Professional Documents
Culture Documents
1 Mathematical Modeling of Piezo Electrical Energy Harvesters On Half Dump Truck Model
1 Mathematical Modeling of Piezo Electrical Energy Harvesters On Half Dump Truck Model
1 Mathematical Modeling of Piezo Electrical Energy Harvesters On Half Dump Truck Model
ISTANBUL UNIVERSITY-CERRAHPASA
INSTITUTE OF GRADUATE STUDIES
Ph.D. THESIS
SUPERVISOR
Prof. Dr. Erol UZAL
Assoc. Prof. Dr. Erman ASLAN Assist. Prof. Dr. Lütfi Emir SAKMAN
Kocaeli University İstanbul University-Cerrahpaşa
Faculty of Engineering Faculty
As required by the 9/2 and 22/2 articles of the Graduate Education Regulation which was
published in the Official Gazette on 20.04.2016, this graduate thesis is reported as in accordance
with criteria determined by the Institute of Graduate Studies by using the plagiarism software
to which Istanbul University-Cerrahpasa is a subscriber.
FOREWORD
The basis for this research originally stemmed from my passion for developing innovative
methods to harvest energy from the ambient environment. In the automotive industry which is
known as one of the major CO2 contributors, there is an increasing necessity to reduce the use
of polluting fossil fuels which are harmful to both the environment and public health.
Piezoelectric transducing is one of the most used energy harvesting methods due to its high-
power density and its accessibility. Most energy harvesting mechanisms present in our daily
life use the piezoelectric effect. Piezoelectric transducers have been used in the automotive
industry for a long time but have not reached their full potential in this sector yet. Piezoelectric
energy harvesting has great potential for the automotive industry by reducing greenhouse gases
emission and boosting vehicles’ energy efficiency.
The main sources for energy harvesting on a vehicle are heat losses, vibrations, and braking
energy. The suspension system is one of the supporting key components of vehicles that
intervene in the vibration absorption process. The vibrations are generated from bumpy and
uneven roads, especially at mining sites. Energy harvested from suspensions serves as a
supplement to the onboard alternator. Indeed, this energy can be used to charge the battery and
to power wireless sensors such as tire pressure, vehicle speed, and strain monitoring sensors.
Energy wasted through vibrations is abundant so a sustainable way to recover this energy is
handy and brings several environmental and economic advantages. Even though the suspension
system of off-road vehicles is subjected to vibrations and strain, limited research has been done
on mathematical modeling of these parts' energy harvesting systems. The purpose of this
research work is to develop a mathematical model to calculate the output charge and voltage
from the tire and suspension piezoelectric energy harvesting system of off-road vehicles in
general and dump trucks particularly. These mathematical models are based on the fundamental
laws of physics. These laws were derived from several years of scientific experiments and
observations. Thanks to this model, the suspension system’s dynamics will be represented and
predicted. In order to analyze off-road vehicles’ dynamics, lumped parameter models can be
used.
iv
TABLE OF CONTENTS
Page
FOREWORD .........................................................................................................................iv
TABLE OF CONTENTS ....................................................................................................... v
LIST OF FIGURES ................................................................................................................ x
LIST OF TABLES ................................................................................................................ xx
LIST OF SYMBOLS AND ABBREVIATIONS ............................................................. xxii
ÖZET ..................................................................................................................................xxvi
SUMMARY ...................................................................................................................... xxvii
1. INTRODUCTION ............................................................................................................. 1
1.1. SIGNIFICANCE ...................................................................................................... 1
1.2. PROBLEM STATEMENT ...................................................................................... 2
1.3. RESEARCH OBJECTIVES .................................................................................... 3
1.4. THESIS OUTLINE .................................................................................................. 4
2. LITERATURE REVIEW ................................................................................................. 5
2.1. BACKGROUND ..................................................................................................... 5
2.2. PIEZOELECTRIC ENERGY HARVESTING ....................................................... 9
2.2.1. History of Piezoelectricity ................................................................................ 10
2.2.2. Piezoelectric Characteristics ............................................................................. 12
2.2.2.1. Direct and Indirect Piezoelectric Effects .................................................. 12
2.2.2.2. Ferroelectric Domains of Crystals and Poling ......................................... 14
2.2.2.3. The Curie Temperature ............................................................................. 15
2.2.2.4. Dielectric Hysteresis Loops of Ferroelectrics .......................................... 17
2.2.2.5. The Coupling Factor k .............................................................................. 22
2.2.3. Piezoelectricity Linear Theory ......................................................................... 24
2.2.3.1. Basic Equations of Piezoelectricity ........................................................... 26
2.2.3.2. Constitutive Equations .............................................................................. 27
2.2.3.3. The Elasto-Piezo-Dielectric Coefficients .................................................. 31
2.2.4. Power Harvested in AC Mode (Alternating Current Impedance Matching
-ACIM) ....................................................................................................................... 34
2.2.4.1. Influence of a Low Coupling on the Mechanical Domain ........................ 38
2.2.4.2. Influence of a High Coupling on the Mechanical Domain ....................... 40
v
2.2.5. The Limitations of The Piezoelectricity Linear Theory ................................... 42
2.2.5.1. Frequency Limitations............................................................................... 42
2.2.5.2. The Electrostriction ................................................................................... 43
2.2.5.3. The Depolarization.................................................................................... 43
2.3. PROCESSING OF PIEZOELECTRIC MATERIALS .......................................... 44
2.3.1. Single Crystals .................................................................................................. 44
2.3.1.1. Ferroelectric Single Crystals .................................................................... 44
2.3.1.2. Non-Ferroelectric Single Crystals ............................................................ 45
2.3.2. Polymer Films .................................................................................................. 46
2.3.3. Ceramics ........................................................................................................... 48
2.4. RESONANT PIEZOELECTRIC TRANSFORMERS .......................................... 51
2.4.1. The Rosen-type resonant piezoelectric transformers........................................ 53
2.4.2. Topologies of the Rosen-type Piezoelectric Transformer. ............................... 56
2.4.3. Influence of a Connected Load Impedance on the Output of a Piezoelectric
Transformer. ............................................................................................................... 59
2.4.4. Influence of Q on the resonant frequency of the transformer........................... 60
2.4.5. Limitations of the Resonant Transformers in Measuring Applications ........... 61
2.5. MAIN COMPONENTS OF A PIEZOELECTRIC ENERGY HARVESTING
SYSTEM ......................................................................................................................... 62
2.5.1. The Piezoelectric Element ................................................................................ 62
2.5.2. The Host Structure ............................................................................................ 63
2.5.3. Electronics ........................................................................................................ 63
2.6. CIRCUITS AND BASICS OF ELECTRONICS .................................................. 63
2.6.1. Basic Equations ................................................................................................ 65
2.6.2. Semiconductor Components ............................................................................. 73
2.6.2.1. Rectifiers and Diodes ................................................................................ 73
2.6.2.2. Field Effect Transistor .............................................................................. 76
2.6.2.3. Bipolar Junction Transistor ...................................................................... 78
2.6.3. The Direct Current Impedance Matching (DCIM) ........................................... 81
2.6.3.1. Schematic representation of the circuit ..................................................... 81
2.6.3.2. Governing Equations in the DCIM Circuit ............................................... 83
2.6.3.3. Influence of a Low Coupling on the DCIM Circuit ................................... 84
2.6.3.4. Influence of a High Coupling on the DCIM Circuit ................................. 86
2.6.3.5. Maximum Harvested Power ...................................................................... 88
2.6.4. Synchronized Switch Harvesting on Inductor Circuit ...................................... 89
vi
2.6.4.1. Schematic Representation of the Circuit ................................................... 89
2.6.4.2. Governing Equations in the SSHI Circuit ................................................. 91
2.6.4.3. Influence of a Low Coupling on the SSHI Circuit ..................................... 93
2.6.4.4. Maximum Harvested Power ...................................................................... 96
2.6.5. The Synchronous Electric Charge Extraction Circuit ...................................... 97
2.6.5.1. Schematic Representation of the Circuit ................................................... 97
2.6.5.2. The Maximum Harvested Power ............................................................... 99
2.7. SUSPENSION SYSTEMS .................................................................................. 100
2.7.1. Main Types of Suspension System ................................................................. 100
2.7.2. Boundary Friction ........................................................................................... 102
2.7.3. Level Control and Adjustment ....................................................................... 103
2.7.4. Spring Characteristics of Suspension Systems ............................................... 105
2.7.4.1. Particular States of a Hydro-pneumatic Suspension System .................. 105
2.7.4.2. Spring Stiffness Rate................................................................................ 105
2.7.5. Damping Characteristics of Hydro-pneumatic Suspensions .......................... 111
2.7.5.1. Fluid Friction Damping. ......................................................................... 112
2.7.5.2. End of Stroke Damping ........................................................................... 114
2.7.6. Mathematical Car Models .............................................................................. 116
2.7.6.1. The Quarter Car Model........................................................................... 116
2.7.6.2. The Half Car Model ................................................................................ 120
2.7.6.3. The Full Car Model ................................................................................. 124
2.8. SUSPENSION ENERGY HARVESTING .......................................................... 126
2.9. TIRE PRESSURE MONITORING SYSTEM .................................................... 135
2.9.1. Direct Tire Pressure Monitoring System ........................................................ 135
2.9.1.1. Active TPMS ............................................................................................ 138
2.9.1.2. Passive TPMS .......................................................................................... 139
2.9.2. Indirect Tire pressure Monitoring System ...................................................... 140
2.9.3. Hybrid Tire Pressure Monitoring System ....................................................... 141
3. MATERIALS AND METHODS .................................................................................. 144
3.1. SUSPENSION SYSTEM PIEZOELECTRIC ENERGY HARVESTER
CRICUITRY ................................................................................................................. 144
3.2. GOVERNING EQUATIONS OF THE MATHEMATICAL MODELS ............ 146
3.2.1. Mathematical Models Excluding the Piezoelectric Material .......................... 146
3.2.1.1. Quarter Car Model Excluding the Piezoelectric Material...................... 146
3.2.1.2. Half Car Model Excluding the Piezoelectric Material ........................... 148
vii
3.2.1.3. Non-Linear Quarter Car Model .................................................................. 150
3.2.2. Mathematical Models Including the Piezoelectric Material ........................... 151
3.2.2.1. Piezoelectric Material Model .................................................................. 151
3.2.2.2. Quarter Car Model Including the Piezoelectric Material (PZT). ........... 156
3.2.2.3. Half Car Model Including the Piezoelectric Material PZT. ................... 158
3.2.2.4. Non-linear Quarter Car Model Including the Piezoelectric Material
(PZT). 165
4. RESULTS ....................................................................................................................... 167
4.1. THE PIEZOELECTRIC STACK AND TRUCK PARAMETERS..................... 167
4.2. STEP INPUT RESPONSE .................................................................................. 170
4.2.1. QCM Step input response ............................................................................... 170
4.2.2. HCM Step İnput Response. ............................................................................ 173
4.3. HARMONIC EXCITATION RESPONSE.......................................................... 175
4.3.1. Road Wavelength Classification .................................................................... 176
4.3.2. Bode Plots ....................................................................................................... 176
4.3.2.1. Bode Magnitude Plot Of The QCM ......................................................... 176
4.3.2.2. Bode Magnitude Plot of the HCM ........................................................... 177
4.3.3. QCM Harmonic Excitation Response ............................................................ 178
4.3.4. HCM Harmonic Excitation Response ............................................................ 181
4.3.5. Non-linear QCM harmonic Excitation ........................................................... 184
4.3.6. Frequency Domain Analysis .......................................................................... 187
4.3.6.1. Advantages of the Frequency Domain Analysis ...................................... 187
4.3.5.1. Laplace Transform of the QCM and HCM equations. ................................ 187
4.3.7. Power Dissipated from the QCM Suspension System in the Frequency
Domain. .................................................................................................................... 197
4.3.8. Power Dissipated from the HCM Suspension System in the Frequency
Domain. .................................................................................................................... 199
4.3.9. QCM Harvested Voltage and Power in the Frequency Domain. ................... 200
4.3.10. HCM Harvested Voltage and Power in the Frequency Domain. ................... 201
4.3.11. QCM Harvested Voltage and Power in the Velocity Domain........................ 201
4.3.12. HCM Harvested Voltage and Power in the Velocity Domain........................ 203
4.4. RANDOM EXCITATIONS ................................................................................ 204
4.4.1. Random Excitation results of the QCM ......................................................... 208
4.4.2. Random Excitation results of the HCM ......................................................... 214
4.4.3. QCM class D random Excitation Harvested Power and Voltage according
to the Truck’s Velocity ............................................................................................. 219
viii
4.4.4. HCM class D random Excitation Harvested Power and Voltage according
to the Truck’s velocity. ............................................................................................. 220
4.5. FULL CAR MODEL SPRUNG DISPLACEMENT ........................................... 222
4.5.1. State space expressions of the full car model ................................................. 226
4.5.2. Full Car Model Bode Magnitude Plot Sprung Mass ...................................... 233
4.5.3. Full Car Model Step Response ....................................................................... 234
4.5.4. Full Car Model Harmonic Excitation Response ............................................. 235
5. DISCUSSION ................................................................................................................. 237
5.1. EFFECTS OF ROAD PARAMETERS ON THE HARVESTED VOLTAGE
AND POWER OF THE HCM. ..................................................................................... 237
5.2. EFFECTS OF PIEZOELECTRIC STACK PARAMETERS ON THE
HARVESTED VOLTAGE AND POWER OF THE HCM. ......................................... 238
5.2.1. Effect of the Force Factor of the Piezoelectric Stack on the Harvested
Power and Voltage. ................................................................................................... 238
5.2.2. Effect of the Capacitance of the Piezoelectric Stack on the Harvested
Power and Voltage. ................................................................................................... 239
5.3. EFFECTS OF THE TRUCK PARAMETERS ON THE HARVESTED
VOLTAGE AND POWER OF THE HCM. ................................................................. 240
5.3.1. Effect of the Suspension Damping Coefficients bsf and bsr on the
Harvested Voltage and Power. ................................................................................. 240
5.3.2. Effect of the Unsprung Damping Coefficients buf and bur on the
Harvested Voltage and Power. ................................................................................. 241
5.3.3. Effect of the Suspension Stiffness Coefficients Ksf and Ksr on the
Harvested Voltage and Power. ................................................................................. 242
5.3.4. Effect of the Unsprung (tire) Stiffness Coefficients Kuf and Kur on the
Harvested Voltage and Power. ................................................................................. 243
5.3.5. Effect of the Sprung Mass MS on the Harvested Voltage and Power. ........... 244
5.3.6. Effect of the Unsprung Masses Muf and Mus on the Harvested Voltage and
Power. ....................................................................................................................... 245
6. CONCLUSION AND RECOMMENDATIONS ......................................................... 247
6.1. CONCLUSION .................................................................................................... 247
6.2. RECOMMENDATIONS FOR FUTURE WORKS ............................................ 249
REFERENCES ................................................................................................................... 251
APPENDICES..................................................................................................................... 267
CURRICULUM VITAE .................................................................................................... 275
ix
LIST OF FIGURES
Page
Figure 2.1: Power density vs voltage comparison plot of different energy harvesting
techniques [37]. ................................................................................................................. 7
Figure 2.2: Piezoelectric energy harvesting and its main applications [48-57]. ..................... 9
Figure 2.5: Piezo ceramic crystals, domains, and domain walls [67]. .................................. 15
Figure 2.7: Perovskite structure (unit cell) of PZT above (a) and below (b) the curie
temperature [71]. ............................................................................................................. 16
Figure 2.10: (a) Strain during poling (b) Strain hysteresis loop [77]. ................................... 21
Figure 2.11: The three kinds of strain during the poling process of ferroelectric
ceramics. .......................................................................................................................... 22
Figure 2.12: Planar oscillations of a thin disc of piezoelectric material [58]. ....................... 23
Figure 2.15: 1D mass spring damper model with a piezo element. ...................................... 35
Figure 2.16: Natural frequency ratio from the open circuit (R=∞) to short circuit (R=0).
