Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications [1 ed.] 9781617615429, 9781617613234

Citation preview

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest Ebook

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

PHYSICS RESEARCH AND TECHNOLOGY

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

CRYOGENICS: THEORY, PROCESSES AND APPLICATIONS

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

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

PHYSICS RESEARCH AND TECHNOLOGY Additional books in this series can be found on Nova‘s website under the Series tab.

Additional E-books in this series can be found on Nova‘s website under the E-book tab.

GENETICS – RESEARCH AND ISSUES Additional books in this series can be found on Nova‘s website under the Series tab.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Additional E-books in this series can be found on Nova‘s website under the E-book tab.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

PHYSICS RESEARCH AND TECHNOLOGY

CRYOGENICS: THEORY, PROCESSES AND APPLICATIONS

ALLYSON E. HAYES Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

EDITOR

Nova Science Publishers, Inc. New York Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2011 by Nova Science Publishers, Inc. All rights reserved. No part of this book may be reproduced, stored in a retrieval system or transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical photocopying, recording or otherwise without the written permission of the Publisher. For permission to use material from this book please contact us: Telephone 631-231-7269; Fax 631-231-8175 Web Site: http://www.novapublishers.com NOTICE TO THE READER The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No liability is assumed for incidental or consequential damages in connection with or arising out of information contained in this book. The Publisher shall not be liable for any special, consequential, or exemplary damages resulting, in whole or in part, from the readers‘ use of, or reliance upon, this material. Any parts of this book based on government reports are so indicated and copyright is claimed for those parts to the extent applicable to compilations of such works.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Independent verification should be sought for any data, advice or recommendations contained in this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage to persons or property arising from any methods, products, instructions, ideas or otherwise contained in this publication. This publication is designed to provide accurate and authoritative information with regard to the subject matter covered herein. It is sold with the clear understanding that the Publisher is not engaged in rendering legal or any other professional services. If legal or any other expert assistance is required, the services of a competent person should be sought. FROM A DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS. Additional color graphics may be available in the e-book version of this book.

LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA Cryogenics : theory, processes and applications / editor, Allyson E. Hayes. p. cm. Includes index. ISBN:  (eBook) 1. Low temperatures. 2. Low temperature engineering. I. Hayes, Allyson E. QC278.4.C79 2010 621.5'9--dc22 2010047661

Published by Nova Science Publishers, Inc. † New York

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

CONTENTS Preface

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Chapter 1

vii Control of the Sonic Boom Generated by a Flying Vehicle by Means of a Cryogenic Impact on the Flow Process V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov and V.F. Volkov

Chapter 2

Liquid Oxygen Magnetohydrodynamics J. C. Boulware, H. Ban, S. Jensen and S. Wassom

Chapter 3

Effect of Cryogenic Treatment on Microstructure and Mechanical Properties of Light Weight Alloys Kaveh Meshinchi Asl and Mehdi Koneshloo

1 39

69

Chapter 4

Cryogenic Treatment and Fatigue Resistance Paolo Baldissera and Cristiana Delprete

Chapter 5

Application of Fiber Bragg Grating Sensors at Cryogenic Temperatures Ines Latka, Tobias Habisreuther and Wolfgang Ecke

105

Cryogenic Grinding: Application for Structural Modification and Formulation Development of Drug Molecules Kunikazu Moribe, Kenjirou Higashi and Keiji Yamamoto

123

Chapter 6

Chapter 7

Chapter 8

Improvements in Tolerance to Cryopreservation Using Shoot-Tips of Chrysanthemum (Dendranthema grandiflorum Kitam.) from Genetically Modified Plants that Accumulate Trehalose María Teresa González-Arnao, José O. Mascorro-Gallardo, Antelmo Osorio Saenz, María del Rocío Valle-Sandoval and Florent Engelmann Cryogenics Vessels Thermal Profiling Using the Boubaker Polynomials Expansion Scheme BPES Investigation Da Hong Zhang

Index

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

93

137

149 165

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

PREFACE Cryogenics is the study of the production of very low temperature (below −150 °C, −238 °F or 123 K) and the behavior of materials at those temperatures. This book presents current research from across the globe in the study of cryogenics, including the effect of cryogenic treatment on microstructure and mechanical properties of light weight alloys; the application of Fiber Bragg grating sensors at cryogenic temperatures; cryogenic grinding; liquid oxygen magnetohydrodynamics; and genetic engineering techniques used to improve tolerance to cryopreservation. Chapter 1 - Cooling of surfaces of various objects is widely used in science and engineering. This process assists in formation of better characteristics of various devices, tools, and instruments by increasing the quality of their operation and reliability. On the other hand, surface cooling and supercooling can generate severe problems in operation of vehicles, sometimes leading to catastrophic situations. Therefore, despite the long story of studying the cooling processes, it is still important because of numerous applications. Chapter 2 - In the cryogenic realm, liquid oxygen (LOX) possesses a natural paramagnetic susceptibility and does not require a colloidal suspension of particles for practical application as a magnetic working fluid. Commercial ferrofluids have performed well in industrial applications, but expanding their workable range to low temperatures requires a suitable selection of the carrier fluid, such as LOX. In this chapter, the equation of motion for the pure fluid is derived and applied to a slug of LOX being displaced by a pulsed magnetic field. Its theoretical performance is compared to actual experimental data with discussion on empirical parameters, sensitivity to measurement uncertainty, and geometric similarity. The 1.1 T pulse of magnetic flux density produced oscillations in the slug of 6-8 Hz, generating up to 1.4 kPa of pressure change in a closed section when the slug acted like a liquid piston. The experiments and theoretical model demonstrate that LOX could be used as a magnetic working fluid in certain applications. Chapter 3 - This chapter mainly focuses on the effects of low temperature (subzero) treatments on microstructure and mechanical properties of aluminum and magnesium alloys. Deep cryogenic treatment on A319 aluminum alloy showed that the abrasion resistance of the alloy was improved after the treatment. This improvement was attributed to the strengthening of the α-aluminum matrix which slows down the propagation of the existing defects. The execution of deep cryogenic treatment on AZ91 magnesium alloy changed the distribution of β precipitates. The tiny laminar β particles almost dissolved in the microstructure and the coarse divorced eutectic β phase penetrated into the matrix. This microstructural modification

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

viii

Allyson E. Hayes

resulted in a significant improvement on mechanical properties of the alloy. The steady state creep rates were measured and it was found that the creep behavior of the alloy, which is dependent on the stability of the near grain boundary microstructure, was improved by the deep cryogenic treatment. After the deep cryogenic treatment, the sliding of grain boundaries was greatly suppressed due to morphological changes. As a result, the grain boundaries are less susceptible for grain boundary sliding at high temperatures. After dry sliding wear tests were performed, the wear resistance of the alloy improved remarkably after deep cryogenic treatment. Furthermore due to interest in the subzero treatments of steels in the past few decades, AISI H13 tool steel was chosen and cryogenic treatment at -72ºC and deep cryogenic treatment at -196ºC were applied and it was found that the execution of low temperature treatments on samples affected the microstructure of the H13 tool alloy to a great extent. By applying the subzero treatments, the retained austenite was transformed to martensite due to the completion of martensite transformation. The cryogenic treatment at a very low temperature and holding the samples for a long time, also lead to precipitation of more uniform and very fine carbide particles. This microstructural modification resulted in a significant improvement on mechanical properties and wear resistance of the alloy. Chapter 4 - Among the various applications of cryogenics, the implementation of cold thermal processes with the aim of enhancing mechanical properties of materials is the most attractive from the perspective of structural component engineering design. In particular, considering the key-role played by the fatigue behavior of materials in this discipline, the development of methodologies that allow to achieve longer service life is an evergreen topic, which concerns many fields of application such as energy production and transportation. Starting from the scientific literature of the last 30 years concerning shallow and deep cryogenic treatment (SCT and DCT) and including the most recent experimental results achieved at the Politecnico di Torino, both direct and indirect evidences of an effective influence on the fatigue behavior of steels are pointed-out. Experimental methodologies and data analysis approaches are detailed with a special focus on the estimation of optimal treatment parameters. In particular, two classes of steels that show a good liability for such processes are discussed in depth: austenitic stainless steels and carburized ones. In both cases, the potential consequence in terms of reliability and service life of structural components is noticeable and can be highlighted through practical design examples (stainless steel springs and carburized gears are discussed in details). In the final part of the chapter, an overall picture of the most promising future applications is given with a particular focus on materials beyond steels such as different alloys (i.e. aluminum, magnesium, titanium), polymers and composites and in consideration of their specific fatigue mechanisms. Chapter 5 - Fiber Bragg grating (FBG) sensors are well known means for the measurement of strain and temperature in broad temperature and strain ranges. The most important features of this sensor type are its small size, light weight, full electrical insulation, negligible interaction with electric and magnetic fields, and the flexible fiber leads of good thermal insulation between sensor and its interrogation unit. In particular, the low thermal conductivity of the optical fiber is an advantage when working with low temperatures. Another important feature is the possibility to include several grating sensors in the same fiber, which can later be interrogated simultaneously via wavelength multiplexing. An FBG can be defined as a periodic modulation of the refractive index along a section of the fiber core. Such gratings can be produced by the irradiation of photosensitive silica

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Preface

ix

fibers with an UV laser and an interferometer setup. The FBG reflects a wavelength, which is dependent on the period of the structure and on the effective refractive index of the fiber. Exposure to temperature or strain changes will affect the period of the structure or/and the refractive index, thereby changing the reflected wavelength. A polychromator based measurement guarantees simultaneous measurements for all sensors in the wavelengthmultiplexed assembly, with equal measurement duration of typically 100 µs. Thus, it is possible to reconstruct exact strain modes or temperature distributions from the actual multipoint results. The wavelength changes can be monitored with a 1ζ repeatability of about 0.1 pm. While conventional electrical resistance strain gages show increasing cross-sensitivities to temperature and magnetic field with decreasing temperature down to liquid helium, it has been found that fiber optic Bragg grating strain sensors show negligible thermo-optic and magneto-optic effects in the cryogenic environment and they allow, therefore, reliable strain measurements. These specific application advantages of optical fiber Bragg grating sensors at low temperatures make them attractive for structural health monitoring of cryogenic devices such as superconductive magnets. Other applications are material characterization, e.g. of superconducting materials or of structural elements, respectively. Flux pinning is one of the most fundamental and interesting properties of type II superconductors. In materials with the strongest pinning, the pinning-induced strain can be so large that it can lead to cracks in the material. With the application of FBG sensors, spatially resolved measurements of magnetostrictive effects in superconducting samples become possible. FBGs can also be used for the measurement of thermal expansion coefficients down to 4 K, or for the change of Young‘s modulus with respect to temperature. Chapter 6 - Cryogenic grinding is used to modify the physicochemical properties of a solid drug. A low-crystalline or amorphous form is formed depending on grinding conditions such as temperature and grinding time. Methods used to prepare amorphous forms sometimes affect the stability and dissolution behavior of drugs, even if the powder X-ray diffraction measurement of each formulation shows a halo pattern. Grinding of a drug in the presence of excipients makes it possible to prepare drug formulations such as solid dispersions, complexes, and cocrystals. Grinding under cryogenic conditions is favorable when formulation components are not suitable for ambient temperature grinding. Grinding temperature affects the molecular states of the components of the system. The physicchemical characterization of samples prepared at different temperatures can be used to effectively determine the mechanism of complex or cocrystal formation. Drug nanoparticles with improved dispersibility and dissolution properties can be prepared by using suitable excipients. Thus, sample grinding under cryogenic conditions appears to be useful in the field of pharmaceutics. Chapter 7 - Intracellular accumulation of organic osmolytes such as sucrose and trehalose is highly correlated with tolerance to stress induced by dehydration and freezing in many plants and microorganisms. Based on the existence of such protective mechanisms, the authors generated genetically modified lines of chrysanthemum (Dendranthema grandiflorum Kitam.) var. Indianapolis with induced capacity to biosynthesize trehalose, and performed cryopreservation experiments with shoot-tips isolated from chrysanthemum in vitro plantlets of two transgenic lines with different endogenous accumulation of trehalose and with shoottips dissected from non transgenic in vitro plantlets. After dissection, apices were precultured

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

x

Allyson E. Hayes

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

on semi-solid MS medium supplemented with 0.3 M sucrose for 4 days, loaded in a 0.4 M sucrose + 2 M glycerol solution for 20–30 min and exposed to PVS2 or PVS3 vitrification solutions for 0, 20, 40 or 60 min at room temperature prior to rapid immersion in liquid nitrogen in cryovials with 1mL of the respective PVS. The highest shoot regeneration after cryopreservation was obtained following exposure to either PVS solution for 40 min. Genetically modified tissues displayed an improved tolerance to cryopreservation in comparison with non transgenic ones. Cryopreserved shoot tips from plants of transgenic lines produced 67% and 48% shoot regeneration after treatment with PVS2 and 54% and 52% with PVS3, while non-transgenic ones showed 33% shoot regeneration after treatment with PVS2 and 36% with PVS3. These results demonstrate that genetic engineering techniques can be a useful biotechnological tool to generate transgenic organisms showing improved tolerance to cryopreservation. Chapter 8 - This chapter investigates temperature dynamical profiling inside vacuuminsulated cryogenic vessels. The proposed temperature, velocity and acceleration profiles in the particular case of cylindrical geometry have been obtained through the application of the Boubaker Polynomials Expansion Scheme (BPES).

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

In: Cryogenics: Theory, Processes… Editor: Allyson E. Hayes, pp. 1-37

ISBN: 978-1-61761-323-4 © 2011 Nova Science Publishers, Inc.

Chapter 1

CONTROL OF THE SONIC BOOM GENERATED BY A FLYING VEHICLE BY MEANS OF A CRYOGENIC IMPACT ON THE FLOW PROCESS V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov and V.F. Volkov

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Khristianovich Institute of Theoretical and Applied Mechanics, Siberian Branch, Russian Academy of Sciences, Novosibirsk, Russia

INTRODUCTION Cooling of surfaces of various objects is widely used in science and engineering. This process assists in formation of better characteristics of various devices, tools, and instruments by increasing the quality of their operation and reliability. On the other hand, surface cooling and supercooling can generate severe problems in operation of vehicles, sometimes leading to catastrophic situations. Therefore, despite the long story of studying the cooling processes, it is still important because of numerous applications. The cooling processes are particularly versatile and important in aeronautics, because they affect flight safety. Aircraft icing caused by collisions of supercooled water drops with the windshield, resulting in rapid crystallization of these drops and formation of ice accretions of various shapes and sizes, is well known. This unsteady process arises when the aircraft enters clouds containing fine droplets of supercooled water in a metastable state at negative temperatures (down to – 40  -60°С). In most cases, aircraft icing occurs when it flies in the atmosphere containing supercooled water drops (i.e., water in the liquid phase at negative temperatures). During their collisions with the frontal surfaces of aircraft elements, the water drops become rapidly crystallized, forming ice accretions of various shapes and sizes. Under icing conditions, ice structures are formed on the frontal surfaces of the wings, rudder and

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

2

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

elevator, propellers, cockpit windows, navigation gauges, and blisters. Unfortunately, the mechanism of rapid crystallization of supercooled drops after their impact on the aircraft surface has not bee adequately studied yet. Despite the increasing effectiveness of various devices, icing is still a factor that exerts a substantial effect on flight safety and regularity. Statistical data on aircraft icing situations in various geographical regions of the Earth show that icing is most probable if the aircraft flies in the temperature range from 0 to –15°С. Some cases of icing at air temperatures of –50°С and lower were registered [1-5]. Inlet channels of aircraft engines can also be subjected to icing at positive (up to +10°С) temperatures. This is explained by condensation and freezing of air moving in inlet channels. Cases of icing of supersonic inlets are also known. At supersonic flight velocities, the temperature of inlet surfaces drastically increases and can exceed 2000°C at Mach numbers М∞= 5÷8. Therefore, it is necessary to cool the compression surfaces and, simultaneously study the influence of cooling on inlet performance. Hypersonic inlets are cooled to ensure their thermal protection and improve their aerodynamic characteristics. These problems were discussed in [6-10]. It was demonstrated that cooling makes the laminar boundary layer on the central body more stable to developed separation and prevents upstream transfer of counter pressure, which leads to reduction and elimination of separation ahead of the inlet and to significant improvement of the throttling characteristics of the inlet. Surface cooling at supersonic velocities exerts a considerable effect on boundary layer stability and its transition to the turbulent state. This research direction has been actively developed at the Khristianovich Institute of Theoretical and Applied Mechanics of the Siberian Branch of the Russian Academy of Sciences (ITAM SB RAS). Some results are published in [11-16]. It was demonstrated in computations and experiments that surface cooling is accompanied by stabilization of the first mode of disturbances, which favors reduction of the range of unstable frequencies, and the neutral stability curve is shifted toward higher Reynolds numbers. At the same time, the second mode of disturbances (high frequencies) is destabilized, i.e., the range of unstable frequencies is expanded, moving toward higher frequencies, whereas the amplification factor increases. Surface cooling does not affect the interaction of acoustic waves with the supersonic boundary layer. Experimental data obtained agree with predictions of the hydrodynamic stability theory. When the aircraft moves with a supersonic velocity, its surfaces are subjected to considerable aerodynamic heating. Therefore, one of the key problems in creating such flying vehicles is cooling of heat-intense elements of the airframe, engine, and onboard equipment, which cannot be solved at the expense of the cooling capacity of liquid fuels, such as kerosene. Therefore, the Tupolev Joint Stock Company developed an experimental aircraft TU-155, which was adapted to use not only liquid hydrogen, but also liquefied natural gas (LNG). It should be noted that liquid hydrogen possesses extremely useful properties: high heat of combustion, tremendous cooling capacity, and environmentally friendly nature, which make it possible to improve the flight performance and to create aircraft with flight velocities M > 6. Thus, the first aircraft in the world operating on cryogenic fuels was developed. Aerospace industry also involves the search for new material resources to create more environmentally friendly fuels that simultaneously ensure a high cooling capacity. Such fuels can be hydrocarbon gases (methane, propane, butane, etc.) obtained from natural and oil gases, and also hydrogen. These gases are rather different in terms of their physical properties, which can substantially affect the aircraft structure, propulsion, and exploitation.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Control of the Sonic Boom Generated by a Flying Vehicle…

3

The above-given examples give only limited illustrations of the spectrum of surface cooling applications and, simultaneously, stimulate the use of these processes in solving other problems. It is known, in particular, that sonic boom (SB) reduction by conventional methods based on searching for an optimal distribution of the volume and lift force along the aircraft has not yet seen much progress. The configuration of the Concorde supersonic passenger aircraft, which is close to optimal in terms of the SB signature, does not ensure an admissible SB level required by ICAO norms. As was shown in [17], the excess pressure level on the Earth‘s surface in the cruising flight of Concorde exceeds 100 Pa. Results of research aimed at the development of the supersonic passenger plane of the second generation [18-20, 30, 36] also show that it is impossible to reach an admissible SB level generated by a conventional configuration of the aircraft with a large takeoff weight without deteriorating the lift-to-drag ratio. These and other studies dealing with passive methods of SB reduction in flight of supersonic transport planes undoubtedly show that it are necessary to search for new methods of controlling the SB parameters. The first investigations of active control of SB parameters by means of mass addition (in the form of air jets oriented with respect to the model in a certain manner) and energy addition (ensured by combustion of a hydrogen-air mixture) near the body [20, 21] show the prospects of this approach. Recent research in the last decade includes various methods of energy supply to a supersonic flow with the use of laser and microwave radiation, electron guns, and spark discharges, aimed at controlling the flow parameters, including the formation of a disturbed flow near the flying vehicle [23-27]. At the same time, the process of energy removal with the help of cryogenic technologies offers many new possibilities of active control of the flow around the aircraft and the level of the sonic boom generated by this aircraft. The present chapter describes the results of numerical and experimental investigations performed at ITAM SB RAS [28–33]. The parameters of the SB generated by power-law bodies of revolutions were systematically studied in experiments and calculations. The result was determining a class of modified (with the help of spherical bluntness of the nose part) bodies that ensure reduction of the bow shock wave intensity, as compared with a nonblunted body with the same aspect ratio in the middle zone of the sonic boom. Cryogenic technologies of controlling the flow around the aircraft were developed to extend the length of the SB minimization region (middle zone) to distances corresponding to the cruising flight altitude. A detailed description of the experimental techniques and facilities is given. A series of experimental and numerical studies was performed, and the possibility of active controlling of SB parameters by means of surface cooling and organization of distributed injection of the coolant into the flow was demonstrated for the first time.

1. PROBLEM OF CREATING A SUPERSONIC PASSENGER PLANE The increasing rate of modern life and business globalization, which require highvelocity transportation, stimulate the research aimed at creating supersonic and hypersonic passenger planes. This is particularly important for intercontinental flights and for internal flights in countries with large territories, for instance, Russia. The time of a non-stop flight of a subsonic aircraft (A-310, TU-204) from Moscow to Vladivostok at a distance of

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

4

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

approximately 6500 km is nine tiresome hours, and the flight to San Francisco takes almost twice as long (16 hours). The flight with a supersonic velocity equal to the doubled velocity of sound will reduce this time by more than a factor of 2. The problem of supersonic aircraft development involves ecological and economical restrictions, which determine the grounds for aircraft exploitation and its competitiveness on the market. Ecological restrictions include the provision of an acceptable noise level, reduction of hazardous wastes of fuel combustion products into the atmosphere, and provision of an acceptable level of the sonic boom generated by the aircraft flying at supersonic velocities. The cost efficiency of the aircraft (the cost of transportation of one passenger) depends substantially on the lift-to-drag coefficient of the aircraft. Specialists unanimously agree than the basic obstacle in creating a supersonic passenger plane is the requirements to the SB signature. Thus, what is the sonic boom? In atmospheric flight with a velocity greater than the velocity of sound (Figure 1), the disturbed flow region is bounded with the bow shock wave (SW) emanating from the nose part of the aircraft and the rear SW formed on the tail part of the aircraft. Near the aircraft (the so-called near zone), there are intermediate SWs, as well as expansion and compression waves generated by individual elements of the aircraft structure. As the disturbances generated by each point of the aircraft surface propagate with a velocity close to the velocity of sound (which is smaller than the aircraft velocity), the SW shape is close to conical. The pressure, temperature, and density of air increase in a jump like manner behind the SW owing to accumulation of disturbances. Owing to nonlinear effects (dependences of the velocity of propagation of disturbances on their amplitude), the flow with distance from the aircraft (far zone) is formed so that the distribution of the excess pressure (with respect to the atmospheric value) acquires an N-shape. The observer on the ground level perceives this Nwave as one or two (depending on the aircraft size and flight altitude) distant explosions. It is this phenomenon generated by drastic differences in pressure on the SW that is called the sonic boom. The adverse SB effect on human beings and animals (both psychological and physiological) and on buildings (destructive) necessitated a restriction on the value of the admissible excess pressure on the SW. This norm was periodically revised as more information on the SB influence on the environment was available. Taking into account the prediction [34] for 2012 (15 Pa), we can state that the ecological requirements became more severe almost by an order of magnitude during the last 40 years. According to the classical theory of Whitham [35] used in design of supersonic planes, the value of the pressure difference on the bow SW is determined by the distribution of the volume and lift force along the aircraft and decreases with an increase in the flight altitude and aircraft length and with a decrease in the aircraft weight. Conditions that ensure SB reduction make it difficult to ensure the aerodynamic and cost efficiency of the aircraft with a given payload and light range. For aircraft with a mass of more than 100 tons (as a result of the increasing contribution of the lift force to the SB), it is problematic to satisfy the modern ecological requirements even at the expense of lower cost efficiency of the aircraft [18, 36]. Supersonic passenger airplanes of the first generation (Concorde, TU-144) in the cruising supersonic flight generated a pressure difference on the bow SW at the ground level of the order of 100 Pa. For this reason, supersonic flights over the USA territory were prohibited from the very beginning.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

5

near zone

Middle zone rear shock wave

far zone

bow shock wave

Figure 1. Sonic boom formation.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Fig.1. Sonic boom formation.

In view of the limited capabilities of conventional methods, ITAM SB RAS researchers investigated the possibility of SB reduction by means of an active impact on the disturbed flow with the use of mass or energy addition or energy removal in the vicinity of the flying vehicle. The objective of the active impact on the flow around the aircraft is to ensure a disturbed pressure distribution in the near zone, which would ensure SB reduction on the ground level in the course of its evolution at large distances. To ensure cost efficiency of the aircraft, its lift-to-drag ratio should be retained at the level provided by the initial configuration.

2. PHYSICAL GROUNDS OF THE CRYOGENIC IMPACT ON THE FLOW The capabilities of active control of the disturbed flow formation near the aircraft are substantially extended by the process of energy removal, which is ensured by various methods of organization of a cryogenic impact on the flow process. The velocity of sound (a) determining the disturbance propagation velocity is known to be proportional to T , where T is the static temperature of the flow. This fact implies that the velocity of sound can be reduced by decreasing the flow temperature. Gas-dynamic relations on the oblique SW yield a direct relation of the SW intensity Ps  ( Ps  P ) / P with the ratio of the static temperatures

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

6

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

of and behind the SW [37]. Here, P is the static pressure, and the subscripts s and  refer to the flow parameters behind and ahead the SW, respectively. Using the equation of state of an ideal gas, we can write the relation

Ts Ps     , T P   s

(1)

where  is the gas density. Using the ratio of the static pressures on the oblique SW

Ps 2  1  M 2 sin 2     Ps  1 , P   1  1

(2)

we can express the normal-to-SW component of the Mach number M n  M  Sin , where



is the SW inclination angle and



M 2 sin 2  

is the ratio of specific heats, via the shock intensity

1  Ps 

k 1 k 1 .

2k k 1

(3)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Using Eq. (3), we transform the ratio of flow densities on the oblique SW: s   1     1

1 2 1 (   1)( M 2 sin 2  )



 1  1

1 4 1   1 (   1)(   1)  1   Ps    1 

(4)

Substituting Eqs. (2), (4) into Eq. (1), we obtain the direct dependence of the ratio of the static temperatures on the SW on the SW intensity:     Ts  1 4  (1   Ps )  1 T  1   1   (   1)(   1)  1   Ps     1   

(5)

Figure 2 shows the dependence (5).

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

7

Ts / T

Ps Figure 2. Ratio of the static temperatures on the oblique SW versus SW intensity.

Fig.2. Ratio of the static temperatures on the oblique SW versus SW intensity Decrease in SW intensity INCREASE IN SW Рис. INTENSITY V Oblique SW

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3.

