sy=0.046

|E*| / MPa

1E+4

1E+3 master curve fit 4.4 deg C 21.1 deg C 37.8 deg C 54.4 deg C

1E+2

1E+1 1E-6

1E-4

1E-2

1E+0

1E+2

1E+4

1E+6

fr / Hz

Fig. 7. Master curve of dynamic modulus of Sup-25 mixture

The master curve of dynamic modulus was illustrated in Fig. 7, and fitting parameters d, a, b and c were −0.9677, 5.5495, −1.6506 and −0.3096 respectively. The predicted fatigue life (Table 5) indicated that when loading speed reduced from 22 to 10 km/h, the number of loading repetition to prevent fatigue cracking decreased from 1.29  1013 to 7.51  1012, which suggested that slower speed is more likely to cause fatigue cracking. Table 5. Predicted fatigue life of asphalt layer under different low velocities Speed/km/h 10 15 22

et/le 6.33 5.52 5.14

f/Hz 1.09 1.63 2.39

|E*|/MPa 4976 5450 5924

Nf 7.51  1012 1.09  1013 1.29  1013

5 Conclusion A test section of semi-rigid base asphalt pavement was built with pre-embedded sensors to study dynamic response via APT device MLS66. Typical vertical and longitudinal strain curves were obtained. Vertical strain curves exhibited a tension-compression process, and longitudinal strain curves showed an alternating compression-tensioncompression process. The pavement was loaded at 10, 15 and 22 km/h respectively, and

Analysis of Dynamic Response for Semi-rigid Base Asphalt Pavement

69

results indicated that ΔeB-C of 10 km/h was 79.59 le (an increase of 31% compared with that of 22 km/h) in vertical direction. Furthermore, the variation of longitudinal strain and their corresponding duration increased when loading speed decreased. The predicted fatigue life of Sup-25 course reduced from 1.29  1013 to 7.51  1012 when loading speed reduced from 22 km/h to 10 km/h. Consequently, low speed leads to larger strain and longer duration of high-strain zone, and further this has a negative effect on the development of rutting and fatigue cracking of asphalt pavement. Acknowledgement. Authors appreciate the support of the scientific research funding from Transportation Science and Technology Project of Jiangsu Province (2016Y01 and 2018Y27).

References Xiao, C.: Evaluation of measured strain responses to in situ vehicular loading for typical asphalt pavements. Eng. Trans. 67(1), 55–73 (2019). https://doi.org/10.24423/EngTrans.970. 20190214 Dong, Z., Liu, H., Tan, Y., Chen, F., Wang, B.: Field measurement of three-direction strain response of asphalt pavement. J. South China Univ. Technol. (Nat. Sci. Ed.) 37(7), 46–51 (2009) Arraigada, M., Pugliessi, A., Partl, M.N., Martinez, F.: Effect of full-size and down-scaled accelerated traffic loading on pavement behavior. Mater. Struct. 47(8), 1409–1424 (2014). https://doi.org/10.1617/s11527-014-0319-2 Qian, Z., Yang, Y., Chen, T., Li, Z.: Dynamic response of asphalt pavement under moving loads with low and variable speed. J. Central South Univ. (Sci. Technol.) 46(3), 1140–1146 (2015). https://doi.org/10.11817/j.issn.1672-7207.2015.03.047 Wu, J., Ye, F., Hugo, F., Wu, Y.: Strain response of a semi-rigid base asphalt pavement based on heavy-load full-scale accelerated pavement testing with fibre Bragg grating sensors. Road Mater. Pavement 16(2), 316–333 (2015). https://doi.org/10.1080/14680629.2014.995211 Rahman, A.S.M.A., Tarefder, R.A.: Dynamic modulus and phase angle of warm-mix versus hotmix asphalt concrete. Constr. Build. Mater. 126, 434–441 (2016). https://doi.org/10.1016/j. conbuildmat.2016.09.068 Assogba, O.C., Tan, Y., Sun, Z., Lushinga, N., Bin, Z.: Effect of vehicle speed and overload on dynamic response of semi–rigid base asphalt pavement. Road Mater. Pavement 8, 1–31 (2019). https://doi.org/10.1080/14680629.2019.1614970 Han, Z., Sha, A., Hu, L., Wu, R., Li, H.: Full-scale investigation on the traffic load influence zone and its dimension for HMA layer in inverted pavement. Constr. Build. Mater. 219, 19–30 (2019). https://doi.org/10.1016/j.conbuildmat.2019.05.110

Performance of Unbound Pavement Materials in Changing Moisture Conditions Marit Fladvad1,2(&) and Sigurdur Erlingsson3,4 1

Department of Geoscience and Petroleum, NTNU – Norwegian University of Science and Technology, Trondheim, Norway 2 Norwegian Public Roads Administration, Trondheim, Norway [email protected] 3 Pavement Technology, Swedish National Road and Transport Research Institute (VTI), Linköping, Sweden [email protected] 4 Faculty of Civil and Environmental Engineering, University of Iceland, Reykjavík, Iceland

Abstract. Expected climate changes will in many areas represent a shift towards increased precipitation and more intense rainfall events. This may lead to increased moisture within road structures and possible overloading of road drainage systems. Pavement design methods must therefore be able to predict the behaviour of pavement materials at increased moisture levels. An instrumented accelerated pavement test (APT) has been conducted on two thin flexible pavement structures with coarse-grained unbound base course and subbase materials using a heavy vehicle simulator (HVS). The two pavement structures were identical except for the grain size distribution of the subbase material, where one had a dense 0/90 mm curve with a controlled fines content, and the other had an open-graded 22/90 mm curve. The APT was conducted using constant dual wheel loading, and three different groundwater levels were induced in order to change the moisture content in the structures. The HVS was stopped regularly for carrying out response measurements from the instrumentation. The analysis is focussed on the response and the performance of the unbound aggregate layers to varying moisture levels in the pavement structure. Keywords: Accelerated pavement test content
Groundwater table


Heavy vehicle simulator
Moisture

1 Introduction Environmental conditions are important inputs to pavement design, combined with traffic load, material choice and layer thicknesses. Pavement design has traditionally been done using empirical methods, based on long-term experience with similar materials and conditions (ARA Inc. 2004). When climate changes cause a shift towards increased precipitation and more intense rainfall events, empirical methods can no longer be used to predict pavement performance. A transition from empirical to

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 70–79, 2020. https://doi.org/10.1007/978-3-030-55236-7_8

Performance of Unbound Pavement Materials in Changing Moisture Conditions

71

mechanistic design is needed, and to achieve this, we need to be able to model the behaviour of the pavement structures under all conditions (Erlingsson 2007). In the Nordic countries, most roads are designed using flexible pavements, with relatively thin hot mix asphalt (HMA) layers above thicker unbound base and subbase providing a substantial part of the bearing capacity. This practice is partly due to the availability of high-quality aggregate resources, making unbound aggregates an affordable solution. In Norway, local aggregate resources originating from tunnels and road cuts are often utilized in construction projects. Aggregates can be produced and used on-site, reducing transportation costs for raw materials. Due to space and time restrictions at the construction sites, such production requires a simple setup. These restrictions, combined with experience with frost heave problems caused by excess fines, have resulted in a practice where all fine material is sorted out from the large-size aggregate used in the subbase layer (Aarstad et al. 2019; Aksnes et al. 2013). In Sweden, on the other hand, aggregate resources are normally transported in from quarries where the production methods allow for full quality assessment of the products. Here, large-size aggregate materials with controlled fines content are used in the subbase layer. Water and moisture have a large influence on the performance of pavement materials (Dawson 2009). As the moisture content increases, the friction between aggregate particles becomes lower, and the resistance to differential particle deformation is reduced, leading to a reduced resilient modulus of unbound aggregates (ARA Inc. 2004; Erlingsson 2010; Lekarp and Dawson 1998). The main objective of the current research is to examine the effect of gradation on the performance of subbase materials when tested at three different groundwater levels. Two subbase materials of equal geological composition are tested, one material containing fines, while all particles 3 m below the surface. Horizontal distance between sensors not to scale

ASGs measure the longitudinal and transversal strain at the bottom of the AC layers. SPCs measuring vertical stress are distributed at the top, middle and bottom of the subbase. eMUs measure vertical strain at 7 different levels; over the bound base layer, unbound base layer, three levels in the subbase and two levels in the subgrade, the lowest reaching to approximately 30 cm below the formation level. Moisture content was measured at four different levels; in the subgrade approximately 15 cm below the formation level, at two levels in the subbase layer, and in the middle of the base layer. The pavement response was measured at regular intervals during the test. Measurements using falling weight deflectometer (FWD) was conducted in each phase. FWD measurements were conducted before the accelerated traffic started and at the end of phase w2 and w3. 2.3

Groundwater Table

The APT was divided into three phases, each with a separate depth of the GWT. In phase w1, GWT was located at great depth, >3 m below the pavement surface. GWT was stable at this level for the first 550 000 load repetitions. For phase w2, GWT was raised to 30 cm below the formation level, corresponding to the depth of the drainage level of a pavement structure in operation. GWT was stable at this level for 368 000 load repetitions. For phase w3, GWT was raised further to a level of about 5 cm above the formation level. This level simulates a situation where the drainage system is overloaded and unable to keep the GWT at the designed drainage level. GWT was stable at this level for 286 000 load repetitions. No traffic load was applied while GWT was raised.

74

M. Fladvad and S. Erlingsson

3 Results 3.1

Water Content

Table 1 shows the volumetric water content for all three phases. The water content in the subgrade increases from 12.8% to 27.7%, corresponding to full saturation. The 0/90 mm subbase and overlying base show an increase in water content when GWT is introduced in the subgrade. For the open-graded structure, it is only at the end of phase w3 that an increase in water content is visible in the base course. Table 1. Volumetric water content in unbound base, subbase and subgrade at the time of FWD measurements 0/90 mm structure 22/90 mm structure w1 w2 w3 w1 w2 w3 Base course 7.7% 8.6% 10.0% 7.1% 7.2% 10.0% Upper subbase 6.1% 6.5% 6.7% 2.1% 2.2% 2.5% Lower subbase 8.6% 9.2% 10.2% 1.5% 1.6% 2.1% Subgrade 12.8% 26.6% 27.7% 12.8% 26.6% 27.7%

3.2

FWD Back-Calculation

FWD measurements were conducted at all three GWT levels. Table 2 shows the back-calculated stiffness moduli for each phase. As the climate chamber had to be removed in order to enable FWD measurements, the temperature could not be controlled at 10 °C for these measurements. To allow the comparison of deflection bowls in Fig. 2, the FWD results have been back-calculated with the resilient modulus for the AC layers adjusted corresponding to 10 °C. In the back-calculation procedure, the subbase was divided into three sublayers and the subgrade into two layers. The 22/90 mm structure has the highest deflection in phase w1, but is less affected by the raised GWT in phases w2 and w3, compared to the 0/90 mm structure. From the stiffness values, it appears that the difference in deflection between phase w1 and w2 in the 22/90 mm structure is only due to the decrease in subgrade stiffness. For the 0/90 mm structure, on the other hand, base and subbase stiffness is also affected by the introduction of a GWT in phase w2. Between phase w2 and w3, stiffness in all layers is affected. The reduced AC stiffness is not directly related to the increased GWT, but a result of pavement degradation due to the 1.2 million load repetitions.

Performance of Unbound Pavement Materials in Changing Moisture Conditions

75

Table 2. Back-calculated resilient moduli at 10 °C in the pavement structure at the three groundwater levels 0/90 mm structure 22/90 mm structure Thickness [mm] Stiffness modulus Thickness [mm] Stiffness modulus [MPa] [MPa] w1 w2 w3 w1 w2 w3 AC layers 104 6500 6500 5671 112 6500 6500 5671 Unbound base 101 250 225 203 121 250 250 203 Subbase 1 163 220 157 149 150 143 143 135 Subbase 2 150 220 143 129 150 143 143 135 Subbase 3 50 220 129 116 50 143 143 135 Subgrade 1 300 82 66 46 300 82 66 46 Subgrade 2 2132 82 53 46 2117 82 53 46

a) 0/90 mm structure

b) 22/90 mm structure

Fig. 2. Adjusted deflection bowl based on back-calculated stiffness parameters from FWD measurements with load 50 kN at 10 °C

3.3

Stress

Figure 3 shows the typical registration of an SPC. Four consecutive registrations as the wheel load pass over the sensor are shown after 377 000, 555 000 and 940 000 load repetitions. The effect of the change in moisture content in the three phases (w1, w2 and w3) is evident as the measured stress decreases significantly between the phases.

Fig. 3. Vertical stress signals registered at a sensor located in the middle of the subbase layer of the 0/90 mm structure after 377 000 (w1), 555 000 (w2) and 940 000 (w3) load repetitions

76

M. Fladvad and S. Erlingsson

Figure 4 shows the development of measured stress throughout the APT. A postcompaction effect is seen in phase w1, as the stress increases gradually. The stress decreases after each GWT raise, before it increases toward the end of the test.

Fig. 4. Development of measured vertical stress at three levels in the two subbase layers. Vertical dashed lines indicate GWT raise. Initial measurement made after 20 000 load repetitions.

3.4

Strain

The measured vertical strain at three levels in the subbase layer is displayed in Fig. 5. The strain increases throughout each phase, while it decreases at each GWT raise. The decrease in measured strain as GWT was raised was not expected. This indicates that the structure is getting stiffer, which disagrees with the SPC registrations in Fig. 3 and Fig. 4, as well as the FWD measurements. Response measurements using 50 kN wheel load was also analysed, and showed the exact same tendency for both stress and strain as Fig. 4 and Fig. 5. Registered strain signals from two sensors in the lower part of the subbase layers are displayed in Fig. 6. No difference in strain is visible between phases w1 and w2, while in phase w3, strain increases by 60–70%.

Fig. 5. Development of measured vertical strain at three levels in the two subbase layers. Vertical dashed lines indicate GWT raise. Initial measurement made after 20 000 load repetitions.

Performance of Unbound Pavement Materials in Changing Moisture Conditions

a) 0/90 mm structure

77

b) 22/90 mm structure

Fig. 6. Vertical strain measured in the lowest part of the subbase layer after 377 000 (w1), 555 000 (w2) and 940 000 (w3) load repetitions

3.5

Rut Development

The surface rut development during the APT is shown in Fig. 7. A post-compaction tendency can be observed from the increased rut development for the initial 20 000– 30 000 load repetitions. The 0/90 mm structure obtains a surface rut depth of 19.9 ± 1.0 mm after 1 233 000 load repetitions, compared to 20.3 ± 0.7 mm for the 22/90 mm structure. The average rate of rut for each phase is calculated in Table 3. The rate of rut increases for each GWT raise.

Fig. 7. Rut development during APT. Each line represents the average of three laser profiles; error bars show max/min measurements. Dashed vertical lines indicate GWT phase transitions Table 3. Average rut rate the last 200 000 load repetitions of each phase [mm per 100 000 load repetitions] w1 w2 w3 0/90 mm structure 0.43 2.14 2.67 22/90 mm structure 0.56 1.78 2.80

78

M. Fladvad and S. Erlingsson

4 Conclusions Two instrumented flexible pavement structures were tested by an HVS where the GWT table was increased twice during the APT. Moisture, stresses and strains in the pavement were measured, along with surface deflection from FWD and rut depth. Both pavement structures proved to be durable, showing about 20 mm rut depth after 1.2 million load repetitions with a 60 kN dual-wheel load (axle load 12 tonnes). The accelerated traffic corresponds to 2.55 million 10 tonne standard axles or an annual average daily traffic of about 3000 vehicles per day over 20 years, assuming the amount of heavy vehicles is 10%. Pavement response and rut development are affected by the GWT level. When the GWT is raised, SPC register lower stress levels; hence, the stiffness is decreased. The vertical strain measurements indicate a decrease in strain as GWT is raised, indicating that the stiffness of the pavement materials increases. This is not in agreement with SPC or FWD measurements. The authors are unable to explain this effect, and these results must be studied further. The rutting accelerates considerably as GWT is introduced to the upper part of the structure. Even though the structures respond differently to the GWT increases, both reach a maximum rut rate of about 2.7 mm per 100 000 load repetitions in phase w3, and similar rut depth when the test is finished. For both structures, flooding results in a substantial increase in pavement degradation. The data provided from the APT presented here can be used to model unbound materials’ response to increased moisture content following increased precipitation and more intense rainfall events. Acknowledgements. This research is funded by the Norwegian Public Roads Administration, with contribution from the Research Council of Norway (project no. 256541). The manuscript is a part of the first author’s PhD degree at NTNU. The authors would like to thank Veidekke for supplying the subbase materials to the APT.

References Aarstad, K., Petersen, B.G., Martinez, C.R., et al.: Local use of rock materials - production and utilization. State-of-the-art. SINTEF, Trondheim (2019). https://www.sintef.no/globalassets/ project/kortreist-stein/012-kortreist-stein-sota-h3-endelig.pdf Aksnes, J., Myhre, Ø., Lindland, T., et al.: Frost protection of Norwegian Roads. Basis for revision of the Norwegian pavement design manual, Statens vegvesens rapporter nr. 338. Norwegian Public Roads Administration, Trondheim (2013). https://www.vegvesen.no/fag/ publikasjoner/publikasjoner/statens+vegvesens+rapporter/_attachment/748672?_ts=14a524b ccb8 ARA Inc.: Guide for the Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures, Final report, NCHRP 1-37A. Washington DC (2004) Dawson, A. (ed.): Water in Road Structures - Movement, Drainage & Effects. Springer, Dordrecht (2009). https://doi.org/10.1007/978-1-4020-8562-8

Performance of Unbound Pavement Materials in Changing Moisture Conditions

79

Erlingsson, S.: Impact of water on the response and performance of a pavement structure in an accelerated test. Road Mater. Pavement Des. 11, 863–880 (2010). https://doi.org/10.1080/ 14680629.2010.9690310 Erlingsson, S.: Numerical modelling of thin pavements behaviour in accelerated HVS tests. Road Mater. Pavement Des. 8, 719–744 (2007). https://doi.org/10.1080/14680629.2007.9690096 Lekarp, F., Dawson, A.: Modelling permanent deformation behaviour of unbound granular materials. Constr. Build. Mater. 12, 9–18 (1998). https://doi.org/10.1016/S0950-0618(97) 00078-0 Li, T., Baus, R.L.: Nonlinear parameters for granular base materials from plate tests. J. Geotech. Geoenviron. Eng. 131, 907–913 (2005). https://doi.org/10.1061/(ASCE)1090-0241(2005) 131:7(907) Norwegian Public Roads Administration: Håndbok N200 Vegbygging. Statens vegvesen Vegdirektoratet, Oslo (2018). https://www.vegvesen.no/_attachment/2364236/binary/1269980 Saevarsdottir, T., Erlingsson, S.: Water impact on the behaviour of flexible pavement structures in an accelerated test. Road Mater. Pavement Des. 14, 256–277 (2013). https://doi.org/10. 1080/14680629.2013.779308 Saevarsdottir, T., Erlingsson, S., Carlsson, H.: Instrumentation and performance modelling of heavy vehicle simulator tests. Int. J. Pavement Eng. 17, 148–165 (2016). https://doi.org/10. 1080/10298436.2014.972957

Performance Evaluation of Asphalt Pavement with Semi-rigid Base and Fine-sand Subgrade by Indoor Large-Scale Accelerated Pavement Testing J. T. Wu1(&) and Y. T. Wu2 1

2

Ningbo Institute of Technology, Zhejiang University, Ningbo, China [email protected], [email protected] School of Geology Engineering and Geomatics, Chang’an University, Xi’an, China

Abstract. A large-scale accelerated pavement testing (APT) with mobile load simulator 66 (MLS66) was carried out in an indoor structure groove, in order to investigate the performance development of asphalt pavement with semi-rigid base and fine-sand subgrade. The test was conducted with 20% overload and external heating system. Rutting deformation, seismic modulus (measured by portable seismic pavement analyzer), damage mode of asphalt layers and volume parameters were measured. In addition, stress and deformation of fine-sand subgrade were also analyzed. Results indicate that shear flow deformation occurred, and no cracking was observed during diagnostic excavation. Relationship between rutting depth and loading applications was piecewise linear. The average seismic modulus decreased by a quarter after loading compaction. Compaction of upper and middle asphalt layers could be also one of critical factors for the resistance of shear flow deformation of asphalt mixture at high average temperature. Moreover, phase rate of rutting deformation and lateral upheaval coefficient (proportion of the maximum height of lateral upheaval to the maximum rutting depth) were appropriate to be evaluation indicators. Keywords: Accelerated pavement testing
Rutting deformation modulus
Fine-sand subgrade
Performance evaluation


Seismic

1 Introduction Since the road transport and large-tonnage vehicles increase dramatically coupled with traffic channeling, rutting, cracking, upheaval, wave break, potholes and other early damages turn up in certain highways after 3 to 5 years’ service (Sun 2005). This indicates that the structural design or pavement materials are uncoordinated with the working life of road. Mechanical response of pavement structure, failure modes and damage mechanisms under overload, low speed or other complex traffic conditions have already been research hotspots. As an advanced method, accelerated pavement testing (APT) can determine the response and performance of pavement in a short term via applying axle loads on a dedicated road under controlled conditions. It becomes an © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 80–89, 2020. https://doi.org/10.1007/978-3-030-55236-7_9

Performance Evaluation of Asphalt Pavement

81

effective solution to overcome shortcomings of indoor routine tests and long-term field observation of roads in service. There are dozens of APT throughout the world currently. These projects play important roles in pavement material optimizing, pavement structure determination and relevant specifications or guidelines development, and so forth (Hugo et al. 2012; Wynand 2012). But most of them were carried out on flexible base asphalt pavement, the research findings of which are not necessarily directly applicable in China. Because more than 90% of asphalt pavements are constructed with semi-rigid base, and there exists great modulus difference between the inorganic binder base and asphalt course. In addition, there are a few APT studies focus on asphalt pavements with fine-sand subgrade (Wu et al. 2015). In order to pre-judge the reasonability of the pavement structure designed for a national highway, an APT was carried out in a large-scale indoor structure groove by the Mobile Load Simulator 66 (MLS66). Rutting deformation and seismic modulus of asphalt layers, as well as the stress response and permanent deformation of subgrade were measured. After completing loading, a structural excavation and drilling core samples were obtained for determining state of pavement structure. Finally, through analyzing the depth distribution and trend of experimental data, failure mode, evaluation indicators and critical factors for asphalt pavement on semi-rigid base and finesand subgrade under accumulative loads are determined.

2 Materials and Methods 2.1

Pavement Structure

The asphalt course consisted of three layers. The upper layer was SMA-13 (short for stone mastic asphalt, with the nominal maximum aggregate size (NMAS) of 13.2 mm) of 40 mm thickness modified by styrene butadiene styrene (SBS) and rubber particles. The middle and lower layers were AC-20C and AC-25C (short for coarse asphalt concretes with the NMAS of 19 mm and 26.5 mm separately) with the corresponding thickness of 60 mm and 80 mm, respectively, both modified with rock asphalt. The modified asphalt macadam waterproof bonding layer with broken stone bestrewed connected asphalt course and semi-rigid base. The base was 360 mm cement stabilized macadam (4% cement) along with 180 mm low-dose cement stabilized macadam subbase (3% cement). The subgrade was compacted fine-sand of 800 mm. The humidity at the height of 200 mm and 400 mm were 14.7% and 10.8%, respectively. Stress sensors were laid at the height of 200 mm, 400 mm, 600 mm and 800 mm of subgrade. And, at height of 200 mm and 600 mm, deformation sensors were placed, respectively. Gravel drainage layer of 480 mm was at the bottom of groove. The road structure was constructed mechanically in a large-scale indoor groove, which was 24 m in length, 12 m in width and 2 m in depth. Temperature and humidity of the pavement structure were controlled through environmental systems. The design of structure size and environmental simulation system was fit for reproducing the conditions of actual road.

82

J. T. Wu and Y. T. Wu

2.2

Loading Test Parameters

In this test, MLS66 simulated traffic overload up to 120 kN, 20% higher than the standard axle load (BZZ-100). Each measured actual contact track footprint area was 0.041 m2, therefore, the calculated average contact pressure was 0.73 MPa. The loading frequency was set at 6,000 loading applications per hour and the lateral wander of wheels was not considered in view of time limit. To simulate the unfavorable condition, the pavement was heated artificially with six heating plates placed next to the tracks in the loading process. The temperature of heating plates was set to 65 °C. And the temperatures measured at outside edge and center of wheel were roughly 60 ° C and 40 °C, respectively. This test lasted for 5 days and after each loading day, the cumulative loading applications were up to 0.12 million, 0.25 million, 0.35 million, 0.42 million and 0.49 million, respectively.

3 Asphalt Pavement Results Analysis 3.1

Rutting Deformation Analysis

Deformation of surface profiles was measured by MLS Profilometer Driver-P2003 after loading each day. The loading area was divided into seven sections marked as 1st–7th with the interval of 1 m along the driving direction. Each profile section was 1.5 m in length, and data were collected at 10 mm interval. 3.1.1 Rutting Deformation After a certain loading, the surface profiles of cross-sections were W-shaped (Fig. 1). Baseline profiles were taken before the APT without loading, as the calculation basis of the follow-up rutting deformation. The part of deformation above the baseline was defined as the upheaval rutting and it was expressed with positive values; correspondingly, the part below the baseline was depression expressed with negative values. And the deformation absolute difference of the highest and lowest points of one profile curve was taken as the rutting depth. The variation of maximum rutting depth of sections 1st–7th with loading applications is shown in Fig. 2. 12

0

490000

6 4 2 0 -2 -4

Depression

Upheaval

Deformation /mm

8

Rutting depth

10

-6 -8 0

150 300 450 600 750 900 1050120013501500 Transverse length (mm)

Fig. 1. Rutting deformation

Fig. 2. Relationship between rutting depth and loading

Performance Evaluation of Asphalt Pavement

83

Seen from the relationship of rutting depth versus loading applications, rutting depth increased with the loading applications as expected. It could be divided into two phases and described with piecewise linear. The increase rate of average rutting depth during 0 to 0.25 million applications was 0.5 mm per 10,000 applications, while it inclined to 0.2 mm per 10,000 applications in the second phase. There were differences in the rutting deformation among sections, and the average upheaval, depression and rutting depth of these sections were almost equal to the corresponding values of the 3rd section. The rutting depths from the 4th, 5th, and 6th sections were greater, followed by the 3rd section, then those of the 2nd and 7th sections were smaller, and that of 1st section was smallest. This is because the 1st and 7th sections both located at end of the wheel tracks, where the wheels began to contact with and get away from the pavement. Thus, force applied was relatively smaller. However, the contact state of wheel and surface at sections locating in the middle of wheel tracks were more stable, which can be selected as the analysis object for performance change in the MLS accelerated loading test. Moreover, the upheaval deformations were greater than depressions, and the difference between them gradually increased with the loading applications. Taken the 4th section as example, the absolute differences in five loading days were 0.218 mm, 0.275 mm, 1.932 mm, 2.561 mm, and 3.063 mm, respectively. During the third loading day, changes of depression and lateral upheaval were 0.160 mm and 1.497 mm, while in the fifth day, they were 0.061 mm and 0.563 mm. Here, the early deformation showed mainly as the compacted rutting, then it converted to shear flow rutting with continues loading, and finally, the absolute difference between upheavals and depressions tended to a stable and low level. A field APT with MLS66 on an actual highway with the same materials and heating temperature condition, which was similar to the test done in the large-scale groove, had been conducted (Wu et al. 2012). In the field test, the fine-sand subgrade was 1.5 m in thickness, and the load was set at 150 kN with wheel contact pressure of about 0.80 MPa, with the same loading frequency as the indoor test. Comparing with the indoor test, the distribution of rutting deformation in sections was different, rutting deformations at the 6th and 7th sections were relatively larger. The reason is that the slope of actual road caused greater wheel pressure in the end of tracks at these two sections, which resulted in a larger rutting deformation. Middle sections could also be used to show the average deformation level during loading test, and the sections for maximum deformation investigation should be determined according to the transverse and longitudinal slope of the road. Values from indoor groove test were greater comparing with the maximum and average of rutting deformation in the field test. With the increase of loading applications, the difference between them became wider, shown in Fig. 3.

J. T. Wu and Y. T. Wu Rutting depth (mm)

84

Related finite element simulation results show that the factors, such as the subgrade thickness, boundary constraint at the bottom of soil, road width restriction and outer lateral restraint, have little effect on the results of tests (Wu 2010). Therefore, the above differences maybe mainly result from the road construction. This point was also proved by subsequent analysis results of excavation and core samples. In brief, it is still important to control quality and compaction degree during construction to ensure rutting deformation at low level.

25 20 15 10 5 0

12

25

35

42

49

Indoor-maximum 9.254 16.613 17.950 22.337 22.961 Field-maximum

7.988 11.409 12.491 13.753 15.188

Indoor-mean

7.058 12.819 14.482 17.632 18.166

Field-mean

6.531

9.726 10.130 10.993 12.362

Fig. 3. Comparison of rutting deformation of indoor and field tests

3.1.2 Rutting Deformation Rate In order to study the formation law of rutting deformation in the accelerated pavement test in the indoor groove, the node rate (cumulative average rate) and phase rate (average rate of each loading day) of rutting deformation were calculated by the following Eqs. 1 and 2, respectively. Results are listed in Table 1. Rn Nn

ð1Þ

Rn  Rn1 Nn  Nn1

ð2Þ

Vnode ¼ Vphase ¼

where Vnode is defined as the node rate of rutting deformation, Vphase is defined as the phase rate of rutting deformation, Rn and Rn−1 are the rutting deformation corresponding to the loading applications of Nn and Nn−1, n is the separate phase, and in this research, n refers to the loading day, i.e. n = 1, 2, 3, 4, 5 in the calculation. The average node and phase rate of rutting deformation both reached the maximum values at the start, then the values gradually descended with the loading applications increasing and maintained stable at last. During the entire loading process, node rates of upheaval deformation was 0.04–0.07 mm/104 greater than that of depression. Phase rates of upheaval were much greater than corresponding values of depression especially during 0.25–0.35 million applications. It helps to explain the significant acceleration of shear flow with the application of load in this phase. The fluctuation of phase rate of rutting depth was more significant. Because of the applicability of reflecting the rutting development and identifying deformation state phase stage, it was suggested to be an evaluation indicator.

Performance Evaluation of Asphalt Pavement

85

3.1.3 Upheaval Coefficient The upheaval coefficient is defined as the proportion of the maximum height of lateral upheaval to the maximum rutting depth. Calculation results are given in Table 2. The average upheaval coefficient of indoor APT test was 0.55, greater than that from field test of 0.40. The comparison shows that the asphalt pavement in groove produced a greater shear flow deformation once again. It is similar to the maximum rutting depth that the average upheaval coefficient of the 1st–7th section was approximately equal to that of section 3rd. But the order of upheaval coefficient value among measured sections, mainly the relative rank of the 2nd and 4th sections, was different. Through comparison of Table 2, the order of upheaval coefficient among sections from big to small was listed as the 2nd, 5th, 6th, 1st, 3rd, 4th and 7th section. Generally, larger rutting depth means larger lateral upheaval coefficient. Considering inhomogeneity of loads and mixture materials, lateral upheaval coefficient, which was consist of more data of section, should be more appropriate to be as an indicator in rutting analysis. Table 1. Development rate (unit: mm
10−4) 4

Table 2. Upheaval coefficients

Applications (10 ) Upheaval Depression Vnode Vphase Vnode Vphase

Section Loading applications (104) 12 25 35 42 49

12 25 35 42 49

1st 2nd 3rd 4th 5th 6th 7th Mean

3.2

0.31 0.28 0.24 0.23 0.21

0.31 0.26 0.14 0.19 0.06

0.28 0.23 0.17 0.19 0.16

0.28 0.18 0.02 0.26 0.01

0.51 0.64 0.47 0.49 0.61 0.54 0.38 0.52

0.57 0.65 0.53 0.51 0.61 0.58 0.41 0.55

0.59 0.67 0.59 0.55 0.63 0.59 0.46 0.58

0.56 0.64 0.57 0.56 0.59 0.55 0.41 0.55

0.57 0.64 0.58 0.57 0.60 0.55 0.43 0.56

Seismic Modulus Analysis

In the test, the seismic modulus of asphalt course was measured by portable seismic pavement analyzer (PSPA), a nondestructive field-deployable device for measuring sonic, ultrasonic, and resonant vibrations. This enables pavement evaluation under loading. Seismic modulus of each section within 40 mm to 180 mm from the surface top was measured after rutting measurement. After the applications of 0, 0.12, 0.25, 0.35 and 0.42 million, the measured values are 8,900, 6,600, 6,500, 6,400 and 7,000 MPa, with the temperature at 40 mm of 18, 40, 40, 42, and 38 °C, respectively. The relationship among seismic modulus, loading applications, and temperature at the depth of 40 mm below the road surface conforms to the regression Eq. 3 as follows. Ms ¼ 5N  142T þ 12147 ðR2 ¼ 0:78Þ

ð3Þ

where Ms is the average seismic modulus of asphalt course /MPa, N is the loading applications /104 and T is the temperature at the depth of 40 mm below the road surface /°C.

86

3.3

J. T. Wu and Y. T. Wu

Excavation and Core Samples Analysis

The damage of trail road was diagnostically studied by excavation with steps from the pavement surface to subgrade top. The excavation cross-section of asphalt layers after loading test helps to confirm that the rutting was the only damage generated in the upper and middle asphalt layers. Meanwhile, the base and subgrade had no visible deformation, and there was no crack on the top of base. Almost all of shear flow were in asphalt layers. It further illustrated that the semi-rigid subgrade efficiently separated the plastic deformation of asphalt mixture from the subgrade and plastic deformation was the main cause of rutting in this structure. Core samples of asphalt layer from the wheel track and unloaded part with a diameter of 100 mm were drilled respectively. Heights of three layers in cores were measured by distinguishing the dividing line between layers. Every sample from unloaded part was cut into separate cylinders according to three layers of surface with different mixture materials. Furthermore, samples from loaded zone were cut into 100 mm roughly to test resilient modulus, remaining part of the upper and middle layer together. The average height of core samples from unloaded area was 164.20 mm. The thickness of lower layer varied from 77 mm to 83 mm, which was close to the design thickness, while both of upper and middle layers were less than the designed values. The water absorption of core samples from most parts, of which dense gradation Marshall samples usually lower than 1.0%, was large. It indicates that the compaction of asphalt course was insufficient in some extent. Calculated void ratios of three asphalt layers were 5.2%, 9.3%, and 6.9% respectively. The corresponding compaction were 94.8%, 90.7%, and 93.3%. Low compaction makes the rutting deformation be more severe in the middle layer. In addition, upper and middle asphalt mixtures were compacted after loaded, and bulk density became larger. Sample analysis mentioned above implied that compaction of upper and middle layers was one of critical factors for rutting resistance of asphalt pavement.

4 Fine-sand Subgrade Analysis 4.1

Dynamic Stress Response

Additional dynamic stress in fine-sand subgrade was shown in Fig. 4. Dynamic stress descended gradually with the depth increasing, which was consistent with the transmission of inner stress under the driving load. From the vertical perspective, the additional stress at the subgrade height of 800 mm (on the surface of subgrade) was close to 21 kPa, and that at the heights of 200 mm and 400 mm was below 7 kPa. It is consistent with the theoretical solution of multi-layer elastic layered system (Ling et al. 2011). Therefore, the fine-sand subgrade with semi-rigid subgrade can diffuse the stress well. The dynamic stress had a trend of slightly increasing with the loading process, which might indirectly reflect the attenuation of structure modulus. But considering that the semi-rigid subgrade isolated the negative effects, the local impact effect caused by unevenness of the wheel track might also contribute to this phenomena.

Performance Evaluation of Asphalt Pavement

4.2

87

Permanent Deformation

From the relationship between the permanent deformation (measured by displacement sensors) of subgrade and loading applications, shown in Fig. 5, it can be known that: (1) In the early loading stage, the permanent deformation developed rapidly. It tended to be stable after the loading applications reached 0.35 million. This means that under the condition of 20% overload, the fine-sand subgrade would not have any problem of dynamic instability under the driving loads. (2) The permanent deformation at each depth of subgrade was controlled within 0.1%, so the permanent deformation of 3 m high subgrade can be within 3 mm, which is far below the control standards. It follows that, the additional stress of asphalt pavement caused by the permanent deformation of subgrade structure under semi-rigid base was almost negligible.

20 60

40 80

Permanent deformation (%)

24

Dynamic stress (kPa)

18 12 6 0

0.0

0.5 1.0 Loading time (s)

1.5

Fig. 4. Additional dynamic stress of subgrade

0.20 0.15 0.10 0.05 0.00

20 60 0

20 40 Loading applications (×104)

Fig. 5. Permanent deformation loading applications subgrade

60

versus

5 Conclusions and Discussions The pavement structure found reliable in the APT has been put into practice to reduce the construction costs. Conclusions can be drawn as follows: (1) Rutting deformation was the only damage form in the upper and middle layers of asphalt course within the loaded zone, and no crack was observed in the asphalt layers, semi-rigid base or subbase during diagnostic excavation. Almost all of shear flow arised in asphalt layers. Relationship between rutting depth and loading applications was piecewise linear. Both of maximum or average rutting depth gained from indoor APT were greater than the corresponding values in field test. And with the loading applications growing, the difference between them kept increase. Furthermore, the upheaval coefficient was up to 0.55, higher than the result of field test. The results of indoor test reflected the performance in a more unfavorable condition. Loading had compaction effect on asphalt layers. And the average seismic modulus decreased by 25% after loading compaction, which made the deformation severer. Compaction of upper and middle layer was one of critical factors for rutting resistance of asphalt pavement.

88

J. T. Wu and Y. T. Wu

(2) The sections in the middle of wheel track (loaded zone) could be selected to analyze the average level of rutting, but the maximum rutting analysis should be chosen by considering the road slope. Phase rate of lateral upheaval coefficient were appropriate to be evaluation indicators, because of their applicability to reflect the rutting development (including depression, upheaval and absolute rutting depth), identify deformation state phase stage and consist of more data of section. (3) Semi-rigid base prevented the transmission of deformation in asphalt mixture layer and isolated the negative effects of loads on upper layers. Therefore, the deformation of asphalt pavement caused by the permanent deformation of subgrade structure under semi-rigid base was almost negligible. (4) Due to the results gotten under the conditions of heating, heavy-traffic and continuous loading, different from the actual field behavior, the loading applications should be amended according to the factors of temperature, lateral wander, axial loads and so on, as follows: ① Temperature: As one of the most critical factors for asphalt pavement rutting, it should be amended. Suppose there are four months in a year with 40 °C as the lower test temperature, therefore the correction coefficient is 4/12; ② Lateral wander of wheels: During the loading test, no lateral wander of wheels was considered. However, for the two-way eight-lane road surface, the lane coefficient of 0.25–0.35 is regulated. 3/10 is taken as the correction coefficient. ③ Loading interval time: The loading was continuous in the full-scale test (6,000 applications per hour, that is 1.7 applications per second for one position), while it was intermittent on the actual road. Given that there is no data for reference, this correction is not considered. ④ Axle loads: the standard single-axle double-wheel load is 100 kN, while 120 kN is used in the indoor test. According to the axial load conversion formula (the conversion index is 4.35) for equivalent deflection in the asphalt pavement design specification, the effect of 120 kN on the vertical deformation of the structure is 2.21 times that of 100 kN, that is (120/100) ^4.35 = 2.21. For the field loading test, the coefficient can be calculated as (150/100)^4.35 = 5.83. Therefore, the loading applications in the indoor APT is equivalent to that of field pavement based on the following formula: Ne ¼ 49  104  4=12  3=10  2:21 ¼ 1; 083  104 , which is close to the lower limit for heavy traffic of 12 million. For the field APT, the corresponding corrected ones can be up to 2,857  104, larger than the limit for extra-heavy traffic of 25 million. In view of the rutting mean of 12 mm (allowable value of 10–15 mm for highway in specification), the asphalt pavement could be regarded reasonable.

Performance Evaluation of Asphalt Pavement

89

References Hugo, F., Arraigada, S., Tian, Z.F., Kim, R.Y.: International case studies in support of successful applications of accelerated pavement testing in pavement engineering. Paper Presented at the 4th International Conference on Accelerated Pavement Testing, APT 2012. University of California, Davis, 19–21 September 2012 (2012) Ling, J.M., Qian, J.S., Ye, F., et al.: Accelerated loading test of pavement with fine-sand subgrade. Tongji University. Shanghai, P.R. China (2011) Sun, L.J.: Structure behavior theory of asphalt pavement. Beijing, China (2005) Wu, J.T., Ye, F., Ling, J.M., Qian, J.S., Li, S.M.: Rutting resistance of asphalt pavements with fine sand subgrade under full-scale trafficking at high and ambient air temperature. Paper Presented at the 4th International Conference on Accelerated Pavement Testing, APT 2012. University of California, Davis, 19–21 September 2012 (2012) Wu, J.T., Ye, F., Hugo, F., Wu, Y.T.: Strain response of a semi-rigid base asphalt pavement based on heavy-load full-scale accelerated pavement testing with fiber bragg grating sensors. Road Mater. Pavement Des. 16(2), 316–333 (2015). https://doi.org/10.1080/14680629.2014. 995211 Wu, Y.B.: Design of laboratory circular track system and FEM simulation of circular track rutting test. Master thesis. Changsha: Hunan University (2010) Wynand, J.V.: Significant findings from full-scale accelerated pavement testing. Transportation Research Board (2012). https://doi.org/10.17226/22699

Accelerated Fatigue Damage Profile of Asphalt Concrete Placed on Semi-rigid Layer Yi Li, Jiahao Li, Liping Liu, and Lijun Sun(&) The Key Laboratory of Road and Traffic Engineering, Ministry of Education, Tongji University, Shanghai, China [email protected]

Abstract. Asphalt concrete (AC) modulus reduction caused by repeated axle loading significantly affects pavement long-term performance; including when built on a semi-rigid layer. However, quantifying this effect is challenging. The primary objective of this paper was to monitor and evaluate modulus reduction and fatigue damage accumulation at various AC depths utilizing data obtained from two semi-rigid pavement sections. During loading, a non-destructive method, portable seismic pavement analyzer (PSPA), was used to predict the modulus ratio. PSPA test results show that the damage is nonlinear with respect to the loading passes. Also, depth and AC thickness can influence the development of damage. A developed model showed that it could predict the aforementioned nonlinear relationship. The model parameters can be used to identify the damage level at various AC depths. Unexpected compared with previous understanding, the damage in AC layers was found to increased first, then decreased, and finally increased with the depth. Since PSPA is cheap, portable, and easy to apply, this method to identify the damage level in AC layers is proven to be applicable and practical. Keywords: Asphalt concrete
Fatigue damage
Modulus reduction
APT test

1 Introduction Asphalt concrete (AC) modulus reduction caused by repeated axle loading can lead to fatigue damage and fatigue cracking in AC layers, including when built on a semi-rigid layer. Evaluating and simulating the in-situ AC modulus deterioration process can help to identify the weak AC layer and proactively prevent the fatigue cracking in AC layers. This importance calls for special efforts in this area. In continuum damage mechanics (Kachanov 1986), the damage is defined as the relative decrease in the active area, that is, as the area lost to microcracking divided by the original area. The damage can also be expressed as the relative decrease in modulus, because the strain in the material is proportional to the stress in the intact part of the area (Ullidtz 1999). Based on this definition, many efforts have been made to predict the AC modulus reduction process. Deacon (1965) is one of the first researchers that introduced linear Miner’s law into a study on fatigue characteristics of asphalt mixtures. The model he proposed assumed that the total accumulation of fatigue damage during the design life © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 90–99, 2020. https://doi.org/10.1007/978-3-030-55236-7_10

Accelerated Fatigue Damage Profile of Asphalt Concrete Placed

91

was the linear summation of cycle ratio. After that, Hopman et al. (1989), Tsai et al. (2002, 2007), Ghuzlan and Carpenter (2006), Shen and Sutharsan (2011) all used Minier’s law to analyze the cumulative fatigue damage. However, with the data from WesTrack, Tsai et al. (2002) found that the linear hypothesis of cumulative fatigue damage was not accurate. (Tsai et al. 2003) used a Weibull density function to describe the nonlinear relationship between stiffness ratio and loading repetitions. Besides, Liu and Sun (2003) proposed the structural behavior function to describe the development of cumulative fatigue damage. In the aspect of obtaining the remaining modulus or stiffness data to analyze the damage process, coring and laboratory tests are the traditional methods. Celaya and Nazarian (2007), Mbarki et al. (2012), and Bell et al. (2012) all took this approach in their studies. However, this approach is time-consuming and is often conducted when all the loadings have been applied. Falling weight deflectometer (FWD) is one of the traditional non-destructive evaluation tools to obtain the remaining modulus of pavements. This method is convenient and can assist in pavement condition evaluation. Sebaaly et al. (1989) showed that the FWD test could be used to measure the progression of damage in asphalt layers at the FHWA accelerated loading facility (ALF). However, the back-calculation of the modulus is complicated and difficult. The portable seismic pavement analyzer (PSPA) is another non-destructive tool to monitor the modulus reduction process. It is an automated device for conducting the spectral analysis of surface waves (SASW) (Jurado et al. 2012). Based on the higher frequency impulse generated by the PSPA loading mechanism, the device can monitor the change in modulus for different AC layers. Celaya and Nazarian (2007), Bell et al. (2012), Jurado et al. (2012) all used this test method in their studies. The PSPA test is proved to be non-invasive, cheap, and convenient. The primary objective of this paper was to describe and evaluate the fatigue damage profile of asphalt concrete placed on the semi-rigid layer. The data used in this study was obtained from two semi-rigid pavement sections. The full-scale accelerated pavement testing (APT) was conducted to induce fatigue damage in the AC layers. Besides, PSPA was selected as the method to monitor the modulus reduction process during the APT test. Then, a developed model was used to fit the damage development process. By performing the regression analysis, the parameter of this model was obtained and normalized to identify the weak layer.

