A NOVEL MAGNETORHEOLOGICAL VALVE WITH MEANDERING FLOWPATH STRUCTURE

FITRIAN IMADUDDIN

UNIVERSITI TEKNOLOGI MALAYSIA

Replace this page with form PSZ 19:16 (Pind. 1/07), which can beobtained from SPS or your faculty.

“We hereby declare that we have read this thesis and in ouropinion this thesis is sufficient in terms of scope and quality for the

award of the degree of Doctor of Philosophy”

Signature :Name : Associate Professor Ir. Dr. Saiful Amri MazlanDate :

Signature :Name : Dr. Hairi ZamzuriDate :

Replace this page with the Cooperation Declaration form, which can beobtained from SPS or your faculty.

A NOVEL MAGNETORHEOLOGICAL VALVE WITH MEANDERING FLOWPATH STRUCTURE

FITRIAN IMADUDDIN

A thesis submitted in fulfilment of therequirements for the award of the degree of

Doctor of Philosophy

Malaysia-Japan International Institute of TechnologyUniversiti Teknologi Malaysia

JUNE 2015

ii

I declare that this thesis entitled “A Novel Magnetorheological Valve with Meandering

Flow Path Structure” is the result of my own research except as cited in the references.The thesis has not been accepted for any degree and is not concurrently submitted incandidature of any other degree.

Signature :Name : Fitrian ImaduddinDate :

iii

To my father, my mother, my wife and my brothers

iv

ACKNOWLEDGEMENT

I wish to praise the Almighty Allah the Most Gracious for the blessing andstrength that have been given to my life. My deepest gratitude goes firstly to my mainsupervisor, Associate Professor, Ir. Dr. Hj. Saiful Amri bin Mazlan for his tremendoussupport during my entire study. His intensive encouragement, enthusiasm and guidancehave made me able to pass through this process easier. I also would like to thankmy co-supervisor Dr. Hairi bin Zamzuri for his advice, interest and support to mywork. I must also express my gratitude to the Malaysia-Japan International Instituteof Technology (MJIIT) for the financial support provided during my study though theMJIIT scholarship. I also would like to thank the Lembaga Pengelola Dana Pendidikan(LPDP) for providing the additional incentive during the completion of my thesiswriting.

Appreciation is also given to the faculty members and colleagues in the VehicleSystem Engineering (VSE) research laboratory, especially Jamal, Yasser and Izyanfor helping me a lot during my earlier time in the university. Thanks also to Mr.Hairullail, Madam Aishah and the remaining MJIIT staffs that have been supportive tome during my study. Particular credit is also given to my Indonesian friends in UTMKuala Lumpur especially Ubaidillah and Burhanuddin for being my family abroad. Iwould also like to thank Aizzat and the developers of the utmthesis LATEX project formaking the thesis writing process a lot easier for me. Special acknowledgement goes tomy previous supervisors Dr. Khisbullah Hudha and Dr. Gunawan Nugroho. They haveinfluenced me with the passion and love to the scientific research. I regret that I cannotmention all the valuable names here, but I believe and pray that Allah will reward allthe good deeds that have been given to me.

Lastly, I would like to express my sincerest gratitude to my parents, Dr. AhmadDahlan and Umi Sholichatin for unlimited love, support, trust and pray that havebrought me to this level. My two little brothers, Zamzam Ibnu Sina and AllahyarhamGhilman Hunafa, for being such a good role model for me. Last but not least, mydearest wife, Vivi Diawati, for the pray, patience, love and understanding that havemade me through this journey without any hesitations.

v

ABSTRACT

The development of a new Magnetorheological (MR) valve with meanderingflow path as a new approach to improve the MR valve performance is presented inthis research. The meandering flow path was formed by the arrangement of multipleannular and radial channel so that the total effective area in an MR valve can beincreased without compromising the size and power requirement of the valve. Themain objective of this research is to explore the achievable pressure drop of theMR valve with meandering flow path. This research was started with the conceptdevelopment where the meandering flow path structure is analytically modeled andnumerically simulated to predict and analyze the effect of variables involved. Theprediction results showed that the meandering flow path structure is able to increasethe achievable pressure drop of an MR valve significantly. The gap size analysisshowed that the size of annular gaps mainly contributed to determine the viscouspressure drop component. Meanwhile, the field-dependent pressure drops were mainlydetermined by the size of radial gaps. The prediction results of the concept was alsoassessed and confirmed by the experimental work using a dynamic test machine. Basedon the experimental data, two hysteresis models, namely the polynomial model andthe modified LuGre model, were developed to model the hysteresis behavior. Theassessment results of the hysteresis models indicated that both model were able toreplicate the hysteresis behavior. However, the modified LuGre model, though 9.5%less accurate than the polynomial model, was showing better consistency in a widerrange of input values. In general, the new concept contributes in the development ofa new type of MR valve that could achieve pressure drop nearly three times than theannular, radial and annular-radial type MR valve.

vi

ABSTRAK

Pembangunan konsep baru injap reologi magnet (MR) dengan menggunakanlaluan aliran yang berliku-liku sebagai pendekatan baru untuk meningkatkan prestasiinjap MR dibentangkan dalam kajian ini. Laluan aliran yang berliku-liku dibentukmelalui beberapa susunan saluran gegelang dan tebaran jejari secara berurutansupaya jumlah kawasan yang berkesan di dalam injap MR boleh ditingkatkan tanpamenjejaskan saiz keseluruhan dan prestasi injap. Tujuan utama kajian ini adalah untukmeneroka kebolehcapaian nilai susutan daripada injap MR dengan menggunakanlaluan aliran yang berliku-liku. Kajian ini bermula dengan pembangunan konsep,di mana injap dengan laluan aliran yang berliku-liku dimodelkan secara analitikaldan disimulasikan secara berangka untuk meramalkan prestasi injap dan juga untukmengambil kira kesan pembolehubah yang terlibat. Keputusan simulasi menunjukkanbahawa konsep injap dengan laluan aliran yang berliku-liku mampu meningkatkankebolehcapaian yang ketara dari segi nilai susutan tekanan daripada injap MR.Berdasarkan kepada analisis saiz saluran telah dijalankan, hasil menunjukkan bahawasaiz saluran gegelang lebih menyumbang kearah menentukan komponen kelikatandari susutan tekanan manakala komponen susutan tekanan akibat medan magnetditentukan terutamanya oleh saiz saluran dari tebaran jejari. Konsep ini turut dinilaimelalui kerja eksperimen menggunakan mesin ujian dinamik, yang telah mengesahkankeputusan yang diramalkan oleh simulasi. Berdasarkan data eksperimen, dua modelhisterisis, iaitu model polinomial dan model LuGre yang telah diubahsuai, telahdibangunkan untuk mengilustrasikan tingkah laku histerisis injap MR. Keputusanpenilaian model histerisis menunjukkan bahawa kedua-dua model dapat mereplikasiciri-ciri histerisis daripada injap MR. Walau bagaimanapun, model LuGre yangtelah diubahsuai, walaupun 9.5% kurang tepat berbanding model polinomial, telahmenunjukkan konsistensi yang lebih baik dalam pelbagai ruang lingkup data masukanyang lebih besar. Secara umumnya, konsep baru injap MR ini dapat memberikanpendekatan baru dalam membangunkan sebuah injap MR yang dapat meningkatkankebolehcapaian susutan tekanan sehingga tiga kali ganda berbanding injap MR jenisgegelang, jejari dan gegelang-jejari.

vii

TABLE OF CONTENTS

CHAPTER TITLE PAGE

DECLARATION iiDEDICATION iiiACKNOWLEDGEMENT ivABSTRACT vABSTRAK viTABLE OF CONTENTS viiLIST OF TABLES xLIST OF FIGURES xiLIST OF APPENDIX xiv

1 INTRODUCTION 11.1 Introduction 11.2 Motivation of Study 21.3 Research Objectives 61.4 Research Scope 61.5 Significance of Research 71.6 Outline of Thesis 7

2 LITERATURE REVIEW 92.1 Introduction 92.2 Magnetorheological Fluid 9

2.2.1 Composition of MagnetorheologicalFluid 10

2.2.2 Operational Modes of Magnetorheologi-cal Fluid 11

2.3 Magnetorheological Valve 142.3.1 Annular Magnetorheological Valve 152.3.2 Radial Magnetorheological Valve 18

viii

2.3.3 Combination of Annular and RadialMagnetorheological Valve 20

2.4 Experimental Assessment Method for Magnetorhe-ological Valve 20

2.5 Modeling Approach for Magnetorheological Valve 222.5.1 Steady-state Model 232.5.2 Dynamic Model 27

2.6 Utilization of Magnetorheological Valve 302.6.1 Linear Magnetorheological Damper 312.6.2 Rotary Magnetorheological Damper 352.6.3 New Magnetorheological-based Actua-

tors 392.7 Summary of Chapter 2 41

3 MAGNETORHEOLOGICAL VALVE CONCEPT 423.1 Introduction 423.2 Design of Magnetorheological Valve 42

3.2.1 Conceptual Design 433.2.2 Design Consideration 453.2.3 Magnetic Simulation 47

3.3 Steady-state Modeling of MagnetorheologicalValve 50

3.4 Simulation of Magnetorheological valve Perfor-mance 54

3.5 Summary of Chapter 3 61

4 EXPERIMENTAL ASSESSMENT 624.1 Introduction 624.2 Experimental Apparatus 62

4.2.1 Magnetorheological Fluid 624.2.2 Magnetorheological Valve 644.2.3 Testing cell 67

4.3 Experimental Set-up 684.4 Experimental Results 71

4.4.1 Off-state and On-state Pressure DropCharacteristics 71

4.4.2 Effect of Gap Size 754.4.3 Effect of Current Input Variation 76

ix

4.4.4 Effect of Excitation Frequency Variation 794.5 Summary of Chapter 4 80

5 HYSTERESIS MODELING OF MAGNETORHEO-LOGICAL VALVE 825.1 Introduction 825.2 Polynomial-based Hysteresis Modeling Approach 825.3 Modified LuGre-based Hysteresis Modeling Ap-

proach 875.4 Model Performance Comparison 905.5 Summary of Chapter 5 95

6 CONCLUSIONS AND RECOMMENDATIONS 966.1 Conclusions 96

6.1.1 The New Magnetorheological Valve Con-cept 96

6.1.2 Gap Size Selection Effect 976.1.3 Experimental Assessment of Magne-

torheological Valve Performance 976.1.4 Hysteretic Modeling of Magnetorheolog-

ical Valve 986.2 Contributions of the Research 996.3 Open Problems and Recommendations for Future

Works 1006.3.1 Pressure Tracking Control System 1006.3.2 Other Open Problems 103

REFERENCES 105Appendix A 120 – 129

x

LIST OF TABLES

TABLE NO. TITLE PAGE

3.1 Materials selection of valve component for valve routing 463.2 List of MR valve parameter 533.3 Performance benchmarking between the proposed MR valve

concept and the counterparts 604.1 Typical properties and material compatibility of MRF-132DG 634.2 The variable arrangement of experimental test using

Shimadzu Fatigue Dynamic Test Machine 715.1 Correlation test results between the model coefficient a and

current input I 855.2 List of coefficients for the polynomial-based parametric MR

valve model 875.3 List of approximated function for different parameters 905.4 Comparison of relative error at 0.75 Hz frequency excitation 915.5 Comparison of relative error at 0.50 Hz and 1.00 Hz

frequency excitations 94

xi

LIST OF FIGURES

FIGURE NO. TITLE PAGE

2.1 Movement of magnetic particles in the MR fluid with andwithout magnetic field 10

2.2 Shear mode 122.3 Valve mode 132.4 Squeeze mode 132.5 Magnetic Gradient Pinch mode 142.6 MR throttle valve 162.7 C-shaped pressure control valve 162.8 Three port MR valve 172.9 Double-coil annular MR valve 182.10 Basic structure of single stage radial MR valve 192.11 Two-way controllable radial MR valve 192.12 Annular-Radial MR Valve 202.13 Typical arrangement of constant flow assessment method 212.14 Typical arrangement of variable flow assessment method 222.15 Illustration of significant variables in an MR valve 252.16 Bouc-Wen model 292.17 Parametric hysteretic polynomial model 302.18 Artificial Neural Network model 312.19 Valve mode MR damper 322.20 Shear mode MR damper 322.21 External coil MR damper 332.22 MR damper with bifold valves 332.23 Bifold MR damper for high impulsive loads 342.24 Bifold MR damper for shock vibration mitigation 342.25 Basic structure of Bypass MR damper 352.26 Bypass MR damper for large scale seismic application 352.27 Vane type MR damper with arc valve 362.28 Vane type MR damper with outer coil valve 372.29 Vane type MR damper with inner coil valve 38

xii

2.30 MR based bellow-driven motion control 392.31 MR hydraulic power actuation system 392.32 Actuation with embedded Terfenol-D pump 402.33 MR based link manipulator 402.34 4/3 way directional MR valve (citation) 413.1 Concept assessment sequence of the new MR valve concept 443.2 Basic concept of MR valve with meandering flow path 453.3 Approximation of yield stress as a function of magnetic flux

density, B 483.4 Two-dimensional axisymmetric model of the MR valve in

FEMM 483.5 Flux lines and contour of magnetic field of the MR valve in

FEMM 493.6 Magnetic flux density along MR fluid flow path for 0.5 mm

gap size with respect to various current input 503.7 Gaps zone in MR valve with multiple annular and radial gaps 523.8 Dimension and variables of MR valve 543.9 Estimation of achievable pressure drop of MR valve with 0.5

mm gap size 553.10 Percentage of pressure drop contribution from each zone (a)

viscous (b) field-dependent at 1 A current input 553.11 Effect of gap size on the pressure drop (a) viscous (b) field-

dependent at 1 A current input 573.12 Comparison of operational range between various gap

configurations 594.1 B-H curve of the MRF-132DG 634.2 Field induced yield stress of the MRF-132DG 644.3 Exploded view of the MR valve prototype 654.4 Failure of the bolt-locking mechanism to withstand internal

pressure 654.5 Modification and comparison of the MR valve prototype(a)

Exploded view of the MR valve design (b) Fabricatedprototype of MR valve 66

4.6 MR valve installation in the testing cell 674.7 Experimental arrangement schematic for the MR valve testing 694.8 MR valve testing set-up using in-house test machine 704.9 Testing cell installation in the Shimadzu Fatigue Dynamic

Test Machine 71

xiii

4.10 Comparison of measured and theoretical off-state peakpressure drop at various flow rates for 0.5-0.5 mm (annular-radial) gaps configuration 72

4.11 Comparison of measured and theoretical on-state peakpressure drop at various current inputs and flow rates for 0.5-0.5 mm (annular-radial) gaps configuration 73

4.12 Comparison of measured peak pressure drop in various gapsize combinations 76

4.13 Comparison of the MR valve dynamic range for each gap sizecombinations 77

4.14 The pressure dynamics of MR valve at various current inputfor 0.5-0.5 mm (annular-radial) gaps configuration, (a) 0.50Hz (b) 0.75 Hz (c) 1.00 Hz 78

4.15 The trend of peak pressure drop at various current input for0.5-0.5 mm (annular-radial) gaps configuration 79

4.16 The pressure dynamics of MR valve at current input of 1A atvarious frequency excitation for 0.5-0.5 mm (annular-radial)gaps configuration 80

5.1 The difference between MR damper model and MR valvemodel excitation 83

5.2 Trend of the normalized coefficient values at the positive flowacceleration (lower loop) to the variations of current input (a)a6, a5 and a4 (b) a3, a2, a1 and a0 85

5.3 Trend of the normalized coefficient values at the negative flowacceleration (upper loop) to the variations of current input (a)a6, a5 and a4 (b) a3, a2, a1 and a0 86

5.4 Trend of estimated parameters with respect to current input 905.5 Comparison between the test data and the model results for

various current input, (a) 0.3 A (b) 0.6 A (c) 0.9 A 925.6 Comparison between the test data and the model results for

current input of 1.0 A at various frequency excitations, (a)0.50 Hz (b) 1.00 Hz 93

6.1 Basic structure of pressure tracking control of MR valve 1006.2 Simulation results of pressure tracking control under various

functions as reference, (a) Sinusioidal (b) Pulse (c) Saw-tooth 102

xiv

LIST OF APPENDIX

APPENDIX TITLE PAGE

A CAD Drawings 120

CHAPTER 1

INTRODUCTION

1.1 Introduction

Magnetorheological (MR) fluid is one of the fluids in the class of fieldresponsive material [1, 2], that has sensitive rheological properties to magnetic field[3–7]. The development of the fluid, together with the progressing research in theunderstanding of its behavior, has convinced researchers and engineers that MR fluidis a promising material for future applications [8–10]. This is because of their adaptiveforce capacity and their inherent ability to provide a simple, fast and robust interfacebetween electronic controls and mechanical components. The fluid was first introducedin Rabinow’s Magnetic clutch in 1948 [11] and has gained in popularity since enteringthe automotive market. MR fluid is very responsive to magnetic field, with an estimatedresponse time of less than 10 milliseconds [12], and requires relatively low power tooperate. The advantages of MR fluid have created great interest in MR based devicedevelopment in a wide range of applications.

One of the most popular devices that utilized the unique characteristics ofMR fluid is MR damper [13], which has been commercially available for high-end passenger vehicles as a semi-active suspension or adjustable suspension [14].The working principle of an MR damper is basically similar to a conventionalviscous damper which employs flow restriction concept to generate damping. Theflow restriction in a conventional viscous damper is normally generated by an orificechannel which act as the valve. Since the gap of the orifice channel is fixed, the flowrestriction that can be generated by the valve of the conventional viscous damper isalso fixed. The MR dampers use different approach by employing MR fluids as itsworking fluid and an MR valve in its flow restriction mechanism. Although the gapsize of the channel in an MR valve also can be fixed, the magnetic field strength in theflow channel of the MR valve can be regulated [15]. Therefore, the flow of MR fluid

2

that pass through the MR valve can be controlled without having to modify the gapsize of the channel. On the other hand, it can be said that the performance of the MRvalve to generate flow restriction highly determines the overall performance of the MRdamper.

Considering the importance of MR valve, many designs of MR valve havebeen proposed. One of the earliest design of stand-alone MR valve was proposed inKordonski et al. [16] which later elaborated by Gorodkin et al. [17]. In the literatures,annular MR valve designs with optimize-able geometry and controllable MR fluid flowresistance were provided. A simpler concept of annular MR valve was proposed byYokota et al. [18], which consisted of annular flow channel and electromagnetic coilinstalled in adjacent to the flow channel. The works were improved by Yoshida et al.[19] by proposing a three-port annular MR valve using permanent magnet. In the sametime, a meso-scale (less than 25 mm outer diameter) annular MR valve were proposedby Yoo and Wereley [20] using internal double coils with counter flux direction. Whilethe advancement of annular type MR valve were continuously explored, Wang etal. [21] started to discuss about the radial type MR valve for the large-scale seismicbypass damper configuration. The benefit of radial MR valve over annular MR valvein terms of pressure drop rating as well as the benefit of external bypass MR valveconfiguration was compared in the literature. Performance assessments of MR valvewere also performed by Grunwald and Olabi [22] through the performance analysisof the annular and orifice type MR valve. The discussions of MR valve type wereextended by Ai et al. [23] and Wang et al. [24] through an MR valve design with bothannular and radial flow path. In their design, both type of resistance channel were usedin an MR valve to increase the on-state resistance force while maintaining valve sizeand power consumption. In order to make an MR valve more applicable to retrofitgeneral hydraulic applications, Yoo and Wereley [25] introduced the installation ofmultiple MR valves in H-bridge configuration to actuate a hydraulic cylinder. The workthen followed by John et al. [26] with the embedded version of H-bridge MR valve andby Salloom and Samad [27] with the introduction of 4/3 way MR valve design.

1.2 Motivation of Study

MR damper for semi-active vehicle suspension systems are among the mostpopular and commercially successful MR fluid devices [28–34]. In general, vehiclesuspension system can be divided into three categories; passive suspension system,semi-active suspension system, and active suspension system [35]. Passive suspension

3

system is the common suspension system installed in most vehicle nowadayswhich typically consists of spring and damper in parallel configuration. Semi-activesuspension system is similar with passive suspension system but the stiffness of thecomponent (spring and/or damper) can be controlled to suit the desired ride or handlingperformance [36, 37]. Active suspension system, on the other hand, is the suspensionsystem with the involvement of active actuators such as hydraulic [38], pneumatic [39]or electro-mechanic [40,41], which could provide external force to the suspension. MRdampers are usually implemented as a semi-active device to retrofit hydraulic dampersto enhance passive suspension performance. Enhancement of suspension performanceis feasible since the performance limitations of passive suspension system occurred dueto a fixed stiffness value of the spring and damper. In this case, MR damper, in contrastto conventional linear hydraulic damper, has the capability to change its dampingstiffness by varying the magnetic field strength inside the damper. Together withembedded control system, MR dampers have gained popularity and proved its potentialto enhance the performance of suspension systems. Aside of dampers, other types ofMR devices have been developed to meet other automotive application demands suchas engine vibration suppressors [42–45], seat suspensions [46–49], brakes [50–53] andclutches [54–57].

According to the location of the valve, the MR damper can be divided into twogroups, the MR damper with internal valve and the MR damper with bypass valve.The MR damper with internal valve typically has MR valve embedded in the pistonof the damper, similarly with the configuration of the valve in a conventional viscousdamper. This configuration is the most common valve installation in an MR dampersince it is neat and compact. However, the internal valve configuration is not withoutsetback. The disadvantages of internal valve configurations are mainly in the spacelimitation of MR valve installation, the complexity of wiring and in the risk of thermalbuild-up from the immersed valve. The MR valve integration to the piston is the mainreason why the construction of the MR damper with internal valve can be really neatand compact. However, since the available space inside the cylinder is very limitedand the MR valve requires sufficient space for electromagnetic coil and magnetizationchannel, the performance range of the damper is very narrow. Moreover, since thecoil is embedded with the piston, the common way of wiring installation is normallymade through the conduit along the rod, which made it prone to leakages and tendsto be costly for fabrication. On the other hand, the heat dissipation, as a result ofkinetic energy conversion into heat, can be more severe in an MR damper than in aconventional viscous damper because the magnetically altered damping stiffness willdefinitely increase the heat dissipation. In the case where the MR valve is immersed inthe MR fluid, the heat dissipation from the MR valve will have to disperse to the MR

4

fluid first, which responsible in the increase of fluid temperature, before eventuallyreleased to the environment. The experimental observation conducted by [58] reportedthat the temperature rise of MR fluid in an MR damper after 400 s of operation atcurrent input of 2 A and frequency excitation of 6 Hz have caused the damping force todegrade in about 38%. However, the same experiment observed that less degradationcan be achieved if the MR damper is properly finned, whereas increase the thermalrelease to the environment.

The practice in the other type of MR damper, known as the bypass MRdamper, is not embedding the valve in its piston since the construction uses nofluid channel in the piston [59]. In the bypass MR damper, the fluids flow betweenchambers through the bypass channel outside the cylinder where MR valve is installed.Therefore, the valve in the bypass MR damper configuration is easier to be installed andmaintained since the construction of the main cylinder is similar with the structure of aconventional hydraulic cylinder. The bypass MR damper is also less prone to over-heatbecause the valve is already located outside the cylinder. Various types of MR valvealso can be implemented in an MR damper with bypass configuration because the valvesize is no longer constrained by the cylinder size nor the piston size. However, theexistence of bypass channel and MR valve outside the cylinder are obviously makingthe bypass construction not as neat and compact as the damper with internal valve.The MR damper with bypass configuration is also difficult to be installed in space-constrained applications since the bypass damper requires more room than the damperwith internal valve. With these characteristics, the bypass configuration is normallyimplemented in the large scale MR damper with high energy dissipation [60–62].

Despite the advantages and disadvantages of each MR damper structure,the technological advancement of the MR valve, as the heart of the MR damperperformance, is not as extensive as the advancement of the MR damper. Regardlessthe types of MR damper, most of them are still using the same MR valve concept. Theonly differences are probably the size, coil configuration and/or MR fluid types. Mostof MR dampers are employed with annular type MR valve as one of the most populartypes of MR valve [29, 46, 49, 63–68]. The annular MR valve is the first generationof MR valve that utilized the annular channel as the effective area. The effective areais the area where the MR fluids are exposed to magnetic flux perpendicularly to theflow direction. There are several variants of annular MR valve that has been proposedby researchers [16–20, 69], but the main concept is basically similar. The annular MRvalve is popular because it is simple to be manufactured and has a high ratio betweenthe on-state and off-state performance. However, the effectiveness of space utilization

5

in the conventional annular MR valve is very low because not all areas of the annularchannel can be utilized as the effective area. Therefore, any improvement effort onthe annular valve performance will typically tend to increase the valve size either inlength, by enlarging the effective area, or in diameter, by enlarging the electromagneticcoil. Thus, in a constrained space device such as in the MR damper with internal valveconfiguration, the desired performance improvements are sometimes difficult to beachieved.

Due to the limitation of the annular MR valve, another type of valve, known asthe radial MR valve, was introduced by [21]. The radial MR valve, as a distinction fromannular MR valve, has radial flow channel inside the valve and utilize it as the effectivearea. The utilization of radial channel as the effective area offers several benefits thanthe effective area of the annular channel, especially in terms of area efficiency sincethe radial channel can be made in multi-stage configuration. Therefore, with multi-stage capability of the radial MR valve, the performance improvement of radial valvetypically has lower implication to the valve size than the one in the annular MR valve.As a result, the radial valve concept has been installed to serve several concepts oflarge scale MR dampers [61, 62, 70, 71]. Recently, another concept of MR valve alsohas been developed by combining both annular and radial valve concept in a singlevalve [23,24]. The combination of both annular and radial channel in an MR valve hasbeen proven effective to improve the performance of MR valve. It has been reportedby [72] that the MR valve with combination of annular and radial channel has higherachievable pressure drop than annular valve with lower power consumption althoughat the cost of lower valve ratio. The MR valve with combination of annular and radialchannel also has been implemented in MR mount design [42] and MR damper design[73].