Fi of controlling the SW intensity, which is This relation demonstrates the possibilities illustrated in Figure 3. g.3.3 The SW intensity can be reduced by increasing the temperature ahead of the SW or decreasing the temperature behind the SW. The opposite procedure leads to enhancement of the SW intensity. It should be noted that a rigorous increment of temperature on the SW front is needed to change the SW intensity, in accordance with Eq. (5) derived for an ideal gas, which is rather difficult to provide. At the same time, Eq. (5) indicates the prevailing effect on the SW intensity of the temperature determining the disturbance propagation velocity over the corresponding changes in the gas density, which reduce the efficiency of the temperature effect. Naturally, the efficiency of the temperature effect is determined by the distance where the flow temperature is changed in practice (from the surface generating the disturbances to the SW front) relative to the ideal conditions.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

8

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

3. OBJECTIVES OF RESEARCH Based on studying the effect of the body shape on the parameters of the SB generated by this body, chirkashenko and yudintsev [29] determined a class of modified (with the use of spherical bluntness of the nose part) of power-law bodies, which ensured a significant (up to 50%) reduction of the bow sw intensity in the middle zone, as compared with the initial body ( rb

 0 ) with an identical aspect ratio. The length of the SB middle zone is determined by

the distance where the bow sw generated by the blunted nose part interacts with the intermediate sw propagating further downstream and giving a velocity greater than the bow sw velocity. The intermediate sw is formed near the body surface owing to interaction with the accelerating flow (owing to spherical bluntness). The intensity of the intermediate sw and its location on the disturbed pressure profile for a given flight mach number м is determined by the geometric characteristics of the modified body: power index n , aspect ratio   l / d m , and relative bluntness radius rb  2rb / d m ( l, d m are the body length and maximum diameter, respectively) Figure 4 shows the behavior of the asymptotic parameters of the bow SW intensity with distance from the body (in terms of the height H ) for different aspect ratios and bluntness radii. Here, K  H / d m is the relative distance from the body (caliber). The jumps of the asymptotic parameters on its dependences on the relative distance (Figure 4) correspond to interactions of the bow and intermediate shock waves. A typical feature of modified power-law bodies [38] is the fact that they ensure (at certain degrees of spherical bluntness) a greater decrease in the drag force, as compared with the initial power-law bodies ( rb  0 ) having an identical aspect ratio. It should be noted that

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

power-law bodies ( rb  0 ) at moderate supersonic velocities are optimal in terms of the wave drag in the class of bodies of revolution with a given aspect ratio. Reduction of the drag force and the SB intensity due to the effect of bluntness indicates the prospects of using such bodies. As the aspect ratio is increased, the effect of bluntness on drag reduction becomes weaker. At the same time, the length of the SB middle zone increases with increasing aspect ratio and degree of body bluntness (Figure 4). Thus, the requirements imposed on the body geometry to ensure a decrease in the SB parameters and drag reduction are contradictory, which results in a small length (over the altitude) of the region of bow SW intensity minimization. In cruising flight of an aircraft with a mid-section diameter of the order of 3 m at the cruising height of 18,000-21,000 m, which corresponds to K  6000  7000 , a mechanism preventing the interaction of the intermediate SW with the bow SW up to these altitudes is needed to increase the length of the SB middle zone with minimum possible expenses for the drag increase (or without them). This can be done by several methods: shifting the area of formation of the intermediate SW in the downstream direction, decreasing the intermediate SW intensity, or eliminating the possibility of intermediate SW formation.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

9

p s  K 3 / 4

K

а

p s  K 3 / 4

K

b

Figure 4. Decay of the intensity with distance from the modified body versus the Fig. 4. Decay ofbow theSW bow SW intensity with distance frompower-law the modified poweraspect ratio and bluntness of the body at М = 2.03: (a) λ = 4,

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

law body versus the aspect ratio and bluntnessrbof=0the body at М = 2.03: (a) λ (1), 0.1 0.2 (3); λ =and 6, rb0.2 =0 (3); (1), 0.1 and(1), 0.3 0.1 (4). (2), 0.2 (3), and 0.3 (4). = 4,(2),rb and =0 (1), 0.1(b) (2), (b)(2), λ =0.2 6, (3), rb =0

Fig. 5. Flow around a modified power-law body.

Figure 5. Flow around a modified power-law body. 1 – modified body, 2 – bow SW, 3 – expansion wave, 4 – compression waves, 5 – hanging SW.

1The – modified body, 2 in– anbow SW,vicinity 3 – expansion wave, 4body – intermediate SW is formed immediate of the modified power-law surface (Figure 5). The flow expanding over the spherical bluntness generates a system of c nging SW. and power-law parts of the compression waves behind the line of junction of the spherical body; interaction of these compression waves leads to the formation of a hanging intermediate SW.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

10

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

As the intermediate SW is formed in an immediate vicinity of the body surface, it becomes possible to control this process by means of surface cooling. The flow temperature is reduced by heat conduction within the limits of the thermal boundary layer. For an additional decrease in flow temperature, the heat transfer process can be intensified by organizing convective heat transfer by means of distributed injection of the coolant into the region of hanging SW formation. The flow around the effective body formed by the injected coolant, however, can generate additional disturbances, which can enhance the bow SW from the blunted nose part. To avoid this adverse factor, it is necessary to ensure an appropriate scheme of injection determined by the distribution of the perforation over the body surface and by the coolant injection regime. Numerical and experimental studies were performed to verify the possibility of using these cryogenic technologies for reduction of the SB generated by the vehicle and the drag force of the latter.

4. TEST CONDITIONS As it is impossible to ensure complete modeling of SB propagation under laboratory conditions (because of the limited size of wind tunnels), a comprehensive study was performed with a combined experimental and numerical method developed at ITAM SB RAS [39,40]. The method is based on measuring the disturbed pressure profiles near the model mounted in the wind-tunnel test section and on further recalculation of these profiles to large distances with the use of the quasi-linear theory [35].

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

4.1. Experimental Facility The experiments were performed in a T-313 supersonic wind tunnel based at ITAM SB RAS. The test section size was 0.6х0.6х2.0m, and the test parameters were М=2.03, Re1=25*1061/m, and Тo=260-300 K [41]. The scheme of the experiment is shown in Figure 6. The model 1 on a tail sting 2 was mounted on a bracket 3 rigidly fixed on the upper wall of the test section 4. Three through orifices on the bracket for mounting the tail sting 2 and a collect clamp 16 of its fixation on the bracket made it possible to change the model position with respect to the measurement plate 5 discretely over the test section height and continuously in the streamwise direction. The measurement plate was placed on the lower wall of the test section, as is shown in Figure 7. The mean measurements were performed with the perforated part of the plate being located at 250 mm from the lower wall of the test section. The structure of the measurement plate used was determined in advance by conditions of experimental studies of interference problems. The width and length of the rectangular plate surface were 440 and 400 mm, respectively. The measurement base of the plate was organized with the use of a detachable perforated insert with attached pressure tubes, which was flush-mounted on the plate surface with the perforation plane being shifted by 38.5 mm with respect to the streamwise axis of symmetry of the plate. The plate could be continuously moved in the transverse direction with its distance from the lower wall of the test section being fixed. The measurement base of the

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

11

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

plate, which included 100 pressure taps 0.5 mm in diameter arranged on the insert with a constant step of 3.5 mm, made it possible to register the pressure profile over a length of 346.5 mm. The first measurement point was located at a distance of 20 mm from the leading edge of the plate. To reduce the length of the pressure tubes and, hence, the time of pressure equalization, the pressure gauges 6 were located in the pressure chamber of the wind tunnel. The length of the pipelines from the measurement insert to the measurement cavity of the gauges was 1.2–1.5 m. The gauges arranged in blocks of five gauges each were placed into an insulated container, which made it possible to reduce the effect of temperature variations in the pressure chamber in the course of the experiments on the gauge readings. These measures allowed us to reduce the time of pressure equalization in the gauge cavities to 20–30 s after the wind-tunnel start-up process and, correspondingly, the amount of compressed air used. The signals from the pressure gauges were fed through an automated commutation system of the wind tunnel to NR34970A 45-channel recording voltmeters 7, which ensured registration of 5.5 decimal digits with subsequent transfer of the digital information to the PC 8 for storage in the database and further processing. Figure 8 shows the location of the elements of the measurement-calculation system in the panel room of the T-313 wind tunnel.

Figure 6. Sketch of the experiment. 1 – model, 2 – tail sting, 3 – bracket, 4 – wind-tunnel test section, 5 – perforated measurement plate, 6 – pressure gauges, 7 – registration system, 8 – PC, 9 – reservoir with liquid nitrogen, 10 – pipeline for liquid nitrogen feeding to the pressure chamber, 11 – pipeline d = 8 mm, 12 – pipeline d = 6 mm, 13 – intermediate sting, 14 – fluoroplastic insert, 15 – replaceable nose part, 16 – collect clamp, 17 – thermal insulator.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

12

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Figure 7. Location of the model and measurement plate in the test section of the T-313 wind tunnel.

In the experiments, we used domestic absolute-pressure strain gauges (TDM9-A-0.1 and TDM9-А-0.06 with nominal values of 0.1 and 0.06 MPa, respectively) and foreign strain gauges (KPY42-A with a nominal value of 0.16 MPa). The gauges were calibrated prior to each series of experiments. The errors of the gauges obtained in calibrations in the pressure range from 0 to 1 absolute atmosphere controlled within 13.4 Pa by a rider-type absolute-pressure meter stayed within the ranges indicated in the table below. The results were processed with a special program in accordance with a developed algorithm.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Gauge type TDM9-А-0.06 TDM9-А-0.1 KPY42-A

 Pmax  Pапр.  Pзад. , Pa 135 140 120

P 

 (P

апр.

Pзад. ) 2 / n  1, Pa

60 63 58

Figure 8. Location of the elements of the measurement-calculation system in the panel room of the T313 wind tunnel. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

13

4.2. Models and Technologies of Their Cooling A special system was developed and fabricated for injection of the coolant (liquid nitrogen) into the model (Figure 6). This system includes a TRZhK-2U reservoir for transportation of liquefied gases (oxygen, nitrogen) 9 with a volume of 0.5 m3 and a feeding pipeline consisting of segments of metallic corrugated hoses 10 (d = 25 mm and total length 5 m) and a tube 11 (d = 8 mm and length 1.5 m) made of 12Kh18N10T steel, which was connected at the entrance to the model cavity with a copper tube 12 (d = 6 mm). The feeding pipeline in the form of a tube (8x1 mm) enters the inner cavity of the model via internal channels in the intermediate sting 13 and the insulating fluoroplastic insert 14 providing thermal decoupling between the model and the sting. Exhaustion of liquid nitrogen from the model cavity and its subsequent exhaustion in the base region of the model into the ambient flow occur along streamwise grooves on the outer surface of the insulating insert 14. To prevent exceeding of the admissible pressure in the model cavity during its cooling, the total cross-sectional area of these grooves was made greater than the cross-sectional area of the feeding pipeline and corresponded to an equivalent orifice with a diameter d = 7 mm. The outer surface of the feeding pipeline was thermally insulated from the ambient atmosphere and metallic elements with a fluoroplastic strip, foam rubber, and fluoroplastic inserts 17. The initial variant of the model in the form of a modified power-law body of revolution (

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

  6 , n = 0.75, and rb  2rb / d m = 0.2) with the mid-section diameter d m = 50 mm was made of 12Kh18N10T steel. It consisted of two parts jointed at a distance of 200 mm from the model tip with a threaded connection. Model variants 1, 2, and 3 with different schemes of perforation for injection of the super cooled gas from the model surface (Figure 9) were obtained with the use of replaceable nose parts 15 (Figure 6) 72 mm long, which were connected with the initial model via a threaded connection. These nose parts were made of LS-59 brass alloy whose coefficient of temperature deformation is close to the material of the initial model [43]. Perforation with the orifice diameter of 0.3 mm was provided on the lower surface of model 1, beginning from the cross section located near the line of junction of the spherical and power-law surfaces to a distance of 40 mm from the model tip, over a sector of 180º (from the side of the measurement plate). The degree of perforation determined by the ratio of the cross-sectional area of the orifices to the total cross-sectional area of the surface S p  S p / S s decreased with distance from the model tip from 8% to 1.3% by changing the distance between the cross sections with an identical number of orifices uniformly distributed over the cross section perimeter. The diameter of the orifice equivalent to the total area of perforation was 4.63 mm. The perforation configuration of model 2 differed from that of model only by the increase in the diameter of some orifices up to 0.35 mm. On model 3, the zone of perforation arranged in the form of orifices 0.5 mm in diameter started at the model tip and ended at a distance of 9 mm from the model tip. The averaged degree of perforation was approximately 40%. The diameter of the orifice equivalent to the total area of perforation was substantially greater than that of model 2 and reached 6.66 mm.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

14

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Thermocouple T3

Initial model configuration 11 1

Model configuration – 1, 2

Model configuration – 3

Figure 9. Location of perforation and thermocouples on theand modelthermocouples variants examined.on Fig. 9. Location of perforation

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

the -5model variants examined. Pm ∙10 , Pa

-5

P∙10

, Pa

Figure 10. Pressure in the cavity of model 2 versus the pressure of liquid nitrogen in the reservoir.

а

P 10-5 Pa

b

P 10-5 Pa

Fig. 11. Flow rate of the injected liquid versus pressure: (a) flow rate through the

Figure 11. Flow rate of the injected liquid versus pressure: (a) flow rate through the perforation, (b) perforation, (b) ratio the flowand rates theofperforation the ratio of the flow rates through theof perforation the through base region the model; 1- and water, 2- base liquidregion nitrogen.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

15

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 12. Cooling of the model on the testbench: (а) view from the side of the nose part; (b) view from the side of the tail part.

To monitor the temperature of the model surface, the nose parts were equipped with copper-constantan thermocouples with a wire diameter of 100 µm. The hot junction was mounted at the level of the outer surface of the model and was insulated from the model mass by a layer (0.3÷0.4 mm) of the VS-9T heat-resistant glue. The cold junctions were brought out from the pressure chamber to the ambient atmosphere with the use of electrode and compensation wires. The signals from the thermocouples were recorded by the NR34970A multichannel integrating voltmeter, which provided the error of temperature measurement smaller than 1.5 degrees, with allowance for measuring the cold junction temperature. Different numbers of thermocouples were used in the test for various reasons; their arrangement corresponded to the schemes in Figure 9. Prior to the wind-tunnel experiments, the cooling technology was developed for each model variant. For this purpose, the system of liquid nitrogen injection with all elements of the model mounting was assembled outside the test section. The parameters measured during liquid nitrogen injection were the temperature of the model surface, the pressure in the TRZhK-2U reservoir, and the excess pressure in the inner cavity of the model (Pm), which was measured by a standard indicating pressure gauge with the measurement range from 0 ÷ 1аtm. connected by a pressure tube with the inner cavity of the nose part of the model. The measurement results are plotted in Figure 10. The ratio of the pressure in the model cavity (Pm) and the pressure in the liquid nitrogen reservoir (P) approximately equal to 0.38 agrees

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

16

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

with the ratio of the cross-sectional area of the feeding pipeline and the total cross-sectional area of nitrogen injection into the flow. To obtain reliable information on the amount of liquid nitrogen used, the flow rate through the coolant injection system was directly measured with the use of water. The measured results are plotted in Figure 11. The flow rates of liquid nitrogen through the perforation (Figure 11а) were obtained by recalculation of results obtained with the use of water. The ratio of the coolant flow rates (Figure 11b) through the perforation and the grooves for liquid exhaustion to the base part of the model is practically independent of pressure in the range examined. Figure 12 illustrates the test bench experiments with the system of liquid nitrogen injection. The following cooling technologies were adopted, based on results of test bench experiments. For the initial variant of the model (without perforation): liquid nitrogen was injected approximately during 5 min with the excess pressure in the reservoir Р= 0.03÷0.05 MPa. This process reduced the initial temperature of the model T 1, T 2  18  20 C to 0 C. After that, the pressure Р = 0.09 ÷ 0.1 MPa was reached, and cooling was continued during 1.5 ÷ 2.0 min. The temperature of the model T2, which is the most inertial one in terms of cooling, decreased to –196 C. After the liquid nitrogen injection was terminated, the model surface temperatures increased during 20 min to T1 = –29 C and

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

T2 = –50 C. It should be noted that the surface temperature of the initially cooled model noticeably increased with a drastic increase in Р, which seems to be caused by formation of zones near the thermocouples, where the coolant was in a two-phase or gaseous state. A thin (1÷1.5 mm) hoarfrost layer starts forming on the model surface, beginning from the nose part of the model. For model variants 1,2,3 (with perforation): the initial level of pressure in the reservoir during cooling remained unchanged (Р = 0.03÷0.05 MPa), but the time needed for the model temperature to decrease to 0 C was more than twice greater because of the greater flow rate of the coolant through the perforation and reached approximately 12 min for model 3. A further decrease in the model temperature was ensured at Р = 0.09÷0.1 MPa during 3÷4 min. Prior to wind-tunnel actuation, the model mounted in the test section was cooled with the ejection responsible for coolant ejection operating at the minimum pressure. Figure 13 shows the results of visualization of cooling the model mounted in the wind-tunnel test section. In visualizing the supersonic flow around the cooled model, no hoarfrost was found on the model surface.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

17

Figure 13. Visualization of cooling the model (variant 2) in the test section of the T-313 wind tunnel.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

4.3. Method of Measurements and Data Processing A principle of superposition of disturbances was used to obtain the useful signal, which was the distribution of the excess relative static pressure behind the reflected SW generated by the model over the control surface. The interaction of the disturbed flow generated by the model with the background pressure distribution on the measurement plate (without the model) was assumed to follow a linear law, which is valid if the useful signal exceeds the background signal by an order of magnitude. Therefore, the useful signal can be determined as the difference between the total signal and the background signal (without the model):

Psig ( x )  Psig bach ( x )  Pbach ( x ) , where

Psig ( x )  [ Psig ( x )  P ] / P ;

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

18

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

P sig  bach ( x )  [ Psig  a ( x )  P ] / P ; Pbach ( x )  [ Pbach ( x )  P ]/ P . A series of tests was performed to obtain an acceptable background distribution of pressure over the measurement plate and free-stream uniformity in the area of disturbance propagation between the model and the plate to ensure required measurement accuracy. It was rather difficult to satisfy these requirements for the model under study, which had the length of the contoured nozzle part equal to 300 mm. This fact is illustrated by the background pressure distributions obtained for two different positions of the plate in the test section (Figure 14). The location of the pressure taps in the longitudinal plane of symmetry of the wind tunnel ( Z m.b.

 0 ) is ensured by shifting the plate by 38.5 mm from its symmetric position with

respect to the wind tunnel. In this case, the length of the measurement area at the excess pressure level P  0.1 is approximately 250 mm. The further drastic increase in pressure is caused by the compression wave whose beginning is located substantially more upstream than the intersection of the characteristic cone emanating from the corner point of the leading edge of the plate with the perforation plane. If the plate is located symmetrically with respect to the test section, the beginning of compression wave formation is shifted downstream to a distance corresponding to the interaction of the disturbed pressure initiated by the leading edge of the plate with the perforation plate. The perforation plate is also shifted with respect to the longitudinal plane of symmetry of the wind tunnel by Z m.b.  8.5 mm. The upstream shift of the zone of formation of the maximum pressure on the Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

measurement base of the plate with its symmetric location ( Z m.b.  0) shows that disturbances forming this zone are induced by the interaction of the side surface of the plate, which was located near the size wall of the test section, with the developed boundary layer.

Figure 14. Distribution of the background pressure over the measurement plate for different locations of the latter with respect to the test section. 1, 2, 3 – Z m.b.

 0: Pr.-2694, 2700, 2705; 4, 5 – Z m.b. 

38.5 mm: Pr.-2695, 2710.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

19

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 15. Excess pressure profile in the vicinity of the initial model (without cooling): 1 – measured profile, K  3.7 , 2 – reconstructed profile.

The results obtained display good agreement of the pressure distributions in different tests with one installation of the plate (Pr-2700, 2705). A noticeable difference is observed in the distributions obtained in tests with repeated installation of the plate (Pr.-2694, 2700, 2705 and Pr.-2695-2710), which is explained by inaccuracy of reproduction of the plate location in the wind tunnel in different tests. Therefore, background pressure measurements were performed for each installation of the plate, and the results of these measurements were taken into account in processing the corresponding measurements with the model. Comparisons of the pressure distributions obtained in multiple measurements during one run show that the scatter does not exceed the maximum values obtained during calibrations of the pressure gauges. Nevertheless, the thus-processed measured pressure profiles (Figure 15) are not free from fluctuations, which are apparently caused by free-stream disturbances. This is confirmed by comparisons with the results obtained by the probing method [29] eliminating the effect of flow inhomogeneities. The distortion of the pressure distribution in the tail part of the positive phase of the pressure profile (Figure 15) is apparently caused by disturbances emanating from the threshold of the lower wall of the wind-tunnel test section. If there is no model in the test section, these disturbances do not reach the measurement base of the plate and, correspondingly, are not included into the background pressure distribution. If the model is mounted in the test section, the disturbances from the threshold (in the form of a compression wave) arrive on the model surface, are reflected from the model surface, arrive on the measurement base of the plate, and are registered in the total signal. As a result, there are distortions in the tail part of the processed pressure profile. To eliminate such effects of the wind tunnel in recalculating the pressure profile to large distances, a smoothing procedure was applied to the initial pressure distributions. This procedure is based on the a priori information in the form of a reliable pressure distribution measured behind the incident SW [29] under the conditions of a constant excess pressure impulse. All measured pressure profiles were recalculated to large distances after the smoothing procedure.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

20

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

4.4. Method of Recalculaton of the Measured Results to Large Distances from the Model Recalculation of the pressure profiles to large distances was performed under the assumption of a homogeneous atmosphere by the method [42] based on the quasi-linear theory [35]. The relations determining the value of the disturbed pressure on the characteristic and the position of this characteristic at an arbitrary distance from the initial profile have the following form in the second approximation with respect to the disturbance intensity: 1/ 2

r   p   p0  0  r

;

r x   r  k1pr01/ 2  r1/ 2  r01/ 2   k 2 p 2 r0 ln    x0  r0  ; Here,  p  p  p  , p

(   1) M 2 1 / 2 2

k2 

2 

, and

2   M 2  1 , M  is the Mach number, k 1  (  1) M  ,





is the ratio of specific heats.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

In recalculating the initial pressure profiles (measured at a distance r0) to the distance r, the positions of discontinuities in the deformed profile were determined on the basis of the known result that the weak SW ( P  1 ) is a bisector of the angle between the incoming characteristics of the undisturbed and disturbed flows [35]. To analyze the possibility of recalculation based on the quasi-linear theory, the method was tested with the use of experimental results in a wide range of disturbed pressure ( P  0.1  1.0 ) and different shapes of the initial pressure profiles [37].

5. RESULTS AND DISCUSSION 5.1. Surface Cooling First, the initial model (without perforation) was tested to determine the pure influence of the decrease in the model surface temperature on the formation of a hanging SW induced by interaction of the overexpanded (in the flow around the spherical bluntness) with the model surface [31]. Figure 16a shows the profiles of the dimensionless excess static pressure behind the reflected bow SW, which were measured near the cooled and non-cooled model surface (K = 3.7). Figure 16b shows an up scaled fragment of these pressure distributions in the region of formation of the hanging shock. Figure 17 corresponding to Figure 16b shows the pressures measured at different times and the confidence interval.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

21

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Model cooling does not involve any significant changes in the flow structure in the bow SW region. On the profile for the cooled model, in the region of formation of the hanging shock (see Figure 16b, 2), however, the excess pressure decreases behind the intermediate shock and increases ahead of the intermediate shock (with respect to the situation for the noncooled model, Figure 16b, 1). As a result, the pressure difference on the intermediate shock near the cooled model surface is almost one half of the corresponding pressure difference on the non-cooled model. The reduced (as compared with the non-cooled model) level of pressure behind the intermediate shock persists in the downstream direction up to the rarefaction wave, which substantially reduces the positive impulse of the SB wave. Thus, cryogenic forcing reduces the intensity of the intermediate shock wave and the positive impulse of the SB wave with an unchanged bow SW intensity. Figure 18 shows the change in the flow temperature on the surfaces of the non-cooled and cooled models, presented as the dependences of temperature on the real time of experiments. The vertical bars on the abscissa axis correspond to the beginning and end of the supersonic regime of wind-tunnel operation. During 1 min after the test time beginning (Figure 18a, 1), the temperature permanently decreases owing to unsteady heat transfer on the surface of the non-cooled model and reaches −5◦C by the end of the test time in the region of the junction between the spherical and power-law surfaces (Figure 11, thermocouple T1 mounted at a distance of 5 mm from the model tip). Steady heat transfer between the model and the incoming flow is established on the surface of the cooled model (Figure 18b) soon after the test regime is stabilized. The temperature in the region of formation of the hanging shock reaches −50 to −55◦C, which is substantially higher than the free-stream static temperature T  143 К . The temperature measured by the second thermocouple located on a less steep segment of the generatrix (Figure 18b, thermocouple T2) is close to the free-stream static temperature. Thus, the flow temperature near the model surface in the region of hanging shock formation can be reduced approximately by 50◦C by model cooling. The ratio of temperatures determining the hanging shock intensity changes from Ts / T  1.87 to Ts / T  1.52 , i.e., by 18.5%.

а

b

Fig. 16. Pressure profiles measured near the model: 1 – non-cooled

Figure 16. Pressure profiles measured near the model: 1 – non-cooled model, 2 – cooled model, 3confidence of measurements. model, 2interval – cooled model, 3- confidence interval of measurements.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

22

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Figure 17. Pressure profiles in the region of intermediate SW formation: Pr 2708 – non-cooled model, Pr 2709 – cooled model. 1- confidence interval of measurements.

Figure 18. Changes in temperature of the model surface; (а) non-cooled model; (b) cooled model.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The estimates calculated by Eq. (5) (Section 2) show that a 13.5% decrease in the static temperature behind the shock wave of intensity P  0,1 leads to a twofold decrease in intensity. As the calculations predict that the temperature ahead of the hanging shock is higher than the free-stream static temperature (approximately by 3–4%), we can assume that the ratio of SW intensities remains unchanged as the hanging shock waves (generated by the non-cooled and cooled models) propagate between the model and the measurement plate. Then, the reduction of the hanging shock intensity is caused by the decrease in temperature behind the shock wave due to model cooling. Therefore, the temperature ratio reach Ts / T  1 , which prevents the hanging shock formation under the test conditions used. Thus the flow temperature near the model surface has to be additionally reduced by 61ºС by means of increasing the flow rate of the coolant, decreasing the thickness of the model wall, or replacing the model material by a material with a higher thermal conductivity. These estimates suggest that the decrease in intensity of the intermediate SW is caused by the lower velocity of propagation of disturbances in the region of their formation. The heat transfer between the model surface and the flow in the region of hanging shock formation proceeds via heat conduction of air in the thermal boundary layer. The result obtained shows that this is sufficient to ensure a noticeable effect on the parameters of the wave structure nucleating in an immediate vicinity of the surface (hanging shock). Under real conditions, the supersonic cruising flight proceeds at an altitude of 18,000 m at T  217 К , which exceeds the static temperature of the flow under the test conditions by a factor of 1.52. Therefore, a given temperature ratio is reached under real conditions owing to a more significant decrease in surface temperature. Thus, to obtain reliable information Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

23

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

about the influence of the disturbance velocity on the hanging shock formation, it is necessary to model real temperature conditions. Schlieren pictures taken during flow visualization with an IAB-451 shadowgraph showed that the flow regimes are identical for the model without coolant supply and for the cooled model, which is demonstrated in Figure 19. The results obtained show that it is possible to control the process of formation of wave structures, such as the hanging shock, by means of surface cooling in the region of nucleation of these structures. The evolution of the disturbed pressure profiles with distance from the non-cooled and cooled models is illustrated in Figure 20. In the case with a distance K = 500 from the noncooled model (Figure 20a), the bow SW is formed owing to interaction of the intermediate SW with the SW generated by spherical bluntness, which leads to a higher bow SW intensity. The intermediate SW generated by the cooled model and propagating with a lower velocity because of its decreasing intensity persists on the pressure profile. On the cooled model, the intensity of the bow SW generated by spherical bluntness is 25% lower than that on the noncooled model.