2 Experimental Method 2.1

Full-Scaled Accelerated Pavement Test

The Mobile Load Simulator 66 (MLS66) machine, shown in Fig. 1, was employed to induce fatigue damage in the AC layers. It applied 6,000 axle passes per hour, and the speed of the wheel was close to 22 km/h. Dual-tire wheels applied the loading directionally without lateral wandering. The wheel load was 75 kN on two tires, and the effective length of the loading area was 6.6 m.

92

Y. Li et al.

Fig. 1. MLS66 applied on: a) Section I, b) Section II

The data in this study was obtained from two semi-rigid pavement sections. Section I has a pavement structure that consists of a 180 mm thick AC layer, a 450 mm thick highly cement-treated base, a 150 mm thick granular subbase, and subgrade. Section II has a 300 mm thick AC layer, a 350 mm thick highly cement-treated base, a 200 mm thick granular subbase, and subgrade. The typical air-void content of the asphalt mixture was 4.2%, and the cement-content of the semi-rigid base was 5%. Both of the sections had fine-grained soils in the subgrade. According to the extracted cores from the sections, there was good bonding between all the layers. Also, as the test sections were newly built before the APT tests, it was believed that there was no cracking in the AC layers or the base. At the end of APT tests, 550,000 passes and 1,840,000 passes of loading were respectively applied to section I and section II. During the tests, both test sections were instrumented with the platinum thermistor sensors to record the AC layer temperature. For section I, the sensors were installed at 80 mm and 160 mm down from the surface, while the sensors were installed at 8 mm, 15 mm, 21 mm and 27 mm for section II. The measured temperatures were used to adjust the measured PSPA modulus to a reference temperature. After the APT tests, the cracking was recorded by visual inspection. For Section I, tiny longitudinal top-down cracking was observed at the outer edge of wheel paths, while no visible cracking was observed for Section II. 2.2

Portable Seismic Pavement Analyzer (PSPA) Test

PSPA was used to monitor the remaining modulus. It is an automated device to conduct the spectral analysis of surface waves (SASW). If the pavement structure has several materials, the surface wave velocity will change with wavelength. The velocities can be converted to moduli by applying Eq. 1 (Li and Nazarian 1994): E ¼ 2q½ð1:13  0:16vÞVR 2 ð1 þ vÞ

ð1Þ

where E means the calculated modulus, VR means the velocity of surface waves, q means mass density, and v means Poisson’s ratio. In this study, the Poisson’s ratio is 0.35, and the surface wave velocity versus the wavelength is determined by the time records measured by the two PSPA receivers. The spacing between the receivers is

Accelerated Fatigue Damage Profile of Asphalt Concrete Placed

93

254 mm. Due to this spacing, the PSPA test can only obtain the moduli of AC layers from 80 mm depth to 280 mm. To analyze the modulus reduction process at different AC depths, the AC layers of the two sections were respectively divided into 11 and 21 sublayers (10 mm for each). Figure 2 manifests the layout of the tests. The sections were divided by five stations. At the stations, the tests were conducted periodically at the center line of the wheel paths (positions 1 and 2) and were performed in the longitudinal and transverse directions. For each direction, the measurements were repeated ten times. To avoid random error, the results of each position and direction were averaged. Station 1 Station 2 Station 3 Station 4 Station 5

Longitudinal direction

Transverse direction 1m Test position 1

1m

1m

1m Center line of the wheel path Center line of the wheel path

Test position 2

Fig. 2. Layout of the PSPA tests on the both test sections

To exclude the temperature effect, the measured PSPA moduli were adjusted to a reference temperature by applying Eq. 2 (Li and Nazarian 1994): E25 ¼ ET =ð1:35  0:014TÞ

ð2Þ

where E25 and ET are the moduli at 25 °C and test temperature T (in degrees Celsius). In this study, both test sections were instrumented with temperature sensors to measure the AC temperatures. Figure 3 shows the recorded temperature during the PSPA tests. Since there was shelter over the pavement sections, the test section was not directly exposed to the sunshine, and the temperature gradient in the AC layers can be ignored. As a result, the average value of measured temperatures during each PSPA test was selected as the test temperature T in Eq. 2.

Fig. 3. Measured AC temperatures: a) Section I, b) Section II

94

Y. Li et al.

3 Results and Discussion 3.1

Development of Cumulative Fatigue Damage

According to continuum damage mechanics (Kachanov 1986; Ullidtz 1999), damage can be expressed as Eq. 3: D ¼ DA=A0 ¼ ðE0  EÞ=E0 ¼ 1  E=E0

ð3Þ

where D is the damage, ΔA means the area lost to microcracking, A0 means the original area, E0 means the initial moduli, and E means the remaining moduli of the material. Since the cumulative fatigue damage is hard to directly obtained from field tests, this study takes the modulus ratio (E/E0) to represent the development of damage. The remaining moduli can be obtained from PSPA tests. After temperature adjustment and calculation, parts of the modulus ratios versus the loading passes are illustrated in Fig. 4 and Fig. 5. Obviously, both of the transverse and longitudinal modulus ratios decrease significantly with the increasing loading passes, but different modulus ratio values are corresponding to the AC depths. In other words, the depth of the sublayer has an effect on the development of the cumulative damage. Meanwhile, compared with section I, the modulus reduction process of section II is much slower. It means the pavement structure or the AC thickness can also influence the cumulative damage. Additionally, it is better to use a nonlinear model to describe the development of the cumulative damage, instead of a linear one such as Miner’s law.

Fig. 4. Modulus ratio at position 1 of Section I: a) 80 mm, b) 110 mm, c) 140 mm, d) 180 mm

Accelerated Fatigue Damage Profile of Asphalt Concrete Placed

95

Fig. 5. Modulus ratio at position 1 of Section II: a) 80 mm, b) 150 mm, c) 220 mm, d) 280 mm

3.2

Developed Model for Modulus Ratios

To further describe the damage, a nonlinear developed model was utilized. This model proposed by Liu and Sun (2003) can be expressed as Eq. 4: ( E ¼ E0

"
 #) "
 # A B E A B 1  exp  ; ¼ 1  D ¼ 1  exp  N E0 N

ð4Þ

where N means APT loading passes, A and B are the model regression parameters. It can be inferred from Eq. 4 that the modulus ratio equals 0.632 as long as the loading passes equals A. It means that A can represent the loading passes at which the modulus decreases to 63.2% of its initial value. In other words, larger A means that, to induce the same damage, the AC layer can bear more repeated loadings. As a result, A can reflect the bearing capacity of AC layers and can help to identify the weak AC layer. In this study, since each APT test was ended in two months, aging was not considered as a partial explanation of the changes of A with time. By performing the regression analysis between modulus ratios and corresponding loading passes, parameter A was obtained. Figure 4 and Fig. 5 show part of the fitting results. It can be seen that the developed model can fit the modulus ratios well. Also, Table 1 gives part of the regression results of parameter A. There is a difference between the bearing capacity of the longitudinal direction and that of the transverse direction. It is confirmed that the testing direction and anisotropic property of AC cannot be ignored in structural behavior analysis. Furthermore, according to Table 1, there is a relationship between parameter A and depths. However, since this parameter has a wide range of values for different positions and depths, it is difficult to compare the regression results and analyze the relationship between the parameter and depths.

96

Y. Li et al.

For this reason, a simple normalization method was introduced into this study. It can help to control the uninteresting variability and narrow the range of values. After normalization, the values of parameter A are in the range of 0–1. This normalization method can be expressed as Eq. 5: xnormalization ¼

x  xmin xmax  xmin

ð5Þ

where xnormalization means the value of parameters after normalization, x means the original value of the parameter, xmin and xmax means the minimum and maximum value of the parameter at various depths. Figure 6 presents the normalized A of the two test sections. Obviously, for various positions and directions, the figures show a similar relationship between A and depths. For both sections, A generally declines first, then increases, and finally decreases with the depth. Meanwhile, Fig. 6(b) shows that, as the AC thickness increases, the normalized A of the bottom AC layer decreases significantly. In other words, for thick asphalt concrete placed on a semi-rigid layer, the bottom AC layer is weakest, and the repeated axle loadings can fast induce damage to it.

Table 1. Part of regression results of parameter A Depth (mm)

80 90 100 110 120 130 140 150 160 170 180

Position 1 Transverse direction I II 721 5396 693 4890 617 4440 616 4125 732 4886 804 5191 858 6527 843 6426 759 6310 753 5946 700 6415

Longitudinal direction I II 2311 3626 2048 3571 1503 3466 1270 3142 1305 3368 1532 4197 3289 4308 2106 4281 2548 4301 2821 4036 2300 4016

Position 2 Transverse direction I II 1491 4167 1204 3971 1074 3735 1066 3522 1309 3714 1911 4840 2595 5251 2412 4906 2132 4543 2099 4144 1836 3650

Longitudinal direction I II 1107 3450 1028 3295 1100 3209 998 3166 1789 3526 1911 4042 2010 4077 3366 4003 3033 4005 2765 3976 2762 3830

Accelerated Fatigue Damage Profile of Asphalt Concrete Placed

97

Fig. 6. Normalized A of: a) Section I, b) Section II

Furthermore, the values of the sublayers from 80 mm to 180 mm were compared. The figures show that the asphalt layer at 100–110 mm depth has a smaller value of A, while the asphalt layer at 140–150 mm depth has the maximum value. It means the asphalt layer at 100–110 mm depth is a weak sublayer, while the asphalt layer at 140– 150 mm depth has the opposite results. Also, Fig. 6 illustrates that the results of the two test sections have a great agreement. It proves that this method to identify the weak AC layer is applicable and practical. Several AC cores were taken from the wheel paths before and after the APT tests to conduct laboratory indirect tensile fatigue tests (IDFT). The asphalt concrete from 6 cm to 10 cm (Layer I) and the asphalt concrete from 12 cm to 16 cm (Layer II) were cut out, respectively. The results show that at different stress ratios, the reduction in fatigue life of Layer I is larger than that of Layer II. The results can validate the PSPA results and the analysis results from the model. The causes of the results were discussed. As the cores extracted after the APT tests show good bonding conditions between all the layers, the bonding cannot explain the weakness of the identified AC layer. Finally, the tensile strains of AC layers that placed on a semi-rigid base were considered, mainly because the profile of A values is contrary to the distribution profile of the strains. The larger strain means the fast damage induced in the AC layer. It may be a possible reason to explain the results of this study.

4 Conclusions In this study, the accelerated fatigue damage of asphalt concrete placed on the semirigid layer was monitored and evaluated. The portable seismic pavement analyzer (PSPA) was used as the monitor method. The test results show that the damage is nonlinear with respect to the loading passes, and the depth and the AC thickness can influence the development of the damage. A nonlinear model, structural behavior function, was used as the evaluation method. By performing the regression analysis, this model was proven to fit the damage process well. Also, the model parameter A can reflect the bearing capacity of

98

Y. Li et al.

AC layers and help to identify the weak AC layer. By analyzing the normalized A, it is found that the damage in AC layers increases first, then decreases, and finally increases with the depth. This finding is unexpected compared with the previous understanding. Besides, as the AC thickness increases, the normalized A of the bottom AC layer decreases significantly. For thick asphalt concrete placed on a semi-rigid layer, the bottom AC layer is weakest, and the repeated axle loadings can fast induce damage to it. Additionally, the results of the two test sections have a great agreement. The laboratory indirect tensile fatigue tests (IDFT) results can validate the PSPA results and the analysis results from the developed model. Finally, the tensile strains were considered as the cause of the weakness of the identified layer. Since PSPA is cheap, portable, and easy to apply, this method to identify the damage level in AC layers is proven to be applicable and practical. It can help to proactively prevent the pavement from fatigue cracking and help to determine the timing, extent, and type of maintenance. The pavement managers do not need to make decisions after the cracks after they become visible. Acknowledgements. The authors gratefully acknowledge financial support from National Key R&D Program of China under grant number 2018YFB1600100.

References Bell, H.P., Haward, I.L., Freeman, R.B., Brown, E.R.: Evaluation of remaining fatigue life model for hot mix asphalt airfield pavement. Int. J. Pavement Eng. 13(4), 281–296 (2012). https:// doi.org/10.1080/10298436.2011.566925 Celaya, M., Nazarian, S.: Stripping detection in asphalt pavement with seismic method. J. Transp. Res. Board 2005, 64–74 (2007). https://doi.org/10.3141/2005-08 Deacon, J.A.: Fatigue of Asphalt Concrete. University of California, Berkeley (1965) Ghuzlan, K.A., Carpenter, S.H.: Fatigue damage analysis in asphalt concrete mixtures using the dissipated energy approach. Can. J. Civ. Eng. 33(7), 890–901 (2006) Hopman, P.C., Kunst, P.A.J.C., Pronk, A.C.: A renewed interpretation method for fatigue measurement, verification of Miner’s rule. Paper Presented at the 4th Eurobitume Symposium, Madrid, 4–6 October 1989 (1989) Jurado, M., Gibson, N., Celaya, M., Nazarian, S.: Evaluation of asphalt damage and cracking development with seismic pavement analyzer. J. Transp. Res. Board 2304, 47–54 (2012). https://doi.org/10.3141/2304-06 Kachanov, L.M.: Introduction to Continuum Damage Mechanics. Martinus Nijhoff Publisher, Leiden (1986) Li, Y., Nazarian, S.: Evaluation of aging of hot-mix asphalt using wave propagation techniques. In: Engineering Properties of Asphalt Mixtures and the Relationship to Their Performance, no. 1265, pp. 166–179 (1994) Liu, L.P., Sun, L.J.: Performance-based structural design method on whole life asphalt pavement. J. Tongji Univ. (Nat. Sci.) 31(9), 1044–1048 (2003) Mbarki, R., Kutay, M.E., Gibson, N., Abbas, A.R.: Comparison between fatigue performance of horizontal cores from different asphalt pavement depths and laboratory specimens. Road Mater. Pavement Des. 13(3), 422–432 (2012). https://doi.org/10.1080/14680629.2012. 685843

Accelerated Fatigue Damage Profile of Asphalt Concrete Placed

99

Sebaaly, P., Tabatabaee, N., Bonaquist, R., Anderson, D.: Evaluating structural damage of flexible pavement using cracking and falling weight deflectometer data. J. Transp. Res. Board 1227, 64–74 (1989) Shen, S., Sutharsan, T.: Quantification of cohesive healing of asphalt binder and its impact factors based on dissipated energy analysis. Road Mater. Pavement Des. 12(3), 525–546 (2011). https://doi.org/10.1080/14680629.2011.9695259 Tsai, B.W., Harvey, J.T., Monismith, C.L.: WesTrack fatigue performance prediction using Miner’s law. J. Transp. Res. Board 1809, 137–147 (2002). https://doi.org/10.3141/1809-16 Tsai, B.W., Harvey, J.T., Monismith, C.L.: Application of Weibull theory in prediction of asphalt concrete fatigue performance. J. Transp. Res. Board 1832, 121–130 (2003). https://doi.org/10. 3141/1832-15 Tsai, B.W., Harvey, J.T., Monismith, C.L.: Calibration of fatigue surface cracking using simplified recursive Miner’s law. J. Assoc. Asph. Paving Technol. 76, 693–735 (2007) Ullidtz, P.: Deterioration models for managing flexible pavements. J. Transp. Res. Board 1655, 31–34 (1999). https://doi.org/10.3141/1655-05

Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic Simulations to Mitigate the Damage Effects on Road Structures Paul Marsac1(&), Juliette Blanc2, Olivier Chupin2, Thomas Gabet1, Ferhat Hammoum1, Navneet Garg3, and Mai Lan Nguyen2 1

2

MAST-MIT, Univ Gustave Eiffel, IFSTTAR, F-44344 Bouguenais, France {paul.marsac,thomas.gabet, ferhat.hammoum}@univ-eiffel.fr MAST-LAMES, Univ Gustave Eiffel, IFSTTAR, F-44344 Bouguenais, France {juliette.blanc,olivier.chupin, mai-lan.nguyen}@univ-eiffel.fr 3 FAA William J. Hughes Technical Center, Egg Harbor Township, NJ 08405, USA [email protected]

Abstract. The concept of truck platooning is to take advantage of the connectivity technologies and automated driving support systems to link trucks in close formation (convoy) in order to increase transport efficiency, reduce fuel consumption and gas emissions while improving road safety. However, closely guided trucks following each other could have a different impact on road structures than the usual truck traffic. Notably, different studies with traffic simulators highlight the significant effect of the wander pattern of traffic on the damage observed on test road structures. On the other hand, the positioning systems could offer the opportunity to choose an optimized wander pattern of the trucks within the platoons, especially designed to mitigate the overall damage. In this context, the study reported in this paper addresses, through numerical simulations, the multi-loading effects of truck platoons on road structures for different configurations of wander patterns. The aggressiveness of these patterns is evaluated on the basis of the viscoelastic response of the pavement structures (strain field) computed using software ViscoRoute© 2.0. Finally, recommendations on the choice of wander pattern configurations to be tested on APT facilities are proposed. Keywords: Truck platooning
Numerical simulation
Traffic damage
Traffic wander pattern

1 Introduction Truck platooning consists in making trucks driving in close formation, without drivers, by using the connectivity technologies and automated driving support. It may increase transport efficiency, reduce fuel consumption and gas emissions while improving road safety. However, closely following guided trucks could have a different impact on road infrastructures than the usual truck traffic. Different experimental studies performed by © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 100–107, 2020. https://doi.org/10.1007/978-3-030-55236-7_11

Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic

101

means of fatigue carrousels (Kerzrého et al. 2012) or other traffic simulators highlighted the significant effect of the wander pattern of traffic on the damage observed on tested road structures (Hayhoe et al. 2004; Simonin and Hornych 2018), particularly in the case where there is no wandering. In the framework of the French rational road design procedure, load wandering is taken into account in an implicit manner. Indeed, the coefficients used to correlate computations and field results are set up by means of experimental tests performed on the fatigue carrousel at Université Gustave Eiffel using a fixed load wandering. In the design method, the wandering coefficient is fixed and the same for any design. When the design procedure was adapted to airfields pavements, the wandering coefficient was explicitly taken into account, to adapt to the specificity of airplane loads wandering and then to better characterize the damage due to airplanes traffic. Contrarily to airplanes, truck platoon may concentrate the loads on a given lane instead of spreading them as airplanes can do. In this context, the first step of this study consists in taking advantage of numerical simulations to get an order of magnitude of the impact of truck platoon on service life of infrastructures. Strains are computed at the bottom of the asphalt concrete layer, for different configurations of wander patterns. Viscoelastic computation is performed by means of the software Viscoroute 2.0. Then, damages induced by the different strain fields are assessed using the French design method (Corte and Goux 1996) and compared to quantify the potential impact on the road structure durability.

2 Numerical Simulations The numerical simulations are performed with the software ViscoRoute© 2.0. 2.1

Overview of ViscoRoute© 2.0

ViscoRoute© 2.0 (Chabot et al. 2010; Chupin et al. 2010; Hammoum et al. 2010) is a software dedicated to the computation of the dynamic response of multilayer pavements subjected to loads moving at constant speed (quasi-stationary assumption). The material layers can be modeled as linear elastic (non-bituminous layers) or viscoelastic according to the Huet (Huet et al. 1963; Huet 1999) or Huet-Sayegh (Sayegh 1965) models which are well adapted to represent the thermo-sensitive behavior of asphalt concrete. The computation is performed by means of a semi-analytical approach that consists in seeking the solution to the mechanical problem first in the wavenumber domain prior to returning to the spatial domain by running the Fast Fourier Transform. The calculation outputs are the displacement, strain and stress fields in the structure. 2.2

Road Structure

The road structure chosen is composed of 2 courses of bituminous materials on an infinite elastic roadbed. A constant Poisson’s ratio (m = 0.35) is assumed for the 3 materials. The Young modulus of the roadbed is 120 MPa. The base-course is a 20 cm thick layer of high modulus asphalt (named EME) and the surface-course is 2.5 cm

102

P. Marsac et al.

thick layer of thin asphalt (named BBTM). The viscoelastic properties of the materials of the 2 asphalt layers are modeled with a Huet-Sayegh model. This model (Fig. 1) is composed of an elastic spring connected in parallel to 2 parabolic dampers in series with an elastic spring. E0 is the static elastic modulus, E∞ is the instantaneous elastic modulus, k and h are exponents of the parabolic dampers (1 > h > k > 0) and d is a positive non-dimensional coefficient balancing the contribution of the first damper (1). h represents the temperature and s is a response time parameter which account for the equivalence principle between frequency and temperature (2). The values of the HuetSayegh model parameters for the 2 materials are given in Table 1. The simulation is made for a temperature of 30 °C.

Fig. 1. Schematic representation of the Huet-Sayegh model

E  ðxsðhÞÞ ¼ E0 þ

E1  E0

1 þ dðixsðhÞÞk þ ðixsðhÞÞh

sðhÞ ¼ eðA0 þ A1 h þ A2 h Þ 2

ð1Þ ð2Þ

Table 1. Values of the Huet-Sayegh model parameters for the 2 asphalt materials Surface course (BBTM) Base course (EME)

2.3

h d A0 A1 A2 E0 (MPa) E∞ (MPa) k 19 19644 0.213 0.628 2.535 3.072 −0.382 0.00165 22

31008

0.186 0.599 2.064 5.865 −0.388 0.0020

Loading Configuration

The loading configurations are chosen in order to simulate a platoon of 3 trucks closely following each other with 2 m interspace. Only one wheel path is considered as the 2 wheel paths are symmetrical. The loading is 32.5 kN for the tractor wheels and 40 kN for the semi-trailer tridem wheels. The wandering effect is simulated through variations of the transversal distance d between the wheel paths of each individual truck according to Fig. 2.

Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic

103

Tractor 1 front wheel Tractor 1 rear wheels Semi-trailer 1 tridem 1 d

1

2

2

2

d

1

y

1=32.5 kN 2=40kN

2m

z

x

Fig. 2. Loading configurations: platoon of 3 tractor/semi-trailer trucks with 2 m interspace distance and variable transversal distance d between the wheel paths of each truck

The simulation are performed in an orthogonal coordinate system (x,y,z) as mentioned in Fig. 2. The transversal strain (eyy) field at the bottom of the EME base course (z = −22.5 cm) is calculated with ViscoRoute© 2.0 for 8 loading conditions corresponding to values of the transversal distance d ranging from 0 to 35 cm with an incremental step of 5 cm. The speed of the trucks is fixed at 20 m/s. 2.4

Strain Field Results

The envelope (maximum value on a transverse profile) of the strain (eyy) along the longitudinal (x) axis is plotted in Fig. 3. For a better readability, only 3 simulations are plotted for wandering distances of d = 0, 15 and 25 cm because the trend is almost linear between 0 and 25 cm and the envelopes are similar for d > 25 cm. A view of the end of the curve, corresponding to the 3 wheels of the last semi-trailer, is plotted in Fig. 4. The Fig. 3 highlights the cumulative effect due to the lack of wandering. No wander leads to reach a transversal strain of 142 µstrains. For a wander of 15 or 25 cm, the transversal strain at the bottom of the AC layer reaches approximately 125 µstrains, as it is shown in the Fig. 4. -4

Strain YY

1.5

1

x 10

d=0 cm d= 15 cm d= 25 cm

0.5

0 20 15 10

Fig. 3. Envelopes (maximum value on a transverse profile) of the strain (eyy) along the longitudinal (x) axis for d = 0, 15 and 25 cm

104

P. Marsac et al. -4

1.5

x 10

d=0 cm d= 15 cm d= 25 cm

Strain YY

1

0.5

0 -13

-14

-15

-16

-17

-18

-19

X (m)

Fig. 4. Envelopes (maximum value on a transverse profile) of the strain (eyy) along the longitudinal (x) axis for d = 0, 15 and 25 cm. Zoom on the end of the curve (trailer of the last truck)

In the same way, the envelopes (maximum value on a longitudinal profile) of the strain (eyy) along the transverse (y) axis are plotted in Fig. 5. -4

1.5

x 10

d=0 cm d= 15 cm d= 25 cm

Strain YY

1

0.5

0 -1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Y (m)

Fig. 5. Envelopes (maximum value on a longitudinal profile) of the strain (eyy) along the transverse (y) axis for d = 0, 15 and 25 cm

Without wandering (d = 0 cm), the peaks of eyy are necessarily concentrated on the same longitudinal profile at y = 0. For d = 25 cm the maxima are distributed on 3 profiles and the cumulative effect is less important with a maximum peak value for eyy

Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic

105

12% lower than for d = 0 cm. Thus we can already assume that wandering can significantly mitigate the damage effect of a platoon.

3 Damage Effect Calculation To quantify the different damage effects, a calculation is performed according to the damage calculation method used in the pavement structure design software Alizé (Balay 2013). The damage D of a platoon is considered as proportional to the sum along a longitudinal profile of the positive strain eyy peaks power the slope of the fatigue law (3). An example of the local peaks and inter-peaks minima of ɛyy along a longitudinal profile is plotted in Fig. 6. D ¼ K 1=b

Xi¼n

1=b

e i¼1 di



Xi¼n1 i¼1

1=b



ð3Þ

euli;i þ 1

With: K: the constant of the fatigue law −1/b: the slope of the fatigue law (in this study, a value of −1/b = 5 is assumed) edi: the i order local peak of eyy on a longitudinal profile euli,i+1: the i order local inter-peaks minimum of ɛyy on a longitudinal profile

-4

Strain YY

1.5

x 10

1

YY strain profile local maxima local minima εd1 εd2

0.5 εul12 0 20

15

εul23 10

5

0

X (m)

-5

-10

-15

-20

Fig. 6. Local peaks and inter-peaks minima of eyy along a longitudinal profile

The second term of the right side of Eq. (3), proportional to D, is plotted along the transverse direction in Fig. 7.

106

P. Marsac et al.

Fig. 7. Second term of the right side of Eq. (3), proportional to D, along the transverse direction

The lack of wandering induces an increase of damage caused by a platoon of 3 trucks. A ratio of about 6 is found between the damage (Fig. 7) without wandering (squares) and the damage with a wandering with d = 25 cm (triangles).

4 Conclusion This numerical analysis highlights the significant potential effect of the wander pattern of the truck platoons on the damage induced on road structures. A ratio of about 6 is found between the damage induced by a 3 trucks platoon without wandering and the damage induced by the same platoon with an optimized wander pattern. This emphasizes the importance of considering the wander pattern in the positioning systems in order to conciliate the benefits associated with the platooning and the road network durability. For that purpose, the numerical simulation software ViscoRoute© 2.0 can contribute to improve accelerated pavement tests efficiency by pre-selecting relevant wander patterns according to the features of the platoons considered before experimental simulations on APT facilities or implementation on positioning systems. Acknowledgements. The activity presented in the paper is part of the research project ENSEMBLE co-funded by the European Union under the Horizon 2020 (H2020) Research and Innovation Program (grant agreement No 769115).

Optimization of Truck Platoon Wander Patterns Based on Thermo-Viscoelastic

107

References Balay, J.M.: Manuel d’utilisation ALIZÉ - LCPC version 1.5 (2013) Chabot, A., Chupin, O., Deloffre, L., Duhamel, D.: ViscoRoute 20: a tool for the simulation of moving load effects on asphalt pavement. Road Mater. Pavement Des. 11(2), 227–250 (2010). https://doi.org/10.1080/14680629.2010.9690274 Chupin, O., Chabot, A., Piau, J.M., Duhamel, D.: Influence of sliding interfaces on the response of a layered viscoelastic medium under a moving load. Int. J. Solids Struct. 47(25–26), 3435– 3446 (2010). https://doi.org/10.1016/j.ijsolstr.2010.08.020 Corté, J.F., Goux, M.T.: Design of pavement structures: the French technical guide. Transp. Res. Rec. 1539, 116–124 (1996). https://doi.org/10.1177/0361198196153900116 Hammoum, F., Chabot, A., Saint Laurent, D., Chollet, H., Vulturescu, B.: Effects of accelerating and decelerating tramway loads on bituminous pavement. Mater. Struct. 43, 1257–1269 (2010). https://doi.org/10.1617/s11527-009-9577-9 Hayhoe, G.F., Garg, N., Dong, M.: Permanent deformations during traffic tests on flexible pavements at the National Airport Pavement Test Facility. In: Airfield Pavements: Challenges and New Technologies, pp. 147–169 (2004) Huet, C.: Etude par une méthode d’impédance du comportement viscoélastique des matériaux hydrocarbonés. Ph.D. thesis Université de Paris, France (1963) Huet, C.: Coupled size and boundary-condition effects in viscoelastic heterogeneous and composite bodies. Mech. Mater. 31(12), 787–829 (1999). https://doi.org/10.1016/s0167-6636 (99)00038-1 Kerzrého, J.P., Hornych, P., Chabot, A., Deloffre, L., Trichet, S., Coirier, G., Gouy, T.: Evaluation of the aggressiveness of different multi-axle loads using APT tests. In: Jones, D., Harvey, J., Al-Qadi, I., Mateos, A., (eds.) Advances in Pavement Design through Full-scale Accelerated Pavement Testing. CRC Press, London, pp. 505–517. https://doi.org/10.1201/ b13000-62, http://www.crcnetbase.com/doi/pdf/10.1201/b13000-62 Sayegh, G.: Contribution à l’étude des propriétés viscoélastiques des bitumes purs et des bétons bitumineux. Ph.D. thesis, Faculté des Sciences de Paris, France (1965) Simonin, J.M., Hornych, P., Nguyen, M.L.: E-Way Corridor - DF4- Conception et construction des systèmes routiers, Phase 3 Options d’aménagement des voies d’autoroute actuelles (2018)

Impact of Mix Design Optimization on HMA Rutting Performance Under Accelerated Pavement Testing Fabrizio Meroni1(&), Wenjing Xue1, Gerardo W. Flintsch1,2, and Brian K. Diefenderfer3 1

2

Center for Sustainable Transportation Infrastructure, Virginia Tech Transportation Institute, Blacksburg, USA [email protected] Department of Civil and Environmental Engineering, Virginia Tech, Blacksburg, USA 3 Virginia Transportation Research Council, Blacksburg, USA

Abstract. To build durable asphalt pavements, one of the main concerns that many transportation agencies have shared in recent years is the low binder content of mixes designed according to Superpave. In order to add more binder and still be able to achieve satisfying performance, adjustments to the design compaction energy and mix aggregate structure can be made. In 2015, the Virginia Department of Transportation started an Accelerated Pavement Testing (APT) program at the Virginia Tech Transportation Institute. Two test lanes in this facility were used to investigate the field performance of two different dense-graded surface mixtures designed with two design gyration levels (50 and 65). Laser profiler, Multi-Depth Deflectometer, and pressure cells were used to monitor pavement response and permanent deformation throughout the experiments. This paper analyzes the instrumentation responses and permanent deformation collected during the APT experiments. In parallel, pavement cores were tested in the laboratory and analyzed in terms of dynamic modulus, flow number, and rutting resistance. The results showed that the optimized surface mixture, which was designed with 50 gyrations, was able to achieve a similar or better rutting resistance than the mixture designed at the traditional 65 gyration level. Keywords: Accelerated Pavement Testing
Gyrations
Compaction
Rutting

1 Introduction The Superpave mix design method was introduced in 1993 with the purpose of creating asphalt mixtures with sufficient binder content for long-term performance and durability (FHWA 2010). However, soon after the nationwide application of Superpave, many transportation agencies found that the mixes they produced using the Superpave design method were low in asphalt content, which resulted in poor durability, especially with respect to cracking resistance, and had shorter service lives. Prowell and

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 108–117, 2020. https://doi.org/10.1007/978-3-030-55236-7_12

Impact of Mix Design Optimization on HMA Rutting Performance

109

Brown (2007) conducted a study on the mixes’ densification and, after collecting material from 40 field projects, recommended a reduction of the gyratory compaction levels in order to allow higher asphalt contents. The Federal Highway Administration also suggested that transportation agencies perform an independent evaluation in order to modify the number of gyrations (FHWA 2010). The Virginia Department of Transportation (VDOT) introduced the Superpave design system in 1997, replacing the Marshall method. The original Superpave Ndesign compaction levels were implemented with 86, 95, and 109 gyrations for low, medium, and high traffic levels, respectively. In 2002, the design gyrations number was reduced to 65 to increase binder content and durability. Maupin (2003) conducted multiple laboratory tests and concluded that an increase in asphalt content would result in improved mix properties. The addition of more binder should improve the resistance to cracking without detrimental effects on rutting performance (Maupin and Diefenderfer 2007). Katicha and Flintsch (2016) evaluated the impact of gyration levels and binder contents on the rutting and cracking performances of dense-graded mixtures using the unconfined Flow Number test and indirect tension strength test. The results showed that binder content could be increased by reducing the number of gyrations from 65 to 50 and reducing the design air voids (i.e., VTM) and that the amount of binder affected the flow number test significantly for all mixtures and the indirect tension strength test for some mixtures. In addition to the changes in the design gyrations number, many state agencies started including additional specification criteria based on the test results obtained with devices such as the Asphalt Mixture Performance Tester (AMPT) and the Asphalt Pavement Analyzer (APA). Currently, the suitability of performance tests to support mix design is part of the balanced mix design framework, in which the mix performance is evaluated at the design stage both in terms of cracking and rutting resistance (West et al. 2018). Khosla and Ayyala (2013) studied asphalt mixes with different asphalt contents and design gyration levels of 50, 75, 100 and 125. The authors evaluated the balance between rutting and fatigue performance, recommending an intermediate Ndesign value of 85. Aguiar-Moya et al. (2007) evaluated the performances of mixes designed with 50, 75, 100, and 125 gyrations and found that the interval between 55 and 85 gyrations is the optimal level for the performance of the asphalt mixes. In 2015, VDOT started an Accelerated Pavement Testing (APT) program at the Virginia Tech Transportation Institute, employing a Heavy Vehicle Simulator (HVS) as its technological centerpiece. Two test lanes (lane 3 and 4) in this APT facility were designed to study the field performance of optimized dense-graded surface mixtures with different gyration levels and binder contents. The two lanes had the same structure with respect to the thicknesses and materials, with the only difference being in the surface layer mix. The surface mixture of lane 3 was designed with 50 gyrations, while lane 4 was designed with 65 gyrations.

110

F. Meroni et al.

2 Objective The primary objective of the study was to compare the performance of two densegraded surface mixtures, mainly with respect to rutting resistance. One mixture was traditional, acting as the control mix, and one was optimized in order to enhance durability. The two mixes were designed using different design gyration levels (65 for the traditional, 50 for the optimized), and tested at the Virginia APT facility. In addition to APT, laboratory testing was conducted to check the properties of the materials on a smaller scale. The two mixes were tested with the AMPT to determine their dynamic modulus and flow number and with the APA to determine their rutting resistance.

3 Equipment The layout and structure of the two test lanes are presented in Fig. 1, Fig. 2, and Fig. 3. The structure was built over a 68.6 cm subgrade layer placed over a rigid foundation. Both lanes featured a 17.8 cm 21-B aggregate subbase, a 10.2 cm IM-19.0 mm (normal maximum aggregate size; NMAS) intermediate mixture base layer, and a 7.6 cm 9.5-mm (NMAS) dense-graded surface layer. The surface mix in lane 3 was designed using 50 gyrations while the mix in lane 4 was designed using the current standard of 65 gyrations. One load cell was installed on the centerline of each cell; the locations of these are shown in Fig. 2 and Fig. 3. In addition, a Multi-Depth Deflectometer (MDD), which is an integration of several linear variable differential transformers (LVDTs), was installed in cells 3A and 4B in order to keep track of the vertical displacement at various depths within the pavement structure. Because the LVDTs in the same pod can’t be too close to each other, two MDDs were used for each test cell to ensure that all the critical depths were covered, as shown in Fig. 2 and Fig. 3.

Fig. 1. Section layout

Impact of Mix Design Optimization on HMA Rutting Performance

111

Fig. 2. Structure and instrumentation in cells 3A and 4B

Fig. 3. Structure and instrumentation in cells 4A and 3B

The heavy vehicle simulator (HVS) used at the Virginia APT facility is a model Dynatest Mark VI. This model has a maximum wheel speed of 20 km/h (±3.2 km/h) for loads that range from 30,025 kN to 100,085 kN. It can achieve 24,000 bidirectional passes or 12,000 uni-directional passes in 24 h (Cooke 2015). The unit features an environmental chamber that allows maintaining a constant temperature at the loaded area. The pavement surface is heated with infrared heaters located along the edge of the test lane within the environmental chamber. A laser profiler mounted on the HVS carriage was used on a daily basis to scan the pavement surface and measure the vertical permanent deformation at the pavement surface. The rut depth measurements were collected across the full width of the wheel path for a distance of 203.2 cm (101.6 cm in. on either side of the center of the wheel path) and along the length for a distance of 5.4-m. The spacing of the measurements was 4 in. in the longitudinal direction and 10.2 cm in the transverse direction.

112

F. Meroni et al.

4 Method The four test cells were loaded one at a time through a designed loading experiment. The two test cells in each lane were considered as replicates. During each test, the HVS operated continuously with the exception of regular daily maintenance and occasional repairs. The temperature of the surface layer was held at 40°F as monitored by a thermocouple embedded at a depth of 5.1 cm from the surface. The loading sequence consisted of applying a number of passes at progressively higher load levels of 40 kN (9,000 lbf), 53.4 kN (12,000 lbf), and 66.7 kN (15,000 lbf). After the wheel load was increased beyond 40 kN, the HVS was operated at a 40 kN wheel load for 15 min every morning to capture pavement responses at this load level throughout the entirety of the test. The loading was applied through a dual tire assembly, with 11.00R22.5 tires inflated at 7.58 bar (110 psi). The assembly was running uni-directionally at a constant speed of 6.44 km/h (4 mph). The 40 kN (9,000 lbf) load level was intended to simulate half of an 80 kN (18,000 lbf) standard axle load. The loading and repetitions were transformed into equivalent single axle loads (ESALs) using the conversion shown in Eq. 1: ESALs ¼

  wheel load 4:2 9;000

ð1Þ

The loading timeline and number of ESALs applied to the pavement within the testing period are provided in Table 1. Table 1. Loading timeline for Lane 3 and Lane 4 Cell 3B 4B 4A 3A

Start date 1/9/2017 3/20/2017 5/24/2017 8/12/2017

End date 3/7/2017 5/13/2017 7/31/2017 10/7/2017

# of passes 273,110 278,274 295,131 296,393

# of ESALs 704,163 641,141 850,014 829,029

5 Materials Laboratory Characterization Since the only difference between lane 3 and 4 was the surface layer, the laboratory testing focused on the material characterization of the surface mixtures. The composition of both the control mix (lane 4) and the optimized mix (lane 3) is reported in Fig. 4, which illustrates how the optimized mix was characterized by a higher asphalt content and a lower natural sand content.

Impact of Mix Design Optimization on HMA Rutting Performance

113

Fig. 4. Composition of mixes

The fundamental engineering properties of the asphalt mixture were characterized using specimens cored from the testing lanes. Specifically, the dynamic modulus and the flow number tests were performed on small-sized specimens with the AMPT in accordance with AASHTO T-378. The results of the dynamic modulus test are presented in Fig. 5 and the results of the flow number test are given in Fig. 6.

Fig. 5. Dynamic modulus test results

Fig. 6. Flow number test results

Figure 5 shows that the optimized surface mixture in lane 3 had a higher modulus than that in lane 4. With respect to the flow number, Fig. 6 shows that the optimized mix went into the tertiary zone with a larger number of load cycles compared to the traditional mix, which indicates that the mixture from lane 3 demonstrated a better rutting resistance than lane 4. The results obtained from the two tests were consistent though unexpected. Surface mixtures with higher design gyration levels are expected to have a better rutting resistance. The overall gradation of the mix, particularly the sand content, was believed to be the possible reason for such behavior.

114

F. Meroni et al.

To further check the rutting performance, full depth cores were tested with the APA. The results are shown in Fig. 7.

Fig. 7. APA test results of surface mixtures

The optimized mixture (lane 3) had an average rutting depth of 6.81 mm and the traditional mixture (lane 4) had an average rutting depth of 7.75 mm. Although the difference is not statistically significant, the traditional mixture presented a rutting depth 13.8% higher than lane 3. This result was consistent with the dynamic modulus and flow number tests.

6 Instrumentation Response The displacement at the layer interphases at multiple depths was measured via the LVDTs integrated into the MDDs (Xue et al. 2020). The permanent deformations of all the functional LVDTs are summarized in Fig. 8, which shows how the MDD installed in lane 3 measured lower deformation levels with respect to the top three layers when compared to lane 4. The subgrade, however, recorded similar levels of deformation in both lanes.

Impact of Mix Design Optimization on HMA Rutting Performance

(a) 3A

115

(b) 4B

Fig. 8. Vertical permanent deformation at multiple depths measured by MDDs

7 Rutting Performance In order to measure the permanent deformation at the surfaces of testbeds, two different methods were used: • The laser profiler mounted on the HVS carriage • The LVDT on MDD installed at the surface level. The laser profiler records many transverse planes along the longitudinal direction (one plane every 4 in.). Each point in the scanned surface represents the surface vertical permanent deformation (SVPD) at the location from the original level. MSVPD is simply the maximum of the SVPDs in a specific transverse plane, while the overall rutting depth incorporates the lifts and the shape of deformation in the transverse plane into the calculation. Compared to MSVPD, rutting depth is a more comprehensive indicator of pavement rutting status and is used more commonly in the pavement engineering field. Thus, rutting depth was calculated from the profiler measurements for each transverse plane of the cell. To obtain the rutting performance of test cells under constant speed, the rutting depths were averaged for every day throughout the experiments, as shown in Fig. 9. With respect to the LVDT on MDD installed at the surface level, the reference line of each fluctuation represents the location of the LVDT, and the deformation is calculated by subtracting the original location from the current location. Compared to the laser profiler, this method is affected by some limitations when it comes to measuring the deformation of pavement surfaces, as shown by Xue et al. (2020).

116

F. Meroni et al.

Fig. 9. Rutting depths for the four test cells

Figure 9 shows that the shapes of the four curves were similar, but their magnitudes were quite different. The two cells in lane 3 (3A and 3B) had lower rutting depths than the two testbeds in lane 4 (4A and 4B). For example, the average rutting depth of 3A and 3B was about 12.5 mm (0.5 in.) when the accumulated ESAL was 500,000, and the corresponding rutting depth of 4A and 4B was 22.5 mm (0.9 in.). Considering the objective of the experiment, the comparison suggests that the optimized mix have similar or better performance than the traditional mix designed for the higher compaction energy. One possible explanation is that the lower compaction energy allowed the designer to achieve the required air voids with a lower percentage of natural sand and slightly less asphalt and this helped improve the rutting performance. The test cells were trenched during the forensic investigation after the APT experiments to verify the deformation of each layer. Combining the thickness measurements with surface deformation provides a way to estimate the final deformation of the layers.

8 Conclusions This paper documents the results of an APT experiment which aimed at assessing the potential of mix optimization with the objective of improving rutting resistance. The study compared a control mix and an optimized mix designed using different design compaction energy. The control mix was designed with 65 gyrations, while the optimized mix was designed with 50 gyrations. The optimized mix had a lower natural sand content, which probably increased the aggregate interlock mechanism in the mix. The testing showed no indication that the optimized mixes would have rutting problems, supporting the implementation of the reduction of the design compaction

Impact of Mix Design Optimization on HMA Rutting Performance

117

energy level. The optimized mix exhibited a similar or superior rutting resistance in the full-scale setting, in the laboratory, and in the forensic investigation. These results also show that small differences in the aggregate structure, such as the different content of natural sand in the mixes studied, can greatly affect the mix performances.