From these explanations, it can be observed that the technological advancementof an MR valve has a significant impact to the advancement of other MR devices.Therefore, particular explorations of the MR valve concept are necessary as a basisto provide knowledge on how to improve the performance of MR devices in general.The concept explorations is not limited to the geometrical and design arrangements ofthe MR valve, but also in terms of behavioral characteristics of the MR valve such asthe identification of the MR valve hysteretic behavior. The hysteretic behavior, as wellas other complex characteristics, in generally in any MR devices is still considered achallenging problem in terms of the modeling technique and controller design [14,74].The hysteresis could be occurred due to magnetic field remnant in steel elementsand due to the viscoelastic properties of MR fluid itself. In terms of controllability,

6

hysteresis behavior is a disadvantage since the controller will face difficulties to trackthe damper behavior. For example, according to Wang and Liao [74], tracking ability ofdamping force is one of the highly important issues that should be considered in orderto get an accurate MR damper controller. However, a controller with such capabilitywill tend to be more complex, require more computational resources, be costly andless robust. Therefore, innovation in the control design is also vital to support thefinal implementation of MR devices. Innovation of the control algorithm will be moredifficult if the model that is used in the controller design phase is not able to simulatethe hysteresis phenomenon accurately. A simple and accurate model of an MR valve, inparticular, is needed in order to design an appropriate controller with good robustness,stability and reliability. Therefore, the advancement of modeling technique that havethe ability to accommodate the hysteretic behavior of MR valve is as important as theadvancement of the MR valve concept.

1.3 Research Objectives

This study embarks on the following objectives:

(a) To develop a new concept of MR valve with meandering flow path to improvethe achievable pressure drop.

(b) To analyze the effect of gap size selection to the achievable pressure drop of MRvalve.

(c) To assess the performance of MR valve using experimental work.

(d) To model the hysteretic behavior of the MR valve.

1.4 Research Scope

In this research, a new concept of MR valve will be investigated. This studyfocuses on the elaboration of a new concept of MR valve utilizing the combination ofmultiple annular and radial gaps that formed a meandering flow path. The new conceptof MR valve is introduced to provide an adjustable pressure drop with a high on-statelimit. In order to examine the capability of the MR valve, the steady-state model ofthe new MR valve concept is derived and the magnetic circuit performance of theMR valve is simulated using Finite Element Method Magnetics (FEMM) software

7

package. The performance of MR valve, in this study, is only evaluated in terms ofthe achievable pressure drop as a function of gap size of the flow channel, magnitudeof current input charged to the coil, and fluid flow rate. This research is also coveringthe experimental evaluation of the MR valve using an MR valve testing cell in variableflow rates, to reveal the hysteretic behavior, with constant current inputs. The measuredperformance of the MR valve is also used to model the hysteretic behavior of the MRvalve, which is not covered in the steady-state model. However, the optimization of theconcept is not discussed in this research and the demonstration of control applicationis only performed.

1.5 Significance of Research

The significance of this research is mainly in terms of general advancementof MR devices and applications especially to answer the demand of smart, simple yethigh performance and reliable new MR devices. The new concept of MR valve withmeandering flow path is expected to provide a new method to improve the design of anMR valve, which will highly influence the design of MR damper as well as other MRbased actuators. Moreover, the concept is expected to be performed as a demonstrationof a generic MR device that can suit various applications. Therefore, the concept canbe anticipated as a modular and re-sizable device so that the range of operation and thecapacity can be adjusted. The significances of this research are summarized as follows:

(a) This study demonstrates a new concept of MR valve using a meandering flowpath structure.

(b) This research provides knowledge of the effect of gap size selection to theachievable pressure drop of the valve which will be further useful for valve sizingprocess.

(c) The hysteretic modeling process of the MR valve introduces a new modelingapproach of MR valve using modified LuGre hysteresis model.

1.6 Outline of Thesis

This thesis is organized in six chapters. Each respective chapter in thisthesis ends with a brief summary outlining the achievements and findings that were

8

established in the chapter. The outline of this thesis is organized as shown:

Chapter 2 covers the theoretical background, which includes the propertiesand the working modes of MR fluids, the basic knowledge of MR valves, the recentadvancement of MR valves, as well as the applications of MR valves.

Chapter 3 explains the new concept of the MR valve with meandering flowpath, the design consideration for the performance assessment, the steady-state modelderivation, the magnetic simulation as well as the performance prediction of the newMR valve with respect to various dependent variables.

Chapter 4 elaborates the experimental evaluation of the MR valve including thedescription of the experimental setup, the experimental procedure and the analysis ofthe experimental results.

Chapter 5 presents the development of two different hysteresis MR valvemodels, the parameter estimation processes and the performance comparison of thesetwo models.

Chapter 6 concludes the work and presents the achieved contribution of theresearch as well as recommends open problems for future work.

CHAPTER 2

LITERATURE REVIEW

2.1 Introduction

This chapter provides an overview of magnetorheological (MR) fluids, theworking modes of MR fluids, the MR valve design and modeling as well as theapplication range of MR valves. The overview is presented in order to elaborate theresearch background about the main idea presented in this research. The first part ofthis chapter presents some of the main issues related to the MR fluids compositionand its working mode. The second part of the chapter elaborates the fundamentals andclassifications of the MR valves. The third part of this chapter discussed about themodeling approach. The final part of the chapter provides the review of the currentapplication that involved MR valves in its working mechanism.

2.2 Magnetorheological Fluid

The MR fluid was first discovered in the Rabinow’s magnetic fluid clutch in1948 [11] and has been classified as one of the class of smart materials [9]. MR fluid isclassified as smart material because of its adaptive behavior and their inherent ability toprovide a simple, fast and robust interface between electronic controls and mechanicalcomponents. When the fluid is subjected to magnetic field, the iron particles startaligning along the magnetic fluxes, as shown in Figure 2.1. Hence, the movementof the fluid is restricted by particle chains thus increasing its apparent viscosity.Consequently, the fluid changes its state from liquid to a viscoelastic solid dependingon the strength of the magnetic field. The advantages of MR fluid have created greatinterest in MR based device development in a wide range of applications such as incivil applications [60, 61, 63, 75–80] medical prosthetic applications [81–84], groundvehicle applications [28,30,31,34,55,85], aerospace applications [86–88] and military

10

applications [33,89–91]. Since the rheological properties of MR fluid can be modifiedmagnetically, an MR fluid based device requires no moving parts and therefore lessvulnerable to wear and tear.

Figure 2.1 Movement of magnetic particles in the MR fluid with and without magneticfield

2.2.1 Composition of Magnetorheological Fluid

MR fluids are commonly referred to a two-phase fluid suspension made froma mixture between magnetizable particle in a non-magnetizable liquid in such a wayso that it has very sensitive rheological properties to magnetic field [3]. Most of MRfluids are using the carbonyl iron powders as the magnetic particles. The carbonyliron powders are normally acquired from the chemical vapor decomposition (CVD)of pentacarbolyl iron, due to its large magnetization saturation and low coercivity[92]. The carbonyl iron powders are also preferred in MR fluid synthesis because itis chemically pure, naturally spherical in shape so that any anisotropy in magneticmoment can be minimized [93]. In contrast with the particles, there are numerous kindof carrier liquid that have been used in MR fluid such as water, mineral oil, siliconoil, hydrocarbon oil, and polyesters [92]. In order to increase the strength of MRfluid, the fraction between the volumetric amount of iron powder over the volumetricamount of MR fluid is highly important. The shear stress of the fluid with the influenceof magnetic field, or known as the on-state MR effect, normally increases with theincrease of volume fraction of iron powder [94]. However, the trade-off appears inthe off-state viscosity value that also increases exponentially with the increase ofvolume fraction of iron powder [95]. The trade-off has encouraged researchers to findsolutions to increase the strength of MR fluid and simultaneously suppress the off-stateviscosity value of the fluid such as to mix powders with different particle size and/orsize distributions [96] or to employ the functional composite particles that is coatedwith magnetic nanoparticles layer using core-shell technique [97]. There are also other

11

approach of MR fluid synthesis which are mixing non-magnetizable particles dispersedin magnetizable liquid (ferrofluid) that is known as the inverse ferrofluid [98] and theCarbonnanotube (CNT)-magnetite based MR fluid [99].

MR fluid are often distincted from ferrofluid from its particle size. Ferrofluid,similarly with MR fluid, is also a class of field sensitive fluid. However, the particlesize of MR fluid are often bigger (few microns) than the nano-sized particle ferrofluidbecause it seek larger yield stress due to stonger dipole moment between the particles.The yield stress of MR fluid represents the peak value of the stress versus strain curve,since the gel-like structure will break when the stress reached this peak value [1].However, since the particle size of the MR fluid is relatively large, it is susceptible tosettling and often caused unwanted abrasion [100]. The ferrofluid is less prone settlingbecause the nano-sized particles exhibit Brownian motion that keeps them dispersedin the carrier fluid. The larger the size of the particle, the lesser the effect of Brownianmotion and the faster the tendency of settling. According to Bell et al. [101], the MRfluids with particle size above 10 microns tend to settle in less than 20 minutes. Thatis why, in order to inhibit sedimentation and particle aggregation, some approachessuch as the addition of thixotropic agents and surfactants [102, 103] or the attempt ofnon-spherical particle shape implementations [104,105] were often given during to thefluid synthesis process.

However, despite of several improvement efforts to the MR fluid composition,there are generally no major modifications on the formulation technique of MR fluidsince the first discovery by Rabinow in the late 1940s. The strength of the very basicfluid formulation can be considered adequate for most MR devices and the stability ofthe fluid is also not a critical factor since most of MR devices such as dampers andbrakes can naturally act as a good mixing device [12].

2.2.2 Operational Modes of Magnetorheological Fluid

There are three known basic operational modes in any employment of MR fluidin a device, shear mode, flow mode and squeeze mode [15, 92, 106, 107]. Recently,another working mode was proposed by Goncalves and Carlson in [108], called themagnetic gradient pinch mode. However, the working modes of MR fluid are notstrictly limited to these four since the MR devices can be operated in a combination ofthese modes [109–114].

12

Shear mode, also known as the clutch mode, occurs when MR fluid is exposedto magnetic field between two parallel magnetic surfaces. One of the surfaces ismoving whilst the other is fixed, as shown in Figure 2.2. The fluid shear area whereMR fluid is exposed to magnetic field is called the effective area of shear mode. Shearmode is commonly used in brakes and clutches but also appears in some particulardesign of dampers, called the shear mode damper.

Figure 2.2 Shear mode

Flow mode, also known as the valve mode occurs, when MR fluid is exposedto a magnetic field whilst the fluid flows between two fixed parallel magnetic surfaces.The concept of flow mode is depicted in Figure 2.3. Similarly with the shear mode,flow mode has the effective area, which is defined as the area where the flowing MRfluid is exposed to magnetic field. Flow mode is commonly used in dampers and inother applications in which the devices requires a valve to control fluid flow.

Despite the concept of squeeze mode has been used in several early rotorvibration damper designs, the fundamental principle of this mode in MR fluid wasnot discussed until later in several publications [115–118]. Squeeze mode occurs whenMR fluid is exposed to magnetic field while at the same time being compressed ordecompressed, as shown in Figure 2.4. Since it involves very small displacement mostof its application appears in vibration suppression devices.

The latest working mode is called the magnetic gradient pinch mode. Thebasic idea of this mode is similar to flow mode but with a different configuration of

13

Figure 2.3 Valve mode

Figure 2.4 Squeeze mode

magnetic circuit design. In the magnetic gradient pinch mode, as shown in Figure2.5, the magnetic poles are arranged axially along the flow path and separated by anon-magnetic material. This kind of poles arrangement will create elliptical magneticfibrils, which will block the flow of MR fluid in the valve gap. One of the uniquecharacteristics of this mode is the slope between pressure and velocity relationship inmagnetic gradient pinch mode will be significantly increased when the magnetic fieldis increased [108]. This is unique, since in the conventional flow mode the slope tendsto remain constant under any magnetic field strength modifications. Another advantage

14

is the possibility to use MR fluid with coarser particles up to 100 microns, since a largerorifice is feasible to be used with the magnetic gradient pinch mode [108].

Figure 2.5 Magnetic Gradient Pinch mode

2.3 Magnetorheological Valve

Valve is the key components in almost any flow control mechanism. Thecommon practice in hydraulic system usually combines solenoid actuation valveelectronic devices in order to regulate fluid flow. However, conventional hydraulicvalve consists of several moving parts inside the valve, which makes it less responsivewhile at the same time more vulnerable to wear and tear. Therefore, an MR fluid basedvalve was introduced in order to improve the performance of valve. Moreover, sinceworking principles of most MR devices are based on the manipulation of fluid flowrate, the key performance of MR device is determined by the performance of MR valve.A successful improvement of MR valve performance can give a significant impact tothe development of other MR devices.

MR valve as one of the application branch of MR fluid typically consist of valvebody, valve core, electromagnetic coil, and fluid channel. The detail configurationof each MR valve could be different depending on the specific design but the basicprinciple is usually similar. The working principle of MR valve utilizes the sensitiverheological properties of MR fluid. The change in MR fluid rheological properties dueto magnetic flux variation from the electromagnetic coil in the valve changes the fluidflow resistance and therefore changes the pressure drop so that the fluid flow can beslowed or even stopped [20].

15

Considering the importance of MR valve, many designs of MR valve havebeen proposed. According to the structure of its flow path, MR valve can be dividedinto three categories, the annular type MR valve, the radial type MR valve and thecombination of annular and radial MR valve.

2.3.1 Annular Magnetorheological Valve

The annular MR valve can be easily distincted as the MR valve with effectivearea in the annular channel. Since the effective area of the valve mode is defined asthe area where the fluxes cross the fluid in perpendicular to the flow direction, thefluxes in annular MR valve is always directed in radial direction. The simplicity ofmechanical structure of the annular MR valve has made it the most common MR valvetype employed in various MR devices. However, the length of annular channel that canbe exploited as the effective area is limited and consequently the magnetorheologicaleffect that can be generated is relatively narrow.

One of the earliest designs of stand-alone MR valve was proposed in Kordonski[16] which later elaborated by Gorodkin et al. [17]. In the literatures, the initialconcept of MR valve, that is called the MR throttle valve, was presented to providea controllable MR fluid flow resistance. These designs basically comprise of two flatannular throttling hydraulic channels, indicated by 1, that were connected via internalchannels, indicated by 2, as shown in Figure 2.6. The induction coil winding, 4, andthe core, 5, were made from magnetoconducting material and together are formedthe magnetoconducting body, 3. The hydraulic channel itself is used as a part of themagnetic circuit, which the gap between the body and the core is acted as the zoneto induce the MR fluid with magnetic field. The design was also apparently preparedto serve various flow connection pattern for this design can be used in both serial orparallel connections depending on which flow port is connected to the flow conduit.The valve is acted as serial valve when each channel, 1, is connected to inlet/outletport, 6, to the inlet/outlet pipe, 8. Meanwhile, the valve is served parallel configurationwhen the pipe inlet, 10, is connected to the channel, 1, to the outlet, 8. The results ofperformance testing of this design were shown the capability to provide pressure dropof nearly 2.5 MPa at flow rate of 40 ml/s and 200 kA/m magnetic field intensity.

Another concept of annular MR valve was proposed by Yokota et al. [18]. Thedesign employed the very basic principles of valve mode using a C-shaped magnetic

16

Figure 2.6 MR throttle valve [17]

core attached to the flow channel as shown in Figure 2.7. The C-shaped magneticcore act as the main flux guiders which conducted the magnetic flux loops to inducethe MR fluid flows in perpendicular direction. Although the experimental results wereonly reported around 0.7 MPa of achievable pressure drop at similar flow rate and fieldintensity with the previous design, the design supposed to be very efficient since fluxloss can be prevented to a very minimum value. However, the bulkiness of the designwould quite inconvenience for application and compared with the previous design,this design is only equipped two flow port which made it only suitable for serial flowconnection.

Figure 2.7 C-shaped pressure control valve [18]

The work of Yokota et al. [18], was improved by Yoshida et al. [19] through the

17

introduction of a three-port annular MR valve as depicted in Figure 2.8. The three-portannular MR valve design comprises of serially connected two-port MR valves similarlywith the one in Yokota et al. [18] concept. However, one supplementary port is addedin between these two-port to ease the application that requires parallel flow connection.A NdFeB permanent magnet is added in between both coil to create a magnetic offsetas well as to regulate the direction of the fluid.

Figure 2.8 Three port MR valve [19]

Another group of researchers have focused to pursue an improvement of MRvalve in terms of size and performance by introducing the meso-scale (less than 25mm outer diameter) annular MR valve [20]. As illustrated in Figure 2.9, the conceptwas proposed using double coils with counter flux direction. The coils were wound inthe bobbin shaft which acted as the main flux core, while the valve casing was alsoacted as the flux return. The efforts were focused in two aspects, the first one was theanalysis of plug thickness, δ, which was defined as the thickness of pre-yield regionand directly related to the relationship between magnetic field strength and the valueMR fluid yield stress, while the second one is related to the material selection of thevalve which influenced the induced magnetic field strength to the MR fluid. The resultsdemonstrated that the approach was successfully generate saturated pressure drop of 2MPa at the flow rate of 40 ml/s.

18

Figure 2.9 Double-coil annular MR valve [20]

2.3.2 Radial Magnetorheological Valve

Although the advancement of annular MR valve were continuously explored,the space limitation to exploit the effective area in the annular channel has attractresearchers to investigate the other types of MR valve known as the radial MR valve.Radial MR valve is identified as the valve with effective area located in the radial flowdirection. In order to perpendicularly cross the radial fluid flow, the magnetic fluxesare typically directed in annular direction. As a result, the length of channel required toexploit the effective area is much smaller than the annular MR valve and therefore themagnetorheological effect of the radial MR valve is potentially wider than the annularMR valve.

The benefit of radial MR valve is introduced by Wang et al. [21] in thediscussion of radial MR valve for large-scale seismic damper application. The basicstructure of radial valve that was presented in the literature is shown in Figure 2.10. Inthe literature, the benefit of radial MR valve was benchmarked using the case of largescale damper applications. As a comparison, two large scale dampers with annularMR valve were compared with a large scale MR damper with radial MR valve. Theresults shown that the large scale MR damper with radial MR valve could providehigher efficiency in shorter valve length. The other advantage of radial MR valvestructure, namely the capability of radial MR valve to be made in modular stage wasalso highlighted in the discussion. The performance improvement of the MR valve wasshown to be linearly related with the number of addition of radial module in the MRvalve.

The work on radial MR valve was extended by the introduction of two-waycontrollable MR valve by Aydar et al. [119] as illustrated in Figure 2.11. Similarconcept of radial MR valve was enhanced through the installation of permanent magnetin adjacent to the electromagnets. The permanent magnet constantly provided initial

19

Figure 2.10 Basic structure of single stage radial MR valve [21]

magnetic flux to the radial gaps so that the off-state condition already has pre-yieldvalue. The two-way condition of the valve defined the ability of the valve to reduce thepressure drop as well as to enhance the pressure drop depending on the resultant of theflux directions. The presented design utilized the manipulation of flux similarly withthe one presented by Yoshida et al. [19]. The combination of permanent magnet andelectromagnet, which knows as the hybrid magnetic circuit, also provide additionaladvantages such as the possibility to enhance the on-state performance, to reverse themagnetorheological effect as well as to provide fail-safe features during power failureto the electromagnet [120].

Figure 2.11 Two-way controllable radial MR valve [119]

20

2.3.3 Combination of Annular and Radial Magnetorheological Valve

The successful performance improvement on radial MR valve approach hasmade researchers continuously explored the advancement of MR valve by combiningthe annular and radial MR valve. Ai et al. [23] was among the first who exploredthe gaps combination approach in MR valve, which later detailed by Wang et al. [24]as shown in Figure 2.12. Basically, the main idea of the approach is maximizing theengagement opportunities between the fluid and the flux by channeling the fluid inannular and radial gaps so that the fluid can cross the magnetic flux perpendicularly.With the combination of annular and radial gaps in an MR valve, the valve efficiencycan be increased in the meaning that the fluid flow blocking force can be increasedwithout any significant expenses to the valve size and energy consumption.

Figure 2.12 Annular-Radial MR Valve [24]

According to the study by Nguyen et al. [72] on the optimization of variousMR valve design, it was reported that the MR valve with annular and radial gapscombination is capable to deliver higher achievable pressure drop than annular MRvalve. In the same value of generated pressure drop, the annular-radial MR valve alsohas the smallest power consumption. However, in terms of the ratio of viscous pressuredrop the the field dependent pressure drop, the annular-radial MR valve has lowerperformance than annular valves.

2.4 Experimental Assessment Method for Magnetorheological Valve

The experimental assessment of the MR valve normally aimed to measure thepressure drop characteristics and validate the predicted performance of an MR valve

21

design. The characteristics of MR valve pressure drop are typically evaluated in termsof response to the variation of flow rate and response to the variation of current input.In general, the experimental assessment method of MR valve can be divided into twotypes. The first type is the constant flow assessment while the second type is thevariable flow assessment.

The constant flow assessment evaluates the performance of MR valve bymeasuring the pressure difference between both MR valve ports with subject tosteady fluid flow rate. This assessment arrangement has been used in several study[18, 27, 121, 122] is capable to characterize the MR valve similarly with the resultsfrom the quasi steady-state model where the pressure drop value varies in quasi steadyflow rate and yield stress. The constant flow assessment is also very useful to evaluatethe time constant of the MR valve, since it can maintain the constant flow rate valuewhile rapidly changing the current input. However, the testing arrangement of theconstant flow assessment method is a little bit complicated since it almost alwaysinvolves fluid pump in the arrangement such as shown in Figure 2.13. The complexityof the measurement comes from the existence of iron particles in the MR fluid and itsabrasive characteristics that limit the option of pump selection.

Figure 2.13 Typical arrangement of constant flow assessment method [18]

The variable flow assessment evaluates the performance of MR valve bymeasuring the pressure difference between both MR valve ports with subject tovariable fluid flow rate [20, 123, 124]. The fluid flow in variable flow assessmentis normally generated by a reciprocating movement of a hydraulic cylinder which

22

simultaneously generates pressure difference between cylinder chambers. The pressuredifference between hydraulic chambers induces alternating flow between thosechambers which the flow line passes through the MR valve. Since the flow isalternating, the variable flow assessment is very ideal to observe the hysteresis behaviorof the MR valve. The result of variable flow assessment of an MR valve is alsovery useful to further predict the dynamic behavior of a damper that utilizes the MRvalve. The measurement arrangement is also relatively simpler than the constant flowassessment given that the dynamic test machine or fatigue test machine can be usedas the source of movement to the hydraulic cylinder such as shown in Figure 2.14.However, the method cannot be used to measure the time constant of the MR valvesince the valve response to rapid change of current input in a constant flow rate cannotbe executed.

Figure 2.14 Typical arrangement of variable flow assessment method [124]

2.5 Modeling Approach for Magnetorheological Valve

In order to satisfy the requirement of control system design, the behavior ofMR based devices should be mathematically modeled. In general, the behavior ofany MR based devices can be described in two ways, the steady state behavior andthe dynamic behavior. The steady state behavior of the MR device is defined as thebehavior when the variation of variables is time-invariant. The dynamic behavior, on

23

the other hand, is defined as the behavior when the influenced variables are changingover time. According to Snyder et al. [125], the linear steady state behavior model isonly sufficient to predict the energy dissipation, however, it cannot accurately portraysthe force response of the MR damper. The force response of the MR damper is knownas a highly nonlinear variable that is dependent to amplitude and frequency of motionand it can only be modeled using nonlinear dynamic model [126–129].

2.5.1 Steady-state Model

In general, any model of MR based devices would fundamentally refer to threeknown models of MR fluid characteristic, Bingham plastic model, biviscous model,and Herschel-Bulkley model [74]. Bingham plastic model is the simplest and the mostpopular model in field controllable fluids such as Electrorheological (ER) fluid andMR fluid. Bingham plastic model describes the rheological behavior of MR fluid in thepre-yield region as a rigid solid body while in the post-yield region as Newtonian fluid[130]. The Bingham plastic model can be described in the stress-strain relationshipequation as follows [74]:

τ = τy (H) sgn (γ) + ηγ (2.1)

where τ is the MR fluid viscosity when no magnetic field applied, γ is the shear-strainrate and sgn (γ) is the signum function for shear-strain rate.

Biviscous model is particularly used for the analysis of MR based devicesoperating in the squeeze flow mode [74], because it has an additional expression forstatic yield stress. The equation for biviscous model can be expressed as follows [6]:

τ =

τy (H) + ηpoγ, for |τ | > τys

ηpreγ, for |τ | < τys(2.2)

The expression of the biviscous model in Equation 2.2 has two states. Forabsolute shear stress |τ | above static yield stress, τys, the expression is similar tothe Bingham plastic model with post-yield viscosity, ηpo as the viscosity componentof the fluid. For shear stress below static yield stress, the expression is similar to aconventional Newtonian fluid with pre-yield viscosity ηpre as the viscosity componentof the fluid. The relationship between dynamic yield stress τy (H) and static yield stress

24

τys can be expressed as [6]:

τy (H) = τys

(1− ηpo

ηpre

)(2.3)

Wang and Gordaninejad [131] are among the first who generalized the useof the Herschel-Bulkley model in MR fluid. The Herschel-Bulkley model can beconsidered as the improvement of the Bingham plastic model since it has parameterto anticipate shear thinning and shear thickening behavior, which only occurs at highshear rates. The Herschel-Bulkley model can be expressed as [6]:

τ =(τy (H) +K |γ|

1m

)sgn (γ) (2.4)

In general the Herschel-Bulkley model is similar to the Bingham plastic modelexcept for the existence of two new parametersK andm.K andm are fluid parametersknown as the consistency parameter and the flow behavior index respectively. K issimilar to off-state fluid viscosity η in Bingham plastic model while m is the index torepresent shear thinning or shear thickening effect. Form > 1, Equation 2.4 representsa shear thinning fluid, while shear thickening fluids are described by m < 1. TheHerschel-Bulkley model can provide exactly the same results as the Bingham plasticmodel when m = 1.