Figure 19. Schlieren pictures of the flow around the initial model (power-law body of revolution ( n  0, 75;   6; r  0, 2; M  2, 03 ); (а) non-cooled model; (b) cooled model. 

з

а

b

c

Fig. 20.profiles Pressure profiles bow SW distances at different distances from Figure 20. Pressure behind the behind bow SWthe at different from the model; 1 – non-cooled model, 2 – cooled model. the model; 1 – non-cooled model, 2 – cooled model.

The recalculated results predict that the intermediate SW on the pressure profile persists up to K = 1400, with the bow SW intensity decreasing to 50%. At a distance K = 1500 from the cooled model (Figure 20b), the bow SW on the pressure profile is formed owing to interaction with the intermediate SW. The bow SW intensity is significantly lower than the intensity of the bow SW generated by the non-cooled model.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

24

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Further decay of this wave induced by the positive pressure gradient behind the wave occurs much slower, as compared with the bow SW decay on the non-cooled model. When the profile shape becomes similar to the N-wave, the process of the bow SW decay reaches almost an asymptotic law, which predicts that the SW intensity is determined by the positive impulse of the SB wave. As the cryogenic forcing reduces the value of the positive impulse, the intensity of the bow SW on the cooled model at large distances (Figure 20c, K = 6000) corresponding to the asymptotic law of decay, the intensity of the bow SW generated by the cooled model is approximately 12% lower than the intensity of the bow SW generated by the non-cooled model. The result obtained shows that it is possible to control the flow around the body by means of cooling its surface and also to control the parameters of the sonic boom generated by this body. Thus, the possibility of controlling the process of formation of wave structures, such as the hanging shock, in an immediate vicinity of the surface inducing these structures by means of surface cooling in the region of nucleation of these structures is demonstrated. The effect of surface cooling persists in the disturbed flow at large distances from the body. Reduction of intensity of the intermediate SW and the positive impulse of the SB wave near a modified power-law body by means of surface cooling allows the zone with a significantly reduced SB level (down to 50%) to be substantially extended, with an additional 12% decrease at large distances. The calculated estimates suggest that the main method of changing the flow structure in the vicinity of the intermediate SW is the reduction of the velocity of disturbances near the surface generating this wave.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

5.2. Coolant Injection from the Surface Reduction of flow temperature by 50◦C, reached by coolant pumping in the model cavity made it possible to reduce the hanging SW intensity almost by a factor of 2 and, thus, to shift the point of its interaction with the bow SW to a distance equal to 1500 body diameters (calibers). This value is substantially smaller than the supersonic airplane cruising flight altitude equal to 6000–7000 calibers. Difficulties arising in further reduction of flow temperature by this method and the necessity of further reduction of temperature under real flight conditions show that this method of controlling the hanging SW intensity is insufficiently effective. To obtain the middle zone length of interest for practice, active controlling of hanging SW parameters was additionally studied to increase the efficiency of reduction of flow temperature in the region where this SW appears [32]. Hanging SW formation was affected by organizing various schemes of distributed injection of the coolant from the model surface to the zone of hanging SW generation. In contrast to surface cooling (without injection), coolant injection allowed a significant decrease in flow temperature in the region of hanging SW formation and, as a consequence, in the velocity of propagation of disturbances owing to convective heat exchange between the injected coolant and the free stream. To avoid possible generation of additional shock waves by the coolant jets, it is necessary to ensure an appropriate distribution of injection intensity over the body surface.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

25

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The perforation schemes of models 1 and 2 (Figure 11) were chosen to ensure coolant injection directly into the region of formation of the hanging SW. There are grounds to believe that the hanging SW, which arises owing to interaction of the flow accelerating in its motion around the spherical bluntness with the model surface, starts to form in the flow around a modified power-law body without coolant injection in the vicinity of the junction between the spherical and power-law surfaces. The hanging shock emerging in an immediate vicinity of the model surface, which is an envelope of the family of converging compression waves, transforms to a hanging SW with distance from the body. As the regime of coolant injection had to ensure effective cooling of the flow between the hanging shock and the model surface and the layer of the supercooled gas formed owing to evaporation of the coolant entering the flow should not generate additional shock waves, the degree of surface perforation was reduced in the downstream direction. To determine the influence of coolant injection from the spherical surface of the body, we performed tests with model 3 (Figure 11). Model 1. The disturbed pressure profiles measured in the flow near the surface of model 1 (Figure 21, K = 3.7) with injection of liquid nitrogen with an initial pressure of 0.135 MPa show that coolant injection exerts a noticeable effect on the process of formation of the hanging SW. Under such cryogenic forcing, the bow SW intensity remains almost constant, but the intensity of the expansion wave following behind the bow SW is reduced in the entire flow region, except for the region of formation of the hanging SW on the non-cooled model. In this region of the flow, which arises under the action of the injected coolant, the flow continues to expand with intensity higher than the intensity of the expansion wave generated by the noncooled model. The expansion zone is closed by the hanging SW whose intensity is approximately equal to the intensity of the intermediate SW near the non-cooled model. The downstream displacement of the hanging SW obtained in this case corresponds to half of the model diameter. Owing to the downstream displacement, the interaction of the hanging SW with the bow SW occurs approximately at a distance K=1500 (Figure 21). A decrease in the bow SW intensity is reduced up to 40%, as compared with the non-cooled model. After this interaction, the bow SW intensity increases and further remains at the level corresponding to the non-cooled model (Figure 21, K=6000). The resultant increase in the SB middle zone length to 1500 calibers, as compared with the initial model (K=500), is comparable with the value obtained with surface cooling (without coolant injection) and is not of interest for practice. Model 2. Figure 22 shows the pressure profiles measured near the surface of model 2 under conditions without coolant injection and with injection of liquid nitrogen with an initial pressure of 0.15 MPa and the results of their recalculation to large distances. In contrast to model 1, this modification of perforation ensures a significant increase in the length of the region and the degree of flow overexpansion behind the bow SW (Figure 22, K = 3.7), which leads to a significant downstream displacement of the compression wave acting as a precursor for hanging SW formation.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

26

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Fig. 21. Dynamics of the disturbed pressure profiles generated by model 1: 1–initial model,

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 21. Dynamics of the disturbed pressureof profiles generated model 1: 1–initial model, 2– with 2– with injection liquid nitrogen (P=by 0.135 МPа). injection of liquid nitrogen (P= 0.135 МPа).

Fig. 22. Dynamics of the disturbed pressure profiles generated by model 2:

Figure 22. Dynamics of the disturbed pressure profiles generated by model 2: 1- with injection of liquid nitrogen (P= 0.15 MPa), 2 – initial model.

A small increase in the bow SW intensity leads to a significant decrease in the impulse of the positive phase of the SB wave. This fact testifies to a decrease in the pressure drag of the body. An increase in coolant pressure leads to enhancement of flow rarefaction and to a greater downstream length of the expansion region with an insignificant increase in the bow SW intensity. In the course of evolution of the pressure profile deformed by cryogenic forcing, the downstream-shifted compression wave, owing to nonlinear effects, becomes transformed already at moderate distances from the model into a pressure shock (Figure 22, K = 500) located at a substantially greater distance from the bow SW than that in the case with model 1. In the course of its propagation, the bow SW induced by the bluntness intensely decays under the action of the following expansion wave before it can start interacting with the

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Control of the Sonic Boom Generated by a Flying Vehicle…

27

hanging SW. The process of decaying of the bow SW generated by the model without coolant injection is rather long owing to interaction of the compression wave propagating behind the bow SW and formed owing to interaction between the bow SW and the hanging SW. The process of intense decaying of the bow SW generated by the model with coolant injection persists up to large distances from the body (Figure 22, K = 6000). According to recalculation results, the total length of the middle zone of this model can be increased to values equal almost to 7000 calibers, while the SB intensity is reduced by more than 40%. Figure 23 shows the flow temperatures measured in the experiment near the surface of model 2 without coolant injection and with injection of the coolant with P = 0.15 MPa. The vertical bars on the abscissa axis indicate the beginning and end of the design supersonic flow regime. Before the beginning of the supersonic flow, the thermocouples mounted on the cooled model register the temperature almost corresponding to the liquid nitrogen temperature. After the beginning of the supersonic flow, stationary heat transfer is rather rapidly established near the model surface. The thermocouple T3 located at the greatest distance from the model tip registers the maximum temperature, while the thermocouple T2 located on the lower surface in the perforated region registers an additional decrease in temperature to t  200oС. The temperature in the region of the junction of the spherical bluntness and the power-law surface, registered by the thermocouple T2, is substantially lower than the temperature on the cooled model in the absence of coolant injection (Figure 18). Apparently, the entire amount of liquid nitrogen injected into the flow does not evaporate, and there is a liquid phase in the supercooled gas layer formed near the model surface. This fact is indirectly confirmed by the values of temperature registered by the thermocouple T2 (see Figure 23b) located directly in the zone where the coolant enters the model surface (Figure 11). Under the test conditions, the temperature t   200oC corresponds to the temperature of liquid nitrogen. An increase in flow temperature near the model surface at the end of the flow regime (τ > 1335, Figure 23b) is caused by the formation of the gas phase in the model cavity owing to an increase in the coolant flow rate by means of increasing its pressure in the course of the experiment. Figure 24 shows the schlieren pictures of the supersonic flow around model 2 under conditions without coolant injection (Figure 24a) and with injection of liquid nitrogen into the flow in the zone of hanging SW formation (Figure 24b).

Figure 23. Changes in the flow temperature near the surface of model 2: (а) without coolant injection, (b) with coolant injection (Р = 0.15 МPа).

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

28

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

S. G.

а.

b

Fig. 24. Visualization of the flow near the surface of model 3:

Figure 24. Visualization of the flow near the surface of model 3: (а) initial model, (b) with coolant injection (Р = 0.15 MPa). (а) initial model, (b) with coolant injection (Р = 0.15 MPa).

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 25. Flow structure near a modified power-law body with coolant injection to the region of hanging SW formation; 1 – modified power-law body; 2 – perforation; 3 – bow SW; 4 – sonic line; 5 – expansion waves; 6 – compression waves; 7 – supercooled gas layer; 8 – hanging SW.

Injection of liquid nitrogen from the surface of the cooled model leads to formation of density gradients in the flow, which propagate over the characteristic surfaces with distance from the body and become fairly equalized. The formation of a supercooled gas (S.G.) layer near the perforated surface of the model (Figure 24b) is observed. The results obtained allow us to gain new knowledge about the flow structure arising under cryogenic forcing of the flow around a modified power-law body. Figure 25 shows the scheme of the flow formed in the case of injection of liquid nitrogen behind the spherical bluntness in accordance with the schemes corresponding to models 1 and 2. Distributed injection of liquid nitrogen through the perforation into the region of hanging SW formation on the model without cryogenic forcing leads to formation of a supercooled gas layer near the model surface with the presence of the liquid phase of the non-evaporated coolant. This layer is gradually entrained downstream by the incoming flow. The remaining liquid nitrogen rapidly evaporates, which is confirmed by improvement of optical transparency of the supercooled gas layer with distance from the perforation region (Figure 23b). Interaction of the flow accelerated owing to its motion around the spherical bluntness with the supercooled gas layer forming an effective body leads to additional expansion of this flow (Figure 22, K = 3.7). Interacting with the supercooled gas layer, the flow accelerated

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

29

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

owing to its expansion induces a family of converging compression waves near the model surface. This interaction forms a hanging shock, which transforms to a hanging SW with distance from the model. Apparently, this process is determined by the increase in the effective radius of curvature of the surface experiencing the action of the supercooled gas layer and also by the reduction of the velocity of sound and, correspondingly, velocity of propagation of disturbances in the region located inside this layer. The specific feature of such a scheme of coolant injection for model 3 is the specific shape of the supercooled gas layer, with the flow past this layer providing effective overexpansion of the flow almost without any increase in bow SW intensity. A comparison of experimental results on controlling the hanging SW parameters reached on model 2 with results for model 1 shows that the perforation configuration is extremely important in organizing cryogenic forcing. As the flow temperature behind the SW affects the intensity of this wave (as predicted by Eq. (5), Section 3), a significant contribution to pressure redistribution is made by intense convective cooling of the flow in the region between the hanging SW and the body surface, which is ensured by means of coolant injection. The results obtained, however, do not allow us to estimate the dependence of the flow formation process on the temperature and mass of the injected coolant. Additional studies are necessary to obtain this information. Model 3. As in the case with model 2, coolant injection from the surface of the spherical bluntness on model 3 increases the length of the flow expansion region behind the bow SW and, hence, leads to a significant downstream displacement of the zone of generation of the hanging SW (Figure 26а). In contrast to model 2, the intensity of the bow SW near model 3 is substantially higher, owing to a greater effective bluntness radius ensured by the formation of the supercooled gas layer on the surface. Because of the downstream displacement of the hanging SW and an increase in velocity of the bow SW caused by its higher intensity, the length of the middle zone formed in the case with model 3 is greater than 6000 calibers (Figure 26b). Despite an increase in the bow SW intensity near the body, this intensity is reduced with respect to the original model by the minimizing effect of the bluntness, beginning from the distance equal to 2000 calibers; this reduction reaches more than 40% at large distances. Figure 27 shows the schlieren pictures of the flow around model 3. Comparison of injection schemes. The main drawback of the scheme of injection from the spherical surface (model 3) is the increase in energy that should be spent on overcoming the increased drag force of the body and the negative thrust of the coolant jets.

а

b Fig. 26. Disturbed pressure profiles generated by model 3.

Figure 26. Disturbed pressure profiles by model 3. 1nitrogen – initial 2 – with injection of 1 – initial model, generated 2 – with injection of liquid (P=model, 0.18 MPa). liquid nitrogen (P= 0.18 MPa). П Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

30

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Figure 27. Flow visualization near model 3: (а) initial model; (b) with coolant injection.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The scheme of coolant injection from the model surface behind the spherical bluntness in the region of hanging SW formation (model 2) allows the energy needed to overcome the negative thrust generated by the coolant jets to be reduced, because the coolant is injected from the surface aligned at smaller (as compared with the spherical bluntness surface) angles to the incoming flow. In this case, an effective temperature action on the hanging shock intensity is provided, because the temperature of the incoming flow impinging on the hanging shock is not reduced, in contrast to the case with coolant injection from the spherical surface on model 3. The decrease in the impulse of the positive phase of the SB wave obtained on model 2 with an insignificant increase in bow SW intensity testifies to a possible decrease in the drag of the body. The numerical simulations of the flow around model 2 with injection of liquid nitrogen (under the assumption of its complete evaporation from the surface) with the FLUENT CFD software system predicted a moderate (2-3%) decrease in the drag force.

Fig. 28. Length of the middle zone of the sonic boom generated by model 2 versus the

Figure 28. Length of the middle zone of the sonic boom generated by model 2 versus the coolant coolant pressure. pressure.

In practical implementation of cryogenic forcing, the flying vehicle should carry a certain amount of the coolant onboard, which reduces the payload. These expenses can be partially compensated by decreasing the drag of the vehicle owing to the use of a modified nose part and reduction of the drag due to coolant injection. To estimate the energy expenses needed to ensure effective reduction of SB intensity, we studied the effect of the coolant flow rate on the length of the SB middle zone. Figure 28 shows the length of the SB middle zone versus the pressure of the coolant injected from the surface of model 2.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Control of the Sonic Boom Generated by a Flying Vehicle…

31

With an increase in the coolant pressure to 0.06 MPa, the middle zone length substantially increases and reaches more than 4000 calibers. With a further increase in pressure, the growth of the middle zone length becomes slower, and the middle zone length is approximately 7000 calibers for P = 0.15 MPa. Model 2 ensures the middle zone length of 6000 to 7000 calibers for an initial pressure of liquid nitrogen equal to 0.14–0.15 MPa. Under the test conditions used, the ratio of the flow rates of the injected coolant and air impinging onto the body mid-section was 7–8%. This ratio could be reduced to 1.5–2.0% with retaining the middle zone length by means of reducing the sector of coolant injection (approximately to 90o) and the longitudinal size of the perforation region. In view of the significant effect of the perforation scheme on pressure redistribution, a further decrease in the coolant flow rate can be reached by choosing an appropriate scheme of distributed injection. The flow rate should be decreased with the drag force of the flying vehicle retained at a level not higher than the initial one (without coolant injection). To solve this problem, based on the physical concepts of cryogenic forcing, it is necessary to determine the contributions of the coolant mass, its temperature, and its evaporation process to the formation of the flow structure that ensures reduction of the SB level and the drag force of the flying vehicle. With this information available, it is possible to search for schemes of distributed injection and coolant injection regimes to satisfy the conditions of the problem posed. Summarizing the results obtained, we can note that distributed injection of the coolant from the blunted body surface exerts a significant effect on the structure of the disturbed flow in the region of hanging SW formation. The cryogenic forcing considered continues to be effective at large distances from the body and depends to a large extent on the perforation configuration and on the coolant injection regime. Thus, the possibility of controlling the parameters of the sonic boom generated by the flying vehicle and the aerodynamic characteristics of the flying vehicle was demonstrated. A method was proposed to increase the SB middle zone to 7000 calibers. In this case, the bow SW intensity is reduced by more than 40%, as compared with the initial model, which is of significant interest for practice. Apparently, the main mechanisms of cryogenic forcing on the flow structure are the formation of a contact surface by the supercooled gas produced by evaporation of liquid nitrogen and convective cooling of the flow behind the hanging SW, responsible for a lower velocity of disturbances. The method of cryogenic forcing considered here can be used independently or can be combined with available active methods of controlling the SB parameters and the aerodynamic characteristics of flying vehicles (injection of a discrete gas jet or lowtemperature plasma, energy supply by laser or microwave radiation, etc.).

6. APPLICATIONS In addition to the SB reduction problem considered above, cryogenic forcing on the flow by means of surface cooling (without coolant injection) can be widely used to control the aerodynamic characteristics of the flying vehicle and flow parameters in gas-turbine engines.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

32

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

Shock wave structures, such as the hanging SW, are observed both in the external supersonic flow around a flying vehicle and in the channel flows, as is illustrated in Figure 29. The specific feature of these wave structures, namely, their formation in an immediate vicinity of the surface inducing them, makes it possible to control the process of their formation by a temperature impact on the body surface in the zone of their nucleation. Thus, cooling or heating the surface in the region of nucleation of hanging shock waves, one can decrease or increase their intensity and the excess pressure impulse, depending on the conditions of the problem posed. The results obtained on controlling the flow around a modified power-law body with the use of distributed injection of the coolant from the body surface to the region of hanging SW formation were used to develop a method of reduction of the sonic boom generated by a supersonic aircraft configuration [33]. Figure 30 illustrates the dynamics of SB generation by a supersonic aircraft (with the nose part made in the form of a modified power-law body) in the cruising flight with an operating system of cryogenic forcing (Figure 30b) and without it (Figure 30а). Hanging SW Bow SW

model

V

channel

V Expansion wave

Compression wave Hanging SW

Compression wave

а

b

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Formation of hanging waves: flow, (а) external flow, flow. Figure 29. FormationFig. of29. hanging shock waves: shock (а) external (b) channel (b) channel flow.

Figure 30. Generation of the SB wave by a supersonic aircraft (with the nose part shaped as a modified power-law body): (а) without coolant injection, (b) with coolant injection; 1 - aircraft, 2-7 – pressure jumps caused by the bow SW (2), SW induced by the wing (3), tail SW (4), hanging SW (5), bow SW + hanging SW (6), and bow SW + hanging SW + SW from the wing (7).

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

33

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The results are presented in the form of excess static pressure distributions in the flight direction at different distances from the aircraft. The numbers indicate the pressure jumps due to shock waves generated by the basic aircraft elements. The downstream shift of the intermediate hanging SW 5 (Figure 30b, H1) and the increase in its velocity (owing to the increase in the SW intensity) with respect to the SW generated by the wing 3, provided by coolant injection, allow the SB middle zone effect to be maintained down to the ground level. As a result, instead of the bow SW generated by the aircraft without the cryogenic forcing system (Figure 30а, H4), the SB wave front is a system of spatially separated shock waves generated by the bluntness of the nose part of the fuselage 2, intermediate hanging SW 5, and SW from the wing 3 (Figure 30b, H4). The intensity of the bow SW 2 (from the blunted nose part) is expected to be 40% smaller than that of the initial modified body. Concerning the bow SW generated by the configuration without the coolant injection system (Figure 30a, H4), with identical contributions of the fuselage and the wing to its intensity, the reduction of the bow SW intensity for the configuration with coolant injection is 75%. Substantially smaller pressure differences on these shock waves, as compared with the bow SW generated by the initial configuration, and consecutive arrivals of these waves allow significant reduction of the SB intensity. The necessity of controlling the location of the intermediate shock arises when the SB level is reduced by upstream injection of the opposing air jet from the nose part of a slender body. Regimes of the flow at M  2 providing reduction of the drag force and the SB level  were determined in [21, 28]. It was shown that regimes of injection of the air jet from a slender body in the upstream direction in a supersonic flow, which involve reduction of the body drag, ensure reduction of the bow SW intensity owing to the effect of bluntness formed by the contact surface of the injected air jet only within the middle zone. The small height of this zone restricts applications of this method in practice. As in the case of a blunted solid body, this is caused by the restriction of the bluntness influence region with the intermediate hanging SW formed as a result of interaction of the incoming flow with the jet flow in the region of its attachment to the surface. Figure 31 shows the schlieren pictures of the flow around an ogive body with upstream injection of a local air jet in a supersonic flow. The regime of developed interaction (Figure 31а,b) with exhaustion from the nozzle with the Mach numbers at the nozzle exit Mn = 1 and 3.8, and the regime of penetration (Figure 31c) generated by the nozzle with Mn = 3.8 are illustrated. A mechanism of controlling the parameters of the hanging SW is needed to increase the middle zone length in jet injection regimes that ensure an acceptable drag force. In view of the analogy of the flow around a body with an air jet (Figure 32) and around a solid body with spherical bluntness, we can assume that cryogenic forcing implemented on a modified powerlaw body can shift the region of hanging SW formation in the downstream direction with the minimum energy expenses and, therefore, increase the middle zone length to distances greater than the cruising flight altitude. The level of the sonic boom generated by an aircraft equipped with an active control system will be substantially reduced, as compared with the initial configuration, with the minimum energy expenses. It should be noted that the formation of the hanging SW was also observed near the body surface in investigations aimed at drag reduction by means of upstream injection of a low-temperature plasma jet from the body into the incoming

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

34

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al.

supersonic flow [44]. The increase in static pressure behind the hanging SW creates an additional drag force, which can be appreciably reduced by organizing distributed injection of the coolant to the region of hanging SW formation. Thus, the efficiency of the method of drag reduction will be increased, and the thermal protection of the aircraft surface will be ensured.

Hanging SW

а

b

c

Fig. 31. Schlieren pictures of the flow around an ogive model with upstream

Figure 31. Generation of the SB wave by a supersonic aircraft (with the nose part shaped as a modified of the local (M  = 2.03). power-law body): (а) without injection coolant injection, (b) jet with coolant injection; 1 - aircraft, 2-7 – pressure jumps caused by the bow SW (2), SW induced by the wing (3), tail SW (4), hanging SW (5), bow SW Pj  Pj / P0  21.1, (а) M 1, dSW / d m = 0.1,  dfrom  1.48, + hanging SW (6), andn =bow + nhanging SWd+b SW wing (7). n =d b / d nthe

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

(b,c) M n = 3.8; d n 10.048, 2d b 33.08: b -4Pj  80 , c - Pj 5 22 .9 .