References Aguiar-Moya, J.P., Prozzi, J.A., Tahmoressi, M.: Optimum number of superpave gyrations based on project requirements. Transp. Res. Rec. 2001(1), 84–92 (2007) Cooke, B.: HVS MK VI I VDOT User Manual. Dynatest Consulting, Atlanta (2015) FHWA: Superpave mix design and gyratory compaction levels. TechBrief (2010) Katicha, S.W., Flintsch, G.W.: Improving Mixture Durability Through Design Gyrations, Air Voids, and Binder Content. Virginia Transportation Research Council (2016) Khosla, N.P., Ayyala, D.J.: A performance-based evaluation of Superpave design gyrations for high traffic surface mixes. Procedia-Soc. Behav. Sci. 104, 109–118 (2013) Maupin, G.: Additional asphalt to increase the durability of Virginia’s superpave surface mixes. Virginia Transportation Research Council (2003) Maupin, G., Diefenderfer, B.K.: Design of a high-binder–high-modulus asphalt mixture. Virginia Transportation Research Council (2007) Prowell, B.D., Brown, E.R.: Superpave Mix Design: Verifying Gyration Levels in the Ndesign Table, vol 573. Transportation Research Board (2007) West, R., Rodezno, C., Leiva, F., Yin, F.J.: Development of a framework for balanced mix design. NCHRP Project (2018) Xue, W., Flintsch, G., Diefenderfer, B.: Measuring Pavement Permanent Deformation in Accelerated Pavement Testing (2020)

Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams from Accelerated Pavement Testing Trial Sameera Pitawala1(&), Arooran Sounthararajah1,2, James Grenfell2,3, Didier Bodin2,3, and Jayantha Kodikara1,2 1

Department of Civil Engineering, Monash University, Clayton Campus, Melbourne, VIC 3168, Australia {sameera.pitawala,arooran.sounthararajah, jayantha.kodikara}@monash.edu 2 Smart Pavement Hub – SPARC, ARC Industrial Transformation Research Hub (ITRH), Department of Civil Engineering, Monash University, Clayton Campus, Melbourne, VIC 3168, Australia 3 Australian Road Research Board (ARRB), Port Melbourne, VIC 3207, Australia {james.grenfell,didier.bodin}@arrb.com.au Abstract. The purpose of this study was to compare the properties of field constructed and laboratory manufactured foamed bitumen stabilised (FBS) beams under different conditions. A full-scale outdoor test pavement section was constructed with in-situ foamed bitumen stabilisation equipment at the accelerated loading facility (ALF) in Dandenong. The field beam specimens were cut from FBS slabs extracted from the trial pavement test section and the laboratory beams were manufactured in a mould using a British pendulum compactor. The flexural modulus of the field extracted beam specimens and laboratory manufactured beams were investigated using a four-point bending apparatus under different test conditions. Also, X-ray CT scanning was conducted on some field and laboratory samples to study the bitumen distribution throughout the aggregate skeleton. Flexural modulus master curves were developed for laboratory manufactured and field extracted beam specimens. Field extracted beam samples showed lower sensitivity to temperature and frequency compared to the laboratory manufactured FBS specimens. Keywords: Foamed bitumen


Accelerated pavement testing
X-ray CT scan

1 Introduction Foamed bitumen stabilised (FBS) materials have been successfully implemented in many roads across Australia and many other countries. Foamed bitumen stabilisation has become popular in recent years due to its distinctive characteristics such as low cost, environmental friendliness, ability to be opened to immediate trafficking and low moisture susceptibility (Ramanujam and Jones 2000; Kendall et al. 1999; Saleh 2007). Fatigue and rutting are generally considered key failure mechanisms of pavement materials Fatigue cracking has been frequently found in bound material pavement © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 118–126, 2020. https://doi.org/10.1007/978-3-030-55236-7_13

Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams

119

structures, while rutting can be observed in both bound and unbound pavement structures (Su et al. 2017). Recent field trials in Australia confirmed that fatigue cracking was the principal distress mode of FBS pavements (Jameson 2018). The four-point bending (4 PB) test is one of the most used to characterise the flexural fatigue performance of bituminous pavement materials in the laboratory (Austroads 2011; Alderson 2011). The performance of laboratory manufactured, and field extracted FBS samples differ due to various reasons. These include differences in mix design, compaction methods, curing mechanisms, and bitumen dispersion throughout the aggregate skeleton. This paper discusses the performance differences between field constructed and laboratory manufactured beam specimens. The reasons for the differences between field extracted and laboratory manufactured beams are also discussed in this paper.

2 Material Characterisation of Both Field and Laboratory FBS Samples The aggregate, bitumen and hydrated lime chosen for the laboratory manufactured specimens were similar as those used in the field construction for the ALF test pavement. However, the foaming agent used for the laboratory is different from the foaming agent used for FBS pavement construction in the field. Each material is described in the following sections in detail. 2.1

Host Aggregate

Virgin granite aggregate provided by Hanson Quarry at Lysterfield, Victoria was selected for both laboratory and field experiments. The particle size distributions (PSDs) of the field aggregate laboratory aggregate were determined as shown in Fig. 1. The optimum moisture content (OMC) and the maximum dry density (MDD) of the aggregates were found to be 6.1% and 2.23 t/m3, respectively (based on modified Proctor compaction) (Austroads 2019). 100

Passing percentage

80 60

VicRoads class 2 crushed rock FBS grading limits Laboratory -PSD Field-PSD

40 20 0 0.01

0.1

1 ParƟcle size (mm)

Fig. 1. Aggregate gradation curves

10

100

120

2.2

S. Pitawala et al.

Bituminous Binder and Foaming Agent

The foaming characteristic of bitumen is a critical parameter in FBS mix designs. A class 170 bitumen supplied by Puma Energy, Hamilton, Queensland was selected as the primary binder for both laboratory sample preparation and full-scale pavement construction for ALF testing, in accordance with Austroads (2017). HiFOAM-B developed by Pavement Technology Centre Pty Ltd was chosen as the foaming agent for the laboratory sample manufacture. The main reasons for choosing this foaming agent were explained by Pitawala et al. (2019). However, the foaming agent TERIC (Huntsman) was used for the field construction. The foaming characteristics are mainly evaluated using Half-life (HL) and expansion ratio (ER). The ER and HL values were found be 20 and 40 s for HiFOAM-B and 12 and 20 s for TERIC, respectively. For both foaming agents, the ER and HL values satisfied the minimum values specified by QTMR (2017), which are 10 and 20 s respectively for ER and HL. The optimisation of foaming characteristics using foaming agent HiFOAM-B was explained by Pitawala et al. (2019). 2.3

Mix Design of FBS Material Characteristics

The FBS mixture of laboratory manufactured and field extracted samples comprised of 3% bitumen and 2% lime by mass of aggregate. Water was added to 88% of the OMC of the aggregate to achieve the maximum dry density of the final mixture under modified proctor compaction (which is 2,110 kg/m3). The final mixture includes aggregate, water, lime, and foamed bitumen. In general, 85% of the OMC of aggregate under modified proctor compaction is used for field FBS pavement construction. Therefore, the water content used for pavement material compaction is consistent for both field and laboratory samples. The FBS pavement at constructed at the ALF site was compacted using a 16-tonne roller compactor and the laboratory specimens were compacted using British pendulum compactor, which simulates similar compaction as the field compaction. However, the field material was compacted to a greater degree and the dry density of field FBS material was found to be 2.28 kg/m3 while that of the laboratory compacted samples was found to be 2.14 kg/m3.

3 Experimental Results 3.1

Temperature and Frequency Sensitivity Evaluation Using Dynamic Modulus Master Curves

The dynamic modulus (|E*|) is commonly used for the characterisation of the time and temperature dependency of bituminous mixes and was assessed for both the laboratory and field FBS mixes. The master curve approach is the conventional way to analyse dynamic modulus of asphalt materials. The dynamic modulus master curves using the sigmoidal function fitted to the measured data through a regression method (NCHRP 2004). The modulus master curves were constructed for both field and laboratory manufactured FBS beam specimens using flexural modulus results under different

Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams

121

frequencies and temperatures, as depicted in Fig. 2. The temperature and frequency sensitivities of 3% bitumen with 2% lime FBS mixture from both field and laboratory, were evaluated. Both the field and laboratory manufactured specimens were tested in the same frequency range (2 Hz to 15 Hz) and the same temperature range (0 °C to 30 °C). It can be observed that the flexural modulus of the beam specimens decreases with increasing temperature for both cases. The gradient of the flexural modulus master curves indicates the temperature sensitivity and the frequency sensitivity of the material. Also, since it is possible to construct a flexural modulus master curve using the flexural modulus values at different frequencies and temperatures, it can be concluded that the flexural modulus variation with frequency has similar behaviour with change of temperature. According to Fig. 2, it is clear that the laboratory specimens with 3% bitumen and 2% lime have higher temperature sensitivity than that of field specimens for a given FBS mixture.

Fig. 2. Modulus master curve comparison for field and laboratory FBS specimens

The laboratory beams were cured for 3 months at room temperature after three days of initial accelerated curing at 40 °C before testing. The initial 3-days accelerated curing was assumed to be 6–12 months curing in the field as the flexural modulus increase in laboratory manufactured beams after three months can be considered negligible. 3 days curing at 40 °C is found to correspond to when the modulus begins to plateau. This corresponds to a similar point in the curing curve of the field extracted cores from the ALF site tested in indirect tensile modulus. The field beams tested in this study were cured for more than 20 months and the samples were extracted from the untrafficked area of the pavement section. The average flexural modulus of the field samples was found to be 3,515 MPa with only a 5% coefficient of variation. The average density of field samples was 6% higher than that of laboratory beam specimens. The higher flexural modulus of field samples is likely to be due to its high density. In addition, the curing time and curing mechanisms can also influence the

122

S. Pitawala et al.

flexural performance of FBS materials. The changes in density and flexural modulus can alter the temperature sensitivity of FBS specimens. Other possible reasons for lower temperature sensitivity of field samples are discussed in the next subsection. 3.2

Visual and Digital Microscopic Analysis of Laboratory Manufactured Beams and Field Extracted Beams

The fractured surfaces of the samples were observed visually and under a light microscope for both the laboratory manufactured and field extracted FBS beams. The visual macrographs and digital microscope images are shown in Fig. 3 and Fig. 4 for laboratory manufactured and field extracted beams, respectively. It is clear that the bitumen dispersion in field beams was not so fine compared to the laboratory manufactured beams. Possible reasons for poor bitumen dispersion throughout the aggregate skeleton include the achievement of poor foaming characteristics and different mixing after stabilisation in the field.

Fig. 3. Digital macrograph and optical microscopic images - laboratory manufactured beams

Fig. 4. Digital macrograph and optical microscopic images - Field extracted beam specimens

Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams

3.3

123

Comparison of X-Ray CT Images of Laboratory and Field Manufactured Beams

Both the field and laboratory manufactured specimens were scanned using a 3D X-ray CT tomography facility (Siemens Somatom) available at Monash University biomedical imaging. The specimens were scanned at approximately 300 microns resolution. After scanning, the digital radiographs were reconstructed using CSIRO XLI CT software to generate 2D image stacks. The 2D image stacks were then analysed using Avizo 9.7 software run in a supercomputing cluster (www.massive.org.au). After the image acquisition, a median filter was applied for noise reduction of the images. The images were then binarised to discrete phases (Bitumen mastic, coarse aggregate and voids). The bitumen mastic could be further segmented into bitumen, fine particles and voids within the mastic if the resolution was higher than 10 microns. Due to the relatively poor resolution, it was not possible to differentiate the bitumen, fine particles and voids inside the bitumen mastic separately. The binarisation step in the material segmentation is a crucial part in image analysis which has a major impact on the quantitative data of each discrete phase extracted from the image. The segmentation was manually conducted by analysing the pixel intensity. After the segmentation, the total material fraction of each component was then compared with known material quantities to validate the segmentation procedure. The objective of the scanning was to evaluate the bitumen dispersion and the coarse particle distribution within FBS specimens. Hence, the homogeneity between laboratory manufactured specimens and field extracted specimens can be assessed. 3.3.1 Laboratory Manufactured FBS Beams The 3D images constructed from the 2D image stacks were segmented for three components; voids, bitumen mastic (bitumen + fine particles), and coarse particles. After the segmentation, different materials were assigned with different colours as shown in Fig. 5. The beam specimen shown in Fig. 5 is a fractured beam. The red colour indicates the air voids and the fracture path is clearly detected. It should be noted that voids detected here are only the bulk voids and most of the voids are within the bitumen mastic which were not detected in this analysis.

Fig. 5. X-ray CT scanning images of laboratory manufactured FBS beams

124

S. Pitawala et al.

The material distribution along the longitudinal direction was first analysed after separating the specimen into two parts (upper and lower) as illustrated in Fig. 6.

Fig. 6. Volume fractions of a laboratory manufactured FBS specimen a.) Upper section b.) Lower section

As the FBS beam specimens were compacted in a single layer, more insight could be gained if further investigation was conducted to explore whether there was aggregate segregation or inhomogeneity in bitumen dispersion. Accordingly, the material distribution along the vertical axis was also examined as illustrated in Fig. 7. It can be seen there is no considerable variation either in coarse particle content or bitumen mastic dispersion in the vertical direction along the beam sample. Similar results were observed for all the laboratory manufactured samples scanned (3 specimens in total). Thus, it can be concluded that the compaction method developed by the Australian Road Research Board and used in this research can achieve uniform compaction although it was carried out in single layer using a modified British pendulum compactor.

Fig. 7. Material distribution along the vertical axis of the beam sample

Characterisation of Laboratory and Field Foamed Bitumen Stabilised Beams

125

3.3.2 Field Extracted FBS Beams The field extracted specimens were evaluated following the same procedure as for laboratory manufactured FBS beam specimens. The same colour code was used to indicate different phases within an FBS specimen as shown in Fig. 8.

Fig. 8. X-ray CT scanning images of field extracted FBS beams

The material distribution along the longitudinal direction was analysed for field extracted beams as shown in Fig. 9. The average volume fraction and standard deviation of coarse particles and bitumen mastic were displayed next to the volume fraction variations of each phase in Fig. 9. The standard deviation values of field extracted beams are higher in comparison with laboratory manufactured beams. This result indicates there is more inhomogeneity in field compaction than in laboratory manufactured beams.

Fig. 9. Material distribution along longitudinal axis of a field extracted beam

126

S. Pitawala et al.

4 Conclusions The performance of laboratory and field manufactured samples can differ due to various reasons, such as mix design differences, compaction method differences, different curing mechanisms and differences in bitumen dispersion throughout the aggregate skeleton. In this study, the bitumen dispersion, coarse aggregate distribution and density of both laboratory and field manufactured specimens were evaluated separately. The temperature and frequency sensitivities in both cases were assessed and the following conclusions were made based on the results obtained in this study. The bitumen dispersion and coarse aggregate distribution through the FBS beams are more consistent in laboratory manufactured beams in comparison with field extracted specimens. The temperature and frequency sensitivities are lower in field extracted beams compared to the laboratory manufactured beams. This was possibly due to the poor dispersion of bitumen throughout the aggregate skeleton in the field extracted beams compared to the laboratory manufactured specimens and their greater density.

References Alderson, A.: Fatigue Properties of Bitumen/Lime Stabilized Materials- Summary Report. City of Canning (2011) Austroads: Review of Structural Design Procedures for Foamed Bitumen Pavements. Austroads, Sydney (2011) Austroads: Guide to Pavement Technology Part 4D: Stabilised Materials, 2nd edn. Austroads, Sydney (2019) Jameson, G.: Design and performance of foamed bitumen stabilised pavements. Austroads Technical report. Austroads, Sydney. 310 p. (2018) Kendall, M., Baker, B., Evans, P., Ramanujam, J.: Foamed bitumen stabilisation. In: 1999 Proceedings Roads at Work-Developing Southern Queensland, Southern Region Symposium, Qld Department of Main Roads, Goondiwindi, pp. 1–4 (1999) NCHRP: Guide for mechanistic-empirical design of new and rehabilitated pavement structures (2004) Pitawala, S., Sounthararajah, A., Grenfell, J., Bodin, D., Kodikara, J.: Experimental characterisation of fatigue damage in foamed bitumen stabilised materials using dissipated energy approach. Constr. Build. Mater. 216, 1–10 (2019) QTMR: Testing of materials for foamed Bitumen Stabilisation, vol. 150, Brisbane. Australia (2017) Ramanujam, J., Jones, D.: Characterisation of foamed bitumen stabilization. In: Proceedings Road System and Engineering Technology Forum, pp. 1–22 (2000) Saleh, M.F.: Cost evaluation of foam bitumen and other stabilisation alternatives. Int. J. Pavement Eng. 8, 157–161 (2007) Su, N., Xiao, F., Wang, J., Amirkhanian, S.J.C., Materials, B.: Characterizations of base and subbase layers for mechanistic-empirical pavement design. Constr. Build. Mater. 152, 731– 745 (2017)

Effect of the Incremental Loading Conditions in the Permanent Deformation in Heavy Vehicle Simulator Tests Erdrick Pérez-González(&), Jean-Pascal Bilodeau, and Guy Doré Department of Civil Engineering, Université Laval, Quebec City, Canada [email protected]

Abstract. Load magnitude is a variable that needs to be adequately defined to simulate real pavement conditions in accelerated pavement tests (APT) sections. The performance of pavement structures under different loading conditions was evaluated. Two asphalt pavement structures were constructed and tested at Laval University (Quebec, Canada) using a heavy vehicle simulator (HVS). Seven load stages were defined considering different load magnitudes and water table levels, applying more than 182.000 cycle repetitions (unidirectional) with loads in the range from 40 kN to 80 kN on a half-single axle. Permanent deformations measured in the upper part of each layer, as well as at the surface of the pavements. Permanent deformation follows a trend related to the magnitude of the load, which was studied under the stress conditions defined during each load stage that was evaluated. The results show that permanent deformation follows an analogous behaviour under the same stress state in the granular and subgrade layers of the pavements studied. A reference for the normalization of this behaviour is presented in this paper. Keywords: Loading


Deformation
Performance
HVS

1 Introduction Pavement materials, including embankments and subgrade, are subjected to dynamic loads of different magnitudes. Pavement materials have been studied experimentally throughout the world, in full-scale models and on samples tested in the laboratory. Defining a stress state that is analogous to those expected in the field is fundamental in the analysis of damage associated with granular materials (i.e. permanent deformation). In real scale tests using heavy vehicle simulators (HVS), different load magnitudes can be used to emulate stress conditions or to reduce the test running time. In practice, it is common to associate the loading conditions defined for APT tests to repetitions of equivalent single axles load (ESAL). However, the equivalence between the effect of different stress levels applied during the same test can be challenging, even more, when the analysis seeks to isolate the impact of loads on a single layer of the pavement (i.e. insulating layer or new material).

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 127–136, 2020. https://doi.org/10.1007/978-3-030-55236-7_14

128

E. Pérez-González et al.

The objective of this paper is to present an assessment of the influence of the load defined in HVS tests on the evolution of permanent deformation in individual pavement layers using the stress state of the material as a reference.

2 Test-Site, Materials and Instrumentation Two flexible pavement structures were built in a concrete test pit inside Laval University APT testing facility. Both structures differ in the thickness of the asphalt concrete (AC) layer, identified as EB150 (AC thickness: 150 mm) and EB75 (AC thickness: 75 mm). The size of the test pit is 2  6  2 m3. Pavement sections were instrumented to monitor the horizontal deformations (longitudinal and transverse) at the bottom of the AC layer and vertical stresses, strains and water content in each unbound layer and the subgrade soil. Materials defined in the specifications of the province of Quebec (Canada) were used in this study. Detailed information on pavement structure and instrumentation layout are shown in Fig. 1.

Fig. 1. Cross-section of pavement structures and instrumentation layout

The resulting density of the materials was measured during construction (see Table 1). In the same way, the modulus of each of the materials was measured in the laboratory (see Table 2).

Effect of the Incremental Loading Conditions in the Permanent Deformation

129

Table 1. Density of materials Structure EB150 cd (kg/m3) cd /cmax AC-Surface 2315 95% AC-Bottom 2170 87% Base (MG20) 2065 95% Subbase (MG112) 2097 93% Subgrade (SM) 2011 92%

Structure EB75 cd (kg/m3) cd /cmax 2318 95% – – 2096 96% 2051 91% 1998 92%

In Table 1, cd is the density measured in the test section, and cmax is the maximum density of the material. For granular materials, cmax was obtained by following the specification ASTM D-1557 (D18 Committee, 2015), and for AC specification ASTM D-2041 (D04 Committee 2019) was used.

Table 2. Pavement material modulus Model

Layer



logjE j ¼ d þ

Mr ¼ k1 pa

1 þ eb þ cðlog fr Þ

 k2  I1 pa

a

soct pa

k 3 þ1

Material

Voids AC-Surface EB10-c 5.3% AC-Bottom EB10-c 11.8% k1 Base MG20 1683 SubBase MG112 1297 Subgrade SM 447

fr (s−1) 0.51 0.45 k2 0.732 0.702 0.740

|E*| (MPa) 3134 2116 k3 −0.144 −0.049 −0.160

Table 2 shows the summary result of the dynamic modulus of the asphalt concrete (|E*|) from the use of Witczak’s model (laboratory calibration) for the temperature and speed of the test (i.e. 9 kmh at 21 °C). It also shows the parameters associated with the Uzan model for the resilient modulus of unbound materials (ARA Inc. 2004; Doré and Zubeck 2009). For this study, the asphalt concrete (AC) layers were built in two sub-layers of 75 mm thick each. AC-Surface corresponds to the upper 75 mm (surface layer) of both structures, EB150 and EB75, and AC-Bottom corresponds just to the lower 75 mm of the asphalt concrete layer of the pavement structure identified as EB150. As can be noted in Table 2, AC-bottom showed high void content (11.8%), so it is expected that this layer is more susceptible to permanent deformation. Given this, the deformation corresponding to the asphalt layers was excluded from the analysis.

130

E. Pérez-González et al.

3 Methodology Two pavement sections were constructed using the planned instrumentation (see Fig. 1), then the HVS was positioned over the test pit, as shown in Fig. 2. The HVS has a set of dual tires and was set to unidirectional loading; a constant speed was used (i.e. 9 km/h) with a tire pressure of 690 kPa. The air temperature was controlled at 21 °C throughout the test. Total deformation at the surface of the pavement was measured using a portable beam equipped with a precision laser (±0.001 mm), readings were always taken on the same transverse lines throughout the tests, and the initial measurement was used as a reference. Pavement response (stress and strains) were collected periodically, with a higher frequency of measurements at the beginning of the cycles (load increase), and reducing the frequency progressively during each load stage. A loading protocol was defined to apply incremental loads, starting from the representation of an ESAL (80 kN on single axle, 40 kN for the half axle) up to twice this load (160 kN on single axle, 80 kN for the half axle). The different loading stages of the test are shown in Table 3.

Fig. 2. Instrumentation and HVS: a) sensors positioning during pavement construction, b) asphalt concrete layer surface, c) HVS in position Table 3. Loading protocol followed in the HVS test. Stage 1 2 3 4 5 6 7

Load (kN) 40 50 60 60 70 80 80

Full axle equivalence (kN) 80 100 120 120 140 160 160

Water table depth – – – 1m – – 1m

Number of passes 50000 50000 50000 20000 5000 5000 2250

ESAL’s 50000 122070 253125 101250 46895 80000 36000

Cumulative ESAL’s 50000 172070 425195 526445 573340 653340 689340

Effect of the Incremental Loading Conditions in the Permanent Deformation

131

Stages 4 and 7 (see Table 3) are excluded from the discussion, as they consider the saturation conditions of the lower layers, which implies the need for additional analyses that exceed the objective of this paper. Stress paths followed by the pavement materials during the tests were described to analyze the stress states in the different pavement layers. A stress path is a representation of the successive states of stress in a specimen during loading or unloading, and are defined by the deviator stress (q) and the mean stress (p), as follows: 0

0

q ¼ r1  r3 0

p¼

ð1Þ 0

r1 þ 2  r3 3

ð2Þ

Where r1 and r3 are major and minor principal stresses, respectively. Since there were no horizontal stress measurements during the testing, to calculate the confinement level (r3 ) needed to describe the stress states in the pavement (Eqs. 1– 2), the resilient modulus measured on site was used to calculate this variable. For granular and soil materials, the on-site resilient modulus was calculated with the following equation: Mr ¼

q e

ð3Þ

Where q = deviator stress (kPa), and e = resilient strain (le). Matching the on-site measurements (Eq. 3) with those described by the material in the laboratory (Table 2) and knowing the deviator stress (measured with vertical pressure sensors), it was possible to determine the value of r3 . Associating the stress state with the deformations measured on-site, the analysis of the results was performed.

4 Results and Discussion 4.1

Total Deformation, Discretized by Load Magnitude

Surface deformation (total deformation) was measured during this study (See Fig. 3a). The trend followed by the total deformation of the pavement shows an incremental behaviour at a constant rate, not being particularly sensitive to the increase of the halfaxle loads. Only an increase in the deformation rate after raising the water table is noticed. In the same way, the EB150 structure shows a higher accumulated deformation due to the high content of voids in the AC-bottom (see Table 2). The time hardening approach was used to discretize the behaviour defined with each load increment, using the following equation to fit the deformation at each loading stage (Rahman and Erlingsson 2015): ep ¼ aðN0 þ Nieq Þ

b

ð4Þ

132

E. Pérez-González et al.

Where ep is the permanent strain or deformation, N0 is the number of cycles in the stress path starting from zero, Neq i is an adjustment factor for cycles in the stress path i, and a and b are regression parameters. Calibration of Eq. 4 for each load level allows the subsequent simulation of complete load cycles on the pavement structure. Parameters for the surface deformation measured during the HVS test are presented in Table 4. Figure 4 shows the results for simulations at different load levels.

Table 4. Equation 4 parameters for pavement surface deformation EB75

a b Neq i R2 EB150 a b Neq i R2

40 kN .086 .300 0 .975 .170 .272 0 .990

50 kN .758 .122 41549 .987 .594 .178 18872 .967

60 kN .747 .140 67020 .861 .372 .226 56324 .958

70 kN .277 .264 41957 .605 .985 .145 123222 .769

80 kN .955 .173 12471 .956 .920 .196 9815 .892

Once the deformation model has been calibrated, considering the time-hardening, the equivalence between the permanent deformations caused at different load levels can be obtained using the following equation: Px dx ¼ Pref dref

ð5Þ

Where, Px and dx are the load and deformation obtained at a load x, and Pref and dref are the load and deformation using the reference load (40 kN for this study). Figure 3b shows the trend of Eq. 5. In the trend described in Fig. 3b, it can be seen that in both cases, the permanent deformation ratio increases linearly up to a Px/Pref = 1.5 (i.e. 60 kN), after which the ratio of deformation start to increase exponentially. This behaviour may also be due to the fact that the water table was raised in previous load cycles, even though it was removed in subsequent load cycles (as in the one under analysis), a residual effect remains.

Effect of the Incremental Loading Conditions in the Permanent Deformation

133

Fig. 3. (a) Surface deformation during HVS test, (b) Ratio of deformation (Pref = 40 kN)

40kN 10

50kN

60kN

70kN

80kN

EB150

8

Surface deformation (mm)

6 4 2 0 0 8

20

40

60

80

100

120

140

160

180

60

80

100

120

140

160

180

EB75

6 4 2 0 0

20

40

Number of passes (103)

Fig. 4. Time-hardening models for surface deformation of the pavement

134

4.2

E. Pérez-González et al.

Deformation in Layers, Discretized by Stress State

Stress paths followed by the different unbound materials during the HVS test were studied. Figure 5 shows the progression of the stress state during the test. Stresses measured at the top of the layers draw a straight line in a q-p space under an incremental loading condition. The increase in the applied load will proportionally increase both the deviator stress (q) and the confinement stress (r3) and the mean stress (p) as a consequence. The slope describing the stress path is defined by the material modulus and its Poisson ratio.

150

100

Subgra de EB7 5

q (kPa)

200

80 kN

Su bB as eE B1 50 Ba se EB 15 0

Subgr ade E B150 Sub Bas eE B75

250

70 kN 60 kN 50 kN 40 kN

75

se Ba

EB

Subgrade EB75 SubBase EB75 Base EB75 Subgrade EB150 SubBase EB150 Base EB150

50

0 0

200

400

600

800

1000

1200

1400

1600

p (kPa) Fig. 5. Stress paths defined by each layer during HVS test

To study the deformation associated with each layer of the pavement, a stress state representative of each load level on the stress path was defined. For this purpose, isolines were drawn over the stress path of each material under each level of loading (see Fig. 5). Data extrapolation was made to complete isolines of 70 kN and 80 kN for the SubBase of EB150 since, at these load levels, the signal received by the pressure sensors in this layer was erratic. The intersection between the isolines of the load levels and the stress paths for each material is considered a stress state representative of the average deformation measured during each load level. Normalization of the stress state, as well as the deformation in each of the layers, was done by dividing the variables under study (q, p and ep ) by the value described by the reference load (qref, pref and epðref Þ ). In this study, the conditions defined by a load of 40 kN are taken as reference, as it represents half of an equivalent single axle load of 80 kN. Figure 6 shows the equivalence between permanent deformations resulting from increased loads. This figure groups the normalization of the response of the different materials and layers tested during the HVS test.

Effect of the Incremental Loading Conditions in the Permanent Deformation

135

Fig. 6. Equivalence between permanent deformations ðep =epðrefÞ Þ

The ep =epðrefÞ trend described by the materials studied during the HVS testing (see Fig. 6) can be described using Eq. 6: ep epðref Þ

¼ ð0:241  b þ 1:281Þ px =pref


0:727

ð6Þ

With, b¼

px =pref qx =qref

ð7Þ

Where ep is the permanent deformation, p is the mean stress, and q is the deviator stress. Subindices x and ref are associated with the load in analysis and the value with the reference load (40 kN), respectively.

5 Conclusions Equivalence between the permanent deformation (total and per layer) in the pavement at different load levels was studied. 40 kN was used as a reference for normalization since it corresponds to the ESAL, which is a commonly used indicator in practice. The time hardening approach was used to study the effect of the load magnitude on the development of permanent deformation. This approach reveals that the ratio between total deformation caused by different load magnitudes follows a linear trend

136

E. Pérez-González et al.

up to a defined magnitude (60 kN in this study), after which the rate of deformation begins an exponential trend. Permanent deformation equivalence for each pavement layer, concerning the deformation caused by a reference load, can be defined through the stress state in the material of the layer. A normalization based on the stresses defined under the application of a reference load allows establishing a unified criterion for the effect of load magnitude on the permanent deformation. An equation describing the equivalent permanent deformation as a function of stress state in the material was presented in this paper. Stress states are a rational reference in the definition of load magnitudes during HVS test planning. Acknowledgment. The authors wish to acknowledge to the partners of Phase 2 of the NSERC industrial research Chair on the interaction of heavy loads, climate and pavements of Université Laval (Chair i3C) for the support given to this research.

References ARA Inc.: Guide for the Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures (Final Report NCHRP 1-37A). Transportation Research Board of the National Academies (2004) D04 Committee: Test Method for Theoretical Maximum Specific Gravity and Density of Bituminous Paving Mixtures. ASTM International (2019). https://doi.org/10.1520/D2041_ D2041M-19 D18 Committee: Test Methods for Laboratory Compaction Characteristics of Soil Using Modified Effort (56,000 ft-lbf/ft3 (2,700 kN-m/m3)). ASTM International (2015). https://doi. org/10.1520/D1557-12 Doré, G., Zubeck, H.K.: Cold Regions Pavement Engineering. ASCE Press, McGraw-Hill (2009) Rahman, M.S., Erlingsson, S.: A model for predicting permanent deformation of unbound granular materials. Road Mater. Pavement Des. 16(3), 653–673 (2015). https://doi.org/10. 1080/14680629.2015.1026382

Comparison of In-situ Response from Companion Asphalt Pavement Sections in Different Climates Michael Vrtis1(&), Benjamin Worel1, and David H. Timm2 1 Office of Materials and Road Research, Minnesota Department of Transportation, 1400 Gervais Avenue, Maplewood, MN 55109, USA {michael.vrtis,ben.worel}@state.mn.us 2 Department of Civil Engineering, Auburn University, 238 Harbert Engineering Center, Auburn, AL 36849, USA [email protected]

Abstract. Full-scale flexible pavement test sections were constructed as part of a national cracking group experiment on the NCAT Test Track (Auburn, Alabama) and at MnROAD (Monticello, Minnesota). These sections were intended to provide field performance data for comparison with various laboratory cracking performance tests to improve testing and cracking specifications. The sections were designed with various cracking potential and thus have differing stiffness and dynamic responses under loading. Embedded response instrumentation (i.e., strain gauges and pressure cells) were installed at both sites to evaluate in situ pavement response. The instruments, data collection and processing procedures were kept as uniform as possible to facilitate valid comparisons between sites. This paper compares dynamic pressure response measurements from similar pavement sections constructed in different climates at two accelerated pavement testing facilities. The measured pavement temperature was used to normalize the pavement response data to a reference temperature. As expected, the measured pressure response was affect by the asphalt stiffness and depth of the sensors. The coordination between testing facilities allowed for a more direct comparison of pavement response data and can be improved upon in future investigations to provide further direct comparisons between additional facilities. Keywords: Pavement instrumentation
In-situ characterization subgrade
Accelerated pavement testing


Base and

1 Background and Introduction In 2015 the National Center for Asphalt Technology (NCAT) and the Minnesota Department of Transportation (MnDOT) began a national pooled-fund study to investigate laboratory cracking tests. The objective of this Cracking Group (CG) study was to determine which cracking test would provide the best indication of field performance for each location’s climatic demands. Seven different laboratory cracking tests were included and the tests were evaluated based on repeatability, equipment © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 137–146, 2020. https://doi.org/10.1007/978-3-030-55236-7_15

138

M. Vrtis et al.

requirements, sample preparation demands, and matching of field performance. To facilitate the field performance comparison, asphalt concrete test sections were designed with varying cracking potentials and were constructed on the NCAT Test Track (TT) and the Minnesota Road Research Facility (MnROAD) in 2015 and 2016, respectively. The TT sections were focused on top-down cracking (TDC) and the MnROAD sections were focused on low-temperature cracking (LTC). MnROAD is a cold climate pavement testing facility that is owned and operated by MnDOT. MnROAD is located parallel to I-94 in Monticello, MN and diverts interstate traffic for loading of the pavement test sections. MnROAD has been in operation for over 25 years and has been influential in the implementation of Superpave mixtures in the Midwest United States, unbonded concrete overlays, and in low temperature cracking performance of asphalt pavements (Worel and Van Deusen, 2015; Worel et al. 2008). The NCAT TT is a closed-access 2.7 km loop located in Opelika, AL. The TT is loaded by five triple-trailer trucks. Each axle is loaded to the legal US axle limit of 9,000 kg (20,000 lbs) resulting in a total vehicle weight of around 68,000 kg (150,000 lbs). The trucks operate 16 h per day, five days a week. The TT has been in operation since 2001 and has been influential in revising material specifications, proving out innovative technologies and developing perpetual pavement concepts (Brown et al. 2002; Timm et al. 2006; West et al. 2018; West et al. 2019). While the two facilities have pavement research in common, Minnesota and Alabama have vastly different climates. In Minnesota, the average January low and high temperatures are −15 °C and −5.6 °C, respectively. In Alabama, the average January low and high temperatures are 2 °C and 14 °C, respectively. The July (summer) temperatures are more similar with Alabama averaging a low of 21 °C and a high of 34 °C while Minnesota averages a low of 17 °C and high of 29 °C (rSSWeather.com). Minnesota averages 77 cm of precipitation and Alabama averages 134 cm of rainfall (US Climate Data). The differing climates result in different environmental demands experienced by the pavements. One of the common criticisms regarding full-scale accelerated pavement testing (APT) facilities is that the implementable results are limited to the specific region and climate that the facility is located in. This is evident in the typical regional nature of funding sources for testing facilities. The instrumentation, data collection, and processing procedure utilized at both facilities was carefully coordinated to facilitate a more direct comparison of in-situ response from the CG test sections and provide results that were applicable to a greater audience.

2 Objective and Scope The objective of this paper is to compare in-situ pressure response measurements collected at the NCAT TT and MnROAD to provide insights into the as-built mixture properties and the long-term performance of the test sections. The pressure response will be compared to climate and expected stiffness of the asphalt.

Comparison of In-situ Response from Companion Asphalt Pavement Sections

139

Similar instrumentation arrays were installed at each facility and a common data processing scheme was used. Operating differences between the facilities created some differences in the analysis approach however effort was taken to limit the differences as much as possible.

3 Section Description and Instrumentation Layout 3.1

NCAT Test Track

Seven test sections were constructed as part of the CG on the NCAT TT in summer 2015. Four of the seven CG sections are discussed in this paper, the excluded sections were testing the effect of construction variables and ground tire rubber. The sections were constructed with 17 cm of hot mix asphalt (HMA) over 15 cm of crushed granite granular base over the native TT subgrade material, as shown in Fig. 1(a). The TT sections were designed to evaluate the HMA’s resistance to TDC and the ability of the various laboratory tests to predict the TDC field performance. The HMA was constructed in three layers with only the mix in the top 4 cm layer varying with the different TDC susceptibly mixes. The two bottom HMA layers were designed at 6.5 cm thick and used a highly modified asphalt (HiMA) binder that has performed well in previous TT cycles (Timm et al. 2013). The HiMA was intended to provide sufficient stiffness to limit flexural strain at the bottom of the HMA and flexibility to achieve good fatigue characteristics. A description of the surface lift used in each of the four TT sections is shown in Table 1, along with the continuous PG of the extracted binder (Pressure Aging Vessel aged) from each mix. It can be seen that the surface mixes featured various recycled materials and PG binders. These results are based on plant-produced mixes collected during construction at each facility. Asphalt strain gauges (ASG) were installed at the bottom of the HMA and Geokon 3500 Dynamic Soil Pressure Cells were installed in the top of the base and subgrade layers. The plan view layout of the instrumentation array is shown in Fig. 1(b). Thermistors were installed to capture temperature at various depths. Hourly maximum, minimum, and average temperatures are stored. Unfortunately, the majority of ASGs used in this study at the TT did not survive construction despite following the procedure used in numerous previous NCAT ASG successful installations (West et al. 2018).

140

M. Vrtis et al.

Fig. 1. (a) Cross section of NCAT TT sections with instrumentation locations (b) Plan view of TT instrumentation

Table 1. Section description and extracted PG Facility Section Description NCAT TT – RAP Stockpile N1 20% RAP N8 20% RAP + 5% RAS S5 35% RAP, PG 58-28 S6 20% RAP HiMA All TT CG Base/Intermediate Lifts MnROAD – RAP Stockpile – RAS Stockpile 17 10% RAP + %5 RAS 19 20% RAP, 3.0 design air 21 20% RAP, PG 58H -34 23 15% RAP, HiMA

3.2

Extracted PG, °C 112.0–13.8 88.6–16.6 107.3–5.44 82.8–23.0 101.4–21.5 102.3–28.8 86.5–19.8 120.7–23.0 73.2–26.2 70.8–25.8 70.2–30.3 72.0–31.7

MnROAD

Eight CG test sections were constructed at MnROAD, four of which were instrumented with strain gauges and soil pressure cells. 13 cm of HMA were built over 30 cm of MnDOT Class 5 base over 30 cm of MnDOT Class 3 sub-base. The MnROAD sections were constructed in three lifts with the same mix design. 18 cm of MnDOT Select Granular material was placed to keep the native, clay-loam subgrade material from intruding into the sub-base. The same type and manufacturer of soil pressure cells were used at MnROAD and NCAT TT. It can be seen in Fig. 2(a) that the pressure cells at MnROAD were installed only at top of the Class 3 layer. Thermocouples were used to monitor temperature at

Comparison of In-situ Response from Companion Asphalt Pavement Sections

141

various depths in the pavement structure. Temperature readings were recorded every 15 min. Similar to the TT instrumentation array, the MnROAD instrumentation is centered on the outside wheelpath and had additional lines of sensors at 61 cm offsets.

Fig. 2. (a) Cross section of MnROAD sections with instrumentation locations (b) Plan view of MnROAD instrumentation

4 Data Collection and Processing High-speed pavement response data from ASGs and pressure cells were collected on a weekly basis at the TT. The collections alternated between morning and afternoon to characterize pavement responses at different points in the diurnal cycle. Collection was conducted during the TT’s normal trucking operations when the trucks were travelling at 70 kmh. Data collection at MnROAD was conducted on a monthly basis from March through November; dynamic testing was not done in winter months when the pavement structure was frozen. The responses from a five axle MnROAD tractor trailer, loaded to 36,000 kg, travelling at 8 kmh and 56 kmh are recorded during each collection. Raw voltage versus time responses from both facilities were processed using a custom program template using DADiSP software. The program operated by using the responses from the pressure cells to establish the vehicle speed. Using the sensor spacing, vehicle speed and axle spacing, the program waas then able to capture the local max-min for each gauge and each axle. More information regarding processing in DADiSP has been documented by Timm (Timm 2016). The use of the same processing program ensured that the responses were being measured and characterized in the same manner.

142

M. Vrtis et al.

5 Pressure Response Results The TT responses from the earthen pressure cells versus mid-depth asphalt temperature are shown in Fig. 3 (a) and (b) for the TT subgrade and base, respectively. Each data point is the 95th percentile of the maximum response from single axle loadings on a given collection date from October 2015 through November 2017. To account for asbuilt thickness variations, the base pressure values have been normalized to a HMA thickness of 17 cm and the subgrade pressure values have been normalized to a thickness of 32 cm (HMA + base). It can be seen that both the base and subgrade pressure were affected by temperature of the HMA. As expected, the base pressure had a higher magnitude, as it is closer to the surface. All coefficient of determination (R2) values are greater than 0.90 indicating that the exponential trend-lines accurately characterize the response with temperature. The pressure values presented match the extracted binder stiffness presented in Table 1. Following mechanistic theory, Section N8 is expected to have the lowest pressure values as it was significantly stiffer than the other sections. Stiffer HMA will widen the stress zone within the HMA layer thus reducing the vertical compressive stress at the bottom of the HMA (Huang, 2004). N8 had the lowest subgrade pressure and the lowest base pressures, as indicated by the y-intercept value for the exponential trend-line shown. The pressure values were normalized to a reference temperature of 20 °C following the method used previously by Vargas and Timm (2013) and are shown versus time in Fig. 4. The linear trend lines show that there is very little change over time for these sections.

Fig. 3. (a) TT subgrade pressure versus temperature (b) TT base pressure versus temperature

Comparison of In-situ Response from Companion Asphalt Pavement Sections

143

Fig. 4. (a) TT subgrade pressure at 20 °C versus date (b) TT base pressure at 20 °C versus date

The MnROAD pressure versus temperature is presented in Fig. 5. There are fewer data points from MnROAD as data were only collected monthly during months when the pavement structure is not typically frozen (March through November). An additional filter was used to remove data points when the base or subgrade was partially frozen during March, April, and November collections. The maximum tandem axle response was used. The MnROAD pressure data presented was the maximum response per truck pass from tandem axles on the middle of the outside wheelpath pressure cells. The magnitude of the responses are lower for the MnROAD sections primarily due to the depth of the sensor (TT at 33 cm under surface and MnROAD at 43 cm under surface). The MnROAD sections have a lower R2 that is likely attributed to fewer data points and less influence of HMA temperature due to depth under the surface. Despite the lower R2 values, a relationship between extracted binder stiffness and pressure is still evident. Section 23 had the lowest PG low temperature grade and had the largest pressure values. The result for 23 is similar to the TT HiMA result for the base layer pressure in which the HiMA has the greatest pressure values indicating a less stiff behavior. The pressure values at a reference temperature of 20 °C versus date are shown in Fig. 5 (b). There is an increasing trend in pressure for Sections 17 and 19 which would indicate the HMA or base is losing stiffness. Section 17 has cracking within the instrumentation array but the nearest cracks are over 61 cm (2 ft.) from the pressure plates. Another contributing factor to the trends in Fig. 5 (b) is influence of freezing within the pavement structure. The circled data points for Section 17 are from March and April and show the erratic nature of responses that still exist in the data despite removing data that had thermocouple readings below 0 °C anywhere in the pavement structure.