......................................................................................................................................... 37
Figure 2.17: The normalized response graph of the displacement for a coupling factor
𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05. ............................................. 39
Figure 2.18: The normalized response graph of the voltage for a coupling factor𝑘𝑒2 =
0.01 and a dimensionless damping ratio 𝜉 = 0.05. ........................................................ 39
x
Figure 2.19: The normalized response graph of the power for a coupling factor 𝑘𝑒2 =
0.01 and a dimensionless damping ratio 𝜉 = 0.05. ........................................................ 40
Figure 2.20: The normalized response graph of the displacement for a coupling factor
𝑘𝑒2 = 0.5 and of a dimensionless damping ratio 𝜉 = 0.05............................................ 41
Figure 2.21: The normalized response graph of the voltage for a coupling factor𝑘𝑒2 =
0.5 and of a dimensionless damping ratio 𝜉 = 0.05. ...................................................... 42
Figure 2.22: The normalized response graph of the power for a coupling factor 𝑘𝑒2 =
0.5 and of a dimensionless damping ratio 𝜉 = 0.05. ...................................................... 42
Figure 2.26: Processing of a PVDF polymer film (a) application of mechanical stress,
mechanical stretching (b) contact poling (c) corona poling (d) electrospinning
process [91]. .................................................................................................................... 48
Figure 2.33: Diagram of the narrow double poled ring piezoelectric transformer. ............... 57
Figure 2.34: Diagram of the thick double poled disc piezoelectric transformer. .................. 57
Figure 2.35: Diagram of the double poled disc piezoelectric transformer. ........................... 57
Figure 2.38: Influence of the connected load impedance on the transformation ratio. ......... 59
xi
Figure 2.41: The symbol of a resistor. ................................................................................... 67
Figure 2.51: The typical voltage curve of a full bridge rectifier. .......................................... 76
Figure 2.59: Waveform of the displacement input at the optimal resistance. ....................... 82
Figure 2.61: Waveform of the voltage at the optimal resistance - 𝑉𝑑𝑐 is represented in
full line while 𝑉𝑝 is represented in dotted lines. ............................................................. 82
Figure 2.62: The normalized response graph of the displacement for a low coupling
factor 𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit. ......... 85
Figure 2.63: The normalized response graph of the voltage for a low coupling factor
𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit. .................... 85
xii
Figure 2.64: The normalized response graph of the power for a low coupling factor
𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit. .................... 85
Figure 2.65: The normalized response graph of the displacement for a high coupling
factor 𝑘𝑒2 = 0.5 and of a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit. ....... 87
Figure 2.66: The normalized response graph of the voltage for a high coupling factor
𝑘𝑒2 = 0.5 and of a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit................... 87
Figure 2.67: The normalized response graph of the average power for a high coupling
factor 𝑘𝑒2 = 0.5 and of a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit. ....... 88
Figure 2.72: The normalized response graph of the displacement for a low coupling
factor 𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit. ........... 94
Figure 2.73: The normalized response graph of the voltage for a low coupling factor
𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit ...................... 94
Figure 2.74: The normalized response graph of the power for a low coupling factor
𝑘𝑒2 = 0.01 and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit. ..................... 95
Figure 2.75: The normalized response graph of the displacement for a high coupling
factor 𝑘𝑒2 = 0.5 and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit. ............. 95
Figure 2.76: The normalized response graph of the voltage for a high coupling factor
𝑘𝑒2 = 0.5 and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit. ........................ 96
Figure 2.77: The normalized response graph of the average power for a high coupling
factor 𝑘𝑒2 = 0.5 and a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit............. 96
Figure 2.80: Waveforms of the voltage of a SECE circuit. The full line represents 𝑉𝑝
while the dotted line represents 𝑉𝑑𝑐. .............................................................................. 99
Figure 2.81: Waveforms of the current of a SECE circuit. The full line represents 𝐼1
while the dotted line represents 𝐼2. ................................................................................. 99
xiii
Figure 2.83: force-displacement curves of mechanical, hydropneumatic, and pneumatic
suspension mechanisms [105]. ...................................................................................... 102
Figure 2.84: Forces equilibrium at the piston of a mono suspension system [105]. ........... 104
Figure 2.87: height of gas column ℎ0𝐹 for different static spring loads 𝐹𝐹1[105]. ........... 110
Figure 2.91: The quarter car model (two degrees of freedom)............................................ 117
Figure 2.92: Experimental model of the quarter car model with a real tire as unsprung
mass [113]. .................................................................................................................... 119
Figure 2.93: Experimental model of the quarter car model with a moving plate as
unsprung mass [114]...................................................................................................... 120
Figure 2.94: The half-car model (four degrees of freedom). ............................................... 121
Figure 2.95: Experimental setup of a half car model with two suspensions connected
by a beam [119]. ............................................................................................................ 123
Figure 2.96: Experimental setup of a half car model with the MRAs and steel frame
[120]. ............................................................................................................................. 123
Figure 2.97: The full car model (seven degrees of freedom). ............................................. 124
Figure 2.98: Quarter car modeled designed with piezoelectric material connected in
parallel [126]. ................................................................................................................ 126
Figure 2.99: half-car model with piezoelectric elements connected in parallel [128]. ....... 127
Figure 2.100: Quarter car model designed with piezoelectric material connected in
series. ............................................................................................................................. 129
Figure 2.101: Dual mass piezoelectric energy harvesting system installed on a tire. ......... 129
Figure 2.102: Quarter car model of car tire piezoelectric energy harvesting system. ......... 130
Figure 2.103: Multilayer piezoelectric stack installed inside the shock absorber [139]. .... 132
xiv
Figure 2.104: Quarter car model with the multilayer (ML PZT VEH) piezoelectric stack
connected in series with the spring [140]. ..................................................................... 133
Figure 2.107: Direct TPMS sensor module fitting in the wheel station [143]. ................... 137
Figure 2.109: Active TPMS transmitter schematic diagram [144]. .................................... 138
Figure 2.110: Active TPMS receiver schematic diagram [144]. ......................................... 138
Figure 2.111: Direct TPMS Piezoelectric ceramic power supply schematic diagram
[144]. ............................................................................................................................. 140
Figure 3.4: Series connection between the suspension spring and the piezoelectric
stack. .............................................................................................................................. 152
Figure 3.5: Parallel connection between the suspension spring and the piezoelectric
stack. .............................................................................................................................. 152
Figure 3.6: Piezoelectric stack and electric circuit set. ....................................................... 153
Figure 3.7: Equivalent electric circuit to the piezoelectric stack and electric circuit set. ... 155
Figure 3.8: Quarter car model with the piezoelectric stack. ................................................ 156
Figure 3.9: Half car model with the piezoelectric stack. ..................................................... 159
Figure 4.1: QCM Simulink representation using transfer functions of the motion
equations. ....................................................................................................................... 169
Figure 4.2: MATLAB Simulink model representation of HCM with piezoelectric stack
....................................................................................................................................... 170
xv
Figure 4.3: QCM step input. ................................................................................................ 171
Figure 4.4: QCM step input car body displacement plot..................................................... 171
Figure 4.5: QCM step input harvested voltage plot (v). ...................................................... 172
Figure 4.6: QCM step input harvested power plot (w). ....................................................... 172
Figure 4.8: HCM step input car body displacement plot..................................................... 174
Figure 4.9: HCM step input harvested voltage plot. ........................................................... 174
Figure 4.10: HCM step input harvested power plot. ........................................................... 175
Figure 4.14: QCM harmonic excitation car body displacement plot................................... 179
Figure 4.15: QCM harmonic excitation harvested voltage plot. ......................................... 180
Figure 4.16: QCM harmonic excitation harvested power plot. ........................................... 180
Figure 4.17: HCM rear tire harmonic excitation input plot................................................. 182
Figure 4.18: HCM harmonic excitation car body displacement plot. ................................. 182
Figure 4.19: HCM harmonic excitation front stack harvested voltage plot. ....................... 182
Figure 4.20: HCM harmonic excitation front stack harvested voltage plot. ....................... 183
Figure 4.22: Bode amplitude plot of the proposed non-linear model (cardody
dysplacement to harmonic excitation ratio). ................................................................. 185
Figure 4.25: Dissipated power from the QCM suspension in the frequency domain. ........ 198
Figure 4.26: Dissipated power from the QCM suspension according to the truck
velocity. ......................................................................................................................... 198
Figure 4.27: Dissipated power from the HCM suspension in the frequency domain. ........ 199
xvi
Figure 4.:28: Dissipated power from the HCM suspension according to the truck
velocity. ......................................................................................................................... 199
Figure 4.29: Maximum harvested voltage and power from the QCM in frequency
domain plots. ................................................................................................................. 200
Figure 4.30: Maximum harvested voltage and power from the HCM in frequency
domain plot. ................................................................................................................... 201
Figure 4.31: Maximum harvested voltage and power from the QCM according to the
truck’s velocity. ............................................................................................................. 202
Figure 4. 32: Maximum harvested voltage and power from the QCM within the truck’s
operating velocities. ....................................................................................................... 203
Figure 4.33: Maximum harvested voltage and power from the HCM according to the
truck’s velocity. ............................................................................................................. 203
Figure 4.34: Maximum harvested voltage and power from the QCM within the truck’s
operating velocities. ....................................................................................................... 203
Figure 4.35: Front tire random excitation source MATLAB SIMULINK representation.
....................................................................................................................................... 206
Figure 4.36: Rear tire random excitation source MATLAB SIMULINK representation.
....................................................................................................................................... 207
Figure 4.37: QCM class A road random excitation input plot. ........................................... 208
Figure 4.38: QCM class D road random excitation input plot. ........................................... 209
Figure 4.39: QCM class G road random excitation input plot. ........................................... 209
Figure 4.40: QCM class H road random excitation input plot. ........................................... 210
Figure 4.41: QCM class A road random excitation harvested voltage (V). ........................ 210
Figure 4.42: QCM class A road random excitation harvested power (W). ......................... 211
Figure 4.43: QCM class D road random excitation harvested voltage (V). ........................ 211
Figure 4.44: QCM class D road random excitation harvested power (W). ......................... 212
Figure 4.45: QCM class G road random excitation harvested voltage (V). ........................ 212
Figure 4.46: QCM class G road random excitation harvested power (W). ......................... 213
Figure 4.47: QCM class H road random excitation harvested voltage (V). ........................ 213
Figure 4.48: QCM class G road random excitation harvested power (W). ......................... 214
Figure 4.49: HCM class A road random excitation harvested voltage (V). ........................ 215
xvii
Figure 4.50: HCM class A road random excitation harvested power (W). ......................... 215
Figure 4.51: HCM class D road random excitation harvested voltage (V). ........................ 216
Figure 4.52: HCM class D road random excitation harvested power (W). ......................... 216
Figure 4.53: HCM class G road random excitation harvested voltage (V). ........................ 217
Figure 4.54: HCM class G road random excitation harvested power (W). ......................... 217
Figure 4.55: HCM class H road random excitation harvested voltage (V). ........................ 218
Figure 4.56: HCM class H road random excitation harvested power (W). ......................... 218
Figure 4.57: QCM Class D random excitation harvested power and voltage according
to the truck’s velocity. ................................................................................................... 219
Figure 4.58: QCM Class D random excitation harvested power and voltage in the
truck’s real operating velocities [20; 65km/h]............................................................... 220
Figure 4.59: HCM Class D random excitation harvested power and voltage according
to the truck’s velocity. ................................................................................................... 221
Figure 4.60: HCM Class D random excitation harvested power and voltage in the
truck’s real operating velocities [20; 65km/h]............................................................... 221
Figure 4.62: FCM sprung mass simulink representation using a transfer function. ........... 233
Figure 4.63: FCM sprung mass Bode magnitude plot. ........................................................ 234
Figure 4.65: FCM step road input sprung mass displacement plot. .................................... 235
Figure 4.67: FCM harmonic road excitation sprung mass displacement response. ............ 236
Figure 5.1: Effect of the road sine wave amplitude on the harvested voltage and power
of the HCM. ................................................................................................................... 237
Figure 5.2: Effects of the piezoelectric stack force factor on the harvested voltage and
power. ............................................................................................................................ 238
Figure 5.3: Effects of the piezoelectric stack capacitance on the harvested voltage and
power. ............................................................................................................................ 239
Figure 5.6: Effects of the suspension stiffness coefficients on the harvested power and
voltage. .......................................................................................................................... 243
Figure 5.7: Effects of the unsprung stiffness coefficients on the harvested power and
voltage. .......................................................................................................................... 244
Figure 5.8: Effects of the sprung mass on the harvested power and voltage. ..................... 245
Figure 5.9: Effects of the unsprung masses on the harvested power and voltage. .............. 246
xix
LIST OF TABLES
Page
Table 2.1: Comparison of energy harvesting methods and power densities [19, 26, 33-
35]. ..................................................................................................................................... 6
Table 2.8: Comparison of the properties of resonant and non-resonant transformers. ......... 62
Table 2.12: comparison of the main piezoelecric energy harvesting systems ..................... 133
Table 4.2: Quarter car model parameters (777F Off-highway Mining Dump Truck)......... 168
Table 4.4: HCM parameters (777F Off-highway Mining Dump Truck). ........................... 169
xx
Table 4.8: QCM harmonic excitation results....................................................................... 183
Table 4.10: ISO 8608 road roughness levels for spatial reference frequency 𝒏𝟎 =
𝟎. 𝟏𝒎 − 𝟏 [162]. ........................................................................................................... 207
Table 4.11: QCM random excitation plot peak values ........................................................ 214
Table 4.12: QCM random excitation plot peak values ........................................................ 219
xxi
LIST OF SYMBOLS AND ABBREVIATIONS
Symbol Explanation
a : Acceleration amplitude
km : kilometer
P : Power (W)
Q : Electric charge (Q)
M : Magnetithatn or magnetic polarization
H : Magnetic field intensity
B : Magnetic field density
P : Polarization
E : Electric intensity
D : Dielectric displacement
°C : degree Celsius
W : Watt
dB : Decibel
F : Farad
C : Coulomb
S : Strain
T : Stress
k : Coupling Factor
H : Henry (inductance)
𝑽𝑨𝑰 : Voltage After Inversion
𝑽𝑩𝑰 : Voltage Before Inversion
Mus : Mass of the unsprung mass (tire assembly)
Ms : Mass of the sprung mass(car body).
Ks : Sprung stiffness coefficient (car body)
bs : Sprung damping coefficient of the linear viscous damper (car body).
Kus : Unsprung stiffness coefficient (tire)
bus : Absorptivity (tire) respectively
YR : Road excitation
Yus : Vertical displacement of the unsprung mass (tire)
xxii
Ys : Vertical displacement of the sprung mass (car body)
Yuf : Vertical motion of the front unsprung mass (front tire)
Yur : Vertical motion of the rear unsprung mass (rear tire)
ϴs : Pitch or angular displacement and bounce
Muf : Front unsprung mass
Mur : Rear unsprug mass
Ksf : Front sprung stiffness coefficient
Ksr : Rear sprung stiffness coefficient
bsf : Front sprung damping coefficient
bsr : Rear sprung damping coefficient
Kuf : Front unsprung stiffness coefficient (front tire)
Kur : Rear unsprung stiffness coefficient (rear tire)
buf : Front unsprung damping coefficient (front tire)
bur : Rear unsprung damping coefficient (rear tire)
Kp : Short circuit stiffness of the piezoelectric stack,
Cp : Piezoelectric capacitance,
Yp : Relative displacement between the sprung and unsprung mass
𝝉 : Time delay (L/u)
L : Wheelbase (m)
u : Car velocity
α : Force factor
Ap : Surface area
hp : Thickness of a layer of the piezoelectric stack.