V

6

7

Fig. 32. around a slender body withinjection upstream injection an air jet into the supersonic Figure 32.Flow Flow around a slender body with upstream of an air jet into theofsupersonic incoming flow (regime of developed interaction). 1- bow SW, 2- contact surface, 3- expansion wave, 4incoming flow (regime of developed interaction). 1- bow SW, 2- contact surface, 3- expansion jet layer, 5- hanging SW, 6- closing SW, 7- stagnant zone. wave ing SW, 7- stagnant zone. головная УВ, 2- контактная поверхность, 3- волна разрежения, 4- струйный слой, 5- висячая

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

35

REFERENCES [1] [2] [3]

[4] [5] [6]

[7]

[8] [9] [10]

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

Mazin, IP. Physical Grounds for Aircraft Icing [in Russian], Gidrometeoizdat, Moscow, 1957. Trunov, OK. Aircraft Icing and Methods of Fighting [in Russian], Mashinostroenie, Moscow, 1965, 247. Tenishev, RKh; Stroganov, BA; Savin, VS; Kordinov, VK; Teslenko, AI; Leontiev, VN. Anti-Icing Systems of Flying Vehicles [in Russian], Mashinostroenie, Moscow, 1967, 320. Lynch, FT; Khodadost, A. Effects of ice accretions on aircraft aerodynamics, Progress in Aerospace Sciences, 2001, 39, 323-362. Bragg, MB; Broeren, AP; Blumenthall, LA. Iced-airfoil aerodynamics, Progress in Aerospace Sciences, 2005, 41, 323-362. Gurylev, VG; Mametiev, Yu. A. Effect of centerbody cooling on inlet starting, low stalling at the entrance, and throttling characteristics of inlets at supersonic and hypersonic velocities, Uch. Zap. TsAGI, VI, No. 1975, 2, 139-146. Grin, VT; Zakharov, NN. Experimental study of tangential injection and cooling of the wall on the flow with separation, Izv. Akad. Nauk SSSR, Mekh. Zhidk. Gaza, No. 6, 1971. Ryder, MO. Turbulent boundary layer, skin friction heat transfer and pressure measurements on hypersonic inlet. Compression surfaces, AFFDL-TR-68-102, 1968. Vasiliev, IYu; Grin, VT; Zakharov, NN. Control of the boundary layer in hypersonic inlets, Proc. III All-Union Conf. on Applied Aerodynamics, Kiev, 1973. Neiland, VYa. Specific features of the boundary layer separation on the central body and its interaction with a hypersonic flow, Izv. Akad. Nauk SSSR, Mekh. Zhidk. Gaza, No. 6, 1973. Lysenko, VI; Maslov, AA. Тhe effect of cooling on supersonic boundary-layer stability, J. Fluid Mech, 1983, Vol. 147, 39-52. Lysenko, VI; Vaslov, AA; Trantition reversal and one of its causes, AIAA J, 1981, Vol. 19, No. 6, 705-708. Lysenko, VI; Maslov, AA. Effect of deep cooling on the transition in a supersonic boundary layer, Izv. Akad. Nauk SSSR, Mekh. Zhidk. Gaza, 1981, No. 2, 43-49. Lysenko, VI; Maslov, AA. Transition of the laminar boundary layer to a turbulent state due to surface cooling, J. Appl. Mech. Tech. Phys., No. 3, 1981, 30-36. Lysenko, VI; Maslov, AA. Effect of cooling on stability of a supersonic boundary layer, Izv. Akad. Nauk SSSR, Mekh. Zhidk. Gaza, 1982, Vol, 264, No. 6, 1318-1321. Lysenko, VI. Stability and Transition of High-Velocity Boundary Layer and Wake Flows [in Russian], Novosibirsk, 2006, 288. Parker, MA. The sonic boom problem, Aircraft Engineering, Vol. 40, No. 8, 30-38. Vasiliev, LE; Popov, SI; Svishchev , GR. Aviation XXI. Predictions and prospects, Tekhn. Vozd. Flota, 1994, Vol. 68, Nos. 1-2, 14 - 17. Zhilin, YuL. Sonic boom generated by a supersonic passenger airplane, Trudy TsAGI, 1983, No. 1489, 41-45. Miles, RB; Martinelli, L; Macheret, SO; Shneider, MN; Girgis, IG; Zaidi, SH; Mansfield, DK. Supersonic of sonic boom by dynamic off - body energy addition and

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

36

[21]

[22]

[23] [24] [25]

[26]

[27]

[28]

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[29]

[30]

[31]

[32]

[33] [34] [35]

V.M. Fomin, V.F. Chirkashenko, A.M. Kharitonov et al. shape optimization, AAIA-2002-0150, 2002, 13. Chirkashenko, VF; Yudintsev, Yu.N. Effect of upstream injection of a jet from the body into a supersonic flow on the parameters of the sonic boom generated by the body and on the drag force, Abstracts of IV Intern. Conf on Nonequilibrium Processes in Nozzles and Jets, IXI Intern. Workshop on Jet, Separated, and Unsteady Flows, St. Petersburg, Russia, 2002, 431 - 432. Garanin, AF; Tretyakov, PK; Chirkashenko, VF; Yudintsev, Yu.N. Control of shock wave parameters by mass and energy addition, Izv. Ross Akad. Nauk, Mekh. Zhidk. Gaza, 2001, No. 5, 186-193. Miles, RB; Macheret, SO; Shneider, MN; Raizer, YuP; Girgis, IG; Zaidi, SH. Steady and unsteady supersonic flow control with energy addition, AAIA, 2003, 2003-3862. Georgievskii, P.Yu; Levin, VA. Supersonic flow with external sources of heat, Pisma Zh. Tekh. Fiz, 1988, Vol. 14, No. 8, 684-687. Tretyakov, PK; Garanin, AF; Grachev, GN; et al. Control of supersonic flow with the use of a powerful optical pulsed discharge, Dokl. Ross. Acad. Nauk, 1996, Vol. 351, No. 3, 339-340. Fomin, VM; Chirkashenko, VF; Volkov, VF. Control of the sonic boom generated by the flying vehicle and of the drag force by means of an active impact on the flow process, in: Achievements of mechanics of continuous media (collected scientific papers) [in Russian], 2009, Dalnauka, Vladivostok, 719-759. Aleksandrov, AF; Chuvashov SN; Timofeev, IB. Method for shock-free supersonic motion of a flying vehicle in the atmosphere and the flying vehicle, Patent RF No. 2107010, Bul. No. 8, Part II, 1998, 374. Chirkashenko, V.F; Yudintsev, Yu.N. Regimes of interaction of the opposing jet with the incoming supersonic flow, in: Gas dynamics and acoustics of jet flows (collected scientific papers) [in Russian], ITAM SB RAS, Novosibirsk, 1979, 75-116. Chirkashenko, VF; Yudintsev, Yu.N. Parameters of shock waves generated by bodies of revolution in a homogeneous atmosphere, Fluid Mechanics - Soviet Research, 1985, Vol. 14, No. 6, 31-39. Fomin, VM; Chirkashenko, VF; Volkov, VF. Numerical study of the influence of the aerodynamic configuration of a supersonic passenger aircraft on the parameters of its sonic boom, Vych. Tekhnol, Vol. 11, Part 2, Special issue, 2006, 64 74. Fomin, VM; Chirkashenko, VF; Volkov, VF; Kharitonov, AM. Controlling the level of the sonic boom generated by a flying vehicle by means of cryogenic forcing. Part 1. Cooling of the vehicle surface, J. Appl. Mech. Tech. Phys., 2008,Vol. 49, No. 6, 962970. Fomin, VM; Chirkashenko, VF; Volkov, VF; Kharitonov, AM. Controlling the level of the sonic boom generated by a flying vehicle by means of cryogenic forcing. Part 2. Distributed injection of a supercooled gas from the vehicle surface, J. Appl. Mech. Tech. Phys., 2009, Vol. 50, No. 2, 284-290. Chirkashenko, VF; Fomin, VM; Kharitonov, AM; Volkov, VF. Method of sonic boom reduction, Patent RF No. 2356796, 2009. Wlezien, R; Veitch, L. Quiet supersonic platform program, AIAA Paper, 2002, 20020143, 17. Whitham, GB. The flow pattern of a supersonic projectile,‖ Comm. Pure Appl. Math, 1952, Vol. 5, No. 3, 301-338.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Control of the Sonic Boom Generated by a Flying Vehicle…

37

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[36] Fomin, VM; Chirkashnko, VF; Volkov, VF. Sonic-boom problem and possible ways to solution, Proc. Intern. Conf. on the Methods of Aerophysical Research, Russia, Novosibirsk 30 June - 6 July, 2008, Abstracts Part, I, 38-39. [37] Fomin, VM; Kharitonov, AM; Chirkashenko, VF; Volkov, VF. Control of the sonic boom level by means of cryogenic forcing on the flow around the flying vehicle,‖ Preprint ITAM SB RAS, No. 5, Novosibirsk, 2007, 40. [38] Blagosklonov, VI; Vasilchenko, VI; Grozdovskii, GL. Aeromechanics of Supersonic Flow around Power-Law Bodies of Revolution and Others [in Russian], Mashinostroenie, Moscow, 1975, 183. [39] Chirkashenko, VF; Yudintsev, Yu.N. Development of the method for measuring the sonic boom parameters in supersonic wind tunnels, Preprint ITAM SB RAS, No. 6, Novosibirsk, 1983, 41. [40] Chirkashenko, VF; Yudintsev, Yu. N. System for computer-aided measuring the sonic boom parameters in wind tunnels, Preprint ITAM SB RAS, No. 21, Novosibirsk, 1983, 25. [41] Kharitonov, AM. Supersonic wind tunnel T-313, in: Techniques and Methods of Aerophysical Experiment, part 1, Novosibirsk State technical University, Novosibirsk, 2005, 217. [42] Ryzhov, OS. Decay of shock waves in steady flows, J. Appl. Mech. Tech. Phys., No. 6, 1961, 36-43. [43] Malkov, MP; Danilov, IB; Zeldovich, AG; Fradkov, AB. Handbook on Physical and Engineering Basis of Deep Cooling [in Russian], Gosenergoizdat, Moscow, 1978, 416. [44] Fomin, VM; Malmuth, N; Maslov, AA; et al. Effect of the opposing plasma jet on the total and distributed aerodynamic characteristics of a blunted body, Dokl. Ross. Akad. Nauk, 1999, Vol. 368, No. 2, 197-200.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

In: Cryogenics: Theory, Processes… Editor: Allyson E. Hayes, pp. 39-68

ISBN: 978-1-61761-323-4 © 2011 Nova Science Publishers, Inc.

Chapter 2

LIQUID OXYGEN MAGNETOHYDRODYNAMICS J. C. Boulware, H. Ban, S. Jensen and S. Wassom Mechanical and Aerospace Engineering Dept., Utah State University, Logan, Utah, USA

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

1. ABSTRACT In the cryogenic realm, liquid oxygen (LOX) possesses a natural paramagnetic susceptibility and does not require a colloidal suspension of particles for practical application as a magnetic working fluid. Commercial ferrofluids have performed well in industrial applications, but expanding their workable range to low temperatures requires a suitable selection of the carrier fluid, such as LOX. In this chapter, the equation of motion for the pure fluid is derived and applied to a slug of LOX being displaced by a pulsed magnetic field. Its theoretical performance is compared to actual experimental data with discussion on empirical parameters, sensitivity to measurement uncertainty, and geometric similarity. The 1.1 T pulse of magnetic flux density produced oscillations in the slug of 6-8 Hz, generating up to 1.4 kPa of pressure change in a closed section when the slug acted like a liquid piston. The experiments and theoretical model demonstrate that LOX could be used as a magnetic working fluid in certain applications.

2. INTRODUCTION Elimination of moving parts and increased subsystem lifetime is a major benefit of actuator systems with magnetically responsive fluids as opposed to those relying on a mechanical driver to instigate flow. Space systems, in particular, could benefit from increased subsystem lifetime as it would increase the overall mission length; however, unlike groundbased magnetic fluid systems, use of magnetically responsive fluids in the low-temperature regime of space requires verification of fundamental principles through basic research and experimentation, since it has never been applied.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

40

J. C. Boulware, H. Ban, S. Jensen et al.

2.1. Magnetic Fluids Magnetism occurs due to the atomic or molecular structure of a material and can be classified as ferromagnetic, diamagnetic, or paramagnetic depending on the behavior of the poles. Ferromagnetic solids have permanently aligned poles and generate their own magnetic fields. Liquids, however, cannot maintain the alignment without a field and are either paramagnetic, in which the poles align with the applied field, or diamagnetic, in which the poles align opposite the applied field. The bulk effect of each is that paramagnetics are attracted to the field (towards an increasing gradient), and that diamagnetics are repelled by it (away from an increasing gradient)[1]. In the 1960s, NASA developed ―ferrofluids,‖ which are a colloidal suspension of ferromagnetic particles in a carrier fluid. A surfactant on the particles prevents their alignment without a field; thus, ferrofluids actually exhibit superparamagnetism since they have an extremely high susceptibility to an applied field. Ferrofluids have found many industrial applications, such as in high-end audio speakers, digital data storage, and resonance imaging. As a working fluid, ferrofluids have been proposed for pumps[2-7], valves[8], actuators[9], heat pipes[10-11], and even optical tuners[12]. The range of applicability of ferrofluids, however, is limited by the thermal characteristics of the carrier fluid, typically water, oil, or a hydrocarbon. While much use has been made of ferrofluids at ambient and high temperatures, freezing of the carrier fluid prevents their use at low temperature. Furthermore, the presence of nanoparticles and surfactants in ferrofluids complicates analyses, mainly due to agglomeration and nonhomogeneity. In the cryogenic realm, liquid oxygen (LOX) presents a potential solution as it possesses a natural paramagnetic susceptibility and does not require particles for practical application.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

2.2. Liquid Oxygen In all phases, the unpaired electrons in an O2 molecule lead to a bulk paramagnetic effect. At room temperature, however, the thermal energy within the molecules may dominate the magnetic alignment with an applied field; hence warm oxygen does not have an appreciable susceptibility. As temperature decreases and thermal energy is reduced, the molecules are more able to align and susceptibility increases. This phenomenon is known as Curie‘s Law, where, essentially, paramagnetic susceptibility increases as temperature decreases. Furthermore, once oxygen condenses (90 K, 1 atm), the volumetric susceptibility, , significantly increases with the density of the fluid. The relationship between volumetric susceptibility, mass susceptibility, mass, and molar susceptibility, molar, is defined through density, , and molecular weight, MW, as

 

1



 mass 

1  molar  MW

.

(1)

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

41

Throughout the remainder of the chapter, ―susceptibility‖ will refer to volumetric susceptibility. Although it is approximately 30 times weaker than a low-end ferrofluid, LOX has the highest known paramagnetic susceptibility of pure fluids at about 0.0042. The lack of magnetic particles eliminates risks such as corrosion and shock, and since LOX is already commonly used for life support, thermal management, and propulsion systems, the integration process is simpler than for a ferrofluid.

2.3. Previous Research

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The basic properties of LOX have been measured under a variety of temperature and pressure ranges[13-15], but unfortunately, very few experiments have studied the influence of a magnetic field, perhaps due to the volatile nature of LOX. Surface tension[16], surface instabilities[17], and levitation phenomena[18, 19] have all been studied under high magnetic fields, but none of these experiments generated a bulk displacement of the liquid. Yerkes[20] measured the wicking heights of LOX heat pipes when augmented by a magnetic field and showed an increase of up to 4 times the capillary pressure for a magnetic flux density of 0.27 T. These experiments are useful in understanding the nature of LOX magnetohydrodynamics as well as experimental research on magnetic fluid pumps, magnetoviscosity, and magnetic fluid pipe flow.

2.3.1. Magnetic fluid pumps Regarding LOX, only one experimental study could be found which generated a high flow rate. Youngquist[21] of the Kennedy Space Center researched the dynamics of a column of LOX in a U-tube when a magnetic field was applied. He measured the displacement of one end of the column when the other was pulsed with a magnetic field induced by a solenoid. Figure 1 shows the experimental setup. An electric current of 30 A was pulsed through the solenoid, generating a magnetic field with a maximum flux density of 0.9 T. With the field applied, the height of the column oscillated about a new mean, reaching a maximum displacement of 4-5 cm. It is worthy of note that pulses of 100 A and 6 T were attempted, but yielded erratic results, often ripping off the top of the column. A theoretical model was created to obtain a one-dimensional, finitedifferenced solution, which employed a second-order, velocity-based damping function relying upon empirical coefficients. Again, this study was the only experimental research found on the magnetohydrodynamics of LOX, but other ferrofluid pumps served well as bases for comparison. Park and Seo[2-4] of Pusan National University have developed a magnetic fluid linear pump for use in infusion pumps and artificial hearts in the medical industry. Employing magnetic yokes to propagate droplets of a magnetic fluid, the device uses surface shear to pump water as shown in Figure 2. Park and Seo report pumping heights equivalent to 2 kPa (0.29 psi) for a maximum flux density of 0.036 T (360 G). While this seems like an extremely small field compared to Youngquist‘s experiment, it is important to note the Park and Seo are using a ferrofluid and not LOX. The research performed by Park and Seo is useful as a study on traveling waves and their effects on the surface dynamics of a magnetic fluid droplet, but difficult to apply to LOX

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

42

J. C. Boulware, H. Ban, S. Jensen et al.

due to the differences in susceptibility and surface tension. Nonetheless, the work serves as a good benchmark for comparison.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Experimental setup for LOX pumping, taken from Youngquist.

Figure 2. Experimental setup for pumping water with a magnetic fluid, taken from Park and Seo.

Hatch[5] of the University of Washington developed a ferrofluidic rotary micropump to enhance lab-on-a-chip MEMS technology. The concept (shown in Figure 3) achieved 1.2 kPa of pressure head using a rotating and stationary permanent magnet with a surface flux density of 0.35 T (3500 G). As in the experimental arrangement of Park and Seo, the device pumps a separate, immiscible fluid, but uses normal pressure instead of shear. The study reports operation at 4 and 8 rpm for 3 days at a time. It was found that the steady-state pressure gradient decreased over time when the plugs were rotated both clockwise and counterclockwise. Pumping speeds greater than 8 rpm generated too much pressure and disrupted the coupling between the permanent magnet and the translating ferrofluidic plug. While Hatch‘s design is intuitive and effective, the rotating permanent magnet is a mechanically moving component and, therefore, negates the goal of creating a system for fluid actuation with no moving parts.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

43

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 3. A rotating permanent magnet to propagate a ferrofluid plug, taken from Hatch.

Figure 4. Two-coil system for pumping a ferrofluid by magnetic surface stress, taken from Krauss.

Moghadam[6] also developed a microscale magnetic fluid pump and eliminated the moving parts. He used a series of solenoids spaced along a tube to drive a magnetic fluid linearly, similar to the method of Park and Seo. However, instead of wrapping the solenoids around the tube, they were offset and orthogonally aligned so that their core could be filled with an iron rod and increase the magnetic flux density. The setup produced 0.64 kPa of pressure head for flow rates of 1.1 cm3/min at 0.45 T. The study compared different working fluids and particles, but relied on the viscous drag of the particles to create fluid motion. Krauss[7] of the University of Bayreuth has used a two-coil system to pump a ferrofluid circularly. The 90° phase difference of the two coils with orthogonal axes produced a net field able to rotate the fluid through the magnetic stress on the fluid surface. The mean diameter of the duct was 100 mm and the system produced a maximum fluid velocity of 70 mm/sec and a magnetic field of 800A/m. Zahn and Greer[22] of the Massachusetts Institute of Technology took a theoretical approach to traveling waves, but without a free surface. They found that the magnetic fluid can actually be pumped backwards if the wave moves too fast. Without the free surface, the field interacts with the nanoparticles inside the ferrofluids, and motion is generated through

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

44

J. C. Boulware, H. Ban, S. Jensen et al.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

the particle spin. They studied the dynamics of a spatially steady field, which, however, varied sinusoidally in time. Their work was followed up by Mao and Koser[23] of Yale University who were able to vary the field in space as well. Their findings showed that a maximum flow velocity was achieved when the product of the applied magnetic field frequency, the wave number, and the height of the channel approached unity. In other words, pumping becomes more efficient as the magnetic field frequency approaches the reciprocal of the relaxation time constant of the magnetic particles in the fluid. Mao and Koser compared their experimental data with numerical results for a 2D solution using FEMLAB and a 1D solution using Matlab. They found that all 3 agreed well until the magnetic field frequency reached about 30 kHz, when the Matlab solution began to diverge. The aforementioned research illustrates the importance of fluctuating magnetic fields for pumping. Without a gradient of the magnetic field, no net force is generated, just as with a pressure gradient. However, as shown by Youngquist, stationary solenoids are still able to create a magnetic field gradient, since their strength lessens with distance. By pulsing the stationary solenoid, a time-varying gradient can also be induced and used for position control of the magnetic fluid.

2.3.2. Magnetoviscosity Viscosity is adherent to fluid motion and can be calculated through its stress and strain rates. The normal and tangential surface force on a differential element due to thermodynamic pressure can be found through a divergence of the stress tensor; likewise, magnetic force can be found through the divergence of the Maxwell stress tensor, but its associated viscosity is much more complicated. Molecular or microscale magnetic particles in a paramagnetic fluid align with the applied field and can induce additional shear as a function of the strength of the field. When aligned, the magnetic torque helps the particles resist rotation, thereby disrupting fluid flow. The magnetoviscous effect is heavily studied in ferrofluids, but questions remain in the case of a pure, paramagnetic fluid like LOX. For the purpose of the current research, LOX is considered as a ferrofluid with angstrom-scale particles, a fill fraction, , of 100%, and a carrier fluid with the same viscosity as non-magnetized LOX. From equations given by Shliomis[24], the full fill fraction approximation leads to a vortex viscosity, , of 1.5 times the non-magnetized shear viscosity, , but the small diameter and viscosity lead to a nearly infinitesimal Brownian relaxation time,  The ratio ultimately leads to a very small increase in the effective viscosity from particle alignment, .

3 2

  

(2)

  3V / kT

(3)

 0 MH  / 4     1   0 MH  / 4 

(4)

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

45

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

where V is the particle volume, k is the Boltzmann constant, T is temperature,  is the permeability of free space, M is the magnetization, and H is the applied field. These equations, however, were written for dilute solutions and may not be applicable for high concentrations. Furthermore, experiments by McTague[25] have shown that particle interactions may also affect the overall viscosity, even in dilute solutions, and that the atomic interactions between particles under a magnetic field generate an increase in the viscosity. The equations above assume particles up to 10,000 times larger than a molecule of oxygen; thus, different forces may be at play. Without an adequate theory, the magnetoviscous effect of LOX cannot be declared either insignificant or significant until a physical experiment measures it. Lastly, use of a high-frequency AC field in the magnetic field may actually induce a ―negative viscosity,‖ as shown by Bacri[26]. As mentioned, a static or low-frequency field will retard flow through particle alignment with the field. In the case of high frequencies, increased fluid motion was observed, indicating a reduction in the viscosity. This effect may be desirable or undesirable depending on the intended application.

2.3.3. Magnetic fluid pipe flow Aside from influencing the rotational viscosity and particle interactions, the field can have a macroscopic effect on the flow of a magnetic fluid through a pipe. Cunha[27] numerically studied the laminar flow through a pipe within a magnetic field with a linear gradient. When the field gradient was opposite the flow direction, the fluid was impeded as expected; however, Cunha noted the drag reduction as the field gradient facilitated fluid flow. He characterized the flow by a magnetic pressure coefficient, Cpm, representing the ratio of the magnetic to hydrodynamic pressures in the flow. Cunha correlated his results to a nonmagnetic friction factor relationship of f = 8/Re and found that, as the magnetic effects arise, f is reduced. The reduction is more pronounced for higher Reynolds numbers, but the study is limited to an asymptotic value near Re = 50. Nonetheless, Cunha showed that as an axial field in the direction of fluid flow increases, drag on the walls decreases. An axial field with a linear gradient is not simple to reproduce in a laboratory experiment; instead, Chen[28] applied a ring magnet and focused on the streamlines for magnetic fluid flow in a tube as the field and magnetoviscous response varied. He characterized the system parameters through a magnetic Reynolds number, Rem, and a viscosity parameter, R. Figure 5 shows the difference in the streamlines as the system parameters vary. The field was applied by a ring magnet at z = 0, but of undisclosed length or gradient. Study of the axial velocity profiles at various locations shows that an adverse gradient occurs even without a magnetoviscous influence. This indicates that even if the viscosity of LOX does not increase with a magnetic field, fluid damping still increases due to flow circulation. Schlichting[29] gives the classical solution of oscillating flow through a pipe, but the presence of a magnetic field and the finite slug length complicate the analysis for the current study. In the case of an infinite slug without a magnetic field, the shear could be doubled during oscillations; it is expected that the augmentation would be greater with the magnetic field and finite slug.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

46

J. C. Boulware, H. Ban, S. Jensen et al.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Streamline patterns for magnetic fluid flow in a tube with (a) Rem = 103 and R = 0 (no viscosity variance with field); (b) Rem = 103 and R = 1; (c) Rem = 105 and R = 1; and (d) Rem = 1.225 * 106 and R = 3.5, taken from Chen.

2.4. Current Test Parameters To determine the viability of LOX as a working fluid in a magnetically driven actuator, the authors conducted studies that were essentially an evolution of Youngquist‘s experiments, but with test objectives focused on advancing the technology towards applied research instead of basic; thus, the experimental principle was different as well. The broad goal of the research was to support the notion that LOX could be used as a working fluid in a magnetic fluid system due to its significant paramagnetic susceptibility. This goal was to be achieved by performing controlled, quantitative experiments, correlating them to a theoretical model, and determining predictable trends from the results. The theoretical model was to limit empirical input (other than initial conditions) and be able to make predictions regardless of system geometry. Most importantly, the final data should be useful to future, applied research. For this purpose, the experiments used a slug of LOX rather than a long column, as in Youngquist‘s experiments. Magnetic pressure on a slug is maximized when one edge is in the center of the solenoid and the other is in a negligible field. While this is achievable with a long column, a smaller slug is nearly as effective. Figure 6

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

47

shows that the magnetic flux density of a 0.6 cm (0.25 in) solenoid drops to 5% of its maximum value of 1.1 T at a distance 1.75 cm from the center of the solenoid (Youngquist‘s column of LOX totaled 36 cm). This benchmark differs depending on solenoid geometry, applied current, and wire spacing, but the example shows that a slug can achieve nearly the same magnetic pressure for a much smaller mass and length. Eliminating mass and length reduces inertia and shear, which would otherwise retard slug motion. The experiment was designed to displace a slug of LOX and correlate its dynamics to a numerical model. In all, the research studied the viability of LOX in a magnetic fluid system with the following objectives:

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

1. Displace a LOX slug using magnetic fields. Experimentally accomplishing this would quantify the potential of a LOX-based magnetic fluid system. 2. Detect the displacement through pressure. Innovative measuring techniques will be required to study how LOX behaves in a magnetic field.

Figure 6. Magnetic flux density along the axis of a solenoid.

3. Simulate the dynamics numerically. A verified numerical algorithm can quantify LOX performance outside the scope of laboratory testing. 4. Perform parametric studies to examine efficiency optimization methods. Information on the sensitivity to uncertainties and geometric variance will help to estimate the potential capability of an optimized system.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

48

J. C. Boulware, H. Ban, S. Jensen et al.

3. THEORETICAL MODEL The theoretical model was based on a simple force balance on the slug. The change in momentum of the slug was a result of the net force from magnetism, pressure, and damping. The equation of motion essentially becomes the Navier-Stokes equations with an additional term for the magnetic force. Rosensweig[1] provides a thorough description of the force due to magnetism, also known as the Kelvin force. The Kelvin force density, fm, can be found through the divergence of the Maxwell stress tensor as a function of the permeability of free space, o, the magnetization vector, M, and magnetic field, H, as

f m   0 M   H .

(5)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The resultant magnetization from an applied field can be described by the Langevin function, where the volumetric paramagnetic susceptibility  is the ratio of the magnetization vector to the applied field vector,  = M / H. By substituting for M, using the vector identity, H·H = (H·H)/2 - H(H), and noting that Ampere‘s Law cancels out the curl of the applied field, Eq. (5) can be derived to

f m   0 H   H ,

(6)

f m   0  H  H  ,

(7)

f m   0  ( H  H)/ 2 - H  (  H)  ,

(8)

f m   0  ( H  H)/ 2  ,

(9)

f m   0  H 2 / 2 .

(10)

With a constant temperature, the relative permeability, , also remains constant. The relative permeability is the ratio between the magnetic flux density, B, and applied magnetic fields,  = B / H, which can also expressed in terms of volumetric susceptibility,  = o(1+). Given these relations, the Kelvin force density is derived as such: 2

B f m   0   / 2 ,  fm 

0  B 2 , 2 2

(11)

(12)

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

0 

fm 

B 2 ,

(13)

 B 2 , 2 20 ( 1   )

(14)

20 ( 1   ) 2

2

1

fm 

49

and the force in the axial direction is

f m ,x 

 d 2 Bx , 2 2  0 ( 1   ) dx 1

(15)

where the subscript x denotes the axial direction. The differential term considers the ends of the slug, and when Eq. (15) is integrated over the entire volume with a one-dimensional approximation, the force due to magnetism in the axial direction, FM, is

FM  f m ,x a 2 L ,

(16)

B x  FM  a 2 L , 2 2  0 ( 1   ) x

(17)

B x ,US  B x ,DS  FM  a 2 L , 2 20 ( 1   ) L

(18)

2

1

2

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

1

FM 

2

a 2  2 2 ( B x ,US  B x ,DS ) , 2 20 ( 1   )

(19)

where a is the tube radius, L is the slug length, and the subscripts US and DS denote the upstream and downstream directions. The magnetic flux density generated by the solenoid is found by summing the contribution of each loop. The magnetic flux density from an individual loop of wire is derived from the Biot-Savart Law as,

B x ,loop 



I ( t ) 0 rloop

2 rloop  dx 2 2

2



3 / 2

.

(20)

Where r is the radius of a single loop of coil, I(t) is the applied current over time, the subscript loop denotes a single loop of the coil, and dx is the axial distance from that loop. The oscillatory motion, finite slug length, and unknown magnetoviscous effects complicated the damping force on the one-dimensional analysis. These effects could be

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

.