144

M. Vrtis et al.

Fig. 5. (a) MnROAD subgrade pressure versus temperature (b) MnROAD subgrade pressure at 20 °C versus date

6 Field Performance Results To complete the field performance portion of the CG experiment, monitoring and loading of the test sections at both facilities are continuing. TDC has been observed in several of the TT sections. Forensic cores were taken to verify that the cracking was initiated from the surface. Section N1 had 10.3% of the lane area cracked; N8 had 16.6%. Sections S5 and S6 did not have any cracking. The TT sections had 10 million ESALs applied during the timeframe presented in this paper. Cracking has been observed in all MnROAD sections however not all sections have developed LTC that the experiment intended to investigate. Previous MnROAD LTC experiments have shown higher amounts of LTC than have occurred and have shown that LTC does not always occur in the first several winters (Marasteanu et al. 2007) however those some of those sections were constructed before the PG binder system. Cracking has developed within the instrumentation arrays of Sections 17, 19, and 23. All sections had cracking along the longitudinal construction joint between the driving lane and shoulder that is within 100 cm of the pressure cells. The full impact of the cracking on the pressure responses is still not completely understood as 17 showed an increase in pressure in 2019 but 19 and 23 did not. Section 21 also had an upward trend in pressure, similar to 17, but does not have any cracking in the instrumentation array.

7 Summary and Conclusions Based on the work presented in this paper the following conclusions were made. • Coordination and planning in the experimental design phase facilitated a useful comparison of in-situ pavement response between MnROAD and NCAT TT. Similar coordination between APT facilities will improve the usability of pavement response data and allow for projects with broader impact.

Comparison of In-situ Response from Companion Asphalt Pavement Sections

145

• Low temperatures at MnROAD limit the amount of useable data that can be collected and analyzed without the influence of frozen layers. • Pressure cells showed reasonable agreement with mixture properties, especially with mixtures than were significantly different that the others, such as RAP/RAS or HiMA. • Continued performance monitoring and pavement response data needs to be collected to validate trends occurring with traffic and temperature. Acknowledgements. The authors would like to acknowledge the instrumentation teams at each facility including Robert Strommen, Dan Roushar, Dave Van Deusen, and Leonard Palek at MnROAD and Brian Waller and Matthew Sasser at NCAT. Their efforts installing and maintaining the instrumentation networks are critical the data quality and analysis presented in this paper. The CG experiment was funded by MnDOT, Alabama Department of Transportation, Alabama Department of Environmental Management, Georgia Department of Transportation, South Carolina Department of Transportation, Tennessee Department of Transportation, Kentucky Transportation Cabinet, Oklahoma Department of Transportation, Missouri Department of Transportation, Maryland Department of Transportation, New York Department of Transportation, Illinois Department of Transportation, Michigan Department of Transportation, and Wisconsin Department of Transportation.

References Brown, E.R., et al.: NCAT Test Track Design, Construction, and Performance. National Center for Asphalt Technology, Auburn (2002) US Climate Data: US climate data (n.d.). https://www.usclimatedata.com/climate/auburn/ alabama/united-states/usal0035. Accessed 24 Nov 2019 Huang, Y.H.: Pavement Analysis and Design, 2nd edn. Pearson Education Inc., Pearson (2004) Marasteanu, M., et al.: Investigation of Low Temperature Cracking. Minnesota Department of Transportation, St. Paul (2007) rSSWeather.com: Climate for Minneapolis-St.Paul, Minnesota (n.d.). http://www.rssweather. com/climate/Minnesota/Minneapolis-St.Paul/. Accessed 24 Nov 2019 Timm, D., et al.: Phase II NCAT Test Track Results. National Center for Asphalt Technology, Auburn (2006) Timm, D., et al.: Field and Laboratory Study of High-polymer Mixtures at the NCAT Test Track: Final Report. National Center for Asphalt Technology, Auburn (2013) Timm, D.H.: Key concepts in dynamic signal processing from instrumented pavement sections. In: 5th International Conference on Accelerated Pavement Testing, San Jose, Costa Rica, 19– 21 September 2016 (2016) Vargas, A., Timm, D.: Physical and Structural Characterization of Sustainbable Asphalt Pavement Sections at the NCAT Test Track. National Center for Asphalt Technology, Auburn (2013) West, R., et al.: Phase V (2012–2014) NCAT Test Track Findings. National Center for Asphalt Technology, Auburn (2018) West, R., et al.: Phase VI (2015-2017) NCAT Test Track Findings. National Center for Asphalt Technology, Auburn (2019)

146

M. Vrtis et al.

Worel, B.J., Jensen, M., Clyne, T.R.: Economic benefits resulting from road research performed at MnROAD. In: Third International Conference on Accelerated Pavement Testing, Madrid, Spain (2008) Worel, B., Van Deusen, D.: Benefits of MnROAD Phase-II Research. Minnesota Department of Transportation, St. Paul (2015)

Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate via Full-Scale Accelerated Testing Xingdong Wu1(&), Mustaque Hossain1, and Greg Schieber2 1

2

Department of Civil Engineering, 1701C Platt St., 2118 Fiedler Hall, Manhattan, KS 66506, USA {chinawu,mustak}@ksu.edu Bureau of Construction & Materials, Kansas Department of Transportation, 700 SW Harrison St., # 420, Topeka, KS 66603, USA [email protected]

Abstract. Tack coat is a thin, bituminous material layer which may be applied to an existing pavement surface to promote interface bonding between new and existing pavement components. This ensures a monolithic structure that carries traffic loads and releases thermal stress. Insufficient tack coat application can result in inadequate interface bonding, delamination, or debonding within pavement layers; conversely, excessive tack coat application can cause weak interface bonding, rutting, or bleeding. A slow setting anionic polymer-modified emulsion (SS-1hP) with a standard application rate of 0.23 L/m2 is commonly used as the tack coat for hot-mix asphalt (HMA) pavement overlay. To investigate optimum tack coat rates, a 6.1 m  4.3 m. HMA section was divided in this study into four test sections and coated with 50%, 100%, 160%, and 240% of 0.23 L/m2 tack rate and overlay. The test sections were then subjected to 2 million load repetitions on an accelerated pavement testing machine with an 89 kN single-axle load. The transverse rutting profile and strain at the overlay interface were measured at intervals of 100,000 load repetitions. Laboratory and in-situ pull-off tests were conducted after 1.5 million load repetitions. The results indicate that the test section with a tack rate of 240% has the smallest rut depth, while the 160% sample shows the greatest strain. However, the test section with a tack rate of 100% (0.23 L/m2) showed optimal performance based on in-situ strain and in-situ bond strength tests. Keywords: HMA overlay
SS-1hP
Accelerated pavement testing
Interface bonding

1 Introduction Tack coats are emulsions which are obtained from an asphalt binder by mixing an emulsified agent with water. When the tack coat is sprayed on clean pavement at ambient temperature, the water evaporates from the emulsion and a thin “bonding” layer of asphalt globules forms on the pavement surface. This thin layer promotes

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 147–156, 2020. https://doi.org/10.1007/978-3-030-55236-7_16

148

X. Wu et al.

bonding between the asphalt lifts and ensures the pavement structure’s performance and durability. The tack coat emulsion must be applied appropriately to construct (or rehabilitate) hot-mix asphalt (HMA) pavements. Applying an adequate tack coat emulsion provides a reliable bond between HMA pavement lifts resulting in a monolithic pavement structure that carries traffic loads over a long service life. Conversely, insufficient tack coat application can lead to premature structural failures like debonding and mat slippage; excessive tack coat application can lead to fatigue cracking, bleeding, and rutting. The tack coat emulsion application rate is one of the most important acceptance criteria to ensure adequate bonding between HMA layers. In Kansas, the Kansas Department of Transportation (KDOT) generally uses a slow-setting anionic polymer-modified emulsion (SS-1hP) with a recommended application rate of 0.23 L/m2 for HMA pavement construction and rehabilitation (KDOT 2007). Recently, engineers in Kansas have noticed premature cracking on newly overlaid HMA pavements. It is yet unclear whether this cracking was caused by inadequate application of SS-1hP. To mitigate this problem, it is necessary to determine the optimal SS-1hP application rate. The objective of this study is to optimize the SS-1hP application rate and evaluate the corresponding HMA pavement performance comprehensively. An accelerated pavement testing (APT) machine was used to program a full-scale test to evaluate the performance of various SS-1hP tack coat application rates. A direct tension pull-off test was used to determine the in-situ interface bond strength. A similar test was conducted in the laboratory, and the results were compared with those from the field tests to validate them.

2 Study of HMA Interface Bond Strength Various HMA interface bond strength evaluation test methodologies have been developed over recent decades. There have been very few previous studies centered on APT. Romanoschi (1999) proposed a direct shear test method in which a cylindrical specimen is placed between two metal cups as such that the interface is positioned in the middle of the cups. This test reveals the relationship between the interface shear strength and the normal stress as-obtained by regression analysis. A pull-off test was developed at the University of Texas at El Paso that can be used to measure the tensile strength of a tack coat before a new overlay is placed, as well as to identify the quality of the tack coat in situ (Deysarkar 2004). An HMA interlayer bonding study on a porous asphalt course interface was conducted by Italian researchers wherein a tack coat was applied at the interface of an existing porous asphalt layer and a newly laid open-graded course. The various tack coat application rates achieved acceptable interlayer bonding, but higher application rates generated some scattering in the results (Canestrari et al. 2005). Researchers at the Illinois Center for Transportation used an APT machine to study the interface bonding between HMA overlays and Portland cement concrete (PCC) pavement. The results indicated that SS-1hP has higher interface bond strength than RC-70, and the 0.23 L/m2 tack coat application rate produced the highest HMA

Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate

149

interface shear strength (Leng et al. 2008). Mealiff (2014) used an APT machine to study HMA pavement rehabilitation with various application rates of SS-1hP and Emulsion Bonding Liquid (EBL). The EBL test section showed lower rutting depth than the SS-1hP sections and all sections showed a similar pavement cracking after 400,000 repetitions of a 89-kN load. The cracking was observed to move top-down through the samples. The ASTM standard test “Standard Test Method for Pull-Off Strength of Coating Using Portable Adhesion Tester” was modified by KDOT to evaluate in-situ bond strength (Rahman 2010). This modified test method was applied to measure the tensile force that causes interface bonding failure. The testing results can be recorded as an interface bonding pass/fail system or by recording the tensile force to split the bonded layers (Mealiff 2014). It was adopted to assess in-situ bond strength in tack coat materials as introduced by the Kansas Standard Test Method, KT-78. This test methodology is currently used for determining the tensile adhesive strength of asphalt pavement tack coat at Kansas (KDOT 2007).

3 In-Situ Test Construction and Tack Coat Layout In this study, an existing 6.10 m by 4.3 m HMA pavement test pit was milled to 50 mm. The milled surface was thoroughly brushed with a power broom and then cleaned with compressed air. The test pit was divided into four test sections, each 3.05 m by 2.13 m. The sections received various tack rates of SS-1hP, as shown in Fig. 1. The four tack coat application rates tested represent 50%, 100%, 160%, and 240% of the current KDOT-recommended rate of 0.23 L/m2. The 50% rate mimics the loss of tack materials due to pickup by HMA delivery trucks after spraying tack at 0.23 L/m2. The heavier tack rate of 160% simulates a situation in which, even after tack pickup by the trucks, the tack rate remains close to 100% (0.23 L/m2). A handpumped pressure sprayer was used to apply the four application rates of SS-1hP. Spray rates were verified by measuring the weight of tack applied with even coverage. To monitor the strain level, eight H-bar strain gages (PML-60-2L, Tokyo Sokki Kenkyujo Co. Ltd.) were installed at the interface between the HMA overlay and the milled asphalt surface. Each strain gage was epoxied to two notched aluminum rods (100 mm  13 mm  6 mm), then the gage and bars were secured to the milled surface with metal staples. The stock lead wires were replaced with high-thermal resistance shielded wiring to better tolerate the heat of the fresh HMA mixture. The soldered connection at the base of the strain gage was also covered by heat shielding. Two strain gages were installed per section in between the wheel paths, the strain gages parallel to the wheelpath (Fig. 2). A compact DAQ (cDAQ) system from National Instruments (NI) was used to collect strain data throughout the test.

150

X. Wu et al.

The test sections were paved with a 12.5-mm Nominal Maximum Aggregate Size (NMAS) Superpave mixture containing 25% recycled asphalt pavement (RAP) (SR12.5A) and PG 58-28 binder. A steel-drum vibratory roller was used to compact the 50 mm overlay at a target 7% in-situ air voids. A nuclear density gage was used to check the HMA mat density after rolling.

Fig. 1. SS-1hP tack coat test section layout

Fig. 2. a) Strain gage installation b) H-bar strain gage orientation

4 APT Programing An APT machine at the Kansas State University (KSU) Civil Infrastructure Systems Laboratory (CISL) was used to simulate realistic pavement loading to examine interface bond performance. The APT load assembly hydraulically placed load to a single axle with dual tires moving across the test pit at about 11.3 km/h. A 89-kN bidirectional loading wheel was rolled back and forth accounting for one repetition. A protective housing insulated the system and enabled air conditioners to maintain steady temperatures throughout the test. This custom-built machine also introduces wander and deviates up to 150 mm either direction with a truncated normal distribution to simulate realistic load application. Figure 3 shows the APT machine and a complete wander cycle consisting of 676 load repetitions.

Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate

151

Fig. 3. a) KSU’s APT machine b) Distribution of 676-load repetitions

Sections were programmed as a full-scale test at a constant temperature of 25 °C under dry conditions. Since the pavement was constructed in pits and the HMA overlay mixture was paved wall-to-wall, the moisture content in the subgrade soil remained constant during the test. This was verified by the lack of change in volumetric moisture content measured by the time domain reflectometry (TDR) gages installed in the subgrade. The transverse rut depth of each section was measured intermittently using a Chicago digital linear variable differential transformer (LVDT) with a roller attachment at the end and mounted to a level aluminum beam. To ensure a consistent location unaffected by loading, two transverse profiles were taken for each section with respect to the referenced point on the concrete floor. The transverse rut depth was derived from the elevations at every 12.5 mm measured across the entire section. The APT machine was run at a constant of 89 kN loading in increasing increments with periodic strain measurements. The sections were also inspected at fixed intervals for any signs of cracking. At the conclusion of the experiment, 1.5 million APT load repetitions had been applied to all sections.

5 Interface Bond Strength Test Interface bond strength was determined by conducting both field (in-situ) and laboratory pull-off tests. The field pull-off test was conducted after 1.5 million APT loading repetitions, where three 50 mm diameter cores were drilled to just below the interface layer of each section near to the wheel path. The cores were tested in situ using a portable, battery-powered ComTen pull-off tester in accordance with the KT-78 procedure. Aluminum pucks were epoxied to the 50 mm diameter cores and pulled in direct tension at 20 mm/min. Figure 4 shows the field pull-off testing preparation and setup.

152

X. Wu et al.

Fig. 4. a) pull-off test cores (100% SS-1hP) b) cores with steel pucks c) cored samples after test

The laboratory pull-off test was conducted in a manner similar to that specified in KT-78 for the field pull-off test. One 150 mm diameter core was extracted from the outer wheel path of each section, then three 50 mm diameter cores were taken and epoxied to a 76 mm by 76 mm steel plate. The specimen was mounted inside the chamber of a universal testing machine (UTM-25) and a constant loading rate at 20 mm/min with a constant temperature of 25 °C. Figure 5 shows the laboratory pulloff test setup.

Fig. 5. Laboratory UTM-25 pull-off tester

6 Results and Discussion 6.1

Permanent Deformation and Rut Depth

Permanent deformation was defined as the difference between identical elevation points at various times (or load repetitions), and was computed and plotted as shown in Fig. 6. The largest difference among identical elevation points is the reported rut depth value, which can be attributed to a lack of noticeable heaving on the sections. Figure 7 shows the rut depth progression on the test sections from initial to 1.5 million load repetitions. SS-1hP sections with 50% and 100% of the tack rate required by KDOT (0.23 L/m2) exhibited the most intense deformation over time up to 800,000 repetitions. The 100%

Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate

153

Fig. 6. Original and permanently deformed profiles after 1.5 million repetitions

Fig. 7. SS-1hP rut depth after 1.5 million repetitions

and 160% tack rate Sects. (0.23 and 0.38 L/m2, respectively) performed similarly with respect to rutting after 1.5 million repetitions. The 240% tack rate Sect. (0.54 L/m2) performed better in terms of rutting. 6.2

Resilient Strain

In this study, although eight gages were installed (two per section), only six gauges were operational after construction potentially due to soldering issues and stresses from the construction equipment. None of the gages on the 240% tack rate section survived, likely due to the heavy tack application causing shorting out. The results shown in Fig. 8 are the median strain values for the 90 s intervals aligned with the baseline for that unit of time. This strain was measured for an entire wander cycle (or 676 load repetitions). The top 50 strain values for each given wander cycle were averaged and plotted as shown in Fig. 8. The strain level at the bottom of the HMA layer is related to the fatigue performance of the HMA pavement. A lower strain level indicates better fatigue performance in the HMA layer; a higher strain value indicates that the test section is more fatigue-prone.

154

X. Wu et al.

The results show that the section with the KDOT-specified tack rate (0.23 L/m2 of SS-1hP) had the lowest strain among all sections after 0.9 million APT load repetitions. It had the best fatigue performance but did not last throughout the overall load repetitions. The section with 160% tack rate (0.38 L/m2) showed a large increase in strain after loading began and another jump in strain after 400,000 repetitions, likely due to slippage. Overall, the section with 160% tack rate (0.38 L/m2) shows the highest strain level and thus the worst fatigue performance. A large increase in strain after 400,000 and 900,000 repetitions was also observed in the section with 50% tack rate (0.12 L/m2). This section has the lowest strain level after 900,000 load repetitions, but the strain reading plot trends indicate relatively unstable behavior. Although the section with 100% SS-1hP (0.23 L/m2) showed one sudden spike at 700,000 repetitions and gradual increases between 900,000 and 1.3 million repetitions, the strain reading after this point returned to a low, stable value.

Fig. 8. Strain during APT loading

6.3

Bond Strength

In-field pull-off tested cores from the sections with 50%, 100%, and 160% tack coat application are shown in Fig. 9. Both pull-off tests were repeated if one of the cores split in the asphalt parts rather than the interface layer. All tested samples were inspected to ensure the interface bonding was 100% failed.

Investigation of Optimum Hot-Mix Asphalt Pavement Overlay Tack Coat Rate

155

Fig. 9. Filed pull-off test results s) 50% b) 100% c) 160%

The results of the laboratory and field pull-off tests were programed into a statistical analysis of analysis of variance (ANOVA) using IMB SPSS Statistics (Version 24.0) software. The results indicated no statistical differences among test results for each section. Figure 10 shows the results of the laboratory and field pull-off tests. The field pull-off results indicate that the KDOT-specified 100% SS-1hP tack rate (0.23 L/m2) has the highest bond strength among test sections, closely followed by the 50% tack rate. The 240% tack rate (0.54 L/m2) shows the lowest bond strength.

1600 1400

Laboratory pull-off test

a

1200

Peak stress (kPa)

Field pull-off test

a a

1000

a

a

a

a a

800 600 400 200 0 50

100 160 Residual application rate (%)

240

Fig. 10. Bond strength after 1.5 million repetitions

A similar trend was observed in the laboratory test results. The KDOT-specified 100% SS-1hP tack rate (0.23 L/m2) shows the highest bond strength, closely followed by the 50% tack rate (0.12 L/m2). The 240% (0.54 L/m2) tack rate shows the lowest bond strength among the test sections.

156

X. Wu et al.

7 Conclusions In this study, a full-scale test was conducted using an APT machine to estimate the optimum HMA overlay tack coat rate. Laboratory and in-situ pull-off tests were conducted to evaluate various SS-1hP application rates. The conclusions can be summarized as follows. • Full-scale test results indicated that the KDOT-recommended rate of 0.23 L/m2 (100%) performs well in terms of the resilient strain, but poorly in terms of rutting and permanent deformation of HMA pavement after a certain amount of loading repetitions. The application rate of 0.54 L/m2 (240%) performed best in terms of both rutting and permanent deformation. • Laboratory and field pull-off tests produced closely correlated results. • The KDOT-recommended rate of 0.23 L/m2 performs well as a tack coat material based on the in-situ bond strength, and laboratory bond strength observed in this study.

References KDOT: Kansas Department of Transportation special provisions to the standard specifications, Topeka, Kansas (2007) Romanoschi, S.A.: Characterization of pavement layer interfaces, Ph.D. Dissertation, Louisiana State University. Baton Rouge, Louisiana (1999) Deysarkar, I.: Test set-up to determine quality of tack coat, Master’s thesis, University of Texas at EI Paso. EI Paso, Texas (2004) Canestrari, F., Ferrotti, G., Partl, M.N., Santagata, E.: Advanced testing and characterization of interlayer shear resistance. J. TBR 1929, 69–79 (2005). https://doi.org/10.1177/ 0361198105192900109 Leng, Z., Ozer, H., Al-Qadi, I.L., Carpenter, S.H.: Interface bonding between Hot-Mix Aspahlt and various Portland cement concrete surface. J. TRB 2057, 46–53 (2008). https://doi.org/10. 3141/2057-06 Mealiff, D.: Evaluation of interface bond of Hot-Mix Asphalt overlay, Master’s thesis, Kansas State University. Manhattan, Kansas (2014) Rahman, F.: Performance evaluation of 4.75-mm NMAS Superpave mixture, Ph.D. Dissertation, Kansas State University, Manhattan, Kansas (2010)

Experimental Methods for Material Selection, Quality Control, and Forensic Investigations of Asphalt Paving Materials Tianhao Yan1, Jhenyffer Matias de Oliveira1, Mugurel Turos1, Mihai Marasteanu1(&), Michael Vrtis2, and Dave van Deusen2 1

University of Minnesota, Twin Cities, 500 Pillsbury Drive SE, Minneapolis, MN 55455, USA [email protected] 2 Minnesota Department of Transportation, 1400 Gervais Ave, Maplewood, MN 55109, USA

Abstract. The Minnesota Road Research Project (MnROAD) and the National Center for Asphalt Technology (NCAT) have formed a partnership to execute asphalt mixture performance testing experiments with a nationwide implementation impact. As part of this partnership, a pooled-fund study, called MnROAD Cracking Group (CG) experiment, was conducted to identify laboratory experiments that can best address low-temperature cracking performance. The construction of the test cells at MnROAD was done in 2016 and original binders and loose mix were collected and used to prepare testing specimens for laboratory experiments. In this paper, the viability of using three test methods for asphalt mixtures and one test method for asphalt binders in the material selection process, quality control, and forensic investigations of asphalt paving materials is discussed. These test methods are the Bending Beam Rheometer (BBR) for creep and strength of asphalt mixtures; low temperature SCB (Semi-Circular Bending) fracture testing for asphalt mixtures; |E*| (Dynamic Modulus) testing of asphalt mixtures using the IDT (Indirect Tensile) configuration; and BBR strength testing of asphalt binders. First, the materials used are described and the test methods are discussed. The experimental results are presented and statistical tools are used to identify significant factors in predicting low temperature cracking resistance of the set of asphalt materials used in the CG experiment. Conclusions are drawn based on preliminary field performance data. Keywords: Asphalt mixture Test methods


Asphalt binder
Low temperature cracking


1 Introduction Low temperature cracking is the main distress in asphalt pavements located in cold temperature regions. Many test methods have been developed to evaluate the low temperature cracking resistance of asphalt materials. In this paper, the viability of four low temperature cracking resistance tests is investigated. The test methods are the Bending Beam Rheometer (BBR) for creep and strength of asphalt mixtures; low © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 157–166, 2020. https://doi.org/10.1007/978-3-030-55236-7_17

158

T. Yan et al.

temperature SCB (Semi-Circular Bending) fracture testing for asphalt mixtures; |E*| (Dynamic Modulus) testing of asphalt mixtures using the IDT (Indirect Tensile) configuration; and BBR strength testing of asphalt binders. First, the materials used are described and the test methods are discussed. The experimental results are presented and statistical tools, including analysis of variance (ANOVA) and Tukey’s method, are used to identify significant factors in predicting low temperature cracking resistance of the set of asphalt materials used in the CG experiment. Based on the results from the statistical analyses and preliminary field performance data, recommendations are made regarding the use the experimental methods investigated.

2 Materials The Minnesota Road Research Project (MnROAD) and the National Center for Asphalt Technology (NCAT) have formed a partnership to execute asphalt mixture performance testing experiments with a nationwide implementation impact. The construction of the test cells at MnROAD was done in the summer of 2016 and original binders and loose mix were collected and used to prepare testing specimens for laboratory experiments. Table 1 summarizes the information of the eight mixtures.

Table 1. Information of the Asphalt mixtures investigated Cell no. Binder 16 17 18 19

PG PG PG PG

64S-22 64S-22 64S-22 64S-22

RAP % 20 10 20 20

RAS % 5 5 0 0

TotalAC % Virgin AC % NCAT Mix ID 5.27 5.43 5.43 5.70

3.17 3.94 4.20 4.46

30–40% ABR with RAS 20–30% ABR with RAS 20% ABR 20% ABR 100 gyration, 3.0% air void, 100 Ndes 20 PG 52S-34 30 0 5.32 3.47 35% ABR with PG 52S-34 21 PG 58H-34 20 0 5.38 4.15 20% ABR with PG 58H-34 22 PG 58H-34 20 0 5.73 4.5 20% ABR with LMS 23 PG 64E-34 15 0 5.23 4.31 20% ABR with PG 64E-34 Note: The data listed are the percentages by the total weight of mixture. RAP and RAS stand for reclaimed asphalt pavement and reclaimed asphalt shingle. AC stands for asphalt content. ABR stands for asphalt binder replacement. LSM stands for large molecular size binder.

3 Test Methods and Results Four low temperature cracking resistance tests were investigated in this study. They are the BBR creep and strength test of asphalt mixtures; low temperature SCB fracture testing for asphalt mixtures; |E*| testing of asphalt mixtures using the IDT configuration; and BBR strength testing of asphalt binders.

Experimental Methods for Material Selection, Quality Control

159

The BBR creep and strength tests of mixture were performed according to the procedure proposed by Marasteanu et al. (2012). The results of this test include the creep stiffness, m-value, strength, and failure strain. Tests were conducted at three temperature levels, 0, −12, and −24 °C. Due to the limitation of space, only the results of creep stiffness and strength are shown in Fig. 1 and Fig. 2 respectively.

20000

Creep Stiffness @60s, MPa

15000

0C

10000

-12C

5000

-24 C

0 cell16 cell17 cell18 cell19 cell20 cell21 cell22 cell23 Fig. 1. BBR creep stiffness at 60 s of all mixtures.

10.0

Strength at Failure, MPa

8.0

0C

6.0

-12C

4.0

-24 C

2.0 0.0 cell16 cell17 cell18 cell19 cell20 cell21 cell22 cell23 Fig. 2. BBR strength of all mixtures.

The SCB fracture tests were performed according to AASHTO TP-105 (2013). Tests were performed at two temperatures, −12 and −24 °C. Tests results include fractural energy and fractural toughness. The results of fractural energy are shown in Fig. 3.

160

T. Yan et al.

Fracture Energy 0.7 0.6

-12°C

-24°C

Gf (kJ/m2)

0.5 0.4 0.3 0.2 0.1 0.0 Cell 16

Cell 17

Cell 18

Cell 19

Cell 20

Cell 21

Cell 22

Cell 23

Fig. 3. SCB fracture energy for all mixtures.

The IDT |E*| tests were performed according to the procedure proposed by Kim et al. (2004). Tests were conducted at 12 °C. The results are shown in Fig. 4.

|E*| Master Curves (Reference Temperature 12 C) 100

|E*| (GPa)

10 Cell 16 Cell 17 Cell 18 Cell 19 Cell 20 Cell 21 Cell 22 Cell 23

1

0.1

0.01 1.E-06

1.E-04

1.E-02

1.E+00 1.E+02 1.E+04 Reduced Frequency (Hz)

Fig. 4. IDT |E*| master curves of all mixtures.

1.E+06

1.E+08

Experimental Methods for Material Selection, Quality Control

161

The BBR creep and strength tests of binder were performed according to the studies of Marasteanu et al. (2017), and Matias et al. (2019). Tests were performed at two temperatures, the PG low temperature plus 4 °C (PGLT + 4) and the PG low temperature plus 10 °C (PGLT + 10). The creep stiffness and strength of the binders are shown in Fig. 5 and Fig. 6, respectively. Since cells 16, 17, 18 and 19 used the same binder, only the cell 16 binder was tested.

Average Creep Stiffness, MPa @ 60sec PGLT+10C CoV_PGLT+10C

PGLT+4C CoV_PGLT+4C

600

15%

500

12%

400

9%

300 6%

200

3%

100 0

0% 16

20

21

22

23

Fig. 5. BBR creep stiffness at 60 s for binders from Cell 16, 20–23.

Average Binder BBR Strength, MPa PGLT+10C CoV_PGLT+10C

PGLT+4C CoV_PGLT+4C 50%

8

40%

6

30% 4 20% 2

10%

0

0% 16

20

21

22

23

Fig. 6. BBR strength for binders from Cell 16, 20–23.

162

T. Yan et al.

4 Data Analysis Statistical analyses are performed to identify significant factors. The tools used include analysis of variance (ANOVA), Tukey’s method, which represents a pairwise comparison technique, and correlation matrices based on Pearson’s correlation. 4.1

Analysis of Variance

One-way ANOVA was performed to determine the statistical significance of the mixture properties. The significance level (a) was set at 0.05. The null hypothesis (H0) assumes that all the sample means are equal. The alternate hypothesis (Ha) states that at least one of the sample means is different. An example is given in Table 2 that shows the ANOVA results for the BBR Creep Stiffness at −12 °C. Table 2. Summary and ANOVA results for creep stiffness at 60 s at −12 °C Summary Groups Cell 16 Cell 17 Cell 18 Cell 19 Cell 20 Cell 21 Cell 22 Cell 23 Anova Source of variation Between groups Within groups Total

Count

Sum

6 5 5 6 5 5 5 5

40.08251 34.7625 35.19701 42.14436 22.49815 26.10168 32.7937 29.15639

SS 32.2618 17.2756 49.5374

df 7 34 41

Average 6.680419 6.952501 7.039402 7.024061 4.49963 5.220336 6.558741 5.831278

Variance 0.519472 0.570816 1.194998 0.575539 0.562762 0.396015 0.122235 0.10331

MS F P-value F crit 4.6088 9.0706 2.82E-06 2.2938 0.5081

The parameters in Table 2 are: Sum of Squares (SS), degrees of freedom (df), Mean Square (MS), F-value, P-value and F-critical. The degrees of freedom are obtained between groups (number of groups minus one) and within groups (number of total samples minus the number of groups). The Sum of Squares is calculated by adding the squared differences between the individual responses and the mean. The Mean Square is calculated dividing the Sum of Squares by the degrees of freedom and the F-value is calculated as the ratio of between-groups mean square to within-groups mean square (Oehlert 2000). The F-value is greater than the F-critical value and the p-value is smaller than the alpha level selected (0.05). We can conclude that there is enough evidence to reject the

Experimental Methods for Material Selection, Quality Control

163

null hypothesis and say that at least one of the eight cells has significantly different means and belongs to a different population. To identify these differences, Tukey’s was performed. 4.2

Tukey Analysis

Since ANOVA only indicates if one or more mixtures have different means, it is necessary to run an additional test to find out the specific differences. Tukey’s method, a pairwise comparison technique, was chosen because it constructs simultaneous confidence intervals for differences of all pairs of means and controls the probability of making one or more Type I errors (Oehlert 2000). The confidence intervals and boxplot were generated for all mixtures and corresponding properties. The boxplot provides a visual interpretation of the confidence intervals in which mixtures are grouped according to their means; mixtures with the same color and letter belong to the same group. Figure 7 shows an example of the Tukey and boxplot results.

Fig. 7. (a) Confidence intervals (Tukey) for BBR creep stiffness at −12 °C.

5 Discussion Through the ANOVA and Tukey analyses, the mixtures can be grouped by different tests at different temperatures. The results are summarized in Fig. 8.

164

T. Yan et al.

Temperature

Mixture BBR Creep Stiffness (60s), GPa

Mixture m-value (60s) Mixture BBR Strength, MPa

Cells 16

17

18

19

20

21

22

23

-12C -24C -12C -24C -24C

Mixture BBR Strain @ Failure, %

-12C

Fracture Energy, kJ/m2

-12C

Fig. 8. Summary of boxplots for mixture results

As shown in Fig. 8 the testing methods investigated provide repeatable results that follow trends similar to the one observed using traditional methods. The results also show that the properties are highly temperature dependent and the ranking observed at one temperature can change at a different temperature. In addition, it is observed that materials with similar rheological properties, such as complex modulus absolute value |E*|, creep stiffness S, and m-value, do not necessarily have the same fracture resistance. These results confirm one more time the need for a fracture/strength test for correctly evaluating cracking resistance of asphalt materials. As also shown in Fig. 8, in general, the mixtures used in this experiment have similar properties, which may indicate similar service performance. The only exceptions appear to be the mixture from cell 20 that has the highest RAP content and the mixture from cell 23 that contains a highly modified binder; for most properties evaluated, the two mixtures were each grouped separately from the other mixtures. The results also indicate that, for some properties, there is no clear separation between the mixtures prepared with the PG-22 binder and the mixtures prepared with the PG-34 binders due to the addition of RAP and RAS in the mix design.

6 Preliminary Field Performance Data The MnRoad test cells 16 to 23 are located on the mainline of I-94 (westbound). Each test cell has a width of 38 feet (11.6 m) and a length of 500 feet (152.4 m). The average traffic volume on those cells was 696059 ESALs (equivalent single axle loads) per year after 2016. The average lowest air temperature in winter is approximately −30 °C. The field performance of these sections has been closely monitored since their construction in 2016, and the results are summarized in Table 3. There is less than

Experimental Methods for Material Selection, Quality Control

165

5 mm of rutting in each section. The IRI values are below 95 in/mile and they have stayed consistent since construction. A wide array of cracking has developed in several of the sections, however, there is very little traditional low temperature transverse cracking that the experiment was intended to investigate. Cell 17 has the greatest amount of transverse cracks observed, while Cell 20 has zero transverse cracks. All sections have longitudinal construction joints (both centerline and driving lane-edge) cracking that were sealed in October 2019. Sections 17, 18, and 23 have significant amounts of fatigue cracking with pumping of base materials through the asphalt concrete. MnROAD is utilizing other tools to forensically identify the cause/and severity of cracking not related to material selection, such as: paver-mounted infrared cameras during construction, falling weight deflectometer testing, instrumentation embedded in the pavement structure (strain gauges, pressure cells, thermocouples, and moisture sensors), pavement cores, ground penetrating radar, and visual distress surveys. Although there is a large amount of cracking in these sections, the IRI is steady and in the “Good” category. Table 3. Preliminary field data Cell number

16 17 18 19 20 21 22 23

Load related Fatigue Longitudinal wheel path (m) (m2) 5.3 36.3 62.5 24.1 33.7 26.5 1.6 10.7 1.3 3.4 7.9 14.3 21.3 78.6 135.4 101.8

Construction related Center line Shoulder joint joint (m) (m) 121.9 137.2 149.3 152.4 152.4 152.4 141.4 133.2 52.4 0.0 152.4 20.7 152.4 152.4 152.4 152.4

LTC Transverse (m) 17.7 21.3 10.7 18.6 0.0 8.5 15.2 13.1

7 Conclusions In this study, four tests were investigated to evaluate the performance of asphalt pavements at low temperatures. It was found that: 1. The testing methods investigated provide repeatable results that follow trends similar to the one observed using traditional methods. 2. The properties are highly temperature dependent and the ranking observed at one temperature can change at a different temperature. 3. It was observed that materials with similar rheological properties, such as complex modulus absolute value |E*|, creep stiffness S, and m-value, do not necessarily have

166

T. Yan et al.

the same fracture resistance. These results confirm one more time the need for a fracture/strength test for correctly evaluating cracking resistance of asphalt materials. 4. Preliminary field data confirms the general conclusion that cells should have similar performance. The exception is Cell 23 for which performance may have been affected by other factors that are not related to asphalt material mechanical properties.

References AASHTO TP105: Standard Method of Test for Determining the Fracture Energy of Asphalt Mixtures Using the Semicircular Bend Geometry (SCB). AASHTO, Washington DC (2013) Oehlert, G.: A First Course in Design and Analysis of Experiments. W. H. Freeman (2000) Kim, Y., Seo, Y., King, M., Momen, M.: Dynamic modulus testing of Asphalt concrete in indirect tension mode. Transp. Res. Rec. 1891, 163–173 (2004) Marasteanu, M., Falchetto, A.C., Turos, M., Le, J.L.: Development of a simple test to determine the low temperature strength of asphalt mixtures and binders (No. NCHRP IDEA Project 151) (2012) Marasteanu, M., Ghosh, D., Cannone Falchetto, A., Turos, M.: Testing protocol to obtain failure properties of asphalt binders at low temperature using creep compliance and stress-controlled strength test. Road Mater. Pavement Des. 18, 1–16 (2017) Matias De Oliveira, J., Yan, T., Turos, M., Ghosh, D., Van Dusen, D., Marasteanu, M.: Simple method to evaluate strength and relaxation properties of asphalt binders at low temperature. Transp. Res. Rec. 2673, 492–500 (2019)

APT of Portland Cement Concrete

Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced with Metal Fibers Alfonso Pérez, Federico Castro(&), Miguel Ángel Franco, and Paul Garnica Pedro Escobedo, Mexico {IMT-aperez,IMT-fcastro,IMT-mfranco, IMT-pgarnica}@imt.mx

Abstract. Whit the emergence of new materials and additives to improve the performance of road infrastructure it is necessary to evaluate, compare and validate the contribution of new materials in pavement performance. This document presents the evolution of the structural performance of a natural scale section constructed of concrete. The section is divided in two areas, the first area constructed with reinforced concrete with metal fibers and the second area constructed of plain concrete. The main purpose of the experiment is the knowledge about the contribution of the metallic fiber to the pavement performance. To accelerate the deterioration of the pavement, both areas were exposed to an accelerated application of damage with the HVS (Heavy Vehicle Simulator) device. To determine the structural evolution of the pavement, surface deformation sensors were installed and deflectometry tests were performed to determine the load transfer between slabs, maximum deflection, back calculation of elasticity modulus and others parameters, the evaluation of pavement performance was made every 500,000 equivalent axles of 80 kN. Keywords: Reinforced concrete performance


Metal fibers
Accelerated test
Structural

1 Background Mexico is the first commercial partner of the United States as a result of an intense and dynamic industrial activity, and the transportation of goods and services that is generated is done in 80% by the national road network. This implies that the volume and weight of cargo vehicles is very high, often overweight, which has caused a challenge in the preservation of pavements, trying to balance the needs of transport companies and the need for structural reinforcement of structures in a large proportion of the more than 40,000 km that make up the main road transport corridors. Due to the above, the almost obsessive search for greater durability of the pavements by the responsible authorities, public or private, has generated a very competitive market of products, additives and in general of new materials that offer to increase the performance of asphalt and hydraulic concrete pavements. © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 169–177, 2020. https://doi.org/10.1007/978-3-030-55236-7_18

170

A. Pérez et al.

In the case of concrete pavements, the use of metallic fibers as reinforcement of concrete is increasingly used in Mexico. The most common applications are the reinforcement of industrial floors, commercial floors, airport runways, tunnel alignments and slope stabilization through shotcrete. According to the specialized literature in different applications, the metal fibers have shown increases in the ductile behavior of concrete, so that, in many cases, incorporating metal fibers allows to reduce the thickness of concrete slabs. Nevertheless, the main objective is to extend the useful life of the concrete elements with better service conditions, cost reduction, reduction of conservation and repair actions. It is in this context that a company expert in industrial concrete floors and steel fiber distributor, requested to the Mexican Ministry of Communications and Transportation, that these be evaluated for use in concrete pavements of the federal highway network with the offer that with a lower slab thickness can be obtained a performance comparable to a standard concrete pavement. To demonstrate this, the Mexican Transportation Institute was asked to conduct a comparative study of two short sections of concrete pavement using an accelerated test of these pavements with the heavy vehicle simulator (HVS) available at our facilities. The following describes the experiment carried out and the main results obtained.

2 Construction For the accelerated pavement test, the construction of a hydraulic concrete Sect. 36 meters long and 6 meters wide was carried out, the section was divided in two areas with the thicknesses and dimensions shown in the Fig. 1. The slab thicknesses shown there were established by the federal highway department authorities, with the agreement and acceptance of the client of the study, in order to have challenging test conditions to the slab with the steel fibers. 18 meters

18 meters 0.25 meters

Plain Concrete

0.20 meters

Hydraulic Base

0.30 meters

Subgrade

Concrete Reinforced with Metal Fibers

0.18 meters

Hydraulic Base

0.20 meters

Subgrade

0.37 meters

Natural Soil Fig. 1. Sketch of a longitudinal cutting of Structural sections of the evaluated pavements

The construction of the lower layers, subgrade and hydraulic base, was carried out simultaneously; In the case of the concrete layer, this was done in two stages, the first stage was constructed the area of plain concrete and the area of the reinforced concrete with fiber was made in a second stage. The manufacturer of the metallic fibers made the design. The modulation of the slabs corresponds to square slabs of three meters of side,

Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced

171

with dowel bars placed in a transversal way. Ready-mixed concrete was used to construct the slabs, in the case of plain concrete a rupture modulus of 48 kg/cm2 was used and a modulus of 35 kg/cm2 was used for the area of fiber reinforced concrete. This contrast between the values of the rupture modules was also on purpose. The dosing and mixing of the fiber with the concrete was carried out on site using a proportion of 3 kg of fiber per cubic meter of concrete (10% to 30% of normal dosage), Fig. 2 shows the dosage of the metallic fiber, the location of dowel bars and placement of the concrete.

Fig. 2. Construction of the section for the accelerated pavement test: a) Placing concrete and dowel bars, b) field dosing of the metallic fibers in the ready concrete mix truck

3 Quality Control During Construction As part of the quality control during the construction, tests were carried out to ensure the quality and to know the final state of the construction, tests were carried out to determine the elastic modulus for each layer that conforms to the pavement, for this the GeoGauge equipment was used which provides information that allows to know the final state of the construction of each layer of the pavement, to later perform analysis and relate the performance of the pavement with the quality of construction, Fig. 3 shows the modulus of elasticity measured with GeoGauge device for Subgrade layer and hydraulic base layer. In addition to quality control tests during construction, laboratory tests were performed to determine the modulus of rupture in beams taken from the ready-mix concrete, Table 1 shows the results obtained in the laboratory.

172

A. Pérez et al. Subgrade Layer

30 25

Length (meters)

Hydraulic Base Layer MPa. < 50 50 – 62 62 – 74 74 – 86 86 – 98 98 – 110 > 110

20 15

35

20

5

5

3

4

< – – – – – >

60 72 84 96 108 120 120

15 10

2

60 72 84 96 108

25

10

1

MPa.

30

Length (meters)

35

5

Width (meters)

a)

1

2

3

4

5

Width (meters)

b)

Fig. 3. Elasticity modulus for the different layers obtained with GeoGauge device: a) elasticity modulus of subgrade, b) elasticity modulus of hydraulic base

Table 1. Modulus of rupture obtained in the laboratory with 4 points loading. Sample Plain concrete CS1 CS2 Reinforced concrete with metal fibers CRF1 CRF2

Modulus of rupture (kg/cm2) 49.8 47.1 36.9 40.1

4 Experiment Features Using the Heavy Vehicle Simulator (HVS) device (Fig. 4), the controlled application of damage was carried out, establishing the load, direction and speed of application of the loads. For this experiment, the application of loads was carried out in a bidirectional way with a total of 322,640 repetitions of a 80 kN load on a dual tire and speed of 8 km/h, which is equivalent, with the law of the fourth power, to a little more than 5 million equivalent axles of 80 kN that are used as reference in Mexico. The load was applied to the side of the longitudinal joint, considering this the most unfavorable condition for the pavement; Fig. 5 shows the effective test area and the point of load application during the experiment.

Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced

173

Fig. 4. View of the heavy vehicle simulator (HVS) on the test sections already built

Effective Test Length = 15.0 meters

Effective Test Area

Tire Footprint

Longitudinal Joint

Fig. 5. Test section and load application area

5 Structural Performance To monitor the performance of the pavement sections, structural evaluations were agreed with the client with an FWD equipment every 500,000 repetitions of the 80 kN equivalent axle, as well as the load transfer efficiencies at the transverse joints. In addition, sensors were placed on the surface of each section to continuously measure vertical deflections and tensile/compression strains (Fig. 6).

a)

LVDT arrangement

b) Strain gage

Fig. 6. View of sensors to continuously measure vertical deflections and tensile/compression strains

174

5.1

A. Pérez et al.

Structural Evaluation with Deflectometry

From the data obtained with an impact deflectometer FWD, the evolution of maximum deflection measured at the center of the slab for both concrete areas was determined, in the graph shown in Fig. 7 the deflections are presented for each area the maximum deflection. It can be seen that the structural capacity throughout the time of the test remained nearly constant, although the deflections were always greater in the fiberreinforced concrete section and this was a direct result of the lower thickness of the slab. However, in the steel fiber pavement section the deflections increase slightly after the application of 1.5 million of equivalent axles, which could mean a certain change in the structural capacity of the structure. In addition to the maximum deflections, the estimation of the load transfer efficiency was made evaluating with impact deflectometer on the edge of the slabs, on the transverse joint. The relationship between the deflection determined in the unloaded slab with respect to the loaded slab represent the load transfer efficiency, the evolution of the load transfer with respect to the increase in the number of equivalent axles are presented in Fig. 8. Good levels of joint transference remained constant over time in the two sections tested.