𝒇 : Time domain frequency
𝒏 : Spatial frequency
𝒏𝟎 : Spatial reference frequency
t : Time
s : Complex variable (frequency domain/s-domain)
𝑮(𝒏𝟎 ) : Road roughness coefficient [m2/m-1]
xxiii
Abbreviation Explanation
xxiv
xxv
ÖZET
DOKTORA TEZİ
İstanbul Üniversitesi-Cerrahpasa
xxvi
SUMMARY
Ph.D. THESIS
Istanbul University-Cerrahpasa
The suspension system is one of the key components of a vehicle that contributes to providing
ride quality and comfort by absorbing shocks due to the roughness of the roads. It consists of
parts that undergo significant energy dissipation underlining its energy harvesting potential.
The combination of parameters such as elastic suspensions, stiff dampers, and rough road
profiles can supply a huge amount of Energy. The main purpose of this research work is to
develop a mathematical model which will allow investigating the potential of dump trucks’
suspension energy harvesting mechanisms. For the sake of conserving the ride quality and
comfort, a series arrangement is used for the connection between the piezoelectric stack and the
suspension spring. The quarter and half truck models with and without incorporated
piezoelectric stack were mathematically modeled and their motion equations were derived. The
Laplace transforms of these equations were obtained in order to identify their transfer function
for analysis in both the time and frequency domains. Simulations were carried out by utilizing
the MATLAB Simulink software. Three different types of inputs were used for the simulations.
They are the step input, the harmonic, and random excitation. The results obtained from this
study revealed that there is great potential in piezoelectric energy harvesting from off-highway
truck suspension systems.
xxvii
1
1. INTRODUCTION
1.1.SIGNIFICANCE
Energy needs have increased dramatically over the last centuries owing to the rapid growth of
the industrial and automotive sectors. This situation has caused the multiplication of energy
crises over time. The world’s current energy system relies essentially on fossil fuels which
pollute the environment and whose reserves are limited. Therefore, a growing interest in the
recovery of energy lost because of vibrations was observed. Indeed, scientists from several
disciplines have been researching sustainable techniques to harvest energy from ambient and
aeroelastic vibrations for powering electronics. These new techniques enabled numerous
applications such as wireless sensors for equipment remote condition monitoring, health
monitoring, lighting, and so on [1].
The automotive and heavy equipment industry is one of the most important economic sectors
in the world. This industry has been experiencing a boom for the past decades resulting in an
increase in air pollution through vehicle emissions which have a bad impact on the environment
and public health. In order to remain competitive and keep increasing their sales, there is an
increasing need for manufacturers to control vehicle pollution by exploring innovative methods
to use clean energy. Various types of renewable energy have been used for this end [2].
According to Williams and Yates [3], there are three main types of vibrations to electric energy
conversion systems which are piezoelectric, electromagnetic, and electrostatic transductions.
As a result of their simplicity, wide frequency, and massive power density, researchers’ interest
in piezoelectricity has been growing over the years. Voigt [4] observed that numerous natural
crystals such as Rochelle salt, and quartz display a piezoelectric effect. Harvesting energy from
ambient vibrations is one of the cleanest and most eco-friendly techniques to acquire this scarce
resource.
Approximately 14% of fossil fuel energy is used to move the vehicle as reported by the US
Energy Department. The rest of the energy is mainly lost through vibrations and heat. Several
investigations were conducted in assessing available waste energy harvesting technologies [5,
6]. Piezoelectric energy harvesting has great potential for the automotive industry by reducing
2
greenhouse gases emission and boosting vehicles’ energy efficiency [7]. Patel [8] revealed that
piezoelectric materials can generate electricity and decrease the fuel consumption of vehicles
by at least 5 %. The main sources for energy harvesting on a vehicle are heat losses, vibrations,
and braking energy [9]. Nowadays, trucks have numerous wireless CM sensors which are
currently powered by disposable and polluting batteries. Piezoelectric energy could be
harvested from the truck suspension system and used to charge rechargeable batteries in order
to power wireless sensors such as TPMS sensors, vehicle speed, and strain monitoring sensors
[10].
All means of transport which consume petroleum products cause environmental pollution
during burning and refueling cycles [11]. In 2019, a report from the Energy Information
Administration (EIA) [12] revealed that 91% of the total energy demands of the US
transportation sector were met by petroleum-based fuels. However, only a small amount of this
energy is utilized by the vehicle to move [5, 6]. The suspension system is a key component of
trucks as they support the weight and provide ride quality and comfort by absorbing road
inconsistencies [13]. Despite this importance, it is one of the main sources of energy loss
through vibration dampening. According to Velinsky and White [14], the amount of energy
dissipated depends on several parameters such as the vehicle’s speed, mass, spring stiffness,
damping coefficients, and road profile.
Indeed, the increase of the mass of the vehicle combined with an augmentation of the damping
and stiffness coefficient leads to higher energy dissipation in the suspension system as observed
in reference [15]. Moreover, the road condition influences the quantity of energy that vanishes
through the dampers. The road profile can be analyzed both macrostructurally (longitudinal
road profile) and microstructurally (transversal road profile) [16]. The former is designated at
a vehicle speed between 25 – 120 km/h while the latter represents the shocks sensed by the
passengers due to the bumps. It was noticed that the microstructural profile has the most
important effect on the VSS. Wei and Taghavifa in their works [17] highlighted 04 categories
of road surfaces in the random excitation configuration. This classification was made based on
the power spectral density (PSD) function. These road categories are rough runway, smooth
runway, smooth highway, and gravel highway. A significant difference in power dissipation
3
was observed between the smooth and gravel highway configurations (respectively 200W and
2000W). This huge gap can be explained by the variation of the suspension system velocity for
the smooth (0.75m/s) and gravel (2m/s) roads. According to the same research work, the
suspension system of a car that is subjected to harmonic excitation experiences notably lower
power dissipation. Indeed, an average power loss of 57.84 W was reported at the speed of 13
km/h and 39 km/h. Abdel Kareem et al. [18] revealed that at a speed of 20 -50 km/h, the power
dissipated by conventional dampers on a flat road is in a range of 10 – 90W against 40 – 140W
for a road with no less than one bump. The findings of the aforementioned works stipulate that
the energy wasted through mechanical vibrations is non-negligible. Consequently, the main
goal of this work is to assess the potential VSS energy harvesting with the aim of turning
vibrations into useful electrical energy.
First and foremost, the main purpose of this study is to construct analytical (quarter and half)
truck models in order to assess energy harvesting mechanisms for the VSS. The second
objective of this work is to investigate the influence of vehicle, road, and piezoelectric
parameter variations on the harvested power and potential difference. Finally, this study aims
to draw a comparison between the models.
To achieve the foregoing objectives, the VSS springs and piezoelectric stacks will be connected
in series for dissipated energy harvesting. The equation of motion will be derived for each
model. In order to perform time domain and frequency domain analyses, the Laplace transform
of the motion equations will be found as well. The energy harvesting system of the VSS will
be represented for each model using Simulink. An analysis of physical quantities of the
harvested energy (power and potential difference) will be carried out in frequency and time
domains. Then a study will be conducted on the car, road, and piezoelectric parameters in order
to observe how the change of variables affects the output power and voltage. Simulations of the
models will be performed under harmonic and random excitation conditions.
4
This work consists of 6 main sections; it starts with an introduction which includes the problem
statement, and research objectives. The second chapter is an introduction to piezoelectric
materials and their parameters. It is also a state-of-the-art review of the available piezoelectric
energy harvesting methods in general and VSS-based piezoelectric energy harvesting
mechanisms in particular. In the third section, the equations of motion for the half and quarter
truck models were derived. These derivations were made for both VSS-only and VSS +
piezoelectric stack configurations. The Laplace transforms of the motion equations were found
in order to carry out the simulations in the MATLAB SIMULINK software. The time domain
and frequency domain analysis results are presented in the fourth chapter. In the fifth section,
the findings of the study and the effects of road and truck parameters are discussed. Last but
not least conclusions are drawn and recommendations are made for future works.
5
2. LITERATURE REVIEW
2.1. BACKGROUND
Numerous energy harvesting techniques have been explored over the past decades in order to
significantly decrease the use of polluting and scarce petroleum-based products as illustrated in
these references [19-26]. Yildiz [19], Cao and Li [26] reported in their works that the energy
harvesting method which furnishes the highest power density is solar energy harvesting in direct
sunlight (100 mW/cm3). Its power density in indoor conditions is 100 μW/cm3. Nevertheless,
the piezoelectric energy harvesting technique remains the most satisfactory one because of its
versatility of use, unlike solar power which suits only outdoor use.
Harvesting energy from the ambient environment is an idea that has received increasing
attention over the past years. Several kinds of mechanical vibrations like fluid flow-induced
vibrations, objects in motion, and vibrating structures exist in nature. A remarkable amount of
energy is wasted through vibrations. This dissipation can occur in several forms such as friction
and heat loss. There are three main available methods to convert this dissipated energy into
useful electrical energy. They are piezoelectric, electromagnetic, and electrostatic harvesting
mechanisms. Piezoelectricity has attracted the most researchers due to its significant power
density, wide frequency range, and its use of the implementation in energy harvesting projects.
Moreover, the availability and ease of fabrication of piezoelectric materials are one of the
reasons that can explain this growing interest.
Table 2.1: Comparison of energy harvesting methods and power densities [19, 26, 33-35].
Energy source Harvested power density
Solar 100 mW/cm3 (outdoor)
100 μW/cm3 (indoor)
Vibrations 37 mW/cm3 (piezoelectric transduction)
4.3 mW/cm3 (electromagnetic transduction)
0.584 mW/cm3 (electrostatic transduction)
Wind 177 mW/cm2
Thermoelectric 60 W/cm2
Temperature gradient 15 μW/cm3 at 10° C gradient
Ambient Radio Frequency 1 μW/cm2
Acoustic noise 0.003 μW/cm3 at 75 dB
0.96 μW/cm3 at 100 dB
Gonzalez et al. mentioned in their work [36] that piezoelectric transduction provides a wider
frequency range than electromagnetic-based generators. Indeed, piezoelectric materials have
the largest power density vs voltage in comparison with the other energy harvesting techniques
as represented in Figure 2.1. Furthermore, the piezoelectric energy harvesting method offers a
higher voltage output which necessitates fewer post-processing stages to be turned into useful
voltage than that of the other techniques. In fact, piezoelectric transduction generates a
significant amount of energy that can be used directly and without any supplemental voltage
injection as opposed to electrostatic transduction which requires an external polarization source
(generally variable capacitors) to initiate the energy conversion operation.
Figure 2.1: Power density vs voltage comparison plot of different energy harvesting techniques [37].
Aljadiri et al. mentioned in their research work [38] the benefits and limitations of the different
vibration-based energy harvesting techniques. Electromagnetic transducers have the ability to
function without any extra voltage input and are not subjected to any mechanical constraints.
Furthermore, they can supply high output current. In contrast, they offer a low voltage output
and display poor performances when they are used in microstructural applications. In addition
to that, they are difficult to integrate with Micro-Electro-Mechanical Systems (MEMS)
fabrication procedures. In comparison, electrostatic transducers can be easily integrated with
MEMS fabrication procedures and can provide a high voltage output. Nevertheless, they display
drawbacks such as the presence of mechanical constraints, the need for external voltage input,
and their low output current. It can be noticed in Table 2.2 that the piezoelectric generators offer
the most advantages such as their simplicity, ability to work without any external voltage input,
wide frequency range, and ease of use in diverse applications.
Energy harvested through piezoelectric transduction can be used in various applications such
as wireless sensors, lighting and battery charging, and so on [1]. Thanks to their decreasing
costs, their increased sensitivity and the advances in radio frequency technology, the use of
sensors in agriculture has been popularized in the last few years [39]. For instance, sensors can
8
Chen et al. performed research work [42] on the integration of control systems with wireless
sensors and actuators networks. An innovative, collaborative, and distributed estimation
approach was proposed. The results indicated that the control objectives were achieved. Christin
et al. conducted a survey [43] on the challenges faced by the acceptance of wireless sensors
used in industrial applications. The outcomes of the evaluation enabled the identification and
discussion of open issues. The popularization of sensors also provided the medical sector with
a revolution in the domain of remote patient monitoring.
Ko et al. reviewed in their work [44] the main applications of wireless sensor networks in the
healthcare sector. Key challenges related to the privacy and security of personal data were
discussed. They also highlighted the fact that these challenges are worsened by the lack of
resources. Prototype systems for physiological and behavioral monitoring were defined as well
as current research works challenges. Furthermore, the use of sensors has experienced a boom
in military and civil applications.
García-Hernández et al. [45] carried out a survey on wireless sensor networks and their
applications. Their low cost and flexibility allow us to find out innovative and new application
areas for remote sensing. Bridges stayed cables are vital for the structural integrity of bridges
9
and their tension was monitored using polyvinylidene difluoride (PVDF) films as sensors as
reported in this reference [46]. Chalard et al. in their research work [47] investigated the
challenges and encouraging applications of wireless sensor devices. Figure 2.2. illustrates
piezoelectric energy harvesting and its main applications.
Figure 2.2: Piezoelectric energy harvesting and its main applications [48-57].
Materials have played a very important role in human history and civilizations [58]. This is the
reason why the key periods of time in history were named after the dominant materials of these
eras. For instance, the Stone Age is a prehistoric era characterized by the widespread use of
stone to make edged weapons and tools. The Bronze and Iron Ages are also eras marked by
extensive use of bronze and iron, respectively. Our current era is known as the Synthetic
Material Age and is the period during which plastics, alloys, composites, and other materials
with exceptional properties and performances were made. The Smart Materials Age can also be
distinguished by the advances in engineering disciplines such as biotechnology and
nanotechnology. Smart materials play the role of both sensors and actuators. They are fully
capable of sensing, actuating, and responding to parameter variations in their environment.
Several types of materials like optical fibers ceramics, polymers, etc. possess these properties.
10
The ever-decreasing size of mini electronics has enhanced research works on new techniques
to develop autonomous devices which need no external energy input to operate [37]. The
piezoelectric energy harvesting method has received great attention because it is the most
suitable and profitable vibration-based mechanical to electrical energy harvesting technique.
This can be explained by its huge power density and ease of implementation in diverse
applications.
This chapter starts with a review of the piezoelectricity history which is followed by an
introduction to piezoelectric characteristics. Finally, the terminology and linear piezoelectric
mathematical theory were presented.
The piezoelectricity term is derived from the Greek expression “piezein” which means to press.
The piezoelectricity notion was discovered by Jacques and Pierre Currie in 1880 through
experimental research on the measurement of surface charges of single crystals [60]. It was
found that some natural materials such as quartz, tourmaline, and sodium-potassium tartrate
display piezoelectric characteristics [61]. Indeed, humankind has unconsciously used
piezoelectricity for years (the flint, a type of quartz was used to produce fire in the Stone Age).