50

J. C. Boulware, H. Ban, S. Jensen et al.

treated as having a combined effect on the wall shear stress through an empirically found damping factor, . The classic relation for laminar wall shear stress in Hagen-Poiseuille flow as given by White[30] and the force due to damping, FD, was calculated as

 w  4 x / a ,

(21)

FD  2a( L  Lhidden ) w ,

(22)

where w is the wall shear stress, x is the velocity of the slug in the axial direction, is the nonmagnetized dynamic viscosity of LOX, L is the visible length of the slug, and Lhidden is the hidden length of the slug in the steel sections. During filling, portions of LOX remained in the plumbing and could not be directly measured, but could be calculated through the frequency of the oscillations. The cause of the hidden slugs and calculation of their length will be discussed in a later section. The pressure force, FP, resulted from the differential pressure on either side of the slug as

FP  a 2 p ,

(23)

where p denotes the pressure differential across the slug. The change in pressure resulted from the compression and expansion of closed volumes on either side of the slug. Thus, with the forces due to pressure, magnetism, and damping, the equation of motion for the LOX slug becomes

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

mx  a 2 p 

a 2  2 2 ( B x ,US  B x ,DS )  2a w ( L  Lhidden ) , 2 20 ( 1   )

(24)

where m is the mass of the visible and hidden slugs and x is the acceleration. This onedimensional force balance assumes that the finite length slug was an incompressible solid and does not account for surface tension, cohesion, instabilities, or breakdown of the slug[31]. Bashtovoi[32] points out that capillary effects are reduced under the influence of a magnetic field and are thus considered negligible during the pulse; however, they must be significant enough to hold the slug in place when nonmagnetized. Gravity was also ignored, because the tube was oriented horizontally. The relationship between the initial magnetic pressure on the slug and its maximum displacement must be nondimensionalized to compare different geometries. The maximum displacement can be nondimensionalized using the cross-sectional area of the tube and the downstream volume as

x* 

dxmax a 2 Vol DS

,

(25)

where dxmax is the maximum displacement of the slug and VolDS is the downstream volume. The initial magnetic pressure on the slug, pm,i, is defined as Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics US

p m ,i   0  M i dH  DS





 2 2 BUS ,i  B DS ,i , 2 20 ( 1   ) 1

51

(26)

where the subscript i represents the initial value before the pulse. Because the initial magnetic pressure is a function of the magnetic flux density at each of the edges, it can be found as a function of the initial slug position. To nondimensionalize the initial magnetic pressure, the Alfven velocity, ua, could be used as,

ua 

Bmax

,

0 

(27)

so that the resulting nondimensional initial magnetic pressure is,

pm  *

p m ,i .5  u a

2

,

(28)

where Bmax is the maximum magnetic flux density and  is the density of LOX. It is also useful to define an average initial velocity, ui, with the maximum displacement and the length of time required to reach that maximum displacement, dt, which occurs during the first oscillation. The average initial velocity is thus,

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

ui 

dx max dt

.

(29)

Then, the Mason number represents the ratio of damping to magnetic forces and is defined for the current study as

Ma 

FD FM



8 ( L  Lhidden ) u i

a 2 p m ,i

.

(30)

4. EXPERIMENTAL APPARATUS The experiments were conducted on a small slug of LOX inside a circular tube, and measurements were made that described the slug dynamics in a variety of test conditions. The tube was oriented horizontally to mitigate the dominance of gravity, and was, therefore, small enough so that the capillary forces allowed slug formation without inhibiting motion. Because LOX is extremely volatile, helium was used as the surrounding gas, since it does not react with oxygen. Also, with a melting point of 4 K at 1 atm, helium could be treated as an ideal gas at the test conditions. Since the test section was part of a closed volume, the slug displacement could be measured through pressure changes on either side of the slug as long as it did not break down.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

52

J. C. Boulware, H. Ban, S. Jensen et al.

Figure 7. Experimental principle of measuring slug displacement through pressure changes.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 8. Photograph and CAD drawing of experimental apparatus, from Boulware[33].

The slug dynamics were sensitive to the following parameters: slug length initial position solenoid geometry applied current system volume tube radius initial system pressure. Experimentally, it was not feasible to vary the tube radius since it affected the capillarity of the slug. Even marginal changes could significantly affect the dominance that surface tension would play; thus, tube radius remained constant throughout the experiments. Likewise, the volatility of LOX precludes high-pressure testing; thus, the initial system pressure remained as close to atmospheric as possible throughout the experiments. The closed volume was placed in a liquid nitrogen bath to prevent LOX boil-off, and test conditions and fluid properties were calculated at 77 K and 1 atm.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Liquid Oxygen Magnetohydrodynamics

53

A photograph and CAD drawing of the experimental system can be seen in Figure 8. The closed volumes on either side of the test section were dubbed the ―upstream‖ and ―downstream‖ sides where the upstream side was the larger volume including the condenser, and the downstream side was the smaller volume. Different geometric configurations required different system volumes. Figure 8 shows a small downstream section of 1.8 cm3 and a small solenoid of 30 gauge wire. Another configuration used a downstream volume of 5.9 cm3, but maintained the upstream volume constant at 337 cm3. Because the downstream volume was significantly less than the upstream volume in both cases, the data from the downstream pressure sensor was used for comparison. The operating pressure was maintained between 100-135 kPa for safety, and the runtime was limited to 0.25 seconds to reduce resistance heating in the solenoids. Before the liquid slug could be precisely positioned in the quartz tube, gaseous oxygen had to be introduced to the system at room temperature. Once the system was closed and submerged in liquid nitrogen, the gas condensed into LOX droplets and fell from the heat exchanger into the horizontal section of the plumbing. From there, a magnetic wand was used to drag portions of the LOX into the transparent quartz tube. While the process allowed for precise measurement of the slug length within 0.8 mm, an unknown amount of LOX remained in the steel sections. The mass of LOX that could not be seen was dubbed the hidden slug length but could be precisely calculated through the frequency of the pressure oscillations as will be shown later. The quartz tube had an inner diameter of 1.9 mm, and the solenoids were powered by a Hewlett-Packard 6268B 900 W DC power supply. The power supply had an upper limit of 30 V or 30 A; therefore, an optimization process for the solenoid sizing could be developed. To maximize the capability of the power supply, a resistance of 1  was desired when the solenoid was in the liquid nitrogen tub; thus, with a known coefficient of temperature resistance of 0.0039 for copper, the corresponding resistance when at room temperature was 6.34 . With a wire of known gauge, the total wire length could be found, and then an iterative scheme using Matlab and Excel could be used to determine the length and outer diameter of the solenoid that produced the highest magnetic field for a constant voltage source. The optimal slug length for a particular solenoid was determined as the length that generated the highest pressure change while accounting for forces due to magnetism, pressure, and damping. The theoretical model was used to create a numerical solution to find this length. Kulite CT-375 analog pressure sensors located upstream and downstream of the slug in the test section were sampled at 5 kHz using a Measurement Computing PCIM-DAS1602/16 A/D card driven by Matlab with Simulink and xPC Target with a combined uncertainty of 0.17 kPa from the effects of nonlinearity, hysteresis, 16-bit analog-to-digital conversion errors, and repeatability. Because the changes in the upstream and downstream pressures were the desired output, the absolute pressure and the measurement uncertainty were not influencing factors. The noise in the raw data was reduced by a Chebyshev Type II lowpass filter, set to 0 db at 45 Hz and -40 db at 50 Hz. The LOX slug formed a concave meniscus with edges measureable within 0.8 mm resolution via notches on the quartz tube.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

54

J. C. Boulware, H. Ban, S. Jensen et al.

5. NUMERICAL SOLUTION To apply the theoretical model to the experiment, a numerical simulation was written in Matlab v7.6.0 (R2008a) on a 2.4GHz Pentium 4 processor with 2GB of RAM. The pulse dynamics were typically solved in less than 2 seconds, allowing for a thorough optimization of system variables through a regression analysis. Fluid properties for LOX were taken from the CRC Handbook of Chemistry and Physics[15], and studies by Hilton[14] indicated that the pressure fluctuations would not significantly affect those values. Experimental measurements and observations were used to determine certain boundary and initial conditions. The magnetic flux density is proportional to the applied current and depends on the temperature of the solenoid over time. Eqs. (31-34) calculated the solenoid temperature and current over time.

Ri  R0 ( 1   ( 293  Ti 1 )) ,

(31)

I i  V / Ri ,

(32)

Pi  I i Ri ,

(33)

2

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Ti  Ti 1 

Pi m sol c p

t ,

(34)

where R is the resistance of the solenoid,  is the coefficient of temperature resistance for copper, T is the temperature of the solenoid, P is the electrical power, the subscript i denotes the time step, and the subscript 0 represents the initial value. Values for the mass of the solenoid, msol, and its specific heat capacity, cp, were measured in the laboratory. Because the solenoids are only powered for 0.25 seconds, convective cooling was shown not to be prevalent, despite being in direct contact with the nitrogen. This is likely due to the nitrogen boil-off creating a vapor bubble around the solenoid. Using data from the previous time step, the position of the slug could then be found. Positive displacement and velocity were considered as in the upstream direction, and the center of the solenoid was considered the origin. Since the slug was initially at rest, the initial velocity and net force were assigned values of zero. The displacement of the upstream edge of the solenoid was found as,

x i  x i 1  x i 1  t 

FT ,i 1 2 a ( L  L hidden ) 2

t 2 ,

(35)

where FT is the total force from the previous time step. Then, the magnetic flux density and force due to magnetism could be found as,

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

B x ,US ,i 

B x ,DS ,i 

1

0 I i 2

0 I i 2

2

( M x 1 )



M r 1



1 n 0 m   ( M x 1 ) 2

1 2

( M x 1 )

M r 1



1





n 0



3/ 2

  , (36) 

 a sol  2 nb r 2 , 2 2 3/ 2  a sol  2 nb r    xi  L  2 mb x   



m   ( M x 1 ) 2

FM ,i

 a sol  2 nb r 2  2 2  a sol  2 nb r    x i  2 mb x 

55

a 2  2 2  ( B x ,US ,i  B x ,DS ,i ) , 2 2 0 ( 1   )

(37)

(38)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

where Mx is the number of turns in the solenoid in the axial direction, Mr is the number of turns in the radial direction, asol is the radius of the solenoid, and b is the wire diameter. Using an isothermal, ideal gas assumption, the force due to pressure was found as,

VDS ,i  VDS ,i 1   xi  xi 1 a 2 ,

(39)

VUS ,i  VUS ,i 1   xi  xi 1 a 2 ,

(40)

PDS ,i  PDS ,i  1

V DS ,i  1

PUS ,i  PUS ,i  1

VUS ,i  1

,

(41)

,

(42)

FP ,i  a 2 ( PDS ,i  PUS ,i ) .

(43)

V DS ,i

VUS ,i

Finally, the force due to damping was calculated along with the Reynolds number and wall shear stress as,

Rei 

 w ,i 

2 x i 1 a

 16 xi  1 Rei

,

(44)

,

(45)

2

2

FD ,i  2a( L  Lhidden ) w ,i ,

(46)

where Re is the Reynolds number.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

56

J. C. Boulware, H. Ban, S. Jensen et al. Together, the pressure, magnetic, and damping forces formed the net force on the slug as

FT ,i  FP ,i  FM ,i  sgn( xi 1 )FD ,i ,

(47)

where damping was always opposite the direction of velocity. To begin the next time step, the velocity was calculated as x i 

FT ,i  FT ,i 1 2  a 2 ( L  Lhidden )

 t  x i 1 .

(48)

The small volume of the downstream volume amplified its pressure fluctuations, so it was used for comparison. The time step of the simulation was chosen as 0.0002, so that it correlated with the sampling frequency of the pressure sensor, and then the absolute residual, , between the experiment and simulation was calculated at each time step as

 i  Pexp,i  Psim ,i ,

(49)

where the subscripts exp and sim denote the experiment and simulation data. The simulation as a whole was characterized by the root mean square deviation (RMSD) of the residuals during the pulse period.  RMS 

1 N

N



2 i

,

(50)

i 0

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

where N is the number of data points during the pulse.

6. RESULTS AND DISCUSSION The authors completed the aforementioned experimental and numerical research and published the results in a series of articles[33-35]. Key findings from the studies are the optimization of the solenoid and slug combination, the maximum attainable pressure change in the downstream section, uncertainty determination from the numerical model, and geometric variance of the test parameters.

6.1. Solenoid / Slug Optimization Using the process described before, two solenoids were constructed of 24- and 30-gauge wire, as shown in Table 1. The actual fabrication of Solenoid B led to a resistance of about 1.3  in the liquid nitrogen; hence, only 23.4 A of current could be drawn from the power supply. The reduced current draw led to a lower total magnetic flux density, but its smaller size generated a higher

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

57

flux density gradient. The magnetic flux density along the axis affected the initial magnetic pressure for a slug based on its length and initial position. Figure 9 and Figure 10 show the maximum pressure change calculated by the numerical simulation for a variety of slug lengths and initial positions. Table 1. Optimized solenoid specifications Solenoid A 24 25 40 25.6 mm 6.3 mm 35.1 mm 128.4 g 1.17 T @ 30 A 89.7 T/m @ 30 A

30 22 22 6.8 mm 6.3 mm 21.3 mm 11.1 g 1.0 T @ 23.4 A 128.5 T/m @ 23.4 A

(a) 2

1.5

Max DP (kPa)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Wire Gauge: Radial Turns: Axial Turns: Length: Inner Diameter: Outer Diameter: Mass: Bmax: dB/dzmax:

Solenoid B

1

0.5

0

(b)

2

4

6

8

10

Slug Length (cm)

Figure 9. Isometric (a) and side view (b) of a surface map generated to find optimal slug for Solenoid A[35]. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

58

J. C. Boulware, H. Ban, S. Jensen et al.

A positive change in pressure represents a compression of the downstream section and a negative change represents an expansion. The dashed lines in Figure 9b and Figure 10b follow the pressure change for various slug lengths regardless of its initial position. The peak of the lines indicates the slug length that would induce the maximum pressure change for that solenoid. This value was termed the optimal slug length and was 2.7 cm for Solenoid A and 1.3 cm for Solenoid B.

(a) 2 1.8 1.6

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Max DP (kPa)

1.4 1.2 1 0.8 0.6 0.4 0.2 0

(b)

0

0.5

1

1.5

2

2.5

3

3.5

4

Slug Length (cm)

Figure 10. Isometric (a) and side view (b) of a surface map generated to find optimal slug for Solenoid B[35].

6.2. Maximum Attainable Pressure Change Figure 11 shows an experimental run versus the numerical simulation and was the basis of the analyses. A 0.25 sec pulse caused an oscillation of a 1.3 cm slug that had one edge initially in the center of the solenoid. The system pressure was about 131.2 kPa when power was switched to the solenoid at 0.01 sec. The downstream pressure fluctuated at approximately 7 Hz, generating a maximum pressure change of 1.2 kPa at about 0.06 sec after the pulse began. The regression analysis found that a hidden slug length of 14.5 cm and a

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

59

damping factor of 6.08 resulted in the lowest RMSD of the absolute error at 30.6 Pa. The solid line in Figure 11 shows the accuracy of the Matlab simulation. To perform the parametric studies mentioned in the third objective, several hundred runs similar to Figure 11 were conducted. Each oscillated at about 6-8 Hz with amplitudes relative to the initial conditions and applied current.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 11. Experimental and simulated pressure oscillations for a 1.3 cm slug[34].

Figure 12. Maximum pressure change for various slug lengths using Solenoid B.

In Figure 11, the maximum pressure change was approximately 1.2 kPa at about 0.07 sec. One edge of the 1.3 cm slug in the center of the solenoid correlates to an initial center displacement of 0.65 cm between the slug and solenoid. As the initial center displacement varied, a trend was apparent in the maximum pressure change attainable in the downstream end. Increasing the offset of the slug from the solenoid resulted in a higher pressure change symmetrically for compression and expansion, as shown in Figure 12 for various slug lengths. Each of the tests in Figure 12 used Solenoid B, which was calculated to have an optimal slug length of ~1.3 cm from the regression analysis. In Figure 12, the 1.3 cm slug length seemed to generate the highest pressure change, thereby verifying the regression analysis. The runs in Figure 12 maximized the capability of the 900 W power supply. For many applications, however, it is essential to consider a low-power system. Figure 13 shows the maximum pressure change versus current for three slug lengths. Because the Biot-Savart Law

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

60

J. C. Boulware, H. Ban, S. Jensen et al.

denotes a linear relationship between applied current and magnetic flux density, Eq. (5) indicates that the trend in Figure 13 should be quadratic; instead, it appears linear. This is likely due to resistance heating and the limited heat capacity of the solenoid. As the solenoid temperature increases, it cannot draw as much current and, thus, cannot generate as high a magnetic field. At high current levels, the solenoid may not stay cool long enough to generate high pressures, even during the 0.25 s pulse.

Figure 13. Experimental data obtained for the maximum differential pressure generated at various currents for 1.3 cm, 1.9 cm, and 2.5 cm slugs with one edge centered in the solenoid.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

6.3. Hydrodynamic Breakdown Each of the data points in Figure 12 represents a run conducted in which the slug maintained its form and remained intact. If the initial position was too far from the solenoid, the displacement of the slug generated a pressure force greater than the magnetic force, and the slug broke down, apparent in the data as well as physically in the experiment. Stationary experiments with a ferrofluid by Perry[31] verified the theoretical prediction that the pressure force could not exceed the maximum magnetic capability of the solenoid. In Perry‘s case, the stationary experiments induced a hydrostatic breakdown, whereas, in the current study, the breakdown took place during dynamic motion and occurred slightly earlier than the static equations predicted. With one edge of the slug centered in the solenoid and the other in a negligible field, Eq. (24) predicted that the breakdown pressure should have been 1.95 kPa. Resistance heating in the solenoid, however, limited the amount of current available at the time the maximum pressure change occurred. Assuming 19.5 A of current at the peak of the curve, the breakdown pressure should have been 1.45 kPa, which matches the experimental data much better, regardless of the fact that it is a static prediction and does not consider fluid properties, such as surface tension, cohesion, contact angle, and viscosity. Prediction of the hydrodynamic breakdown must also consider the uneven force distribution along the slug, pressure differential about the slug, and gravity, to be completely understood. The Reynolds number for the bulk slug motion remained under 1500 for all of the runs, but a more aggressive study may have to consider slug velocity and internal flow dynamics.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

61

A probabilistic method to predict risk of failure should be used to predict hydrodynamic breakdown, as it is likely due to a combination of the uneven force distribution along the slug, pressure differential about the slug, gravity, and low surface tension of LOX. For higherspeed tests, the rapid oscillations may also induce turbulent internal flow dynamics that cause the slug to break down.

6.4. Determining Uncertainty

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

By extracting the frequency and amplitude of the oscillations in the experimental data, the Matlab model could be used to calculate the hidden slug length, damping factor, and precise initial position of the slug. The authors performed detailed analyses on the uncertainties and found a logarithmic correlation between frequency and hidden slug length and also between amplitude and damping factor, as shown in Figure 14. The initial position of the slug also affected the amplitude; however, unlike the damping factor, it did not cause the mean to decrease over time. Based on the oscillations, these correlations could be used to precisely calculate the hidden slug length, damping factor, and initial position of the slug within the given uncertainty bounds of experimental measurements.

(a)

(b) Figure 14. (a) Hidden slug length vs. frequency and (b) damping factor vs. amplitude[34].

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

62

J. C. Boulware, H. Ban, S. Jensen et al.

6.5. Geometric Dependence A variety of geometries were tested to determine if an optimal configuration existed. As mentioned, the Matlab model could be used to predict the slug length which would produce the maximum pressure change for a particular solenoid. The authors confirmed the theoretical results by testing three configurations with Solenoids A and B as follows:

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Configuration 1: Solenoid A with an optimized slug length of 2.7 cm Configuration 2: Solenoid B with an optimized slug length of 1.3 cm Configuration 3: Solenoid B with a nonoptimized slug length of 2.5 cm Figure 15 shows the magnetic pressure for each configuration as the center of the slug relative to the center of the solenoid varies. The maximum magnetic pressure was proportional to the magnetic flux density each solenoid could produce. Thus, since Solenoid B was used for Configurations 2 and 3, the maximum magnetic pressure was the same for both. The peaks correlate to one edge of the slug in the center of the solenoid and, therefore, occur further out for Configurations 1 and 3 because they used longer slugs than Configuration 2. The experiments performed on each configuration were compared to determine any trends relative to solenoid geometry. Figure 16 - Figure 18 show experimental data for the maximum pressure change versus initial position for each configuration, as well as points of hydrodynamic breakdown and correlation to the numerical simulation. Since Configuration 1 had a larger downstream volume, the pressure change caused by slug displacement did not correlate with the same displacement in Configurations 2 and 3. Although the maximum magnetic pressure in Figure 15 indicates that the slugs should have broken down at 2.11 kPa for Solenoid A (Configuration 1) and 1.54 kPa for Solenoid B (Configurations 2 and 3), the experimental data showed a breakdown around 1.5 kPa for all three. Again, this is likely due to the aforementioned differences between hydrostatic and hydrodynamic breakdown.

Figure 15. Center position of slug relative to solenoid versus magnetic pressure on the slug for each solenoid[35]. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

63

Maximum Pressure Change (kPa)

2 1.5 1 0.5 0 -4

-3

-2

-1

\ 0

1

2

3

4

-0.5 -1 Experimental Data Hydrostatic Breakdown

-1.5

Simulation -2

Center of Slug Relative to Solenoid (cm)

Figure 16. Experimental and simulated data for the maximum pressure change in the downstream section as the initial position of the center of the slug varied for Configuration 1[35]. 2

Maximum Pressure Change (kPa)

1.5 1 0.5 0

-15

-10

-5

0

5

10

15

-0.5 -1 Experimental Data Hydrostatic Breakdown

-1.5

Numerical Simulation -2

Figure 17. Experimental and simulated data for the maximum pressure change in the downstream section as the initial position of the center of the slug varied for Configuration 2[35]. 2

Maximum Pressure Change (kPa)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Center of Slug Relative to Solenoid (mm)

1.5 1 0.5 0 -2.5

-1.5

-0.5

0.5

1.5

2.5

-0.5 -1 -1.5

Experimental Data Hydrostatic Breakdown Numerical Simulation

-2

Center of Slug Relative to Center of Solenoid (cm)

Figure 18. Experimental and simulated data for the maximum pressure change in the downstream section as the initial position of the center of the slug varied for Configuration 3[35].

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

64

J. C. Boulware, H. Ban, S. Jensen et al.

It was observed that throughout a series of runs conducted within the same time frame, the hidden slug length and damping factor were relatively unchanged. Thus, these uncertainties were found for a few points and then applied over the whole range of initial positions to obtain the numerical simulation. Since the simulation did not consider hydrodynamic breakdown, it could study points unachievable in the experiment. The study showed that the maximum pressure change eventually decreased as the initial position increased due to a lack of magnetic pressure far away from the solenoid. The peaks of the numerical simulation correlate to the magnetic pressure curve in Figure 15 reaching a nearnegligible region. This point occurs at about 3.25 cm for Configuration 1, 1.0 cm for Configuration 2, and 1.75 cm for Configuration 3. To apply the studies to other geometries, the findings must be nondimensionalized. Using Eq. (11) and (14), the nondimensional maximum displacement and initial magnetic pressure were seen to have a common trend, as seen in Figure 19. Each of the configurations held a nearly linear relationship regardless of their differences in overall size. The average slope of a linear fit of the three data sets was 2.38 and could be applied to other geometries with similar physical phenomena. For the current study, the capillary and gravity forces were negligible compared to the pressure, damping, and magnetic forces; thus, the results found could only be applied to systems of the same scale. Furthermore, a different paramagnetic fluid would have a different susceptibility and fluid properties, resulting in a different slope as well. Nonetheless, the correlation found is useful as a guideline during the preliminary design state of a magnetic fluid actuator. Non-Dimensional Maximum Displacement

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

0.015

0.01

0.005

0 -0.004

-0.003

-0.002

-0.001

0

0.001

0.002

0.003

0.004

-0.005 Config #1 -0.01

Config #2 Config #3

-0.015

Non-Dimensional Initial Magnetic Pressure

Figure 19. Nondimensional maximum displacement versus nondimensional initial magnetic pressure[35].

Each configuration used the same power source and generated approximately the same pressure change. Treating the pressure change as an indicator of work performed by the fluid would entail approximately the same output per power input for each configuration. However, since fluid damping is detrimental to performance, it could be seen as a form of exergy destruction. Exergy is the potential work of a system and is used to determine its second law

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

65

efficiency. High damping of the fluid reduces the potential work and, therefore, should be factored into a measure of the system‘s second law efficiency. This measure can be found through a nondimensional number known as the Mason number, the ratio of damping to magnetic forces, and is shown in Eq (30). Figure 20 shows pm,i versus ui for the three configurations. The slope of the linear trend line for each set of data provides an important characteristic of each combination, as it is a component of the Mason number. A lower Mason number indicates a more efficient system, since less damping would exist for a given magnetic force (achievable through power input). While it seems counterintuitive that a slower average initial velocity would imply a more efficient system, Solenoid A was nearly 4 times longer than Solenoid B; thus, the magnitude of velocity was relative. Instead, the Mason number is representative of exergy destruction and should, ideally, be as low as possible.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

(a)

(b)

(c) Figure 20. Average initial velocity versus initial magnetic pressure for (a) Configuration 1, (b) 2, and (c) 3[35].

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

66

J. C. Boulware, H. Ban, S. Jensen et al.

The near-linearity of the trend lines indicates that the Mason number is consistent for each configuration. This implies that with a given slug length and profile of the magnetic flux density along the axis of the solenoid, ui can be calculated through a proportionality constant of pm,i. When Eq. (9) is calculated for each run, the average Mason number for each configuration can be found as shown in Table 2, assuming each configuration has an average damping factor of 6 and hidden slug length of 13 cm. The correlations in Table 2 reveal two methods for improving system efficiency through geometric measures. By optimizing the slug length for a solenoid and by minimizing the overall geometry, damping can be lessened for a given magnetic force. While this seems particularly beneficial for applications with MEMS technology, it is important to remember that capillarity has not been introduced to the system. Nonetheless, the experiments and analyses performed have verified the theoretical model and are therefore useful to future research and applications. Table 2. Average Mason number Config. 1 2 3

# of Runs 41 26 37

Avg. Mason # 0.247 0.091 0.143

Std. Dev. 6.62 * 10-2 2.03 * 10-2 3.54 * 10-2

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

7. CONCLUSIONS The case for a LOX-based magnetic fluid system to replace mechanically moving parts in a satellite system has been argued. The experiment alone satisfied the first two objectives by confirming the potential of a LOX-based magnetic fluid system through innovative measuring techniques. The experiment and numerical simulation also verified the theoretical model, thus satisfying the third objective. The fourth objective was satisfied through the parametric studies which examined the maximum pressure change attainable, effect of uncertainty, and geometric variance. Accomplishing the objectives aided the overall goals by establishing the following conclusions: an optimal slug size for a specific geometry exists that maximizes the attainable pressure change for a given power source; a low-power system may perform more efficiently due to less resistance heating in the solenoid; unknown portions of LOX in the fluid system and undetermined physical phenomena can be empirically calculated through the theoretical model; an optimized slug and solenoid configuration will result in a more efficient system in terms of work performed by the fluid with a minimal amount of damping; reducing the physical scale of the experiment increases its efficiency, as long as capillarity is still negligible.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Liquid Oxygen Magnetohydrodynamics

67

The conclusions of this research support the overall goals in aiding the design of a LOXbased magnetic fluid system with no moving parts for small satellite applications.