Fig. 7. Evolution of maximum deflection measured at the load application point for the plain concrete region and the concrete section reinforced with metal fibers

Fig. 8. Evolution of load transfer efficiency in evaluated at every 0.5 million equivalent axles of 80 kN, for plain concrete and concrete reinforced with metal fiber.

Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced

175

With the data obtained from the deflectometry tests, the back-calculated elasticity modulus for each type of concrete evaluated were obtained, which can be seen graphically in Fig. 9 that represents the evolution of the parameter in function of the equivalent axles accumulated during the experiment. An increase in time of the retrocalculated modules is observed up to a value of the order of a million equivalent axes, possibly still related to the hardening of the concrete.

Fig. 9. Back-calculated elasticity modulus of the slab for the area of plain concrete and the area of reinforced concrete

Similarly, the evolution of the modulus of reaction of the subgrade is shown in Fig. 10. This is consistent with the structural evaluation of any concrete pavement where thickness is a major contributor to the value of the back-calculated parameters.

Fig. 10. Backcalculation subgrade reaction modulus calculated from the deflections obtained with an impact deflectometer for intervals of 0.5 million equivalent axes.

5.2

Surface Deflections and StrainsMeasured During Load Applications

To better determine the performance and structural evolution of a pavement, it is necessary to have more information, so two instrumented beams were installed to measure in a continuous way the deflection suffered by the plain concrete area and the reinforced concrete area during the load applications. Figure 11 shows the evolution of the maximum deflection values and Fig. 12 the evolution of maximum strain values.

176

A. Pérez et al.

Fig. 11. Evolution of maximum surface deflection with the number of equivalent axles of 80 kN applied to the pavement.

Fig. 12. Evolution of maximum strains with the number of equivalent axles of 80 kN applied to the pavement.

3.5 M -7m

-6m

-5m -4m

-3m

-2m

-1m 0m

1m

2m

3m

4m

5m

6m

7m

4.0 M 1.0 M

0.5 M

Fig. 13. Final pattern of cracking in the test sections.

A classical analysis with the Kenslabs software (Huang 2004), considering a slab model supported on a Winkler foundation allows to obtain the maximum values of vertical deflections and deformations indicated in Figs. 11 and 12. In general there is a good agreement between the measured and calculated values. Figure 13 shows the final pattern of cracking in the test sections.

Accelerated Pavement Test to Evaluate the Performance of Concrete Reinforced

177

6 Conclusions The design of accelerated pavement tests must respond to the needs and requirements of the road agencies in each country. In Mexico, a unique equipment like the Heavy Vehicle Simulator (HVS) is asked to give quick answers to the projects proposed by the interested clients, which finally are the ones that allow financing the studies like the one presented in this article. That is why in each project it is necessary to contrast more than one variable at a time and to seek the best form of analysis. In this particular case, in the two concrete pavement sections evaluated, there were changes in thickness and in the properties of the concrete resulting from the addition of steel fibers. In the end, the evolution of all measured parameters and the good comparison with the theoretical calculations, seems to demonstrate that the client of the study overestimated the potential benefits of adding steel fibers to the concrete and offer that with lower slab thickness the same performance could be obtained compared to a conventional concrete pavement. At the time of writing this article, another similar study is already underway but with the use of acrylic fibers. This will be another story.

References Huang, Y.: Pavement Analysis and Design, 2nd edn. Pearson Prentice Hall (2004) U.S. Department of Transportation, Federal Aviation Administration: Use of Nondestructive Testing in the Evaluation of Airport Pavements, Washington, DC (2011) U.S. Department of Transportation, Federal Aviation Administration: Airport Pavement Design and Evaluation, Washington, DC (2016)

The Challenges of the Accelerated Testing of Jointed Concrete Pavements Angel Mateos2(&), Rongzong Wu1, John Harvey1, Julio Paniagua1, Fabian Paniagua1, and Robel Ayalew1 1 2

Pavement Research Center, University of California, Davis, USA Pavement Research Center, University of California, Berkeley, USA [email protected]

Abstract. One of the main challenges of the accelerated pavement testing (APT) is reproducing the distress mechanisms that will cause the structural failure of the pavement in the field. This challenge is particularly difficult for jointed concrete pavements since some of their main distresses are driven by critical combinations of concrete hygrothermal conditions, slab’s support, and traffic loading that are very difficult to reproduce during the accelerated testing. Two such distresses are top-down cracking, either transverse or longitudinal, which is driven by simultaneous loading at distant locations, and faulting, which is driven by fines pumping produced by wheels moving at relatively high speed. Furthermore, while traffic loading is the main action that can be relatively easily controlled in the accelerated pavement testing, temperature and moisture-related shrinkage actions, just by themselves or in combination with traffic loading, result in slabs deformations and concrete tensile stresses that cannot be ignored. This paper presents a discussion of these and other limitations of the accelerated testing of jointed concrete pavements. The discussion is supported by modeling results and by experimental data collected during the testing of concrete pavements with the Heavy Vehicle Simulator in California. Keywords: Jointed plain concrete pavement
Temperature gradients
Drying shrinkage
Top-down cracking
Faulting
Ambient environment action

1 Introduction The accelerated pavement testing (APT) can be defined as the controlled application of wheel loading to pavement structures for the purpose of simulating the effects of longterm in service loading conditions in a compressed time period (Hugo and Martin 2004). This is done while environmental effects on the pavement are typically controlled and measured (Steyn 2012). The goal of the simulation of the long-term in service loading conditions is to reproduce the distress mechanisms that will cause the structural failure of the pavement in the field. This is why, for example, the evaluation of the rutting performance of asphalt materials is typically conducted at high temperatures (Wu et al. 2008), the evaluation of subgrade rutting performance is typically conducted with a near-surface water table (Mateos and Soares 2014), the simulation of bottom-up asphalt fatigue cracking typically requires testing of thin flexible pavements © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 178–185, 2020. https://doi.org/10.1007/978-3-030-55236-7_19

The Challenges of the Accelerated Testing of Jointed Concrete Pavements

179

(Mateos et al. 2011), etc. Similar challenge is applicable to the APT of jointed concrete pavements. The main structural distresses that the jointed concrete pavements undergo in the field are cracking and faulting (AASHTO 2015). The loss of longitudinal smoothness can be mainly regarded a consequence of these two main distresses. Truck axle configuration plays a major role in top-down cracking while truck speed plays a major role in faulting (NCHRP 2004). Nonetheless, most APT facilities cannot reproduce realistic truck axle configuration and speed. Furthermore, while concrete hygrothermal conditions play a major role in both cracking and faulting (NCHRP 2004), the simulation of these conditions in the APT is not a trivial task. This paper explores how the limitations of most APT devices/facilities impact the evaluation of jointed concrete pavements. Scope. This paper is focused on accelerated pavement testing devices, either linear or circular, typically known as test tracks. The Heavy Vehicle simulator (HVS), the Ifsttar carrousel, and the Australian Accelerated Loading Facility are examples of these devices. The accelerated testing of jointed concrete pavements in test roads like NCAT or MnROAD present very different challenges that are not discussed in this paper.

2 Identification of Limitations of the APT of Jointed Concrete Pavements 2.1

Simulation of Thermal Actions

Except during the early age of concrete, when the heat released by the cement hydration process is considerable, the temperature of the slab is determined by the heat exchange between the slab and the ambient environment. The energy from the sun radiation plays a major role on this exchange. Obviously, the impact from the sun radiation does not exist when the APT testing is conducted indoors. Even when the testing is conducted outdoors, most APT devices are large enough to cover most of the section being tested. This shedding may result in undesirable differential heating of the slabs, which may considerably distorts the outcome of the test. In an APT experiment that was recently concluded at the University of California Pavement Research Center (Mateos et al. 2019), the top, south, east, and west panels of the Heavy Vehicle simulator (HVS) environmental chamber were installed to prevent the differential heating of the slabs (Fig. 1). The north panels were left open since no sun radiation could come in from that side. While the opening allowed air temperature on the tested section to change as the outdoors temperature, the lack of sun radiation resulted in a much smaller variation of the slab mean temperature and the equivalent linear temperature difference (ELTD), compared to the sections that were not affected by the shed of the HVS (outdoors series in Fig. 1).

180

A. Mateos et al. 50

Temperature (°C)

40 30 20 10 0 -10 29-Jun

a)

4-Jul

9-Jul

14-Jul

Mean (Outdoors)

ELTD (Outdoors)

Mean (Test sec on)

ELTD (Test sec on)

19-Jul

b) Fig. 1. a) Comparison of slab mean temperature and ELTD between uncovered concrete section (Outdoors series) and section under the HVS shed (Test section series), b) tested section, environmental chamber is open on its north side

2.2

Simulation of Hygral Actions

Concrete drying shrinkage results in slab contraction and concave upwards curvature, similar to negative (top of the slab cooler than bottom) temperature gradients. For any particular concrete pavement, these effects considerably vary versus time, depending on concrete age and weather conditions, as shown on Fig. 2 example. This figure includes the differential drying shrinkage (top versus bottom of the slab) in a set of concrete slabs built with different concrete mixes (Mateos et al. 2019). Because the APT experiments are typically conducted in a relatively short time period, in the order of weeks, the outcome of the test may be highly dependent on the time of the year when it is conducted.

The Challenges of the Accelerated Testing of Jointed Concrete Pavements

181

0

Differen al shrinkage (μɛ)

-100

-200

-300

-400

-500

-600 20-Feb

10-Apr

30-May

19-Jul

7-Sep

27-Oct

16-Dec

4-Feb

26-Mar

15-May

Rainfall day

Type II/V portland cement

Type II/V portland cement

Type III portland cement

Calcium sulfoaluminate cement

Type II/V internally cured

Fig. 2. Differential drying shrinkage (top versus bottom of the slab) in a series of concrete slabs built with different concrete mixtures; concrete slabs built on February 23–25

2.3

Transverse Joints Opening

The opening of the transverse joints is known to impact the load transfer efficiency (LTE) attributed to the aggregate interlock (NCHRP 2004). Among other factors, this opening is determined by the mean temperature of the slabs and the drying shrinkage. As explained in Sect. 2.1 and 2.2, the simulation of these two variables is not a trivial task. Furthermore, the opening of the transverse joints also depends on the thermal contraction that the concrete experiences after setting. This thermal contraction results in initial deployment of transverse joints at relatively large intervals, 10 to 45 m (ACPA 2002). Very frequently, the length of the APT sections is smaller than this interval and this may result in little opening of the transverse joints compared to the field, i.e., unrealistically high LTE. 2.4

Faulting of Transverse Joints

Transverse joint faulting is the result of the loss of support under the leave slab due to fines pumping from beneath the leave slab to beneath the approach slab. The pumping of the fines requires, in addition to an erodible base and free moisture under the joint, that the truck wheels travel at a relatively high speed. Unfortunately, the speed of the wheel is very low for most APT devices. Most linear devices, which are the most common worldwide, cannot reach 10 mph (Jones et al. 2012). Consequently, most APT devices cannot simulate transverse joint faulting. On the other hand, the loss of LTE due to the wearing of the aggregate interlock and the dowel-concrete interface can be adequately simulated in APT since this wearing is directly related to the stresses that

182

A. Mateos et al.

develop under the wheel loading. An example of successful evaluation of the loss of LTE under accelerated loading is summarized in Fig. 3 (Mateos et al. 2019).

Differential defl. under HVS wheel (mm)

0.1

40 kN dry

0

60 kN dry

80 kN dry

80 kN wet

100 kN wet

-0.1

Thick asphat base

-0.2 Thick asphalt base with fatigue cracking Thin asphalt base

-0.3 -0.4 -0.5

Thick rubberized aphalt base

-0.6 -0.7

Number of wheel passes

-0.8 0

100 000

200 000

300 000

400 000

Fig. 3. Loss of LTE during HVS testing of thin bonded concrete overlay of asphalt, 6  6 slabs, undoweled

2.5

Top-Down Cracking

Top-down transverse cracking is the result of longitudinal tensile stresses that occurs at the top of the slab, at or close to the slab longitudinal edge, under certain loading conditions (NCHRP 2004). These loading conditions typically include the combined action of a truck steering and drive axles, each acting at one side of a slab, together with negative thermal gradient and drying shrinkage in the slab (Fig. 4). Similarly, longitudinal cracking is the result of transverse tensile stresses that occur at the top of the slab, close to the transverse joint, under certain loading conditions (Hiller and Roesler 2005). In this case, the loading conditions include, in addition to negative thermal gradient and drying shrinkage in the slab, the combined action of left and right wheels of the truck axle (Fig. 4). Unfortunately, to the best knowledge of the authors, only one APT device (at Kansas State university) uses a full axle (left and right wheels) and no APT device would be able to reproduce the combined action of steering and drive axles of the trucks. These facts, together with the limited simulation of thermal and drying shrinkage actions, result in limited capability of the APT to reproduce top-down cracking of jointed concrete pavements.

The Challenges of the Accelerated Testing of Jointed Concrete Pavements

183

Principal Stresses 2200 kPa

0 kPa -450 kPa

Fig. 4. Tensile stresses at the top of the slab that may result in longitudinal and transverse topdown cracking; loading includes the combined action of distant wheels and negative ELTD; truck loading is 80 kN (18000 lb) steering axle and 170 kN (40000 lb) drive axle; ELTD is −11 °C (−20 °F)

3 Discussion Based on the limitations presented in Sect. 2, it is clear that reproducing certain distress mechanisms that may affect jointed concrete pavements poses a difficult challenge. Faulting is one of those mechanisms. As explained above, the onset of faulting requires pumping of the fines that cannot occur when the wheel speed is as low as it is in most APT devices. Fortunately, the loss of LTE can be adequately reproduced in APT. Modeling and field calibration seems to be the only option to extrapolate the loss of LTE in APT to loss of LTE and faulting in the field. Reproducing top-down cracking represents a difficult challenge as well. In part because this type of cracking is highly sensitivity to the curvature of the slabs, either due to negative temperature gradients or drying shrinkage, or both, and in part because the combined action of distant wheels cannot be reproduced by the APT devices. In theory, it is possible to find combination of single wheel loading and slab’s negative curvature (concave upwards) that results in tensile stresses at the top and bottom of the slab that are close to the stresses that would occur under the join action of a real truck steering and drive axles (Fig. 5). Similarly to faulting, modeling would be the tool to make the extrapolation from the top-down cracking in the APT experiment (very likely corner cracking) to the top-down cracking in the field. In this case, the limited

184

A. Mateos et al.

knowledge of concrete drying shrinkage in the field, concrete tensile creep, slab-base interaction, and other phenomena that play a role in top-down cracking would represent a major limitation in order to make the extrapolation. Furthermore, the simulation of the negative curvature of the slabs in the APT experiment would represent a practical challenge. At this moment, it does not seem like the APT community has fund a practical solution for such simulation. It must be indicated that increasing the load level is not the solution for simulating top-down cracking. If a negative curvature is not induced on the sabs, the increase in load level will likely trigger bottom-up cracking before top-down cracking takes place.

Principal Stresses

2200 kPa

0 kPa -350 kPa

Fig. 5. Modeling of a single wheel loading (right wheel of the steering axle in Fig. 4); ELTD is −24 °C (−44 °F)

4 Conclusions The simulation of faulting and top-down cracking in the accelerated testing of jointed concrete pavements presents a number of challenges, for three main reasons. The first one is the low wheel speed of most APT devices, which results in the inability to reproduce the fines pumping that the onset of faulting requires. The second reason is the half-axle loading of most APT devices, which results in limitations to simulate the combined action of distant wheels of the real trucks. The third reason is the practical difficulty of imposing the negative curvature on the slabs that, together with the combined action of distant wheels of the real trucks, trigger top-down cracking in the field. Modeling is required to extrapolate APT results to faulting and top-down cracking the field. The development of an approach to induce negative curvature in the slabs during the APT experiments remains a challenge for the APT community. Acknowledgements. This paper describes research activities that were requested and sponsored by the California Department of Transportation (Caltrans). Caltrans sponsorship is gratefully acknowledged. The technical review by Caltrans, led by Deepak Maskey and Dulce Feldman from the Office of Concrete Pavement, and oversight by Joe Holland, of the Division of Research, Innovation and System Information, is appreciated.

The Challenges of the Accelerated Testing of Jointed Concrete Pavements

185

Disclaimer. The contents of this paper reflect the views of the authors and do not necessarily reflect the official views or policies of the State of California or the Federal Highway Administration. This paper does not represent any standard or specification.

References AASHTO: Mechanistic-empirical pavement design guide: A manual of practice. American Association of State Highway and Transportation Officials (2015) ACPA: Early Cracking of Concrete Pavement—Causes and Repairs. American Concrete Pavement Association: IS405.01P (2002) Hiller, J.E., Roesler, J.R.: Determination of critical concrete pavement fatigue damage locations using influence lines. J. Transp. Eng. 131(8), 599–607 (2005) Hugo, F., Martin, A.E.: Significant Findings from Full-Scale Accelerated Pavement Testing, vol. 325. Transportation Research Board (2004) Jones, D., Harvey, J., Al-Qadi, I.L., Mateos, A. (eds.): Advances in Pavement Design Through Full-Scale Accelerated Pavement Testing. CRC Press (2012) Mateos, A., Ayuso, J.P., Jáuregui, B.C.: Shift factors for asphalt fatigue from full-scale testing. Transp. Res. Rec. 2225(1), 128–136 (2011) Mateos, A., Soares, J.B.: Characterization of the stiffness of unbound materials for Pavement design: do we follow the right approach? J. Transp. Eng. 140(4), 04014001 (2014) Mateos, A., Harvey, J., Paniagua, F., Paniagua, J., Wu, R.: Accelerated testing of full-scale thin bonded concrete overlay of asphalt. Transp. Res. Rec. 2673(2), 404–414 (2019) NCHRP: AASHTO Mechanistic-Empirical Design Guide, NCHRP Project 1-37a. Transportation Research Board (2004) Steyn, W.J.: Significant Findings from Full-Scale Accelerated Pavement Testing, vol. 433. Transportation Research Board (2012) Wu, R., Harvey, J.T.: Evaluation of the effect of wander on rutting performance in HVS tests. In: Proceedings of the 3rd International Conference on Accelerated Pavement Testing, February 2008

Accelerated Pavement Testing for the Evaluations of Structural Design and Safety Performance of an Innovative Road Coating Mai Lan Nguyen1(&), Pierre Hornych1, Minh Tan Do1, Thierry Sedran1, and Duc Tung Dao2 1

MAST, Univ Gustave Eiffel, F-44344 Bouguenais, France [email protected] 2 R&D Lafarge Holcim, Saint-Quentin Fallavier, France [email protected]

Abstract. Pervious concrete use has increased since more than a decade thanks to its main feature of high water permeability. However the use of this material is still limited to light traffic areas, like car parks or walkways in urban areas. Recently LafargeHolcim Research Center has developed an innovative concept of concrete pavement structure using this type of material for heavy traffic highways. An ultra-thin innovative pervious concrete layer is laid directly (wet on wet) on top of an optimized concrete layer just after its placement. This innovative solution allows ensuring both structural performance and surface characteristics for road users in terms of safety, riding comfort and aesthetic. In order to evaluate the actual performance of this innovative pavement structure, an accelerated pavement test (APT) was carried out using the FABAC heavy traffic simulator of IFSTTAR in Nantes. A total of five million cycles of 6.5 tons half-axle load were applied. Structural performance and lifetime of the pavement were evaluated by comparison with a design process using the French-standarddesign-software called Alizé-LCPC. The evaluation of the bonding strength at the interface between the pervious wearing course and the structural concrete bottom layer showed good durability throughout the entire test, which could allow possible consideration of the surface layer in the whole pavement design process. Finally, dynamic friction tests conducted on the innovative wearing course during the APT allowed to evaluate and predict the evolution of its skid resistance throughout the pavement lifetime. Keywords: Pervious concrete
APT
Pavement design Dynamic friction test
Innovative Road Coating


Interface bonding


1 Introduction Road maintenance budgets have a strong economic impact. In general, developed countries are spending in maintenance yearly amounts between 0.2% and 0.4% of their GDP. With bitumen price rising, concrete pavements become competitive with asphalt pavements, with lower maintenance costs during their service life. Indeed, maintenance © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 186–195, 2020. https://doi.org/10.1007/978-3-030-55236-7_20

Accelerated Pavement Testing for the Evaluations of Structural Design

187

cost, including the economic loss caused by traffic disruption due to maintenance works, represents an important budget (FEHRL 2008). Improving the riding comfort and safety of concrete pavements (evenness, noise, and rain water evacuation) has been a major challenge during the last 20–30 years. The development of the Next Generation Concrete Surface (NGCS), also known as grindand-groove technique, is an example of the concrete pavement community response to such challenge. Recently, LafargeHolcim Innovative Center (LHIC) has developed an innovative coating as wearing course for concrete pavements, to offer: a durable solution with concrete structural capacity; expected standards in terms of comfort and safety; technical innovations such as color choices and easy placing. This new wearing course with a thickness of 3–4 cm, which is placed “fresh on fresh” on conventional concrete, aims to offer an interface between the structural function and the user perception. It is designed to have equivalent costs at construction level as classical concrete pavements, and no structural maintenance over the lifetime, (only surface maintenance, to be assessed). To validate this innovative concept, a fullscale accelerated test was carried out, with the FABAC traffic simulator of IFSTTAR.

2 Material Properties The properties of the ultra-thin pervious concrete, based on polymer modified cement, were studied in the first stage by laboratory mechanical tests, including 3-point flexural strength, compressive strength and durability (scaling resistance). The concrete mixture was designed with Ordinary Porland cement, modified with different type of polymer, with a very fine sand and a monosize coarse aggregate as described in the patent EP 3222780 A1. The strength of concretes were then measured by 3-point flexural tests and compressive tests on 4 cm  4 cm  16 cm size prisms. Figure 1 shows that the mechanical strength of concretes, at a constant rheology, is significantly different, depending on the type and dosage of polymer. In general, polymer improves the flexural strength of concrete (up to 40% compared to the control concrete without polymer). The influence of polymer on the compressive strength needs to be studied further, as it can enhance or reduce the latter. The influence of polymer on the resistance to scaling of concrete was also studied with 2 different families: an acrylic acid ester copolymer which proved very efficient to improve the flexural strength and classically recommended polymer for concrete, Styrene Butadiene emulsion. The scaling test used the most severe method, developed by LRC, in which the pervious concrete is submerged in salty water containing 3% of NaCl and is submitted to freeze-thaw cycles (cf. Fig. 2) until complete failure.

188

M. L. Nguyen et al.

Fig. 1. Influence of polymer type on the compressive strength at 28 days (Rc28d) and the 3point flexural strength (Rfl28d), with load application on the un-molded surface of samples.

Fig. 2. a) 24 h freeze-thaw cycle used for scaling testing b) Scaling results and comparison with a classical pervious concrete

The polymer pervious concrete performance was compared to that of classical pervious concrete tested simultaneously. The classical pervious concrete used a 6/10 mm crushed aggregate within cement paste volume of 160 L per cubic meter for a Water/Cement ratio of 0.34. The results, presented on Fig. 2, show that polymer can improve significantly the scaling resistance of pervious concrete.

3 FABAC Traffic Simulator and APT Pavement Structure The FABAC machine is a linear traffic simulator, capable of applying heavy vehicle loads on full-scale pavement sections. Initially developed for testing of the continuous reinforced concrete pavement concept (Aunis and Balay 1998), since then this traffic simulator has been successfully used for assessment of pavement performance in many research and development projects, including the most recent one on crack propagation in asphalt pavement (Nguyen et al. 2020). The machine has a total length of 10 m, and is equipped with four wheel modules, driven by a chain and electric motor (see Fig. 3).

Accelerated Pavement Testing for the Evaluations of Structural Design

189

Each wheel module can be equipped with single or dual wheels, and apply loads varying between 30 and 75 kN. The length of the loaded section is 2 m. The loading speed can vary between 1 and 5 km/h.

Fig. 3. Principle of the FABAC machine

The experimental pavement was built in September 2016. It consisted of a 3 cm thick porous concrete surface course, a 25 cm thick concrete base, and a 10 cm thick unbound granular subbase. It was built on a 70 MPa subgrade with a total length of 4 m and total width of 2 m. A classical road concrete, of class BC5 of 38 MPa as 28 day compressive strength, was used for the base course. A wooden formwork was installed on the granular subbase, for the construction of the concrete structure (Fig. 4).

Fig. 4. Compaction of the porous concrete layer with a striker roller, and then a vibrating plate

The BC5 concrete was placed in the formwork, leveled and vibrated. Then, the pervious concrete was laid on top of the base layer immediately (fresh on fresh). A 4 cm thick layer was put in place. As the pavement would be protected by the

190

M. L. Nguyen et al.

FABAC frame, no transverse slope was needed to facilitate the water rain drainage (which is the case in the real pavement design). The concrete was first vibrated with a striker roller (Fig. 4), but the compaction was not very efficient. Therefore, it was decided to modify the compaction method: a wooden board was placed on the porous concrete, and the compaction was carried out using a vibrating plate, placed on the board (Fig. 4). 24 h after the construction, two joints were created, by sawing the pavement structure up to a depth of 9 cm. One joint was created in the middle of the structure, and the second one 50 cm apart. The surface texture of the porous concrete and the joint are shown on Fig. 5. The experimental pavement was also instrumented with 8 temperature probes, measuring the air temperature, and the temperatures at different depths in the structure.

Fig. 5. Surface texture of the porous concrete, sawed joint, and sealing of the joint (Sikaflex resin)

4 Test Program The experimental pavement section was subjected to 5 million load cycles with the FABAC machine, using dual wheels, loaded at 65 kN (French standard equivalent axle load), and a loading speed of 3,6 km/h. The test was performed between November 2016 and June 2017. The monitoring of the pavement included: • Visual inspections, to assess pavement deterioration; • Transverse profile measurements, carried out with a profilometer equipped with a laser displacement sensor, with a vertical resolution of 1 mm; • Air temperature and pavement temperature measurements (8 sensors). • Mean texture depth measurements, using the sand patch method; • Skid resistance measurements, using the Skid Resistance Tester (SRT) and the Dynamic Friction Tester (DFT); • Coring, outside the loaded area, to perform in situ pull-off tests, to evaluate the bonding strength between the porous concrete and the base layer.

Accelerated Pavement Testing for the Evaluations of Structural Design

191

5 Results 5.1

Bonding Strength

To verify the bonding strength at 28 days of the 2 layers, pull off tests (NF EN 1542 standard) were carried out on the slab at different positions. The results are shown in Table 1. The National Concrete Pavement Technology Center (NCPTC 2014) recommends a bonding strength from 0.5 MPa to 2.5 MPa for a durable two-lift concrete pavement. Table 1. Bonding strength of porous concrete versus base course concrete Mean

Standard Coeff of Min. value Max. value deviation variation Direct tensile strength 2.00 MPa 0.49 MPa 24.5% 1.09 MPa 2.82 MPa

5.2

APT Structure Analysis

5.2.1 Temperature Measurements and Visual Inspections Figure 6 presents the surface temperature variations recorded during the whole test period. On the surface, the temperatures varied between 0 °C in the winter and 40 °C in the summer. These conditions are representative of the Nantes climate.

Fig. 6. Temperature variations at the surface of the pavement structure

Visual inspections were performed at different stages of the test, to detect surface damage (cracking, surface wear). As practically no evolution was observed, only the results obtained at the end of the test are presented. Figure 7 shows the aspect of the surface of the porous concrete, and of the joints, after 5 million load cycles. The marks on the surface correspond to rubber deposited by the tires, and to grease coming from the FABAC machine. No cracking or deterioration of the surface was detected at the end of the test. The joints also presented no significant damage (no wear or breaking of the edges of the joints).

192

M. L. Nguyen et al.

Fig. 7. View of the porous concrete surface, and of one transverse joint at the end of the test

5.2.2 Transverse Profile Measurements Transverse profile measurements were made at 4 different positions on the experimental pavement, at different stages of the test, using a profilometer with a non contact laser displacement transducer. An example of transverse profiles measured after construction and at the end of the test, near the center of the test section is shown on Fig. 8. The measurements indicate absolutely no permanent deformation of the concrete. The same results were obtained for the other transverse profiles.

Height (cm)

Transverse profile for x = -0.25 m

Length (cm)

Fig. 8. Transverse profiles in the center of the experimental section, after construction and after 5 million load cycles

5.2.3 Service Life of the Evaluated Structure Calculations were performed with the French pavement design method, to evaluate the design life of the tested structure. Indeed, we considered that the polymer pervious concrete would have the same behavior as standard concrete (BC5) in term of fatigue resistance with a lower elasticity modulus (26 GPa). These calculations were performed considering different traffic classes, (and design hypotheses for structuring road networks (VRS) – a higher category of road and main utilities in French road classification). The results are summarized on Fig. 9: for each traffic class, the figure indicates the design life (in number of years), and compares it with the applied traffic (also

Accelerated Pavement Testing for the Evaluations of Structural Design

193

converted to number of years). The tested structure has a design life of 23 years, for a traffic class T1 (500 heavy vehicles/day), which is validated by the results of the FABAC test (no significant deterioration after this level of traffic). Naturally, this type of structure could be designed for a heavier traffic, by increasing its thickness.

Design life (VRS)

Nb years with simulated traffic

Service life (years)

1000.0 100.0 10.0

111.6 52.7

23.2 21.1

1.0

4.2

8.8

3.5 0.4

0.1 200 T2

1.8 0.2

500

1200

3000

5920

T1

T0

TS

TEX

Traffic class (heavy vehicles/day)

Fig. 9. Comparison of design life and observed service life (in years) for different levels of traffic.

5.3

Skid Resistance Evolution

To evaluate the evolution of the surface characteristics of the porous concrete, three types of tests were performed: • Mean texture depth measurements (MTD), by the volumetric patch test (standard NF EN 13036-1); • Skid resistance measurements, using the Skid Resistance Tester (SRT, standard NF EN 13036-4) and the Dynamic Friction Tester (DFT, ASTM E 1911 standard). The DFT measures the friction force between the surface and three rubber pads attached to a rotating disc. A water supply unit provides a water depth of 1 mm on the surface. A friction-speed curve is recorded from 80 km/h to complete stop. Generally, values of friction coefficient, at 20, 40 and 60 km/h (DFT20, DFT40 and DFT60) respectively, are extracted for analyses. The mean texture depth and friction measurements were performed before loading, and after 1, 3 and 5 million load cycles at 4 different points on the pavement. The results (mean values) of the different tests are summarized in Table 2. The mean texture depth of the porous concrete is relatively high, and remained practically constant throughout the test. The lower value obtained at 1 million loads is probably due to the particular test conditions (humid weather and wet pavement).

194

M. L. Nguyen et al. Table 2. Mean texture depth and friction coefficient measurements

Number of cycles MTD (mm) SRT (friction coefficient l) DFT20 (l) DFT40 (l) DFT60 (l)

Initial (0 loads) 0.85 0.74

1 million load cycles 0.75 0.75

3 million load cycles 0.85 0.52

5 million load cycles 0.91 0.54

0.60 0.59 0.55

0.55 0.52 0.48

0.46 0.45 0.43

0.41 0.41 0.40

Figure 10 presents the evolution of the SRT and DFT measurements with the number of load applications. The SRT tests indicated a high initial friction coefficient (0.74), followed by a decrease after 3 million loads, and then a stabilization. After 5 million loads, (which represents a high level of traffic), the friction coefficient is equal to 0.54, which remains satisfactory, according to French practice.

Fig. 10. Evolution of the SRT and DFT friction coefficients during the experiment.

The DFT tests show a similar trend to the SRT, with a decrease of the friction coefficient in the first part of the experiment, followed by a stabilization. For this test developed in the USA, there are no specified values in France. However, the final value of DFT20 of 0.41 meets the requirements of the state of Louisiana for a conventional mix with a life of 15 years.

6 Conclusions In the APT test, the Innovative Road Coating was submitted to 5 million loads, without presenting significant damage. The macrotexture did not evolve after 5 million cycles. The microtexture has decreased as expected, but still presented a satisfactory level after 5 million cycles. The SRT test gave relatively stable results between 3 and 5 million

Accelerated Pavement Testing for the Evaluations of Structural Design

195

cycles, while the DFT friction test showed a continuous wear of the microtexture, with still acceptable values at the end of the test. Measurements of cross-sectional profiles showed no significant deformation after 5 million passages, and no evolution of the joints. The design calculations confirmed the durability of the structure tested here, which can support 20 years of traffic, with 500 trucks per day per lane, without any maintenance due to safety or structural requirements. The new material could then represent an excellent alternative as wearing course for concrete pavements for upcoming projects. However, the APT could not reproduce the long-term effect on the daily and seasonal curling, warping, and interface deboning neither the top down fatigue cracking resistance linked to the dry-hot climates. A real scale jobsite with continuous survey would be realized to answer to these questions.

References ASTM E 1911: Standard Test Method for Measuring Surface Frictional Properties Using the Dynamic Friction Tester Aunis, J., Balay, J.: An applied research program on continuous reinforced concrete pavement: the Fabac project. In: 8th International Symposium on Concrete Roads, Lisbonne (1998) EP 3222780 A1: Concrete pavement structure comprising a concrete base layer and an elastomer improved concrete wearing layer. European patent application, March 2017 FEHRL: New road construction concepts (2008) NCPTC: Guide to Concrete Overlays: Sustainable Solutions for Resurfacing and Rehabilitating Existing Pavements, 3rd edn. National Concrete Pavement Technology Center (2014) NF EN 1542: Products and systems for the repair and protection of concrete structures. Test methods. Measurement of bond strength by pull-off, July 1999 NF EN 13036-1: Road and airfield surface characteristics - Test methods - Part 1: measurement of pavement surface macrotexture depth using a volumetric patch technique, September 2010 NF EN 13036-4: Road and airfield surface characteristics - Test methods - Part 4: method for measurement of slip/skid resistance of a surface: the pendulum test, March 2012 Nguyen, M.L., Chupin, O., Blanc, J., Piau, J.M., Hornych, P., Lefeuvre, Y.: Investigation of crack propagation in Asphalt pavement based on APT result and LEFM analysis. J. Test. Eval. 48(1), 161–177 (2020)

Experimental Investigation of Wheel-Load Induced Strain Responses in Roller Compacted Concrete Pavements Moinul Mahdi, Yilong Liu, Zhong Wu(&), and Tyson Rupnow Louisiana Transportation Research Center, 4101 Gourrier Ave, Baton Rouge, LA, USA [email protected], [email protected]

Abstract. A recently completed study at the Pavement Research Facility of Louisiana showed that, thin roller compacted concrete (RCC) pavements has great potential to be used for low to medium volume roadway applications where a large amount of heavy and/or overweight trucks are often encountered. However, the existing RCC pavement design procedures are solely empiricallybased and only applicable for heavy industrial pavement’s thickness design. The objective of the current study is to develop a mechanistic-empirical (M-E) based RCC thickness design procedure through the evaluation of cracking mechanism and joint performance of RCC test sections under accelerated pavement testing. Two 203-mm thick RCC pavement sections were constructed for this study: one section built over a 216-mm soil cement base and the other over a 305-mm cement treated soil base. Each section was instrumented with one fiber optic strain plate, over which 16 fiber optic strain gages and 3 temperature gages were embedded. Those sensors were positioned apart on the plate to measure wheel load-induced transverse and vertical strains at two critical positions (i.e. the top and the bottom of the RCC layer). This paper documents an investigation of the strain responses on RCC pavements under the APT loading with regard to different load positions, levels, repetitions and temperatures. The measured strains were then used to compare, verify and calibrate a developed finite element model, which will eventually be used for the RCC fatigue performance and thickness design. Keywords: Fiber optic plate
Roller Compacted Concrete response
Finite element modeling


Pavement

1 Introduction Roller Compacted Concrete (RCC) pavement has drawn a great deal of attention in recent years due to its ease of construction, cost effectiveness and potential application for roadways with heavy truck traffic. A recently conducted accelerated pavement testing (APT) study in Louisiana on four full-scale RCC pavement sections including two RCC slab thicknesses (152 mm and 102 mm) over different base types indicated that all RCC tests had very high load carrying capacity. The results from this study also showed that, the impact of the underlying substrate has significant influence on the © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 196–205, 2020. https://doi.org/10.1007/978-3-030-55236-7_21

Experimental Investigation of Wheel-Load Induced Strain Responses in RCC

197

performance of the RCC pavement sections (Wu et al. 2017). Based on the findings from the APT study, RCC was used for a project in Lafayette Parish, which marked the first usage of RCC pavement in the roadway system in Louisiana. RCC pavement design methods employ the same basic strategy as for conventional concrete pavements: keeping the pavements flexural stress and fatigue damage caused by wheel loads within allowable limits. Thickness design procedures for RCC pavements for heavy industrial applications (such as ports and multimodal terminals) have been developed by the Portland Cement Association (PCA) (PCA 2004) and US Army Corps of Engineers (USACE 2000). Because the critical stresses in RCC are flexural, fatigue due to flexural stress is usually used in the thickness design. The PCA fatigue model, which is based on data derived from Portland Cement Concrete (PCC) beam fatigue tests conducted by PCA is mainly used for the RCC fatigue analysis. The model is a conservative lower bound estimate of the allowable number of load applications at a given stress ratio (Titus-Glover et al. 2005). However, the existing RCC pavement design procedures are solely empirically-based and only applicable for heavy industrial pavement’s thickness design. The objective of the current study is to develop a mechanistic-empirical (M-E) based RCC thickness design procedure through the evaluation of cracking mechanism, joint performance and load induced responses of RCC test sections under accelerated pavement testing. This paper describes experimental investigation of load induced strain responses under Heavy Weight Deflectometer (HWD) and ATLaS30 dual tire loading using an innovative instrumentation technique based on fiber optic sensors.

2 Experimental Design Two 203 mm RCC sections (each of 21.8 m long and 4 m wide) over treated and stabilized soil bases were constructed at the Pavement Research Facility (PRF) of Louisiana and are undergoing accelerated loading testing. The RCC pavement sections are to study the performance and design procedure of RCC pavement for Louisiana low to medium volume roadways with heavy truck traffic. Figure 1 shows the two RCC sections constructed at the PRF site and cross sections of the pavement structure. A heavy vehicle load simulation device, ATLaS30, is being used for the accelerated loading of RCC test sections. The ATLaS30 equipped with a dual tire half-axle load has a capacity of applying wheel loads up to 135 kN. through hydraulic cylinders in a channelized, bi-directional mode. Due to high load carrying capacity, the ATLaS30’s dual-tire wheel load at different levels (e.g., 40, 72, 89, 98, and 112 kN) and repetitions will be used in this experiment in order to fail each RCC section within a reasonable time frame.

198

M. Mahdi et al.

Fig. 1. Pavement Structure of the RCC test section

2.1

Pavement Instrumentation

Both RCC pavement test sections were instrumented with innovative polymeric plate technology. Each test section was retrofitted with a thin polymeric plate positioned perpendicularly to the loading direction. The plate was instrumented with fiber optic gages and fixed inside a thin saw cut in the RCC layer with a slow curing epoxy glue. The fiber-optic strain gages working principle is based on the White Light Polarization Interferometry technology. This technology uses a signal conditioner to sense the path length difference inside a Fabry-Perrot interferometer of a known cavity length and delimited by two dielectric mirrors. With proper calibration, the path length difference can be related to strain and temperature measurements (Grellet et al. 2013). Several validation tests were performed shortly after installation to ensure good quality strain measurements. The strains measured by different sensors showed similar shapes and amplitudes under similar loading conditions. Figure 2 shows the installation of the fiber optic strain plate on the RCC pavement test sections.

(a) Saw cutting

(b) Cleaning of saw cut

(c) Installation

(d) Pouring epoxy

Fig. 2. Installation of fiber optic strain plate in RCC pavement test section

The plate width and thickness are 763 mm and 5 mm with a height of 203 mm to fit the thickness of the RCC layer at each test section. Each plate has sixteen fiber optic strain gages and three temperature gages at different depths. The sensors were positioned apart along the plate in order to measure critical strains under dual tire accelerated loading. Figure 3 shows the layout of the strain and temperature gages on the plate: five transversal strain gages and three vertical strain gages 5 mm below the top of the plate; five transversal strain gages and three vertical strain gages 5 mm above the bottom of the plate and three temperature gages at the top, mid depth and bottom of the plate. This will allow the measurement of vertical and transverse strains in the upper

Experimental Investigation of Wheel-Load Induced Strain Responses in RCC

199

and lower parts of RCC layer under accelerated loading along with the temperature profile throughout the RCC slab.

Fig. 3. Strain plate dimensions and sensor locations correspond to wheel load

One DAQ equipped with 8 channels is being used to collect the data of the 19 gages on one plate. The DAQ sends and receives the light and the software interprets and transforms the received signal into physical quantitative values. These values are compiled in a text file at a specified frequency. For the purpose of this project, a 500 Hz data collection frequency was used. 2.1.1 Effect of Plate Installation on the Pavement Structure HWD tests were performed on both the RCC pavement test sections to study the uniformity of pavement structures before and after retrofitting the fiber optic plate. Table 1 shows the change in average deflection at the center of the HWD load plate (D0) before and after the plate installation at the installation location under 40kN HWD load. There is no significant change observed at the center deflection indicating no substantial damage on the pavement structure.

Table 1. HWD deflection under the load plate before and after plate installation Before plate installation After plate installation %difference Average D0 (mils) Average D0 (mils) Section-1 2.91 3.01 3.32% Section-2 2.72 2.83 3.89%

The HWD test was also conducted under four different load magnitudes (26, 40, 55 and 72 kN) with three drops each at three different locations: one feet left of instrumented plate (station A), on top of instrumented plate (station B) and one feet right of instrumented plate (station C). Figure 4 shows the HWD testing plan and the deflection basins for both RCC sections.

200

M. Mahdi et al.

Fig. 4. Deflection basin under HWD loads at different station

As can be seen from Fig. 4, installation of the fiber optic plates did not cause any weakening or strengthening of the pavement structures and behaved as an integral part of the RCC pavement test section. Strain responses and surface deflection were also recorded under HWD loading at different load magnitudes along with pavement temperature. Based on some early pavement response results, it was observed that for the same load magnitude the pavement surface deflection varied significantly with change in temperature profile along the RCC layer. The strain responses observed from the fiber optic plate also resembled the similar trend. With an increasing deltaT (difference between top to bottom temperature of RCC slab), the surface deflection increased and the critical transverse strain at the bottom of the slab also increased. This could be due to the downward curling effect of the slab. Further investigation will be conducted during the experiment to understand this phenomenon. Figure 5 shows the strain basin at the bottom of the RCC layer under 72 kN HWD load at different locations (station A, B and C).

Fig. 5. Strain basin at the bottom of RCC layer under HWD loads

Experimental Investigation of Wheel-Load Induced Strain Responses in RCC

201

The results presented in Fig. 5 also confirms that the strain basin patterns were the same for a HWD loading plate positioned on either side of the strain plate. Same responses were observed at the top transverse sensors and under different load magnitudes. This indicates that a good bonding of the epoxy with the RCC layer was achieved and the plate was able to transfer the applied loads to the entire strain plate indicating continuity of the RCC pavement layer. 2.2

Strain Responses Under Accelerated Loading

Load induced strain responses were measured under ATLaS30 dual tire loading at different load magnitude for both static and dynamic loading condition. For static loading, the ATLaS30 dual tires were manually positioned on top of the strain plate and load was applied hydraulically on the RCC pavement section at four different load levels (40, 72, 89 and 112 kN). For dynamic loading, ATLaS30 dual tire loads were applied bi-directionally at 4mph speed with different load magnitudes for multiple passes. The recorded data was used to produce strain basins under different loading conditions. To analyze the repeatability of a strain measurement, three or more passages with the same load magnitude were recorded. Figure 6 shows the strain basin at the top and bottom of the RCC layer under 112 kN static dual tire loading for both pavement sections. As expected general trend was observed through the results for both RCC pavement sections.