In the case of MR valve, most of its steady-state behavior are also modeledbased on Bingham plastic model equation, although in some conditions the moresophisticated models, such as the Herschel-Bulkley model, are more preferred [117].In particular, the MR valve behavior normally declared by the following equation:

∆P = ∆Pviscous + ∆Pyield (2.5)

where the behavior of MR valve is described as the value pressure drop of an MRvalve that is consist of two parts, the pressure drop from viscous properties of the fluidand pressure drop from the field dependent yield stress of the fluid. The most knownmodel of MR valve is the mathematical model based on the annular flow channelconfiguration, which also became the common understanding of the flow mode of MR

25

fluid. The annular based MR valve model can be described as follows [23]:

∆Pviscous =6ηQL

πd3R(2.6)

∆Pyield =cτ (H)L

d(2.7)

where the pressure drop from viscous properties in Equation 2.6 is proportionallyrelated to the fluid base viscosity (η), the volumetric flow rate (Q), and the annularchannel length of the valve (L) but inverse-cubical to the valve gaps (d) and inverseproportional to the channel radius (R). Meanwhile, the field dependent pressure dropin Equation 2.7 is proportionally associated with the field dependent yield stress value(τ (H)) of MR fluid, annular channel length (L), and flow-velocity profile coefficient(c) but inverse-proportionally related to the gap size (d). The coefficient c is normallydefined as the ratio between field dependent pressure drop and viscous pressure dropwhich its value is in the range of 1 to 3.

1

2

Figure 2.15 Illustration of significant variables in an MR valve [132]

Note should be taken that the model shown in Equation 2.6, described theviscous pressure drop using the constant of 6. In some other literature [22, 27], theviscous pressure drop was described using the constant of 12. The reason behind thedifference is in the understanding of the annular channel length (L). In a commonannular valve design, there will always be two annular channel length (L1 and L2) that

26

are separated by the coil winding as shown in Figure 2.15. Some researchers, preferto define both length separately and therefore the pressure drop expression of eachannular channel length is defined using the constant of 6. Some others, assume bothchannel length identical (L1 = L2 = L) and therefore since the length of L defined as2L, the constant of 6 became 12.

The other form of annular MR valve model was also presented using ananalytical terms called the nondimensional plug thickness, δ, as shown in the followingequation [20, 133]:

∆P =12ηQL

bd3(1− δ

)2(

1 + δ2

) (2.8)

The nondimensional plug thickness is defined as δ = 0 if the fluid flows as theNewtonian fluid and defined as δ = 1 if the fluid is completely blocked by the valvewith infinite blocking pressure. However, the concept of plug thickness is difficult to beimplemented in the pressure drop prediction since the relation between the magneticstrength and yield stress of the fluid to the pressure drop are not direcly expressed inthe equation. The plug thickness concept will be more suitable for reverse calculationof yield stress using measured pressure drop data or to model the valve behavior usingsteady-state equation.

The other steady-state model that has been developed is the radial MR valvemodel. Wang et al. [21] proposed the radial valve model as described in the followingequation:

∆Pviscous =6ηQ

πd3ln

(Do

Di

)(2.9)

∆Pyield =2(Do−Di

2

)τ (H)

d(2.10)

where the valve gap size (d) in Equations 2.9 and 2.10 refers to the size of radial gapswhile the Do and Di refer to the outer radius of the radial gaps and the inner radiusof the radial gaps respectively. The field dependent yield stress value (τ (H)) of MRfluid in Equations 2.9 and 2.10 is also specifically refers to the yield stress value of inthe radial gaps. Similarly with the annular valve model, the gap size is also inverse-cubically related with the viscous pressure drop and inverse proportionally related withthe field dependent pressure drop.

27

In the same form with the Equation 2.9 and 2.10, the radial valve model wasalso presented in the following form [24]:

∆Pviscous =6ηQ

πd3ln

(Do

Di

)(2.11)

∆Pyield =cτ (H)

d(Ro −Ri) (2.12)

where the main differences between these models are only in preference of usingdiameter (D) and radius (R) of the radial gaps in the equation and also in the constantselection of the field dependent pressure drop where in equation 2.10 the moderateconstant of 2 is used while in equation 2.12, the flow-velocity profile coefficient, c, ischosen.

According to these equations, it can be summarized that the pressure dropof an MR valve is determined by several factors. The fluid base viscosity, flow rateare the parameters that proportionally related only to the viscous pressure drop. Theflow path length, on the other hand, is the only variable that influencing both viscouspressure drop and yield pressure drop simultaneously. Meanwhile, the channel gap canbe considered as the most significant parameter, since, according to the equations, itinverse-cubically determine the value of the viscous pressure drop as well as inverse-proportionally influence the yield pressure drop. However, intrinsically, the influenceof channel gap in determining the pressure drop is even more significant since it alsodetermine the magnetic field strength. According to Wang et al. [24], the fluid channelgap has magnetic resistance one thousand times than the resistance of the magneticmaterials of the valve core. Therefore, the alteration of gap size will dramaticallychange the magnetic circuit performance which finally will influence the yield stressvalue.

2.5.2 Dynamic Model

While the steady state modeling approaches for MR valve has been established,the dynamic modeling technique of MR valve as a single component is rarelydisseminated. This can be understood, since most of the literatures were only focusedtheir modeling in the prediction of the valve peak performance. However, to fulfillother purpose such as the control design process, the steady-state model only willnot adequately represent the device behavior. The necessity of dynamic modelingof MR valve also can be considered higher than MR damper because the MR valve

28

and obviously the model can be used in the wider range of applications aside ofMR damper. For example, the knowledge of MR valve behavior and its virtualrepresentation can be used as a reference to design and predict the performance ofan MR damper or to help further development of new concept of actuators that requireMR fluids flow control [25, 128, 134].

In general, the dynamic modeling of an MR based device can be divided intotwo approaches, the parametric modeling and non-parametric modeling. Parametricmodeling is the modeling approach with a collection of mechanical elements asa hypothetic representation of the device so that the device behavior can bemathematically modeled using the dynamic relationship between elements. Parametricapproach has the advantages due to the ability to provide a generalized form of themodel, which means that the same form can be used repeatedly in other devices.Nevertheless, there are pitfalls in the accuracy of the results since the model abilityto follow the device behavior is limited and in order to mimic the characteristicsof a particular device in a good match, the involvement of a rigorous parameteridentification method is mandatory [135, 136]. On the contrary, the non-parametricapproach is able to avoid the problem in modeling accuracy. The non-parametricmodeling is approaching the modeling problem with the analytical expression todescribe the device characteristics based on the testing data [74]. However, there isno generalized form since the model structure will be unique for each device and thedevelopment is highly dependent on a specific testing data. Hence, the process flow togenerate the non-parametric model is typically more complicated than the parametricmodel. Moreover, the validity of the generated model is only limited to a particulardevice that is incorporated in the modeling process.

There are various kinds of parametric model that have been developed. Forexample, Sakai et al. [137], proposed a simple parametric model based on the LuGrefriction operator as an inverse model of an MR damper as an actuator for vibrationcontrol. The LuGre friction model also explored and modified by Jimenez and Avarez-Icaza [138] to model an MR damper as a part of a semi-active control system. The othertype of parametric model, which also apparently the most popular parametric model,known as the Bouc-Wen model has been extensively explored as the parametric modelof MR damper [136,139–143]. The Bouc-Wen model approximate the device behaviorthrough a set of hypothetic spring and damper as shown in Figure 2.16.

Another example of parametric modeling approach for MR device is the classof algebraic model, namely the hyperbolic tangent model, that was proposed by

29

Figure 2.16 Bouc-Wen model

Gavin et al. [144]. The method was then utilized by Kwok et al. [135] to modelan MR damper incorporating particle swarm optimization (PSO) as the parameteridentification method. Jiang and Christenson [145] also use similar model in a real-time hybrid simulation environment for a large scale MR damper. A performancecomparison study of various parametric modeling approach that has been conductedby ahin et al. [146] reported that the class of algebraic model is able to provide betterperformance fit than differential model. Although, according to comprehensive reviewby Wang and Liao [74], the accuracy of the parametric model is not just rely on thebasic form of the model itself but also highly determined by the identification methodof the parameters. In general, Wang and Liao [74] inferred that model with highernumber of parameters and equipped with an appropriate identification method willhave better accuracy than the one with less parameters.

The other parametric method that is also often explored in the MR dampermodel is the polynomial model. Choi et al. [147] introduced the approximation of theforce-velocity curve of an MR damper using polynomial equations. The method issimple yet unique because there were two separated 6th order polynomial equationsfor each positive and negative directions as shown in Figure 2.17. to compensatethe inequality of the force-velocity curve in positive and negative direction due tohysteresis. The method was then applied by Ubaidillah et al. [148] in the modelingapproach of an MR damper for automotive semi-active suspension system. Thepolynomial model is favored because it is convenient to use and easy to be adapted inan open-loop system [149]. However, according to Sahin et al. [146], the polynomialmodel has weakness in characterizing the device behavior in low velocity region dueto the absence of variables that characterize the pre-yield force.

30

Figure 2.17 Parametric hysteretic polynomial model [147]

Nevertheless, while the achievable accuracy of a parametric model is alwaysbounded by the model complexity, if the issue of model accuracy became morestringent, the non-parametric model often chosen as a better option. There are severalmethods that have been executed to model the MR damper behavior with non-parametric approach. One of the most extensively discussed is the artificial neuralnetwork [150–152]. The artificial neural network was often chosen as the base methodof non-parametric model due to its ability to automatically learn from the trend ofmeasurement data and generate an emulated functions afterward as depicted in Figure2.18. The artificial neural network is also able to be combined with various learningmethod such as lazy learning [153], and ANFIS [154].

2.6 Utilization of Magnetorheological Valve

Despite the lack of specific literature discussed about MR valve as a stand-alone device, in fact, it is the most vastly used MR devices. The MR valve has becomethe most influential component in the MR damper, which is commercially the mostsuccessful MR device. Recently, the development progress of new MR actuators alsostimulated the rapid development of MR valve. In this section, the utilization of MR

31

Figure 2.18 Artificial Neural Network model [151]

valve is elaborated in three different MR devices, namely the linear MR damper, therotary MR damper and the new MR actuator.

2.6.1 Linear Magnetorheological Damper

Linear MR damper is one of the most discussed MR devices so far. The rangeof application of MR damper is widely ranged from automotive to civil applications.According to the working modes, there are two basic contruction of MR damper. Themost common design is the valve mode MR damper with valve located in the piston asshown in Figure 2.19. The other MR damper equipped with shear mode as depicted inFigure 2.20.

Since both types of damper regulated flow using the narrow gaps in the piston,these narrow gaps are functioned as the valve that is embedded in the damper piston.Although these structures are the most common structure of MR damper due to itscompactness the fact that the coil should be wound in the piston has increased thecomplexity of wire assembly as well as increasing the weight and thickness of thepiston head [14]. The internal coil placement also made the heat dissipation from thecoil more difficult and potentially increase the temperature of the MR fluid [58].

The alternatives of coil placement has been proposed by Chen et al. [155]using external coil structure (see Figure 2.21). The construction of the damper notjust allow the heat dissipation to be directly released to the atmosphere but also madeit possible to add the self-powered and self-sensing features [157]. The coil also can

32

Figure 2.19 Valve mode MR damper [9]

Figure 2.20 Shear mode MR damper [75]

33

Figure 2.21 External coil MR damper [155]

Figure 2.22 MR damper with bifold valves [28]

be easily accessed for maintenance without having to interfere with the internal part ofthe cylinder nor the fluid. However the basic construction of the external coil designis quite complicated and the performance of the damper is apparently less than theinternal coil design due to difficulties in optimally routing the magnetic flux.

The other approach of coil placement are using the bifold valve design shownin Figure 2.22 [28] where the coils are statically installed in both ends of the cylinder.The bifold valve damper gain the advantage in design compactness similarly with thedamper with internal valve. Good damping performance was also reported especiallyin high shock velocity. However, the design is also quite complicated and the valve

34

Figure 2.23 Bifold MR damper for high impulsive loads [156]

Figure 2.24 Bifold MR damper for shock vibration mitigation [158]

effective area is very limited and thus the performance expansion cannot be executed.Since the design can provide advantage in high shock velocity, the design has beenexplored for vibration mitigation application by Facey et al. [156] and Mao et al. [158]as illustrated in Figure 2.23 and 2.24.

If the design simplicity and the ease of capacity expansion are the priority,then the bypass structure can be a better option. The bypass structure where the valve

35

Figure 2.25 Basic structure of Bypass MR damper [8]

are separated from the piston and physically located outside the cylinder as depictedin Figure 2.25 [8]. The separation between the cylinder and the valve has made theassembly process of the bypass damper is far simpler than the other damper sincevariety of commercially available hydraulic cylinder can be used. In the other words,the customization requirement of damper components is very minimum. Though theassembly process is simpler, the mounted valve outside the cylinder typically made thesize of bypass damper larger than the damper with internal valve. Nevertheless, mostof applications of bypass MR damper is on a large scale damping component such asthe seismic dampers shown in Figure 2.26 [60].

Figure 2.26 Bypass MR damper for large scale seismic application [60]

2.6.2 Rotary Magnetorheological Damper

The rotary MR damper is another type of MR damper that operates in angularmovement. For some applications that require large stroke, the rotary MR damper canbe a better option than linear damper since the risk of buckle is very minimum. The

36

rotary damper also more suitable in a hostile working environment since the dampershaft in a rotary damper is sealed better than the damper rod in a linear damper.However, the rotary MR damper that equipped with valve is so far is limited to thevane type MR damper only. The other devices that can be classified as rotary MRdamper, e.g. MR brake, are typically not equipped with valve and operate mainly inshear mode.

In terms of device advancement, unfortunately, vane type MR damper has notbeen discussed in many publications. Only three written publications were found overthe last 10 years. The first one is the work of Zhang et al. [159], which adaptedthe conventional structure of hydraulic vane damper as shown in Figure 2.27. In aconventional hydraulic vane damper, the fluid would be pumped from one chamberto another across gaps between the static vane and the rotary vane when the damperrotates. In the proposed design, these gaps were utilized as valve, using the effectivearea of shear mode by incorporating them with electromagnetic coil. In order to installthe electromagnetic coil, the static vane or in this case was called the clapboard, wasmade in a hollow structure. Meanwhile, the effective area of shear mode of the valvewas made in an arc configuration and therefore it was called as an arc MR valve.

Figure 2.27 Vane type MR damper with arc valve [159]

Preventing leakages between chambers is the main challenge in the design ofa vane type damper. Therefore, the main strength of the proposed design in [159] isits capability to use the potentially leaked gap between chambers into the flow controlzone. However, since the coil was located inside the damper, the generated heat fromthe coil would be difficult to effectively dissipate. Maintenance of the damper is alsodifficult, if there is a problem with the coil, the whole damper has to be dismantled.Unfortunately, given that the intention of the publication was only to demonstrateoptimization analysis of the new vane type MR damper design, no disclosed results

37

on damper performance were presented.

The second publication of vane type MR damper was written by Giorgetti etal. [32]. As shown in Figure 2.28, the second vane type MR damper, consisted ofmoving vane, indicated by A, static valve body, B, the valve core, C, the vane dampercasing, D, and the electromagnetic coil seat, E. As with the previous design by Zhanget al. in [159], it incorporated an internal valve design. However, in this design, thevalve was not located in the gaps between static and rotary vane, but was intentionallymade in the static vane body. The new configuration in [32] has changed the natureof internal valve mechanism from shear mode to flow mode. Leakages that potentiallyoccurred at the gaps between the static and rotary vane still occur, but are neglectedsince they are very small.

Figure 2.28 Vane type MR damper with outer coil valve [32]

Generally, the vane type MR damper design by Giorgetti et al. in [32] hasshown simplicity and compactness. The experimental results showed that the dampercould provide up to 200 Nm damping torque at 3 A current input in a feasible geometryto be applied in passenger vehicle suspension. However, the incorporated internal valvecould have the same complexity in terms of coil maintenance as that described in [159].

Yang et al. [160] discussed a vane type MR damper design with flow typeinternal valve. The design has shown similarities with the work done by Giorgetti etal. in [32] in the context of annular flow mode valve design, where the effective areasare located in line with the direction of fluid flow. The main difference was in theplacement of the coil, as shown in Figure 2.29. In Giorgetti et al. design [32], the coilwas located in the outer radial side of the fluid path, while in Yang et al. design [160],the coil was located in the inner radial side of the fluid path.

38

Figure 2.29 Vane type MR damper with inner coil valve [160]

Since Yang et al. [160] put the coil in the inner radius, a lower resistance valueof the coil could be obtained leading to lower power consumption. The valves in thedesigns of Yang et al. [160] and Giorgetti et al. [32] have similar dimension and numberof coil turns as well as the wire gauge, the coils in the Giorgetti et al. design [32] wouldsimply have higher resistance than the coils in the Yang et al. design [160]. Higherresistance in coil means higher power consumption for the same current input, sincepower relates proportionally to resistance and quadratically to current. Therefore, if itis assumed that both valves have the same effective area and gap width, then both valvewould have the same performance but the Yang et al. design [160] would consume lesspower.

The disclosed experimental results show that the achievable damping forcethat can be generated by Yang et al. [160] could be up to 3 kN at 4A current input.However, since the damping performance was evaluated in different measurement unitsthe results are difficult to be compare with those of Giorgetti et al. [32]. Giorgetti etal. [32] demonstrated the damper performance in torque unit, while Yang et al. [160]showed them in force unit. Although it is known that the achievable damping forcecan be converted to torque by multiplying them with the length of rotating hub, thelength of rotating hub was not specifically mentioned in any part of their publications.Nevertheless, logically, both designs should be in the same class in terms of achievabledamping torque, since both of them were designed for passenger vehicle suspension.Both of them were also capable of providing almost the same ratio between on-stateand off-state damping at around 6.

39

2.6.3 New Magnetorheological-based Actuators

Aside from the device component, the MR valve also has been developed as astand-alone device. The stand-alone MR valve is an independent component that canbe used to retrofit the hydraulic valve or to complete a newly developed actuators.There are several discussions about a new MR based actuator that can be functioned asan active actuator similarly with a hydraulic or pneumatic actuator such as the bellow-driven motion controller shown in Figure 2.30. [19], the MR hydraulic power actuationsystem in Figure 2.31 [25], the MR actuation system with embedded terfenol-D pumpin Figure 2.32 [26], and the MR based link manipulator in Figure 2.33 [134].

Figure 2.30 MR based bellow-driven motion control [19]

Figure 2.31 MR hydraulic power actuation system [25]

40

Figure 2.32 Actuation with embedded Terfenol-D pump [26]

Figure 2.33 MR based link manipulator [134]

One of the unique working application of the MR valve is the concept of 4/3way directional MR valve proposed by Saloom and Samad [27] as depicted in Figure2.34. The 4/3 way directional valve is one of the common functional hydraulic valvethat serve the hydraulic actuator and determine the direction of force actuation. The4/3 way directional valve normally implemented in servo hydraulic actuation but inorder to operate in higher performance, a sophisticated valve actuation is required. Thedesign of 4/3 way directional MR valve was one of the proposed solution that utilizingthe rapid response of MR fluid to magnetic field. The proposed 4/3 way MR valveconcept was basically constructed of four single MR valves that each of the valveelement was arranged in between the valve ports. There are five ports in this valvedesign, one inlet pressure port, two actuator ports, and two return ports. By activatingand adjusting the strength of magnetic field in different combination of annular valve,the direction and rate of MR fluid flow can be regulated so that the actuation control offorce and direction of hydraulic can be achieved.

41

Figure 2.34 4/3 way directional MR valve [27]

2.7 Summary of Chapter 2

The overview of MR fluids, MR valve structure, MR valve modeling techniqueand the variety of MR valve applications have been presented. The explanations of MRfluids have been elaborated based on its composition and its working modes. Thereare many types of MR fluids but generally, researchers have made a clear distinctionof MR fluid from other types of magnetically responsive fluid from its micron-sizedparticle. The MR valve explanations were conducted based on the categorization offlow channel types, namely the annular MR valve, radial MR valve and combinationof annular and radial MR valve. The annular MR valve, as the most popular andcommonly used MR valve in the MR damper, has limitation in performance expansioncapability which promote the discussion of radial MR valve that has advantages inmulti-stage configuration. The most recent study that explored the MR valve withcombination of both annular and radial flow channel also has been covered in theoverview. The modeling overview of the MR valve was also presented and dividedinto two parts, the steady-state and dynamic model. The steady-state model discussioncovered the mathematical approaches that often used to model the physical behavior ofthe MR valve, while the dynamic model explored two general approach of hysteresismodeling of MR devices, the parametric model and the non-parametric model. Inthe final section of the chapter, various devices and applications that incorporatesMR valve as its main component were presented. According to the review, it can besummarized that the current advancements of MR valve are still unable to improve theachievable pressure drop in a compact valve size and low power consumption. Throughthe review it was also shown that although the dynamic models of MR damper havebeen widely studied, the dynamic models of MR valve, in particular, have never beenthoroughly discussed.

CHAPTER 3

MAGNETORHEOLOGICAL VALVE CONCEPT

3.1 Introduction

A clear understanding in MR valve behavior and characteristics are importantaspects in the design process of an MR valve. In this chapter, a methodology to designa new concept of MR valves is carried out. The concept development process is dividedinto two phases, the first phase is mainly related to the design consideration includingthe magnetic circuit simulation while the second phase is related to the derivation ofthe MR valve steady-state model to predict the performance of each valve design.The Finite Element Method Magnetics (FEMM) version 4.2 is utilized to performthe electromagnetic circuit design simulation for the MR valve, which results arethen combined with the steady-state mathematical model of MR valve to predict theachievable pressure drop that can be generated by the valve. The results of performanceprediction are then used to select the valve design for experimental assessment.

3.2 Design of Magnetorheological Valve

The elaboration of MR valve design is needed to explain the technicalconsiderations that are taken to develop the concept. There are several aspects inthe design process of the MR valve that are critical such as the determination of thematerial for the valve component, the strucural design consideration and the simulationof the magnetic circuit of the MR valve.

43

3.2.1 Conceptual Design

There are two disclosed MR valve concepts that provided less than 50 mmdiameter MR valve with achievable pressure drop more than 1.5 MPa at 40 ml/sflow rate. The first concept was proposed by Yoo and Wereley [20] using double coilannular configuration while the second concept was introduced by Ai et al. [23] usingcombination of annular and radial configuration which experimentally elaborated [24].The double coil annular concept, as depicted previously in Figure 2.9, was proposed asa high performance MR valve that can generate pressure drop up to 2 MPa at fluid flowrate of 40 ml/s with outer diameter size of 25.4 mm and uses annular flow path channelwith internal coil attachment. The valve size still can be reduced up to 40% without anyperformance degradation if high permeability material such as Hiperco alloy is used.Meanwhile, the combination of annular and radial concept, shown in Figure 2.12, wasintroduced as an effort to improve the performance of conventional MR valve design.The first concept of MR valve with annular and radial flow path has outer diameter sizeof 45 mm and capable to perform in similar pressure drop range with the double coilannular MR valve.

Learning from these examples, the principles of pressure drop improvement inthe MR valve basically can be done by maximizing the yield stress of MR fluid ina larger effective area. Effective area of MR fluid is the area where the rheologicalproperties of MR fluid can be effectively regulated by the activation of magneticfield. Yield stress of MR fluid can be augmented by increasing the magnetic fluxdensity. However, magnetic flux density is highly related to the magnetic field intensityfrom the coil and permeability of magnetic material that is used as a medium. If theutilization of high permeability material is exempted, the common consequences ofincreasing magnetic flux density are higher dimension of coil, which lead to larger sizeof MR valve, and higher power consumption. The other way to increase yield stress ofMR fluid is by increasing the effective area by enlarging the valve dimension and/orby extending the flow path of the MR fluid.

When the valve size matters, any improvement efforts which involveenlargement of valve dimension are not preferable. For that reason, extending theflow path of MR fluid is the feasible option. The combination of annular and radialconcept by Ai et al. [23] is a good example of effective area expansion by extendingthe flow path. Another study by Nguyen et al. [72] has also confirmed that the MRvalve with combination of annular and radial gaps could provide higher achievablepressure drop than other types of annular MR valve with the same outer radius and

44

power consumption. However, since the current design uses only a pair of annularand radial gaps, the total effective area improvement is still very narrow; thus, theenhancement of pressure drop that can be achieved is also limited. Therefore, in orderto expand the total effective area more, the length of flow path should be increasedusing multiple combination of annular and radial gaps forming a meandering flow path.The conducted study in this research will explore the effectiveness of the meanderingflow path concept to improve the pressure drop performance of the MR valve in thesequence shown in Figure 3.1.

Figure 3.1 Concept assessment sequence of the new MR valve concept

The detailed structure of MR valve with meandering flow path that is exploredin this study is shown in Figure 3.2. The MR valve structure can be divided intothree components, the casing, the coil and the valve core. The casing resembles theouter shell of the valve with two main functions. The first function is to secure thewhole structure of the valve together as well as to connect the valve with other devicesthrough its embedded fittings. The second function is to properly guide the magneticflux from the coil so that the loss of flux to the air, which reduce the efficiency, can beminimized. The coil is the main component that generates and determines the magneticfield strength of the MR valve. In this proposed MR valve structure, the coil is turnedin a coil bobbin that is also functioned as the wall of the flow channel. The core is thecomponent that formed the annular and radial flow channel, which can be divided intofour parts, the side core, the center core, the orifice core and the spacer.