In addition to that, the flint displays a very good cutting property and it was used to fabricate
tools and edged weapons. In fact, the “lighter” property of the flint was accidentally found
during the tool and weapon manufacturing process [59]. It is the asymmetry of the crystals that
allows them to exhibit this effect. When compressive or tensile stress is applied to the crystals,
11
a dissociation of positive and negative charges occurs and this leads to a polarization of the
crystals [62]. These materials have the ability to generate charges when they are subjected to
mechanical deformation. The intensity and direction of the output current depending on the
direction and magnitude of the mechanical stress or strain. When no stress is applied to these
materials, crystals have neutral charges owing to negative and positive charge polarization. In
this case, crystals yield dipole moment. However, stress application on these materials
There are two piezoelectric effects which are direct and converse piezoelectric effects. The
direct effect is when pressure produces electricity through piezoelectric materials; mechanical
energy is converted into electrical energy. This is the effect utilized by generators and sensors.
In case potential difference causes width and length variations of the materials, it is the converse
piezoelectric effect; electrical energy is converted into mechanical energy. This is the
configuration of piezoelectric actuators. The converse piezoelectric effect was mathematically
deduced by Lippman in 1881 using the fundamental principles of thermodynamics.
One of the first applications of piezoelectric materials was in the military sector in the 20th
century. During World War I, Langevin and his research team developed a composite ultrasonic
transducer for submarine detection in 1917. It consisted of an assembly of quartz crystals
located between two steel plates. The set was mounted in a waterproof casing. They were able
to reach a resonant frequency of 50kHz. They were able to reach their goal of measuring the
depth by timing the return echo of a previously emitted high-frequency sound after the end of
the war. This invention represents a turning point in the development of ultrasonic transducers
and systems. The discovery of polycrystalline piezoelectric materials allowed to improve the
performances of these devices.
During the Second World War, superpowers such as Japan, the United States of America, and
the Union of Soviet and Socialist Republics funded research works on the improvement of
actuator materials. It was found that some artificial polycrystalline materials (ceramic
ferroelectrics) obtained through the sintering of metallic oxide powders displayed dielectric
constants 100 times greater than that of common cut crystals like the barium titanate. However,
the random distribution of the crystals did not allow them to exhibit a macroscopic piezoelectric
effect. In 1946, Cady [63] stated that this kind of polycrystalline materials could acquire
macroscopic piezoelectric materials through poling which is the exposure to an electric field.
12
The discovery of the lead zirconate titanate (PZT) was made in 1954 by Jaffe et al. [64]. This
material displays improved piezoelectric characteristics. One of the most important advantages
of the PZT is the morphotropic phase boundary (MPB) which characterizes its phase diagram.
This characteristic of the PZT is the one that allows it to be used in several applications. New
additives have been mixed with the Basic PZT in order to improve its properties.
In 1969, the poling of uniaxially drawn polyvinylidene difluoride (PVDF) allowed it to acquire
impressive piezoelectric and pyroelectric properties. It was also discovered that although other
plastic materials like polyvinyl chloride (PVC) and Nylon exhibit piezoelectric effects, they do
not perform as well as the PVDF.
(converse effect). This will result in the increase and the decrease of the length of the crystal
[66].
Various natural and synthetic single crystal materials possess piezoelectric properties and they
are characterized by a crystalline structure. The crystals are connected through ionic bonds.
The quartz is unique in that no polarization remains in its structure after the cancellation of
the external mechanical stress/strain or electric field. Unlike quartz, other single-crystal
piezoelectric materials such as barium titanate, Rochelle salt, topaz, tourmaline, etc. exhibit
some remnant polarization after the removal of the external mechanical or electric field.
Hysteresis is when the current state of a system depends on its history and previous treatments
and states it went through. These single-crystal materials are subjected to a hysteresis field
when the remaining polarization is reversed using an inverse electric field. In this case, they
become ferroelectric.
There is a number of uniform polarization regions in ferroelectric crystals. These regions are
the ferroelectrics or Weiss domains. The particularity of a domain is that all its electric dipoles
point in the same direction. Numerous domains may exist in the same crystal and they are
separated by domain walls. The polarization vectors of domains are either at right angles (90°)
or antiparallel (180°) to one another, creating 90° and 180° domain walls, respectively (Figure
2.5). All ferroelectric domains in a crystal can be merged into a single domain through the
poling process (exposure to a high electric field).
15
Figure 2.5: Piezo ceramic crystals, domains, and domain walls [67].
Prior to poling, ferroelectric ceramics such as the PZT do not have any piezoelectric properties
due to the random orientations of the ferroelectric domains. Poling is an operation during which
a high electric field is applied in order to pole or force the Weiss domains to be oriented in the
same direction as the applied electric field. Throughout the polling process, the domains of
crystals are forced to point. However, the polarization vectors cannot perfectly parallel the field
direction unless the geometry of the crystal or grain is aligned in the same direction as the
electric field (as shown in Figure 2.6). The field causes some strain (∆S) in the crystals. After
the removal or cancellation of the electric field, there is some remnant polarization and strain
in the crystals (∆Sr). In order to reverse the polarization direction in the domains, a very strong
electric field can be applied. This process is called switching.
Ferroelectrics are much solicited in the sensing sector thanks to their piezoelectric properties
[69]. The Curie temperature is the point above which ferroelectric materials lose their
16
The term perovskite designates any material with a crystalline structure close to that of the
mineral called perovskite [70]. This mineral, which consists mainly of calcium titanate oxide
(CaTiO3), was first discovered by the German mineralogist Gustavus Rose in Russia in 1839.
Then the material was named after Perovski, a Russian mineralogist.
As represented in Figure 2.8 (a), PZT unit cell possesses a symmetric cubic structure above the
Curie point. While Figure 2.8 (b) illustrates the tetragonal cell unit structure of the PZT below
the Curie temperature with Pb2+, Ti4+ / Zr4+ moved relative to the O2- ions [71].
Figure 2.7: Perovskite structure (unit cell) of PZT above (a) and below (b) the curie temperature [71].
BaFe12O19 453
Ba (Ni, Zr)0.4 Fe11.6 O19 434
Ba (Co, Zr)1.2Fe10.8O19 374
Ba (Co, Zr)2.0Fe10.0O19 315
BaxSr1-xTiO3 (BST) Linear from 0 to 120
BaTiO3 125(variable according to the grain size)
α-Fe 770
ε’-Fe2.2C (hcp) 450
χ-Fe5C2 (monoclinic) 257
ϴ-Fe3C 212
Ga(Mn)As -100
KxNa1-xNbO3 (KNM) (perovskite) 410
Li2Ge7O15 (Lithium heptagermanate, LGO) 20
25% Mn-HfO2 300
Mn-ZrO2 390
NaNbO3 (Sodium Niobate) 63
NH4H2PO4 -125
PbZrO3 (Lead Zirconate, PZ) 233
PbTiO3 490
ZnTiO3 5.0
Zn(Zr0.01Ti0.99)O3 11.5
Zn(Zr0.05Ti0.95)O3 13.0
Zn(Zr0.10Ti0.90)O3 13.5
Zr1-xMnxO2-y From –400 to 600
The dielectric hysteresis phenomenon of ferroelectrics represents the lag between the action
of changing physical quantities such as the electric polarization (P), and the effect of this
change (in this case, the change in the polarizing electric field). In other words, the response
of the system does not depend only on the current intensity or magnitude of the external
change, but also on the system’s history. This phenomenon was firstly reported by Vasalek in
18
1921 in his research work [75]. He observed that this type of hysteresis is quite similar to the
magnetic one. The magnetization or magnetic polarization (M), the magnetic field intensity
(H), the magnetic field density (B) can be replaced by the polarization (P), the electric
field/intensity (E), and the dielectric displacement or electric flux density (D), respectively.
He found out that the electric displacement D in the direction of the crystal parallel to the x-
axis did not only depend on the present electric field (E) but also on its preceding values [76].
There is a linear relationship between the polarization (P) and the dielectric displacement or
electric flux density (D) is linear as shown in equation (2.1).
Di = Pi + ε0. Ei (2.1)
Where i represents any of the 3 axes (x, y, z); ε0 represents the permittivity of free space 8.854
x 10-12 F/m (Farad/meter).
The exposure of some initially unpolarized ferroelectric material like PZT to a high and growing
electric field at a temperature slightly below its curie temperature will lead to an increasing
alignment of the electric dipoles with the direction of the applied electric field. In this case, the
polarization curve (dielectric hysteresis loop) is similar to the loop in Figure 2.9.
In case the applied electric field is not high enough (OA), there is no crystal domain switching
which means that there is no change in the material behaviors; it will keep its normal dielectric
properties (paraelectric). In this case, a linear relationship exists between the polarization P and
the applied electric field E resulting in a straight line-like curve (OA).
As the applied electric field E increases (AB), several domains with negative polarizations will
start to be gradually switched into positive polarization (the direction of the applied electric
field). As a result, if the applied electric field E is increased, there is a sharp rise of the
polarization P within the AB segment.
If the polarization goes beyond the non-remanent polarization (BC section, above the
spontaneous polarization Ps), it is in the saturation polarization section. This means that all the
domains are aligned in a positive direction (the direction of the applied electric field). In this
saturation state, the crystal only consists of a single domain with a unique polarization direction.
20
This case is represented by the saturation polarization section of the dielectric hysteresis curve
represented in Figure 2.9 (BC).
As the applied electric field decreases until cancellation, the polarization (P) decreases too.
However, it stops at the remnant polarization point + Pr (point D); it does not return to zero.
The domains will remain aligned in a positive direction. In order to remove the remnant
polarization and bring the polarization P to 0, it is necessary to apply an electric field of opposite
direction (negative) and whose value reaches the coercive field strength (-Ec in this case). If
the electric field keeps increasing in the negative direction, the electric dipoles will start being
reversed until the cycle is completed (point F).
The relative deformation (butterfly) loop of ferroelectric materials is represented in Figure 2.10.
It depicts the variation of the strain according to the change of the applied electric field. This
loop illustrates the hysteresis effect which was exhibited in the case of the polarization P (Figure
2.9).
Figure 2.10: (a) Strain during poling (b) Strain hysteresis loop [77].
During the polling process of ferroelectric material (curve “a” in Figure 2.11), there are three
main types of strain:
The “virginal” strain (c) occurs when the applied electric field strength is low [77]. This
stage is characterized by the absence of strain as the applied electric field increases.
During this step, only 180° domain switching, which does not require any strain, is
performed as shown in Figure 2.12.
The piezoelectric effect strain (d) within the aligned domain is the second type of strain
that occurs during poling. This type of strain is quasi-linear unlike the third type of
strain.
In the third kind of strain (e), there is a quadratic relationship between the strain and the
applied electric field. At the end of this stage, there is an electrostriction phenomenon.
It is a property that all dielectric materials possess owing to the ion displacement of the
crystal lattice after exposure to an applied electric field.
22
Figure 2.11: The three kinds of strain during the poling process of ferroelectric ceramics.
It is possible to couple mechanical and electric fields in piezoelectric materials. With these
materials, some electrical energy input can be used to obtain some mechanical energy or vice
versa. Thus, an electromechanically coupling coefficient is necessary for this conversion. This
coefficient allows evaluating the effectiveness with which electrical energy can be converted
into mechanical energy or vice versa. It is usually determined at frequencies that are below the
resonant frequency of the piezoelectric material.
2 𝐶𝑜𝑛𝑣𝑒𝑟𝑡𝑒𝑑 𝐸𝑛𝑒𝑟𝑔𝑦
𝑘𝑒𝑓𝑓 = (2.2)
𝐼𝑛𝑝𝑢𝑡 𝐸𝑛𝑒𝑟𝑔𝑦
𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦
(𝑑𝑖𝑟𝑒𝑐𝑡 𝑝𝑖𝑒𝑧𝑜𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐 𝑒𝑓𝑓𝑒𝑐𝑡)
2 𝑀𝑒𝑐ℎ𝑎𝑛𝑖𝑐𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦
𝑘𝑒𝑓𝑓 = {𝑀𝑒𝑐ℎ𝑎𝑛𝑖𝑐𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦 (2.3)
(𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑒 𝑝𝑖𝑒𝑧𝑜𝑒𝑙𝑒𝑐𝑡𝑟𝑖𝑐 𝑒𝑓𝑓𝑒𝑐𝑡)
𝐸𝑙𝑒𝑐𝑡𝑟𝑖𝑐𝑎𝑙 𝐸𝑛𝑒𝑟𝑔𝑦
An analysis of the 𝐾𝑒𝑓𝑓 revealed that modern piezoelectric materials can convert up to 50% of
2
the received energy at low frequencies. The values of 𝑘𝑒𝑓𝑓 given in the tables are theoretical
and were obtained through vibration modes of ideal (unrealistic) specimens of the material. The
actual coupling coefficients of practical transducers are generally lower than those of the ideal
ones.
23
𝑘𝑝 is labeled as the planar coupling factor. The planar coupling of a thin disc is the coupling
between the electric field in direction 3 1 (parallel to the disc axis) and the radial mechanical
vibrations in directions 1 and 2. It is also known as radial coupling [58].
Moreover, the coupling factor 𝑘𝑒𝑓𝑓 can be expressed as a quotient of energy densities:
1 𝑇 2 1 𝑆 2
𝜀 𝐸 − 𝜀 𝐸
2 2
𝑘𝑒𝑓𝑓 = 1 𝑇 2 (2.4)
𝜀 𝐸
2
1
Where, 2 𝜀 𝑇 𝐸 2 represents the total stored energy density for the freely deforming piezoelectric
1
body (T=0, no constraint). 2 𝜀 𝑆 𝐸 2 represents the total stored energy density when the object
(piezoelectric body) is constrained (S=0). The numerator of the quotient is the difference
between these two densities and it represents the converted mechanical energy.
Sometimes, it is possible to extract the converted energy and to recover the unconverted energy
as well. The higher the coupling factor k is, the more efficient the transduction is. However, it
24
should be noted that 𝑘 2 should not be considered as a measure of the system’s efficiency
because it is the ratio of the useful converted power to the input power. In piezoelectric
materials, the unconverted energy is not necessarily lost. For instance, unconverted energy can
be stored in the piezoelectric material as heat and can be recovered through several techniques.
The real efficiency is the ratio of the useful energy to the energy dissipated by the transducer.
In general, a properly tuned transducer which was well adjusted and operating in its resonance
region can achieve efficiencies of over 90%. In case it is far from its resonance region, the
efficiency could be very low.
Figure 2.13: Relation between mechanical and electrical properties of piezoelectrical materials [78,
79].
As previously stated, ferroelectric materials exhibit a lag phenomenon between applied external
actions and the response of these materials to these external changes. For instance, a lag exists
between the variation in the polarizing electric field and the polarization. The characteristics of
ferroelectric materials vary depending on the history of the material. The dielectric hysteresis
loop of ferroelectrics, which represents the behaviors of these materials, is shown in Figure 2.9.
However, the use of ferroelectrics is limited to applications in the linear section of the loop
which can be approximately considered as the part of the curve between C and D (Figure 2.9).
This section starts with an introduction to the linear theory of piezoelectricity. This linear theory
describes the behaviors of piezoelectric materials in the linear range. Although the theory is
25
valid for all piezoelectric materials, it is very precise for non-ferroelectric materials such as
quartz. When it comes to ferroelectric materials, it is necessary to consider the limitations of
the piezoelectricity linear theory. It should be noted that the piezoelectricity linear theory is not
valid for non-linear characteristics such as depolarization, resonance, and hysteresis.