REFERENCES [1] [2] [3]

[4] [5]

[6]

[7] [8] [9]

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[10]

[11]

[12] [13] [14]

[15] [16]

[17]

Rosensweig, RE. Ferrohydrodynamics, Dover, New York, 1985. Park, GS; Seo, K. ―A Study on the Pumping Forces of the Magnetic Fluid Linear Pump,‖ IEEE Transactions on Magnetics, Vol. 39, No. 3, 2003, 1468-1471. Park, GS; Seo, K. ―New Design of the Magnetic Fluid Linear Pump to Reduce the Discontinuities of the Pumping Forces,‖ IEEE Transactions on Magnetics, 2004, Vol. 40, No. 2, 916-919. Seo, K; Park, GS. ―A Research on the Pumping Forces in the Magnetic Fluid Linear Pump,‖ IEEE Transactions on Magnetics, 2005, Vol. 41, No. 5, 1580-1583. Hatch, A; Kamholz, AE; Holman, G; Yager, P; Bohringer, KF. ―A Ferrofluidic Magnetic Micropump,‖ Journal of Microelectromechanical Systems, 2001, Vol. 10, 215-221. Moghadam, ME; Shafii, MB; Dehkordi, EA. ―Hydromagnetic Micropump and Flow Controller. Part A: Experiments with Nickel Particles Added to the Water,‖ Experimental and Thermal Fluid Science, 2009, Vol. 33, No. 6, 1021-1028. Krauss, R; Liu, M; Reimann, B; Richter, R; Rehberg, I. ―Pumping Fluid by Magnetic Surface Stress,‖ New Journal of Physics, 2006, Vol. 8, No. 18, 1-11. Goldstein, SR. “Magnetic Fluid Actuated Control Valve, Relief Valve and Pump,” U.S. Patent 4053952, 1977. Kamiyama, S. ―A Magnetic Fluid Actuator,‖ Advanced Robotics, 1986, Vol. 1, No. 2, 177-185. Ming, Z; Zhongliang, L; Guoyuan, M; Shuiyuan, C. ―The Experimental Study on Flat Plate Heat Pipe of Magnetic Working Fluid,‖ Experimental and Thermal Fluid Science, 2009, Vol. 33, No. 7, 1100-1105. Jeyadevan, B; Koganezawa, H; Nakatsuka, K. ―Performance Evaluation of Citric IonStabilized Magnetic Fluid Heat Pipe,‖ Journal of Magnetism and Magnetic Materials, 2005, Vol. 289, 253-256. Liao, W; Chen, X; Chen, Y; Pu, S; Xia, Y; Li, Q. ―Tunable Optical Fibers with Magnetic Fluids,‖ Applied Physics Letters, 2005, Vol. 87, 1-3. Celik, D; Van Sciver, SW. ―Dielectric Coefficient and Density of Subcooled Liquid Oxygen,‖ Cryogenics, 2005, Vol. 45, 356-361. Hilton, DK; Van Sciver, SW. ―Absolute Dynamic Viscosity Measurements of Subcooled Liquid Oxygen from 0.15 MPa to 1.0 MPa,‖ Cryogenics, 2008, Vol. 48, 5660. Lide, DR. (ed.), CRC Handbook of Chemistry and Physics, 89th Edition, CRC Press/Taylor and Francis, Boca Raton, 2009. Takeda, M; Nishigaki, K. ―Measurements of the Surface Tension of Liquid Oxygen in High Magnetic Fields,‖ Journal of the Physical Society of Japan, 1992, Vol. 61, No. 10, 3631-3635. Catherall, AT; Benedict, KA; King, PJ; Eaves, L. ―Surface Instabilities on Liquid

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

68

[18]

[19]

[20] [21]

[22]

[23]

[24] [25] [26] [27]

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[28]

[29] [30] [31] [32] [33]

[34]

[35]

J. C. Boulware, H. Ban, S. Jensen et al. Oxygen in an Inhomogeneous Magnetic Field,‖ Physical Review E, 2003, Vol. 68, 1-3. Catherall, AT; López-Alcaraz, P; Benedict, KA; King, PJ; Eaves, L. ―Cryogenically Enhanced Magneto-Archimedes Levitation,‖ New Journal of Physics, 2005, Vol. 7, No. 118, 1-10. Hilton, DK; Celik, D; Van Sciver, SW. ―Subcooled Liquid Oxygen Cryostat for Magneto-Archimedes Particle Separation by Density,‖ Advances in Cryogenic Engineering, CP985, 2008, Vol. 53, AIP, 1517-1522. Yerkes, KL. ―Use of Magnetic Fields to Augment Wicking of Oxygen Heat Pipes,‖ Topics in Heat Transfer HTD-ASME, 1992, Vol. 3, 115-120. Youngquist, RC; Immer, CD; Lane, JE; Simpson, JC. ―Dynamics of a Finite Liquid Oxygen Column in a Pulsed Magnetic Field,‖ IEEE Transactions on Magnetics, 2003, Vol. 39, No. 4, 2068-2073. Zahn, M; Greer, DR. ―Ferrohydrodynamic Pumping in Spatially Uniform Sinusoidally Time-Varying Magnetic Fields,‖ Journal of Magnetism and Magnetic Materials, 1995, Vol. 149, 165-173. Mao, L; Koser, H. ―Ferrohydrodynamic Pumping in Spatially Traveling Sinusoidally Time-Varying Magnetic Fields,‖ Journal of Magnetism and Magnetic Materials, 2005, Vol. 289, 1999-2002. Shliomis, MI. ―Effective Viscosity of Magnetic Suspensions,‖ Soviet Journal of Experimental and Theoretical Physics, 1972, Vol. 34, No. 6, 1291-1294. McTague, JP. ―Magnetoviscosity of Magnetic Colloids,‖ Journal of Chemical Physics, 1969, Vol. 51, No. 1, 133-136. Bacri, JC; Perzynski, R; Shliomis, MI. ―Negative-Viscosity Effect in a Magnetic Fluid,‖ Physical Review Letters, 1995, Vol. 75, No. 11, 2128-2131. Cunha, FR; Sobral, YD. ―Asymptotic Solution for Pressure-Driven Flows of Magnetic Fluids in Pipes,‖ Journal of Magnetism and Magnetic Materials, Vol. 289, 314-317. Chen, CY; Hong, CY; Wang, SW. ―Magnetic Flows in a Tube with the Effects of Viscosity Variation,‖ Journal of Magnetism and Magnetic Materials, 2002, Vol. 252, 253-255. Schlicting, H. Boundary-Layer Theory, 7th ed; McGraw-Hill, New York, 1979, 436438. White, FM. Viscous Fluid Flow, McGraw-Hill, New York, 1991, 116-118. Perry, MP; Jones, TB. ―Hydrostatic Loading of Magnetic Liquid Seals,‖ IEEE Transactions on Magnetics, 1976, Vol. MAG-12, No. 6, 798-800. Bashtovoi, V; Kuzhir, P; Reks, A. ―Capillary Ascension of Magnetic Fluids,‖ Journal of Magnetism and Magnetic Materials, 2002, Vol. 252, 265-267. Boulware, JC; Ban, H; Jensen, S; Wassom, S. ―Experimental Studies of the Pressures Generated by a Liquid Oxygen Slug in a Magnetic Field,‖ J Magn Magn Mater, 2010, Vol. 322, 1752-1757. Boulware, JC; Ban, H; Jensen, S; Wassom, S. ―Modeling of the Dynamics of a Slug of Liquid Oxygen in a Magnetic Field and Experimental Verification,‖ Cryogenics, 2010, doi:10.1016/j.cryogenics.2010.03.004. Boulware, JC; Ban, H; Jensen, S; Wassom, S. ―Geometric Influence on Liquid Oxygen Magnetohydrodynamics,‖ Article in Press, accepted for publication in Exp Therm Fluid Sci.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

In: Cryogenics: Theory, Processes… Editor: Allyson E. Hayes, pp. 69-91

ISBN: 978-1-61761-323-4 © 2011 Nova Science Publishers, Inc.

Chapter 3

EFFECT OF CRYOGENIC TREATMENT ON MICROSTRUCTURE AND MECHANICAL PROPERTIES OF LIGHT WEIGHT ALLOYS Kaveh Meshinchi Asl1 and Mehdi Koneshloo2 1

School of Materials Science and Engineering, Clemson University, Clemson, SC, USA 2 Department of Materials Science and Engineering, Sharif University of Technology

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

ABSTRACT This chapter mainly focuses on the effects of low temperature (subzero) treatments on microstructure and mechanical properties of aluminum and magnesium alloys. Deep cryogenic treatment on A319 aluminum alloy showed that the abrasion resistance of the alloy was improved after the treatment. This improvement was attributed to the strengthening of the α-aluminum matrix which slows down the propagation of the existing defects. The execution of deep cryogenic treatment on AZ91 magnesium alloy changed the distribution of β precipitates. The tiny laminar β particles almost dissolved in the microstructure and the coarse divorced eutectic β phase penetrated into the matrix. This microstructural modification resulted in a significant improvement on mechanical properties of the alloy. The steady state creep rates were measured and it was found that the creep behavior of the alloy, which is dependent on the stability of the near grain boundary microstructure, was improved by the deep cryogenic treatment. After the deep cryogenic treatment, the sliding of grain boundaries was greatly suppressed due to morphological changes. As a result, the grain boundaries are less susceptible for grain boundary sliding at high temperatures. After dry sliding wear tests were performed, the wear resistance of the alloy improved remarkably after deep cryogenic treatment. Furthermore due to interest in the subzero treatments of steels in the past few decades, AISI H13 tool steel was chosen and cryogenic treatment at -72ºC and deep cryogenic treatment at -196ºC were applied and it was found that the execution of low temperature treatments on samples affected the microstructure of the H13 tool alloy to a great extent. By applying the subzero treatments, the retained austenite was transformed to martensite due to the completion of martensite transformation. The cryogenic treatment at a very low temperature and holding the samples for a long time, also lead to precipitation of more uniform and very fine carbide particles. This microstructural

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

70

Kaveh Meshinchi Asl and Mehdi Koneshloo modification resulted in a significant improvement on mechanical properties and wear resistance of the alloy.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

INTRODUCTION Cryogenic treatment also known as cold or subzero treating is a very old process and is widely used for high precision parts. The use of extreme cold to strengthen metals has been used since long time ago for centuries. Swiss watch-makers stabilized the delicate components of their timepieces by storing them for several years in mountain caves to obtain maximum performance and precision. In general, unlike surface treatments, the cryogenic treatments influence the core properties of the materials. Scientists have been experimenting with the use of extreme cold to strengthen metals since the mid-1800s. However the process of experimentation and understanding of the cryogenic treatment of steels really got under way during World War II. Steel tooling was immersed in liquid nitrogen for a brief period of time, then warmed up and placed into service. Unexpectedly, these cryogenically treated steels would experience a greatly enhanced service life. Cryogenic treatment has been shown to result in significant increases in the wear resistance of steels such as D-2 and M-2. It has been believed that the increase in wear resistance is a direct result of the reduction in the amount of retained austenite and change in the carbide morphologies. In order to avoid confusion, a distinction will be drawn between "cold treatment (CT)", at temperatures down to around -72ºC (dry ice temperature) and "deep cryogenic treatment (DCT)", near liquidnitrogen (-196ºC) temperatures. Deep cryogenic treatment improves certain properties beyond the improvement obtained by normal cold treatment [1-11]. The increasing demand for reducing weight for aerospace and automotive applications has been the motivation for many research programs concerning light weight alloys. There has been tremendous effort in the past decades to improve the properties of Aluminum, Magnesium and Titanium alloys for light weight applications. In certain applications, lightweight alloys are subjected to sliding motion. Thus, wear resistance is also becoming a key factor in these alloys. Recently there have been some attempts to apply cryogenic treatment to improve the properties of light weight alloys and extend the use of this process from steels to other materials. Magnesium alloys have been identified as ideal materials for lightweight structures in automotive, aerospace and railway industries with great development potential because of their low density, high specific strength and stiffness, superior damping capacity, good electromagnetic shielding characteristics and good machinability. It is expected that a lot of kinds of magnesium alloys have been used for structural materials and the use of magnesium castings has been expanded at impressive rate. Therefore it is important to develop the magnesium alloys having excellent mechanical properties. As yet, the number of commercially available magnesium alloys is still limited especially for applications at elevated temperatures [12-16]. Magnesium alloys offer lightweight alternatives to conventional metallic alloys because of their low density. For critical automobile applications, wear properties of magnesium alloys are important. Despite the attractive range of bulk mechanical properties, a relatively poor resistance to fracture and wear is a serious hindrance to wider application of these alloys. The wear properties can be varied substantially through changes in the microstructure and the morphology or volume fraction, mechanical properties and the nature of the interface between

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Effect of Cryogenic Treatment on Microstructure and Mechanical…

71

matrix and the reinforcing phase. The most commonly used magnesium alloy, being used in approximately 90% of all magnesium cast products, is AZ91 which contains 9.0 wt.% Al and 0.9 wt.% Zn. While this alloy has good castability and exhibits adequate tensile yield strengths at ambient temperature, it faces strong challenges such as creep and wear resistance [17-23]. Thus, in this investigation deep cryogenic treatment was applied to this alloy to study its effect on mechanical properties and creep resistance of AZ91 alloy. Aluminum alloys are also characterized as one of the light weight metals which provide a high strength to weight ratio. These alloys have the most applications between the light metals and are used in a wide range of applications from constructions to aerospace and etc. Thus some parameters such as strength, hardness, high temperature properties, thermal expansion and wear resistance become a key factor in the extensive use of these alloys. One of the most commonly used aluminum-silicon (Al-Si) alloys is the A319 aluminum alloy which contains manganese, silicon and copper as the main alloying elements. Due to presence of silicon which results in Al-Si eutectic, this group of aluminum alloys has a very high fluidity. This alloy has some applications in automotive industries and thus is commercially important. It could be used in fabrication of engine blocks, cylinder heads and pistons. This is due to its good castability characteristics, good mechanical properties and low cost of the production. Recently there has been extensive studies to improve the properties of this alloy because of the demand for the lighter vehicles. Thus the deep cryogenic treatment was also applied to this alloy to study its effects on some mechanical properties and wear resistance of the alloy. Tool steels have high hardness, high toughness, good machinability and low deformation during heat treatment and thus are used for the manufacture of hot forging dies. H13 hot work tool steel is normally used as the die material in precision molds of manufacturing tools for die casting. Die failure is an important issue in industries and many operating difficulties and loss of production arise through die wear which is costly. The microstructure and wear properties of tools are thus essential elements in quality and production control of the parts. While this alloy exhibits good properties, it faces strong challenges such as wear resistance [9-11]. Thus, deep cryogenic treatment can be applied to this alloy to study its effect on microstructure and mechanical properties such as abrasion characteristics before and after cryogenic treatment. The beneficial effects of cryogenic treatment on improved performance of a majority of metallic and non metallic materials have not been presented until now. However, the special cryogenic conditions which are indispensable for obtaining the observed beneficial results in these materials have previously been explained as homogenization and stabilization of internal microstructure [5-11]. The definitive explanation has yet to be offered. In this study, the effects of subzero treatments on three different systems is investigated. The AZ91 and A319 alloys were chosen as representatives for the magnesium and aluminum alloys to study the effects of applying cold temperatures on light metals. Furthermore since the few investigations on the effects of the subzero temperatures are limited to a few different steels, the AISI H13 hot work tool steel was chosen as a new steel alloy system. This helps us to compare the characteristics of the light weight metallic systems and steels alloys before and after applying the process of subzero treatment. This leads to a better understanding of the changes that the materials undergo during this treatment.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

72

Kaveh Meshinchi Asl and Mehdi Koneshloo

EXPERIMENTAL Cryogenic treatment of samples has been performed by placing specimens in an isolated alumina chamber which was designed by means of the heat transfer equations to estimate the thermal gradient of the chamber (Figure 1). The top of the chamber was covered by insulator wool after placing the samples in the chamber. This chamber was progressively immersed in a liquid nitrogen reservoir by means of an electric motor. The sample temperature was monitored by a K type thermocouple which was used to operate a step motor which lowered the sample and maintained a temperature decline at the rate of 0.5 ºC min-1. Steps were about 1 mm and approximately 8 hours was taken to reach to about -196ºC. This painstaking method eliminates the probability of thermal shock and micro cracking. Specimens were held at either cryogenic temperature at -72°C (193K) or deep cryogenic temperature at -196°C (77K) for 8 hr and then slowly brought up to approximately +25°C to study the effect of tempering treatment on the properties of the samples. For all the samples tested in this investigation, the samples were cut from the same place of the cast ingot and had the same microstructure in terms of grain size and precipitate distribution.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

AZ91 Magnesium Alloy Commercial pure magnesium, Aluminum and zinc (>99.9%) were used to prepare AZ91 alloy. Manganese was added as Aluminum-25%Manganese master alloy. Alloy was melted in a mild steel crucible and during the melting process, surface of the charge and the melt were protected by Magrex 36 flux. Finally after complete defluxing, the melt was poured at the pouring temperature of 750˚C into a permanent mold which was preheated up to 300˚C. The chemical composition of the samples was determined by wet chemical analysis as shown in Table 1. Mechanical properties of round tensile samples (ASTM-B557M standard) were measured using an Instron 4400 machine (1 mm/min strain rate). The hardness was measured using Instron Brinell hardness tester. The averages of three measurements were reported for each experiment. Creep tests were performed in air using a Santam creep test machine (ASTME139 standard). The displacement was measured using a strain gauge in the experiments. By plotting the displacement at various times, the strain curve was obtained and the steady state strain rate was calculated. The microstructure and phase distribution were characterized by Philips XL 40 scanning electron microscope (SEM). The elemental contents of various phases in the polished samples were determined by the energy dispersive X-ray spectroscopy (EDS) system of SEM. Dry sliding wear test without lubricant according to ASTM G99 was conducted using a pin-on-disk apparatus. Pin specimens of diameter 5 mm and length 15 mm were prepared by rods grinding up to 1200-grit with silicon carbide, polishing with 0.05 µm diamond paste and alcohol, followed by cleaning with ultrasonic equipment in pure acetone. M35 hardened tool steel of 160 mm diameter and 65 HRc hardness was used as a counterface. After each test the disc surface was ground against 1200-grit SiC paper and cleaned with acetone. The experiments were performed under 10, 20, 30, 50 and 100N stationary normal loads and

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

73

sliding speeds of 0.25, 0.5 and 1.0 m/s at various sliding distances from 250 to 2000m. Weight loss values were determined from weight differences before and after the tests using a precise electronic balance with an accuracy of ±0.1 mg. Weight loss versus sliding distance curves were plotted and the wear rates were calculated.

Figure 1. Schematic configuration of cryogenic treatment system.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

‍Table 1. Chemical compositions of samples Alloy Code AZ91

Al 9.10

Zn 0.90

Mn 0.25

RE -

Mg Rem.

The worn pins surfaces were examined using a scanning electron microscopy (SEM) and energy dispersive X-Ray spectroscopy (EDAX) to verify dominant wear mechanisms.

A319 Aluminum Alloy Ingot was melted in a clay graphitic crucible at 730 ºC using an electrical resistance furnace. Degassing of the melt was carried out by an argon rotary degassing installation (15 minutes). Castings were produced from both melts using an ASTM B108 permanent mold. Mechanical properties of round tensile samples (ASTM-1357M - 12.5 mm gage length) were measured using an Instron 4400 machine (1 mm/min cross head). The hardness was measured using Instron Brinell hardness tester (5 mm ball - 250 Kg load). The averages of three measurements were reported for each experiment. As-cast hardness of the samples was measured after three days to allow the mechanical properties of the alloy to stabilize which could change during the natural aging.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

74

Kaveh Meshinchi Asl and Mehdi Koneshloo

Metallographic samples were prepared by mounting the abraded sections in cold setting resin. Microstructures and fracture surfaces were studied by Olympus PME3 optical microscope and Philips XL 40 scanning electron microscope. Table 2 shows the chemical composition of the samples. Abrasion tests were carried out to study the effect of cold treatment cycles on erosion wear properties of the alloy (Figure 2). Cylindrical samples (10 mm diameter, 100mm length) were prepared for abrasion tests. Specimens were prepared by grinding rods up to 1200-grit with silicon carbide, polishing with 0.3 and 0.05 µm alumina powder suspensions and alcohol, followed by cleaning with ultrasonic equipment in pure acetone. Samples were immersed 50 mm in a silica sand containing chamber and rotated in contact with abrasive SiO2 particles (0.4-0.6 mm) using an electric motor (60 rpm). The angularity number of the silica particles was 1.33 according to AFS standard. The silica grain bath was replaced every two hours by new silica particles. Weight loss values were determined from weight differences before and after the tests using a precise electronic balance with an accuracy of ±0.1 mg and compared in different conditions.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 2. Abrasion test equipment used in the experiments.

Table 2. Chemical compositions of samples Si 5.9

Cu 3.54

Mg 0.38

Fe 0.62

Mn 0.35

Ni 0.28

Ti 0.11

Al Rem.

‍Table 3. Chemical compositions of the H13 tool steel used in the experiments compared with ASTM A-671standard Element C nM iS rC oM V P S eF

H13 (Wt%) 0.36 0.38 0.96 4.82 1.19 0.86 0.017 0.004 Rem.

ASTM A-671 (Wt%) 0.32-0.45 0.20-0.60 0.80-1.26 4.75-5.50 1.10-1.75 0.80-1.20 Max. 0.03 Max. 0.03 Rem.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

75

AISI H13 Hot Work Tool Steel The chemical composition of the annealed H13 tool steel rod with a diameter of 20mm which was used in this study is presented in Table 3. The composition is in the range of H13 tool steel based on the ASTM A-671 standard. Mechanical properties of round tensile samples (ASTM-B557M standard) were measured using an Instron 4400 machine (1 mm/min strain rate). The hardness was measured using Wolpert HRC hardness tester. The averages of three measurements were reported for each experiment. For evaluation of the fracture toughness, the impact V-notch samples were prepared. The impact test samples were prepared using the ASTM-A370 standard. The microstructure and phase distribution were characterized by Philips XL 40 scanning electron microscope (SEM). The elemental contents of various phases in the polished samples were determined by the energy dispersive X-ray spectroscopy (EDS) system of SEM. Abrasion tests were carried out using the same procedure that was used for the A319 aluminum alloy. All the samples used in these experiments were held at 1040°C for 30 min for austenitizing followed by air quenching before any treatments. Table 4 shows the different treatment sequences tested on the samples in this investigation.

‍Table 4. Heat treatment sequences of tool steel samples used in this study

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Sample Processing sequences 1

Heat treated at 1040oC for 30 min followed by air quench

Tempered at 560°C for 2hr

-

2

Heat treated at 1040oC for 30 min followed by air quench

Tempered at 560°C for 2hr

Cryogenic treatment at -72°C for 8hr

3

Heat treated at 1040oC for 30 min followed by air quench

Tempered at 560°C for 2hr

Deep Cryogenic treatment at 196°C for 8hr

4

Heat treated at 1040oC for 30 min followed by air quench

Tempered at 560°C for 2hr

Deep Cryogenic Tempered at treatment at 560°C for 2hr 196°C for 8hr

-

RESULTS AZ91 Magnesium Alloy The microstructure of as-cast AZ91 alloy consists of α-Mg matrix and β (Mg17Al12) phase. In the as cast AZ91 alloy, the β phase exhibited two morphologies. The first one, coarse with irregular morphology, is eutectic β. The second, tiny laminar shaped particles and surround the eutectic β phase, is a precipitate of β due to decrease of aluminum solubility with

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

76

Kaveh Meshinchi Asl and Mehdi Koneshloo

decreasing temperature (Figure 3). According to the Mg-Al phase diagram and previous investigations [24,25], these two particles have the same composition. In the previous work, it was demonstrated that although the β phase has the main strengthening effect on the Mg-Al based alloys at room temperature, it has a low melting temperature and is the main reason for poor mechanical properties of these alloys at elevated temperatures [24]. a

c

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

b

d

Figure 3. (a) SEM micrograph of as-cast structure, (b) SEM micrograph showing two kinds of β particles, divorced eutectic and tiny lamellar particles, (c) SEM micrograph of as-cast cryogenic treated structure, (d) SEM micrograph showing changes in morphologies of two kinds of β particles.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

Element Magnesium K K munimulA

77

At. % 60.14 39.86

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 4. Energy dispersive X-ray analysis of the eutectic β phase penetrated into the matrix.

Non-equilibrium solidification is the reason that β particles exhibited two different kinds of morphologies [25]. According to the binary Mg-Al phase diagram, the composition of AZ91 alloy is close to the hypoeutectic Mg-Al alloys. Since the mold for casting in this investigation was from mild steel, the cooling rate of the alloys was high, hence, it was easy to produce divorced eutectic. After alloy solidification, the precipitation of β occurred during continuous cooling and caused the formation of the second kind of β, the tiny laminar particles. After deep cryogenic treatment was applied to as-cast specimens, the microstructure of AZ91 alloy altered. The morphology of β phases in the microstructure changed impressively after cryogenic treatment as shown in Figure 3 (The same sample was used by taking micrographs of as-cast sample, then cryogenically treating the sample and taking some micrographs again). The coarse divorced eutectic β phase penetrated the matrix (Figure 4). As a result of this change in the microstructure, the mechanical properties of cryogenic treated samples increased compared with the untreated as-cast specimens as shown in Table 5. This improvement was attributed to the strengthening of the matrix against propagation of the existing defects which is due to the important role of β precipitates in the microstructure which are the main strengthening effect at room temperatures. Since the β precipitates are mainly distributed at grain boundaries, the morphology of the β particles after cryogenic treatment and the stabilization of internal microstructure are both key factors in strengthening these alloys.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

78

Kaveh Meshinchi Asl and Mehdi Koneshloo

The creep tests were carried out to investigate the effect of cryogenic treatment on creep properties of the alloys. Creep tests were performed on specimens over the temperature range of 135ºC to 200ºC and at an applied stress of 48-96 MPa. By measuring the slope of the typical creep curves, the steady creep rate of the alloys can be calculated and it is shown in Figure 5 for different alloys at 200ºC. The most important result caused by cryogenic treatment was the improvement in creep resistance of the alloy. The creep resistance of AZ91 alloy was relatively poor but it was observed that cryogenic treatment resulted in remarkable improvement of creep resistance of the alloy.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 5. Steady state creep rates of the alloys for different stresses in 200ºC.