Fig. 6. Strain basin at top and bottom of RCC layer under ATLaS30 dual tire loading

The strain data collected with the plate was either negative (compression) or positive (tension). As shown in Fig. 6 for the transversal strains (top and bottom), compression was observed at the top and tension at the bottom due to the bending flexural behaviour of the RCC layer. It was observed that the critical compressive strain at the top of the RCC layer occured at location 3 (middle of dual tire loading), whereas the critical bottom tensile strain occured at the middle of each individual tire at location 2 or 4. Similar strain basins were observed under dynamic loading and different load magnitudes.

202

M. Mahdi et al.

It was also observed that section-2 exhibits less critical tensile strain compared to section-1 for the same loading and weather condition. According to the pavement structure, the section-2 was supposed to be a weaker section compared to section one due to having a weaker base support resulting in higher tensile strain at the bottom. However, based on field core samples, it was reported in the previous study that section-2 were built thicker than the designed thickness (Wu et al. 2017). That explains why section-2 has less tensile strain compared to section-1. The HWD surface deflection under the HWD plate also showed that, section-2 had less center surface deflection compared to section-1. 2.2.1 Static and Dynamic Loading Condition Wheel loading is commonly modelled as static loads in concrete pavement design guidelines although it has a dynamic nature. Previous studies indicated that dynamic analysis may result in a more significant pavement response in different circumstances and hence cannot be neglected for pavement design. Figure 7 shows the critical pavement responses at the top and bottom of the RCC layer under different ATLaS30 dual tire loads for both static and dynamic loading.

Fig. 7. Pavement strain responses under different ATLaS30 load magnitude

For both, static and dynamic loading, the compressive strains at the top of the RCC layers showed an increasing strain with the increase of the load level, as well as a decreasing strain with the increase of the RCC layer thickness. Similarly, the tensile strains at the bottom of the RCC layers showed an increasing strain with the increase of the load level, as well as a decreasing strain with the increase of the RCC layer thickness. The dynamic pavement responses were observed to be lower than the static loading for top compressive strains for both pavements. However, for the bottom tensile strains, the dynamic pavement response was higher on section-1 and lower on section-2. This could be due to the different RCC thickness and base stiffness. It is known that dynamic structural response depends on the ratio of load frequency to natural frequency of the pavement structure. According to a previous study, increase in the subgrade stiffness or a decrease in base thickness results in an increase of fundamental natural frequency of the pavement. For usual sub-grade strengths and vehicular load frequencies, this implies that the thicker the concrete base, the more sensitive it is to dynamic loads (Yousefi Darestani et al. 2006). Further investigation will be conducted during the experiment to explain this phenomenon and to understand the dynamic behaviour of RCC pavement under accelerated loading.

Experimental Investigation of Wheel-Load Induced Strain Responses in RCC

203

The vertical strains responses are not presented in this paper due to having very low magnitude at the bottom of the RCC pavement; the results were also more scattered near the surface because of the higher signal to noise ratio. However, the tensile strains data collected at the top and bottom of the RCC layer are good indicators that the strain plates performance is good overall.

3 Finite Element Modeling and Preliminary Fatigue Analysis A 3D finite element (FE) model was developed using ABAQUS software to investigate the structural behavior of RCC pavements under accelerated loading. This FE model will be used to predict the critical stresses of the RCC pavements under accelerated loading for different load magnitude. The RCC layer is created as individual layer with self-weight and applied hard contact interaction with base layer, in which the coefficient of friction is considered as 1.0. Based on the symmetry of the pavement layers, quarter pavement models were built in ABAQUS. In order to investigate the effect of temperature distribution on RCC slab, the coupled thermal-displacement analysis step was adopted and 8-node thermally coupled brick, trilinear displacement and temperature element (C3D8T) was chosen for RCC layer. The three sets of measured temperatures on RCC pavement sections were applied as boundary conditions on top and bottom surfaces of RCC slab, and the strain of RCC slab due to self-weight and thermal expansion was obtained; then tire load was applied to obtain the strain response under the combination of tire pressure, self-weight and thermal expansion. Then the difference of these two bottom slab transverse responses were calculated and compared with measured strains. The RCC pavements were considered as elastic layers and the material properties were considered based on the HWD back-calculation results. The model was first verified with HWD testing results for two HWD load levels (40 and 72 kN) for different locations. To verify the model, the RCC pavement thickness for section-1 was considered to be 203 mm and 228 mm for section-2. It was observed that the numerical model has sound consistency with the RCC pavement performance under different HWD load levels. Based on the verification, the duel tire loads were applied on this FE model to predict the critical stress within the RCC layers. The bottom transverse strain responses on the two RCC sections with various tire load levels were also simulated and plotted in Fig. 8. It is shown that the simulation results are fairly matched with the measured transverse responses, especially for the maximum transverse strain located under the tire loading area. The slight different in the measured versus predicted strain basin could be due to the anisotropic behavior of the RCC layer. The maximum tensile stress of RCC slabs were selected and the critical stresses were identified for both RCC pavement sections. The predicted critical stresses then used to predict number of load repetition of the RCC pavements under accelerated loading using the PCA fatigue model (Titus-Glover et al. 2005). Based on the previous laboratory test results, a flexural strength of 4.5 MPa was considered to calculate the stress ratio.

204

M. Mahdi et al.

Fig. 8. Measured versus predicted strain responses at the bottom under accelerated loading

In this study, the ATLaS30 dual tire loads were applied at the center of the slab with no wandering. However, for pavement design, the critical stress generally occurs at the slab edge. To identify the critical stresses on the edge of the RCC pavement sections, the FE model was further modified for edge loading. The predicted critical stresses under dual tire loading at the center and edge of the pavement along with the predicted number of load repetition are presented in Table 2. Table 2. Critical stress and predicted number of load repetition Section-1

40 kN 72 kN 89 kN Section-2 40 kN 72 kN 89 kN

Center loading Critical stress MPa 0.70 1.10 1.35

No. of load repetition Unlimited Unlimited 9.5E+13

Edge loading Critical stress MPa 1.55 2.68 3.32

No. of load repetition 1.5E+10 1067 77

0.46 0.76 0.93

Unlimited Unlimited Unlimited

1.21 2.04 2.51

9.22E+17 373814 3203

The predicted number of load repetition will be useful to determine the accelerated loading sequence to fail the RCC pavement sections within a reasonable time frame. The FE model will be further modified to account for dynamic analysis to accurately predict pavement failure under accelerated loading along with the effects of daytime and night time temperature gradients.

4 Summary Fiber-optic strain plates were installed at PRF site to study the load induced RCC pavement responses under accelerated loading. The HWD measurements showed that the installation of strain plates does not affect the pavement structure and performs as an integral part of the pavement. The strain plates allowed capturing complete strain distribution of the RCC layer under different ATLaS30 load magnitudes along with the

Experimental Investigation of Wheel-Load Induced Strain Responses in RCC

205

temperature profile of the RCC slab. The measured strain responses and pavement temperature profile were then used to validate a FE model. The FE model was useful in predicting the critical stresses under different load magnitudes. Based on the critical stress results, the number of load repetitions to failure of the RCC pavement sections were predicted and the prediction results will be used in determining the accelerated loading sequence for this experiment. Acknowledgement. This study was supported by the Louisiana Transportation Research Center and the Louisiana Department of Transportation and Development. The authors would like to express thanks to all those who provided valuable help in this study. The authors express their full gratitude to Guy Doré (Laval University), Jean-Pascal Bilodeau (Laval University) and Charles Leduc (OpSens) for their continues support during the planning and installation of fiber optic strain plates for this project.

References Wu, Z., Rupnow, T., Mahdi, M.: Roller compacted concrete over soil cement under accelerated loading. LTRC. Report Number FHWA/L.A.16/578 (2017) Grellet, D., Doré, G., Bilodeau, J.-P., Gauliard, T.: Wide-base single-tire and dual-tire assemblies: comparison based on experimental pavement response and predicted damage. Transp. Res. Rec. 2369(1), 47–56 (2013) Yousefi Darestani, M., Thambiratnam, D., Nata-atmadja, A., Baweja, D.: Dynamic response of concrete pavements under vehicular loads. In: Scientific Committee, IABSE (eds.) Proceedings IABSE Symposium - Response to Tomorrow’s Challenges in Structural Engineering, pp. 104–105 (2006) Titus-Glover, L., Mallela, J., Darter, M.I., Voigt, G., Waalkes, S.: Enhanced Portland cement concrete fatigue model for StreetPave. TRB 1919(1), 29–37 (2005) Portland Cement Association: Guide specification for construction of roller-compacted concrete pavements. PCA Document IS009 (2004) U.S. Army Corps of Engineers: Roller-compacted concrete. USACE Engineer Manual EM 11102-2006 (2000)

APT for Airfield Pavements

Failure Modes of Rapid-Setting Concrete Repairs Under Accelerated Aircraft Traffic Lulu Edwards(&), Haley P. Bell, and Jeb S. Tingle U.S. Army Engineer Research, Development Center, Vicksburg, MS, USA {Lulu.Edwards,Haley.P.Bell, Jeb.S.Tingle}@usace.army.mil

Abstract. Maintenance and rehabilitation of concrete pavements for roads and airfields can be a logistical challenge when the ability to stop traffic is difficult, especially in high-traffic areas. For this reason, rapid-setting concrete is an efficient alternative to ordinary portland cement concrete (PCC) for pavement repairs such as slab replacements and large patches. Ordinary PCC used for rehabilitation often requires lengthy cure times, a plant to produce the concrete, and the ability to transport the PCC to the construction site. Rapid-setting concrete can be prepared on site, reduces the time required to fully cure to 2 h after placement, and can be utilized in remote areas. In this study, slab replacements and large patches were constructed by using rapid-setting concrete over various backfill materials. The surface thickness of the rapid-setting concrete caps varied from 15 to 36 cm. The slab replacements and large patches were trafficked with simulated aircraft load carts and evaluated structurally and visually during and after trafficking. This paper presents the typical failure modes and pavement distresses of rapid-setting concrete repairs caused by heavy aircraft loads as a function of repair size. Results indicate that failure modes are similar regardless of repair size. Keywords: Rapid-setting concrete


Pavement repair
Airfield pavement

1 Background Since 2006, researchers at the U.S. Army Engineer Research and Development Center (ERDC) have been conducting research to develop expedient airfield concrete pavement repair techniques in an effort to update repair guidance. Damaged pavements must be repaired using efficient methods and durable materials to reduce the total time that the pavement is removed from service. Cementitious, rapid-setting concrete repair materials have been successfully used for repairing damaged concrete pavements. Since 2006, more than 150 slabs have been repaired using proprietary, rapid-setting concrete materials. Data collected during numerous field trials have indicated that the use of several different commercial products can produce a repair that will return an airfield to traffic following a short curing time of 2 h. These rapid-setting concrete materials can be used for a variety of repair types including spall repairs, small and large patches, full slab replacements, and small and large crater repairs (Priddy 2011).

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 209–217, 2020. https://doi.org/10.1007/978-3-030-55236-7_22

210

L. Edwards et al.

The objective of this paper was to identify the typical failure modes and pavement distresses caused by accelerating heavy aircraft loads as a function of repair size when using a rapid-setting concrete surface. Two groups of repairs were completed, i.e., 2.6m-by-2.6-m large patches and four adjacent slab replacements, which included 9.1-mby-9.1-m repairs.

2 PCC Repair Materials Repair areas were excavated to the existing subgrade and backfilled so that pavement structures could be controlled, characterized, and compared relative to one another. Rapid-setting concrete was the surface material for each repair. The main cementitious component in the rapid-setting concrete mix was Rapid Set Cement, a proprietary, calcium-sulfoaluminate-based material that accelerates the hardening time. The aggregate used in the mix was 0.95-cm maximum size pea gravel. The factory-blended material was stored in large 1,361-kg super sacks fashioned from woven geotextile fabric and lined with plastic. The material was pre-blended and required only water for placement. Equipment suitable for mixing the material includes buckets with paddle mixers, mortar mixers, commercial volumetric mixers, and rotating drum transit trucks. The material was mixed by using a simplified volumetric mixer, which was based on prior research conducted by Priddy et al. (2013) that showed that volumetric mixing was the fastest and most consistent method of mixing and placing rapid-setting concrete. Unlike ordinary PCC, this material can sustain heavy aircraft traffic within 2 h after placement. Laboratory test results collected during initial material certification testing during trial field placements have shown that this material achieves an unconfined compressive strength (ASTM C39) in excess of 20.7 MPa after 2 h and over 34.5 MPa after 28 days. In addition to compressive strength, flexural strengths (ASTM C78) obtained using this material are in excess of 2.4 MPa after 2 h and 4.5 MPa after 28 days (Priddy 2011). For this experiment, a crushed aggregate or a rapid-setting flowable fill were used as the base. A #610 crushed limestone (GW) base material was selected for its high compacted strength and availability. The target California Bearing Ratio (CBR) was 50% for the patch repairs, while the target CBR was either 50% (medium strength) or 80 to 100% (high strength) for the adjacent slab replacement repairs. The rapid-setting flowable fill base material used for some repairs in this research consisted of a dry blend of rapid-setting cement and fine aggregates stored in large 1,361-kg super sacks fashioned from woven geotextile fabric and lined with plastic. As with the rapid-setting concrete, the pre-blended flowable fill material requires only the addition of water. When used with a rapid-setting concrete cap, the flowable fill provides sufficient bearing capacity for heavy aircraft pavement applications, as demonstrated in numerous field experiments (Carruth et al. 2015) and generally provides an unconfined compressive strength of 1.7 MPa after 30 min of cure time and 5.2 MPa after 3 h of cure time. A CBR of approximately 100% is typically achieved after 24 h of cure time.

Failure Modes of Rapid-Setting Concrete Repairs

211

3 Repair Procedures The same general procedure was used for all repairs and included saw cutting the parent slabs to establish repair boundaries, concrete pavement breaking, sublayer excavation, sublayer material placement, and rapid-setting concrete cap placement. The material type and thicknesses varied from the subgrade to the base course. The variations occurred because the repairs were part of a previous study with a different objective than what is described in this paper. The sublayers were characterized in situ by using a nuclear-density gauge and a dynamic cone penetrometer (DCP). All repairs were surfaced with varying thicknesses of rapid-setting concrete, depending upon the specific objectives of the test sequence. Repairs were completed throughout the span of 3 years, and temperatures ranged from 15–32 °C. 3.1

Large Patches

Twenty-one large patches were completed in existing 46-cm-thick airfield-quality PCC. The patched repairs were each approximately 2.6 m by 2.6 m and completed in the center of 6.1-m-by-6.1-m slabs. No dowels or rebar were used. A silt material was used to replace the existing silty clay (CL) subgrade for all large patch repairs. The subgrade material was placed in 15- to 20-cm-thick lifts and compacted to a strength of 4 to 6 CBR. On average, 30 to 46 cm of subgrade material was placed before the base or concrete placement, depending on the test plan. The base layers for the large patch repairs consisted of either no base, 15 or 30 cm of crushed limestone (GW), or 15 or 30 cm of flowable fill. If the repair included a base layer, then the base material was placed until the target elevation for the surface concrete layer was achieved. For limestone placement, the base layers were compacted in 8- to 10-cm-thick lifts and reached CBR strengths of 40 to 50% immediately after placement. For patches in which the rapid-setting flowable fill was used as a base, an extendable boom forklift was used to lower the super sacks over the repair void, and the material was dispensed directly into the excavation in 10- to 15-cm-thick lifts. Approximately 151 L of water per lift were dispensed over the surface of each lift by using a water truck. The process was repeated until the target backfill depth was achieved. For the final lift of flowable fill, only 114 to 132 L of water was typically dispensed to reduce the amount of standing water on the surface of the flowable fill. Large amounts of standing water could affect the consistency of the overlying rapidsetting concrete cap during placement. The rapid-setting flowable fill base had CBR strengths of 80 to 100%. The 15-, 20-, 25-, 30-, or 36-cm-thick rapid-setting concrete surface caps were placed by using a simplified volumetric mixer. The super sacks of rapid-setting concrete were continuously loaded in the mixer while the mixed material was dispensed in the repair. The rapid-setting concrete was placed in a manner similar to ordinary concrete; however, it was placed expediently since the material began to harden and set within 15 to 30 min. The rapid-setting concrete cap was finished with a 3-m-long magnesium screed bar. Minimal hand finishing was completed due to the fast hardening times.

212

3.2

L. Edwards et al.

Full Slab Replacements

Six 9.1-m-by-9.1-m repairs were completed inside a full-scale test section consisting of 38-cm-thick airfield-quality PCC. No dowels or rebar were used. A low plasticity clay material was used as the subgrade material. The subgrade material was placed in 15- to 20-cm-thick lifts and compacted to a strength of 30 to 40 CBR. On average, 25 cm of subgrade material was compacted in place before the backfill was placed. The base layers for the slab replacements consisted of either 36 to 50 cm of crushed limestone (GW) or 28 to 34 cm of flowable fill. For limestone placement, the base layers were compacted in 8- to 10-cm-thick lifts and achieved CBR strengths of either 50% (medium strength) or 80 to 100% (high strength). Base layers comprised of flowable fill were completed as described in Sect. 3.1. No formwork was used for the base layers inside the 9.1-m-by-9.1-m repairs. It is difficult to place large sections at one time due to the nature of the rapid-setting concrete. Formwork comprised of plastic forms and expansion boards was required for the replaced slabs to create four smaller placements when capping the surface with rapid-setting concrete. This step was necessary to ensure a quality repair. No expansion boards or dowels were used between the replaced slabs and the parent slabs. A simplified volumetric mixer was used for all repair capping activities as described in Sect. 3.1. The rapid-setting concrete was placed one quadrant at a time and placed diagonally so that the plastic forms could be removed once the concrete around each quadrant had set. Each quadrant was 4.6 m by 4.6 m. The surface thickness of the rapid-setting concrete varied from 15 to 36 cm. The rapid-setting concrete cap was finished with a 6-m-long magnesium screed bar. Minimal hand finishing was completed due to the fast hardening times and to prevent excessive shrinkage cracking.

4 Accelerated Traffic Simulation Heavy aircraft traffic is considered as one of the most damaging types of traffic to pavement surfaces because of its small footprint, heavy loads, and high tire pressures. Two types of heavy aircraft traffic were simulated over the rapid-setting concrete repairs by using specially designed load carts. The large patches were subjected to 15,876 kg with 2,241 kPa tire pressure (single-wheel gear, Fig. 1a). The full slab replacements were subjected to 122,470 kg and 965 kPa tire pressure (six-wheel gear, Fig. 1b). A simulated lateral normally distributed traffic pattern was applied to the pavement repairs. Lanes were designed to simulate the traffic distribution pattern, or wander width, of the main landing gear wheels when taxiing to and from an active runway. Traffic was applied bi-directionally by driving the load cart forward and then backward over the length of the repairs and then shifting the path of the load cart laterally approximately one tire width on each forward path. The procedure was continued until one pattern of traffic was completed. Trafficking operations began approximately 2 h after the completion of each repair and were continued until failure occurred or the maximum number of accelerated traffic

Failure Modes of Rapid-Setting Concrete Repairs

213

passes was achieved. Traffic was capped at 10,000 passes for the large patch repairs and 3,500 passes for the slab replacements.

Fig. 1. a) single-wheel gear, b) six-wheel gear

5 Failure Criteria The pavement repairs were expected to fail because of surface deterioration of the rapid-setting concrete cap under traffic. Visual inspections were performed at selected traffic intervals to identify specific pavement distresses associated with foreign object debris (FOD) or tire hazard potential. Distresses were monitored in accordance with traditional condition survey procedures. Structural failure of the concrete pavement was defined as the identification of any of the following distresses, i.e., high-severity shattered slabs, cracks, or corner breaks measured by using the pavement condition index (PCI) inspection procedure (ASTM D5340). If high-severity corner or joint spalls occurred, then the repair was considered functionally failed. Spalling severity was based on the presence of fragmented pieces that might cause FOD or the extent to which the removed pieces of spalled material might cause tire damage hazards. Joint spalls were defined by using their dimensions after any loose FOD was removed. For comparative analysis, repair failure for this project was quantitatively defined by a high-severity shattered slab or spalling greater than 61-cm long, 15-cm wide, and 5-cm deep at any point along the length of the spall. As FOD was produced during trafficking, it was removed to prevent tire hazards. Loose material in spalls was removed if it could easily be dislodged by hand or broom.

6 Repair Performance Results Prior to trafficking, the surface of each repair was inspected for any pre-traffic distresses. The repairs were then inspected at varying pass levels until they failed or until the maximum number of test passes for the type of repairs was achieved. Each repair surface was also surveyed by using a rod and level prior to and after trafficking. The survey data were used to measure any permanent deformation that occurred. No significant temperature fluctuation occurred for each repair construction and trafficking sequence.

214

6.1

L. Edwards et al.

Large Patches

Twenty-one large full-depth patches (2.6 m by 2.6 m) were used for PCC repairs and subjected to the accelerated heavy aircraft simulator load cart with high tire pressures. Table 1 presents the general results of the large patches trafficked with heavy aircraft traffic. Table 1. General results of traffic tests over rapid-setting concrete large patch repairs Passes to Failure mode failurea

None None None None None None None

Base thickness (cm) – – – – – – –

15 20 20

GW GW GW

15 15 15

1,300 2,500 3,500

11

25

GW

15

10,000+

12 13 14 15

15 20 25 15

Repair no.

Base material

1 2 3 4b 5 6b 7

Cap thickness (cm) 15 20 25 30 30 36 36

8 9c 10

550 1,750 3,000 3,000 9,000 2,500 10,000+

High-severity joint spall High-severity joint spall High-severity joint spall High-severity joint spall High-severity joint spall High-severity joint spall Close to failure due to high-severity joint spall Tire hazard and FOD Depth of joint spall and tire hazard Tire hazard; length and width of joint spall; depth was 4.8 cm Close to failure due to high-severity joint spall High-severity joint spall High-severity joint spall High-severity joint spall High-severity joint spall

GW 30 1,500 GW 30 8,000 GW 30 8,500 Flowable 15 3,900 fill 16d 20 Flowable 15 10,000+ – fill 17 20 Flowable 15 10,000+ – fill 18 25 Flowable 15 10,000+ – fill 19 15 Flowable 30 10,000+ – fill 20 20 Flowable 30 10,000+ – fill 21 25 Flowable 30 10,000+ – fill a Failure defined as high-severity shattered slab or spalls greater than 61 cm long, 15 cm wide, and 5 cm deep; traffic capped at 10,000 passes b Repair repeated due to wet sublayer conditions noted after failure c Repair repeated due to questionable early failure d Repair repeated due to low-severity shattered slab occurring at 550 passes

Failure Modes of Rapid-Setting Concrete Repairs

215

The data in Table 1 show that large patch repairs can be completed by using rapidsetting concrete as long as the underlying base or subgrade material is uniform and compacted. The primary mode of failure for the failed repairs was spalling along the transverse joints (joints perpendicular to the direction of travel). Other typical modes of distress for PCC pavements such as corner breaks, etc. were not prevalent. All repairs with no base material failed between 550 and approximately 10,000 passes. No matter the thickness of the rapid-setting concrete cap, all repairs failed by high-severity joint spalling along the transverse joints. Additional noted surface distresses included shrinkage cracks, low-severity corner breaks, and low-severity linear cracks. However, these types of distresses were not prevalent and did not control the structural performance of the repairs. The shrinkage cracks began appearing approximately 20 min after each repair was completed and were likely the result of overworking the rapid-setting material. All repairs with crushed limestone bases failed by 1,300 to 8,500 passes due to high-severity joint spalling, with the exception of Repair 11. Traffic on Repair 11 was stopped at 10,000 passes. At 10,000 passes, the repair was close to failure with a highseverity joint spall that was 236 cm long, 15 cm wide, and 4.8 cm deep. Typical distresses on the rapid-setting concrete caps with crushed limestone backfill were shrinkage cracks, low-severity corner breaks, and low-severity linear cracks. The repairs with rapid-setting flowable fill as the base material did not reach the failure criteria, with the exception of Repair 15. Repair 15 included a 15-cm-thick cap and failed by a high-severity joint spall at approximately 3,900 passes, indicating that at least a 20-cm-thick cap is required for 10,000+ passes. The typical surface distresses observed on the repairs with flowable fill base layers included shrinkage cracks, lowseverity linear cracks, and joint spalls. Repair 16 had a low-severity shattered slab; however, the shattered slab did not progress in severity and did not result in the repair’s failure. Figure 2 shows a typical joint spall created by the aircraft traffic on the large patches.

Fig. 2. Typical high-severity joint spalling on the trafficked edge of a large patch repair completed with rapid-setting concrete

216

6.2

L. Edwards et al.

Full Slab Replacements

Six full-slab replacements (9.1-m-by-9.1-m), placed in four adjacent partial slabs, were repaired by using rapid-setting concrete and trafficked with a heavy aircraft simulator (six-wheel gear) load cart with high tire pressures. Table 2 presents the general results of the six full slab repairs trafficked with heavy aircraft traffic. These data show that PCC slabs replaced with rapid-setting concrete will fail by high-severity joint spalls. Table 2. General results of traffic tests over rapid-setting concrete adjacent slab replacements Repair no. 1 2 3 4 5

Cap thickness (cm) 25 27 25 36 28

Base material

Base thickness (cm) 36 51 51 41 34

Passes to failurea

Failure mode

GW 200 High-severity joint spall GW 200 High-severity joint spall GW 200 High-severity joint spall GW 2,000 High-severity joint spall Flowable 3,500 Length and depth of fill joint spall; tire hazard 6 15 Flowable 34 450 High-severity joint spall fill a Failure defined as high-severity shattered slab or spalls greater than 61 cm long, 15 cm wide, and 5 cm deep

The typical surface distresses observed for the majority of the repairs, regardless of the base material or surface thickness, included shrinkage cracks, joint spalls, and lowseverity shattered slabs. Also, with the exception of Repair 4, which had the largest cap thickness, all slab replacements had corner breaks in each quadrant’s interior corner, or the center of the entire repair. Figure 3a shows typical shattered slabs and joint spalls and Fig. 3b shows interior corner breaks for the full slab replacements. Increasing the base thickness of the GW backfill did not improve the performance of the repair or change the failure mode. The parent slabs along the trafficking joints produced spalls. The center joints between the slabs typically did not spall. Spalling occurred only at the joints adjacent to the parent slabs in the trafficking area.

Fig. 3. a) Typical shattered slabs and joint spalls in adjacent slab replacement with rapid-setting concrete and b) Corner breaks in the center of the repair of the adjacent slab replacement repair

Failure Modes of Rapid-Setting Concrete Repairs

217

7 Summary and Conclusions The mode of failure for the large full-depth patches and full-slab replacements, regardless of the surface thickness or backfill material, was primarily functional failure rather than structural failure due to high-severity joint spalling along the aircrafttrafficked repair joints. No settlement or faulting of the repairs occurred. Shrinkage cracking appeared on the surface of almost all repairs. This was likely due to the fast curing and overworking of the rapid-setting concrete material. Minimal spalling occurred on the inside joints of the full-slab replacements. Lowseverity shattered slabs and interior corner breaks typically occurred in the full slab replacement repairs but did not result in the repairs’ failure. Airfield pavements must be high-quality to handle the large loads and high tire pressures. The results of this study indicate that using rapid-setting concrete for large and small PCC repairs will not immediately result in structural failure of a pavement and can structurally support aircraft traffic. Acknowledgement. The tests described and the resulting data presented herein, unless otherwise noted, were obtained from research sponsored by the U.S. Air Force and performed by the U.S. Army Engineer Research and Development Center. Permission was granted by the Director, Geotechnical and Structures Laboratory, to publish this information.

References Carruth, W.D., Edwards, L., Bell, H.P., Tingle, J.S., Griffin, J.R., Rutland, C.A.: Large Crater Repair at Silver Flag Exercise Site, Tyndall Air Force Base, Florida. ERDC/GSL TR-15-27. U.S. Army Engineer Research and Development Center, Vicksburg (2015) Priddy, L.P.: Development of Laboratory Testing Criteria for Evaluating Cementitious, RapidSetting Pavement Repair Materials. ERDC/GSL TR-11-13. U.S. Army Engineer Research and Development Center, Vicksburg (2011) Priddy, L.P., Tingle, J.S., Edwards, M.C., Griffin, J.R., McCaffrey, T.J.: CRATR Technology Demonstration: Limited Operational Utility Assessment 2. ERDC/GSL TR-13-39. U.S. Army Engineer Research and Development Center, Vicksburg (2013)

Evaluation of Warm Mix Asphalt (WMA) Technologies for Use in Airport Pavements Navneet Garg1(&), Hasan Kazmee2, and Lia Ricalde2 1

NAPMRC, Federal Aviation Administration, Atlantic City, NJ, USA [email protected] 2 Applied Research Associates, Egg Harbor Township, NJ, USA

Abstract. Federal Aviation Administration (FAA) currently does not allow use of warm mix asphalt (WMA) on Airport Improvement Program (AIP) funded airport runway and taxiway pavement projects due to the lack of pavement performance data under heavy aircraft loads on WMA. FAA’s National Airport Pavement and Materials Research Center (NAPMRC) was established to evaluate performance of new asphalt material technologies under heavy aircraft loading at high pavement temperatures. As part of Test Cycle 2 (TC2), six test lanes were constructed – four outdoors and two indoors, each encompassing three different test sections. In outdoor lanes, three different WMA additives (chemical, waxy, and hybrid) were used in Lanes 2, 3, and 4 respectively. Lane1 is the control section with FAA standard P401 specification hot mix asphalt (HMA). Heavy weight deflectometer (HWD) tests were performed on the constructed test lanes to characterize the pavements. Extensive laboratory tests are planned on WMA’s and HMA (field cores and loose mixes). The test lanes will be subjected to accelerated pavement tests (APT) using custom designed airport heavy vehicle simulator (HVS-A) to study rutting performance (at high pavement temperatures) and fatigue behavior. This paper will present construction of test lanes, asphalt mix designs, results from HWD tests on test lanes, and results from response tests on test lanes. Keywords: Airport pavements simulator
HWD


HMA
WMA
Airport heavy vehicle

1 Introduction In United States, the Airport Improvement Program (AIP) provides grants to public agencies, and in some cases to private owners and entities, for the planning and development of public-use airports that are included in the National Plan of Integrated Airport Systems (NPIAS). Airports using AIP funding are required to use Federal Aviation Administration (FAA) advisory circulars (AC’s) that are maintained by Office of Airports (AAS-100). The guidelines for asphalt mix design are provided in FAA AC 150/5370-10H, Item P-401 (FAA 2018). Both Marshall mix design and Superpave mix design methodologies are allowed. This advisory circular also provides guidance of selection of Performance Graded (PG) binders. FAA AC 150/5370-10 does not include © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 218–227, 2020. https://doi.org/10.1007/978-3-030-55236-7_23

Evaluation of WMA Technologies for Use in Airport Pavements

219

specifications for Warm Mix Asphalt (WMA), and this has resulted in very limited use of WMA on airport pavement in USA. A recent research study Test Cycle-1 (TC-1) carried out at FAA’s National Airport Pavement and Materials Research Center (NAPMRC) showed that WMA technology has potential for use in airport pavements (Garg et al. 2018). The study involved performing accelerated pavement tests (APT) using Heavy Vehicle Simulator (HVS-A) and comparing performance of P401 HMA with WMA. A chemical additive (blended in bitumen) was used to produce WMA. To study the performance of WMA’s produced using different types of additives, as part of Test Cycle 2 (TC2), six test lanes were constructed – four outdoors and two indoors, each encompassing three different test sections. In outdoor lanes, three different WMA additives (chemical, organic, and hybrid) were used in Lanes 2, 3, and 4 respectively. Lane-1 is the control section with FAA standard P401 specification hot mix asphalt (HMA). Reclaimed Asphalt Pavement (RAP) is used in Lanes 5 and 6. The objectives of TC2 are to assess the performance of representative types of WMA’s technologies and compare them with the traditional FAA P-401 HMA under high tire pressure at both high and ambient temperatures. a) Compare WMA performance with HMA performance • Rutting (at high temperature) • Fatigue (at ambient temperature) b) Evaluate different WMA technology types • Wax/organic additive • Chemical additive • Hybrid additive c) Compare WMA/RAP performance with P401 HMA performance • Rutting (at high temperature) • Fatigue (at ambient temperature) This paper summarizes the construction of test lanes, asphalt mix designs, results from HWD tests on test lanes, and results from response tests on outdoor test lanes.

2 National Airport Pavement and Materials Research Center NAPMRC is located at FAA’s William J. Hughes Technical Center in Atlantic City, NJ. A state-of-the-art airport heavy vehicle simulator (HVS-A) is used to conduct accelerated pavement tests (APT) on test lanes at NAPMRC. Figure 1 shows the sixth generation HVS-A custom designed and built to FAA specifications. Additional details about NAPMRC and HVS-A can be found elsewhere (Garg et al. 2018). HVS-A has the capability to heat the pavement surface up to a temperature of 150 °F (67 °C) and maintaining it throughout the test duration. Figure 2 shows the layout of test lanes at NAPMRC. Each of the test lanes were subdivided in three different test sections: north, center and south. North test sections will be used to study rutting performance of

220

N. Garg et al.

asphalt concrete (HMA, WMA, and WMA + RAP) at high pavement temperature (120 °F [49 °C] at 2 in. [51 mm] below pavement surface), South test sections will be used to study fatigue performance at (68 °F [20 °C] at 2 in. [51 mm] below pavement surface), and Center test sections will used for performing additional tests (to be decided in near future).

Fig. 1. Heavy Vehicle Simulator – Airfields (HVS-A)

The pavement cross section consists of 9 in. (229 mm) of asphalt concrete surface (P-401 HMA in Lane 1, WMA’s with chemical (CHM), organic (ORG) and hybrid (HYB) additives in Lanes 2 through 4, respectively, WMA with Organic additive and RAP (SPE) in Lanes 5 & 6), 8 in. (203 mm) of P-209 crushed stone base, 12 in. (305 mm) of P-154 subbase, over a sandy subgrade (CBR-15). P-401, P-209, and P-154 are FAA specifications for HMA, crushed stone base, and subbase respectively (FAA 2018). Pavement cross section was designed using FAARFIELD, FAA pavement thickness design software (FAA 2009), so that failure does not occur in base, subbase, and subgrade, and is limited to asphalt concrete surface.

3 Mix Designs A job mix formula (JMF) was developed for the control mix (HMA PG76-22). This JMF complied with all the requirements in the TC2 Construction Specifications ITEM P-401MR Hot Mix Asphalt (HMA) Pavements and Warm Mix Asphalt (WMA) Pavements. Once the JMF was approved for the control mix (HMA), the different WMA technologies additives were added in the dosage specified by the manufacturer. Mix design criteria are shown in Tables 1 and 2. Table 3 provides the summary of different mixes used in TC-2. Previous literature indicates that the asphalt content of a particular mix does not change upon the addition of warm mix additives (Kheradmand et al. 2014; Rubio et al. 2012). Since the contractor was having difficulty in achieving the target air voids with the designated 5% asphalt content while developing the JMF, the range of air voids was shifted for the organic and hybrid additive mixes as noted in Table 2.

Evaluation of WMA Technologies for Use in Airport Pavements

Fig. 2. TC-2 plan view and test pavement cross section Table 1. Aggregate gradations Sieve size Percent passing (%) 1 inch (25.0 mm) 100 3/4 inch (19.0 mm) 90–100 1/2 inch (12.5 mm) 68–88 3/8 inch (9.5 mm) 60–82 No. 4 (4.75 mm) 45–67 No. 8 (2.36 mm) 32–54 No. 16 (1.18 mm) 22–44 No. 30 (600 µm) 15–35 No. 50 (300 µm) 9–25 No. 100 (150 µm) 6–18 No. 200 (75 µm) 3–6 Voids in Mineral Aggregate (VMA) 14 Asphalt percent: 5.0 ± 0.5 Stone or gravel (PG 76-22 HMA & WMA (CHE, ORG & HYB additives). Stone or gravel (PG 64-22 WMA SPE RAP & SPE) 5.5 ± 0.5 Minimum construction lift thickness 3 inch

221

222

N. Garg et al. Table 2. Mix design criteria

Test property

Value

Number of blows/gyrations Air voids (%) for HMA, WMA CHE, WMA SPE & WMA SPE RAP Mixes Air voids (%) for WMA ORG Mix

75 3.5 (2 to 5)

Test method

ASTM D3203 4.4 (2.7 to 5.7) ASTM D3203 Air voids (%) for WMA HYB Mix 4.2 (2.9 to 5.9) ASTM D3203 Percent voids in mineral aggregate (VMA), min. 14% ASTM D6995 ASTM TSR not less than 80 at a D4867 saturation of 70–80%a Asphalt Pavement Analyzer (APA)b

(3) uz;i ¼ uz;i þ 1 > < ur;i ¼ ur;i þ 1 r ¼ rzz;i þ 1 > > : zz;i srz;i ¼ srz;i þ 1 8 (4) Partially bonded > uz;i ¼ uz;i þ 1 > < rzz;i ¼ rzz;i þ 1 srz;i
¼ srz;i þ 1 > >  : srz ¼ Krz ur;i  ur;i þ 1 8 Perfectly sliding > (5) uz;i ¼ uz;i þ 1 > < rzz;i ¼ rzz;i þ 1 s ¼ srz;i þ 1 > > : rz;i srz;i ¼ 0

3 Application to a Four-Layered Pavement Structure 3.1

Pavement Structure and Materials Properties

The studied structure is the one from the STAC test facility (Bonneuil-sur-Marne, France). It has been designed following the French rational method for airfield pavements (STAC 2016). It consists in four layers and three layer interfaces. Layers thicknesses and mechanical properties are specified in Table 2. The two granular layers, GNT and SOL, are considered as linear elastic. E modulus values are specified in Table 2. As the layer SOL is thicker than 6 m, it can be considered as infinite in depth. Bituminous materials, BBA and GB, have been tested in University of Lyon/ENTPE with the tension-compression complex modulus test on cylindrical samples. As an example, master curves of the material GB is presented in Fig. 5. Interfaces GNT/SOL and GB/GNT are considered as perfectly bonded in any case study. Interface BBA/GB is considered either fully sliding, either partially bonded, or perfectly bonded. Table 2. Pavement layers and material mechanical properties Layer material Airfield type bituminous mixture (BBA) Base type bituminous mixture (GB) Untreated graded aggregate (GNT) Subgrade (SOL)

Thickness [cm] 14.7

Modulus [MPa]  ðBBAÞ E2S2P1D

Poisson’s ratio [−] 0.30

Bulk density [kg/m3] 2500

15.6

 E2S2P1D ðGBÞ

0.30

2500

51.5

275

0.30

2000

165

0.30

1800

Spectral Element Simulation of Heavy Weight Deflectometer Test

Experimental data

2S2P1D model

Norm Phase angle

60

10000

50 40

Norm

1000

30 Phase angle

100

20

10

Phase angle [ °]

100000

Norm of E* [MPa]

663

10

1

0

10-9

10-5

10-1

103

107

Frequency [Hz]

Fig. 5. Experimental results of a tension-compression complex modulus test on material GB in terms of master curves and fitted 2S2P1D curves at the reference temperature 15 °C.

3.2

Interface Influence on the Radial Strain

Three cases have been investigated for the BBA/GB interface: perfectly bonded, partially bonded and fully sliding. Radial strains (err) at the BBA/GB interface (z = −14.7 cm, r = 0 cm) are observed in Fig. 6. For each interface condition case, radial strain at the bottom of BBA layer and at the top of GB layer are plotted. It is observed that, for the perfectly bonded case, the radial strain is the same above and under the BBA/GB interface, close (but not equal) to zero. This is due to the imposed continuity conditions at the interface. The GB layer sustains then less extension. In contrast, for the fully sliding and partially bonded cases, radial strains on both sides of the BBA/GB interface are not of the same sign. Indeed, extension is observed at the bottom of BBA and contraction at the top of GB. This fact means that both layers

350

Contraction

HWD Load [kN]

300

Extension

250 200 150 100 50

bottom BBA

top GB

0 -50 -270

3 MPa/m) Linear Elastic (K (Krz 10^3 MPa/m) rz ==10

-135

0

Bonded

135

Sliding

270

Radial strain [µstrain]

Fig. 6. Radial strain at the BBA/GB interface (r = 0 m, z = −14.7 cm) vs HWD load.

664

J.-M. Roussel et al.

(BBA and GB) are acting separately. A radial strain gap is then observed, contrarily to the perfectly bonded case. 3.3

Interface Effects on Deflection Basins

In order to assess effects of the BBA/GB interface bonding conditions on HWD deflections, the pseudo-static deflection basin (defined in Sect. 1) under HWD load has been simulated (Fig. 7). This also allows to numerically evaluate the HWD ability to detect interface defects. In Fig. 7, for the bonded case, the central deflection is 624 lm and jumps to 966 lm for the sliding case. This 35% increase shows that the central deflection (r = 0 m) is very sensitive to BBA/GB bonding condition, which consolidates results obtained in Sadoun (2016). More generally, it can be noticed that deflections calculated at a radial distance lower than 1.20 m are sensitive to the bonding quality. In most cases, HWD devices are equipped with at least 6 geophones located in this area (generally at 0 cm, 30 cm, 40 cm, 60 cm, 90 cm and 120 cm). Figure 8 shows the maximum central deflection as a function of the shear modulus (Krz) used for the interface. It is observed that for low values of K (106 MPa/m) simulates well a perfectly bonded interface.

Deflection [µm]

1000

Bonded

800

Sliding

600

Partially bonded 3 MPa/m) (Krz 10^3 MPa/m) (Krz == 10

400

Geophones locations (basic configuration)

200 0 0

0.6

1.2

1.8

2.4

Radial distance [m]

Central deflection d0 [µm]

Fig. 7. Pseudo-static deflections basins for three BBA/GB interface conditions 1000 950 900 850 800 750 700 650 600

Sliding interface

Bonded interface

10-1

103

107

Interface stiffness Krz [MPa/m]

Fig. 8. Central deflection as a function of shear stiffness of layer interface BBA/GB

Spectral Element Simulation of Heavy Weight Deflectometer Test

665

4 Conclusion In this paper, bituminous layers interfaces conditions modelling has been presented and integrated within a SEM for the HWD test. Used SEM allows dynamic and linear viscoelastic computations. It has been found that all bonding qualities (between fully sliding and perfectly bonded) can be simulated by considering an elastic shear modulus at the interface. This paper also demonstrates that HWD has potentialities for delamination detection by measuring the deflection basin close to the load centre. Further work will consist in determining if HWD signals are sufficient to backcalculate the interface stiffness. This opens prospects for a better pavement structural assessment.