Standard values of piezoelectric material constants are utilized throughout this research work
[80]. According to the same standard, the (X, Y, Z) and (1, 2, 3) coordinate systems are used to
determine the elasto-piezo-dielectric coefficients of the piezoelectric materials. Among these
orthogonal axes, the polarization direction is represented by the Z-axis. (X, Y) is the plan which
is perpendicular to the polarization direction. The rotation around the orthogonal axes (X,Y,Z
or 1,2,3) of the coordinate system are labeled as 4,5 and 6. They are the mechanical shear stress
which is tangent to plans defined in the coordinate system as depicted in Figure 2.15. The main
properties of the material along X, Y, and Z axes are designated by using 1, 2, and 3 indexes,
respectively. For instance, in the case of the piezoelectric charge constant d31, the first index 3
means that the electrodes are perpendicular to the axis 3/Z while the second index 1 shows that
the stress or strain is applied along axis 1/X [71].
Figure 2.14: The direction index of electromechanical constants in piezoelectric materials [71].
26
The direct piezoelectric (sensor effect) observed in a single crystal can be depicted by a matrix.
This matrix represents the polarization that is developed by the crystal when an external
mechanical stress/strain is applied to the piezoelectric material. In case the applied external
electric field is equal to (E=0), the generated polarization is equal to the free charge q which is
present in the electrodes. This is a particular case of the behavior of the material in the direct
piezoelectric effect. The relation is given by equation 2.5.
Pi = qi = dij. Ti (2.5)
Where, Pi is the linear polarization generated at the surface which is normal to the i axis, qi is
the linear free charge developed at the surface which is normal to the i axis, dij is the
piezoelectric charge constant whose electrodes are perpendicular to the axis i and whose stress
is applied along axis j. The matrix expression of equation 2.5 is:
𝑇1
𝑇2
𝑃1 𝑞1 𝑑11 𝑑12 𝑑13 𝑑14 𝑑15 𝑑16
𝑇
{𝑃2 } = {𝑞2 } = [𝑑21 𝑑22 𝑑23 𝑑24 𝑑25 𝑑26 ] . 3 (2.6)
𝑞3 𝑇4
𝑃3 𝑑31 𝑑32 𝑑33 𝑑34 𝑑35 𝑑36
𝑇5
[𝑇6 ]
As depicted in equation 2.5 and equation 2.6, the polarization vector is equal to the free charge
in the electrodes because of the assumption that the applied electric field is equal to zero (E=0,
piezoelectric shorted). For each direction, the piezoelectric coefficient determines the intensity
of polarization. This relation is true for single crystals.
In the case of ferroelectrics, the polarization is replaced by the change in the spontaneous
polarization (Ps)i. Ps is considered only in the poling direction because it is negligible in the
transversal direction. By taking the linear range into account, the expression for ferroelectric
materials is:
The converse piezoelectric effect (actuator effect) designates the strain created in a piezoelectric
material when it is subjected to an external Electric field. In case the material is not subjected
27
to any mechanical constraint (free displacement; Tij = 0), the converse effect can be represented
by the relation:
Si = dijϴ, T. Ej (2.8)
Where Si is the linear strain induced at the surface which is normal to the i axis, dijϴ, T is the
charge piezoelectric coefficient which indicates the intensity of the deformation in the i-
direction when an electric field is applied in the j direction. The unit used for the charge
piezoelectric constant is m/v. The charge piezoelectric constant in the converse piezoelectric
effect is identical to those of the piezoelectric effect. The matrix expression of equation 2.8 is:
A linear dielectric can simultaneously undergo both mechanical stress and electric field in
specific conditions of temperature. It is possible to observe the electric and mechanical
properties of the materials in order to combine these results.
The mechanical strain S indicates the mechanical linear behaviors of a piezoelectric material
which is exposed to a mechanical stress T, an electric field E and a thermal variation ∆𝜃. This
relationship is also known as the Hook law approximation. The matrix expression of the strain
is:
28
Where, α is the thermal expansion coefficient. The expression of the thermal coefficient is:
𝜕𝑆 1
𝛼𝑖 𝐸,𝑇 = [ 𝜕𝜃𝑖 ] = [°𝐾] (2.11)
𝐸,𝑇=𝑐𝑜𝑛𝑠𝑡
When an electric field is applied to a piezoelectric material there is an electric response which
can be described by the linear polarization P generated in the crystals owing to electrical,
mechanical, and thermal deformation. The expression of the electric response is::
𝑇1
𝑑11 𝐸,ϴ
𝑑12 𝐸,ϴ
𝑑13 𝐸,ϴ
𝑑14𝐸,ϴ
𝑑15 𝐸,ϴ
𝑑16𝐸,ϴ 𝑇2
𝐷1
𝑇3
{𝐷2 } = [𝑑21 𝐸,ϴ 𝑑22 𝐸,ϴ 𝑑23 𝐸,ϴ 𝑑24 𝐸,ϴ 𝑑25 𝐸,ϴ 𝑑26 𝐸,ϴ ] . 𝑇 +
4
𝐷3 𝑑31 𝐸,ϴ 𝑑23 𝐸,ϴ 𝑑33 𝐸,ϴ 𝑑34 𝐸,ϴ 𝑑35 𝐸,ϴ 𝑑36 𝐸,ϴ 𝑇5
{𝑇6 }
𝜀11 𝑇,𝜃 𝜀12 𝑇,𝜃 𝜀13 𝑇,𝜃 𝐸1 𝑝1 𝑇,𝐸
[𝜀21 𝑇,𝜃 𝜀22 𝑇,𝜃 𝜀23 𝑇,𝜃 ] . {𝐸2 } + {𝑝2 𝑇,𝐸 } . ∆𝜃 (2.12)
𝜀31 𝑇,𝜃 𝜀32 𝑇,𝜃 𝜀33 𝑇,𝜃 𝐸3 𝑝3 𝑇,𝐸
𝜕𝐷 𝐶
𝑃𝑖𝑇,𝐸 = [ 𝜕𝜃𝑖 ] = [°𝐾𝑚2 ] (2.13)
𝑇,𝐸=𝑐𝑜𝑛𝑠𝑡
29
The piezoelectric coefficients dij are identical for both mechanical and electric responses.
Therefore, an electromechanical interaction occurs during the piezoelectric activity. So, it is
possible to combine both mechanical and electric response in a single expression, as shown in
equation 2.14.
Equation 2.14 is the elasto-piezo-dielectric matrix. It is also known as the d-form constitutive
equation. Its compact form is:
The d-form matrix is an extensive piezoelectric relation type. The independent variables in the
d-form are the matrixes [T] and [E].
The second constitutive equation for a piezoelectric material is the h-from matrix which is an
intensive piezoelectric relation type:
The independent variables in the h-form are the matrixes [S] and [D].
The third constitutive equation for a piezoelectric material is the g-form matrix which is a mixed
piezoelectric relation type:
30
The independent variables in the g form are the matrixes [T] and [D].
The fourth constitutive equation for a piezoelectric material is the e-form matrix which is a
mixed piezoelectric relation type:
The independent variables in the e-form are the matrixes [S] and [E].
The electric tensors [D] and [E] are vectors (first order tensors); The mechanical tensors [S]
and [T] are 3x3 matrices (second order tensors); The dielectric coefficients [𝜀] and [𝛽] are 3x3
matrices (second order tensors); The piezoelectric coefficients [d], [g], [e] and [h] are 6x3
matrices (third order tensors); The elastic coefficients [S] and [C] are 6x6 matrices (fourth order
tensors).
The thermal effect was not taken into account in the aforementioned constitutive equations. It
must be considered if the material is pyroelectric.
Piezoelectric coefficients
The piezoelectric coefficient dij is labeled as the piezoelectric strain coefficient. Two equivalent
expressions can be used to define it due to the fact that it remains identical for both direct and
converse piezoelectric effects.
𝜕𝑆 𝑚
= [ 𝑗] = (𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑒 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑎𝑐𝑡𝑢𝑎𝑡𝑜𝑟)
𝜕𝐸𝑖 𝐷,𝜃=𝑐𝑜𝑛𝑠𝑡 𝑣
𝜃
𝑑𝑖𝑗 = { 𝜕𝐷 𝐶
(2.19)
= [𝜕𝑇 𝑖 ] = 𝑁 (𝑑𝑖𝑟𝑒𝑐𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑠𝑒𝑛𝑠𝑜𝑟)
𝑗 𝑆,𝜃=𝑐𝑜𝑛𝑠𝑡
Piezoelectric materials are generally anisotropic and because of that their constants such as the
piezoelectric, permittivity and elasticity coefficients are tensors which relate to both the
direction of the applied stress/electric field … and to the directions perpendicular to them.
Therefore, the afore-mentioned coefficients come generally with two subscript indices which
indicate the direction of the two related quantities (for instance, the displacement and electric
field for permittivity, stress and strain for elasticity). The quantity which is kept constant is
indicated using a superscript index.
Piezoelectric coefficients (which generally refer to electric and mechanical quantities) are
accompanied by two (02) subscripts indices. The first subscript index shows the direction of
the electrical quantity (electric field or displacement) and the second subscript index represents
the mechanical quantity (stress or strain). The d-matrix of the three main types of piezoelectric
materials is given below:
0 0 0 0 𝑑15 0
𝑑𝑃𝑍𝑇 = [ 0 0 0 𝑑15 0 0] (2.21)
𝑑31 𝑑31 𝑑33 0 0 0
0 0 0 0 𝑑15 0
𝑑𝑃𝑉𝐷𝐹 = [ 0 0 0 𝑑24 0 0] (2.22)
𝑑31 𝑑32 𝑑33 0 0 0
𝜕𝑆 𝑚2
= [𝜕𝐷𝑗 ] = (𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑒 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑎𝑐𝑡𝑢𝑎𝑡𝑜𝑟)
𝑖 𝐸,𝜃=𝑐𝑜𝑛𝑠𝑡 𝐶
𝜃
𝑔𝑖𝑗 = 𝜕𝐸 𝑉.𝑚
(2.23)
= [𝜕𝑇 𝑖 ] = (𝐷𝑖𝑟𝑒𝑐𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑠𝑒𝑛𝑠𝑜𝑟)
{ 𝑗 𝑁
𝑆,𝜃=𝑐𝑜𝑛𝑠𝑡
𝜕𝐸 𝑉
= [𝜕𝑆 𝑖 ] = 𝑚 (𝐷𝑖𝑟𝑒𝑐𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑠𝑒𝑛𝑠𝑜𝑟)
𝑗 𝐷,𝜃=𝑐𝑜𝑛𝑠𝑡
𝜃
ℎ𝑖𝑗 ={ (2.24)
𝜕𝑇𝑗 𝑁
= [𝜕𝐷 ] = (𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑒 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑎𝑐𝑡𝑢𝑎𝑡𝑜𝑟)
𝑖 𝑆,𝜃=𝑐𝑜𝑛𝑠𝑡 𝐶
𝜕𝐷 𝐶
= [𝜕𝑆 𝑖 ] = 𝑚2 (𝐷𝑖𝑟𝑒𝑐𝑡 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑠𝑒𝑛𝑠𝑜𝑟)
𝑗 𝐸,𝜃=𝑐𝑜𝑛𝑠𝑡
𝜃
𝑒𝑖𝑗 = { 𝜕𝑇 (2.25)
𝑁⁄𝑚
= [𝜕𝐸𝑗 ] = (𝑐𝑜𝑛𝑣𝑒𝑟𝑠𝑒 𝑒𝑓𝑓𝑒𝑐𝑡 − 𝑎𝑐𝑡𝑢𝑎𝑡𝑜𝑟)
𝑖 𝑆,𝜃=𝑐𝑜𝑛𝑠𝑡 𝑉
Elastic coefficients
Two elastic coefficients are used to express the relation between mechanical stress and strain.
These coefficients are stiffness (C) and compliance (S).
The elastic stiffness C indicates the response of the crystals of piezoelectric material to applied
external stress or strain. In other words, it shows the extent to which the material can resist
33
strain in response to an externally applied force (stress) [81]. This quantity can be measured
under conditions of electric field constant or electric charge constant.
𝜕𝑇 𝑁
𝐶𝑖𝑗𝐸,𝜃 = [ 𝜕𝑆𝑗] = (𝑆𝑡𝑖𝑓𝑓𝑛𝑒𝑠𝑠 𝑎𝑡 𝐸 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑒 𝑜𝑟 𝑧𝑒𝑟𝑜 − 𝑠ℎ𝑜𝑟𝑡 𝑐𝑖𝑟𝑐𝑢𝑖𝑡)(2.26)
𝑖 𝐸,𝜃=𝑐𝑜𝑛𝑠𝑡 𝑚2
𝜕𝑇 𝑁
𝐶𝑖𝑗𝐷,𝜃 = [ 𝜕𝑆𝑗 ] = (𝑆𝑡𝑖𝑓𝑓𝑛𝑒𝑠𝑠 𝑎𝑡 𝐷 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑒 𝑜𝑟 𝑧𝑒𝑟𝑜 − 𝑜𝑝𝑒𝑛 𝑐𝑖𝑟𝑐𝑢𝑖𝑡) (2.27)
𝑖 𝐷,𝜃=𝑐𝑜𝑛𝑠𝑡 𝑚2
Where, the first subscript i indicates the direction of the strain and the second one indicates the
𝐷
direction of the stress. For instance, 𝐶32 is the stiffness for normal stress about axis 2 with a strain in
direction 3 under conditions of electric charge constant.
The elastic compliance S of a material represents the amount of strain generated per unit of
applied stress. The measurement of this quantity can be made at a constant electric field or
electric charge.
𝜕𝑆 𝑚2
𝑆𝑖𝑗𝐸,𝜃 = [𝜕𝑇𝑖 ] = (𝐶𝑜𝑚𝑝𝑙𝑖𝑎𝑛𝑐𝑒 𝑎𝑡 𝐸 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑒 𝑜𝑟 𝑧𝑒𝑟𝑜 − 𝑠ℎ𝑜𝑟𝑡 𝑐𝑖𝑟𝑐𝑢𝑖𝑡)
𝑗 𝑁
𝐸,𝜃=𝑐𝑜𝑛𝑠𝑡
(2.28)
𝜕𝑆 𝑚2
𝑆𝑖𝑗𝐷,𝜃 = [𝜕𝑇𝑖 ] = (𝐶𝑜𝑚𝑝𝑙𝑖𝑎𝑛𝑐𝑒 𝑎𝑡 𝐷 = 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑒 𝑜𝑟 𝑧𝑒𝑟𝑜 − 𝑜𝑝𝑒𝑛 𝑐𝑖𝑟𝑐𝑢𝑖𝑡)
𝑗 𝑁
𝐸,𝜃=𝑐𝑜𝑛𝑠𝑡
(2.29)
Where the first subscript i indicates the direction of the strain and the second one indicates the
𝐸
direction of the stress. For example, 𝑆12 is the compliance for normal stress about axis 2 with a strain
in direction 1 under conditions of electric field constant.
Dielectric coefficients
The dielectric constant represents the dielectric displacement per unit of the electric field. It is
also known as absolute permittivity. The first subscript index indicates the direction of the
dielectric displacement while the second subscript index gives information about the direction
34
of the electric field. The measurement of this physical quantity can be made at blocking force
(S=0) or at free displacement (T=0) as shown in equations 2.30 and 2.31.