Table 5. Ultimate Tensile strength, Yield strength and hardness of samples Sample

Yield Strength (MPa)

Ultimate Tensile Strength (MPa)

As-cast

91

170

After cryogenic

98

187

ssendraH )NHB( 59 67

The conventional power law equation relating the minimum creep rate, ε min to the applied stress is: .

ε min = A (ζ/G)n exp(-Q/RT)

Both n and Q are parameters of the material and together may be used to identify the dominant creep mechanism for the material. By plotting logarithmically the minimum creep strain rate versus the applied stress ζ, we can calculate the stress exponent, n. Plot of log ε. versus 1/T will yield the apparent activation energy, Q [26,27].

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Effect of Cryogenic Treatment on Microstructure and Mechanical…

79

The stress exponent n~2 is generally reported for grain boundary sliding, while n=3~7 is for dislocation climb controlled creep, however high n values for high stresses often indicate of power law breakdown [27,28]. The activation energy for self diffusion in magnesium can be taken as 138 kJ mol-1. The value of activation energy for boundary diffusion is estimated to be in the range 70-100 kJ mol-1 [27]. At low stresses, the stress exponent for as cast AZ91 was 5.6 which is in the range of dislocation climb mechanism however the activation energy was measured 121 kJ mol-1 which lies between that for self diffusion and grain boundary diffusion. Together these two factors may indicate a mixed mode of creep behavior, with some grain boundary effects contributing to the overall behavior. For the cryogenic treated AZ91 alloy, the stress exponent was 3.4 and the activation energy was measured 132 kJ mol-1. The values of both parameters fit the values for dislocation climb controlled creep. The activation energy for cryogenically treated alloy is slightly larger than AZ91 alloys due to its better creep resistance. The discontinuous precipitation of β dominant at grain boundaries, weakens the grain boundaries at elevated temperatures. Although they are the main strengthening factor of the alloy at room temperature, they can readily soften and coarsen with increase of temperature. Thus they can have deleterious effects on the high temperature properties of the alloy. Therefore sliding of grain boundaries is an important factor in deformation of Mg-Al alloys at elevated temperatures. After the cryogenic treatment, the microstructure of the alloy was modified. The tiny laminar β particles were almost dissolved in the microstructure and the coarse divorced eutectic β phase penetrated the matrix. Thus suppressing the grain boundary sliding to great extent. As a result of this change in the microstructure, the creep properties of cryogenic treated samples increased compared with the as-cast specimens with no treatments. Figure 6 represents the corresponding wear rate under different normal loads for various sliding speeds. At constant sliding speed, application of higher normal load increases the wear rate generally. On the other hand, under invariable conditions of load, enhancing sliding speed increases the wear rate. It should be noted that, improving sliding speeds from 0.25 to 0.5m/s lead to amplification of wear rates however, experimenting at 1.0m/s raises wear rates significantly more than 0.5m/s. Observations reveal that wear rates increase abruptly between 30 and 50N vertical loads at 0.25 and 0.5m/s sliding speeds, but gradually at 1.0m/s. Diagrams indicate that pure AZ91 magnesium alloy examined at 0.25m/s exhibits the least wear rate except under 100N normal load. In this fashion, deep cryogenic treatment had no positive effect on AZ91 wear resistance except in high loads. Tests performed at 0.5m/s show a similar behavior unless AZ91 preserves its minor wear rate up to 50N. It seems the deep cryogenic process influences the wear rate, especially at higher loads. This fact may be related to the new morphological microstructure of β phase which has better wear resistance than α phase. Indeed, deep cryogenically treatment samples when tested at 1.0m/s display superior wear resistance compared with pure AZ91 alloy. Further investigation is in progress. Traces of parallel grooves can be distinguished on surfaces of most samples. These scratches may be attributed to hard asperities of counterface or detached particles that are removed from disk or pin and placed in surface contact [29]. Figure 7 illustrates typical feature of such a mechanism. Abrasion is the prevailing wear mechanism at lower loads. With wear process advancement, grooves become wider and deeper at sufficiently longer sliding distances.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

80

Kaveh Meshinchi Asl and Mehdi Koneshloo

As cast AZ91

Cryogenic treated AZ91

32

Wear rate x10 -4 (g/m)

28 24 20 16 12 8 4 0 10

20

30

50

100

Force (N)

a)

Wear rate x10

-4

(g/m)

As cast AZ91 32 28 24 20 16 12 8 4 0 10

20

30

50

100

Force (N)

b) As cast AZ91

Cryogenic treated AZ91

32

Wear rate x10 -4 (g/m)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Cryogenic treated AZ91

28 24 20 16 12 8 4 0 10

20

30

50

100

Force (N)

c) Figure 6. Wear rate values for a) 0.25, b) 0.5 and c) 1.0m/s sliding speed for Cryogenic and AZ91 alloy. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

81

Extensive loads at long sliding distances induce subsurface crack formation and growth. If propagated cracks joined the surface large craters will be produced [30]. Characteristic sketches of delamination were identified for long sliding distances (Figure 8). High loads at long sliding distances, predominantly for greater sliding speeds, induce severe plastic deformation on materials [29,30]. Deep cryogenic treatment stabilizes the internal microstructure due to new morphology of β phase particles that acquires suitable mechanical properties that enhance the load bearing restriction of AZ91 alloy. At more severe wear conditions of highest normal load and sliding speed, specimens that had undergone cryogenic treatment show superior wear resistance particularly to plastic deformation.

A319 Aluminum Alloy

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

The microstructure of this alloy consists of α aluminum matrix (dendrite network), eutectic Si particles located between the dendrites (interdendritic region) and intermetallic phases due to presence of the alloying elements. Alloying elements such as Mg and Cu are often added to improve the high temperature properties of this alloy.

Figure 7. Characteristic Scheme of abrasive wear for a pure AZ91 alloy experimented at 0.5m/s under 20N normal load after 1000m.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

82

Kaveh Meshinchi Asl and Mehdi Koneshloo

b

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

(a)

(b) Figure 8. Delamination mode of wear for a) pure AZ91 alloy tested at 0.25m/s, 50N and 1000m, b) pure AZ91 alloy tested at 0. 5m/s, 100N and 1000m.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

83

‍Table 6. The results of the abrasion test of samples Sample Reference After cryogenic

Mean weight loss after 12 hrs (g) 0.0282 0.0150

Mean weight loss after 24 hrs (g) 0.0305 0.0186

After the deep cryogenic treatment the tensile strength of the samples were increased however the most important effect of the deep cryogenic treatment on A319 aluminum alloy was on the wear resistance as presented in Table 6. This could be attributed to the improvement of the resistance of the alloy against the propagation of the existing defects. Worn surface of a deep cryogenically treated sample is presented in Figure 9. Analysis of the white grains in the microstructures shows that they are the silica abrasive grains which have penetrated in the soft dendritic  - aluminum matrix. The dimples result from detachment of the silica particles. In the cryogenically treated samples particles did not detach readily from the matrix due to strengthening of the matrix by the cryogenic treatment. The abrasives have not penetrated the eutectic silicon grains due to their higher hardness.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

AISI H13 Hot Work Tool Steel The annealed H13 tool steel samples used in this investigation were held at 1040°C for 30 min for austenitizing followed by air quenching before any treatments. This is the conventional hardening treatment of H13 tool steels and the microstructure of the alloy based on the phase diagram consists of needle type martensite, retained austenite and carbides (Figure 10). The conventional hardening treatment results in significant amount of retained austenite which has some damaging effects on mechanical properties of tool steels such as machinability, wear, hardness and most important of all on dimensional stability of tool steels. The latter could be a very significant factor in the case of using the tool steels for die material applications. Thus different treatment cycles were applied on the samples to study the effects of low temperature treatments on H13 tool steel. This was done by cooling the samples at temperatures well below the Mf temperature of the H13 tool steel and holding the samples in this temperature range. Figure 11 shows scanning electron microscopy images of the microstructure of alloy 1 to 4 after different heat treatment sequences (Table 4). The execution of cryogenic treatment had a significant effect on the microstructure of the alloy and led to transformation of retained austenite to martensite. As the cryogenic temperature is lowered, more austenite is transformed to martensite. However besides the transformation of retained austenite to martensite, it could be seen that the martensite laths are smaller and distributed more uniformly in the microstructure after holding the samples for a long time at the deep cryogenic temperatures. The microstructure modification can be very important in terms of mechanical properties of and dimensional stability of the tool steel. As stated earlier, the microstructure of this alloy consists of large carbide particles in the matrix. After tempering, significant carbide refining occurred in the microstructure of deep cryogenically treated samples. Further examination of the microstructure indicates that there

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

84

Kaveh Meshinchi Asl and Mehdi Koneshloo

are mainly two different kinds of carbides, chromium (black) and molybdenum (white) carbides (Figure 12-a). The chromium carbides are the dominant carbides in microstructure. The energy dispersive X-ray analysis of both kinds of carbides is presented in Figure 13.

Silicon abrasives

Detached areas

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 9. Worn surface of the cryogenically treated degassed sample after abrasion test.].

Martensite laths

Carbide particles

10µm

Figure 10. Optical image of the microstructure of sample 1 which consists of martensite laths, retained austenite and carbides.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

1

2

3

4

85

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

2 Figure 11. Scanning electron microscopy images of the samples 1 to 4 after different heat treatment 3 4 cycles. The cryogenic treatment at a very low temperature and holding the samples for a long time, results in precipitation of very fine and more uniform distribution of carbide particles in the microstructure after tempering (Figure 12). The carbides also tend to be more spherical in the deep cryogenically treated sample. This could be very important factor since more uniform and spherical carbides result in uniform hardness which is essential in tool steel applications. This is due to low temperature transformation of austenite to martensite and finer distribution of martensite needles in the microstructure. Since carbide precipitation in tempering treatment needs short range diffusion of carbon atoms, it seems that applying the process of deep cryogenic treatment at low temperatures on the H13 tool steel which results in formation of finer martensite laths in the microstructure, initiates nucleation sites for precipitation of fine carbide particles which results in the enhancement of mechanical properties of the alloy. It is also well known that this process is diffusion dependant and it seems that soaking for 8 hours at very low temperatures followed by tempering can fully modify the microstructure.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

86

Kaveh Meshinchi Asl and Mehdi Koneshloo

a)

b)

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 12. Scanning electron microscopy images showing the effect of deep cryogenic treatment on the Chromium (black) and molybdenum (white) carbides distribution of sample 1 (a) and samples 4 (b).

(a)

(b) Figure 13. Energy dispersive X-ray analysis of the chromium (a) and Molybdenum (b) carbides.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

87

The hardness and tensile strength of the samples after each treatment is presented in Table 7. After soaking the samples at cryogenic temperatures, the treatment resulted in increasing the hardness and tensile strength of the samples which is due to transformation of austenite to martensite and also finer shapes of the laths in the microstructure. Lowering the temperature results in lower change in hardness and tensile strength after all the austenite is transformed. This microstructural modification also increases the toughness properties of the alloy. This results in higher impact energies for cryogenically treated samples compared with non-cryogenically treated ones. The study of the fracture surfaces can indicate a quasi cleavage fracture which is typical for tempered steels. A crack can initiate at rough martensite laths or big carbide particles. This leads to propagation of cracks at facets which also leads to a very small plastic deformation at crack interface [31]. Modifying the carbide and martensite distribution thus also changes the fracture behavior after the cryogenic treatment. The most important effect of tempering the deep cryogenically treated samples was improving the wear properties of the alloy. The better distribution of martensite laths along with the more uniform and finer distribution of Table 7. Ultimate Tensile strength and hardness of samples

Sample

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

1 2 3 4

Ultimate Tensile Strength (MPa) 1580 1640 1695 1720

Hardness (HRC)

Impact Energy (J/cm2)

49 51 55 59

15.4 16.1 17.3 18.2

carbides increase the wear properties especially at longer times and higher distances (Figure 14). The improvement of wear properties is more noticeable at the samples worn after 30 hours compared with shorter times, which is probably due to rapid propagation of initial cracks that form at earlier stages of wear. It seems that the smaller and more uniform distribution of carbides plays an important role in inhibiting the formation growth of cracks due to strengthening of the microstructure. This behavior can also be related to the retained austenite present in the structure and its transformation into martensite under cold processing conditions as stated earlier. The amount of untransformed austenite is proportional to the difference between the MS and cryogenic temperature. Scanning electron microscopy images of the worn surfaces are presented in Figure 15. As it could be seen, the cracks grow faster in the alloy without any subzero treatments. For the deep cryogenically treated alloy, although the cracks are formed, the crack propagation rate is much slower compared to noncryogenically treated sample.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

88

Kaveh Meshinchi Asl and Mehdi Koneshloo

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 14. Wear rate values at different times for alloys 1 to 4.

(a)

(b) Figure 15. Scanning electron microscopy images of the worn surface of alloy 1 (a) and 4 (b) after 100h.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

89

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

CONCLUSION 1. Deep cryogenic treatment changes the morphology of β particles in AZ91 magnesium alloy. The coarse divorced eutectic β phase penetrated into the matrix which resulted in the improvement of mechanical properties of cryogenic treated samples compared with as-cast samples. This improvement was attributed to the strengthening of the matrix against propagation of the existing defects 2. In AZ91 alloy the results from creep curves indicate a mixed mode of creep behavior, with some grain boundary effects contributing to the overall behavior. However after deep cryogenic treatment which resulted in modifying the microstructure, dislocation climb controlled creep is the dominant creep mechanism. 3. It seems that deep cryogenic process influences the wear rate especially at higher loads in AZ91 magnesium alloy. The wear resistance of deep cryogenically treated samples improved significantly at higher loads and sliding speeds which was due to the stabilization of internal microstructure and the new morphology of β particles as the main strengthening effect at room temperatures. 4. Cryogenic treatment of A319 Aluminum alloy improves the abrasion resistance of the cryogenically treated samples. The wear tests result in penetration of the abrasive silica particles into the soft  - aluminum matrix. This improvement was attributed to the strengthening of the α-aluminum matrix against propagation of the existing defects. 5. Supplementing the deep cryogenic treatment to conventional heat treatment processes may help to achieve more durable tool steel parts. After the subzero treatments, the retained austenite was transformed to martensite. Applying lower temperatures also led to smaller and more uniform martensite laths in the microstructure. The cryogenic treatment at a very low temperature and long sample holding times, also led to precipitation of more uniform and very fine carbide particles. It seems that applying the process of deep cryogenic treatment initiates nucleation sites for precipitation of fine carbide particles and changes the carbide morphologies in the microstructure which results in enhancement of mechanical properties of the alloy. However the most important effect of tempering the deep cryogenically treated samples was the improving of the wear properties of the H13 tool steel. More investigations are underway to understand thoroughly the mechanisms of beneficial effects of deep cryogenic treatment.

ACKNOWLEDGMENT The authors acknowledge Gary Kaufmann at the School of Materials Science and Engineering of Clemson University for his discussions which were thought provoking and inspiring.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

90

Kaveh Meshinchi Asl and Mehdi Koneshloo

REFERENCES [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26]

Meshinchi Asl, K; Tari, A; Khomamizadeh, F. Materials Science and Engineering A, 523, 27-31. Molinary, A; Pellizzari, M; Gialanella, S; Straffelini, G; Stiasny, KH. J. of materials processing technology, 2001, 118, 350. Huang, JY; Zhu, YT; Liao, XZ; Beyerlein, IJ; Bourke, MA; Mitchell, TE. Materials Science and Engineering, 2003, A339, 241-244. Bensely, A; et al., Materials Characterization, 2007, 58, 485-491. Akhbarizadeh, A;. Shafyei, A; Golzar, MA. Materials and Design, 2009, 30, 32593264. Mohan Lal, D; Renganarayanan, S; Kalanidhi, A. Cryogenics, 2001, 41, 149-155. Zhirafar, S; Rezaeian, A; Pugh, M. Journal of Materials Processing Technology, 2007, 186, 298-303. Kheirandish, S; Noorian, A. Journal of Iron and Steel Research, International. 2008, 15(4), 61-66. Jae-Ho Lee et al., Trans. Nonferrous Met. Soc. China, 2009, 19, 917-920. Chen, ZW; et al., Materials Science and Engineering, 1999, A260, 188-196. Khan, TI; et al., Wear, 2000, 244, 154-164. Khomamizadeh, F; Nami, B; Khoshkhooei, S. Metallurgical & Material Transaction A, 2005, 3489. Wang, Y; Guan, S; Zeng, X; Ding, W. Materials Science and Engineering A, 2006, 416, 109. Wu, G; Fan, Y; Gao, H; Zhai, C; Zhu, YP. Materials Science and Engineering A, 2005, 408, 225. Hirai, K; Somekawa, H; Takigawa, Y; Higashi, K. Materials Science and Engineering A, 2005, 403, 276. Wenwen, D; Yangshan, S; Xuegang, M; Feng, X; Min, Z; Denguan, W. Materials Science and Engineering A, 2003, 356, 1. Mahmudi, R; Ghasemi, HM; Faradji, HR. Heat Treatment of Metals, 2000, 3, 69-72. Meshinchi Asl, K; Tari, A; Khomamizadeh, F. Materials Science and Engineering A, 2009, 523, 1-6. Meshinchi Asl, K; Masoudi, A; Khomamizadeh, F. Materials Science and Engineering A, 2010, 527, 2027-2035. Blau, PJ; Walukas, M. Tribology International, 2000, 33, 573. Chen, H; Alpas, AT. Wear, 2000, 246, 106. Guangyin, Y; Su. Yangshan, Wenjiang, D. Materials Science and Engineering A, 2001, 308, 38. Mehta, DS; Masood, SH; Song, WQ. Journal of Materials Processing Technology, 2004, 155-156, 1526. Mehta, DS; Masood, SH; Song, WQ. Journal of Materials Processing Technology, 2004, 155-156, 1526. Wenwen, D; Yangshan, S; Xuegang, M; Feng, X; Min, Z; Dengyun, W. Materials Science and Engineering A, 2003, 356, 1. Dae Kang, H; Sung Park, S; Nack Kim, J. Materials Science and Engineering A, 2005,

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Effect of Cryogenic Treatment on Microstructure and Mechanical…

[27] [28]

413-414, 555. Bettles, CJ. Materials Science and Engineering A, 2003, 348, 280. Somekawa, H; Hirai, K; Watanabe, H; Takigawa, Y; Higashi, K. Materials Science and Engineering A, 2005, 407, 53. Lim, CYH; Lim, SC; Gupta, M. Wear, 2003, 255, 629. Sharma, SC; Anand, B; Krishma, M. Wear, 2000, 241, 33. Mahmudi, R; Ghasemi, HM; Faradji, HR. Heat Treatment of Metals, 2000, 3, 69-72.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[29] [30] [31]

91

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

In: Cryogenics: Theory, Processes, and Applications ISBN: 978-1-61761-323 c 2011 Nova Science Publishers, Inc. Editor: Allyson E. Hayes, pp. 93-104

Chapter 4

C RYOGENIC T REATMENT AND FATIGUE R ESISTANCE Paolo Baldissera∗ and Cristiana Delprete † Politecnico di Torino, Corso Duca degli Abruzzi, Torino, Italy

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Abstract Among the various applications of cryogenics, the implementation of cold thermal processes with the aim of enhancing mechanical properties of materials is the most attractive from the perspective of structural component engineering design. In particular, considering the key-role played by the fatigue behavior of materials in this discipline, the development of methodologies that allow to achieve longer service life is an evergreen topic, which concerns many fields of application such as energy production and transportation. Starting from the scientific literature of the last 30 years concerning shallow and deep cryogenic treatment (SCT and DCT) and including the most recent experimental results achieved at the Politecnico di Torino, both direct and indirect evidences of an effective influence on the fatigue behavior of steels are pointed-out. Experimental methodologies and data analysis approaches are detailed with a special focus on the estimation of optimal treatment parameters. In particular, two classes of steels that show a good liability for such processes are discussed in depth: austenitic stainless steels and carburized ones. In both cases, the potential consequence in terms of reliability and service life of structural components is noticeable and can be highlighted through practical design examples (stainless steel springs and carburized gears are discussed in details). In the final part of the chapter, an overall picture of the most promising future applications is given with a particular focus on materials beyond steels such as different alloys (i.e. aluminum, magnesium, titanium), polymers and composites and in consideration of their specific fatigue mechanisms.

PACS 81.40.Np, 07.20.Mc, 81.40.Cd. Keywords: Fatigue of cryotreated materials. Key Words: cryogenics, fatigue, precipitation hardening AMS Subject Classification: 74E99, 74P10, 74-05. ∗ †

E-mail address: [email protected] E-mail address: [email protected]

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

94

1.

Paolo Baldissera and Cristiana Delprete

Introduction

The use of cold treatments as a practice to improve the mechanical properties of materials is born at the beginning of the 20th century with the so called subzero treatment, which were performed on tool steels. From the end of the 1990s the interest about the Cryogenic Treatment (CT) is rising with the support of scientific literature production and through the communicative power of internet, where many companies offer such treatments on a wide range of materials and applications claiming almost miraculous effects. A step forward in this practice was determined by the introduction of the treating system developed by Ed Busch (CryoTech, Detroit, MI) in the late 1960s and later improved by Peter Paulin (300 Below Inc., Decatur, IL) with a feedback control on cooling and heating rates, which allows to perform effective and crackless CT until very low temperatures. Hardness and wear durability enhancements given by CT on tool steels were widely investigated at the end of the 20th century [1–5]. Then, researches were focused also on other classes of steel [6–13], on other alloys [14–22] and on composites [23, 24]. Though the mechanism of the claimed changes in ferrous alloys is not fully clarified, different hypotheses and microstructural observations, have been reported in literature: • Complete transformation of the retained austenite into martensite [12, 25, 26]; • Fine dispersed carbides precipitation [12, 26–28];

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

• Changes of the residual stress state [10]. A quite different mechanism is claimed in [29–33] for austenitic stainless steels, where the improvements in fatigue resistance seems to be related to the formation and dispersion of nanosized martensite, which is supposed to reduce the dislocation motion. In general, the effectiveness of CT in changing mechanical properties is no more an hypothesis today, being confirmed by several scientific researches on different classes of steel. As a consequence, there are three main questions for further research: 1. what really happen during the treatment in order to induce the observed changes? 2. which classes of materials are affected by such kind of treatment? 3. which treatment parameters have an influence on the final results and on which material properties are they effective? The questions are focused on causes (1) and on effects respectively (2 and 3), thus needing for investigation by different expertise areas. However, both the theoretical oriented focus (causes) and the practical approach (effects) can contribute to the complete understanding of the CT, by reciprocal information sharing. As each independent research is usually focused on a specific aspect and on a particular material, it is important to produce periodical resumes in order to consolidate the acquired knowledge and to give an overall picture. The aim of this paper is to give such global overview about the relationship between CT and fatigue properties by considering the literature and the author research work.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Cryogenic Treatment and Fatigue Resistance

2.

95

Treatment Parameters

In the field of CT, two main treatment categories are usually considered based on the parameters of the cooling-warming cycle. In [12], two main families depending on the minimum temperature are categorized: • Shallow Cryogenic Treatment (SCT) or Subzero Treatment: the samples are placed in a freezer at about 190 K and then they are exposed to room temperature; • Deep Cryogenic Treatment (DCT): the samples are slowly cooled below 160 K, helddown for many hours and gradually warmed to room temperature. Two typical SCT and DCT temperature profiles are shown in Figure 1 with their respective typical parameter ranges. The main process parameters are: minimum temperature (Tmin ), soaking time, cooling and warming rate. In literature, different values of these parameters are applied on various materials, but anyway it is possible to infer some general indications: • The actual Tmin is higher than the nominal one because of thermal insulation limits. • Soaking time over 36 hours does not lead to significant improvements and, in most cases, 24 hours are enough to obtain results. • Shorter soaking periods are usually employed in SCT, while DCT requires more time to produce results.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

• Cooling rates have a limited range of values (from 0.3 K/min to 1.2 K/min) in order to prevent thermal-shock cracking. • In many cryogenic systems the warming rate is not closely controllable and little importance to this parameter is given in literature despite of some suggested hypothesis about carbides precipitation during the warming phase. A value of 0.25 K/min can be considered as generic reference. In [34], the DCT critical parameters were optimized through the Taguchi Design of Experiment (DOE) method on a 18% Cr martensitic stainless steel. From the Analysis of Variance (ANOVA) of the wear test, the soaking temperature has emerged as the most important factor (72% in contribution), followed by the soaking time (24%) and by the cooling rate (10%). The contribution of the post-DCT tempering process performed after DCT has emerged as non significant. In the same paper, the authors calculate the following optimal combination: 89 K for the soaking temperature, 36 hours for the soaking time, 1 K/min for the cooling rate and 1 hour tempering at 523 K after the DCT. However, these optimal parameters were obtained by focusing on wear resistance and on hardness, which not necessarily lead to a better fatigue behavior. In addition, the post-DCT tempering process temperature could be not significant, but the treatment sequence among DCT and tempering is an important factor that determines the final result, as shown in [7, 8]. Here, the most significant hardness improvement on 18NiCrMo5 carburized steel was obtained by performing the DCT before the final tempering, while the highest tensile strength and fatigue limit were achieved by shifting the DCT at the end of the whole process, after the tempering.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

96

Paolo Baldissera and Cristiana Delprete

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

Figure 1. Typical SCT and DCT treatment profiles.

A methodology for fatigue focused optimization of the CT parameters is being developed at the Politecnico di Torino in the case of a solubilized AISI 302 stainless steel. In general, since it allows to take in accout for both failures and run-outs, the maximum likelihood method is suitable for this kind of purpose. A model for fatigue life estimation must be hypothesized in order to have the number of cycles that depends on the treatment parameters to be analyzed (each one with a weighting factor) and on the stress amplitude. Then, by maximizing the likelihood of the model over experimental data, the weight of each parameter in the model is estimated. At last, the combine of parameters that produce the maximum fatigue life can be evaluated with the obtained model. As the methodology is still under development and requiring a large collection of experimental data, it is early to claim for results, but the first applications are giving encouraging outputs and the authors wish to publish a more detailed and extensive explanation in the next future. Another approach to the choice of the CT soaking temperature is proposed in [31]. Here, the martensite start temperature of an austenitic steel was determined through the acoustic emission technique and the SCT was performed 3 K above such temperature. The obtained fatigue improvements are encouraging, but also in this case no statistical information are reported, resulting in a weak reliability from the engineering design perspective. In general, a preliminary investigation is due in order to identify the optimal parameters for the improvement of a specific property on a definite class of steel.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Cryogenic Treatment and Fatigue Resistance

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

3.