45

Figure 3.2 Basic concept of MR valve with meandering flow path

3.2.2 Design Consideration

To facilitate the design process, the consideration of materials for eachcomponent as well as the consideration of coil properties, valve size and fluid typesare needed. In order to select the component materials, the understanding of fluxrouting concept is required. The flux path can be modified through the flux routedetermination process by combining the materials with higher permeability (magneticmaterials) and the materials with lower permeability (non-magnetic materials) in sucha way so that the magnetic fluxes will penetrate the MR fluid channel. The commonunderstanding in magnetic circuit design for MR application usually recommendspenetration of magnetic fluxes in perpendicular direction to the direction of fluid flow.Quantitatively, the flux routing process can be evaluated by calculating the magneticcircuit performance for each combination of materials using analytical or numericalanalysis. However, in order to simplify the routing process, a qualitative evaluation ofthe material selection and combination for the three main components of the proposedMR valve is conducted as summarized in Table 3.1.

In Table 3.1, the combination summary of all possible material arrangement forthe three main components of the valve is shown. The main goal of these combinationsis to find the best material selection that capable to guide the fluxes passing through

46

Table 3.1: Materials selection of valve component for valve routing

No.Valve components

ResultsCasing Coil bobbin Core

1 Magnetic Magnetic MagneticFlux looped through thecoil bobbin and casing

2 Magnetic Magnetic Non-MagneticFlux looped through thecoil bobbin and casing

3 Magnetic Non-Magnetic MagneticFlux looped through casing

and core with a high fluxdensity in the fluid gaps

4 Magnetic Non-Magnetic Non-MagneticFlux looped through thecoil bobbin and casing

5 Non-Magnetic Magnetic MagneticFlux looped through thecoil bobbin and casing

6 Non-Magnetic Magnetic Non-MagneticFlux looped through thecoil bobbin and casing

7 Non-Magnetic Non-Magnetic MagneticFlux looped through casing

and core with low fluxdensity in the fluid gaps

8 Non-Magnetic Non-Magnetic Non-MagneticFlux looped through the

coil bobbin and casing withthe lowest field intensity

the fluid gaps with sufficient magnetic field strength. From these eight possiblecombinations, only two combinations show the possibility of flux to pass through thefluid flow gaps. However, only the combination of magnetic casing and core with non-magnetic coil bobbin is capable to provide a high magnetic field strength in the fluidgaps. From the remarks in Table 3.1, it can be concluded that in order to successfullyguide the fluxes to pass through the fluid gaps, the permeability of the core componentshould be higher than the permeability of the coil bobbin. However, to get a highermagnetic field strength in the fluid gaps, the permeability of the casing should be atleast similar to the permeability of the core.

In this study, the selected material of the casing is the mild steel with propertiesthat compatible with the AISI 1010. The design of the casing consists of two identicalparts, which each part has four threaded holes as the holder for the locking bolts andone female BSPP (British Standard Pipe Parallel) 1/4” fluid channel port. Since thecoil bobbin has to be made from the low permeability material, aluminum is chosenfor the material of the coil bobbin. The coil bobbin is also designed with a pair ofgrooves for O-rings installation and has the threaded holes in pair to the holes in thecasing. Meanwhile, the valve core is separated into two side cores, two orifice cores, acenter core and six spacers. All parts in the valve core are also made from mild steel

47

that comply with AISI 1010 material except for the spacers, which are made fromaluminum.

The outer diameter size of the valve is determined of 50 mm with overall lengthof 77 mm. On the other side, the inner diameter size, that is the inner diameter ofthe coil bobbin that becomes the bore size of the flow channel is determined of 15mm. The core parts are then arranged to form the annular and radial gaps alternatelyin such a way it meanders the flow channel. As one of the subject of the study, thesize of annular and radial gaps are both varied from 0.5 to 2.0 mm. Since the valveis designed to generate high pressure drop, the casing should be strong enough towithstand such pressure and for that purpose the thickness of the casing is determinedof 5.0 mm. As a consequence, the space available for the coil winding that is made from23 SWG copper wire is sufficient for around 545 turns with total resistance of around3.03 Ω. In this study, the current input is limited to 1.0 A therefore the maximum powerconsumption of the valve is also limited to around 3.03 W.

3.2.3 Magnetic Simulation

In order to predict the effect of current input to the MR effect strength ofMR fluid in the effective area, the permeability (B-H curve) and the yield stress, τ ,need to be determined. There are specific B-H curve and τ curve for each type ofMR fluid. In this study, the MR fluid that will be used is the Lord Corporations’sMRF-132DG, which τ curve’s was already approximated by Nguyen et al. in [72],as depicted in Figure 3.3, using polynomial approximation as a function of B in thefollowing function:

τ (B) = 52.962B4 − 176.51B3 + 158.79B2 + 13.708B + 0.1442 (3.1)

where B is the magnetic flux density in Tesla which is highly determined by currentinput, coil turns and dimension, and total reluctance of the medium. The permeabilitycurve of the MRF-132DG, together with the other permeability value of the materialsused in the MR valve can be used in the calculation to determine the magnetic fluxdensity, B, at the effective area. Since the number of turns of the coil is alreadydetermined and the permeability of magnetic material was assumed to follow the B-Hcurves of AISI 1010, the magnitude of magnetic flux density can be determined bythe magnitude of current input supplied to the coil. However, due to the complexityof magnetic structure and non-linear permeability of the materials, the magnitude of

48

magnetic flux density for each zone is difficult to be calculated analytically [111].Thus, in order to determine the value of B for each corresponding current input, afinite element method based software for magnetic simulation called FEMM is used.The results from FEMM are shown in Figure 3.4 and 3.5.

Figure 3.3 Approximation of yield stress as a function of magnetic flux density, B

Figure 3.4 Two-dimensional axisymmetric model of the MR valve in FEMM

Figure 3.4 shows two dimensional axisymmetric meshed model in the FEMMusing triangular element with total elements number of 103771 and total nodes numberof 52281. Figure 3.5 shows the contour and flux lines of magnetic flux density for the

49

Figure 3.5 Flux lines and contour of magnetic field of the MR valve in FEMM

whole magnetic circuit. Notes should be taken that the flux that can influence yieldstress of MR fluid is only the fluxes that passed through MR fluid gaps. The MR fluidgaps area that crossed by the magnetic flux is known as the effective area. Accordingto the result in Figure 3.5, it can be observed that the flux lines have passed across theouter annular gap, outer radial gap and inner radial gap. Therefore, only MR fluid thatpassed through these three gaps can be rheologically influenced by the changing of thecurrent input. Meanwhile, the inner annular and orifice gaps are not intensively exposedto magnetic field and therefore the influence of magnetic field to yield stress of MRfluid in these areas can be neglected. Yet, although the inner annular and orifice gapsdo not provide field-dependent pressure drop, these gaps still have the contribution tothe viscous pressure drop. Nevertheless, more focus should be taken at the flux linesin the outer annular gap and outer radial gap since the structural configuration of outerannular and outer radial gaps have made these gaps formed as a parallel channel for theflux lines. Thus if flux-loss is neglected, the total number of magnetic flux that passedthrough the outer annular and outer radial gaps are equal to the number of magneticflux that passed through the inner radial gap. This is why the inner radial gaps havehigher magnetic flux density than the outer annular gaps or the outer radial gaps. Thedifference of magnetic flux density between the outer radial gaps and the inner radialgaps is also the reason why the radial gaps are divided into the outer and inner zones.

The distribution of magnetic flux density value along the flow path channel for0.5 mm gap size is shown in Figure 3.6 where the flux density is seen to be increasedto the augmentation of current input to the coil. In this study, the maximum current

50

Figure 3.6 Magnetic flux density along MR fluid flow path for 0.5 mm gap size withrespect to various current input

input is limited to 1.0 A while the maximum flux density that can be reached along theflow path channel is around 0.7 Tesla at the inner radial gaps. As a result of parallelmagnetic channel, the second highest flux density is reached at the outer radial gapsof around 0.5 Tesla while the flux density of the outer annular gaps is only around 0.2Tesla. The magnetic flux density values for each corresponding current input is used topredict the yield stress of MR fluid using Equation 3.1. Therefore, from these results,it can be predicted that the pressure drop of the inner radial zone will be the mostsignificant contributor to the total MR valve pressure.

3.3 Steady-state Modeling of Magnetorheological Valve

The assessment of MR valve performance is done using both simulationand experimental work. The simulation of electromagnetic circuit is one of themost important phases in the simulation work and is done using Finite ElementMethod Magnetics (FEMM) software, an open-source software package for magneticsimulation. The derivation of quasi-steady phenomenological model of MR valvedesign is also important in the first stage of simulation to predict the performanceof the MR valve. Typically, the quasi-steady model of the valve will be expressed interms of the achievable pressure drop, which is consist of two parts, pressure drop from

51

viscous properties of the fluid and pressure drop from the field dependent yield stressof the fluid. The basic expression of pressure drop in an MR valve can be declared byfollowing quasi-steady equations [24]:

∆P = ∆Pviscous + ∆Pyield (3.2)

∆Pviscous =6ηQL

πd3R(3.3)

∆Pyield =cτ (B)L

d(3.4)

where expressed that the pressure drop from viscous properties in Equation 3.3 isproportionally related to the fluid base viscosity (η), flow rate (Q), and annular channellength of the valve (L) but inverse-cubical to the valve gaps (d) and inverse proportionalto the channel radius (R). Meanwhile, the field dependent pressure drop in Equation3.4 is proportionally associated with the field dependent yield stress value (τ ) of MRfluid, annular channel length (L), and flow-velocity profile coefficient (c) but inverse-proportionally related to the gap size (d). The coefficient c is obtained by calculatingthe ratio between field dependent pressure drop and viscous pressure drop using theapproximation function as defined by [72] in the following equation:

c = 2.07 +12Qη

12Qη + 0.8πRd2τ (B)(3.5)

However, the expressions in Equations 3.3 and 3.4 are only valid for MR valvewith annular gaps. For radial gaps, the viscous pressure drop and the yield pressuredrop can be expressed as [24]:

∆Pviscous =6ηQ

πd3ln

(Ro

Ri

)(3.6)

∆Pyield =cτ (B)

d(Ro −Ri) (3.7)

where the valve gap size (d) in Equations 3.6 and 3.7 refers to radial gaps while theand refer to the outer radius of the radial gaps and the inner radius of the radial gapsrespectively.

The mathematical expression for the orifice gaps is slightly different, becauseno field dependent yield stress is incorporated into the equation since the fluxes did notpass through the orifice gaps. Therefore, the pressure drop equation for the orifice gaps

52

is only expressed by the equation of viscous resistance of the fluid [22].

∆P =8ηQL

πR4 (3.8)

Since the proposed MR valve design has both annular and radial gaps, themathematical expressions of pressure drop should include all the expressions inEquations 3.3 to 3.8. However, to make the derivation simpler, the valve gaps isseparated into five different zones, the outer annular gaps zone, the outer radial gapszone, the inner annular gaps zone, the inner radial gaps zone, and the orifice gaps zoneas shown in Figure 3.7. The valve needs to be clustered into five zones because eachzone was presumed to have different magnetic flux density. These five zones also couldalso be divided into two categories, the zones that categorized as effective area and thezones that have only viscous resistance. The zones that categorized as effective areaare the outer annular gaps zone, the outer radial gaps zone, and the inner radial gapszone while the zones that have only viscous resistance are the inner annular zone andthe orifice zone.

Figure 3.7 Gaps zone in MR valve with multiple annular and radial gaps

From Figure 3.7, it can be seen that there are two outer annular gaps, twoouter radial gaps, three inner annular gaps, four inner radial gaps and two orifice gaps.Using the expressions of annular and radial pressure drop in Equation 3.3 to 3.8, thequasi-steady pressure drop of MR valve with multiple annular and radial gaps can be

53

described in the following equations:

∆Pvalve = ∆Pannular outer + ∆Pradial outer + ∆Pannular inner

+ ∆Pradial inner + ∆Porifice (3.9)

∆Pannular outer = 2

[6ηQLaoπd3

aoRao

+caτa (B)Lao

dao

](3.10)

∆Pradial outer = 2

[6ηQ

πd3ro

ln

(R1

R0 outer

)+croτro (B)

dro(R1 −R0 outer)

](3.11)

∆Pannular inner = 36ηQLaiπd3

aiRai

(3.12)

∆Pradial inner = 4

[6ηQ

πd3ri

ln

(R1

R0 inner

)+criτri (B)

dri(R1 −R0 inner)

](3.13)

∆Porifice = 28ηQLo

πR0 inner4 (3.14)

The complete parameters for the MR valve with multiple annular and radialgaps are shown in Table 3.2 while the dimensions and the specific location of eachcorresponding variable of the proposed MR valve are shown in Figure 3.8.

Table 3.2: List of MR valve parameterParameter Unit Value

η (MRF-132DG) Pa s 0.092Q ml/s 40-80dao mm 0.5

dr = dro = dr mm 0.5dai mm 1.0Lao mm 10.0Lai mm 6.0-2drLo mm 5.0R1 mm 6.5

R0 outer mm 3.0R0 inner mm 2.5

54

Figure 3.8 Dimension and variables of MR valve

3.4 Simulation of Magnetorheological valve Performance

Figure 3.9 illustrates the results of maximum achievable pressure drop of themultiple annular and radial MR valve at 40 ml/s flow rate with gap size of 0.5 mmand current input 1.0 A is around 5.3 MPa. As a comparison, the achievable pressuredrop of the previous design of compact MR valve at 40 ml/s flow rate is only around 2MPa [20,24]. In other words, the proposed MR valve design has shown the contributionto improve the achievable pressure drop of a compact MR valve. Furthermore, it isshown that the pressure drop from the inner radial gap zone is the highest contributor tothe total achievable pressure drop of the MR valve. Surprisingly, although the magneticflux density of the outer annular zone was already known to be the lowest from theother effective area, the maximum pressure drop of the outer annular zone is higherthan the outer radial zone. Meanwhile, the total viscous pressure drop of the MR valveis shown to be nearly 10 MPa which mostly contributed from the outer annular zone,followed by the inner radial zone and the outer radial zone. Since the contribution ofviscous pressure drop from the inner annular zone and the orifice zone are relativelysmall, these zones are not included in the graph in Figure 3.9. The percentage ofcontribution from each zone to the MR valve pressure drop is illustrated in Figure3.10.

55

Figure 3.9 Estimation of achievable pressure drop of MR valve with 0.5 mm gap size

The illustrations of pressure drop percentages from each zone in Figure 3.10can be used to explain why the pressure drop at the outer annular zone can outperformthe outer radial zone although the magnetic flux density in the outer annular zone islower than the outer radial zone. The viscous pressure drop of outer annular zone hasbeen shown to be the highest contributor of viscous pressure drop of the MR valve. Themagnitude of viscous pressure drop at the outer annular zone has contributed morethan half of the total viscous pressure drop of MR valve while the outer radial andthe inner radial zone have only contributed around 11% and 28% respectively. Thechannel length, channel radius, gap size and fluid flow rate, as shown in Equation 3.10,

(a) (b)

Figure 3.10 Percentage of pressure drop contribution from each zone (a) viscous (b)field-dependent at 1 A current input

56

are the variables that responsible in determining the magnitude of the viscous pressuredrop at the outer annular zone. However, among them, gap size is known as the mostsignificant variable since it inverse-cubically influences the viscous pressure drop.

As the most significant variable that determines the viscous pressure drop, theeffect of gap size variations on the pressure drop of MR valve should be evaluated.Since the flow path has been divided into five zones, there are five different gaps to bevaried. However, to simplify the combinations, the gap size for both outer radial andinner radial zones are determined to be equal (dro = dri) and varied from 0.5 mm to2.0 mm. The same variations and interval are also used at the outer annular gap size.Meanwhile, for zones that have only viscous resistance such as the inner annular gapand the orifice gap, the gap size is kept constant. The effect of each combination on theviscous and field dependent pressure drop of MR valve are shown in Figure 3.11:

Figure 3.11a and 3.11b show that the gap size has a very significant effectto the magnitude of viscous pressure drop as well as the magnitude of field-dependentpressure drop. However, the most significant influence of the annular gap to the viscouspressure drop of the MR valve is seen in the interval between 0.50 to 0.75 mm wherethe viscous pressure drop of the MR valve could be reduced in nearly 0.4 MPa.Obviously, the change of annular gap also has some effect to the magnitude of field-dependent pressure drop. Figure 3.11b shows that the increase of annular gap sizefrom 0.50 to 0.75 mm reduces the field-dependent pressure drop of around 0.8 MPa.Meanwhile, the enlargement of radial gap seems to have more influential effect to thereduction of field-dependent pressure drop magnitude than to the viscous pressure dropmagnitude. From Figure 3.11b, it can be observed that increasing radial gap size from0.50 to 0.75 mm reduces the viscous pressure drop of only 0.25 MPa but responsibleto the decrease of field-dependent pressure drop for almost 2.5 MPa.

The results in Figure 3.11 also illustrate an interesting fact regardingrelationship between gap size and the magnitude of field-dependent pressure drop.As shown in Equations 3.10, 3.11 and 3.13, the relationship rule between gap size andfield-dependent pressure drop are inverse-proportional while the relationship betweengap size and viscous pressure drop are inverse-cubical. Therefore, any modifications tothe gap size supposed to have more significant effect to the viscous pressure drop thanthe field-dependent pressure drop. However, the results in the Figure 3.11 show that theeffects of gap changing to the pressure drop apparently are not consistently followedthe rule stated in Equations 3.10, 3.11 and 3.13. For example, the radial gap reductionfrom 0.50 to 0.75 mm degrades the field dependent pressure drop around 40% but

57

(a)

(b)

Figure 3.11 Effect of gap size on the pressure drop (a) viscous (b) field-dependent at1 A current input

58

only reduces viscous pressure drop about 27%. The reason of this can be explainedby considering the effect of gap size changing to the magnetic circuit performance.Since the magnetic resistance of MR fluid in the gaps is a lot higher than the magneticresistance of magnetic material in casings and cores, small change in the gap size willsignificantly change the magnetic circuit and influence the magnetic flux density [24].Therefore, although in Equations 3.10, 3.11 and 3.13 the gap size seems to onlyinverse-proportionally related to the field-dependent pressure drop, in actual, the gapsize also influences the field-dependent pressure drop by simultaneously changing themagnitude of field-dependent yield-stress of MR fluid.

However, in terms of overall MR valve performance, high achievable field-dependent pressure drop only is not sufficient to show that the MR valve is workingin a decent performance. Since, MR valve with high field-dependent pressure dropbut also with high viscous pressure drop value means that the operational range of MRvalve is low. The operational range of MR valve can be determined using the followingexpression.

λ =∆Pyield

∆Pviscous(3.15)

Equation 3.15 expressed the operational range in terms of amplification ratio ofpressure drop that can be generated using MR valve. For instance, it can be said that theoperational range is high if the field-dependent pressure drop is high and the viscouspressure drop is low while the operational range is low if the field-dependent pressuredrop is low and the viscous pressure drop is high. The expression is an inverse equationto the valve ratio term in [161] which defined as the ratio of the viscous pressure dropto the field dependent pressure drop of the MR valve. As it is desirable to have a smallvalve ratio, it is also desirable to have a large the operational range of MR valve. Theexpression in Equation 3.15 is also similar with the non-dimensional parameter calledBingham number which defined as the ratio of the dynamic field-dependent yield stressto the viscous shear stress [133]. Bingham number is a very useful parameter and hasbeen used as a generic term to analyze the performance range of various MR devicessuch as dampers and valves [162–164]. However, in order to calculate the Binghamnumber, specific value of yield stress, fluid velocity and gap size are required whichmakes Bingham number calculation for MR valve with gap types combination moredifficult than by just comparing the achievable pressure drop.

From Figure 3.12, it can be observed that the operational range of the valve

59

Figure 3.12 Comparison of operational range between various gap configurations

consistently increased when the gap size at the outer annular zone is increased. Themost significant operational range improvement occurred when the outer annular gapsize is increased in the interval between 0.50 to 0.75 mm. Since the increase ofannular gap size provides more significant impact to the viscous pressure drop thanthe field-dependent pressure drop, consequently, the occurrence of operational rangeaugmentation in this interval can be considered as a result of significant viscouspressure drop reduction. In the contrary, it is clear that the increase of radial gapsize will have inverse relationship with the operational range since the increase ofradial gap size will cause significant degradation to the field-dependent pressure drop.However, some exceptions are found for the combinations of 1.0, 1.5 and 2.0 mmannular gap sizes with radial gap size in the interval between 0.50 to 0.75 mm. Theanomaly of trend in these intervals appears because the performance degradation ofthe viscous pressure drop in these gap combinations are higher than the performancedegradation of the field-dependent pressure drop. The higher viscous pressure dropdegradation is also shown to be increased in the combination with higher annular gapsize. However, as the size of radial gap increased, the dominance of field-dependentpressure drop degradation is started to overlap the degradation of viscous pressuredrop. Nevertheless, it still can be concluded that the increase of annular gap sizegenerally has positive effect to the operational range while the increase of radial gapsize tend to reduce the operational range. In this case, due to trend anomaly andsignificant impact to the change of pressure drop value, the gap size interval in between0.5 to 1.0 mm require more attention and will be further explored in the experimental

60

assessment.

As a benchmark with previous MR valve concept, Table 3.3 show a comparisonbetween the performance of the meandering flow path MR valve with the other typesof MR valve. The source of the comparison are combined from several literatures[21,24,165,166] and compared in terms of the ratio between the maximum achievablepressure drop of each valve with its corresponding power consumption and gross valvesize (volume). The valve performance can be considered more efficient if the ratioof pressure drop to power and volume are higher. According to the table it can beconcluded that the MR valve with meandering flow path structure has a very significantimprovement over the annular and radial types of MR valve in terms of pressuredrop per power value (more than 100 times improvement). The meandering flow pathconcept also show enhancement over the annular-radial concept with nearly 150%improvement of the pressure drop to volume ratio.

Table 3.3: Performance benchmarking between the proposed MR valve concept andthe counterparts

200 kN 400 kN 250 kN Annular- ProposedSpecification MR damper MR damper MR damper Radial MR MR valve

[165] [166] [21] valve [24]Valve type Annular Annular Radial Annular-Radial MeanderingMaximumpressure 3.8 16.1 9.6 2.5 5.3

drop (MPa)MaximumPower (W) 262 1920 900 -NA- 3

Pressuredrop per

Power input 14.5 8.39 10.7 -NA- 1766(kPa/W)

GrossVolume 1 31 19 0.1 0.15(dm3)

Pressuredrop perVolume 3.8 0.52 0.48 24.87 35.08

(kPa/cm3)

61

3.5 Summary of Chapter 3

A new concept for a MR valve with meandering flow path was presented inthis chapter. The new MR valve is presented by elaborating the conceptual designof the valve followed by the explanation about the consideration of the designand magnetic field simulation. The derivation of the MR valve steady-state modeland the performance prediction of the MR valve were also presented in separatesection. The novelty of the new MR valve concept is mainly justified by the uniquearrangement of the effective area which makes it capable to deliver higher pressuredrop capacity than its counterparts. The steady-state modeling and simulation work ofthe achievable pressure drop in accordance to the viscous and field-dependent pressuredrop contributions have been conducted to predict the achievable performance of theMR valve. The effect of gap size selection in the annular and radial channel to themagnitude of viscous and field-dependent pressure drop also has been analyzed toevaluate the significance of each channel to the overall performance of the MR valve.According to the assessment results, the outer annular channel was shown as the mostinfluential region to the characteristics of viscous pressure drop. Meanwhile, the innerradial channel was more influential to determine the characteristics of field-dependentpressure drop. The assessments of the gap size effect in each regions were alsoanalyzed in terms of dynamic range of MR valve. The gap size of annular channel hasbeen reported to have positive correlation with the dynamic range of MR valve whilethe gap size of radial channel were reported to have inverse effect. The knowledge ofthese characteristics is important as a reference to perform the experimental testingthat will be discussed in the next chapter.

CHAPTER 4

EXPERIMENTAL ASSESSMENT

4.1 Introduction

The advantage of the meandering flow path concept has been showed inChapter 3 through the comparison between the achievable pressure drop of themeandering valve concept with its predecessors, which typically has maximumpressure drop up to 2.5 MPa. In this chapter, the concept will be experimentallyexplored to prove and demonstrate its performance capability in improving pressuredrop. This chapter is arranged in three sections, which is started with the explanationabout the experimental apparatus, followed by the elaboration of the experimental set-up, which explains the measurement methodology, and finally the discussion aboutthe experimental results. The chapter is discussed in four different subsections, theoff-state characteristics, the on-state characteristics, the analysis about the frequencyexcitations effect, and the analysis about the current input effect. This chapter is endedby the last section that concludes the chapter.

4.2 Experimental Apparatus

4.2.1 Magnetorheological Fluid

The specific MR fluid that is used in this study is the MRF-132DG, an MRfluid made by Lord Corporation for general use in controllable, energy-dissipatingapplications such as shocks, dampers and brakes [167]. The fluid consists ofmicron-sized magnetizable particles in hydrocarbon fluid. According to [168], themagnetizable particles contained in the fluid are between 1 to 20 microns with weightratio of 80.98%. The typical properties and material compatibilities of MRF-132DG

63

are listed in Table 4.1 while its magnetic properties, presented in the B-H curve, andthe τ curve are shown in Figure 4.1 and 4.2 respectively.