𝜕𝐷 𝐹
𝜀𝑖𝑗𝑆,𝜃 = [𝜕𝐸 𝑖 ] = 𝑚 (𝑝𝑒𝑟𝑚𝑖𝑡𝑡𝑖𝑣𝑖𝑡𝑦 𝑎𝑡 𝐷 = 𝑐𝑜𝑛𝑠𝑡 𝑜𝑟 0; 𝑏𝑙𝑜𝑐𝑘𝑖𝑛𝑔 𝑓𝑜𝑟𝑐𝑒) (2.30)
𝑗 𝑆,𝜃=𝑐𝑜𝑛𝑠𝑡
𝑇,𝜃 𝜕𝐷 𝐹
𝜀𝑖𝑗 = [𝜕𝐸 𝑖 ] = 𝑚 (𝑝𝑒𝑟𝑚𝑖𝑡𝑡𝑖𝑣𝑖𝑡𝑦 𝑎𝑡 𝑇 = 𝑐𝑜𝑛𝑠𝑡 𝑜𝑟 0; 𝑓𝑟𝑒𝑒 𝑑𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡) (2.31)
𝑗 𝑇,𝜃=𝑐𝑜𝑛𝑠𝑡
𝜕𝐸 𝑚
𝛽𝑖𝑗𝑆,𝜃 = [𝜕𝐷𝑗 ] = (𝑖𝑚𝑝𝑒𝑟𝑚𝑖𝑡𝑡𝑖𝑣𝑖𝑡𝑦 𝑎𝑡 𝐷 = 𝑐𝑜𝑛𝑠𝑡 𝑜𝑟 0; 𝑏𝑙𝑜𝑐𝑘𝑖𝑛𝑔 𝑓𝑜𝑟𝑐𝑒) (2.32)
𝑖 𝑆,𝜃=𝑐𝑜𝑛𝑠𝑡 𝐹
𝜕𝐸 𝑚
𝛽𝑖𝑗𝑇,𝜃 = [𝜕𝐷𝑗 ] = (𝑖𝑚𝑝𝑒𝑟𝑚𝑖𝑡𝑡𝑖𝑣𝑖𝑡𝑦 𝑎𝑡 𝑇 = 𝑐𝑜𝑛𝑠𝑡 𝑜𝑟 0; 𝑓𝑟𝑒𝑒 𝑑𝑖𝑠𝑝𝑙𝑎𝑐𝑒𝑚𝑒𝑛𝑡)
𝑖 𝑇,𝜃=𝑐𝑜𝑛𝑠𝑡 𝐹
(2.33)
𝑉𝑝
𝜃𝑢̇ − 𝐶𝑝 𝑉𝑝 ̇ = ⁄
𝑅 (2.35)
Equation 2.35 can be used to find the voltage 𝑉𝑝 in terms of 𝑢. By substituting with a complex
harmonic excitation and response:
Where, 𝐹0 , 𝑉𝑝0 and 𝑢0 are complex amplitudes which can be expressed in the form of 𝑎 + 𝑖𝑏.
By substituting the aforementioned expressions in the electrical domain, 𝑉𝑝0 can be written
as:
𝜃𝜔𝑅(𝐶𝑝 𝜔𝑅+𝑖)
𝑉𝑝0 = 𝑢𝑜 (2.37)
(𝐶𝑝 𝜔𝑅)2 +1
By substituting the voltage in equation 2.37 for the mechanical domain. An assumption for
the displacement function is also made. Then, in order to group the real and imaginary terms,
the expression is rewritten:
2 (𝜃𝜔𝑅)2 𝐶𝑝 𝜃 2 𝜔𝑅 𝐹0
−𝑀𝜔 + 𝑘𝑠 + + 𝑖𝜔 (𝐶𝑠 + )= (2.38)
(𝐶𝑝 𝜔𝑅)2 +1 (𝐶𝑝 𝜔𝑅)2 +1 𝑢0
The mass, damping, and stiffness terms can be easily identified in equation (2.38). It is
important to note that the real piezoelectric term is dependent on the frequency. Therefore,
observations will be made for the two extreme particular cases 𝑅 = ∞; 𝑅 = 0.
36
For 𝑅 = ∞, the expression is turned into the open circuit equation. The stiffness term
2
becomes 𝜃 ⁄𝐶 while the damping term completely vanishes.
𝑝
For 𝑅 = 0, the expression is turned into the open circuit equation where all coupling terms
with 𝜃 vanishes. The remaining expression is known as the second-order equation for a mass-
spring-damper system.
𝜃2
𝑘𝑆 +
𝜔𝑛,𝑂𝐶 = √
𝐶𝑝 𝑘𝑆
; 𝜔𝑛,𝑆𝐶 = √ ; (2.39)
𝑀 𝑀
Where, 𝜔𝑛,𝑂𝐶 is the open circuit resonance frequency and 𝜔𝑛,𝑆𝐶 is the short circuit
resonance frequency.
𝜔𝑛,𝑂𝐶 𝜃2 2
= √ 1 + = √1 + 𝑘33 = √1 + 𝑘𝑒2 (2.40)
𝜔𝑛,𝑆𝐶 𝐶𝑝 𝑘𝑠
R is used as a variable in equation 2.35. During its variation between 0 and ∞ is studied by
𝑘 𝐶𝑠 𝜔
𝜔𝑛 = √ 𝑠, 𝜉 = ,Ω= , 𝑟 = 𝐶𝑝 𝜔𝑛 𝑅 (2.41)
𝑀 2𝑀𝜔𝑛 𝜔𝑛
Where 𝜔𝑛 is the short circuit natural frequency, 𝜉 is the dimensionless damping ratio, Ω is the
normalized excitation frequency and 𝑟 is the normalized resistance. Therefore, the normalized
equation of motion can be written as:
𝑘𝑒2 𝑟 2 Ω2 𝑘𝑒2 𝑟 𝐹0
(1 − Ω2 + ) + 𝑖Ω (2𝜉 + )= (2.42)
(𝑟Ω)2 +1 (𝑟Ω)2 +1 𝑢0
37
Firstly, only the real part of this equation will be studied. In order to evaluate the undamped
natural frequencies, the real part will be equated to 0. So, the solution of a fourth-order
equation in Ω with even powers must be found. Figure 2.17 illustrates the set of solutions on a
graph against 𝑘𝑒2 and log10 (𝑟).
Figure 2.16: Natural frequency ratio from the open circuit (R=∞) to short circuit (R=0).
In this case, the coupling coefficient 𝑘𝑒2 in the range between 0 and 0.7. Indeed, the highest
recorded value of the coupling coefficient for a 33 node- ceramic operation is 0.7.
It is also important to study the behavior of physical quantities such as the displacement,
voltage and power from high to low coupling, and at different excitation frequencies. The
combination of equations 2.37 and 2.38 allows drawing the expression of the voltage in the
function of the excitation. Then, the power can be obtained by multiplying the voltage by the
complex conjugate of the current [82].
Nonetheless, the average dissipated power in a resistive load is the voltage root mean square
(RMS) divided by the resistance for a sinusoidal voltage signal. From that, the following
expression can be drawn:
𝑉0 ̅̅̅
𝑉0
𝑃= (2.43)
2𝑅
38
It is also possible to find the expression of the normalized voltage, the normalized power, and
the normalized displacement.
𝑉0 𝑘𝑒2 𝑟Ω(rΩ+i)
𝑉′0 = = 𝑢′0 (2.44)
𝐹0 ⁄𝜃 (𝑟Ω)2 +1
𝑢0 1
𝑢′0 = = 𝑘 2 𝑟2 Ω 2 𝑘2
(2.46)
𝐹0 ⁄𝑘𝑠 𝑒𝑟 )
(1−Ω2 + 𝑒 2 )+𝑖Ω(2𝜉 +
(𝑟Ω) +1 (𝑟Ω)2 +1
The effects of the coupling on the mechanical domain are negligible in case the coupling is low
(𝑘𝑒2 < 0.01). This means that the harvested energy represents only a part of the one dissipated
through damping and that the electromechanical coupling itself is low (𝑘𝑒2 < 0.01) [83].
The normalized response plots of the displacement, voltage and power for these values of the
coupling factor 𝑘𝑒2 = 0.01 and of the dimensionless damping ratio 𝜉 = 0.05 are illustrated from
Figure 2.18 to Figure 2.20 [82].
39
Figure 2.17: The normalized response graph of the displacement for a coupling factor 𝑘2𝑒 = 0.01 and
a dimensionless damping ratio 𝜉 = 0.05.
Figure 2.18: The normalized response graph of the voltage for a coupling factor𝑘2𝑒 = 0.01 and a
dimensionless damping ratio 𝜉 = 0.05.
40
Figure 2.19: The normalized response graph of the power for a coupling factor 𝑘2𝑒 = 0.01 and a
dimensionless damping ratio 𝜉 = 0.05.
It can be observed in Figure 2.18, that there is a resonance peak at Ω = 1 in the mechanical
domain. Figure 2.19 indicates that for a low resistance (𝑟 ≪ 1) the voltage is null while for a
high resistance (𝑟 ≫ 1) there is a voltage resonance at Ω = 1 in the electrical domain. Finally,
a single power peak can be observed in Figure 2.20 at Ω = 1and at 𝑟 = 1 by using the derivative
of the power function to 𝑟 and Ω. The optimum resistance for a low coupling can be calculated:
1 1
𝑟𝑜𝑝𝑡 = ; 𝑅𝑜𝑝𝑡 = (2.47)
Ω 𝐶𝑝 𝜔
This kind of system can be solved sequentially owing to the fact that for a low coupling there
is no significant change in the dynamic stiffness and natural frequency of the system. Firstly,
the problem solving can start with a mechanical domain through the traditional method of
solving equation systems. It should be noted that the viscous damping is always considerably
higher than the harvested power. This requirement can be mathematically expressed as:
𝑘𝑒2 ⁄𝜉 ≪ 1 (2.48)
The solution of the displacement found by solving the equation system in the mechanical
domain can be used in the electrical domain to find the power. To this end, equation 2.44 to
equation 2.46 can be used.
The normalized response plots of the displacement, the voltage and the average power for a
high coupling are depicted in Figure 2.21, Figure 2.22, and Figure 2.23, respectively. The
coupling is considered to be high when 𝑘𝑒2 ≫ 0.01. It can be observed in Figure 2.21 that the
response of the displacement is purely mechanical for 𝑟 ≪ 1. Nevertheless, there is a significant
growth of the coupling as 𝑟 approaches 1. When 𝑟 goes beyond 1, there is a considerable shift
of the amplitude and the natural frequency. The behavior of the voltage in high coupling is quite
similar to that of the low coupling except for the fact that there is a shift of the voltage peak
towards the open circuit natural frequency. In the high coupling, there are two power peaks
41
instead of the single for the low coupling. This can be explained by the impedance matching
phenomenon [83, 84].
Figure 2.20: The normalized response graph of the displacement for a coupling factor 𝑘2𝑒 = 0.5 and
of a dimensionless damping ratio 𝜉 = 0.05.
42
Figure 2.21: The normalized response graph of the voltage for a coupling factor𝑘2𝑒 = 0.5 and of a
dimensionless damping ratio 𝜉 = 0.05.
Figure 2.22: The normalized response graph of the power for a coupling factor 𝑘2𝑒 = 0.5 and of a
dimensionless damping ratio 𝜉 = 0.05.
All objects in our environment possess a natural vibration frequency. When periodic series of
electrical, mechanical, and acoustic pulses are applied to objects with a frequency that is close
to the physical system’s natural frequency, an oscillation of these objects will occur with a very
high magnitude. Generally, the mechanical excitation of regular materials leads to mechanical
resonance. However, an applied mechanical field can generate electric resonance in
piezoelectric materials. Conversely, it is also very important to note that an electrical field may
generate a massive amount of mechanical strain in piezoelectric materials. Thus, a series of
periodic electric impulses with a frequency near the mechanical natural frequency of the
physical system will lead to resonance. The typical frequency response of a piezoelectric
material (¨PZT) is illustrated in Figure 2.24. A linear response area can be observed on this
43
curve. The frequency of resonance depends on the electrical and mechanical conditions of the
environment as well as the characteristics of the material.
2.2.5.2.The Electrostriction
The electrostriction is considered to be a property of dielectrics that causes their shape to change
when they are exposed to an electric field. This property of piezoelectric materials is explained
by the ion displacement in the crystal lattice when in presence of an electric field. The
displacement of positive ions is in the same direction as the applied electric field while the
displacement of negative ions is in the opposite direction. The accumulation of the displacement
leads to a strain on the material. The response of a piezoelectric component usually consists of
a linear and a quadratic component. The quadratic component is determined by the
electrostrictive coefficient.
2.2.5.3.The Depolarization
The poling process turns the PZT into a permanently polarized ceramic. However, it is
important to take steps in order to prevent the material from being depolarized as this could
result in the partial or total loss of piezoelectric properties. 3 main sources of depolarization of
piezoelectric materials exist: thermal, electrical, and mechanical depolarization.
44
Thermal depolarization: Heating a piezoelectric material to its curie point will create
a disorder of the domain orientations resulting in a complete depolarization of the
material. Thus, it is important to use piezoelectric materials in temperatures far below
their curie point. In general, the safe operating temperature range for a piezoelectric
material is [0, curie temperature/2].
Electrical depolarization: The application of a very strong electric field of opposite
polarity to the poling field causes the depolarization of the piezoelectric material.
Factors such as the material properties, the time of exposure to the field … determine
the necessary field strength to depolarize the piezoelectric material.
Mechanical depolarization: The application of very high mechanical stress may
disturb the orientation of the domains leading to a depolarization of the material. The
safe operating stress range varies with the material grade.
Nowadays, numerous natural and synthetic piezoelectric materials are available and used in
piezoelectric and pyroelectric devices and applications. Piezoelectric materials fabrication
processes have been integrated with microelectromechanical systems (MEMS) in both actuator
and sensor applications [85].
Single crystal materials are capable of exhibiting piezoelectric properties without further
processing steps owing to their natural asymmetric structure. There are two main types of single
crystals: Ferroelectric single crystals and non-ferroelectric single crystals.
Ferroelectric single crystals possess a remanent polarization without the presence of any applied
external stress. The polarization direction could be reversed through the application of a strong
electric field in the opposite direction. The main feature of these materials is the presence of a
spontaneous polarization without the application of an electric field. It should be noted that
45
The first material in which ferroelectricity was observed is Rochelle salt. The first theories of
the ferroelectric phenomenon were developed between 1930 and 1934 in the Soviet Union.
Ferroelectric properties of Rochelle salt can only be observed on mono crystal samples of this
material and along one definitive crystalline axis. In 1934, the Soviet physicist B.Vul made the
important discovery of barium titanate BaTiO3. Barium titanate provides several advantages
over formerly known other single crystal ferroelectrics such as the conservation of ferroelectric
properties within a broad range of temperatures, resistance to heat and moisture, high
mechanical strength, and ease of manufacturing. This material has been used worldwide i
several applications.
The main manufacturing step of barium titanate is the roasting of equimolecular quantities of
barium oxide (BaO) and titanium dioxide (TiO2). Later on, several ferroelectric materials
similar to the barium titanate such as lead zirconate, zinc titanates, strontium ad cadmium were
discovered.
Non-ferroelectric single-crystal materials exhibit zero polarization when the applied external
electric or mechanical field is canceled. Besides, it is not possible to switch the polarization
direction of the domains. The most widely used non-ferroelectric single crystal piezoelectric
material is quartz which is a polymorph of SiO2. Hence, it is not possible to switch the
polarization direction of quartz.
Quartz is required for high precision application owing to its particular properties. It should be
noted that quartz crystals are the most stable of all piezoelectric materials. Indeed, its
piezoelectric sensitivity is independent of mechanical stress, time, and temperature. For
example, resonators are applications that require high precision. Therefore, quartz is used in
these components.
Polymers such as polyvinylidene fluoride (PVDF) possess ferroelectric properties owing to the
special arrangement of the chain units in the crystalline phase. Thanks to its advantages such as
good response and high sensitivity, PVDF is used in several applications. The piezoelectric
properties of polymers such as polyvinylidene fluoride are fully developed when the material
undergoes both electric and mechanical stress. Indeed, the piezoelectric properties of the
material can be optimized by poling the polymer film. The main purpose of the polling process
is to align the randomly organized dipolar moments against the applied field. One of the main
47
advantages that the polling process provides to the polymers is the increase of the dielectric
constant due to the decrease of dielectric loss and the alignment of the main dipoles [87]. The
polyvinylidene fluoride crystallizes in α non-polar form like the cast from the melt.