References Al-Khoury, R., Scarpas, A., Kasbergen, C., Blaauwendraad, J.: Spectral element technique for efficient parameter identification of layered media. I. Forward calculation. Int. J. Solids Struct. 38(9), 1605–1623 (2001) Broutin, M.: Assessment of flexible airfield pavements using heavy weight deflectometers, development of a FEM dynamical time-domain analysis for the backcalculation of structural properties. Ecole des Ponts ParisTech (2010) Chupin, O., Chabot, A., Piau, J.M., Duhamel, D.: Influence of sliding interfaces on the response of a layered viscoelastic medium under a moving load. Int. J. Solids Struct. 47(25–26), 3435– 3446 (2010) Doyle, J.F.: Wave Propagation in Structures-Spectral Analysis Using Fast Discrete Fourier Transforms. Springer - Mechanical Engineering Series. Springer US, New York (1997) Grenier, S., Konrad, J.-M., LeBœuf, D.: Dynamic simulation of falling weight deflectometer tests on flexible pavements using the spectral element method: forward calculations. Can. J. Civ. Eng. 36(6), 944–956 (2009) Olard, F., Di Benedetto, H.: General “2S2P1D” model and relation between the linear viscoelastic behaviours of bituminous binders and mixes. Road Mater. Pavement Des. 4(2), 185–224 (2003) Pouget, S., Sauzéat, C., Di Benedetto, H., Olard, F.: Modeling of viscous bituminous wearing course materials on orthotropic steel deck. Mater. Struct. 45(7), 1115–1125 (2012) Roussel, J.M., Sauzéat, C., Di Benedetto, H., Broutin, M.: Numerical simulation of falling/heavy weight deflectometer test considering linear viscoelastic behaviour in bituminous layers and inertia effects. Road Mater. Pavement Des. 20(Suppl. 1), S64–S78 (2019) Sadoun, A., Broutin, M., Simonin, J.M.: Assessment of HWD ability to detect debonding of pavement layer interfaces. In: 8th RILEM International Conference on Mechanisms of Cracking and Debonding in Pavements, vol. 13, pp. 763–769. Springer, Dordrecht (2016) STAC: Rational design methodology for flexible airfield pavement. Technical Guide (2016)

Backcalculation of Airfield Pavement Layer Moduli Under HWD Testing Hao Wang1(&), Pengyu Xie1, and Richard Ji2 1

Rutgers, The State University of New Jersey, New Brunswick, USA [email protected], [email protected] 2 Federal Aviation Administration, Washington, D.C., USA [email protected]

Abstract. This study focused on evaluating airfield pavement in-situ condition using backcalculated layer moduli from Heavy Weight Deflectometer (HWD) testing. Field testing was conducted using HWD devices at the National Airport Pavement Test Facility (NAPTF) of the Federal Aviation Administration (FAA). Two backcalculation methods based on traditional multi-layer linear elastic (MLE) method and dynamic finite element model (FEM) were used in the analysis, respectively. It was found that the backcalculated moduli were slightly different between MLE and FEM methods. However, significant discrepancies were observed for the backcalculated moduli as compared to the laboratory measured moduli. The study findings indicate that the backcalculated moduli need be adjusted for further use in pavement structural analysis. Keywords: Heavy Weight Deflectometer
Backcalculation pavement
BAKFAA
Finite element model


Airfield

1 Introduction Falling weight deflectometers (FWD) apply dynamic impulse load to pavement by dropping free-falling mass onto the plate with rubber springs. The heavy weight deflectometer (HWD) uses the greater load magnitudes than FWD and is more suitable for airport pavements. The magnitude of the impulse load can be varied by changing the mass and/or drop height so that it is similar to that of wheel load on the truck or on the main gear of an aircraft. The surface deflections measured under FWD or HWD testing can be used to assess structure capacity of pavement system and backcalculate the moduli of pavement layers. Previous studies have used MLE to backcalculate layer moduli from FWD testing and found that backcalculation results may not be always reliable (Rada et al. 1992; Mehta and Roque 2003). The discrepancies between the backcalculated moduli and the moduli obtained from laboratory measurements were observed, especially for the resilient moduli of unbound materials. Ong et al. (1991) found that the base and/or subgrade modulus was overestimated because dynamic effect of FWD loading was not

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 666–675, 2020. https://doi.org/10.1007/978-3-030-55236-7_69

Backcalculation of Airfield Pavement Layer Moduli

667

considered in traditional backcalculation methods. Similarly, Siddharthan et al. (1991) and Rahim et al. (2003) found that the backcalculated moduli for base and subgrade layers showed high error as compared to laboratory measurements. Von Quintus and Killingsworth (1998) found that for the unbound layer under asphalt surface, the ratios between the backcalculated moduli from standard procedures and the laboratorymeasured values ranged from 0.35 to 0.62 due to the dynamic effect. Boutin (2010) found that the dynamic behavior of FWD can be magnified with the existence of bedrock as a rigid layer underlying the subgrade. Al-Qadi et al. (2010) and Li et al. (2017) found that that dynamic analysis could cause slightly greater deflections on flexible pavement surface as compared static analysis. HWD tests have been conducted at the National Airport Pavement Test Facility (NAPTF) of Federal Aviation Administration’s (FAA) to evaluate airfield pavement structural capacity in different ways. Garg and Marsey (2002) compared surface deflections from HWD tests and static load tests performed on the flexible pavement at the NAPTF and found the comparison results varied depending on pavement structure. The study also illustrated the complexity of relating pavement surface deflections to pavement life. Gopalakrishnan (2006) analyzed HWD surface deflections acquired at different stages of traffic testing as compared to resilient displacements obtained from multiple-depth deflectometer (MMD) and demonstrated that surface deflections can be used to evaluate structural deterioration of airport flexible pavements. Gopalakrishnan and Thompson (2007) investigated the relationship between surface deflections from HWD tests and the power function parameter of rutting model. Li et al. (2016) monitor the performance of flexible pavement structure during accelerated pavement testing using the ratio of the deflection basin shape factor between traffic and non-traffic area. The results indicate that there is strong correlation between asphalt concrete layer deterioration and increase of certain deflection basin shape factors. This study focused on evaluation of accuracy of backcalculated pavement moduli based on HWD tests conducted at NAPTF. Backcalculation analysis was conducted using the backcalculation software (BAKFAA) developed by FAA and the customized finite element (FE) model for a wide range of airfield pavement structures The backcalculated moduli of pavement layers were compared between different calculation methods and as well as the laboratory measured results.

2 Development and Validation of Finite Element Model The FE model for F/HWD loading is axisymmetric in order to simulate the circular loading area used in the testing. Dynamic analysis was applied to consider the impulse loading applied through dropping weights of F/HWD. The F/HWD loading was simulated in time domain with uniform stress applied on pavement surface within circular area. The loading pulse function was modeled by half-sine curve with 0.03-s duration, which was believed as the reasonable simplification of the real load-time history based on trial analysis of pavement surface deflections using different load shapes (Ruan et al. 2018). Figure 1 illustrated the FE model of pavement with mesh configurations.

668

H. Wang et al.

Fig. 1. Illustration of developed FE model for F/HWD loading

The finite domain of FE model consisted of four-node bilinear axisymmetric solid elements. The mesh density decreased gradually as the position of interest moving away from the loading area for reducing computational time. Infinite elements were used in the horizontal boundaries of the model and at the bottom of subgrade to reduce the degrees of freedom at far field and absorb stress waves for dynamic analysis. Sensitivity analysis was carried out to determine the appropriate mesh sizes in the finite domain and the locations of infinite element (Li et al. 2017). The length of elements close to the loading area was selected at 12.7 mm (0.5 in.) in the radial direction. The final selected finite domain size has an axisymmetric dimension of 8.2  9.3 ft (radius  depth) to achieve the balance between computation cost and accuracy. The horizontal movement was restricted in the symmetric axis and there is no horizontal or vertical translation allowed at the bottom nodes.

3 HWD Testing at Airfield Pavement Sections A series of HWD tests were conducted at construction cycle-7 (CC-7) sections and National Airport Pavement and Materials Research Center (NAPMRC) sections at NAPTF. The HWD tests were conducted at multiple locations of CC-7 and NAPMRC sections with three loading magnitude (12, 24, and 36 kips). The air temperatures during HWD tests were 79 to 80°F for CC-7 sections and 57 to 64°F for NAPMRC section, respectively. Figure 2 shows the schematic diagrams of pavement structures of CC-7 sections. Four pavement test sections were constructed on the north side for construction cycle-7 at NAPTF, which was mainly used to evaluate the effect of asphalt layer thickness on

Backcalculation of Airfield Pavement Layer Moduli

669

structure failure and extend perpetual pavement concept to airport pavements (Garg et al. 2018a). The thicknesses of asphalt layer ranged from 8 to 15 in. The construction materials met the P-401 specifications for asphalt layer (PG76-22), P-154 specification for subbase layer (uncrushed aggregate) and DuPont Clay with CBR value of 5–6 for subgrade.

Fig. 2. Illustration of pavement structures of CC-7 sections (north side)

Fig. 3. Schematic diagram of pavement structure of NAPMRC sections

Figure 3 shows the schematic diagrams of pavement structures of NAPMRC sections. The NAPMRC sections were built for evaluation of performance of hot mix asphalt (HMA) and warm mix asphalt (WMA) under high aircraft tire pressure (Garg et al. 2018b). The thickness of asphalt layer is 5 in. and the base and subbase layers are both 12 in. Two asphalt concrete materials, HMA and WMA, were used for asphalt layer in NAPMRC sections. The construction materials met the P-401/P-403 specification for asphalt layer, P-209 specification for base layer (crushed aggregate), P-154 specification for subbase layer (uncrushed aggregate). The CBR value of subgrade soil was 15.

670

H. Wang et al.

4 Analysis of Backcalculation Results 4.1

Comparison of Backcalculated Moduli from Different Methods

Two different backcalculation methods were used in the analysis: traditional method and FEM-based method. For the traditional method, the commonly used backcalculation software (BAKFAA) developed by FAA was used to backcalculate pavement layer moduli based on multilayer elastic theory. For the FEM-based method, the developed FE model with dynamic analysis was used to backcalculate layer moduli. Elastic material properties with damping ratio of 5% were used for all pavement layers in the FE model. Figure 4 and Fig. 5 show the comparisons of backcalculated moduli using two different methods, respectively, for asphalt surface layer and unbound material layers. The results were illustrated in box-and-whisker plots, which could display the overall characteristics of scattered data. The upper and bottom boundaries of box indicate the third and first quartile, respectively. Two whiskers outside the box extend to the highest and lowest data point. The median and mean value are marked inside the box by horizontal line and cross, respectively. The backcalculated moduli appear reasonable with practical observations. In NAPMRC sections, the moduli of WMA were found smaller than those of HMA. This is probably because the lower production temperature of WMA caused less aging of asphalt binder and thus smaller modulus. The HMA moduli in NAPMRC sections were greater than those in CC-7 sections, which were contributed by the lower temperature during HWD testing. For the convenience of backcalculation, the four-layer structure in NAPMRC sections were considered as three-layer structure to reduce the uncertainty brought by unknown moduli of multiple layers. The existence of crushed aggregate layer increased the moduli of the equivalent base layer in NAPMRC sections, as compared to the uncrushed aggregate layer in CC-7 sections. The moduli of subgrade layer in NAPMRC sections were also greater than those in CC-7 sections, which were consistent with the CBR values.

Fig. 4. Comparison of backcalculated moduli of asphalt surface layer

Backcalculation of Airfield Pavement Layer Moduli

671

Fig. 5. Comparison of backcalculated moduli of base and subgrade layers

In terms of the difference of backcalculated moduli from two approaches, the backcalculated moduli of asphalt surface layers from BAKFAA and FEM-based method fell into the similar range. However, some discrepancies were observed between the backcalculated moduli of unbound materials from two approaches. As compared to the traditional method, the FEM-based backcalculation method caused slightly smaller modulus for the granular base but slightly greater modulus for the subgrade. These discrepancies could be caused by the mass inertia and damping effects considered in dynamic analysis that was implemented in the FEM model. 4.2

Comparison Between Backcalculated and Laboratory Measured Moduli

The backcalculated moduli at CC-7 sections were compared to the laboratory measured material properties. The asphalt mixture with PG76-22 was used in CC-7 sections and the measured dynamic modulus mater curve at different reference temperatures was shown in Fig. 6. Based on the typical pulse duration of HWD loading (0.03 s), the equivalent frequency is about 33 Hz or 5.25 Hz (considering the effect of 2p). The average temperature across the whole asphalt layer was around 81°F, which was used as the representative temperature for comparison. The dynamic modulus of asphalt mixture at 33 Hz and 5.25 Hz was found being 1,229 ksi and 807 ksi, respectively, which were both greater than the backcalculated modulus from HWD tests (400–700 ksi as shown in Fig. 4). The backcalculated modulus was closer to the dynamic modulus of asphalt mixture at 5.25 Hz. The P-154 is the uncrushed aggregate that is usually used as based or subbase material. The subgrade soil at the CC-7 sections is DuPont clay with CBR of 5.5. Triaxial tests have been conducted to measure resilient modulus of P-154 and DuPont clay at different levels of confining pressure and deviatoric stress. The measured modulus was fitted using the generalized nonlinear model that describes resilient modulus as a function of bulk stress and shear stress (ARA 2004).

672

H. Wang et al.

Fig. 6. Dynamic modulus of asphalt mixture (PG76-22) at different reference temperatures

In order to estimate the modulus of unbound material in pavement structure, the stress state need be known. In this study, the stress states on the top of P-154 base layer and subgrade were extracted from the FE model and used to calculate the corresponding modulus using the nonlinear modulus function obtained from laboratory test results. The comparison of laboratory measured and backcalculated modulus is shown in Fig. 7 and Fig. 8, respectively, for P-154 material and subgrade soil.

Fig. 7. Comparisons of laboratory measured and backcalculated modulus for P-154

Backcalculation of Airfield Pavement Layer Moduli

673

Fig. 8. Comparisons of laboratory measured and backcalculated modulus for subgrade soil

It was found that as asphalt layer thickness decreased (from LPF-1 to LPF-4) or the HWD loading magnitude increased, the stress level increased. The estimated modulus of P-154 using the nonlinear modulus function exhibited stress-hardening behavior as the stress level increased. However, the stress-softening behavior of subgrade soil were not observed, which was probably caused by the small contribution of octahedral shear stress under thick asphalt and base layers. In general, the comparison results showed that the ratios of backcalculated and measured moduli were about 1.0 to 3.0 for P-154 and 2.0 to 3.0 for subgrade soil. The finding of backcalculated subgrade modulus is consistent with another independent study conducted by the French Civil Aviation technical center (STAC) (Broutin et al. 2019). The ratios were found not constant across different pavement structures due to the variation of stress states in the base layer and subgrade. It is noted that the ratio of backcalculated and measured moduli for unbound materials is dependent on the stress state that is needed to estimate elastic modulus from the nonlinear modulus function.

5 Conclusions This study conducted backcalculatioin analysis of pavement layer moduli using two methods that are based on multi-layer elastic (MLE) theory and dynamic finite element model (FEM), respectively. It was found that the backcalculated moduli using MLE and FEM methods were similar in general, although the dynamic analysis used in FEM resulted in slightly smaller modulus for the granular base and slightly greater modulus for the subgrade. However, significant discrepancies were observed between the backcalculated moduli and the laboratory measured moduli. The backcalculated moduli of asphalt concrete was found smaller than the dynamic modulus at the loading frequency of HWD loading. On the other hand, the ratios of backcalculated and measured moduli of unbound materials were found generally greater than one, which varied depending on pavement structures and loading magnitudes.

674

H. Wang et al.

The study findings indicate that the backcalculated moduli need be adjusted for further use in pavement structural analysis. Future research need be conducted to verify the study findings for a wide range of airfield pavement structures.

References ARA: Guide for mechanistic-empirical design of new and rehabilitated pavement structures. NCHRP 1-37A Final Report, Washington, DC (2004) Al-Qadi, I.L., Wang, H., Tutumluer, E.: Dynamic analysis of thin asphalt pavements by using cross-anisotropic stress-dependent properties for granular layer. Transp. Res. Rec. 2154, 156– 163 (2010) Broutin, M.: Assessment of flexible airfield pavements using heavy weight dflectometers. Ph.D. Dissertation, LCPC Paris (2010) Broutin, M., Sadoun, A., Duprey, A.: Comparison between HWD backcalculated subgrade dynamic moduli and in-situ static bearing capacity tests. In: International Airfield and Highway Pavement Conference 2019, Chicago, USA (2019) Garg, N., Li, Q., Haggag, M.: Accelerated pavement testing of perpetual pavement test sections under heavy aircraft loading at FAA’s national airport pavement test facility. In: 2018 International Society for Asphalt Pavement (ISAP) Conference, Brazil (2018a) Garg, N., Kazmee, H., Ricalde, L., Parsons, T.: Rutting evaluation of hot and warm mix asphalt concrete under high aircraft tire pressure and temperature at NAPMRC. In: Transportation Research Board 97th Annual Meeting, Washington DC, USA (2018b) Garg, N., Marsey, W.H.: Comparison between falling weight deflectometer and static deflection measurements on flexible pavements at the NAPTF. In: Proceedings of 2002 FAA Airport Technology Transfer Conference (2002) Gopalakrishnan, K.: Condition monitoring of airport pavements subjected to repeated dynamic aircraft loading. Baltic J. Road Bridge Eng. 1(3), 135–142 (2006) Gopalakrishnan, K., Thompson, M.R.: Use of nondestructive test deflection data for predicting airport pavement performance. J. Trans. Eng. 133(6), 389–395 (2007) Li, Q., Cary, C., Garg, N., Rutter, R.: Use of heavy weight deflectometer in monitoring flexible pavement deterioration during accelerated pavement testing. In: Proceeding of Top of Form NDE/NDT for Highways & Bridges: SMT 2016 (In CD) (2016) Li, M., Wang, H., Xu, G., Xie, P.: Finite element modeling and parametric analysis of viscoelastic and nonlinear pavement responses under dynamic FWD loading. Constr. Build. Mater. 141, 23–35 (2017) Mehta, Y., Roque, R.: Evaluation of FWD data for determination of layer moduli of pavements. J. Mater. Civ. Eng. 15(1), 25–31 (2003) Ong, C.L., Newcomb, D.E., Siddharthan, R.: Comparison of dynamic and static backcalculation moduli for three-layer pavements. Transp. Res. Rec. 1293, 86–92 (1991) Rada, G., Richter, C., Stephanos, P.: Layer moduli from deflection measurements: software selection and development of strategic highway research program’s procedure for flexible pavements. Transp. Res. Rec. 1377, 77–87 (1992) Rahim, A., George, K.: Falling weight deflectometer for estimating subgrade elastic moduli. J. Transp. Eng. 129(1), 100–107 (2003)

Backcalculation of Airfield Pavement Layer Moduli

675

Siddharthan, R., Norris, G.M., Epps, J.A.: Use of FWD data for pavement material characterization and performance. J. Transp. Eng. 117(6), 660–678 (1991) Von Quintus, H., Killingswork, B: Analyses relating to pavement material characterizations and their effects on pavement performance. Final report to FAA, Mclean, VA (2012) Ruan, L., Luo, R., Hu, X.D., Pan, P.: Effect of bell-shaped loading and haversine loading on the dynamic modulus and resilient modulus of asphalt mixtures. Constr. Build. Mater. 161, 124– 131 (2018)

SMS Structure Measurement System to Optimize Accelerated Pavement Testing APT Bastian Wacker(&) and Dirk Jansen Federal Highway Research Institute (BASt), Bergisch Gladbach, Germany [email protected]

Abstract. Countries with high traffic volumes are of great importance for national and international freight transport. Construction site impairments have a major influence on the flow of traffic. It is therefore very important to optimize the road structure and materials so that they last longer and minimize maintenance in this sensitive network. The next generation of road research is to improve the overall structure and individual materials in the different layers. After a multitude of theoretical and laboratory tests, a large-scale demonstrator is a very good way to demonstrate the benefits of research before it is used in the road network. An important tool here is a coordinated APT (Accelerated Pavement Testing) program in which the demonstrator is exposed to realistic wheel loads. These APT programs consist of the loading facility, the research infrastructure and the investigation program. The investigation program is one of the main factors in the assessment of the APT program. As a rule, it is necessary to relocate the loading equipment for measurements, resulting in a loss of loading time. In addition, only individual points can be examined and thus do not give a complete overview. To optimize the use of different measurement techniques, a new continuous measurement system (SMS Structure Measurement System) will be developed, tested under controlled conditions and implemented in APT programs. Keywords: Accelerated Pavement Testing APT
SMS Structure Measurement System
duraBASt
Mobile Load Simulator MLS30

1 Introduction Countries with high traffic volumes are of great importance for national and international freight transport. Construction sites influence the flow of traffic and freight traffic is affected by delays. National transit routes are necessary to control all traffic safely and quickly. It is therefore very important to optimize the road structure and the materials that they last longer and minimize maintenance in this sensitive network. This will be a major concern for the next generation of road research. The first steps are theoretical considerations and small tests with laboratory equipment. Before new ideas are implemented in the network, a large-scale demonstrator is a valuable way to demonstrate the benefits of development. An important tool here is an APT (Accelerated Pavement Testing) program in which the demonstrator is loaded with realistic © Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 676–685, 2020. https://doi.org/10.1007/978-3-030-55236-7_70

SMS Structure Measurement System to Optimize APT

677

wheel loads (Steyn 2012). APT programs consist of a loading facility (e.g. Heavy Vehicle Simulator (HVS) and Mobile Load Simulator (MLS)), a research infrastructure, and a coordinated investigation program (Wacker 2019). The combination of all parts should be standardized in order to achieve connections from different projects. Various small changes can be made to support the project objective. All data can be used for the overall research to understand the behavior of road structures. The investigation program is one of the main factors for the evaluation of the APT program. Usually it is necessary to move the loading facility in order to use measuring instruments such as the Falling Weight Deflectometer (FWD) at several points and with different measurement setups. These measurements reduce the loading time and measure only individual points instead of a continuous distance measurement. In order to optimize the use of different measurement techniques, a new measurement system for continuous measurements will be developed, tested under controlled conditions and implemented in APT programs. With the help of the new technology, further investigations and specific evaluations can be used for in-depth research. Another reason for using continuous measurement is to evaluate the change over time. The changes in the road structure are usually gradual and not sudden. This paper covers current APT programs and provides a detailed description of the development and functional testing of the new structure measurement system (SMS). The last part of the paper will show the first implementation of the measurement system in different projects.

2 Current APT Programs The general description of Accelerated Pavement Testing (APT) programs is called “APT is the setup of wheel loads on special constructed or in service pavements. Under controlled and accelerated conditions, the pavement response and performance can be observed in a short period of time (Steyn 2012). APT is important to develop new strategies to analyze pavement structure and innovative materials. Elaborate laboratory tests and full scale APT should be done before testing innovative materials under real road conditions” (Jansen and Wacker 2018). Information on the behavior and performance of the road structure can be provided by non-destructive testing before, during and after the loading phase. In most cases, continuous measurements with integrated sensors are possible. The results obtained are further analyzed by laboratory tests on drill cores after the load tests have been completed. The international Accelerated Pavement Testing conference in 2016 organized by the Transport Research Board (TRB) showed that more and more countries are using APT facilities to develop or calibrate their design methods and to analyze road construction. In addition, APT programs are effective in identifying the most cost-effective method for new developments (Aguiar-Moya et al. 2016). The foundation for the following descriptions is provided by the APT program of the Federal Highway Research Institute (BASt) with the Mobile Load Simulator MLS30 (Jansen and Wacker 2018). In addition to the loading facility complete APT programs consider two more key factors to obtain the best research results.

678

B. Wacker and D. Jansen

First, two different research infrastructures (indoor or outdoor) can be used, depending on the research objective and state of development of the research. When selecting the appropriate research infrastructure, the aim of the research project must be taken into account. APT programs in indoor areas without disturbing factors such as temperature changes and other weather conditions can primarily determine the influence of the load on the road construction. The BASt operates various test halls and can investigate a wide variety of structures. Using the APT program outdoors under normal conditions is more realistic, but the evaluation of measurement data can be more complicated because conditions change regularly. The demonstration, investigation and reference area of BASt (duraBASt) is one possible outdoor test facility and has been located in Germany since 2017. Secondly, the BASt standard APT investigation program comprises systems for recording the bearing capacity (FWD), transverse evenness and surface inspection using images. All these measurements are organized in a fixed measurement sequence, where the time of measurement is defined by the loading cycle. During loading, only continuous measurements are performed on instrumented loading surfaces. There are two main problems with instrumented loading surfaces: firstly, the calibration after the construction process and secondly, the effects on the surrounding material. In addition to the standard APT investigation program, new methods can be used alongside known systems and thus validated. Figure 1 shows the location of the various measurement systems during the APT investigation program. There are five profiles for transverse evenness measurements and six areas for surface inspection images (s1r to s3l). The bearing capacity measurements with the Falling Weight Deflectometer (FWD) are carried out with two different measurement setups. The first measuring setup with the three main measuring points is carried out on the central axis. In the other measuring setup, the ambient situation is also recorded. Each main measuring point is flanked by two further measuring points (half loaded and half unloaded = 25, fully unloaded = 50). The most important measuring system for assessing the entire road structure is the FWD, because the load pulse can also address the deeper layers. These measurements take place at intervals of 100 cm (39.37 in). All other measuring systems provide data for the evaluation of permanent changes (ruts, cracks).

Fig. 1. Measurement positions of the BASt APT investigation program

SMS Structure Measurement System to Optimize APT

679

3 Improve APT Programs with Structure Measurement Systems SMS The development of the Structure Measurement System (SMS) began with the idea of making the APT programs even more meaningful and efficient. Various projects were evaluated and the potential for improvement was identified (Wacker et al. 2016; AguiarMoya et al. 2016). The continuous recording of measured values and the distance-related measurement (stationary measurement) were regarded as improvements. The implementation process took place in various steps and began with the analysis of possible measurement systems. In the further steps, the measurement system was developed, including mounting and measurement data acquisition. In the end, the entire system was used in a functional test to investigate the effects of different load speeds and sampling rates of the sensors. In addition, the implementation was carried out in running APT programs. 3.1

Idea

With standardized APT investigation programs, the measurements are carried out at defined measuring points. These measurement points are measured at fixed times based on the loading progress. In the case of changes, it can only be determined that something has happened between these two measurement periods. But the change process cannot be described in detail. A second aspect is the stationary measurement due to the predefined measurement points. This means that only small areas are analyzed, not the entire load area. There are no direct measurements between the FWD measuring points in Fig. 2 and therefore no information of the road construction available. The idea is to carry out measurements next to the loading wheels. This setting is possible when using twin or super single tires. If twin tires were used, it would also be possible to measure between the two tires. The measuring point would be very close to the maximum load and record the maximum deformation. However, the positioning of the sensor and also the regular control is difficult to realize due to the small distance between the tires. The BASt decided to use super single tires in the APT program because of the bigger impact on the structure. COST 334 (COST334 2001) described that the smaller contact area of a super single tire puts twice as much strain on the road structure as that of twin tires. The measurements were to be carried out with a non-contact system with defined starting positions and options for operating the system with interval or flexible measurements. Based on these measurements, it should be possible to detect and locate changes during loading.

Fig. 2. Concept for SMS Structure Measurement System

680

B. Wacker and D. Jansen

No further calculations are to be made with the current data. The system is to be used exclusively to indicate changes in the structure of the tests and to use established measurement systems in a targeted manner. A new parameter for analyzing the data has been defined (b1). This parameter describes the angle set for a linear connection of the front (x1) and rear (x2) sensors. Therefore, a value per 1.0 cm (0.39 in) should be used for the evaluation. The value should be calculated from the measured signals recorded at a high sampling rate. Due to the different load speeds, different sampling rates are required to ensure high resolution. Each calculated value should consist of at least five data points, which means that the distance between data points must be less than 0.2 cm (0.08 in). 3.2

Realization

The implementation process consists of two phases. The first phase focused on the feasibility study, while the second phase focused on the final implementation of the overall system. Various existing systems were analyzed and evaluated with regard to the measurement idea. A triangulation based laser system was chosen due to the high individuality and the possibility to install the system in the loading device. The space conditions played a decisive role and thus measuring systems requiring more space (Pavement Profile Scanners, Doppler laser sensors) were excluded (Wacker 2019). Due to the fixed installation, the same measurement setup can be used at any time, even if the APT system has to be moved. During the preparation of the sensor mounting system, the first idea in Fig. 2 was rejected. Geometric requirements and measurement quality in relation to the same measurement line were decisive in the decision. Figure 3 shows the end position and structure of the sensor mounting system.

Fig. 3. Position of holder system (left and middle schematic, right final)

The resolution of the selected laser sensors is 3 µm at 20 kHz. The measuring range is between 13–33 cm (5.12–12.99 in) and thus includes the measuring distance of 20 cm (7.87 in) in the developed measuring setup. Various factors have resulted in the first prototype being installed on only one side of the loading wheel. Preliminary investigations with accelerometers showed that a measuring length of 200 cm is possible without significant vibrations (Wacker 2019).

SMS Structure Measurement System to Optimize APT

681

For the feasibility study, three different loading speeds at a sampling rate of 1.5 kHz were used. A wireless system was set up to transmit the measurement data to the control computer. All measuring data within the measuring range were transmitted and stored. Two surfaces (dense and porous) were measured with explainable and repeatable anomalies. On one of the two loading areas, the loading track ended on an already damaged structure (damage between surface and binder layer) caused by an earlier loading program. Due to this damage, the value of the front sensor has changed significantly. The experience with data transfer during the loading process and the damage detection allowed the second phase of the development for the Structure Measurement System (SMS). In the second phase of development, all the experience gained was taken into account and used for the final description of the measurement system. It was defined that one measurement is the average of five load cycles in a row with the measurement system. The measurements are to be performed synchronously by both sensors and the starting point for storing the data should be provided with a trigger at a fixed position. During evaluation, the recorded data files of each individual combination (sampling rate and load speed) are averaged and reduced to 200 information points. For example, a measurement with a sampling rate of 5 kHz and a speed of 2,000 mm/s, a total of 5000 data points (distance 0.04 cm (0.02 in)) is stored. By averaging over 1 cm (0.39 in), the point density is reduced by a factor of 25.

In addition to the measurement details, requirements for the data acquisition system were defined in order to operate the system under stable conditions. In this case, a standardized storage system was also specified to simplify the evaluation process. At the end of the project, the evaluation process was modified to use standardized values. Due to the normalization process, only the changes based on the first usable measurement are displayed. This makes it easier to identify positions where further investigation makes sense. 3.3

Functional Test

After the implementation of the system in the loading facility, various tests were carried out to verify repeatability, stability and evaluation procedures using a reference track. In order to avoid vibrations caused by different conditions, these tests were carried out on concrete surface, which did not change during the tests. Only the reference track at the top of the loading device was used for the evaluation process (Fig. 4). The test setup for the functional test was chosen with five different loading speeds and three different sampling rates. Each combination was measured three times (45 experiments in total). For each experiment, five interval measurements were performed every five minutes, followed by two flexible measurements. Under these conditions, one experiment lasts about 30 min (Wacker 2019).

682

B. Wacker and D. Jansen

Fig. 4. Reference track beneath the roof of the loading facility

The evaluation process showed that different load speeds at the same sampling rate have no influence on the degree of accuracy. Different sampling rates at the same load speeds have a minimal influence on the accuracy level. The calculated difference for the parameter b1 was 0.04°. In view of this small difference, the system was defined as stable and without relevant influence of different setups.

Fig. 5. Anomalies during 45 experiments

SMS Structure Measurement System to Optimize APT

683

As expected, problems with data provision were detected during the first application. One problem was that not enough usable data was stored. To calculate the parameter for the measurement, it must be possible to use at least three load cycles per sequence. The following problems were detected: – – – – –

rear sensor stored the wrong data (status 1), front sensor stored the wrong data (status 2), front and rear sensor stored the wrong data (status 3), not enough data are available (status 4), no data are available (status 5).

Figure 5 shows the distribution of anomalies per experiment. Across all experiments, 78% of the data can be used (status 0). 11% of the data are not available (status 5) and 8% of the data have wrong inputs for the rear sensor (status 1). All other anomalies are not significant with a total of 3%. In order to describe the reason for the various anomalies, each experiment was evaluated in detail. For a good 80% of the first interval measurement (int1) at least 3/5 load cycles are available, which can be used for further calculation steps. For the following interval measurements, this value is continuously reduced to almost 70% until the fifth interval measurement. As already described, two flexible measurements were performed after the interval measurement. With this type of measurement, almost 90% of the measurements can be used for further evaluation processes. These anomalies led to an upgrade of the data acquisition system. Before flexible measurements begin, the sensor cache is emptied, resulting in a more stable data transfer. This procedure was also programmed for the start of each interval measurement. 3.4

Implementation in APT Programs

After the functional test, the system was integrated into real APT programs. For quality control reasons, the road surface and the reference track under the roof were measured simultaneously. A program was carried out in the test hall to optimize the evaluation process. A big advantage was the loading of two constructions during one test. Especially the thin construction showed as expected changes compared to the first rollover. Thus the functionality of the system could be shown very well. Figure 6 presents the final situation after the experiment. Due to the normalized values it is possible to detect the significant positions quickly.

684

B. Wacker and D. Jansen

Fig. 6. Correlation between measurement data and different structures

4 Conclusion APT programs are complex structures that are influenced by various factors. A new continuous measurement system has been developed to support non-destructive testing equipment and procedures during the test period. The new system was tested under controlled conditions before being used in APT programs. In particular, the functional test showed that the SMS Structure Measurement System can be used and provides repeatable data for the evaluation process. During the test, minor errors (status 1 to 5) were discovered which could be minimized by adjustments in the data acquisition system. In a revised version, the sensor holding system can also be mounted on other APT facilities. In future, this prototype system will be further developed and more data will be generated from ongoing projects. Thus, valuable data for the optimization of APT programs can be obtained.

References Jansen, D., Wacker, B.: Toolbox for APT program monitoring at BASt and optimized data handling on new test facility duraBASt. In: Presented at 97th Annual Meeting of Transport Research Board, Washington D.C. (2018)

SMS Structure Measurement System to Optimize APT

685

Steyn, W.: NCHRP synthesis 433: significant findings from full-scale accelerated pavement testing. Transport Research Board, Washington, D.C. (2012) Aguiar-Moya, J.P., Vargas-Nordcbeck, A., Leiva-Villacorta, F., Loría, L.G.: The roles of accelerated pavement testing in pavement sustainability. Springer (2016). https://doi.org/10. 1007/978-3-319-42797-3 COST 334: Effects of Wide Single Tyres and Dual Tyres. Brussels (2001) Wacker, B., Scherkenbach, M., Rabe, R., Golkowski, G.: Belastungseinrichtung Mobile Load Simulator MLS30: Sensorik zur Beanspruchungsdetektion im ersten gemeinsamen Versuchsbetrieb (Load equipment Mobile Load Simulator MLS30: Sensors for load detection in the first joint test operation). Report S101 of the Federal Highway Research Institute (2016) Wacker, B.: Zeitraffende Belastungsversuche mit integriertem Einsatz zerstörungsfreier Messsysteme (Accelerated Pavement Testing with integrated use of non-destructive measuring systems). Ph.D. manuscript - not published yet (2019)

New Pavement Concepts and APT Challenges

Evaluation of a Solution for Electric Supply of Vehicles by the Road, at Laboratory and Full Scale Pierre Hornych1(&), Thomas Gabet1, Mai Lan Nguyen1, Fabienne Anfosso Lédée1, and Patrick Duprat2 1

MAST-LAMES, Univ Gustave Eiffel, IFSTTAR, F-44341 Bouguenais, France [email protected] 2 Alstom, Saint-Ouen, France

Abstract. Nowadays, the development of electric heavy goods vehicles is hindered by the limited capacity of batteries. An alternative solution to batteries is to develop electrified road systems, capable of supplying and/or recharging electric vehicles while they are driving. In this project, an Electric Road System (ERS), i.e. a solution for electric supply by a feeding track embedded in the pavement, is studied. The system consists of metallic conductive segments, fixed on a rubber support, installed in a groove in the middle of the pavement. Laying such a device in a bituminous pavement requires an efficient bonding of the rubber with the pavement materials, and a good resistance to traffic and climate. Studies were made at laboratory scale and at full scale to select appropriate joint materials and assess the load resistance and durability of the rail/pavement system. Two types of laboratory tests were performed on asphalt concrete samples integrating the rubber elements: wheel tracking tests, at temperatures of 20, 40 and 60 °C, and climatic tests (temperature cycles between −20 °C and 20 °C). Modelling of the tests was also carried out. After evaluation of different products, an appropriate joint material was selected. Then, an accelerated test was performed with the FABAC traffic simulator, to validate the behavior of the electric supply system inserted in a bituminous pavement, at full scale. The pavement structure was instrumented, submitted to 500 000 dual wheel load cycles at 65 kN, and monitored using deflection measurements and profile measurements. The test confirmed the satisfactory behavior obtained at laboratory scale. Keywords: Electric Road System
Full scale testing
Mechanical behavior Thermal behavior
Bonding




1 Introduction Electric Road Systems (ERS) are mainly developed to decarbonize road freight transport on long distances. This is why this type of infrastructure will be installed on highways and main roads. The ERS developed by Alstom, a ground-based feeding system, is based on a technical solution used in railways, for power supply of

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 689–698, 2020. https://doi.org/10.1007/978-3-030-55236-7_71

690

P. Hornych et al.

catenaryless tramways (Fort 2013). In the railway solution, the feeding rails are integrated inside the track slab. In this project for road application, the goal is to adapt the ERS to heavy goods vehicles, by installing a mobile current collector underneath the vehicle, which will allow to collect the electric current from the feeding track while driving. One of the main challenges in this ERS development is the integration of the feeding track inside the road pavement (Chen 2016). To accept this type of infrastructure, the road authorities want to be certain that it will not damage the road pavement, that it will be safe for the users and that it will not impact the maintenance works. The integration in the road pavement and the interface with the vehicle are shown on Fig. 1.

Fig. 1. Principle of the Electric Road System and integration of the track in the road pavement

The objectives of the study presented in this paper were to find the best solution for ERS integration in the road, and to validate that it will not affect the mechanical durability of the road structure. For this purpose, laboratory tests were first performed to select an appropriate joint material for integrating the ERS in a bituminous pavement, and to verify its bonding strength, and its resistance to mechanical and thermal loading. Then, a full-scale test was performed to validate the durability of the selected solution under wheel loading.

2 Laboratory Testing The installation of the electric road system (ERS) in an asphalt pavement required to find appropriate solutions to ensure a good bonding between the asphalt materials and the rubber support of the steel rails. For that purpose, it was decided to assess several possible bonding materials. It was decided to test the resistance of the asphalt/bond/rubber assembly according to both mechanical and thermal loadings, in order to simulate pavement loading conditions. For that purpose, laboratory tests were first carried out on small slabs of asphalt concretes in which rubber blocks were incorporated, simulating the electric rail system. Two types of tests were performed: wheel tracking tests, which consisted in submitting the slab to bi-directional wheel

Evaluation of a Solution for Electric Supply of Vehicles

691

loading, at controlled temperatures (20, 40, and 60 °C); thermal loading, which consisted in subjecting the slab to cyclic temperature variations, from +20 °C to +60 °C, and from −20 °C to +20 °C. 2.1

Wheel Tracking Tests

50

50

100

The wheel tracking tests were performed using the French large wheel tracking apparatus (standard NF-EN 12697-22). This apparatus allows to test simultaneously 2 samples, of 500  180  100 mm, under the loading of a wheel with a contact pressure of 600 kPa, and a loading rate of 1 cycle/second. It should be noted that this test is a qualification test for materials, carried under very severe loading and temperature conditions, not meant to represent the effective field conditions. For the project, it was decided to perform the tests at 3 different temperatures, 20 °C, 40 °C, 60 °C. The temperature of 60 °C, used in the standard, represents an extreme temperature seldom reached in pavement surface layers in Europe. An asphalt concrete of AC10 type with a 30/50 pen grade bitumen, typical of wearing courses used on French highways, was chosen for the tests. The asphalt concrete slabs were manufactured, using a laboratory plate compactor. Then each slab was sawn, to incorporate the rubber elements, either in the center of the slab, or on one side, as shown on Fig. 2. This allowed to apply the wheel loads on the rubber rail, or on the joint between the rubber rail and the asphalt material. The interface between the asphalt concrete and the ERS support was filled with an elastomeric resin. The wheel tracking test procedure included 3 load sequences, of 30 000 load cycles each, at increasing temperatures (20 °C, 40 °C and 60 °C), during which the rutting of the specimens was measured. The tests presented here were performed with an elastomeric resin used for sealing of rail track joints, and with the two positions of the rubber element. Figure 3 shows the specimen with the rubber on the side. Rut depths were measured on each specimen at 15 different positions (see Fig. 3).

90

120 180

180

Fig. 2. Slab configurations used for the laboratory wheel tracking tests (dimensions in mm) – in grey the rubber elements

692

P. Hornych et al.

Fig. 3. Wheel tracking test on the slab with the embedded rubber element, and location of the rut depth measurements on each slab

The sample with the centered rubber element gave very good results. The rutting value did not exceed 0.5 mm at 60 °C after 30 000 cycles, which is very low. Indeed, the maximum permissible rut depth for this asphalt concrete at 60 °C is 10 mm after 30 000 cycles. The sample with the rubber element on the side presented more rutting. Figure 4 presents rut depths measurements obtained on the slab with the rubber element on the side, during the test at 60 °C. The final average rut depths reached 3.4 mm on the joint, and 9.3 mm on the asphalt material (with a local maximum value of 12.7 mm). These results were considered good, as the average rutting didn’t exceed the maximum permissible value of 10 mm. It should be mentioned again that this temperature of 60 °C is extreme, and that the effective number of vehicles passing on the interface will in reality be very low.

Fig. 4. Rut depths measured during the wheel tracking test at 60 °C on the specimen with the rubber element on the side.

2.2

Thermal Cycling Tests

Another possible deterioration mechanism of the embedded ERS is thermal cracking, due to the different thermal expansion coefficients of the asphalt concrete and rubber. Firstly, in order to evaluate the risk of such thermal cracking, finite element simulations were

Evaluation of a Solution for Electric Supply of Vehicles

693

carried out. A simplified, thermo-elastic behavior was considered, with an elastic modulus of 20 000 MPa for the asphalt material and 1000 MPa for the rubber, and thermal expansion coefficients of 30 µe/K and 110 µe/K respectively. A 600 mm  200 mm  120 mm slab, with an embedded rubber element (representing a laboratory specimen) was simulated and subjected to a temperature decrease of 40 °C (from 20 °C to −20 °C). Figure 5 present the results of the calculations. As shown on this figure, the different thermal properties of the rubber and asphalt concrete lead to a higher contraction of the rubber element, thus generating in particular compressive stresses under the groove inside the AC material, and tensile stresses at the corner of the groove. Such tensile stress in the AC material may cause cracking, especially under repeated thermal cycles. On Fig. 5, the maximum tensile stresses generated are of the order of 0.2 MPa. Although these simulations are very simplified, and do not describe correctly the longterm viscoelastic behavior of the asphalt mix, it was concluded that a significant risk of thermal cracking exists. It was decided to perform laboratory thermal cycling tests to investigate this phenomenon. Several thermal tests were performed to assess different types of interface and climate conditions. Only the results obtained with the thick interface filled with elastomeric resin are presented here (Fig. 6). The sample was submitted to 88 temperature cycles between −20 °C and 20 °C, at a rate of 2 cycles per day. This simulates a very cold climate. After 88 cycles, the sample presented no crack (see Fig. 6), in contrary to previous samples with other interface materials, not presented here. This was attributed to the high elasticity, low stiffness and good adhesion of the resin, which was able to resist to the high strain levels generated by the temperature cycles. After these tests, it was decided to retain the resin for full scale testing.

Fig. 5. On the left: numerical model of the asphalt slab with the embedded rubber element. On the right: compressive/tensile stresses induced in the specimen by a temperature decrease of 40 °C.

Fig. 6. View of the specimen with elastomeric resin after 88 temperature cycles

694

P. Hornych et al.

3 Full Scale Testing 3.1

Experimental Pavement Section and Test Program

Due to the limitations of the laboratory tests, which cannot reproduce exactly field conditions, it was considered important to validate the behavior of electric road system by a full-scale test, carried out with the FABAC traffic simulator (Aunis et al. 1998). This small linear traffic simulator has four wheel modules, driven by a chain and electric motor (Figs. 7 and 8). Each wheel module can be equipped with single or dual wheels, and apply loads varying between 30 and 75 kN. The length of the loaded section is 2 m. The loading speed can vary between 1 and 5 km/h, and the maximum loading rate, under continuous operation, is 300 000 load cycles per week. The pavement section in which the electric supply rails were embedded was 20 m long and consisted of 3 asphalt layers: a 2.5 cm thin asphalt concrete wearing course, a 5 cm binder course and a 12 cm thick asphalt base course. It was built on a 30 cm thick granular subbase, with a bearing capacity of about 70 MPa. The rails were placed at an angle of 7° with the wheel path of the FABAC machine, to simulate the loading of a vehicle crossing the rails when changing lane (see Fig. 9), which was considered as the worst loading situation. To install the electric supply rails in the pavement, a trench was first milled in the middle of the lane, then the bottom of

Fig. 7. Principle of the FABAC machine

Fig. 8. View of the FABAC machine

Evaluation of a Solution for Electric Supply of Vehicles

695

the trench was filled with the elastomeric resin. The electric supply track was then put in place, and finally, the joints were filled with resin. The test program consisted in applying 500 000 load cycles, with dual wheels, loaded at 65 kN, and a loading speed of 3.6 km/h. This number of load cycles represents worst case conditions. On a road, loads will cross the rail only during overtaking or lane change maneuvers, and the number of loads applied effectively to the rail will be much lower. The test was performed in April 2019. Figure 10 shows the rails after installation in the pavement. The monitoring of the pavement included: • Visual inspections of the pavement and rail system, to detect any deteriorations. • Measurements of transverse profile, before loading, and after 250 000 and 500 000 load cycles, at 3 different longitudinal positions. • Temperature measurements, at 4 different depths. • Benkelmann beam deflection measurements, at regular intervals, on the rails, at 3 different longitudinal positions.

Fig. 9. Positioning of the electric rails on the experimental pavement

Fig. 10. View of the experimental pavement section with the rail system, and position of the transverse profile measurements (lines P1, P2 and P3)

696

3.2

P. Hornych et al.

Temperature Measurements

During the whole test, temperatures were measured continuously (at 10 min intervals) at 4 different depths in the pavement structure (see Fig. 11).