Table 4.1: Typical properties and material compatibility of MRF-132DG [167, 168]FLUID PROPERTIESBase Fluid HydrocarbonSolid Content by Weight, % 80.98Density, g/cm3 (lb/gal) 2.95-3.15 (24.6-26.3)Operating Temperature,C (F) -40 to 130 (-40 to 266)Flash Point,C (F) >150 (>302)Viscosity, Pa-s @ 40C (104F) 0.092 ± 0.015calculated as slope 800-1200 sec−1

MATERIAL COMPATIBILITIESBuna N (Nitrile) GoodButyl PoorEPDM/EPR PoorFluoroelastomer GoodNatural Rubber PoorNeoprene GoodSilicone FairIron GoodStainless Steel GoodAluminum GoodPolyurethane Good

Figure 4.1 B-H curve of the MRF-132DG [167]

64

Figure 4.2 Field induced yield stress of the MRF-132DG [167]

4.2.2 Magnetorheological Valve

As mentioned earlier in Section 3.2.1, the MR valve structure is divided intothree main components, which are the casing, the coil and the valve core. The casing,which should be made from magnetic material, is made from mild steel that compatiblewith AISI 1010 and consists of two identical parts, which each of them also has afluid channel with female BSPP 1/4´ port and four holes for locking bolts. The coilconsists of copper wire windings and an aluminum coil bobbin, which has a grooveon each side for O-ring installation as a sealing mechanism. The coil bobbin also hasfour M2 threaded holes in each side as a bolt-locking mechanism with the casing.Viton is selected as the O-ring material rather than rubber due to poor compatibility ofrubber with hydrocarbon as the carrier liquid of the MRF-132DG (see Table 4.1). Theselection of Viton is referred to several references [50, 169], which utilized the samematerial for their sealing for hydrocarbon based MR fluid. The last component is thevalve core which consists of several parts namely two mild steel side cores, two mildsteel orifice cores, a mild steel centre core and six aluminum spacers. The spacers areused to maintain the clearance between core parts. The exploded view of the MR valvedesign is illustrated in Figure 4.3.

In order to experimentally explore the achievable pressure drop of the MR valvewith meandering flow path, only four different gap size configurations of the MR valvethat are experimentally assessed in this work, namely the MR valve with with 0.5 mm

65

Figure 4.3 Exploded view of the MR valve prototype

radial and 0.5 mm annular gaps, the 0.5 mm radial and 1 mm annular gaps, the 1mm radial and 0.5 annular gaps, and the 1 mm radial and 1 mm annular gaps. Thesegap configurations, according to the simulation results in Section 3.4, were reported topotentially have the highest achievable pressure drop.

Figure 4.4 Failure of the bolt-locking mechanism to withstand internal pressure

66

However, after some initial tests, it was found that the design has weaknessin the strength of bolt-locking mechanism between the casing and the coil bobbin.As shown in Figure 4.4, the M2 female thread in the aluminum bobbin is apparentlynot strong enough to hold the casings together under high internal fluid pressure whichcaused failure. Therefore, the modification is made to the design, especially to improvethe strength of casings assembly joint. The modifications is made by changing the bolt-locking mechanism from the casings to the aluminum bobbin to a directly-coupledthread between casings as shown in Figure 4.5. With the modified design, betterstrength of casing assembly can be achieved and less component is required in thevalve structure.

(a)

(b)

Figure 4.5 Modification and comparison of the MR valve prototype(a) Exploded viewof the MR valve design (b) Fabricated prototype of MR valve

67

4.2.3 Testing cell

MR fluid flow need to be induced acrossed the MR valve,in order to observethe dynamic behavior. There are various methods that have been used by researchersto induce flow of MR fluid such as demonstrated in [19,24,170], however, the methodchosen to induce fluid flow is similar with the method used in [20], where the MR valveis installed in the bypass channel of a hydraulic cylinder, which is fully filled withMR fluid. The fluid flow was then induced by introducing movement to the hydrauliccylinder, which act as the MR valve testing cell, using dynamic test machine. In thisstudy, the testing cell is made from a double rod hydraulic cylinder with maximumstroke length of 70 mm, bore size of 30 mm and rod diameter of 18 mm (the netpiston area is around 452.4 mm2). The double rod cylinder is used as the testing cellto eliminate the requirement of accumulator since no volume compensation is neededduring operation. The prototype of MR valve with meandering flow path is installed inthe testing cell as shown in Figure 4.6.

Figure 4.6 MR valve installation in the testing cell

For instance, the arrangement of the MR valve in the testing cell has made thedevice looks similar with a bypass linear damper. The main difference, in this case, isthat the dynamic force-velocity behavior of the testing cell is not the point of interestssince the behavior of the valve is mainly related to the valve pressure drop and the fluid

68

volumetric flow rate. In order to measure the valve pressure drop, two pressure sensorsare installed in the nearest point to both MR valve ports to ensure that the pressuredrop effect from the flow conduits does not alter the pressure measurements. Thevolumetric flow rate is measured indirectly by multiplying the measured velocity ofthe piston and the effective piston area. The velocity data is obtained by differentiatingthe displacement data of the piston that is measured by the native displacement sensorof the dynamic test machine.

4.3 Experimental Set-up

A simultaneous measurement of pressure drop generated by valve inaccordance to the variations of MR fluids flow rate and magnetic field is conductedto measure the performance of the MR valve. In order to generate the flow, the testingcell is actuated by the dynamic test machine and the variation of fluid flow can berecorded through the measured stroke of the testing cell. Meanwhile, the variation ofmagnetic field is represented by the changing of the current input to the coil.

The experimental setup for the MR valve performance measurement is shownin Figure 4.7 where the pressure drop of the MR valve is measured as well as thedisplacement of the testing cell. The measured displacement data then converted toflow rate by differentiating and multiplying the data with the piston net cross-sectionalarea. The realization of these setup is shown in Figure 4.8 where the MR valve isinstalled in the testing cell with pressure sensors (Parker Hannifin PTDSB1001B1C1,0-10 MPa) tapped in each port of the valve. An in-house steel platform powered bya geared electric AC induction motor equipped with an adjustable crank to generatesinusoidal movement in the platform. An inverter is used to control the rotationalspeed of the electric motor while a wire displacement transducer (Celesco SP2-25)measures the length of platform displacement. All measurement data from the sensorsare connected to the computer through the NI cDAQ-9174 and LabView.

The platform is used to provide the oscillatory movement to the testing cellwhich will result in pressure changing inside the cylinder and will periodically pumpthe MR fluid to flow through the MR valve. The pressure sensor measures the real-time changing of the fluid pressure in each port of the valve. The pressure differencebetween the inlet port pressure and the outlet port pressure represents the MR valvepressure drop. In order to form the relationship between pressure drop and flow

69

Figure 4.7 Experimental arrangement schematic for the MR valve testing

rate as well as the current input, the measurement was conducted in three differentinverter-controlled frequencies namely 5.0 Hz, 7.5 Hz, and 10 Hz, which, accordingto the measurement, are approximately equal to peak flow rate of 37 ml/s, 55 ml/sand 73 ml/s. The peak flow rate and its associated pressure drop will be plotted toform the relation. The approach is similar with the common method in any dampercharacterization process to form the force-velocity curve [154].

Since the force induced to the testing cell is also proportional to the pressuredrop of the valve, the larger the pressure drop is, the higher the force required tomove the testing cell. Unfortunately, the in-house dynamic test machine, which usethe inverter-controlled AC motor as the prime-mover of the platform, is not able toprovide sufficient force to move the testing cell in low frequency excitation. Thespeed reduction of the motor with inverter control is proportional to the reductionof motor power, which obviously decrease the torque of the motor. Therefore, evenwhen the platform is moving, it is difficult to keep the excited frequency in constantvalue, especially during the increased pressure drop of the MR valve. As a result, themeasurement range of the in-house dynamic test machine is very limited.

Due to these limitations, the results of the in-house test machine will onlybe used to evaluate the peak off-state pressure drop of the valve and the peak on-state pressure drop. The dynamic pressure drop, on the other hand, is evaluated with

70

Figure 4.8 MR valve testing set-up using in-house test machine

Shimadzu Hydraulically actuated Fatigue Dynamic Test Machine. The machine isequipped with a 20 kN force sensor and a displacement sensor. The installation ofthe testing cell in the Shimadzu Hydraulically-Actuated Fatigue Dynamic Machineis depicted in Figure 4.9. Similarly with the installation arrangement in the in-housetest machine, the alternating pressure-induced flow is generated every time the rod iscompressed and extended by the dynamic machine. The magnitude of the flow rate isassumed to be proportional to the velocity of the piston strokes so that the velocity ofthe actuation can represent the flow rate variation.

During experimental testing with the Fatigue Dynamic Test Machine, thetesting cell is subjected to a sinusoidal wave, where one end of the testing cell isfixed and the other end is actuated. Since the sinusoidal wave can be expressed inu = A sin (2πft)), the amplitude, A, is fixed to +/- 25 mm, while the excitationfrequency, f , is varied from 0.50 to 1.00 Hz. In order to demonstrate the MR effect,the magnetic field strength is also varied by the changing the magnitude of the currentinput. The summary of variable arrangement for the experimental test is shown in Table4.2.

71

Figure 4.9 Testing cell installation in the Shimadzu Fatigue Dynamic Test Machine

Table 4.2: The variable arrangement of experimental test using Shimadzu FatigueDynamic Test Machine

Excitation Frequency Excitation Amplitude Current Input Cycles(Hz) (mm) (Ampere)0.50 ± 25 0-1 250.75 (interval 0.1)1.00

4.4 Experimental Results

4.4.1 Off-state and On-state Pressure Drop Characteristics

The off-state pressure drop can be defined as the pressure drop generated by theMR valve when the electromagnet is charged with 0.0 A current. In the other words,the off-state pressure drop is the pressure drop with only viscous contribution withoutany magnetorheological effect from the MR fluid. The peak value of the measured off-

72

state pressure drop across the MR valve with meandering flow path as a function offlow rate is shown in Figure 4.10.

Figure 4.10 Comparison of measured and theoretical off-state peak pressure drop atvarious flow rates for 0.5-0.5 mm (annular-radial) gaps configuration

Figure 4.10 shows that generally the trend of the theoretical off-state pressuredrop value is in accordance with the measured one, in the range of flow ratemeasurement between 37-73 ml/s, the pressure drop increased proportionally to theflow rate. In specific, the exact value of the measured off-state pressure drop is slightlylower than the theoretical off-state pressure drop from the model if the assumedviscosity value is equal to 0.092 Pa s, which is the specified viscosity value of theMRF-132DG from the manufacturer. However, since according to the documentationfrom Lord Corp. [168], the exact viscosity value also can be varied up to +/- 0.015Pa s from the specified value, thus the theoretical off-state pressure drop value alsocan be dropped lower than the measured value if lower viscosity value is selectedin the calculation. If the range of viscosity value is taken into consideration, then thetheoretical off-state pressure drop as a function of flow rate will not form into a straight-line curve, but will be more like a band that representing the range of predicted off-state pressure drop values. Figure 4.10, in this case, shows the agreement of the rangeof predicted off-state pressure drop values with the measured off-state pressure dropvalues in the measurement range between 37-73 ml/s.

73

The on-state pressure drop is the pressure drop that is generated when thevalve coil is charged with a current. The performance of the valve during the on-statecondition is measured in the range of current input between 0.2 to 1.0 A with 0.2 Ainterval. The relationship between the measured on-state peak pressure drop and thecurrent input for three different flow rate variations, along with its comparison to thepredicted value from the model are shown in Figure 4.11.

Figure 4.11 Comparison of measured and theoretical on-state peak pressure drop atvarious current inputs and flow rates for 0.5-0.5 mm (annular-radial) gaps configuration

In Figure 4.11, the trend line agreements of the pressure drop to the currentinput curve between theoretical and measurement are shown. Due to the powerlimitation of the test rig prime mover, the measurement data for 37 ml/s flow rateare only able to be recorded up to 0.6 A current input. From Figure 4.11, the measuredpressure drop values are shown smaller than the theoretical value when the currentinput is lower than 0.8 A. However, during and above 0.8 A, the measurement resultsfrom the experiment show higher pressure drop than the predicted one. The reasoncan be related to the explanation of the off-state pressure drop behavior where theviscosity value plays an important role in the accuracy of the predicted value. Becausealthough the graph in Figure 4.11 is showing the on-state performance of the valve,the off-state pressure drop is actually still exists as an initial pressure drop. However,the percentage of off-state pressure drop to the total on-state pressure drop will be

74

varied to the magnitude of the current input. Typically, the percentage of the off-statepressure drop is the highest at the lowest current input but will be getting lower as thecurrent input is increased. For example, at the 73 ml/s theoretical curve, the percentageof the off-state pressure drop is around 62 % for the total on-state pressure drop at0.2 A current input but dropped to 23 % at 1.0 A current input. Therefore, it can beconsidered that the effect of the base viscosity value in the total pressure drop is stillquite significant in the measured pressure drop below 0.8 A.

However, the superiority of the measurement results over the predicted resultsin the current input higher than 0.8 A is quite surprising. Normally, the measured on-state pressure drop will be lower than the estimated pressure drop due to unpredictedflux loss and prediction error in the field strength value of the FEM software. Thereason is not clear now, but several causes can be the source of the phenomenon. Thefirst cause can be assumed as a result of the thickening effect of the fluid, which inseveral literatures were mentioned to be occurred due to spalling of the particles insidethe MR fluid [12, 22]. The thickening effect of the fluid is currently not considered inthe model since the MR valve model is based on the Bingham fluid model. However,the thickening effect can be modeled if the Herschel-Bulkley fluid model is used asthe basis of the MR valve modeling [74]. The second cause is assumed due to theenhanced yield stress phenomenon due to aggregation process [4], which occurredwhen several single-chain structure of magnetic particle joined together and formeda thick column. As discussed by Tao [5], the yield stress enhancement of MR fluidcan be done by changing the microstructure arrangement of magnetic particle duringon-state condition. As already shown in Figure 4.2, The MRF-132DG that is used inthis study has saturated yield stress value in around 48 kPa. However, the relationshipcurve between yield stress and the field strength is derived based on the assumptionthat the particles are arranged in a single chain structure [94]. If the microstructurearrangement of magnetic particle inside the MR fluid is no longer arranged into asingle chain structure but rather a multiple chain structure, the yield stress value of theMR fluid can be potentially enhanced. However, the validity of this assumption needfurther confirmation since the occurrence of multiple chain structure that was discussedin [4] and [5] is, so far, only reported in compression mode, which is possible due tofluid-particle separation phenomena [171]. The yield enhancement in valve mode, tothe knowledge of the author, has not been reported yet. Nevertheless, the MR valveconcept with meandering flow path is already proving an improvement by deliveringhigher achievable pressure drop than the previous concept of MR valve.

75

4.4.2 Effect of Gap Size

The testing results for all gap combinations at flow rate of 37 ml/s are shownin Figure 4.12. It is shown that the pressure drop of all gap size combinations isconsistently increasing with the increase of current input. The slope, however, isslightly gradual in the condition where the current input is lower than 0.2 A. Thedifference in the slope is actually expected considering the nonlinearity of the fluidyield stress curve provided by the fluid manufacturer as shown in Figure 4.2. From thecurve, it is shown that the yield stress is only appeared when the flux density value isabove 0.08 Tesla. Meanwhile, from the previous Section 3.4, it has been shown thatthe flux density of the MR valve at the effective area with the excitation current to thecoil below 0.2 A are below 0.1 Tesla. Due to that reason, the slope of the pressuredrop curve for the MR valve with all gap size combinations below 0.2 A of currentexcitation is not as steep as the slope of the curve above 0.2 A.

Furthermore, it can be seen that the smaller the gap size, the higher the pressuredrops of the MR valve. This effect is obvious, because the smaller the gap, the higherthe effect of wall friction to the fluid, which increased the viscous pressure drop andfield-dependent pressure drop simultaneously. However, a specific note should be takenthat the change of the gap size of the radial channel provide more significant effect tothe pressure drop than the gap size of the annular channel. This is the proof of thesimulated performances that have been presented Section 3.4, which can be explainedin two reasons. The first reason is regarding the number of the radial channel, whichis twice the number of the annular channel, while the second reason is regarding thehigher flux density at the radial channel than the flux density at the annular channel.

Another important aspect that can be evaluated from the data in Figure 4.13 isthe dynamic range of the MR valve. The dynamic range of MR valve is defined as theratio between the on-state and the off-state pressure drops of the valve as described inEquation 3.15. Using the expression in Equation 3.15, the dynamic ranges of the MRvalve for each gap size combination are shown in Figure 4.13. The dynamic range isan important variable as an indication that the valve can be controlled in a wide rangeof performance. Ideally, a good MR device is expected to provide the highest on-stateperformance while capable to suppress the off-state to the lowest value.

Based on the dynamic range values shown in the Figure 4.13, it can beconcluded that among four types of gap size combinations, the MR valve with 1.0mm annular gap size will always have a higher dynamic range than the MR valve with

76

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

1

2

3

4

5

6

7

Current Input (A)

Pre

ssur

e D

rop

(MP

a)

0.5 mm annular − 0.5 mm radial1.0 mm annular − 0.5 mm radial0.5 mm annular − 1.0 mm radial1.0 mm annular − 1.0 mm radial

Figure 4.12 Comparison of measured peak pressure drop in various gap sizecombinations

0.5 mm annular gap size. On the contrary, the increase of radial gap size, although notas significant as the change in annular gap, has a weakening effect to the dynamic rangeof the MR valve. Both of these effects are in agreement with the prediction shown inSection 3.4. The knowledge of both factors, the achievable pressure drop value and thedynamic range, will be useful to decide which gap size combination to be used whendesigning an MR valve for a particular application.

4.4.3 Effect of Current Input Variation

Figures 4.14a to 4.14c detail the characteristics of valve with current inputvariations at 0.50 Hz, 0.75 Hz and 1.00 Hz of frequency excitations. As predicted, theMR valve with meandering flow path is consistently showing pressure drop adjustmentwith the variation of the current input. As seen in Figure 4.14a, the peak pressure dropis around 0.65 MPa at the current of 0.0 A with 0.50 Hz excitation frequency. Whenthe current increases in the interval of 0.1 A, the peak pressure drop are also steadilyincreased with average increments of 0.58 MPa up to the maximum pressure drop ofaround 6.42 MPa at the current of 1.0 A. Similarly, in Figure 4.14b, at the current of 0.0

77

Figure 4.13 Comparison of the MR valve dynamic range for each gap sizecombinations

A with 0.75 Hz excitation frequency the maximum pressure drop that can be achievedis around 0.95 MPa. However, with 0.1 A of current increment, the pressure drop arenormally increased in around 0.56 MPa until the maximum observed pressure dropof 6.53 MPa is achieved at 1.0 A current. Lastly, at 1.00 Hz excitation frequency asshown in Figure 4.14c, the maximum off-state pressure drop is around 1.35 MPa withthe average increments of pressure drop are seen around 0.55 MPa with each 0.1 Aincrement of current so that the peak pressure drop at 1.0 A is measured at 6.86 MPa.The relationships between peak pressure drop in each current increment for 0.50 Hz,0.75 Hz and 1.00 Hz excitation frequency are depicted in Figure 4.15.

As shown in Figure 4.15, the rise of peak pressure drop is almost linear tothe increment of the current input. However, there is an interesting phenomenon thatcan be observed from the curves where the slopes between 0.0 to 0.2 A and 0.8 to1.0 A are slightly lower than the other regions. The phenomenon can be explainedusing the curve of yield stress versus magnetic field strength of the MRF-132DG inFigure 4.2 where the slope is leaner at the lower magnetic field strength value and alsoleaner when it almost reached saturation in the higher magnetic field strength value.

78

−40 −30 −20 −10 0 10 20 30 40−8

−6

−4

−2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

0.0 A

1.0 A0.9 A0.8 A

0.7 A

0.6 A0.5 A

0.4 A

0.3 A

0.2 A

0.1 A

(a)

−60 −40 −20 0 20 40 60−8

−6

−4

−2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

0.0 A0.1 A

0.7 A

0.2 A

0.3 A

0.4 A

0.6 A

0.5 A

0.8 A0.9 A1.0 A

(b)

−80 −60 −40 −20 0 20 40 60 80−8

−6

−4

−2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

0.2 A

1.0 A

0.1 A0.0 A

0.3 A

0.4 A

0.5 A0.6 A

0.7 A

0.8 A0.9 A

(c)

Figure 4.14 The pressure dynamics of MR valve at various current input for 0.5-0.5mm (annular-radial) gaps configuration, (a) 0.50 Hz (b) 0.75 Hz (c) 1.00 Hz

79

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−8

−6

−4

−2

0

2

4

6

8

Current Input (A)

Pre

ssur

e D

rop

(MP

a)

0.50 Hz0.75 Hz1.00 Hz

Figure 4.15 The trend of peak pressure drop at various current input for 0.5-0.5 mm(annular-radial) gaps configuration

Although, the declaration that the 1.0 A current input is already reached saturationregion would be too premature, since the magnetic simulation results in Section 3.2.3showed that the peak magnetic flux density at 1.0 A current input is only around 0.85Tesla. The saturation region of the yield stress, according to Figure 3.3 is started toappear above 0.9 Tesla. However, the slope of the yield stress curve is already seento gradually decrease after the magnetic flux density reached 0.7 Tesla. In the otherwords, it is valid to predict that the attempt to increase the current input above 1.0 Awill not give any significant improvement to the pressure drop performance of the MRvalve, due to the yield stress saturation of the MR fluid.

4.4.4 Effect of Excitation Frequency Variation

The relationship between valve pressure drop against flow rate of the MR fluidsat the coil current of 1.0 A at various frequencies is shown in Figure 4.16. The resultsare the typical characteristics from the fifth cycle onwards since the characteristics ofthe first until the fourth cycle are usually unstable and inconsistent as similarly reportedby Li et al. [172]. From Figure 4.16, it can be concluded that the peaks of the valvepressure drop are increased with the increase of maximum flow rate. Similarly, theshifting of peak pressure drop also shifts the width of the hysteretic region. In otherwords, the width of the hysteretic region of pressure drop is highly dependent to thevalue of peak flow rate as typically found in the characteristics of MR damper in itsanalogous force-velocity pattern.

80

−80 −60 −40 −20 0 20 40 60 80−8

−6

−4

−2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

0.50 Hz0.75 Hz1.00 Hz

Figure 4.16 The pressure dynamics of MR valve at current input of 1A at variousfrequency excitation for 0.5-0.5 mm (annular-radial) gaps configuration

There is an interesting note regarding the bend of the curve shape in each time itpasses through the 0 MPa of pressure drop. The bends of the curve were interpreted asthe pressure lag effect similarly with the ones reported in the behavior of MR dampersby Yang [173] and Zhang et al [174] which caused by the uncompressed air pocketthat inevitably exists in the testing cell. The existence of air pocket inside the cylindercreates force lag during the transition between the positive and negative direction ofthe piston. In fact, these effect were not a unique case in the MR fluids device since itsappearance was also reported in the characteristics of a passive hydraulic damper [175].In order to minimize the air pocket effect, a vacuum technique is often used in thefluid filling process as well as the utilization of the accumulator to pre-compress theair pocket in the damper assembly [123, 173]. However, since no report has stronglymentioned about side effects of the air pocket to the overall damping characteristicsother than just the force lag, in this study, the pressure lag effect will not be taken intoconsideration.

4.5 Summary of Chapter 4

The elaboration of experimental assessment procedure and the performanceevaluation of MR valve with meandering flow path has been presented in this chapter.The experimental assessment has been elaborated in terms of the materials andequipments involved in the testing, the MR valve prototype and the testing methodto assess the MR valve performance. The performance of the MR valve is measuredusing testing cell installed in an inhouse dynamic test platform and a commercial

81

type Fatigue Dynamic Test Machine. The experimental results are assessed in termsof off-state pressure drop characteristics, on-state characteristics, the effect of currentvariations and the effect of frequency excitation to the MR valve performance. Themeasurement results of the off-state characteristics concluded that the prototype coulddeliver off-state pressure drop in the range that is already predicted by the steady-state model, while the on-state characteristics of the prototype with both 0.5 mmannular and radial gaps proved that the concept successfully increased the achievablepressure drop of MR valve more than 6 MPa. The effect of flow rate and currentvariations to the dynamic characteristics of the MR valve were also demonstratedthrough the experimental assessment where the increment of flow rate and currentresult in consistent increment of the pressure drop. In addition the hysteretic behavior,which is also captured in the measurement, can be used as a source to develop thedynamic model of the MR valve which will be covered in the next chapter.

CHAPTER 5

HYSTERESIS MODELING OF MAGNETORHEOLOGICAL VALVE

5.1 Introduction

The experimental results of MR valve have showed that the relationshipbetween pressure drop and fluid flow rate exhibit hysteretic behavior, which cannot becovered in the steady-state model. In order to represent the hysteretic behavior, a newmodel has to be derived based on the experimental data. In this chapter, two types ofparametric hysteresis model, the polynomial model and the modified LuGre hysteresismodel are evaluated. The performance of both model are assessed by comparing themodel results with the experimental data.

5.2 Polynomial-based Hysteresis Modeling Approach

Polynomial based modeling, as one of the parametric hysteresis modelingapproach for MR damper, is chosen as the first approach to model the hysteretic effectof MR valve in this study. The polynomial model was first proposed by [147] to modelthe hysteresis behavior of MR damper. The method also employed by [148] in themodeling of MR damper to demonstrate the force tracking control of MR damper.

The polynomial modeling process is basically conducted by separating thehysteresis curve into two different curves: the curve with positive acceleration (lowercurve) and the curve with negative acceleration (upper curve). Each curve is thenapproximated using a unique polynomial equation, which can be expressed in thefollowing term:

F =n∑i=0

aixi , n = order of polynomial (5.1)

83

where ai is the coefficient that should be empirically determined using the curve fittingof the experimental data. In both [147] and [148],the changing trend of each coefficientas a variable to the magnitude of current input is then approximated using linearregression so that

ai = bi + ciI , i = 0, 1, ..., n (5.2)

where the b and c are the tracking coefficient for the a as a function of current input, I .By substituting the equation 5.2 to equation 5.1, the damping force can be representedas:

F =n∑i=0

(bi + ciI) xi , n = order of polynomial (5.3)

MR Damper model

MR Valve model

x.