Stretching the material up to 5 times its initial length will cause the polymer to switch to the β
configuration which is characterized by a planar zig-zag arrangement with all hydrogen atoms
on one side and fluorine atoms on the other side.
Poling the polymer in an electric field will lead to parallel orientation of the dipoles. The higher
the electric field in the poling process is, the better the electric output response of the PVDF
film will be.
Moreover, mechanically stretching and electrically poling the PVDF film in high-temperature
conditions causes a big increase in the piezo and pyroelectric constants. Corona poling is one
of the techniques used in optoelectronics. Corona poling is a non-contact poling process for
ferroelectric materials performed through corona discharges. These discharges can be described
as partial breakdowns of air and are initiated by a discharge in an inhomogeneous electric field.
This kind of poling process provides optimum results when performed at high temperatures
[88]. When the process is performed under optimum conditions, an improvement of 100% of
the piezoelectric and pyroelectric constants was obtained in comparison with the results
obtained under non-favorable conditions [89].
A corona poling apparatus can be observed in Figure 2.26. This device consists of a DC HV
(direct current high voltage) supply, polytetrafluoroethylene (PTFE) shields for the system’s
safety, a corona needle, and a sample back electrode through which the corona discharges are
generated, the piezoelectric material which is to be poled and a conductive plate.
48
Figure 2.26: Processing of a PVDF polymer film (a) application of mechanical stress, mechanical
stretching (b) contact poling (c) corona poling (d) electrospinning process [91].
2.3.3. Ceramics
Single crystal piezoelectric materials are often difficult and costly to fabricate. This is why
polycrystalline piezoelectric materials such as barium titanate and similar materials have been
49
preferred for several applications. They are labeled as ferroelectric ceramics and are fabricated
using conventional ceramic fabrication processes.
During the fabrication of a ferroelectric ceramic such as lead zirconate titanate (PZT), the
required amounts of zirconium, titanium, and oxides with few percent of additives such as
Sb2O5, CaO, Nb2O5, MnO are mechanically mixed. Then the mixture is dried and calcinated at
temperatures between 800° - 1000°C. Organic binders are added to the obtained stuff to produce
some slurry which will go through a drying process. After that, the ceramics are shaped
according to their application using pressing and firing techniques at temperatures ranging from
1100 to 1300°C. However, the random polarization of individual crystals does not allow these
ceramics to exhibit piezoelectric properties. For that purpose, the ceramics needs to go first
through a poling process. This poling process will allow the ceramic materials to remain with
a remnant polarization. It should be noted that transparent piezoelectric materials are produced
through the addition of lanthanum oxide to the PZT process and long hot-pressing operations
in order to obtain lead-lanthanum zirconate titanate (PLZT).
50
Batch weighing
Drying
Calcination (PZT
formation)
Spray drying
Piezoelectric materials are used in the several applications. One of these applications is the use
of piezoelectric materials in transformers.
A piezoelectric transformer can be defined as a piezoelectric body separated into two main
regions which are the driving and the driven region. The primary circuit of the transformer is
found in the driving part while the secondary circuit is found in the driven part of the
transformer. These regions are electrically isolated but mechanically linked. In the driving part
of the material, mechanical vibrations are generated from the input electric current. These
mechanical vibrations cause the excitation of the driven part of the material where it is
reconverted back into an electrical signal of different voltage, current, or impedance. To sum
up, a piezoelectric transformer’s working principle is characterized by a first piezoelectric
conversion of electrical energy into mechanical energy in the driving region and then a
reconversion of the mechanical energy into electrical energy in the driven region. Transformers
take advantage of the profitable properties of piezoelectric materials operating in resonance.
Because of that, these transformers are labeled as resonant piezoelectric transformers. Resonant
piezoelectric transformers are passive electrical energy transfer transducers that use the
52
piezoelectric properties of materials operated at one of their resonance frequencies. The purpose
of this process is to achieve voltage, current or impedance transformation. Thanks to their
simplicity, ease of use, and fabrication, this kind of transformer has replaced the traditional
electromagnetic transformers. Figure 2.29 illustrates the structure of a resonant piezoelectric
transformer.
Mechanical
Driving Driven
Electrical vibrations (output) Electrical
(input)
output
Load
input region region
Charles A. Rosen and Keith Fish in their patent [92] were the first scientists to introduce the
notion of piezoelectric transformers in 1954. This transformer was named after Rosen and is
known as a Rosen-type transformer. The Rosen-type transformer can be described as a thin
piezoelectric rectangular body separated into two regions that are transversely polarized with
respect to each other as depicted in Figure 2.30. The driving and the driven regions can be
manufactured separately and assembled together using some cement such as epoxy resins.
However, the obtained results are better if they are made from a single piece of piezoelectric
material. The polarization is longitudinal while the one in the driven region is longitudinal. The
polarization vectors in the two regions are at right angles with respect to one another.
This transverse strain causes the occurrence of a strain in the longitudinal dimension of the
body. This causes longitudinal vibrations in the driving region of the body. In case the body is
required to resonate with an anti-node at either end in the fundamental mode at a frequency
equal to the frequency of the applied potentials, a periodic potential appears between the
electrodes of the driven (output) region. Linking these electrodes will supply these periodic
potentials to the high impedance load device.
54
If the frequency of the driving alternating current is non-resonant, there are losses during the
vibration transmission. Because of that, the output voltage will be lower than the input voltage.
In contrast to the non-resonant mode, in case the driving alternating current to the piezoelectric
transformer has the same frequency as the mechanical frequency of the system, the vibration
generated in the driving region will be very high. This will result in high voltage production.
The particle displacement and the mechanical stress distribution in a resonant piezoelectric
transformer excited at the fundamental vibrational node are represented in Figure 2.31. The
selection of a high mechanical quality factor Qm, allowed to obtain a particle displacement that
is almost sinusoidal. In fundamental excitation mode, the displacement possesses a node exactly
in the middle of the resonant transformer [94].
Parameters such as the effective coupling factor keff, the mechanical quality factor Qm and the
dimensions of the piezoelectric body determine the unloaded voltage transformation ratio. This
transformation ratio can be defined as [95]:
55
The PXE43 is a hard lead zirconate titanate (PZT) ceramic which is generally used in power
applications. Indeed, this material’s power density is higher than 25Watts/cm3 without going
through any thermal issue. Other parameters of the PXE43 are given in Table 2.7.
A sinusoid or a sine wave signal can be defined as a continuous mathematical curve that
represents a smooth periodic oscillation. The most efficient way to drive piezoelectric
transformers is the use of a sine wave signal. Nevertheless, a square wave voltage is not enough
to generate a sine wave, more complex electronics are needed. Thus, it is necessary to use the
square wave voltage technique together with a technique to minimize switching loss.
This other technique is labeled Zero Volta Switching (ZVS). The mechanism which is required
to use this technique consists of a DC power supply. The supplied DC voltage is converted into
a square wave voltage which is transmitted to the resonant piezoelectric transformer. Depending
on the system requirements, a rectifier can be put at the output of the piezoelectric transformer.
The generated and/or rectified output voltage is furnished to the load. In comparison with
electromagnetic transformers, piezoelectric ones are very sensitive to load variations. These
load variations can cause resonant frequency shifts. Because of that, it is necessary to track the
56
frequency through a feedback connection from either the transformer or the load as depicted in
Figure 2.32.
Rosen and Fish proposed several models and configurations of transformers in their patent [92].
They proposed five main configurations of Rosen-type piezoelectric transformer.
Figure 2.33: Diagram of the narrow double poled ring piezoelectric transformer.
The thick double poled disc piezoelectric transformer:
Figure 2.34: Diagram of the thick double poled disc piezoelectric transformer.
The double poled disc piezoelectric transformer:
Figure 2.38: Influence of the connected load impedance on the transformation ratio.
The output impedance of a piezoelectric transformer can be described by the following
expression:
1 1
𝑍𝑜𝑢𝑡 = = 2 ).𝜀 .𝜀 . 𝑊.𝑡 .2𝜋.𝑓
(2.50)
𝐶𝑑2. 𝜔 (1−𝑘33 0 𝑟𝐿
𝑜𝑢𝑡
Where, 𝐶𝑑2. is the constrain capacitance of secondary side of piezoelectric transformer.
𝐿𝑜𝑢𝑡 is the interelectrode distance of the secondary side of the piezoelectric transformer
The matching load is the load where the resonant piezoelectric transformer output impedance
matches the load impedance. It is also the load that leads to a sudden increase in the output
characteristics. For optimal performance of the transformer, it is necessary that the transformer
output impedance be inferior to the connected load impedance. So, if it is not already the case,
the piezoelectric transformer output impedance must be reduced in order to meet this
requirement.
𝐶𝑒𝑛𝑡𝑒𝑟 𝑓𝑟𝑒𝑞𝑢𝑒𝑛𝑐𝑦
𝑄= (2.51)
𝐵𝑎𝑛𝑑𝑤𝑖𝑑𝑡ℎ
While designing a piezoelectric transformer, an important aspect that should be taken into
account is to ensure that the transformer is operated at its resonant frequency. A significant
difference between electromagnetic and piezoelectric transformers is that the Q of piezoelectric
transformers is far higher than that of the electromagnetic ones. The higher the Q is, the
narrower the bandwidth is. This means that it is easier to maintain the resonant frequency of an
electromagnetic transformer than that of a piezoelectric one.
In order to obtain the best possible performance, the transformer should be driven at or very
near its resonant frequency. Sometimes, even an error of ±0.5% can have very bad
consequences on the proper operation of the transformer. That is why it is very important to
track the frequency of the transformer through voltage feedback. Numerous factors can affect
the resonant frequency of a piezoelectric transformer during its operation. These factors can be
the output load variation, temperature, drive level …
61
Changing or switching the driving and driven electrodes of a Rosen-type transformer can allow
us to use them in measuring and protection systems. This switch will allow obtaining a very
low voltage output from a high voltage input through the transformer.
Nevertheless, although the voltage transformation for measuring and protecting systems can be
handled by resonant piezoelectric transformers, there are some aspects that limits the use of
Rosen-type transformers in measuring and protecting systems. The first aspect is that the main
frequency range of high voltage currents that are to be measured is between 50 and 60 Hz which
is too low to operate the piezoelectric transformers at their resonant frequency. Secondly, it
should be noted that the crucial factor during measurements is very high accuracy which cannot
be provided by piezoelectric transformers operated at their resonant frequency. Thirdly, it
should be noted that resonant piezoelectric transformers are narrow band transformers and can
be properly operated only at or near their resonant frequency. The above-mentioned facts limit
the use of resonant piezoelectric transformers in measuring systems.
Non-resonant High accuracy and Actuators and sensor Suitable for measuring
Piezoelectric linearity devices applications as they
transformers Wide frequency band provide very high
Low transformation accuracy.
ratios
Low energetic
efficiency
The interaction between diverse domains such as the mechanical and electrical ones illustrates
the complexity of piezoelectric energy harvesting. This complexity is emphasized by the non-
linearities of the circuitry used to process the harvested energy. Indeed, a piezoelectric energy
harvesting system consists of very distinct components whose effects will be discussed.
It is the key component that is directly involved in the mechanical-to-electric energy conversion
operation. Piezoelectric elements are used in two configurations which are the direct and
indirect or converse effects as shown in Figure 2.3 and Figure 2.4, respectively. In the direct
effect configuration, the piezoelectric element is fed with vibrations that originated from a
source such as wind, human motion, or a motor [96]. In the converse effect configuration, the
63
exact opposite happens: the piezoelectric element is fed with some electrical energy and
generates some vibrations.
It is the vibration source (direct effect) and the mechanical structure (converse effect) of the
whole system. These vibrations are generated from random excitations, pulses, and indirect
sources such as water or wind [97, 98].
Examples of host structures are shoes, the human body, the floor, roads, motors bridges, and so
on.
2.5.3. Electronics
Electronics represent the set of electrical circuitry present in a piezoelectric energy harvesting
system. Indeed, advanced electrical circuitry is required to maximize the harvested power.
Passive circuits are generally used in microscale piezoelectric harvesters while active circuits
which require some power to function are used in high output harvesters.
The main advantage of active circuitry over the passive one is that it artificially increases the
electromechanical coupling through the modification of the voltage of the piezo material. The
common characteristic of these circuits is that they possess a storage element (capacitor) that
buffers an. In addition to this storage element, more electronics are used to condition the voltage
for the sensor node. In this research work, the conditioning circuitry is represented by a resistor
called the electric load where the harvested energy is dissipated.
In the energy harvesting field, electronics play an important role. One of the advantages of using
advanced electronic circuits is the possibility to significantly increase the output power. Indeed,
the piezo voltage can be manipulated thanks to electronics. Good knowledge of electronics is
required to understand to fully understand the power harvesting notion. Concepts such as
current, voltage, linear components, and charge should be understood. Good knowledge of
semiconductor components and their functions in the circuit is very important.
64
The main variables in the mechanical domain are the velocity and sum of the forces applied to
a body. In the electrical domain, they are the voltage and the current [99]. Current can be defined
as the flow of electrons which are charge carriers. The electrons are charged negatively.
Conventionally, the electrons from the positive side to the negative side. However, the electrons
flow actually from the negative side to the positive side. The integral of current can be
considered as the number of accumulated electrons and it is known as the charge 𝑄.
𝑄 = ∫ 𝐼𝑑𝑡 2.52
The charge 𝑄 is expressed in coulomb [C]. This physical quantity represents in piezoelectricity
the charge displacement within the material.
𝜎 = 𝐶 𝐸 𝜀 − 𝑒 𝑇 𝐸𝑣 (2.53)
𝐷 = 𝑒𝜀 + 𝜖 𝜀 𝐸𝑣 (2.54)
By using the electrode area and then deriving to time, the expression of the current can be found.
So, the current equilibrium equation can be expressed as:
𝐼 = 𝜃𝑢̇ + 𝐶𝑃 𝑉̇ (2.55)
There are two types of current: the direct current (DC) and the alternating current (AC).
65
The direct current possesses only a single flow direction. The direct current is very common in
battery-powered systems and circuits. Analyzing a DC current-powered system allows solving
static mechanical problems (static solution).
The alternating current is a current that flows in both directions (from the negative to the
positive side and vice versa). It is characterized by amplitude, frequency, and phase. Alternating
current can be found in electrical outlets and in most home appliances. Analyzing an AC
current-powered system comes to solving a dynamic mechanical problem (which requires the
use of differential equations).
For a piezoelectric material, it can be stated that the application of a static displacement causes
the displacement of a fixed amount of charge. In order to obtain continuous power generation,
it is necessary to regularly cycle the displacement. This cycling will inevitably lead to the
generation of AC power. This power must generally be processed because numerous measuring
systems and electronic circuits are powered by DC current. In order to achieve this result, some
devices can be used.
𝐼1 − 𝐼2 − 𝐼3 = 0 (2.56)
The second law of Kirchhoff which is known as the voltage Kirchhoff law states that the sum
of all voltages around a closed network equals 0. Nevertheless, it is necessary to specify the
positive direction before applying the Kirchhoff voltage law.