Fig. 11. Temperature variations at different depths in the experimental pavement section

The FABAC machine is an outdoor equipment, with no temperature control system, and hence the temperatures depend on the local climatic conditions. During the test, the temperatures were mild, varying between 5 °C and 26 °C at the surface, and between 7 °C and 21 °C at a depth of 5 cm (close to the bottom of the ERS rail). 3.3

Transversal Profile Measurements and Visual Inspections

Transverse profile measurements were made before the start of the test, and then at 250 000 and 500 000 load cycles, with a profilometer equipped with a laser displacement sensor. The profiles measured at mid-length of the rail (profile P2) are presented on Fig. 12. The metallic conductive segments are clearly visible on these profiles. The maximum rut depth after 500 000 load cycles reached approximately 6 mm, which is a satisfactory value, given the high number of loads applied, which largely exceeds the number of loads expected on a real road. On national roads in France, the permissible rut depth level is typically 10 mm.

Fig. 12. Transverse profile measurements after 0, 250 000 and 500 000 load cycles.

Evaluation of a Solution for Electric Supply of Vehicles

697

Visual inspections were also made regularly during the test, and Fig. 13 shows a view of the test section after 500 000 load cycles. The observations showed that rutting developed mainly during the first 100 000 load cycles, and was mainly due to permanent deformations of the asphalt material. They also indicated no deterioration of the electric supply track, a good adhesion of the elastomer joint on the asphalt and rubber, and no cracking of the joint. 3.4

Deflection Measurements

Deflections were measured approximately every 100 000 load cycles, at 2 different longitudinal positions, on the rubber element. The measurements were made in the middle of the wheel path, using a Benkelman beam. The results, presented on Fig. 14, indicate a slight increase of deflections during the first 100 000 cycles, and then no significant evolution during the rest of the test, except some variations due to temperature changes (the measured temperatures are indicated on Fig. 14). The deflection levels varied between 20 and 40 mm/100. They were lower than the deflection estimated by linear elastic calculations for this structure (54 mm/100 at a temperature of 12 °C), which indicates a good stiffness of the experimental structure.

Elastomer joint Supply track

Fig. 13. View of the supply track after the accelerated loading test

Fig. 14. Benkelman beam deflection measurements on the experimental section

698

P. Hornych et al.

4 Conclusions A technical solution for the insertion of the ERS system in a bituminous pavement was developed and tested. Standard or traditional tools for assessing the mechanical behavior and durability of asphalt pavements were adapted and successfully applied to select and validate the technical solution. The test results indicated satisfactory performance, with limited rutting, and no visible deterioration of the supply track. The next step is now to build a 50-m long full-scale test track, to validate the safety of the solution using test vehicles. In case of success, the final step before large-scale deployment on roads will be to build a demonstrator of a few kilometers on a circulated road.

References Fort, T.: Essai Alstom. Evaluation sous trafic lourd des systèmes d’alimentation des tramways par le sol. Rev. Gén. Routes Aménage. 914, 81–85 (2013) Chen, F.: Sustainable implementation of electrified roads: structural and material analyses. Ph.D. thesis, KTH Royal Institute of Technology, Sweden, November 2016, p. 91 (2016) Aunis, J., Balay, J.-M.: An applied research programme on continuous reinforced concrete pavements: the FABAC project. In: 8th International Symposium on Concrete Roads, Lisbon, 13–16 September 1998 (1998)

Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes Bertrand Pouteau1(&), Kamal Berrada1, Sandrine Vergne1, Mai Lan Nguyen2, Stéphane Trichet2, Thierry Gouy2, and Pierre Hornych2 1

2

Eurovia Management, Technical and Research Direction, Mérignac, France [email protected] MAST-LAMES, Univ Gustave Eiffel, Ifsttar, F-44344 Bouguenais, France

Abstract. Road construction companies are developing innovative road solutions which integrate new functions, such as the ability to generate energy. The authors focus here on the use of road solutions harvesting solar and geothermal heat energy. In 2014, PIARC’s report entitled “Alternative solutions for fossil fuels for the road system” listed the main limitations blocking the development of these innovations among which the resistance to traffic. The authors present a recent accelerated test at scale one of a pavement that incorporates geothermal tube networks in the first 10 cm of the pavement. The main goal is to assess the impact of the tubes on the durability of the overlay and structure. The first part of the paper describes the experimental site, located at the toll station of Saint Arnoult, next to the A10 highway, near Paris. It is composed of 4 trial sections. Tubes were buried at construction stage at 2 different depths. The second part presents the experimental setup: (i) the FABAC traffic simulator used for the tests and (ii) the survey methods used to follow the evolution of the structures. The third part present the measurements made during the tests and the fourth part is dedicated to the analysis of the results and conclusions. Keywords: Energy harvesting
Tube network
Full scale testing
Durability

1 Introduction French roads represent 1.2% of the territory (about 6000 km2). A large part of them are paved with black asphalt materials, which leads them to capture the sun’s rays and increase their temperature. Road construction companies are developing innovative road solutions which integrate new functions, such as the ability to generate energy. A recent study shows that it is entirely appropriate to build pavements to harvest solar and geothermal heat energy (Pouteau et al. 2019). In 2014, PIARC’s report entitled “Alternative solutions for fossil fuels for the road system” listed the main limitations blocking the development of these innovations, among which the resistance to traffic (Comité technique AIPCR 2014). Since energy harvesting roads are in their early development, it is necessary to perform accelerated tests at scale one to acknowledge the durability of new solutions.

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 699–707, 2020. https://doi.org/10.1007/978-3-030-55236-7_72

700

B. Pouteau et al.

The aim of the study is to assess the impact of the tubes on the durability of the overlay and structure at full scale. This paper presents tests performed on an experimental track that incorporates geothermal tube networks in the first 10 cm of the pavement. Descriptions of the track and of the experimental setup are provided in Sects. 2 and 3. The experimental results are discussed in Sect. 4.

2 Experimental Site Description Experimentations on the Power Road solution, described in Sect. 2.1, are carried out with the support of ADEME within the framework of a national call for projects and the French government’s Programme d’Investissement d’Avenir (PIA). In 2017, an experimental site was built on the Cofiroute network (Vinci Motorways). Section 2.2 presents the entire site and underlines the part which was setup to enable accelerated testing. A description of the test track construction is provided in Sect. 2.3. 2.1

Power Road Description

The tested solution is branded Power Road. The heat is harvested by the flow of a liquid (water and glycol) in tubes buried in the road. The tubes are laid in the asphalt material in the top 10 cm of the pavement under the wearing course. Tube’s standalone life time is expected to be around 50 years. But tubes may weaken the pavement. Two construction methods can be used to achieve tube integration. The indentation method consists of applying prefabricated elements of the exchanger (network of tubes) on the uncompacted hot asphalt mix, then compacting the mix (paving between the passage of pavers and compactors). The exchanger is integrated into the binder course layer during its installation as shown in Fig. 1. The grooving method consists of grooving the binder course using a specific milling machine, then laying the exchanger before installing a filling course and the wearing course. Only the indentation method was tested in the framework of this study.

Fig. 1. Installation of the tubes by indentation (left: positioning of a grid) (right: view of the pavement after indentation of all the grids)

Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes

701

The later rehabilitation of the pavement is not in the scope of this study since it was discussed in a communication (Pouteau et al. 2019). The pavement is designed to be recycled with the tube in them: no metallic components are laid in the pavement structure. A scale one rehabilitation test carried in 2018 demonstrates that recycled products complies with the construction guide and can be reused as is. 2.2

Site Location

The experimental site is located on the access road to the service area of the A10 motorway toll barrier in Saint-Arnoult-en-Yvelines (Pouteau et al. 2019). Figure 2 shows the layout of the various components of the installation: 1/ 500 m2 of pavement equipped with tubes under the wearing course; 2/ A technical room for the control system; 3/ A field of geothermal probes for energy storage; F/ an experimental area dedicated to performing an accelerated traffic test with the FABAC test machine.

100m

200m

Fig. 2. Layout of experimental site

2.3

Experimental Track Description

The experimental track was built on 11th September 2017. It is 18 m long and 2 m wide. It was divided in two zones (Z1 and Z2) with different thicknesses of layers (see Fig. 3). The wearing course and binder course are EB10 with 35/50 bitumen. HMA corresponds to the existing pavement.

Fig. 3. Experimental track structure description.

702

B. Pouteau et al.

3 Experimental Setup Description To study the mechanical behavior of the Power road test sections, it was decided to perform full scale accelerated tests, using the FABAC traffic simulator of IFSTTAR (Chabot et al. 2008). 3.1

FABAC Traffic Simulator and Test Program

The FABAC machine is a small linear traffic simulator, which was developed initially for testing the fatigue behavior of concrete slab pavements. The FABAC machine has four wheel modules, driven by a chain and electric motor (see Figs. 4 and 5). Each wheel module can be equipped with single or dual wheels, and apply loads varying between 30 and 75 kN. The length of the loaded section is 2 m. The loading speed can vary between 1 and 5 km/h, and the maximum loading rate, under continuous operation, is 300,000 load cycles per week. The FABAC machine is 10 m long and 3 m high and was designed so that it can be transported by truck to test sites located on real roads. Thus, in this project, the FABAC machine was transported to the A10 test site, in order to test the two experimental sections (Z1 et Z2). The machine was installed on the site on September 14th, 2017 and was moved back to IFSTTAR in May 2018. A view of the FABAC machine installed on the A10 test site is shown on Fig. 6.

Fig. 4. Principle of the FABAC machine

Tests were performed on the two experimental pavement sections, with 65 kN dual wheel loads, and with a loading speed of 1 m/s (3.6 km/h). Section Z1 was tested between October 10, 2017 and February 1st, 2018. 2,686,000 dual wheel loads were applied on this section. Then, section Z2 was tested between February 1st, 2018 and April 15th, and 447,000 load cycles were applied.

Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes

703

Fig. 5. View of the FABAC machine

Fig. 6. View of the FABAC machine installed on the A10 track

3.2

Survey Methods

The monitoring of the experimental sections included: • visual distress surveys; • measurements of the transverse profile, carried out at 3 positions, located 50 cm apart; • taking photographs of the test sections. The transverse profile measurements were performed with a profile measurement system consisting of a metal beam, supporting a mobile carriage, equipped with a laser displacement transducer. The system can measure the transverse profile over a length of 1.5 m, and with a vertical measurement accuracy of 1 mm. A view of the profilometer is presented in Fig. 7.

704

B. Pouteau et al.

Fig. 7. View of the laser sensor profile measurement system

4 Experimental Results 4.1

Performance of Section Z1

On section Z1, inspections and profile measurements were performed after 923,000, 1,521,000, 1,835,000 cycles, and at the end of the test, after 2,686,000 cycles. Figure 8 presents a picture of the test section, at the end of the test, and a closer view of the aspect of the pavement surface, in the middle of the wheel path. In this figure, it can be seen that at the end of the test, the test section presented some permanent deformation on the wheel path (the pattern of the tires is clearly visible), but no cracking, and no other significant deterioration.

Fig. 8. View of section Z1 at the end of the full scale test, and close-up view of the deformations in the wheelpath

To analyze the evolution of permanent deformation, Fig. 9 presents transverse profiles measured at different numbers of load cycles, and at two different positions on the test section: position P1, located at a distance of 45 cm from the beginning of the

Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes

705

loaded area, and section P2, located at a distance of 95 cm (close to the middle of the loaded area). The results indicate that: • Permanent deformation developed mainly at the beginning of the test. After 923,000 cycles, the maximum permanent deformation reached 7 mm at position P1, and 6 mm at position P2. • Then, during subsequent loading, permanent deformation tended to stabilize, and only a small increase was observed; At the end of the test (2,686,000 loadings), permanent deformation attained 10 mm at position P1, and 6 mm at position P2. • Permanent deformation was higher at the beginning of the loaded area (position P1) than on the rest of the test section.

Fig. 9. Transverse profile measurements at different numbers of load cycles at two different locations on section Z1.

Because the test was carried out during a cold period (between end of October and end of January), and because the pavement surface presented only downward deformation, and no uplift along the rut, the observed permanent deformation was attributed to a post compaction of the asphalt layer, which could be due to the relatively high initial void content of the asphalt mix (initial mean value of 8.1%). No specific deformation was observed above the tubes located in the asphalt layer, indicating that there was no damage induced by the heating tubes. The maximum final permanent deformation depth (10 mm) remains lower than the maximum acceptable rut depth for low to medium traffic pavements (15 mm).

706

4.2

B. Pouteau et al.

Performance of Section Z2

On section Z2, inspections and profile measurements were performed only at the end of the test, after 447,000 load cycles. The final aspect of section Z2, at the end of the test, is shown in Fig. 10. Like section Z1, section Z2 presented only limited rutting on the wheel path, and no other type of deterioration (no cracking, no deformation caused by the heating tubes).

Fig. 10. View of section Z2 at the end of the full-scale test, and detail of the wheel path

Figure 11 presents the transverse profile measurements made on section Z2, after construction and at the end of the test, at two different positions. On this section, which received 447,000 load cycles, the maximum permanent deformation depth attained 2 mm in the center of the section (profile P2), and was higher at the end of the section, attaining 5 mm. As with section Z1, it is probable that this permanent deformation is due to post compaction of the asphalt layer.

Fig. 11. Transverse profile measurements at two different locations on section Z2, before and after loading

Full Scale Testing of an Energy Harvesting Test Road Integrating Tubes

707

In conclusion, results on section Z2 confirmed those obtained on section Z1, indicating only a moderate permanent deformation, most probably due to a postcompaction of the asphalt material, due to its high initial void content.

5 Conclusions After 2.6 million loading cycles of 65 kN dual wheel loads applied on the experimental track, the following conclusions can be made: – no cracks related to the presence of tubes introduced into the pavement by the process were observed. A very fine surface crack was observed around 2 million cycles, but it was no longer visible during the final observation; – no permanent deformation related to the presence of tubes in the pavement was found. – integration of tubes by the “indentation method” seems to have moderate to no impact on the resistance of the pavement to traffic loading. – Tests were done on a new pavement. Aging that leads to material stiffening can have an impact on the cracking behavior. The long-time performance of the pavement, considering such aging phenomena is beyond the scope of this study. This study also acknowledges that the FABAC machine remains a convenient device to investigate durability of innovative pavements. Acknowledgements. The work presented in the paper is part of a research carried out with the support of ADEME (the French Environment & Energy Management Agency) and the Investments for the future program of the French government.

References Chabot, A., Balay, J.M., Pouteau, B., de Larrard, F.: FABAC accelerated loading test of bond between cement overlay and asphalt layers. In: 6th RILEM International Conference on Cracking in Pavements, Chicago, United States, 16 June 2008, pp. 671–680. Taylor & Francis (2008) Comité technique AIPCR: Solutions alternatives aux combustibles fossiles pour le réseau routier. 2014R01FR (2014). ISBN 978-2-84060-343-6 Pouteau, B., Berrada, K., Vergne, S.: Positive energy road: feedback on heat harvesting through pavement solution. Paper Presented at the XXVIth World Road Congress, Abu Dhabi, United Arab Emirates, 6–10 October 2019 (2019)

Future APT – Thoughts on Future Evolution of APT Wynand J. vdM. Steyn(&) University of Pretoria, Private Bag X20, Hatfield 0083, South Africa [email protected]

Abstract. Accelerated Pavement Testing (APT) is defined as the controlled application of a wheel loading, at or above the appropriate legal load limit, to a pavement system to determine pavement response in a compressed time period. In a 4th Industrial Revolution age, issues around the datafication of life, data science, big data, transportation evolution through Mobility as a Service (MAAS), optimization of logistic and supply chains and automation of various aspects of life, including vehicles, become everyday features. It is important to re-evaluate the future of APT. Traditional APT approaches applied traffic load volumes to a section of the pavement in a much accelerated methodology, due to the combination of continuous load application of often increased load levels, and the relatively low traffic volumes on most roads. Currently, many roads carry volumes in excess of the abilities of APT devices to keep up with normal load applications. The number of international APT facilities grew constantly over the last decade, and therefore large masses of APT-based road response data are generated. In this future APT environment, aspects such as the ability of APT to still provide acceleration of traffic loading, as well as the judicious data fusion and analysis of data from various disparate sources/programmes to provide solutions to increasing need for the provision of economical and safe pavements, are addressed in this paper. Keywords: Future APT


Data science
Collaborative analysis

1 Introduction Accelerated Pavement Testing (APT) is defined as the controlled application of a wheel loading, at or above the appropriate legal load limit, to a pavement system to determine pavement response in a compressed time period (Steyn 2012). APT has been used for decades to evaluate road pavement structures with the main objective to evaluate pavement material response and performance (Hugo and Epps 2004). Traditional APT approaches applied traffic loading using equipment that was designed to provide tyre loads in various configurations to a defined pavement section in an accelerated format. This trafficking typically led to deterioration of the pavement structure, which was monitored using a range of sensors and instruments. These data were analysed to obtain an improved understanding of the behaviour of the pavement structure. Analysis of such APT data with laboratory information and field

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 708–717, 2020. https://doi.org/10.1007/978-3-030-55236-7_73

Future APT – Thoughts on Future Evolution of APT

709

understanding, typically led to improved design, maintenance and management models for pavement structures. Acceleration of the traffic loads applied to the pavement structure was typically done through application of the loads to a relatively short test section (enabling a high number of applications applied in a relatively short period) and increase of tyre loads to make use of the exponential damage principle that states that the damage to a pavement has an exponential relationship with the applied loads. This combination of load applications provided an accelerated loading environment which, due to the relatively low traffic volumes on most roads, were adequate to develop pavement deterioration models in a relatively short period (Metcalf 1996; Hugo and Epps 2004). Acceleration of loads on pavement structures are mainly done to enable a quicker understanding of the anticipated behaviour of a pavement structure than waiting for the normal life of the pavement to play out before typical deterioration will take place (Metcalf 1996). Therefore, the application of accelerated trafficking should not affect the response mechanism of the pavement, but should only shorten the waiting period of the expected deterioration outcome. Obviously, the APT environment also allows for the incorporation of smaller, managed changes to the load and environment to develop an improved understanding of the effect of such changes to the expected pavement behaviour. Through the last decades of the 20th century and the first two decades of the 21st century, global road vehicle volumes increased exponentially (Fig. 1). This exponential growth in traffic volumes caused the design traffic for road pavement structures to increase, with a subsequent increase in the rate at which traffic is applied to road pavements. Per illustration, if an APT device that can apply 6 000 passes per hour was used to apply all traffic loads in 1960, it would take 2.4 years to complete the test (full 24 h/365 day operation). In 2030, 43.8 years would be required to conduct the same test. Although this is an illustrative example, it does indicate that merely using the traditional way of APT may not provide timely results in the future.

Fig. 1. World total vehicle numbers 1960 to 2019, extrapolated to 2030

The number of international APT facilities grew constantly over the last 3 decades, and therefore large volumes of APT-based road response data are generated. In a future

710

W. J. vdM. Steyn

APT environment, the judicious use of these volumes of test facilities and resultant increased volumes of test results, through data fusion from various disparate programmes, may be supportive to the future pavement engineer in providing a continuous service regarding expected behaviour of pavements under increased load volumes.

2 4IR and the Environment 2.1

4th Industrial Revolution

The 4th Industrial Revolution (4IR) has been defined in many ways. Schwab (2016) introduced the term as a phase of industrial development that is characterized by a more ubiquitous and mobile Internet, smaller, cheaper and more powerful sensors, and artificial intelligence and machine learning, leading to a world with intertwined virtual and physical systems. In this 4IR age where issues around the datafication of life, data science, big data, transportation evolution through Mobility as a Service (MAAS), optimization of logistic and supply chains and automation of various aspects of life, including vehicles, becomes everyday features, it is important to re-evaluate the future of APT. Although much is being debated around changes in human transportation, the facts are that there is currently still growing demand for high quality road pavement structures, be it new construction or maintenance and rehabilitation of such structures. The near future will thus require a population of trained pavement engineers that understand road pavement materials, behaviour and all its related issues to ensure that the world economy can remain functional. Changes in freight movement, especially on local roads together with a continued growth of trade globalization, will impact on the volumes and loads on local roads. This may be either a positive (more, lighter vehicle) or negative (exponentially increased volumes) impact on local roads. While traditional APT approaches applied traffic load volumes to a section of the pavement in an accelerated methodology, current realities require a different approach where pavement engineers can still do focused research and investigation on specific individual issues, but also grasp the potential of the 4th IR methodologies and techniques to beneficiate existing data in new ways. In this future APT environment, aspects such as the ability of APT to still provide acceleration of traffic loading, as well as the judicious data fusion and analysis of data from various disparate sources to provide solutions to increasing need for the provision of economical and safe pavements, are addressed in this paper. 4IR technologies allow integration between the physical world and digital and intelligent engineering (Muhuri et al. 2019; World Economic Forum 2019). This fusion of reality and virtual environments span across disruptive transportation systems, and smart and intelligent infrastructure. These systems generate unimaginable quantities of data that are characterized by the 5 Vs - velocity, variety, veracity, value and volume (Núñez et al. 2014). Judicious use of such data should support the notion of future APT through intelligent design of mass data consumption models that links objectives and

Future APT – Thoughts on Future Evolution of APT

711

outcomes from distributed road evaluation sources into a new, more accelerated understanding of road pavement behaviour under growing traffic demand. 2.2

Environment

It is a well-debated fact that the climate is changing over time. Regardless of the reasons for such changes (which is outside of the scope of this paper), it is important to appreciate that any change in climate will have an effect on the behaviour of road pavement structures. Climate models indicate changes in all aspects that influence road pavement materials (Easterling 2002). As part of the discussion (and slightly besides the focus of this paper) it should be appreciated that due to the long lives of pavement infrastructure in general, and the relatively slow climate change ratio, appropriate and timeous evaluation of maintenance methodology on infrastructure that is potentially affected by climate should enable improved methodologies and resultant lower impacts of climate change on road infrastructure. In an APT environment, this is especially important, as the long-term effects of climate is typically not directly incorporated into a typical APT test in terms of seasonal changes. The standard method is to conduct several tests under different climate conditions (hot/cold/wet/dry/aged, etc.) (Steyn and Denneman 2008; Steyn and Du Plessis 2008) and incorporate such outcomes in the pavement analysis and modelling. With global climate change expectations, the effects of various climate change models on the behaviour of local road pavement structures needs to be incorporated into the synthesis of expected performance of the road pavement structure. A clear understanding of such expected local climate changes is thus required to ensure an applicable interpretation of the APT data.

3 Accelerated Pavement Testing (APT) 3.1

Philosophy and Principles

APT is defined as the controlled application of a wheel loading, at or above the appropriate legal load limit, to a pavement system to determine pavement response in a compressed time period. It is important to revisit the objectives of APT from time to time to re-establish the reasons and expected outcomes of APT. Overall, these objectives may be summarised as developing an improved understanding of pavement behaviour and performance, with the ultimate goal of providing more economic and sustainable land transportation options to society. 3.2

Traditional APT

Traditional APT approaches applied traffic load volumes to a section of the pavement in a much accelerated methodology, due to the combination of continuous load application of often increased load levels, and the relatively low traffic volumes on most roads. Currently, many roads carry volumes in excess of the abilities of APT devices to keep up with normal load applications.

712

W. J. vdM. Steyn

Traditional APT programmes focused on the identification of a problem, planning of an experiment to obtain data to solve the problem, execution of the APT (and related laboratory testing) programme, and analysis of these data with incorporation of related own research outcomes and published research findings. This process relies on the ability of a researcher to access all the available data and findings related to the research question, and synthesis of an appropriate response to the research question from the available data. If the research question is novel and unique, the related published research may be limited to the extent that a research group may have access to, and have the time and resources to study and incorporate such information into their research. However, with growing numbers of APT programmes, the volume of research findings have grown to the extent that access to and detailed analysis and synthesis of such information may become an effort too large for the typical APT programme to handle. This may lead to exclusion of data that may have supported the research, merely due to the huge volumes of available data. Table 1 provides an indication of the volume of published APT research data, summarising the results of a Google Scholar search for some typical APT related terms. It should be clear that it requires a superhuman effort to sieve through this amount of publications to ensure that appropriate knowledge is gained. Obviously, more specific search terms will provide more detailed results. However, the sheer volume of published information (articles, reports and datasets) in the APT field remains large. With this as background, it is opportune to discuss some options for future APT. Table 1. Result of Google Scholar search for typical APT results Year of publication Search terms APT Rut APT Fatigue 2019 1 060 3 300 2010 to 2018 7 390 22 200 2000 to 2009 4 380 16 900 All records 27 100 125 000

4 Future APT In a future APT environment, aspects such as the ability of APT to provide information of the response of a road pavement structure to tyre loads under different environmental conditions will remain important. However, with new techniques in data mining, data availability from often disparate sources and the need for judicious data fusion, analysis and synthesis, much more should be possible in terms of developing road pavement structure behaviour and performance models. In this section, some of these approaches and novel techniques are briefly discussed with an indication related to possible APT applications.

Future APT – Thoughts on Future Evolution of APT

4.1

713

Connection Between APT, LTPP Programmes and PMS Data

Cooperation between different APT programmes have been discussed often, and examples of such cooperation exist (e.g. MMLS User group (Steyn and Hugo 2016), HVS International Alliance (Harvey et al. 2008), NCAT/MnDOT partnership (Powell and Worell 2016), etc.). Most of the cooperation in these groups focuses on joint planning of APT tests in selected cases, and access to data. The different formats of raw data complicate this process. Incorporation of recent data analysis and mining techniques (Sects. 4.3 and 4.4) may allow deeper analysis of such databases for advanced synthesis of findings from previous results. The use of LTPP data to calibrate APT-derived models have been discussed and done over the years (Jones et al. 2004; Prozzi and Guo 2007; Steyn et al. 2012). However, limited results of the application of these techniques have been forthcoming. This may again be related to the required effort to merge disparate datasets from LTPP and APT sources into a combined database that can be interrogated. Again, the use of recent data analysis and mining techniques (Sects. 4.3 and 4.4) may allow a more stream-lined and effective approach in this regard. Incorporation of Pavement Management System (PMS) databases into this analysis process would add another real-life dataset for improved model development. Such an approach where disparate databases can be combined may bring into play LTPP and PMS data from the various countries – especially when these data are combined with appropriate APT results from these environments. This may lead to a global virtual LTPP/APT/PMS network to drive innovations in the understanding of road pavement structure behaviour. 4.2

Incorporation of Sensor Network Data

Recent advances in sensor and data acquisition technologies has led to the proliferation of sensors that can be incorporated into and next to road pavement structures for collection of environmental, visual and structural response data (Alavi et al. 2015; Inyim et al. 2016). This, combined with advances in the field of Civiltronics (Steyn and Broekman 2020) provide pavement engineers with a host of opportunities to gather road pavement structure and related data on a continuous basis. Further, advances in vehicle tracking technologies provides the opportunity to gather data on VehiclePavement Interaction (VPI) also on a continuous basis (Wessels and Steyn 2018). When combining these data sources in a judicious way, changes in pavement conditions, response and effects can be combined from a multitude of sources to improve the understanding of road pavement behaviour. This type of data analysis can be viewed as a subsection of APT as it provides data at the speed at which vehicles uses the actual road, which can be used together with LTPP data but at much reduced costs. Obviously, data validation should form a major part of this type of analysis as much less control exists regarding the actual applied loads and speeds of the tyre loads onto the pavement. However, one of the benefits of huge volumes of data is that appropriate statistical treatment of such data can assist in this validation process, providing not only mean data responses as well as outliers that may in some cases be the conditions where most of the change in behaviour may occur.

714

4.3

W. J. vdM. Steyn

Data Mining

Data mining consists of an array of techniques with the objective of extracting usable data from a set of raw data through a process of data pattern analysis. It may incorporate processes for automatic pattern predictions (based on trend and behaviour analysis) and clustering (based on finding and visually documented groups of facts not previously known) (Gopalakrishnan 2018). Through this process attempts are made to move from large data sets (Big Data) to mass knowledge, and ultimately focused implementation of new knowledge. Data mining may also require data fusion, incorporating integration of multiple data sources to produce consistent, accurate, and useful information – such as collections of APT, LTPP, pavement sensor and VPI data. These techniques provides for a new realm of potential solutions to standard pavement engineering problems (rutting and fatigue of typical road pavement structures as a perennial example). The need for pavement engineers to check outcomes for applicability and appropriateness in terms of fundamental pavement engineering is nonnegotiable. 4.4

Text Data Mining

Not all information on road pavement experiments are available and published in terms of data sets. The majority of APT outcomes are reported in the form of papers and reports with redacted versions of the trends and behaviour observed by the researchers. Text data mining defines the process of deriving high-quality information from natural text, identifying patterns and relationships within large amounts of text. The techniques has been used previously for analysing trends in conference proceedings (Boyer et al. 2017) and can in a similar way be used to identify trends and relationships from a collection of reports and publications related to APT. With reference to the volume of outputs related to APT (Table 1) this can assist the APT engineer in identifying novel trends from a range of publications that can then be used for more detailed analyses. An example of such an output is shown in Fig. 2 where a Correspondence Analysis of the abstracts of the 5th international APT conference is shown, showing words with similar appearance patterns. 4.5

Skills Required

A discussion about future APT and the techniques and options available to the pavement engineer to conduct advanced, novel and innovative analyses of disparate APTbased road pavement structure data, cannot be concluded without a note on the skills required to conduct such novel analyses (Steyn 2019). There cannot be any argument against the requirement of a fundamental understanding of road pavement materials, mechanics and behaviour for an engineer to be able to conduct APT analysis. However, the techniques that are described in this paper also requires advanced data analysis skills, that may not be part of the average APT engineer’s background or skills set. In this regard the need to incorporate such skills into the general training of APT engineers is obvious, but also the requirement to engage professionals in the fields of

Future APT – Thoughts on Future Evolution of APT

715

Fig. 2. Text analytics conducted on 2016 APT conference abstracts – correspondence analysis

data science, mining and related disciplines to assist with the advanced analysis of disparate APT and related data. A balanced approach is required where the two skills sets complement each other in identifying innovative trends and patterns from the available data without disregarding fundamental pavement engineering concepts.

5 Conclusions Growth in population, traffic volumes and road transportation needs leads to an increased need for provision, maintenance and rehabilitation of road pavement structures. Also, an improved understanding of road pavement structure and VPI data can provide better and more cost effective construction, maintenance and rehabilitation options. In this regard, future APT looks at a combination of traditional APT and laboratory tests, combined with the judicious use of data mining and text analytics techniques to allow improved analyses of available APT data. This should support the APT objective of developing an improved understanding of pavement behaviour and performance, with the ultimate goal of providing more economic and sustainable land transportation options to society.

References Alavi, A.H., Hasni, H., Lajnef, N., Chatti, K.: Continuous health monitoring of pavement systems using smart sensing technology. Constr. Build. Mater. 114, 719–736 (2015)

716

W. J. vdM. Steyn

Boyer, R.C., Scherer, W.T., Smith, M.C.: Trends over two decades of transportation research: a machine learning approach. Transp. Res. Rec. J. Transp. Res. Board 2614(2017), 1–9 (2017). https://doi.org/10.3141/2614-01 Easterling, D.R.: Observed climate change and transportation. In: The Potential Impacts of Climate Change on Transportation, Federal Research Partnership Workshop, 1–2 October 2002. U.S. Department of Transportation Center for Climate Change and Environmental Forecasting, U.S. Environmental Protection Agency, The U.S. Global Change Research Program of the U.S. Climate Change Science Program, U.S. Department of Energy (2002) Gopalakrishnan, K.: Deep learning in data-driven pavement image analysis and automated distress detection: a review. Data 3, 28 (2018). https://doi.org/10.3390/data3030028 Harvey, J.T., Sadzik, E., Coetzee, N.F., Mahoney, J.P.: Developing international collaborative efforts in APT: the HVSIA experience. In: 3rd International Conference on Accelerated Pavement Testing, Madrid, Spain, 1–3 October 2008 (2008). Spanish ISBN NIPO 163-08062-2 Hugo, F., Epps Martin, A.: Synthesis of Highway Practice 325: Significant Findings from FullScale Accelerated Pavement Testing. Transportation Research Board, National Research Council, Washington, D.C. (2004) Inyim, P., Pereyraa, J., Bienvenub, M., Mostafavia, A.: Environmental assessment of pavement infrastructure: a systematic review. J. Environ. Manag. 176, 128–138 (2016) Jones, D., Paige-Green, P., Sadzik, E.: The development of a protocol for the establishment and operation of LTPP sections in conjunction with APT sections. In: 2004 2nd International Conference on Accelerated Pavement Testing, Minneapolis, Minnesota, USA (2004) Metcalf, J.B.: Synthesis of highway practice 235: application of full-scale accelerated pavement testing. Transportation Research Board, National Research Council, Washington, D.C., 117 p. (1996) Muhuri, P.K., Shukla, A.K., Abraham, A.: Industry 4.0: a bibliometric analysis and detailed overview. Eng. Appl. Artif. Intell. 78, 218–235 (2019). https://doi.org/10.1016/j.engappai. 2018.11.007 Núñez, A., Hendriks, J., Li, Z., De Schutter, B., Dollevoet, R.: Facilitating maintenance decisions on the Dutch railways using big data: the ABA case study. In: IEEE International Conference on Big Data, Washington, DC, USA, pp. 48–53 (2014). https://doi.org/10.1109/bigdata.2014. 7004431 Powell, R.B., Worell, B.: Research partnership between the National Center for Asphalt Technology (NCAT) and the Minnesota Department of Transportation MnROAD facility for a nationwide pavement performance experiment. In: Aguiar-Moya, J., Vargas-Nordcbeck, A., Leiva-Villacorta, F., Loría-Salazar, L. (eds.) The Roles of Accelerated Pavement Testing in Pavement Sustainability. Springer, Cham (2016) Prozzi, J.A., Guo, R.: Reliability-based approach for using LTPP and APT test results for estimating fatigue performance. In: TRB 86th Annual Meeting Compendium of Papers, Washington, D.C. (2007) Schwab, K.: The Fourth Industrial Revolution: What it Means, How to Respond. World Economic Forum, Davos (2016) Steyn, W.J.vdM., Denneman, E.: Simulation of temperature conditions on APT of HMA mixes. In: 3rd International Conference on Accelerated Pavement Testing, Madrid, Spain, 1–3 October 2008 (2008). Spanish ISBN NIPO 163-08-062-2 Steyn, W.J.vdM., Du Plessis, J.L.: Moisture related test protocols for HVS testing. In: 3rd International Conference on Accelerated Pavement Testing, Madrid, Spain, 1–3 October 2008 (2008). Spanish ISBN NIPO 163-08-062-2

Future APT – Thoughts on Future Evolution of APT

717

Steyn, W.J.vdM., Anochie-Boateng, J.A.B., Fisher, C., Jones, D., Truter, L.: Calibration of fullscale accelerated pavement testing data using long-term pavement performance data. In: 4th International Conference on Accelerated Pavement Testing, Davis, California, USA (2012) Steyn, W.J.vdM.: Full-scale accelerated pavement testing, 2000 to 2011. NCHRP Synthesis 433. National Cooperative Highway Research Program (NCHRP), Transportation Research Board (TRB), National Research Council, Washington D.C. (2012) Steyn, W.J.vdM., Hugo, F.: Perspectives on trends in international APT research. In: 5th International Conference on Accelerated Pavement Testing. San Jose, Costa Rica, 19–21 September 2016 (2016) Steyn, W.J.vdM.: Training and educating pavement engineer 4.0. In: Proceedings of 12th Conference on Asphalt Pavements for Southern Africa, Sun City, South Africa, 13–16 October 2019 (2019) Steyn, W.J.vdM., Broekman, A.: Civiltronics: fusing the civil and electronics engineering in the 4IR era. SAICE Civ. Eng. 28(1) (2020). ISSN 1021-2000 Wessels, I., Steyn, W.J.vdM.: Continuous, response-based road roughness measurements utilising data harvested from telematics device sensors. Int. J. Pavement Eng. (2018). https:// doi.org/10.1080/10298436.2018.1483505 World Economic Forum (2019). https://www.weforum.org/focus/fourth-industrial-revolution. Accessed Nov 2019

Author Index

A Aguiar-Moya, Jose, 428, 640 Aguiar-Moya, José P., 369 Ahmed, Abubeker, 280 Airey, Gordon, 545 Akhalwaya, I., 251 Alabaster, David, 479 Ali, Ayman, 339, 409 Al-Qadi, Imad L., 418 Arraigada, Martin, 291 Ávila-Esquivel, Tania, 369, 428, 640 Ayalew, Robel, 178 B Bahrani, Natasha, 526, 592 Baltazart, V., 615, 632 Barriera, Maria, 488 Baudru, Yvan, 649 Bell, David, 228 Bell, Haley P., 209 Berrada, Kamal, 699 Berraha, Youness, 51 Bilodeau, Jean-Pascal, 51, 127, 497 Blaineau, Fabrice, 649 Blanc, Juliette, 100, 261, 488, 526, 545, 592, 649 Bodin, Didier, 118, 300 Brissaud, Laurent, 329 Brito, Lélio Antônio, 319 Broutin, Michaël, 516, 658 Büchler, Stephan, 357

Bueno, M., 291 Buzz Powell, R., 3, 40 C Camacho-Garita, Edgar, 369, 428, 640 Cao, Rongji, 61 Carr, Harold T., 507 Castro, Federico, 169 Ceratti, Jorge Augusto Pereira, 319 Chabot, Armelle, 329, 389, 516 Chailleux, Emmanuel, 261, 488 Chatti, Karim, 545 Chazallon, Cyrille, 329 Cheng, Huailei, 379 Chirva, Dmitry, 623 Chupin, Olivier, 100, 389 Coca, Ana-Maria, 12 Coirier, Gilles, 649 Cox, Ben C., 339 D da Silva Filho, Luiz Carlos Pinto, 319 Dao, Duc Tung, 186 de Almeida, Adosindro Joaquim, 555 de Melo Gevaerd, Bruno, 555 de Oliveira, Jhenyffer Matias, 157 de Souza, Rafael Aleixo, 555 del Barco, Ana Jimenez, 261 Dérobert, X., 615, 632 Di Benedetto, Hervé, 658 Diefenderfer, Brian, 270

© Springer Nature Switzerland AG 2020 A. Chabot et al. (Eds.): Accelerated Pavement Testing to Transport Infrastructure Innovation, LNCE 96, pp. 719–722, 2020. https://doi.org/10.1007/978-3-030-55236-7

720 Diefenderfer, Brian K., 108 Dinegdae, Yared, 399 Do, Minh Tan, 186 Doligez, Daniel, 329 Dong, Zejiao, 536 Dopeux, Jérome, 467 Doré, Guy, 51, 127 Doué, S., 632 Du Plessis, L., 251 Duprat, Patrick, 689 Durand, O., 632 E Edwards, Lulu, 209 Erlingsson, Sigurdur, 70, 280, 399 F Fladvad, Marit, 70 Flintsch, Gerardo, 270 Flintsch, Gerardo W., 108 Franco, Miguel Ángel, 169 Francois, Andrae, 409 G Gabet, Thomas, 100, 689 Gao, Ying, 61 Garg, Navneet, 100, 218, 418 Garnica, Paul, 169 Gerbaud, Charlotte, 467 Gharbi, Maissa, 516 Gkyrtis, Konstantinos, 438 Godard, Eric, 329 Gonzalez, Carlos R., 458 Gonzalez, Erdrick Perez, 497 Gouy, Thierry, 649, 699 Graeff, Ângela Gaio, 319 Greenslade, Frank, 479 Grenfell, James, 118, 300 Guan, Wei, 21, 602 Gungor, Osman Erman, 418 H Hammoum, Ferhat, 100 Harvey, John, 178 Hellman, Fredrik, 280 Hornych, Pierre, 186, 261, 329, 526, 592, 649, 689, 699 Hossain, Mustaque, 147 Hu, Yue, 379 Huang, Yi, 61

Author Index I Isaev, Eugeny, 623 J Jameson, Geoff, 300 Jansen, Dirk, 32, 676 Jeremy Robinson, W., 507, 564 Ji, Richard, 666 Jones, David, 309 K Kazmee, Hasan, 218 Kluttz, Robert, 228 Kodikara, Jayantha, 118 Koval, Georg, 329 Kwon, Jayhyun, 564 L Lajnef, Nizar, 545 Laurent-Matamoros, Piero, 428, 640 Le Gal, Yves, 329 Le, Xuan Quy, 649 Lebental, Bérengère, 488 Lédée, Fabienne Anfosso, 689 Lefeuvre, Yann, 389 Lein, Wade, 339 Leischner, Sabine, 479 Levenberg, Eyal, 592 Li, Jiahao, 90 Li, Yi, 90, 379 Liu, Liping, 90, 379 Liu, Yilong, 196 Lo Presti, Davide, 261, 545 Loizos, Andreas, 438 Loria-Salazar, Luis Guillermo, 369, 428, 640 Louw, Stefan, 309 Lv, Zhenglong, 61 M Ma, Xianyong, 536 Mahdi, Moinul, 196 Manosalvas-Paredes, Mario, 545 Marasteanu, Mihai, 157 Marsac, Paul, 100 Mateos, Angel, 178 Mehta, Yusuf, 339, 409 Menant, Fabien, 526 Meroni, Fabrizio, 108, 270 Metrope, Mickael, 467 Millien, Anne, 467

Author Index Mironchuk, Sergey, 623 Mollenhauer, Konrad, 357 N Nasser, Hanan, 389 Nelson, Jason, 348 Nguyen, Mai Lan, 100, 186, 329, 389, 649, 689, 699 Nguyen, Minh Duc, 649 Norgeot, C., 632 Norwood, Gregory, 564 Núñez, Washington Peres, 319 O Olard, François, 261 Otto, Gustavo Garcia, 555 Oubahdou, Yamina, 467 P Palek, Leonard, 583 Paniagua, Fabian, 178 Paniagua, Julio, 178 Partl, M. N., 291 Pérez, Alfonso, 169 Pérez-González, Erdrick, 127 Perraton, Daniel, 51 Petit, Christophe, 467 Piau, Jean-Michel, 389 Picoux, Benoît, 467 Pitawala, Sameera, 118 Planche, Jean-Pascal, 261 Plati, Christina, 438 Popescu, Constantin, 12 Porot, Laurent, 228, 261 Pouget, Simon, 261, 488 Pouteau, Bertrand, 699 Powell, Buzz, 348 R Rahman, Shafiqur, 280 Remek, Lubos, 448 Reynaud, Philippe, 467 Ricalde, Lia, 218 Rodriguez, Camilo Andrés Múñoz, 319 Romanoschi, Stefan A., 12 Roussel, Jean-Marie, 658 Rupnow, Tyson, 196 Rushing, Timothy W., 238 Rust, F. C., 251 S Sahli, Mehdi, 329 Saidi, Ahmed, 339 Sauzéat, Cédric, 658

721 Schieber, Greg, 147 Schmalz, Michael, 357 Schneider, Thomas, 516 Scholten, Erik, 228 Sedran, Thierry, 186 Shan, Ling Yan, 21 Shinohara, Keyla Junko Chaves, 555 Simonin, J.-M., 615, 632 Simonsen, Erik, 280 Siroma, Rodrigo Shigueiro, 649 Smit, M. A., 251 Solodov, Vitaly, 623 Sotoodeh-Nia, Zahra, 261 Sounthararajah, Arooran, 118 Stache, Jeremiah M., 458 Stephane, Maindroult, 516 Steyn, Wynand J. vdM., 708 Sun, Lijun, 90, 379 T Talebsafa, Mohsen, 12 Tautou, Rémi, 467 Thom, Nick, 545 Timm, David, 574 Timm, David H., 137 Tingle, Jeb S., 209, 564 Todkar, S., 615 Todkar, S. S., 632 Trichet, Stéphane, 649, 699 Turos, Mugurel, 157 V Vaillancourt, Michel, 51 Valaskova, Veronika, 448 Valente, Amir Mattar, 555 van Deusen, Dave, 157 Van Rompu, Julien, 488 Vargas-Nordcbeck, Adriana, 348 Vergne, Sandrine, 699 Vrtis, Michael, 40, 137, 157, 583 W Wacker, Bastian, 32, 676 Wang, Hao, 666 Wang, Tongxu, 536 Wang, Xu Dong, 21, 602 Wang, Yuhong, 379 Wayne, Mark H., 564 Wetekam, Jens, 357 Williams, Chris, 261 Wistuba, Michael, 357 Worel, Benjamin, 40, 137, 583 Wu, Chunying, 61 Wu, J. T., 80

722 Wu, Wu, Wu, Wu,

Author Index Rongzong, 178, 309 Xingdong, 147 Y. T., 80 Zhong, 196

X Xiao, Qian, 21, 602 Xie, Pengyu, 666 Xue, Wenjing, 108, 270

Y Yan, Tianhao, 157

Z Zhang, Lei, 21, 602 Zheng, Bingfeng, 61 Zhou, Xing Ye, 21, 602 Ziegler, Thomas, 357