97

Materials and microstructure

All the CT mechanisms highlighted by literature (Ref. Introduction) are related to one of the most important properties of materials from the perspective of engineering design: the fatigue resistance, which involves nucleation and propagation of cracks under cyclic loading conditions. The fatigue of materials is one of the main topics of the 20th and 21th centuries in the fields of mechanical, aerospace and material engineering. The understanding of damage accumulation laws, of crack propagation control and the improvement of the material fatigue resistance play a key-role in most frontier technologies for transportation and for energy production. As perceived by intuition during the 20th century, early fatigue processes act on a sub-grain scale, involving lattice defects such as dislocations and their possibility to slip and to accumulate at grain boundaries under cyclic loading. In this perspective, the CT induced fine precipitation of carbides can play an important role. Despite of that, only few research works were focused on fatigue effects of CT until now. In addition, when dealing with the fatigue of materials, the stochastic nature of this property must not be disregarded, being of primary importance at any scale. At the subgrain scale, the damage accumulation through dislocation pile-up is related to the size and spatial distribution of the precipitates on the slip-plane. At the grain scale, the weakestlink concept deals with lattice orientations with respect to the external load direction and is strictly related to the statistical representation of the whole specimen/component grains. At the macro-scale, the experimental procedures that are usually applied for fatigue testing, such as the stair-case method, reflect its probabilistic nature, which is a direct consequence of the above explained microstructure stochastic. In [9], both SCT and DCT effects on the fatigue behavior of the En 353 carburized steel were evaluated. In this case, the fatigue resistance was increased by balancing the fine carbide precipitation and the retained austenite reduction in the material, as it becomes evident by plotting the results against austenite fraction Figure 2. On the one hand, by lowering the soaking temperature (DCT) the carbide precipitation becomes finer and more homogeneous, with beneficial effects on fatigue crack nucleation. On the other hand, the lower soaking temperature of DCT leads to a reduction in retained austenite on the case hardened layer, which worsen the fatigue behavior. Indeed, it was shown that the presence of such retained austenite is worthwhile since its cyclic transformation during fatigue service causes a redistribution of the compressive residual stresses and generates an increased surface hardness [35]. The balancing of these two opposite effects is of great importance in the perspective of fatigue focused optimization of the CT. In addition, the effects of the two mechanisms have to be evaluated also on the fatigue data scattering, which is a critical point for an overall assessment. The fine carbide precipitation inside the material grain plays an important role in this perspective. As reported in [7], the main achievement of DCT on carburized 18NiCrMo5 is a strong reduction (−80%) of the fatigue limit standard deviation that is measured through stair-case method. As a consequence, the positive contribution of DCT in fatigue life extension is better appreciated at higher reliability levels (i.e. at 99.7% of reliability an increase of 26% in fatigue limit was obtained), which are more significant from the perspective of engineering design. In order to give a quantification of the possible gap between results at different reliability, Figures 3 and 4 show the fatigue data fitting obtained for untreated and cryotreated 18NiCrMo5 steel by considering B50 (50%

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

98

Paolo Baldissera and Cristiana Delprete

Figure 2. Fatigue limit against retained austenite, from the results reported in [9].

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

in reliability) and B10 (90% in reliability) respectively. In the light of the above results, any evaluation of the CT influence on fatigue resistance is to be carefully considered if the statistical aspect of fatigue is not taken into account. Even a fatigue limit improvement calculated at 50% in reliability could be misleading if the related standard deviation is not reported by the authors.

4.

Engineering design applications

Since two of the most frequently cryotreated materials are the carburized steels [7, 9, 13, 36, 37] and the austenitic stainless ones [6, 29, 31, 32, 38], it can be useful to analyze the perspective for the use of CT on their typical application fields. Hereafter, the specific cases of carburized gears and of stainless steel springs will be discussed.

4.1.

Carburized gears

The potential of CT application on transmission manufacturing is widely explained in [39], where both the improvement on tool life for the production of gearings and the increased performances of transmission themselves are discussed. As reported in [40], tooth bending fatigue is the most frequent cause of gear failure in automotive. Referring to UNI 8862 standard, the tooth bending fatigue verification (and design) is performed by comparing the equivalent stress at the tooth root σF with the maximum allowable stress σF P , whose expressions can be resumed as follows: σF =

Ft · Y1 · K b · mn

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

(1)

Cryogenic Treatment and Fatigue Resistance

99

950 900

Smax [MPa]

850 800 750 700

B50 curves

650 600

DCT no DCT

550

5

10

6

N

10

7

10

Figure 3. S-N curves at 50% reliability for the 18NiCrMo5 with and without DCT.

950

850 Smax [MPa]

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

900

800 750 700

B10 curves

650 600 550

DCT no DCT 5

10

6

N

10

7

10

Figure 4. S-N curves at 90% reliability for the 18NiCrMo5 with and without DCT.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

100

Paolo Baldissera and Cristiana Delprete

σF,lim · Y2 (2) SF,min where Ft is the mean tangential load, SF,min is the safety factor, σF,lim is the fatigue limit of the material, b is the tooth face width (usually the design target) and mn is the normal module. Y1 , Y2 and K are given by the product of a series of parameters, which take into account for tooth geometry, surface finishing and loading conditions. Tooth bending will be not dangerous if σF ≤ σF P . For a given tooth geometry, surface finishing and loading condition, and considering a target number of cycles ≥ 3·106, the tooth width only depends on the fatigue limit of the material and on the required safety factor: σF P =

b∝

SF,min σF,lim

(3)

It can be observed that the greater is the reliability level considered for the fatigue limit (i.e. 99.7% as usual in many applications), the lower will be its value, leading to a larger value of b. At the same time, the use of an higher reliability level allows for a smaller safety factor (the polite term for factor of ignorance), as the designer ignorance about the fatigue limit stochastic is already compensated. From this point of view, the reduction in σF,lim standard deviation of carburized steels subjected to DCT allows to strongly reduce the tooth width (and the gear weight) at the same reliability level. A similar approach is given in other standards and methodologies for gear design [41], being the endurance limit at a given reliability level the reference property for the allowable tooth stress. Hence, the above considerations about DCT effects are still valid independently by the applied design procedure. In addition, the DCT induced hardness improvement is usually associated with an increase of the wear resistance, which is another fundamental property for such specific component.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

4.2.

Stainless steel coil springs

Austenitic stainless steels are largely employed in the manufacturing of springs for many applications. Also for these components, cyclic loads are the usual working conditions and the fatigue behavior is of great importance as confirmed by the consolidated use of shot-peening in order increase their endurance [42, 43]. The introduction of CT as a standard spring treatment must be evaluated in the light of the literature results, which currently show conflicting effects from the perspective of this component. On the one hand, in [29, 30] positive effects on the fatigue behavior of such class of steels in hardened condition are claimed, without changes on the tensile strength. On the other hand, such improvement were not confirmed in a recent investigation for the hardened state [44], while a unexpected drop of the elastic modulus was measured after DCT on both hardened and solubilized AISI 302 [45]. Since the stiffness is one of the key-feature in spring design, especially when they are employed in precision valves and mechanisms, a change in the Young’s modulus can lead to important drawback and must be accurately evaluated in the design phase. In addition, these kind of steels usually show a small standard deviation for the fatigue limit and the potential CT effect on such aspect is thus strongly limited with respect to the carburized steel experience. Maybe the specific soaking temperature choice that were proposed in [31–33] could lead to more repeatable results, but the claimed fatigue improvements still need for a wide

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Cryogenic Treatment and Fatigue Resistance

101

statistical confirmation and for an in-depth analysis of the drawback on the elastic modulus. As a consequence, considering the current state of the art, the application of CT on austenitic stainless steel for spring manufacturing looks promising, but not yet ready to be introduced as a standard practice.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

5.

Beyond steel

As modern manufacturing techniques are gradually replacing steels with lighter materials in many applications, it is interesting to evaluate if the CT will have a chance to be effective also on radically different microstructures. By Looking at the literature, some traces of such interest are given by the recent studies on Magnesium [14], on Nickel-Titanium alloys [15,16] and on plastic materials [23,24]. However, these works are mostly focused on wear rather than on fatigue behavior. Concerning the two metallic alloys, it is reasonable the hypothesis of a reinforcement mechanism related to the formation of a fine and homogeneous precipitate phase inside the material grain such as for steels. The fatigue mechanism of these kind of materials is still related to dislocation pile-up at grain boundaries and such kind of precipitate fields should produce similar effects on the nucleation stage. Also in this case, such as for steels, an extensive analysis is needed in order to discriminate facts from hypotheses. In the fatigue behavior of composite plastics, a key-role is played by the interface between fiber and matrix, where crack nucleation mostly occurs. However, in many typical application of such materials, crack propagation is the ruling mechanism and the ability of the structure to respond to the almost unavoidable presence of micro-cracks (toughness) is of great importance. Under this point of view, while it is difficult to hypothesize a reinforcement mechanism though CT over these kind of materials, the hardness increase measured in [23] for some reinforced polymers, looks not really promising in terms of fatigue life extension. Indeed, hardness rising is usually related to a decrease in fracture toughness, which mean a worst fatigue behavior for composite plastics. However, a wide study concerning CT effects on composite fatigue has never been performed and it can not be excluded that other unexpected microstuctural mechanisms would lead to a reinforcement.

6.

Conclusion

Fatigue focused CT literature show that this kind of treatment has the potential to be introduced as a standard practice for some cyclic loaded components such as carburized gears. In particular, the beneficial reduction of fatigue scatter is an important achievement from the perspective of such component reliability design and in-depth investigations about this aspect must be performed in the future in order to better understand worthwhile applications of such treatment. Also stainless steels typically used in the production of springs have shown encouraging results, but some of them still need for confirmation and potential drawbacks on other main properties (i.e. the elastic modulus) suggest to be cautious with claiming positive effects before a more detailed and extensive analysis.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

102

Paolo Baldissera and Cristiana Delprete

The application of CT on modern light materials such as Magnesium and Titanium or fiber reinforced polymers is still in embryonic phase and only weak hypotheses can be made about its potential effects on their fatigue behavior. However, considering the growing interest for such materials, more specific studies are probably under development or, at least, are being planned for the future.

References [1] Barron, R. F. Progress in Refrigeration Science and Technology - Proceedings of the 13th International Congress of Refrigeration , 1973; pp 529–534. [2] Barron, R. F. Cryogenics 1982, 22, 409–413. [3] Meng, F.; Tagashira, K.; Azuma, R.; Sohma, H.; Role, H. Isij Int 1994, 34(2), 205– 210. [4] Barron, R. F. Proceedings of Conference Manufacturing Strategies , 1996; pp 535– 548. [5] Molinari, A.; Pellizzari, M.; Gialanella, S.; Straffelini, G.; Stiasny, K. H. Proceedings of Conference on Advances Materials Processes Technologies, Dublin, 2-6 August 1999, 1999; pp 1461–1469. [6] Baldissera, P.; Delprete, C. Key Eng Mat 2010, 417-418, 793–796.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[7] Baldissera, P. Mater Design 2009, 30, 3636–3642. [8] Baldissera, P.; Delprete, C. Mater Design 2009, 30, 1435–1440. [9] Bensely, A.; Shyamala, L.; Harish, S.; Lal, D. M.; Nagarajan, G.; Junik, K.; Rajadurai, A. Mater Design 2009, 30, 2955 – 2962. [10] Bensely, A.; Venkatesh, S.; Mohan Lal, D.; Nagarajan, G.; Rajadurai, A.; Junik, K. Materials Science and Engineering: A 2008, 479, 229–235. [11] Bensely, A.; Senthilkumar, D.; Mohan Lal, D.; Nagarajan, G.; Rajadurai, A. Mater Charact 2007, 58, 485–491. [12] Bensely, A.; Prabhakaran, A.; Mohan Lal, D.; Nagarajan, G. Cryogenics 2006, 45, 747–754. [13] Preciado, M.; Bravo, P.; Alegre, J. J Mater Process Tech 2006, 176, 41–44. [14] Asl, K. M.; Tari, A.; Khomamizadeh, F. Materials Science and Engineering: A 2009, 523, 27 – 31. [15] Vinothkumar, T. S.; Miglani, R.; Lakshminarayananan, L. J Endodont 2007, 33(11), 1355–1358. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Cryogenic Treatment and Fatigue Resistance

103

[16] Kim, J. W.; Griggs, J. A.; Regan, J. D.; Ellis, R. A.; Cai, Z. Int Endod J 2005, 38(6), 364–371. [17] Lulay, K.; Khan, K.; Chaaya, D. J Mater Eng Perform 2002, 11, 479–480. [18] Chen, P.; Malone, T.; Bond, R.; Torres, P. Proceedings of The 4th Conference on Aerospace Materials, Processes, and Environmental Technology, NASA Center for AeroSpace Information (CASI), February 2001 , 2001. [19] Liu, H.-h.; Wang, J.; Yang, H.; Shen, B.-l. Materials Science and Engineering: A 2008, 478, 324–328. [20] Liu, H.-h.; Wang, J.; Shen, B.-l.; Yang, H.-s.; Gao, S.-j.; Huang, S.-j. Mater Design 2007, 28, 1059–1064. [21] Liu, H.-h.; Wang, J.; Yang, H.-s.; Shen, B.-l.; Gao, S.-j.; Huang, S.-j. Journal of Iron and Steel Research, International 2006, 13(6), 43–48. [22] Yang, H.-S.; Jun, W.; Bao-Luo, S.; Hao-Huai, L.; Sheng-Ji, G.; Si-Jiu, H. Wear 2006, 261, 1150–1154. [23] Indumathi, J.; Bijwe, J.; Ghosh, A. K.; Fahim, M.; Krishnaraj, N. Wear 1999, 225, 343–353. [24] Trieu, H. H.; Morris, L. H.; Kaufman, M. E.; Hood, R.; S, J. L. Proceedings of the 1997 Sixteenth Southern Biomedical Engineering Conference , 1997; pp 90–91. [25] Leskovˆsek, V.; Ule, B. Heat Treat Met 2002, 3, 72–76.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[26] Zhirafar, S.; Rezaeian, A.; Pugh, M. J Mater Process Tech 2007, 186, 298–303. [27] Huang, J. Y.; Zhu, Y. T.; Liao, X. Z.; Beyerlein, I. J.; Bourke, M. A.; Mitchell, T. E. Materials Science and Engineering A 2003, 339, 241–244. [28] da Silva, F. J.; Franco, S. D.; Machado, A. R.; Ezugwu, E. O.; Souza, Jr, A. M. Wear 2006, 261(5-6), 674–685. [29] Singh, P. J.; Guha, B.; Achar, D. R. G. Eng Fail Anal 2003, 10(1), 1–12. [30] Singh, P. J.; Mannan, S.; Jayakumar, T.; Achar, D. Eng Fail Anal 2005, 12(2), 263– 271. [31] Myeong, T.; Yamabayashi, Y.; Shimojo, M.; Higo, Y. Int J Fatigue 1997, 19(93), 69–73. [32] Shimojo, M.; Takashima, K.; Higo, Y.; Inamura, T.; Myeong, H. Metallurgical and Materials Transactions A 2001, 32(2), 261–265. [33] Ianamura, T.; Abe, R.; Myeong, T. H.; Shimojo, M.; Higo, Y. Proceedings of the 7th International Fatigue Congress, Beijing, P R China, 8-12 June 1999 , 1999. [34] Darwin, J.; Mohan Lal, D.; Nagarajan, G. J Mater Process Tech 2008, 195, 241–247. Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

104

Paolo Baldissera and Cristiana Delprete

[35] Jeddi, D.; Lieurade, H.-P. Procedia Engineering 2010, 2, 1927 – 1936, Fatigue 2010. [36] Jordine, A. Proc. IMMA Conf. ‘The Heat is On!’, Melbourne, Victoria, Australia, 2425 May, 1995, pp. 107-111, 1996; pp 418–353. [37] Paulin, P. The Journal of Gear Manufacturing 1993, Mars/April, 26–29. [38] Anooshiravani, D.; Svec, T.; White, K.; Powers, J. J Endodont 1999, 25, 282–213. [39] Alava, L. A. 6th International CTI Symposium - Innovative Automotive Transmissions , 2007. [40] Alban, L. E. Number 1 Gear Failure - Tooth Bending Fatigue , SAE Technical Paper 841088, 1984. [41] Mechanical Engineer’s Handbook ; Marghitu, D. B., Ed.; Academic Press, 2001. [42] Carlson, H. Spring designer’s handbook ; Dekker, NY, 1978. [43] Spring design manual - prepared under the auspices of the SAE Spring Committee ; Committee, S. S., Ed.; Warrendale, SAE, 1996. [44] Baldissera, P.; Delprete, C. Materials & Design 2010, 31, 4731 – 4737.

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

[45] Baldissera, P. Materials & Design 2010, 31, 4725 – 4730.

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

In: Cryogenics: Theory, Processes… Editor: Allyson E. Hayes, pp. 105-122

ISBN: 978-1-61761-323-4 © 2011 Nova Science Publishers, Inc.

Chapter 5

APPLICATION OF FIBER BRAGG GRATING SENSORS AT CRYOGENIC TEMPERATURES Ines Latka, Tobias Habisreuther, and Wolfgang Ecke Institute of Photonic Technology (IPHT), Jena, Germany

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

ABSTRACT Fiber Bragg grating (FBG) sensors are well known means for the measurement of strain and temperature in broad temperature and strain ranges. The most important features of this sensor type are its small size, light weight, full electrical insulation, negligible interaction with electric and magnetic fields, and the flexible fiber leads of good thermal insulation between sensor and its interrogation unit. In particular, the low thermal conductivity of the optical fiber is an advantage when working with low temperatures. Another important feature is the possibility to include several grating sensors in the same fiber, which can later be interrogated simultaneously via wavelength multiplexing. An FBG can be defined as a periodic modulation of the refractive index along a section of the fiber core. Such gratings can be produced by the irradiation of photosensitive silica fibers with an UV laser and an interferometer setup. The FBG reflects a wavelength, which is dependent on the period of the structure and on the effective refractive index of the fiber. Exposure to temperature or strain changes will affect the period of the structure or/and the refractive index, thereby changing the reflected wavelength. A polychromator based measurement guarantees simultaneous measurements for all sensors in the wavelength-multiplexed assembly, with equal measurement duration of typically 100 µs. Thus, it is possible to reconstruct exact strain modes or temperature distributions from the actual multi-point results. The wavelength changes can be monitored with a 1ζ repeatability of about 0.1 pm. While conventional electrical resistance strain gages show increasing crosssensitivities to temperature and magnetic field with decreasing temperature down to liquid helium, it has been found that fiber optic Bragg grating strain sensors show negligible thermo-optic and magneto-optic effects in the cryogenic environment and they allow, therefore, reliable strain measurements. These specific application advantages of optical fiber Bragg grating sensors at low temperatures make them attractive for structural health monitoring of cryogenic devices

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

106

I. Latka, T. Habisreuther, W. Ecke such as superconductive magnets. Other applications are material characterization, e.g. of superconducting materials or of structural elements, respectively. Flux pinning is one of the most fundamental and interesting properties of type II superconductors. In materials with the strongest pinning, the pinning-induced strain can be so large that it can lead to cracks in the material. With the application of FBG sensors, spatially resolved measurements of magnetostrictive effects in superconducting samples become possible. FBGs can also be used for the measurement of thermal expansion coefficients down to 4 K, or for the change of Young‘s modulus with respect to temperature.

INTRODUCTION Fiber Bragg gratings (FBG) are widely used as wavelength-selective dielectric mirrors in the core of an optical fiber. These gratings consist of a periodic refractive index modulation in a short section of the fiber core. The function of such a grating can be understood as the coherent interaction of the Fresnel reflections caused by every single refractive index step. The index modulation is typically achieved by exposure of the fiber core to an intense UV interference pattern [1]. Fiber Bragg gratings have found many applications as wavelengthselective filters in optical telecommunication, laser and sensor technology, e.g., as wavelength demultiplexers [2], laser mirrors [3] or as strain [4] and temperature sensing elements [5].

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

1. FIBER BRAGG GRATINGS – FABRICATION, INTERROGATION, SENSING CAPABILITIES An optical fiber Bragg grating (FBG) can be defined as a periodic modulation of the refractive index along a section of the fiber core (see Figure 1). The FBG acts as a narrow notch filter and reflects a Bragg wavelength λB, which is dependent on the period Λ of the structure and the effective refractive index neff of the fiber: λB = 2 Λ neff

Eq. 1

Applied strain or temperature changes will change the neff and ; the resulting change of the Bragg wavelength is the sensor signal. The sensitivity of B to relative strain L/L is dependent on the effective coefficient p of the photo-elastic interaction in the fiber core [6]: 𝛥𝜆𝐵 = 𝜆𝐵 1 − 𝑝 𝛥𝐿 𝐿

Eq. 2

As a typical value for p in Ge doped silica glass optical fiber, p=0.23 [6] can be assumed. In an operation wavelength range around 830 nm, B amounts to 0.64 pm per applied relative strain of 10-6 (=1µm/m =1µε). The temperature sensitivity varies with the temperature dependence of the refractive index of the fiber (thermo-optic effect) and with the thermal expansion of the fiber (or of its substrate, respectively).

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Application of Fiber Bragg Grating Sensors at Cryogenic Temperatures

107

Figure 1. Sketch of a fiber Bragg grating (Graph: M. Trutzel, DaimlerChrysler Research).

Relative I ntensity

1.00 0.75 0.50 0.25

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

0.00 800

820 840 860 Wavelength [nm] Figure 2. Reflection spectrum of a fiber Bragg grating sensor array with 30 gratings along a single fiber.

The Bragg gratings are produced by irradiation of photosensitive silica fibers by an intense UV laser light fringe pattern. The fringe pattern can be generated in an interferometer setup [7] or by a phase mask [8], respectively. The particular single pulse inscription technique immediately during the fiber drawing process enables the manufacturing of sensor arrays with high numbers of FBG sensors and excellent mechanical strength [8], [9]. For such single pulse gratings, reflectivity values of 30% were achieved [8]. Using other (multi pulse) methods reflectivity of more than 99 % can be reached [8]. The mechanical reliability is dependent on the treatment of the optical fiber for FBG inscription [[9], [10]]. The mechanical strength is known to suffer when the original protective coating is stripped prior to inscription [11]. Single pulse FBGs inscribed during the fiber drawing procedure [9] were shown to have a similar mechanical strength (reversible, elastic strain up to about 6% elongation) as the pristine fiber. An important feature for sensor applications is the possibility to inscribe several FBG sensors in the same fiber (Figure 2), which can then be interrogated simultaneously by wavelength multiplexing (i.e., each Bragg grating is inscribed with a different nominal period Λ, e.g. [12]). The number of gratings, as well as their spectral and geometrical positions can be adapted to the requirements of the structure under inspection. The typical length of the

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

108

I. Latka, T. Habisreuther, W. Ecke

sensing element is about 6 mm. For special applications, the sensing length can be reduced to 1 mm. The sensing signal is the reflected Bragg wavelength of the sensor; its intensity only influences the signal/noise ratio. This spectral interrogation avoids measurement errors by intensity drifts, a common error source for intensity encoded fiber-optic sensors. The fiber distance between the sensor site and signal processing can amount to several kilometers.

2. SENSOR SIGNAL PROCESSING AND MULTIPLEXING

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

We can analyze the Bragg wavelength of FBG sensors by several means. One of the most advantageous setups is based on the illumination of the sensor array by a broadband light source and spectral read-out of the reflected light by a polychromator. The polychromatorbased measurement guarantees simultaneous measurements for all sensors in the wavelengthmultiplexed sensor array, with equal measurement durations of typically 100 µs. Thus, it is possible to reconstruct exact strain mode distributions from the actual multi-point strain results, or to monitor fast temperature variations. This property is in contrast to FBGinterrogators using tunable light sources (e.g. tunable laser [13], [14]) or detectors (e.g. scanning Fabry-Perot [15]). The opto-electronic signal processing in [[16], [17]] is based on broadband super-luminescence diode (SLD) illumination in the 800 nm wavelength range, which allows the use of a cost-effective CCD photo-receiver line in the polychromator (Figure 3). The fiber coupled SLD is connected to a 2x2 fiber-optic coupler and the transmission ports can be connected to the sensor arrays. The reflected light enters the polychromator via the same coupler. During operation, the sensor Bragg wavelengths have to be maintained within the spectral range of the SLD; its full width at half maximum can reach 60 nm (e.g., 805 .. 865 nm). The spectral distance between the sensor Bragg wavelengths must be large enough, approximately 1 nm, in order to avoid their spectral overlap.

Figure 3. Sketch of the signal processing unit. Reproduced from [[18]]: "A fibre Bragg grating sensor system monitors operational load in a wind turbine rotor blade", (K. Schroeder at al., Meas. Sci. Technol., 17(5), 1167-1172, 2006).

Cryogenics: Theory, Processes and Applications : Theory, Processes and Applications, Nova Science Publishers, Incorporated, 2010. ProQuest

Application of Fiber Bragg Grating Sensors at Cryogenic Temperatures

109

The sensor interrogation system in [18] measures changes in the Bragg wavelengths of fiber Bragg grating (FBG) sensor arrays over time within the wavelength range of the SLD from 810-860 nm. The employed FBGs have a spectral width of 0.1 to 0.3 nm and a reflectivity of 30 to 90 %. The interrogation software controls the spectral FBG sensor measurements and performs search and calculation of the sensor reflection peaks (Gaussian peak fit for sub-pixel approximation). The sensor values are transmitted via an Ethernet cable to a PC for their visualization, optionally with adaptive Kalman filtering and storage. Bragg wavelengths are measured with a 1ζ repeatability of 0.1 pm (equivalent to a strain value of about 0.15 µε).

3. FBGS AT LOW TEMPERATURES

Copyright © 2010. Nova Science Publishers, Incorporated. All rights reserved.

As already mentioned in section 2, the Bragg wavelength is sensitive to both temperature and strain. This means that strain measurements have to be corrected for the possible influence by temperature changes ΔT - the so-called cross-sensitivity. Vice versa, temperature measurement results may be subject to strain cross sensitivity. In order to calculate and evaluate the effect of such cross-sensitivities (to temperature, magnetic field, and other measurands), we apply the concept of apparent strain εapp, which presents the erroneous strain value that is obtained by neglecting these cross-sensitivities. This problem of temperature cross-sensitivity is well known for FBG sensors at room temperature [11]. The shift ΔλΒ of the Bragg wavelength λB for FBG sensors depends on the thermo-optic effect δn/n.δT in the quartz fiber of the effective refractive index n, on the linear thermal expansion (temperature coefficient α), and on the applied true strain ε. All strain is weighted in its effect on ΔλB/λB by a factor (1-p), p - effective coefficient of photo-elastic interaction in the fiber core, p ~ 0.20..0.25. ∆𝜆 𝐵 λB

= 𝜀 1 − 𝑝 + Δ𝑇

𝛿𝑛 𝑛 𝛿𝑇

+ 𝛼(1 − 𝑝) = 𝜀 + 𝜀𝑎𝑝𝑝 (1 − 𝑝)

Eq. 3

In a silica glass fiber, the main temperature effect on the Bragg wavelength B originates from the dependence of the effective refractive index neff(T) on temperature since the coefficient of thermal expansion of silica at room temperature is very low,   0.5∙10-6 /K. The thermo-optic effect at room temperature is about B/T  5.5 pm/K (at B  850 nm). It tends towards zero when close to cryogenic temperatures, provided that the protective coating does not add its thermal expansion effect. The temperature-induced apparent strain may amount to several 10 μ per K (which equals a wavelength shift of about 6 pm per K), depending on the thermal elongation of the substrate (Figure 4). A common means of compensation for the temperature effect is the use of an additional, strain-decoupled FBG temperature sensor, as demonstrated in [17]. As the temperature decreases, so does the temperature sensitivity of the FBG sensors. Below 40 K the sensitivity of a silica fiber Bragg grating to temperature drops to