I

F

Q

I

DP

Figure 5.1 The difference between MR damper model and MR valve model excitation

Since the determination of the coefficient is conducted individually for theupper curve and the lower curve in each current input, the number of unique sets ofcoefficient for each curve will be dependent to the order of polynomial, n. However,according to the work of Choi et al. [147], the order of polynomial should be at leastsix or above in order to capture the hysteresis behavior well.

Adopting the method to the case of MR valve requires some changes in theinput and output variables, because there are some differences between the modelingrequirement of MR damper and MR valve as shown in Figure 5.1. Therefore, the

84

generalized form of the polynomial-based parametric model for the MR valve is asfollows:

∆P =n∑i=0

aiQi , n = 6 (5.4)

where the ∆P is the pressure drop of MR valve, similarly to the force F of MR damperand Q is the volumetric flow rate of the MR fluid across the MR valve, similarly to thepiston velocity x of the MR damper. Sixth order of polynomial is also chosen to modelthe MR valve with reference to [147] and [148].

In order to find the set of coefficient a, the experimental data from the frequencyexcitation of 0.75 Hz is selected as a reference. There are 11 sets of curve thatrepresenting the pressure drop response for current inputs of 0.0 to 1.0 A with intervalof 0.1 A. There are 1000 points of measurement for each curve which can be evenlydivided into two regions, the region with positive flow acceleration (lower loop) andthe region with negative flow acceleration (lower loop). Each region is then fitted withthe sixth order polynomial equation to find the specific sets of coefficient ai for eachcurve. The results of curve fittings for both lower loop and upper loop are shown infigure 5.2 and 5.3 respectively.

The trends of coefficients are presented in normalized values to enhance thevisibility of the variations. The method of presentation using normalized value alsohas been used in [127, 146]. The first impression when noticing the trend of valuesin figure 5.2 and 5.3 are that the linear regression approach, such as implemented in[147] and [148], is no longer suitable to represent most of the relationship betweenthe coefficient a and the current input I . The fluctuations of coefficient values for bothlower and upper loops are mostly too large to be approximated using linear function.The suitability of these coefficient trends to be represented by linear relationship alsocan be tested using the Pearson correlation coefficient that is expressed by

r =

∑(I − I

)(a− a)√∑(

I − I)2∑

(a− a)2(5.5)

If the correlation coefficient, r is close enough to +1 or -1, the relationship canbe considered suitable to be represented by linear approximation. However, if the ris equal to zero, then there are no correlation between the corresponding coefficient awith the current input I . The results of correlation test for each coefficient with respect

85

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−15

−10

−5

0

5

10

15

20

25

Current Input (A)

Coe

ffici

ent V

alue

a

6*2*10 9

a5*10 8

a4*−6*10 5

(a)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−5

0

5

10

15

20

25

30

35

Current Input (A)

Coe

ffici

ent V

alue

a

3*−1*10 4

a2*2*10 2

a1*10

a0*−1

(b)

Figure 5.2 Trend of the normalized coefficient values at the positive flow acceleration(lower loop) to the variations of current input (a) a6, a5 and a4 (b) a3, a2, a1 and a0

to the current input variability are shown in table 5.1 It can be seen that most of thecoefficient has non-linear correlation with the current input and therefore cannot berepresented by linear function.

Table 5.1: Correlation test results between the model coefficient a and current input I

Model coefficient Correlation coefficientLower loop Upper loop

a0 -0.9863 0.9906a1 0.8396 0.8511a2 0.5486 -0.8251a3 0.0068 -0.1176a4 -0.1181 0.5095a5 -0.2821 -0.1492a6 -0.0722 -0.3090

86

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−15

−10

−5

0

5

10

15

20

25

Current Input (A)

Coe

ffici

ent V

alue

a

6*−2*10 9

a5*10 8

a4*6*10 5

(a)

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1−5

0

5

10

15

20

25

30

35

40

45

Current Input (A)

Coe

ffici

ent V

alue

a

3*−1*10 4

a2*−2*10 2

a1*10

a0

(b)

Figure 5.3 Trend of the normalized coefficient values at the negative flow acceleration(upper loop) to the variations of current input (a) a6, a5 and a4 (b) a3, a2, a1 and a0

Since most of the coefficient cannot be represented by linear relationship,another polynomial approximation is conducted to fit the coefficient trend. On thebasis of trial and error, it has been observed that 4th order of polynomial is decentlyrepresenting the relationship between ai and I , so that the relationship can be expressedby

ai = biI4 + ciI

3 + diI2 + eiI + fi , i = 0, 1, ..., 6 (5.6)

and the pressure drop can be represented by

∆P =n∑i=0

(biI

4 + ciI3 + diI

2 + +eiI + fi)Qi , n = 6 (5.7)

87

Using the same curve fitting process, the fourth order polynomial coefficientdetermination is conducted for a6, a5, a4, a3, a2, a1 and a0 with respect to current inputvariations. The specific values of bi, ci, di, ei and fi are listed in table 5.2.

Table 5.2: List of coefficients for the polynomial-based parametric MR valve modelCurve Order of Coefficient ValuesRegion Polynomial b c d e f

6 -3.337e-08 1.156e-08 1.641e-08 8.011e-10 6.406e-105 -1.149e-06 8.699e-07 2.292e-07 -1.139e-07 8.383e-09

Lower 4 1.209e-04 4.760e-06 -1.053e-04 -7.317e-06 -3.218e-06Loop 3 5.091e-03 -3.820e-03 -9.721e-04 1.661e-04 -3.785e-05

2 -0.0197 -0.224 0.239 0.0151 4.698e-031 -3.2792 -1.2161 4.9471 0.5715 0.21410 -170.0049 335.074 -193.1929 -4.9548 -1.63896 4.634e-08 -4.068e-08 -3.928e-09 -3.063e-10 -4.239e-105 -2.286e-06 2.938e-06 -8.588e-07 5.549e-08 1.565e-09

Upper 4 -1.715e-04 1.138e-04 5.632e-05 3.617e-06 2.291e-06Loop 3 9.737e-03 -0.0123 3.502e-03 -5.047e-04 -9.416e-06

2 0.0514 0.1260 -0.2001 -8.136e-03 -3.707e-031 -7.7430 7.0455 0.6173 1.1776 0.18620 199.877 -373.947 211.433 1.279 1.9134

5.3 Modified LuGre-based Hysteresis Modeling Approach

The second approach to parametrically model the hysteretic behavior of theMR valve is based on the LuGre hysteresis model. The LuGre model is the hysteresismodel that initially developed to model the friction dynamics [176,177] and is selectedas the base model for the MR valve. The selection of LuGre model as the base model ismainly because hypothetically the flow dynamics is easier to be represented as frictiondynamics rather than as a set of springs and dampers. The primary form of LuGrehysteresis model for MR damper is as follows:

F (t) = σ0z + σ1z + σ2x (5.8)

where the σ0, σ1, and σ2 are the parameters of the model and z is the variable that canbe expressed as:

z = x− |x|g (x)

z (5.9)

88

where g (x) is the additional function that depends on various factors such as materialproperties and temperature.

Further implementations of the LuGre model have brought some modificationsto the primary form such as the modification proposed by Sakai et al. [137] thatexpressed the LuGre model as:

F (t) = σaz + σ0zv + σ1z + σ2x+ σbxv (5.10)

z = x− σ0a0 |x| z (5.11)

where σa, σ0, σ1, σ2, σb, and a0 are the parameters of the model while v is the voltageinput to the coil and the z is interpreted as the variable that expressed the average MRfluids transient deformation generated when the direction of force is changing or alsoknown as the evolutionary variable.

The other modification to the LuGre model was also proposed by Jimenez andAvarez-Icaza [138] in the following form:

F (t) = σ0bzv + σ1z + σ2ax (5.12)

z = x− σ0bαa0 |x| z (1 + a1v) (5.13)

where σ0b, σ1, σ2a, α, a0 and a1 are the parameters of the model, while v is similarlythe applied voltage to the coil as well as the z as the evolutionary variable.

It can be seen that both forms use the voltage, v, as the input variable that isrelated to the coil magnetization strength. In another model, the magnetization strengthis expressed with the magnitude of the current, i, given to the coil [127]. Physicallyboth can be considered as equal term given that the change of coil resistance duringmagnetization is neglected. However, both current and voltage are just the indirectvariables to express the units of magnetization strength. In other words, the value ofvoltage or current that is expressed in the model will not be universally applicable sincethe strength of magnetic field that influences the MR effect is also highly subjected tocoil turns and dimensions.

In similar reasons with the modifications of the polynomial-based parametricmodel as shown in Figure 5.1, some modifications to the base form of the LuGrehysteresis model are also needed to transform the form of damper model to the form

89

of valve model. The modifications of the base model are also needed to reduce theparameter and simplify the model. Therefore, the generalized form of the modifiedLuGre based MR valve model proposed in this paper is as follows:

∆P = Az +Bz + CQα (5.14)

z = Q− a0 |Q| z (5.15)

thus, there are five independent parameters that need to be identified, and these set ofparameters are defined as follows:

Θ = [A,B,C, α, a0]

As a general form, these five parameters can be dependent or independent tothe flow rate and/or the current input depending on the selected method of parameteridentification. However, in common, at least one of these parameters should bedependent to the current input since the current input applied, as one of the independentvariables, is not yet accommodated in the generalized form of the model.

In order to identify the values of model parameters, the characteristics of MRvalve with 0.5 mm annular and 0.5 mm radial gaps at frequency excitation of 0.75 Hzis taken as a sample. The identification process of the model parameter is conductedusing the gradient descent method through the Parameter Estimation Tool (PET) inMATLAB. The method is used to determine the optimum values of the five differentparameters to match the model predictions and the experimental data. Since the curveis unique for different current input, there are also unique sets of parameter values foreach curve. The collection of each parameter values for different current input are thenretraced to find its trend line. The approximated functions of the trend line will be theempirical function for to determine the parameter values with respect to the variationof the current input.

Figure 5.4 present the trends of normalized estimated parameters at variouscurrent inputs. There are similarities of parameter trends in these three differentfrequency excitation where the A, B and α tend to increase, and C tends to decrease.The a0 is intentionally made invariant to the current input and proportionally variableto the frequency excitations. From these figures it can be seen that although there aresome fluctuations in the estimated parameters, these fluctuations did not qualitativelychange the overall trend of the parameters.

90

0 0.2 0.4 0.6 0.8 10

2

4

6

8

10

12

14

Current Input (A)

A*20B/5C*2.5α*5a0*5

Figure 5.4 Trend of estimated parameters with respect to current input

The trend of each parameter is used to generate an approximation function tomimic the trend of each parameter, similarly with the approach demonstrated by Jiangand Christenson [145]. The approximated function will be specific for each parameterfor a bonded range of the current input. The functions are needed so that the model,although case specific, can be used in a more compact form of mathematical functionsand more practical to be used in control design process. The approximated functionsfor each parameter are shown in Table 5.3

Table 5.3: List of approximated function for different parametersParameters Approximated Functions

A −1.1612i3 + 1.397i2 + 0.3613i+ 0.0391B −22.33i5 + 181.31i4 − 324.92i3 + 192.26i2 + 7.37i+ 2.035C 2.268i3 + 0.6812i2 + 1.1267i+ 0.0683α 3.6414i4 − 9.1188i3 + 8.032i2 − 3.247i+ 1.2009a0 1.0

5.4 Model Performance Comparison

The simulation results of both polynomial-based hysteresis model and theLuGre based hysteresis model in comparison with the experimental data for frequency

91

excitation of 0.75 Hz at current inputs of 0.3 A, 0.6 A and 0.9 A are shown in Figure5.5a to 5.5c respectively.

Generally, in these figures, the results from both models are in a goodagreement with the experimental data, especially in terms of the peak pressure dropvalue. Though there are some compromises when the curve is reaching the pressuredrop of 0 MPa due to the air pocket effect as mentioned in section 4.4.4. These airpocket effects, unfortunately, failed to be represented in both models and apparentlyhave been tampering the parameter values during the parameter identification process.

Although in general both model are able to represent the hysteretic behavior ofthe MR valve, in order to compare the model accuracy, a specific measurement need tobe taken. In this study, the relative error, that is adopted from [127], is used to measurethe level of accuracy of each model. The relative error is expressed in the followingequation:

RE =

∑ni=1

∣∣∆P expi −∆Pmodel

i

∣∣∑ni=1 |∆P

expi |

(5.16)

where n is the number of measurement points, ∆P expi is the experimental data of

pressure drop at i-th point and ∆Pmodeli is the i-th pressure drop obtained from the

model. The comparison of relative error between the polynomial-based hysteresismodel and the LuGre based hysteresis model are presented in table 5.4

Table 5.4: Comparison of relative error at 0.75 Hz frequency excitationCurrent Input Relative Error

(A) Polynomial-based model LuGre based model0.0 0.0813 0.08050.1 0.0805 0.10480.2 0.0686 0.05660.3 0.0609 0.06620.4 0.0963 0.10270.5 0.0639 0.05220.6 0.0641 0.06120.7 0.0518 0.04180.8 0.0510 0.05560.9 0.0644 0.09441.0 0.0792 0.1186

Average 0.0693 0.0759

According to table 5.4, most of the results from both model are showing relative

92

-60 -40 -20 0 20 40 60-3

-2

-1

0

1

2

3

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

Experimental DataPolynomial-based Hysteresis ModelLuGre based Hysteresis Model

(a)

-60 -40 -20 0 20 40 60-6

-4

-2

0

2

4

6

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

Experimental DataPolynomial-based Hysteresis ModelLuGre based Hysteresis Model

(b)

-60 -40 -20 0 20 40 60-8

-6

-4

-2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

Experimental DataPolynomial-based Hysteresis ModelLuGre based Hysteresis Model

(c)

Figure 5.5 Comparison between the test data and the model results for various currentinput, (a) 0.3 A (b) 0.6 A (c) 0.9 A

93

error less than 10%, which can be considered good since the results from the modeldeveloped by [127] were reported to have even higher relative error in around 15%.In average, the relative error of the polynomial-based hysteresis model is also lowerthan the relative error of the LuGre based hysteresis model, which is demonstratingthe advantage of the polynomial model in terms of accuracy as already explained inchapter 2.

-40 -30 -20 -10 0 10 20 30 40-8

-6

-4

-2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

Experimental DataPolynomial-based Hysteresis ModelLuGre based Hysteresis Model

(a)

-80 -60 -40 -20 0 20 40 60 80-8

-6

-4

-2

0

2

4

6

8

Flow Rate (ml/s)

Pre

ssur

e D

rop

(MP

a)

Experimental DataPolynomial-based Hysteresis ModelLuGre based Hysteresis Model

(b)

Figure 5.6 Comparison between the test data and the model results for current inputof 1.0 A at various frequency excitations, (a) 0.50 Hz (b) 1.00 Hz

However, while it is obvious that the models, that have been specificallyidentified using the 0.75 Hz experimental data, can mimic the 0.75 Hz measurementdata well, the real challenge is actually to perform at the frequency excitations otherthan 0.75 Hz. As an example. the performance comparisons of each model with theexperimental data at 0.50 Hz and 1.00 Hz frequency excitations for current input

94

of 1.0 A are depicted in Figure 5.6. The results in these figures have shown thatthe deviations are apparent for both models in the 0.50 Hz and 1.00 Hz frequencyexcitations. However, the deviations of the polynomial-based hysteresis model is seenmuch larger than the deviations of the LuGre based hysteresis model, which are alsoshown by the comparison of relative error in table 5.5. These large deviations can beexplained by considering that the polynomial equations, at some values, will reach itsextreme points. In this case the extreme points of the pressure drop are visible at the1.00 Hz frequency excitations, which occurred at the flow rate slightly below +/- 55ml/s.

Table 5.5: Comparison of relative error at 0.50 Hz and 1.00 Hz frequency excitationsCurrent Relative Error

Input Polynomial-based model LuGre based model(A) 0.50 Hz 1.00 Hz 0.50 Hz 1.00 Hz0.0 0.1268 0.5871 0.0767 0.06740.1 0.1240 0.5132 0.1053 0.09500.2 0.1582 0.5029 0.0685 0.05440.3 0.1670 0.6500 0.0894 0.05070.4 0.1891 0.8125 0.1274 0.08030.5 0.1550 0.9269 0.0603 0.06820.6 0.1503 1.0091 0.0669 0.10660.7 0.1405 0.9001 0.0614 0.07570.8 0.1794 0.6162 0.1098 0.04230.9 0.2073 0.3196 0.1424 0.08211.0 0.3033 0.8669 0.1427 0.1369

Average 0.1728 0.7004 0.0955 0.0781

According to the model verification results, it can be concluded that thepolynomial-based hysteresis model is actually providing better accuracy than theLuGre-based hysteresis model but only as far as the model inputs are similar withthe inputs used in the model development. The LuGre based hysteresis model, on theother hand, although cannot compete with the polynomial-based model in terms ofaccuracy, is shown the capability to show a decent performance and smaller deviationsto the measurement data in wider range of model inputs. Additionally, the number ofparameters of the LuGre based hysteresis model is also smaller than the polynomial-based hysteresis model so that the parameter identification process is relatively easier.Although it should be noted that generally the accuracy of the parametric model isnot just determined by the model form but also by the identification method of theparameters [74]. The good agreement between the valve experimental data and themodel showed that the model can be used for developing further applications of the

95

MR valve including actuator design and damper design as well as the development ofcontrol system for suspension and vibration isolation devices that involves MR fluidsflow control.

5.5 Summary of Chapter 5

The hysteretic modeling of the MR valve has been discussed in this chapter.The hysteretic modeling is presented in two types of parametric modeling approaches,namely the polynomial-based hysteresis modeling approach and the modified LuGre-based hysteresis modeling approach. The performance of these two approaches inmodeling the hysteretic behavior of the MR valve is compared with the experimentaldata. According to the performance comparison results, it can be concluded thatthe polynomial-based hysteresis model has better accuracy than the LuGre-basedhysteresis model but only at the range of inputs that is recognized during theidentification process. In the wider range of inputs, the LuGre-based model show bettercapability to consistently mimic the behavior of MR valve than the polynomial-basedmodel. Moreover, the LuGre based model has fewer parameters to identify than thepolynomial based model so that the identification process is relatively simpler.

CHAPTER 6

CONCLUSIONS AND RECOMMENDATIONS

6.1 Conclusions

Magnetorheological (MR) fluids based devices are offering advanced featuresthrough the fluids sensitive rheological properties to magnetic field. The device haspenetrated various field of applications that requires energy dissipation componentsuch as damping and brakes. There are three known ways to utilize the uniquecharacteristics of MR fluids in a device. The first is known as the shear mode, whichoften utilized in brakes, the second is known as the flow mode, which commonly usedin dampers and the last one is the squeeze mode, which mostly appeared in vibrationmountings. An MR valve, in particular, is the main component that responsible ofdetermining the performance of MR devices that work in flow mode. In this study,a new concept of MR valve with the aim to achieve higher pressure drop withoutany major requirement to enlarge the valve size was introduced. The concept waspresented in both simulation, with the aid of Finite Element Method Magnetics(FEMM) software, and experimental work, using the dynamic test machines. Themathematical models of the new concept of MR valve also have been developed andassessed with the experimental data. The detailed conclusions of the study are listed inthe following:

6.1.1 The New Magnetorheological Valve Concept

In this study, the new concept of MR valve has been introduced using thecombination of multiple annular and radial gaps to form a meandering flow path. Themeandering flow path approach is used to increase the path length of MR fluid flowinside the valve so that the opportunity of the MR fluid to be exposed with magneticfield is higher. In order to predict the performance of the new MR valve concept, the

97

steady-state model of the MR valve with multiple annular and radial valve is derived.According to the simulation results of the steady-state model, the new MR valve withmeandering flow path concept is able to reach pressure drop of around 6 MPa in acomparable valve size and power consumption of a conventional 2.5 MPa annular MRvalve. The performance of the MR valve was also analyzed by assessing the effect ofcurrent input variations, flow rate variations and gap size variations to the generatedpressure drop of the MR valve.

6.1.2 Gap Size Selection Effect

Among several variables that influencing the performance of MR valve, gapsize can be considered as the most significant variable to the overall performance of theMR valve. The gap size has been known to have inverse-correlation with the viscouspressure drop at the power of three. At the same time, the gap size is also capable toinfluence the magnetic field strength at the effective area which eventually will relate tothe field-dependent pressure drop of the MR valve. According to the simulation results,the reduction of gap size at the annular channels of the proposed MR valve concept hasa very significant effect to the increase of viscous pressure drop value of the MR valve.Obviously, it also has similar effect to the improvement of the field-dependent pressuredrop but the impact was not as high as the viscous pressure drop. On the other hand,the gap size reduction at the radial channels has greater impact to the improvementof field-dependent pressure drop than the annular channel, but has lower effect to theviscous pressure drop value. Further, the actual selection of the gap size will highlydetermined by the designated applications of the MR valve.

6.1.3 Experimental Assessment of Magnetorheological Valve Performance

The predicted results of the steady-state model are well confirmed by theexperimental work with small deviations in the off-state pressure drop value. Theexperimental work is conducted using a double rod hydraulic cylinder, as the testingcell, actuated by the dynamic test machine. The experimental test was conducted inthree different frequency excitations with constant amplitude of 25 mm to emulate thevariations of peak fluid flow rate. In each frequency excitations the current input isvaried from 0.0 A to 1.0 A with 0.1 A increment. Using the reciprocating sinusoidalmovement from the dynamic test machine, the fluid flow rate across the MR valve

98

is also varied in sinusoidal value, which allow the observation of hysteretic pressuredrop behavior. According to the experimental results, the hysteretic pressure behavioroccurs when there are different measured pressure drop at the same flow rate withnon-zero flow acceleration at different flow direction.

6.1.4 Hysteretic Modeling of Magnetorheological Valve

The hysteretic behavior cannot be shown in the steady-state model. Therefore,another model of MR valve, which can represent the hysteretic behavior, is needed. Inorder to develop a model that can mimic the hysteretic behavior, the measured pressuredrop from the experimental test, which show the hysteretic behavior of the MR valve,is used. Two hysteretic model are developed and compared with the experimentaldata. The first model is developed using the polynomial-based parametric approachand the second model is derived based on LuGre hysteresis parametric operator.The comparison results have shown that the polynomial-based parametric model isproducing more accurate results than the LuGre-based parametric model for the sameinputs that are used in the parameter identification process. Meanwhile, at the otherrange of inputs, the results from the LuGre-based parametric model is more consistentand robust than the results from the polynomial-based parametric model.

Since the verified hystetic model is a very good representation of the MR valve,the model can be used to design the control system for further applications of MRvalve. In this study, the process of designing a pressure tracking control system usingPI controller is presented as a demonstration on how the MR valve model can be usedto serve various applications. The assessment of performance of the pressure trackingcontrol is evaluated by introducing three different types of reference input, whichrepresenting the class of continuous and discontionuous functions, to the controller.The simulation results have shown that the controller is able to track these three signalswell. Further, the pressure tracking control system can be used as the embedded innerloop controller of the MR valve that act to ensure that the generated pressure drop ofan MR valve can be accurately delivered as commanded by the outer loop controller toserve the application demands.

99

6.2 Contributions of the Research

In this research, the new way to improve the achievable pressure drop of MRvalve has been demonstrated with the meandering flow path concept. The new concepthas been studied thoroughly by evaluating the MR valve behavior in steady-state anddynamic environment. The results that were obtained from the study showed that thenew concept can be potentially used in various applications that requires flow controlas well as enhancing the performance of MR devices that utilize valve as one of itsmain component. Nevertheless, the contributions of this research are listed as follows:

(a) This research contributed to the introduction of the new concept of MR valve.The novelty of the new MR valve concept is mainly justified by the uniquearrangement of the effective area, known as the meandering flow path structure,which makes it capable to improve the achievable pressure drop of the MRvalve. The meandering flow path structure was formed by the use of multiplecombination of annular and radial flow channel in a single MR valve structure.The performance improvement that can be achieved with the meanderingflow path structure has been proven in this research through simulation andexperimental assessment.

(b) The second contribution of this research was related to the knowledge on theeffect of gap size selection to the achievable pressure drop of the valve. Thespecific selections of the gap size have a very significant effect to the pressuredrop characteristics of the valve where the annular gaps have shown to beresponsible in the determination of off-state pressure drop while the radial gapswere highly influencing the achievable on-state performance. In terms of valvedynamic range, it has been confirmed in both simulation and experimental workthat the reduction of annular gap size is highly affecting the degradation of valvedynamic range while in contrary occured with the reduction of radial gap size,although the effect is not significant. Further, the knowledge on how to selectthe size of both gaps are very important in the sizing process for particularapplications.

(c) The third contribution was regarding the development of new parametric modelbased hysteretic MR valve model. The hysteresis was identified during theexperimental assessment of MR valve. The model is the first parametric modelderived specifically to model the hysteretic behavior of MR valve.

100

6.3 Open Problems and Recommendations for Future Works

6.3.1 Pressure Tracking Control System

There are wide ranges of applications that can utilize the advance features ofMR valve, since most of actuators that rely on flow control mechanism will requirevalve as their main control element. However, in order for the MR valve to bewell functioned as the final control element of an actuator, the pressure drop thatis generated by the valve has to be controllable accurately. Therefore, the pressuretracking control system is needed as the valve inner-loop controller and integratedmodule. As a case study for future reference, the explanation of the basic structureof the pressure tracking controller and its simulated performance using the MR valvehysteresis model in the case of tracking various types of input references are presented.

MRValve

CurrentDriver

PIController

Desiredpressuredrop

Actualpressure

drop

Controlsignal

Errorsignal

Currentinput

Flowrate

-

+

Figure 6.1 Basic structure of pressure tracking control of MR valve

The basic structure of the pressure tracking control of the MR valve are shownin figure 6.1, which depicts a closed-loop control system with PI controller. The PIcontroller is chosen because it is simple, easy and widely recognized as the mostcommon controller in industrial application [178]. The input of the controller is theerror signal, which is acquired from the subtraction between the reference signal andthe feedback signal. The reference signal in the case of MR valve pressure trackingcontrol is the desired pressure drop and the feedback signal is the actual pressure drop.Therefore, the formulation of the PI controller for the pressure tracking control can beexpressed as follows

u (t) = Kpe (t) +Ki

∫e (t) (6.1)

e (t) = ∆Pd (t)−∆Pa (t) (6.2)

101

where the u(t) is the control signal which drives the current driver to charge theelectromagnetic coil of the MR valve with current, the Kp and Ki are the proportionaland integral gain respectively as well as the ∆Pd and ∆Pa as the desired pressure dropand actual pressure drop.

The current driver is needed to adjust the value of the output signal from thecontroller with the signal input that can drive the MR valve. Since the control signalwill not always satisfy the signal requirement of the MR valve. For example, the valueof ∆Pa can be higher or lower than the value of the ∆Pd. Therefore, the value of e(t)can be positive or negative which concurrently cause the control signal u(t) to be inpositive or negative value. However, the positive of negative sign of the current inputto the electromagnetic coil of the MR valve in this study practically has no effect to themagnitude of the pressure drop. Therefore, it is the task of the current driver to adjustthe sign, which simply can be done using absolute function.

In order to assess the performance of the pressure tracking control, the LuGre-based hysteresis MR valve model is used with several types of reference signal, namelythe sinusoidal signal, the pulse signal and the saw-tooth signal. The selection of thesesignals as the reference input are based on the assumption that if the system cantrack these signal well, then the system is considered robust to track various typesof continuous and discontinuous functions [148]. The simulation results of pressuretracking control under various reference functions are shown in figure 6.2. The resultsare the outcome of pressure tracking control with controller parameters of Kp = 0.08

and Ki = 4, which acquired from MATLAB auto-tuning tools, and Kp = 0.05 andKi = 2, which acquired from trial and error method. From figure 6.2, it can beconcluded that the pressure tracking control of the hysteresis model of MR valve isable to track various types of continuous and discontinuous functions well.

Although the explanation of the pressure tracking control in this section onlyinvolved a closed-loop PI controller, the further application of inner-loop controllerof the MR valve is not just limited to the PI controller. Various control strategy canbe developed and integrated with the MR valve. The demonstrated pressure trackingcontrol system using PI controller shown in this section is merely an illustration ofthe kind of applications that the MR valve can serve. With the involvement of theinner-loop controller, the MR valve can serve both semi-active and active applicationaccurately. Considering the necessity of the study, the further exploration of thepressure tracking control system of the meandering type MR valve is recommendedfor future works.

102

2 3 4 5 6 7 8 9 102.5

3

3.5

4

4.5

5

5.5

Time (s)

Pre

ssur

e D

rop

(MP

a)

Desired Pressure DropControlled Pressure Drop (PI Autotuning)Controlled Pressure Drop (Trial and Error)

(a)

2 3 4 5 6 7 8 9 102.5

3

3.5

4

4.5

5

5.5

Time (s)

Pre

ssur

e D

rop

(MP

a)

Desired Pressure DropControlled Pressure Drop (PI Autotuning)Controlled Pressure Drop (Trial and Error)

(b)

2 3 4 5 6 7 8 9 102.5

3

3.5

4

4.5

5

5.5

Time (s)

Pre

ssur

e D

rop

(MP

a)

Desired Pressure DropControlled Pressure Drop (PI Autotuning)Controlled Pressure Drop (Trial and Error)

(c)

Figure 6.2 Simulation results of pressure tracking control under various functions asreference, (a) Sinusioidal (b) Pulse (c) Saw-tooth

103

6.3.2 Other Open Problems

There are also several problems encountered that were not able to be discussedin this study and can be recommended for future works as a continuation of this study.These problems are as the following:

(a) The conceptual design of the MR valve with meandering flow path was notconsidering any optimization on valve design. The gap sizes of annular and radialflow channel, the coil properties, the MR fluid types and the overall valve size inthis study were selected merely to serve the observation purposes and thereforenot considering any optimum value. In the future, it will be necessary to put theconsideration on the optimization on these variables such as to conduct the studyon the critical thickness of the valve casing with respect to magnetic reluctanceand mechanical strength to withstand the internal pressure of the valve.

(b) The fundamental knowledge of the flow behavior in radial valve, in particular,was also found interesting by the author. The interesting part existed in the stressbehavior of the MR fluid when radially flowing in concave and convex directionsince the shear rate of the fluid is not constant even in a constant volumetric flowrate. Due to these characteristics, it is make sense to suspect that the commonknowledge of MR fluid flow behavior in a parallel channel is no longer validto explain the flow behavior in a radial channel. Therefore, a thorough study isnecessary to reveal the flow behavior in a radial channel in the future, which inparticular, can contribute to improve the model of MR valve with meanderingflow path.

(c) In terms of experimental assessment, in this study the MR valve has experiencedtesting where the flow rate is constantly changing due to the reciprocatingmovement of the testing cell. With these kind of assessment, the transientbehavior of the MR valve due to the variations of fluid flow rate are observable.However, in these testing conditions, the current input remains constant andtherefore the transient response of the MR valve to the change of current inputhas not been investigated yet. The knowledge of transient response of the MRvalve with respect to the current input is necessary to develop a dynamic modelof the MR valve with accurate time constant to serve the control design purposes.

(d) The meandering flow path structure that consists of the combination of annularand radial flow channel inevitably formed many sharp corner similarly with theeffect of elbow. The fluid flow in elbow, in general, has been widely known as acomplex phenomenon due to frictional and separation effects [179,180]. Though

104

for the flow case of non-Newtonian fluid the friction loss coefficient has beenshown lower than in the case of Newtonian fluid [181], the losses in a small flowchannel such as in an MR valve may become significantly large [124]. Therefore,it may be necessary in the future to consider the elbow effect in the meanderingflow path structure in the pressure drop characteristics of MR valve to furtherexamine the concept.

(e) The results of polynomial-based hysteresis model have revealed that thepolynomial model is actually cannot act as a stand-alone model. The limitation ofthe model to track the characterization results can be solved by adding anothersystem that capable to continuously update the polynomial model parameterswhen flow rate input is changing. The addition of such system will provide avery promising hysteresis model since the polynomial model has been shown asa very accurate model in a limited range of flow rate. The further study of on-lineparameter tuning for polynomial-based hysteresis model is an interesting topicand will be beneficial in the field of modeling of hysteretic system.

REFERENCES

[1] Bossis, G., Lacis, S., Meunier, A. and Volkova, O. Magnetorheologicalfluids. Journal of Magnetism and Magnetic Materials, 2002. 252: 224–228.

[2] Bica, I., Liu, Y. D. and Choi, H. J. Physical characteristics ofmagnetorheological suspensions and their applications. Journal of

Industrial and Engineering Chemistry, 2013. 19(2): 394–406.[3] Jolly, M. R., Bender, J. W. and Carlson, J. D. Properties and applications

of commercial magnetorheological fluids. Journal of Intelligent Material

Systems and Structures, 1999. 10(1): 5–13.[4] Tang, X., Zhang, X., Tao, R. and Rong, Y. Structure-enhanced yield stress of

magnetorheological fluids. Journal of Applied Physics, 2000. 87(5): 2634.[5] Tao, R. Super-strong magnetorheological fluids. Journal of Physics:

Condensed Matter, 2001. 13(50): R979–R999.[6] Choi, Y. T., Cho, J. U., Choi, S. B. and Wereley, N. M. Constitutive models

of electrorheological and magnetorheological fluids using viscometers.Smart Materials and Structures, 2005. 14(5): 1025–1036.

[7] Chen, S., Huang, J., Shu, H., Sun, T. and Jian, K. Analysis and testingof chain characteristics and rheological properties for magnetorheologicalfluid. Advances in Materials Science and Engineering, 2013. 2013(Ci): 1–6.

[8] Bolter, R. and Janocha, H. Design rules for MR fluid actuators in differentworking modes. Proceedings of SPIE Vol. 3045. SPIE. 1997, vol. 3045.148–159.

[9] Carlson, J. and Jolly, M. R. MR fluid, foam and elastomer devices.Mechatronics, 2000. 10(4-5): 555–569.

[10] Zipser, L., Richter, L. and Lange, U. Magnetorheologic fluids for actuators.Sensors and Actuators A: Physical, 2001. 92(1-3): 318–325.

[11] Rabinow, J. The magnetic fluid clutch. Transactions of the American

Institute of Electrical Engineers, 1948. 67(2): 1308–1315.[12] Carlson, J. Critical factors for MR fluids in vehicle systems. International

Journal of Vehicle Design, 2003. 33(1/2/3): 207.[13] Milecki, A. and Hauke, M. Application of magnetorheological fluid in

industrial shock absorbers. Mechanical Systems and Signal Processing,

106

2012. 28: 528–541.[14] Zhu, X., Jing, X. and Cheng, L. Magnetorheological fluid dampers: A

review on structure design and analysis. Journal of Intelligent Material

Systems and Structures, 2012. 23(8): 839–873.[15] Wang, J. and Meng, G. Magnetorheological fluid devices: principles,

characteristics and applications in mechanical engineering. Proceedings

of the Institution of Mechanical Engineers, Part L: Journal of Materials:

Design and Applications, 2001. 215(3): 165–174.[16] Kordonski, W., Gorodkin, S., Kolomentsev, A., Kuzmin, V., Luk’ianovich,

A., Protasevich, N., Prokhorov, I. and Shulman, Z. Magnetorheologicalvalve and devices incorporating magnetorheological elements, 1994.

[17] Gorodkin, S., Lukianovich, A. and Kordonski, W. Magnetorheologicalthrottle valve in passive damping systems. Journal of Intelligent Material

Systems and Structures, 1998. 9(8): 637–641.[18] Yokota, S., Yoshida, K. and Kondoh, Y. A pressure control valve using MR

fluid. Proceedings of the JFPS International Symposium on Fluid Power,1999. 4: 377–380.

[19] Yoshida, K., Takahashi, H., Yokota, S., Kawachi, M. and Edamura, K. Abellows-driven motion control system using a magneto-rheological fluid.Proceedings of the JFPS International Symposium on Fluid Power, 2002.(5-2): 403–408.

[20] Yoo, J.-H. and Wereley, N. M. Design of a high-efficiencymagnetorheological valve. Journal of Intelligent Material Systems and

Structures, 2002. 13(10): 679–685.[21] Wang, X., Gordaninejad, F., Hitchcock, G. H., Bangrakulur, K., Fuchs, A.,

Elkins, J., Evrensel, C. A., Dogruer, U., Ruan, S., Siino, M. and Kerns,M. Q. A new modular magneto-rheological fluid valve for large-scaleseismic applications. Wang, K.-W., ed. Proc. SPIE 5386. Bellingham, WA:SPIE. 2004, vol. 5386. 226–237.

[22] Grunwald, A. and Olabi, A. G. Design of magneto-rheological (MR) valve.Sensors and Actuators A: Physical, 2008. 148(1): 211–223.

[23] Ai, H. X., Wang, D. H. and Liao, W. H. Design and modeling of amagnetorheological valve with both annular and radial flow paths. Journal

of Intelligent Material Systems and Structures, 2006. 17(4): 327–334.[24] Wang, D. H., Ai, H. X. and Liao, W. H. A magnetorheological valve with

both annular and radial fluid flow resistance gaps. Smart Materials and

Structures, 2009. 18(11): 115001.[25] Yoo, J.-H., Sirohi, J. and Wereley, N. M. A nagnetorheological

piezohydraulic actuator. Journal of Intelligent Material Systems and

107

Structures, 2005. 16(11-12): 945–953.[26] John, S., Chaudhuri, A. and Wereley, N. M. A magnetorheological actuation

system: test and model. Smart Materials and Structures, 2008. 17(2):025023.

[27] Salloom, M. and Samad, Z. Design and modeling magnetorheologicaldirectional control valve. Journal of Intelligent Material Systems and

Structures, 2011. 23(2): 155–167.[28] Lee, H.-S. and Choi, S.-B. Control and response characteristics of a

magneto-rheological fluid damper for passenger vehicles. Journal of

Intelligent Material Systems and Structures, 2000. 11(1): 80–87.[29] Yao, G. Z., Yap, F. F., Chen, G., Li, W. H. and Yeo, S. H. MR damper

and its application for semi-active control of vehicle suspension system.Mechatronics, 2002. 12(7): 963–973.

[30] Lee, H. G., Sung, K. G. and Choi, S. B. Ride comfort characteristics withdifferent tire pressure of passenger vehicle featuring MR damper. Journal

of Physics: Conference Series, 2009. 149: 012069.[31] Choi, S.-B., Seong, M.-S. and Ha, S.-H. Vibration control of an MR vehicle

suspension system considering both hysteretic behavior and parametervariation. Smart Materials and Structures, 2009. 18(12): 125010.

[32] Giorgetti, A., Baldanzini, N., Biasiotto, M. and Citti, P. Design and testingof a MRF rotational damper for vehicle applications. Smart Materials and

Structures, 2010. 19(6): 065006.[33] Ha, S. H., Seong, M.-S. and Choi, S.-B. Design and vibration control of

military vehicle suspension system using magnetorheological damper anddisc spring. Smart Materials and Structures, 2013. 22(6): 065006.

[34] Bai, X.-X., Hu, W. and Wereley, N. M. Magnetorheological damper utilizingan inner bypass for ground vehicle suspensions. IEEE Transactions on

Magnetics, 2013. 49(7): 3422–3425.[35] Fischer, D. and Isermann, R. Mechatronic semi-active and active vehicle

suspensions. Control Engineering Practice, 2004. 12(11): 1353–1367.[36] Poussot-Vassal, C., Sename, O., Dugard, L., Gaspar, P., Szabo, Z. and Bokor,

J. A new semi-active suspension control strategy through LPV technique.Control Engineering Practice, 2008. 16(12): 1519–1534.

[37] Ubaidillah, Hudha, K. and Jamaluddin, H. Simulation and experimentalevaluation on a skyhook policy-based fuzzy logic control for semi-activesuspension system. International Journal of Structural Engineering, 2011.2(3): 243.

[38] Sam, Y. and Hudha, K. Modelling and force tracking control of hydraulicactuator for an active suspension system. 2006 1ST IEEE Conference on

108

Industrial Electronics and Applications. IEEE. 2006. ISBN 0-7803-9513-1.1–6.

[39] Imaduddin, F., Hudha, K., Mohammad, J. I. and Jamaluddin, H. Simulationand experimental investigation on adaptive multi-order proportional-integral control for pneumatically actuated active suspension system usingknowledge-based fuzzy. International Journal of Modelling, Identification

and Control, 2011. 14(1/2): 73–92.[40] Gysen, B., Paulides, J., Janssen, J. and Lomonova, E. Active

electromagnetic suspension system for improved vehicle dynamics. IEEE

Transactions on Vehicular Technology, 2010. 59(3): 1156–1163.[41] Gysen, B. L. J., van der Sande, T. P. J., Paulides, J. J. H. and Lomonova, E. A.

Efficiency of a regenerative direct-drive electromagnetic active suspension.IEEE Transactions on Vehicular Technology, 2011. 60(4): 1384–1393.

[42] Nguyen, Q. H., Choi, S. B., Lee, Y. S. and Han, M. S. Optimal designof high damping force engine mount featuring MR valve structure with bothannular and radial flow paths. Smart Materials and Structures, 2013. 22(11):115024.

[43] Phu, D. X., Shah, K. and Choi, S.-B. A new magnetorheological mountfeatured by changeable damping gaps using a moved-plate valve structure.Smart Materials and Structures, 2014. 23(12): 125022.

[44] Phu, D. X., Choi, S. B., Lee, Y. S. and Han, M. S. Design of anew engine mount for vertical and horizontal vibration control usingmagnetorheological fluid. Smart Materials and Structures, 2014. 23(11):117001.

[45] Phu, D. X. and Choi, S.-B. Vibration control of a ship engine systemusing high-load magnetorheological mounts associated with a new indirectfuzzy sliding mode controller. Smart Materials and Structures, 2015. 24(2):025009.

[46] Choi, S.-B., Nam, M.-H. and Lee, B.-K. Vibration control of a MR seatdamper for commercial vehicles. Journal of Intelligent Material Systems

and Structures, 2000. 11(12): 936–944.[47] Mcmanus, S., St. Clair, K., Boileau, P., Boutin, J. and Rakheja, S. Evaluation

of vibration and shock attenuation performance of a suspension seat witha semi-active magnetorheological fluid damper. Journal of Sound and

Vibration, 2002. 253(1): 313–327.[48] Choi, S.-B. and Han, Y.-M. MR seat suspension for vibration control of a

commercial vehicle. International Journal of Vehicle Design, 2003. 31(2):202.

[49] Hiemenz, G. J., Hu, W. and Wereley, N. M. Semi-active magnetorheological

109

helicopter crew seat suspension for vibration isolation. Journal of Aircraft,2008. 45(3): 945–953.

[50] Karakoc, K., Park, E. J. and Suleman, A. Design considerations for anautomotive magnetorheological brake. Mechatronics, 2008. 18(8): 434–447.

[51] Park, E. J., da Luz, L. F. a. and Suleman, A. Multidisciplinary designoptimization of an automotive magnetorheological brake design. Computers

& Structures, 2008. 86(3-5): 207–216.[52] Nguyen, Q. H., Jeon, J. C. and Choi, S. B. Optimal design of a

hybrid magnetorheological brake for middle-sized motorcycles. Applied

Mechanics and Materials, 2011. 52-54: 371–377.[53] Nguyen, Q. H. and Choi, S. B. Optimal design of a T-shaped drum-type

brake for motorcycle utilizing magnetorheological fluid. Mechanics Based

Design of Structures and Machines, 2012. 40(2): 153–162.[54] Lee, U., Kim, D., Hur, N. and Jeon, D. Design analysis and experimental

evaluation of an MR fluid clutch. Journal of Intelligent Material Systems

and Structures, 1999. 10(9): 701–707.[55] Neelakantan, V. A. and Washington, G. N. Modeling and reduction

of centrifuging in magnetorheological (MR) transmission clutches forautomotive applications. Journal of Intelligent Material Systems and

Structures, 2005. 16(9): 703–711.[56] Kavlicoglu, B., Gordaninejad, F., Evrensel, C., Fuchs, A. and Korol, G. A

semi-active, high-torque, magnetorheological fluid limited slip differentialclutch. Journal of Vibration and Acoustics, 2006. 128(5): 604.

[57] Wang, D., Tian, Z., Meng, Q. and Hou, Y. Development of a novel two-layermultiplate magnetorheological clutch for high-power applications. Smart

Materials and Structures, 2013. 22(8): 085018.[58] Dogruoz, M. B., Wang, E. L., Gordaninejad, F. and Stipanovic, A. J.

Augmenting heat transfer from fail-safe magneto-rheological fluid dampersusing fins. Journal of Intelligent Material Systems and Structures, 2003.14(2): 79–86.

[59] Kim, H.-C., Oh, J.-S. and Choi, S.-B. The field-dependent shock profilesof a magnetorhelogical damper due to high impact: an experimentalinvestigation. Smart Materials and Structures, 2015. 24(2): 025008.

[60] Sodeyama, H., Suzuki, K. and Sunakoda, K. Development of large capacitysemi-active seismic damper using magneto-rheological fluid. Journal of

Pressure Vessel Technology, 2004. 126(1): 105.[61] Gordaninejad, F., Wang, X., Hitchcock, G., Bangrakulur, K., Ruan, S. and

Siino, M. Modular high-force seismic magneto-rheological fluid damper.Journal of Structural Engineering, 2010. 136(2): 135–143.

110

[62] Liao, C. R., Zhao, D. X., Xie, L. and Liu, Q. A design methodology for amagnetorheological fluid damper based on a multi-stage radial flow mode.Smart Materials and Structures, 2012. 21(8): 085005.

[63] Dyke, S., Spencer, B., Sain, M. and Carlson, J. Modeling and controlof magnetorheological dampers for seismic response reduction. Smart

Materials and Structures, 1996. 5(5): 565–575.[64] Milecki, A. Investigation and control of magnetorheological fluid dampers.

International Journal of Machine Tools and Manufacture, 2001. 41(3): 379–391.

[65] Li, W. H., Du, H. and Guo, N. Q. Design and testing of an MRsteering damper for motorcycles. The International Journal of Advanced

Manufacturing Technology, 2003. 22(3-4): 288–294.[66] Milecki, Andrzej and Sedziak, Dariusz and Ortmann, Jaroslaw and Hauke,

M. Controllability of MR shock absorber for vehicles. International Journal

of Vehicle Design, 2005. 38(2/3): 222–233.[67] Yu, M., Liao, C. R., Chen, W. M. and Huang, S. L. Study on MR semi-

active suspension system and its road testing. Journal of Intelligent Material

Systems and Structures, 2006. 17(8-9): 801–806.[68] Batterbee, D. C., Sims, N. D., Stanway, R. and Rennison, M.

Magnetorheological landing gear: 2. Validation using experimental data.Smart Materials and Structures, 2007. 16(6): 2441–2452.

[69] Huang, J., He, J. and Zhang, J. Viscoplastic flow of the MR fluid in acylindrical valve. Key Engineering Materials, 2004. 274-276: 969–974.

[70] Sahin, H., Liu, Y., Wang, X., Gordaninejad, F., Evrensel, C. and Fuchs,A. Full-scale magnetorheological fluid dampers for heavy vehicle rollover.Journal of Intelligent Material Systems and Structures, 2007. 18(12): 1161–1167.

[71] Guo, P., Guan, X. and Ou, J. Physical modeling and design method of thehysteretic behavior of magnetorheological dampers. Journal of Intelligent

Material Systems and Structures, 2014. 25(6): 680–696.[72] Nguyen, Q., Choi, S. and Wereley, N. Optimal design of magnetorheological

valves via a finite element method considering control energy and a timeconstant. Smart Materials and Structures, 2008. 17(2): 025024.

[73] Bai, X.-X., Wang, D.-H. and Fu, H. Principle, modeling, and testing of anannular-radial-duct magnetorheological damper. Sensors and Actuators A:

Physical, 2013. 201: 302–309.[74] Wang, D. H. and Liao, W. H. Magnetorheological fluid dampers: a review of

parametric modelling. Smart Materials and Structures, 2011. 20(2): 023001.[75] Yang, G., Spencer, B. F., Carlson, J. D. and Sain, M. K. Large-scale MR fluid

111

dampers: modeling and dynamic performance considerations. Engineering

Structures, 2002. 24(3): 309–323.[76] Yang, G., Spencer, B. F., Jung, H.-J. and Carlson, J. D. Dynamic modeling

of large-scale magnetorheological damper systems for civil engineeringapplications. Journal of Engineering Mechanics, 2004. 130(9): 1107–1114.

[77] Wang, J., Ni, Y., Ko, J. and Spencer, B. Magneto-rheological tuned liquidcolumn dampers (MR-TLCDs) for vibration mitigation of tall buildings:modelling and analysis of open-loop control. Computers & Structures, 2005.83(25-26): 2023–2034.

[78] Bitaraf, M., Ozbulut, O. E., Hurlebaus, S. and Barroso, L. Application ofsemi-active control strategies for seismic protection of buildings with MRdampers. Engineering Structures, 2010. 32(10): 3040–3047.

[79] Tu, J. W., Liu, J., Qu, W. L., Zhou, Q., Cheng, H. B. and Cheng, X. D. Designand fabrication of 500-kN large-scale MR damper. Journal of Intelligent

Material Systems and Structures, 2011. 22(5): 475–487.[80] Weber, F. BoucWen model-based real-time force tracking scheme for MR

dampers. Smart Materials and Structures, 2013. 22(4): 045012.[81] Carlson, J. D., Matthis, W. and Toscano, J. R. Smart prosthetics based on

magnetorheological fluids. Proceedings of SPIE. SPIE. 2001, vol. 4332.308–316.

[82] Li, W. and Du, H. Development of an ankle physiotherapy device usingan MR damper. The International Journal of Advanced Manufacturing

Technology, 2004. 25(3-4): 205–213.[83] Gudmundsson, K. H., Jonsdottir, F. and Thorsteinsson, F. A geometrical

optimization of a magneto-rheological rotary brake in a prosthetic knee.Smart Materials and Structures, 2010. 19(3): 035023.

[84] Chen, J. Z. and Liao, W. H. Design, testing and control of amagnetorheological actuator for assistive knee braces. Smart Materials and

Structures, 2010. 19(3): 035029.[85] Wang, D. H. and Liao, W. H. Semi-active suspension systems for railway

vehicles using magnetorheological dampers. Part I: system integration andmodelling. Vehicle System Dynamics, 2009. 47(11): 1305–1325.

[86] Choi, Y.-T. and Wereley, N. M. Vibration control of a landing gear systemfeaturing electrorheological/magnetorheological fluids. Journal of Aircraft,2003. 40(3): 432–439.

[87] Hu, W., Wereley, N. M., Chemouni, L. and Chen, P. Semi-active linearstroke magnetorheological fluid-elastic helicopter lag damper. Journal of

Guidance, Control, and Dynamics, 2007. 30(2): 565–575.[88] Dong, X. M. and Xiong, G. W. Vibration attenuation of magnetorheological

112

landing gear system with human simulated intelligent control. Mathematical

Problems in Engineering, 2013. 2013: 1–13.[89] Li, Z. C. and Wang, J. A gun recoil system employing a magnetorheological

fluid damper. Smart Materials and Structures, 2012. 21(10): 105003.[90] Kozlowska, J. and Leonowicz, M. Processing and properties of

magnetorheological fluids for prospective application in a passive armour.IEEE Transactions on Magnetics, 2013. 49(8): 4721–4724.

[91] Singh, H. J. and Wereley, N. M. Optimal control of gun recoil in direct fireusing magnetorheological absorbers. Smart Materials and Structures, 2014.23(5): 055009.

[92] de Vicente, J., Klingenberg, D. J. and Hidalgo-Alvarez, R. Magnetorheo-logical fluids: a review. Soft Matter, 2011. 7(8): 3701.

[93] Genc, S. and Phule, P. P. Rheological properties of magnetorheologicalfluids. Smart Materials and Structures, 2002. 11(1): 140–146.

[94] Ginder, J. M. and Davis, L. C. Shear stresses in magnetorheological fluids:Role of magnetic saturation. Applied Physics Letters, 1994. 65(26): 3410.

[95] Ierardi, R. F. and Bombard, A. J. F. Off-state viscosity and yieldstress optimization of magneto-rheological fluids: A mixture design ofexperiments approach. Journal of Physics: Conference Series, 2009. 149:012037.

[96] Chiriac, H. and Stoian, G. Influence of particle size distributions onmagnetorheological fluid performances. Journal of Physics: Conference

Series, 2010. 200(7): 072095.[97] Rodrıguez-Lopez, J., Shum, H. C., Elvira, L., Montero de Espinosa, F. and

Weitz, D. A. Fabrication and manipulation of polymeric magnetic particleswith magnetorheological fluid. Journal of Magnetism and Magnetic

Materials, 2013. 326: 220–224.[98] Skjeltorp, A. One and two-dimensional crystallization of magnetic holes.

Physical Review Letters, 1983. 51(25): 2306–2309.[99] Samouhos, S. and McKinley, G. Carbon nanotubemagnetite composites,

with applications to developing unique magnetorheological fluids. Journal

of Fluids Engineering, 2007. 129(4): 429.[100] Rosenfeld, N., Wereley, N. M., Radakrishnan, R. and Sudarshan,

T. S. Behavior of magnetorheological fluids utilizing nanopowder iron.International Journal of Modern Physics B, 2002. 16(17-18): 2392–2398.

[101] Bell, R. C., Miller, E. D., Karli, J. O., Vavreck, A. N. and Zimmerman, D. T.Influence of particle shape in the properties of magnetorheological fluids.International Journal of Modern Physics B, 2007. 21(28 & 29): 5018–5025.

[102] Li, F. and Tao, C. Research on magneto-rheological technology and its

113

application. 2011 Chinese Control and Decision Conference (CCDC), 2011:4072–4076.

[103] Xu, Y., Gong, X., Xuan, S., Zhang, W. and Fan, Y. A high-performancemagnetorheological material: preparation, characterization and magnetic-mechanic coupling properties. Soft Matter, 2011. 7(11): 5246.

[104] Upadhyay, R. V., Laherisheth, Z. and Shah, K. Rheological properties ofsoft magnetic flake shaped iron particle based magnetorheological fluid indynamic mode. Smart Materials and Structures, 2014. 23(1): 015002.

[105] Shah, K., Xuan Phu, D. and Choi, S.-B. Rheological properties of bi-dispersed magnetorheological fluids based on plate-like iron particles withapplication to a small-sized damper. Journal of Applied Physics, 2014.115(20): 203907.

[106] Olabi, A. G. and Grunwald, A. Design and application of magneto-rheological fluid. Materials & Design, 2007. 28(10): 2658–2664.

[107] Mazlan, S. A., Ismail, I., Zamzuri, H. and Fatah, A. Y. A. Compressive andtensile stress of magnetorheological fluids in squeeze mode. International

Journal of Applied Electromagnetics and Mechanics, 2011. 36: 327–337.[108] Goncalves, F. D. and Carlson, J. D. An alternate operation mode for MR

fluidsmagnetic gradient pinch. Journal of Physics: Conference Series, 2009.149: 012050.

[109] Hong, S., Wereley, N., Choi, Y. and Choi, S. Analytical andexperimental validation of a nondimensional Bingham model for mixed-mode magnetorheological dampers. Journal of Sound and Vibration, 2008.312(3): 399–417.

[110] El Wahed, A. K. and Mcewan, C. A. Design and performance evaluationof magnetorheological fluids under single and mixed modes. Journal of

Intelligent Material Systems and Structures, 2011. 22(7): 631–643.[111] Yazid, I., Mazlan, S. A., Zamzuri, H., Mughni, M. and Chuprat, S.

Parameters consideration in designing a magnetorheological damper. Key

Engineering Materials, 2013. 543: 487–490.[112] Yang, C., Fu, J., Yu, M., Zheng, X. and Ju, B. A new magnetorheological

elastomer isolator in shearcompression mixed mode. Journal of Intelligent

Material Systems and Structures, 2015. 26(10): 1290–1300.[113] Yazid, I. I. M., Mazlan, S. A., Kikuchi, T., Zamzuri, H. and Imaduddin,

F. Design of magnetorheological damper with a combination of shear andsqueeze modes. Materials & Design, 2014. 54: 87–95.

[114] Mughni, M. J., Mazlan, S. A., Zamzuri, H., Yazid, I. I. M. and Rahman,M. A. A. Simulation study of magnetorheological testing cell design byincorporating all basic operating modes. Smart Structures and Systems,

114

2014. 14(5): 901–916.[115] Mazlan, S. A., Ekreem, N. B. and Olabi, A. G. The performance of

magnetorheological fluid in squeeze mode. Smart Materials and Structures,2007. 16(5): 1678–1682.

[116] Mazlan, S. A., Issa, A., Chowdhury, H. A. and Olabi, A. G. Magnetic circuitdesign for the squeeze mode experiments on magnetorheological fluids.Materials & Design, 2009. 30(6): 1985–1993.

[117] Farjoud, A., Ahmadian, M., Mahmoodi, N., Zhang, X. and Craft, M.Nonlinear modeling and testing of magneto-rheological fluids in low shearrate squeezing flows. Smart Materials and Structures, 2011. 20(8): 085013.

[118] Wang, H., Bi, C., Kan, J., Gao, C. and Xiao, W. The mechanical propertyof magnetorheological fluid under compression, elongation, and shearing.Journal of Intelligent Material Systems and Structures, 2011. 22(8): 811–816.

[119] Aydar, G., Wang, X. and Gordaninejad, F. A novel two-way-controllablemagneto-rheological fluid damper. Smart Materials and Structures, 2010.19(6): 065024.

[120] Bose, H. and Ehrlich, J. Performance of magnetorheological fluids in a noveldamper with excellent fail-safe behavior. Journal of Intelligent Material

Systems and Structures, 2009. 21(15): 1537–1542.[121] Kostamo, E., Kostamo, J., Kajaste, J. and Pietola, M. Magnetorheological

valve in servo applications. Journal of Intelligent Material Systems and

Structures, 2012. 23(9): 1001–1010.[122] Hu, G., Long, M., Yu, L. and Li, W. Design and performance evaluation

of a novel magnetorheological valve with a tunable resistance gap. Smart

Materials and Structures, 2014. 23(12): 127001.[123] Cook, E., Hu, W. and Wereley, N. M. Magnetorheological bypass damper

exploiting flow through a porous channel. Journal of Intelligent Material

Systems and Structures, 2007. 18(12): 1197–1203.[124] Sahin, H., Wang, X. and Gordaninejad, F. Magneto-rheological fluid flow

through complex valve geometries. International Journal of Vehicle Design,2013. 63(2/3): 241–255.

[125] Snyder, R. A., Kamath, G. M. and Wereley, N. M. Characterization andanalysis of magnetorheological damper behavior under sinusoidal loading.AIAA Journal, 2001. 39(7): 1240–1253.

[126] Spaggiari, A. and Dragoni, E. Efficient dynamic modelling andcharacterization of a magnetorheological damper. Meccanica, 2012. 47(8):2041–2054.

[127] Domnguez-Gonzlez, A., Stiharu, I. and Sedaghati, R. Practical hysteresis

115

model for magnetorheological dampers. Journal of Intelligent Material

Systems and Structures, 2014. 25(8): 967–979.[128] Case, D., Taheri, B. and Richer, E. Dynamical modeling and experimental

study of a small-scale magnetorheological damper. IEEE/ASME

Transactions on Mechatronics, 2014. 19(3): 1015–1024.[129] Balamurugan, L., Jancirani, J. and Eltantawie, M. A. Generalized

magnetorheological (MR) damper model and its application in semi-activecontrol of vehicle suspension system. International Journal of Automotive

Technology, 2014. 15(3): 419–427.[130] Susan-Resiga, D. A rheological model for magneto-rheological fluids.

Journal of Intelligent Material Systems and Structures, 2009. 20(8): 1001–1010.

[131] Wang, X. and Gordaninejad, F. Flow analysis of field-controllable, electro-and magneto-rheological fluids using Herschel-Bulkley model. Journal of

Intelligent Material Systems and Structures, 1999. 10(8): 601–608.[132] Li, W., Du, H. and Guo, N. Finite element analysis and simulation evaluation

of a magnetorheological valve. The International Journal of Advanced

Manufacturing Technology, 2003. 21(6): 438–445.[133] Wereley, N. M. and Pang, L. Nondimensional analysis of semi-active

electrorheological and magnetorheological dampers using approximateparallel plate models. Smart Materials and Structures, 1998. 7(5): 732–743.

[134] Yoshida, K., Soga, T., Kawachi, M., Edamura, K. and Yokota, S. Magneto-rheological valve-integrated cylinder and its application. Proceedings of the

Institution of Mechanical Engineers, Part I: Journal of Systems and Control

Engineering, 2010. 224(1): 31–40.[135] Kwok, N., Ha, Q., Nguyen, T., Li, J. and Samali, B. A novel hysteretic

model for magnetorheological fluid dampers and parameter identificationusing particle swarm optimization. Sensors and Actuators A: Physical,2006. 132(2): 441–451.

[136] Kwok, N. M., Ha, Q. P., Nguyen, M. T., Li, J. and Samali, B. Bouc-Wenmodel parameter identification for a MR fluid damper using computationallyefficient GA. ISA transactions, 2007. 46(2): 167–79.

[137] Sakai, C., Ohmori, H. and Sano, A. Modeling of MR damper with hysteresisfor adaptive vibration control. Proceedings of 42nd IEEE International

Conference on Decision and Control. Maui, Hawaii: IEEE. 2003, December.ISBN 0-7803-7924-1. 3840–3845.

[138] Jimenez, R. and Avarez-Icaza, L. LuGre friction model for amagnetorheological damper. Structural Control and Health Monitoring,2005. 12(1): 91–116.

116

[139] Ikhouane, F. and Rodellar, J. On the hysteretic BoucWen model part I:forced limit cycle characterization. Nonlinear Dynamics, 2005. 42(1): 63–78.

[140] Ikhouane, F. and Rodellar, J. On the hysteretic BoucWen model part II:robust parametric identification. Nonlinear Dynamics, 2005. 42(1): 79–95.

[141] Ikhouane, F., Hurtado, J. E. and Rodellar, J. Variation of the hysteresis loopwith the BoucWen model parameters. Nonlinear Dynamics, 2006. 48(4):361–380.

[142] Ikhouane, F. and Dyke, S. J. Modeling and identification of a shear modemagnetorheological damper. Smart Materials and Structures, 2007. 16(3):605–616.

[143] Rochdi, Y., Giri, F., Ikhouane, F., Chaoui, F. Z. and Rodellar, J. Parametricidentification of nonlinear hysteretic systems. Nonlinear Dynamics, 2009.58(1-2): 393–404.

[144] Gavin, H. P. Multi-duct ER dampers. Journal of Intelligent Material Systems

and Structures, 2001. 12(5): 353–366.[145] Jiang, Z. and Christenson, R. E. A fully dynamic magneto-rheological fluid

damper model. Smart Materials and Structures, 2012. 21(6): 065002.[146] Sahin, I., Engin, T. and Cesmeci, S. Comparison of some existing

parametric models for magnetorheological fluid dampers. Smart Materials

and Structures, 2010. 19(3): 035012.[147] Choi, S.-B., Lee, S.-K. and Park, Y.-P. A hysteresis model for the field-

dependent damping force of a magnetorheological damper. Journal of

Sound and Vibration, 2001. 245(2): 375–383.[148] Ubaidillah, Hudha, K. and Kadir, F. A. Modelling , characterisation and

force tracking control of a magnetorheological damper under harmonicexcitation. International Journal of Modelling, Identification and Control,2011. 13(1/2): 9–21.

[149] Du, H. P., Sze, K. Y. and Lam, J. Semi-active H-infinity control ofvehicle suspension with magneto-rheological dampers. Journal of Sound

and Vibration, 2005. 283(3-5): 981–996.[150] Xia, P.-Q. An inverse model of MR damper using optimal neural network

and system identification. Journal of Sound and Vibration, 2003. 266(5):1009–1023.

[151] Wang, D. H. and Liao, W. H. Modeling and control of magnetorheologicalfluid dampers using neural networks. Smart Materials and Structures, 2005.14(1): 111–126.

[152] Tudon-Martınez, J. C., Lozoya-Santos, J. J., Morales-Menendez, R. andRamirez-Mendoza, R. A. An experimental artificial-neural-network-based

117

modeling of magneto-rheological fluid dampers. Smart Materials and

Structures, 2012. 21(8): 085007.[153] Boada, M., Calvo, J., Boada, B. and Dıaz, V. Modeling of a

magnetorheological damper by recursive lazy learning. International

Journal of Non-Linear Mechanics, 2011. 46(3): 479–485.[154] Zeinali, M., Mazlan, S. A., Abd Fatah, A. Y. and Zamzuri, H. A

phenomenological dynamic model of a magnetorheological damper using aneuro-fuzzy system. Smart Materials and Structures, 2013. 22(12): 125013.

[155] Chen, C. and Liao, W.-H. A self-powered, self-sensing magnetorheologicaldamper. 2010 IEEE International Conference on Mechatronics and

Automation. IEEE. 2010. ISBN 978-1-4244-5140-1. 1364–1369.[156] Facey, W. B., Rosenfeld, N. C., Choi, Y.-T., Wereley, N. M., Choi, S. B.

and Chen, P. Design and testing of a compact magnetorheological Damperfor high impulsive loads. International Journal of Modern Physics B, 2005.19(7-8-9): 1549–1555.

[157] Chen, C. and Liao, W. H. A self-sensing magnetorheological damper withpower generation. Smart Materials and Structures, 2012. 21(2): 025014.

[158] Mao, M., Hu, W., Choi, Y.-T. and Wereley, N. M. A magnetorheologicaldamper with bifold valves for shock and vibration mitigation. Journal of

Intelligent Material Systems and Structures, 2007. 18(12): 1227–1232.[159] Zhang, J. Q., Feng, Z. Z. and Jing, Q. Optimization analysis of a new vane

MRF damper. Journal of Physics: Conference Series, 2009. 149: 012087.[160] Yang, L., Chen, S. Z., Zhang, B. and Feng, Z. Z. A rotary

magnetorheological damper for a tracked vehicle. Advanced Materials

Research, 2011. 328-330: 1135–1138.[161] Nguyen, Q., Han, Y., Choi, S. and Wereley, N. Geometry optimization of

MR valves constrained in a specific volume using the finite element method.Smart Materials and Structures, 2007. 16(6): 2242–2252.

[162] Choi, Y.-T. and Wereley, N. M. Comparative analysis of the time response ofelectrorheological and magnetorheological dampers using nondimensionalparameters. Journal of Intelligent Material Systems and Structures, 2002.13(7-8): 443–451.

[163] Rosenfeld, N. C. and Wereley, N. M. Volume-constrained optimizationof magnetorheological and electrorheological valves and dampers. Smart

Materials and Structures, 2004. 13(6): 1303–1313.[164] Hong, S., Choi, S., Choi, Y. and Wereley, N. Non-dimensional analysis and

design of a magnetorheological damper. Journal of Sound and Vibration,2005. 288(4-5): 847–863.

[165] Sodeyama, H., Sunakoda, K., Fujitani, H., Soda, S., Iwata, N. and Hata,

118

K. Dynamic Tests and Simulation of Magneto-Rheological Dampers.Computer-Aided Civil and Infrastructure Engineering, 2003. 18(1): 45–57.

[166] Fujitani, H., Sodeyama, H., Tomura, T., Hiwatashi, T., Shiozaki, Y., Hata,K., Sunakoda, K., Morishita, S. and Soda, S. Development of 400kNmagnetorheological damper for a real base-isolated building. Proc. SPIE,2003. 5052: 265–276.

[167] Lord Corp. Lord technical data: MRF-132DG magneto-rheological fluid,2011.

[168] Lord Corp. Lord product selector guide: Lord magneto-rheological fluids,2008.

[169] Tan, K., Stanway, R. and Bullough, W. Braking responses of inertia/load byusing an electro-rheological (ER) brake. Mechatronics, 2007. 17(6): 277–289.

[170] Salloom, M. Y. and Samad, Z. Experimental test of magneto-rheologicaldirectional control valve. Advanced Materials Research, 2011. 383-390:5409–5413.

[171] Ismail, I., Mazlan, S. A., Zamzuri, H. and Olabi, A. G. Fluid particleseparation of magnetorheological fluid in squeeze mode. Japanese Journal

of Applied Physics, 2012. 51: 067301.[172] Li, W. H., Yao, G. Z., Chen, G., Yeo, S. H. and Yap, F. F. Testing and steady

state modeling of a linear MR damper under sinusoidal loading. Smart

Materials and Structures, 2000. 9(1): 95–102.[173] Yang, G. Large-scale magnetorheological fluid damper for vibration

mitigation: modeling, testing and control. Phd dissertation. University ofNotre Dame. 2001.

[174] Zhang, H. H., Liao, C. R., Yu, M. and Huang, S. L. A study of an innerbypass magneto-rheological damper with magnetic bias. Smart Materials

and Structures, 2007. 16(5): N40–N46.[175] Yun, Y.-W., Lee, S.-M. and Park, M.-K. A study on the efficiency

improvement of a passive oil damper using an MR accumulator. Journal

of Mechanical Science and Technology, 2010. 24(11): 2297–2305.[176] Canudas de Wit, C., Olsson, H., Astrom, K. and Lischinsky, P. A new

model for control of systems with friction. IEEE Transactions on Automatic

Control, 1995. 40(3): 419–425.[177] Lischinsky, P., Canudas-de Wit, C. and Morel, G. Friction compensation for

an industrial hydraulic robot. IEEE Control Systems Magazine, 1999. 19(1):25–32.

[178] A strom, K. and Hagglund, T. The future of PID control. Control

Engineering Practice, 2001. 9(11): 1163–1175.

119

[179] Crawford, N. M., Cunningham, G. and Spedding, P. L. Prediction ofpressure drop for turbulent fluid flow in 90 bends. Proceedings of the

Institution of Mechanical Engineers, Part E: Journal of Process Mechanical

Engineering, 2003. 217(3): 153–155.[180] Spedding, P., Benard, E. and Mcnally, G. Fluid Flow through 90 Degree

Bends. Developments in Chemical Engineering and Mineral Processing,2004. 12(1-2): 107–128.

[181] Etemad, S. G. Turbulent flow friction loss coefficients of fittings for purelyviscous non-newtonian fluids. International Communications in Heat and

Mass Transfer, 2004. 31(5): 763–771.

APPENDIX A

CAD DRAWINGS

121

ISO

MET

RIC

VIE

W

BB

16 +

0.0

2 TH

RU A

LL

40 -

0.02

4X M

4X0.

7 3

.5O

-ring

Gro

ove

1.09

4X M

4X0.

7 3

.5

10

10.

0°

5

38

22

.64

2

15

16

+ 0

.02

16

.98

SEC

TION

B-B

LEFT

VIE

WRI

GHT

VIE

WB C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R:

0.0

1

AN

GUL

AR:

0

.01

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:1

:1SH

EET

2 O

F 2

A4

C

Alu

mun

ium

WEI

GHT

:

FI

A2

Coi

l Bob

bin

Figure A.1 CAD drawing of coil bobbin

122

AA

50

15

3/8

inch

1

2

6

+ 0.

02 T

HRU

ALL

4X M

4X0.

7 TH

RU A

LL

40 +

0.0

2

50

16

- 0.

02

19

- 0.0

2 5

12

5 -

0.02

24

15

SEC

TION

A-A

ISO

MET

RIC

VIE

W

TOP

VIE

WBO

TTO

M V

IEW

B C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R:

0.0

1

AN

GUL

AR:

0

.01

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:1

:1SH

EET

1 O

F 2

A4

C

Mild

Ste

el 1

010

WEI

GHT

:

FI

A1

Val

ve C

asin

g

Figure A.2 CAD drawing of valve casing version 1

123

50

39

15

24

12.5 11.5

A A

11

.811

(BSP

P 1/

4 FE

MA

LE TH

REA

D

50

49

22

6 + 0.02

6

+ 0.

02

16

- 0.

02

40

+ 0

.1

44

5 - 0.02

14 + 0.02

10

CUT-FACE A-A

Valve Casing Female

Casing2_rev_ver2

29/4/13FI

WEIGHT:

Mild Steel 1010A4

SHEET 1 OF 1SCALE:1:1

DWG NO.

TITLE:

REVISIONDO NOT SCALE DRAWING

MATERIAL:

DATESIGNATURENAME

DEBUR AND BREAK SHARP EDGES

FINISH:UNLESS OTHERWISE SPECIFIED:DIMENSIONS ARE IN MILLIMETERSSURFACE FINISH:TOLERANCES: LINEAR: ANGULAR:

Q.A

MFG

APPV'D

CHK'D

DRAWN

Figure A.3 CAD drawing of valve casing version 2 - female

124

50

50

46

15

23 12

A A

50

49

22

11

.811

(BSP

P 1/

4 FE

MA

LE TH

REA

D)

6 + 0.02

5 - 0.02

25 + 0.02

10

40

+ 0

.1

16

- 0.

02

6

+ 0.

02

CUT-FACE A-A

Valve Casing Male29/4/13FI

WEIGHT:

MILD STEEL 1010A4

SHEET 1 OF 1SCALE:1:1

DWG NO.

TITLE:

REVISIONDO NOT SCALE DRAWING

MATERIAL:

DATESIGNATURENAME

DEBUR AND BREAK SHARP EDGES

FINISH:UNLESS OTHERWISE SPECIFIED:DIMENSIONS ARE IN MILLIMETERSSURFACE FINISH:TOLERANCES: LINEAR: ANGULAR:

Q.A

MFG

APPV'D

CHK'D

DRAWN

Casing2_rev_ver2_2

Figure A.4 CAD drawing of valve casing version 2 - male

125

ISO

MET

RIC

VIE

W

AA

R5.

5

R6.

5 +

0.03

7

0°

20°

16

- 0.

02

0.5

3 -

0.02

2.5

- 0.

02

1 +

0.0

3

SEC

TION

A-A

TOP

VIE

WBO

TTO

M V

IEW

B C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R: 0

.05

A

NG

ULA

R:

0.0

5

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:3

:1SH

EET

1 O

F 2

A4

C

Alu

mun

ium

WEI

GHT

:

FI

B1

Bush

0.5

mm

Figure A.5 CAD drawing of spacer - 0.5 mm

126

ISO

MET

RIC

VIE

W

BB

R5.

5

R6.

5 +

0.03

7

0°

20°

16

- 0.

02

0.5

3 -

0.02

2 -

0.02

1

+ 0

.03

SEC

TION

B-B

TOP

VIE

WBO

TTO

M V

IEW

B C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R:

0.0

5

AN

GUL

AR:

0

.05

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:3

:1SH

EET

2 O

F 2

A4

C

Alu

mun

ium

WEI

GHT

:

FI

B2

Bush

1m

m

Figure A.6 CAD drawing of spacer - 1 mm

127

ISO

MET

RIC

VIE

W 5

- 0.

02

10.

5

15.

5

5

- 0.0

2

13

- 0.

02

TOP

VIE

W

B C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R: 0

.05

A

NG

ULA

R: 0

.05

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:3

:1SH

EET

1 O

F 4

A4

C

Mild

Ste

el 1

010

WEI

GHT

:

FI

C1

Side

Cor

e 1

Figure A.7 CAD drawing of side core

128

5 -

0.02

16 -

0.02

5 TH

RU A

LL

B C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R:

AN

GUL

AR:

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:3

:1SH

EET

1 O

F 2

A4

C

Mild

Ste

el 1

010

WEI

GHT

:

FI

D1

Orif

ice

Cor

e

Figure A.8 CAD drawing of orifice core

129

5 -

0.02

13 -

0.02

4 -

0.02

13 -

0.02

B C D

12

A

32

14

BA

56

DRA

WN

CHK

'D

APP

V'D

MFG

Q.A

UNLE

SS O

THER

WIS

E SP

ECIF

IED

:D

IMEN

SIO

NS

ARE

IN M

ILLIM

ETER

SSU

RFA

CE

FIN

ISH:

TOLE

RAN

CES

:

LINEA

R:

AN

GUL

AR:

FINIS

H:D

EBUR

AN

D

BREA

K SH

ARP

ED

GES

NA

ME

SIG

NA

TURE

DA

TE

MA

TERI

AL:

DO

NO

T SC

ALE

DRA

WIN

GRE

VISI

ON

TITLE

:

DW

G N

O.

SCA

LE:3

:1SH

EET

2 O

F 2

A4

C

Mild

Ste

el 1

010

WEI

GHT

:

FI

D2

Mid

dle

Cor

e

Figure A.9 CAD drawing of center core