𝑉(𝑡) − 𝑉1 + 𝑉2 − 𝑉3 = 0 (2.57)
Several similarities exist between the mechanical and the electrical domain. On one hand, in
the electric domain, Ohm’s law plays a very important role. It states that the flow of a current I
(in amps) through a resistor and the presence of a voltage V (in volts) leads to the following
relationship:
67
𝑅 = 𝑉𝐼 (2.58)
𝐹 = 𝑐𝑣 (2.59)
In addition to the resistor, two other components are frequently used in electronics. The first
one is capacitor C. The capacitance is expressed in Farad (F). A capacitor generally consists of
two parallel plates separated by a non-conductive medium. The symbol of a resistor is
represented in figure 2.43. The role of a capacitor is to accumulate charge as the voltage across
it increases. Its equivalent in the mechanical domain is compliance. A capacitor is governed by
the following equation:
68
𝑉 = ∫ 𝐼𝑑𝑡⁄𝐶 (2.60)
𝑉 = 𝐿𝐼 ̇ (2.61)
The equivalent of the inductance in the mechanical is the mass. The similarity is clear after
observing the second law of Newton:
𝐹 = 𝑚𝑣̇ (2.62)
69
Voltage V Force F
Current I Velocity 𝑢̇
Inductance L Mass M
Current I Force F
Voltage V Velocity 𝑢̇
Capacitance C Mass M
The similarities between the mechanical and electrical domains are even more striking during
the comparison between a mass-spring-damper system and an RLC circuit. An RLC circuit
consists of a resistor, an inductor, a capacitor, and a voltage source connected in series as
represented in Figure 2.45.
1
𝑉 = 𝐿𝐼 ̇ + 𝑅𝐼 + ∫ 𝐼 𝑑𝑡 (2.63)
𝐶
The application of the second law of Newton to the mass-spring-damper system allows writing
the following equation:
It is quite clear that equation 2.63 and equation 2.64 are quite similar. Indeed, they are both
second-order differential equations and can be solved analytically by assuming a harmonic
solution. The natural frequency in the mechanical domain is 𝜔𝑚 = √𝑘⁄𝑀 . In the electrical
domain, the natural frequency is 𝜔𝑒𝑙 = 1⁄√𝐿𝐶 .
Another type of frequently used circuit is the RC circuit which consists of a resistor, capacitor
and voltage source connected in series as represented in Figure 2.47.
𝑉𝑐
𝐶𝑉𝑐̇ + (2.65)
𝑅
In a case the circuit is excited by a step voltage of amplitude 𝑉0 at t = 0, equation 2.65 can be
solved using an exponential function. During the charging step (𝑉(𝑡) from 0 to 𝑉0), the
following asymptotic response function can be used:
During the Discharging step (𝑉(𝑡) from 𝑉0 to 0), following asymptotic response function can
be used:
The quantity RC can be considered as an indicator of the reactivity of the capacitor to a variation
of voltage. For instance, when the system is excited using a step voltage, at t = RC, the voltage
has reached 63% of the applied voltage during the charging step or has decreased 63% during
the discharging step. This time is labeled as the RC time of an RC circuit.
It is possible to express the electrical model of a piezo element using equation 2.55 which is the
current equilibrium equation( 𝐼 = 𝜃𝑢̇ + 𝐶𝑃 𝑉̇). Let’s consider the piezoelectric material as a
generator whose current equals 𝜃𝑢̇ . A capacitor is connected in parallel to this piezo generator.
The obtained circuit is illustrated in Figure: 2.48.
𝐼𝑃 (𝑡) is the current delivered by the piezoelectric element while 𝑉𝑃 (𝑡) represents the potential
difference between the two electrodes of the circuit. It should be noted that the current function
𝐼𝑃 depends on the nature of what is connected to the electrodes of the piezoelectric element.
The connection of a resistor R to the electrodes will lead to:
𝐼𝑃 = 𝑉𝑃 ⁄𝑅 (2.68)
More complex circuitry or other components can be connected to the piezoelectric element
resulting in a more complicated of the current 𝐼𝑃 .
73
Diodes can be defined as components that let the current pass only in a single direction. The
working principle of diodes is very simple. When the applied voltage is in the same direction
as the “through” direction of the diode, it causes the passage of current which is almost
unimpeded (no impedance/resistance). Nevertheless, there is a very small voltage drop 𝑉𝑑 . In
Current will be conducted through the diode only when the voltage is directed in the “through”
and when it goes beyond the small voltage drop 𝑉𝑑 . The voltage drop varies according to the
type and the manufacturing material of the diode. For instance, the voltage drop of a silicon
diode is 0.6V, while that of a germanium diode varies between 0.2-0.3V. The voltage drop of
Shottky type diodes is 0.2V.
In case the voltage is applied in a direction that is opposite to the “through” direction, the diode
blocks the passage of any current. Its perfect equivalent in this configuration would be a resistor
with infinite resistance. There is a negative voltage threshold below which the diode fails. The
schematic symbol of a diode is represented in Figure 2.49 [101]. As depicted in the symbol, the
current is only conducted in the direction indicated from the positive to negative contact.
74
𝑉𝑑
𝐼 = 𝐼𝑆 (𝑒 𝑛𝑉𝑇 − 1) (2.69)
𝐾𝑇
𝑉𝑇 = (2.70)
𝑞
𝑇 is the temperature
The field-effect transistor (FET) is a switch that utilizes the voltage to control its switching
characteristics. The working principle of a FET is explained by the creation of a conductive
channel. When there is a voltage field in the FET, a conductive channel opens, and the larger
the voltage field, the larger the conductive channel. In other words, the conductive channel is
like a variable resistor whose resistance is determined by the voltage field. The schematic
symbol of the variable resistor model of a field transistor is represented in Figure 2.53. A FET
generally consists of four (04) ports. Actually, two (02) of these ports are permanently
connected resulting in three (03) ports.
There are several types of field-effect transistors. Each type has its working principle, polarity
enhancement, and depletion mode. One of the most used and best-known n-channel metal-
oxide-semiconductor field-effect transistors (MOSFET). In this type of transistor, a positive
77
gate-source voltage 𝑉𝐺𝑆 opens the channel between the drain and the source and in case gate-
source voltage 𝑉𝐺𝑆 becomes equal to zero (0), the channel gets completely blocked.
The source (S), the gate (G), the body (B), and the drain are shown in Figure 2.52 and Figure
2.53. The body is not explicitly indicated but is the center to which the arrow points. It is also
connected to the source (S). There are four (04) main regions in metal-oxide-semiconductor
field-effect transistor (MOSFET): the “off” region, the sub-threshold region, the linear region,
and the saturation region.
The “off” is the region where the gate-source voltage is not sufficient, leading to an absence of
channel from drain to source. The subthreshold region is when 𝑉𝐺𝑆 (gate-source voltage) allows
only leakage of the current but is insufficient to allow a full passage of current. The linear region
is the zone where the current, the gate-source, and the drain-source voltage are linearly
proportional. The saturation region is where the maximum drain-source current is reached,
which means that increasing the drain-source voltage will not increase the drain-source current.
It should be noted that of these regions, the linear region is the most interesting one for coupling
purposes with active circuitry.
78
The bipolar junction transistors are used to switch signals and amplify power. They are capable
to use a small amount of control current to let a larger current pass. The ratio of these currents
is labeled as the common-emitter current gain. It is denoted ℎ𝐹𝐸 . A bipolar junction transistor
mainly consists 3 ports. These ports are the emitter (E), the collector (C) and the base (B) as
depicted from Figure 2.55 to Figure 2.57.
There are two main transistor types : the NPN type and the PNP type. The difference between
these two types of bipolar junction transistor is the direction of the current.
In the NPN configuration, a small current is passed from the base to the emitter in order to allow
the passage of a greater current to pass from the collector to the emitter. In addition to the
current requirement, similarly to the diodes, there is a small voltage drop. This voltage drop
consists of two (02) components which are the base-emitter voltage drop 𝑉𝐵𝐸 and the base-
collector voltage drop 𝑉𝐵𝐶 . 𝑉𝐵𝐶 is not very important because there is generally no current
between B and C. 𝑉𝐵𝐸 is the most relevant one as the main current flows between C and E.
79
The cut-off region is the region where the blocks the passage of any current. 𝐼𝐵𝐸 is so low in
this region that it does not allow any current passage from the collector to the emitter. The active
region is the region where 𝐼𝐶𝐸 ≈ ℎ𝐹𝐸 𝐼𝐵𝐸 . The saturation region is the one where 𝑉𝐵𝐸 > 𝑉𝐶𝐸 and
any current of up to ℎ𝐹𝐸 𝐼𝐵𝐸 can flow from the collector to the emitter. It is the region of interest
for energy harvesting. The collector-emitter has only a very minimal equivalent resistance. In
this region, the bipolar junction transistor acts nearly like a short circuit in which the current is
limited by the ratio of currents ℎ𝐹𝐸 and the other components. It should be noted that in this
region, the ratio of currents is determined by the other components. The reverse active region
is the one where the current goes from the base to the collector. This configuration is rarely
used in electronic applications.
80
The PNP BJT is operated similarly to the NPN one. The direction of the currents is the only
difference. A small emitter-base is needed to allow the passage of a larger emitter-collector
current. The notation of the currents is the same as for the NPN transistor but they all always
have a negative value. It should be noted that the PNP transistor consumes a small amount of
the main current, in order to allow most of it to pass unimpeded. The PNP BJT has also the
same operation regions as the NPN one (active, cut-off, saturation and reverse-active). As for
the NPN, the saturation region is theone that attracts our interest in piezoelectric power
harvesting. If 𝑉𝐵𝐸 < 𝑉𝐶𝐸 , the ratio of the currents is 𝐼𝐶𝐸 < ℎ𝐹𝐸 𝐼𝐵𝐸 . The diode model and the
schematic symbol of the PNP BJT are illustrated in Figure 2.55 and Figure 2.56, respectively.
A direct current impedance matching (DCIM) circuit generally consists of a rectifier, storage
capacitor 𝐶𝑆 and load resistor 𝑅, as depicted in Figure 2.59.
It is necessary that the storage capacitor 𝐶𝑆 be large enough in order to keep a quasi-constant
DC voltage. So, the RC time of the storage circuit must be higher than the cycle time of the
piezoelectric element. The voltage, current, and displacement curves of the circuit at the optimal
resistance are represented from Figure 2.60 to Figure 2.62.
82
Figure 2.60: Waveform of the current at the optimal resistance - 𝜃𝑢̇ is represented in full line while 𝐼𝑟
is represented in dotted line.
A sinusoidal current is obtained from the piezoelectric element because sinusoidal displacement
is applied, only that there is a 90 degrees phase difference.
Figure 2.61: Waveform of the voltage at the optimal resistance - 𝑉𝑑𝑐 is represented in full line while
𝑉𝑝 is represented in dotted lines.
Two main phases of operation exist. The first phase represents the situations in which the AC
voltage of the piezoelectric element oscillates between ±𝑉0 as it is depicted in Figure 2.62
83
where the piezo element voltage oscillates between the extrema. It can be seen in Figure 2.61
too, where the rectifier current is zero.
The second phase is called the conductive phase. This is when the circuit conducts at near-
constant voltage when the threshold voltage of the rectifier is reached. It should be noted that
the threshold voltage varies according to parameters such as the diode drops of the rectifier and
the DC voltage. For instance, in Figure 2.61, the conductive phase is on anytime the rectifier
current is different from 0. It should be noticed in Figure 2.62 that the DC voltage remains
almost constant due to the high value of the capacitor 𝐶𝑠 .
It is not possible to solve analytically the governing equation of a rectifier owing to its
discontinuous conduction. This means that an exact analytical solution cannot be found. Shu
and Lien in their research works [83, 102] analyzed the DCIM circuit in detail. By assuming
that we are dealing with a sinusoidal motion, it is possible to find an approximate analytical
solution through the one-half cycle approach. This can be done by using the energy balance and
current conservation notions. The non-dimensional parameters can be recalled:
𝑘 𝐶𝑠 𝜔
𝜔𝑛 = √ 𝑠, 𝜉 = ,Ω= , 𝑟 = 𝐶𝑝 𝜔𝑛 𝑅 (2.41)
𝑀 2𝑀𝜔𝑛 𝜔𝑛
By considering neglecting the diode voltage drop is negligible and by using the same non-
dimensional parameters in equation 2.41 which is repeated below, the equations of the
displacement, voltage and harvested power are:
𝑢0 1
𝑢′0 = = 1⁄2 (2.72)
𝐹0 ⁄𝑘𝑠 𝑘 2 𝑟Ω 2𝑘 2𝑟
𝑒
[(1−Ω2 +𝑟Ω+𝜋⁄2)2 +Ω2 (2𝜁+ 𝑒
2 )]
(𝑟Ω+𝜋⁄2)
𝑉0 𝑟Ω
𝑉′0 = = 𝑘 2 𝑢′ (2.73)
𝐹0 ⁄𝜃 𝑟Ω+𝜋⁄2 𝑒 0
The analysis is made only at t = 0 s and t = T/2 s, where T is the period of excitation. Generally,
the exact voltage and displacement signals do not affect the power output for low coupled
systems. Indeed, it should be noted that small deviations from a sinusoidal do not considerably
influence the output. After using the non-dimensional parameters from equation 2.41, it
becomes clear that these solutions (ACIM) are not mathematically much different from those
of equation 44 to equation 46. Only a few values are different in the two quotients. For instance,
1 is replaced by π/2 in the denominator of the piezoelectric stiffness term of 𝑢′0 . There is also
a difference in the denominator of the piezoelectric damping term.
The analysis of the governing equations allows illustrating the normalized displacement,
voltage, and power plots as represented in Figure 2.63, Figure 2.64, and Figure 2.65. As for the
ACIM circuit, in the low coupling, the piezoelectric material and electrical domain do not have
much influence on the mechanical domain. It can be explained by the fact that no power is
harvested in closed and open circuit conditions. Because of that, the results are identical. One
of the other similarities with the alternating current impedance matching ACIM is the presence
of an optimal load. Nonetheless, the value of the normalized resistance 𝑟 is different.
85
Figure 2.62: The normalized response graph of the displacement for a low coupling factor 𝑘2𝑒 = 0.01
and a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
Figure 2.63: The normalized response graph of the voltage for a low coupling factor 𝑘2𝑒 = 0.01 and a
dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
Figure 2.64: The normalized response graph of the power for a low coupling factor 𝑘2𝑒 = 0.01 and a
dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
86
𝜋
𝑅𝑜𝑝𝑡 = (dimensional optimal electric load) (2.75)
2𝐶𝑝 𝜔
𝜋
𝑟𝑜𝑝𝑡 = (non-dimensional optimal electric load) (2.76)
2Ω
This difference in the normalized resistance between DCIM and ACIM can be found in the
useful power harvesting region which is near the optimal resistance and the natural frequency
(Ω = 1). There is no apparent difference in the displacement response. The voltage response is
quite similar but slightly lower than that of ACIM circuits. This results in a small decrease of
power as well. At the same excitation, the difference between the maximum normalized power
of the DCIM and ACIM circuits means that the optimal resistance load is also different.
The normalized displacement, voltage and power are represented in Figure 2.66, Figure 2.67
and Figure 2.68.
87
Figure 2.65: The normalized response graph of the displacement for a high coupling factor 𝑘2𝑒 = 0.5
and of a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
In the displacement plot, there is a variation of the natural frequency. Indeed, the peak in the
frequency Ω shifts from 1 to √1 + 𝑘𝑒2 as the normalized resistance increases.
Figure 2.66: The normalized response graph of the voltage for a high coupling factor 𝑘2𝑒 = 0.5 and of
a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
88
Figure 2.67: The normalized response graph of the average power for a high coupling factor 𝑘2𝑒 = 0.5
and of a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
There are similarities with the ACIM case for the rest of the quantities. The voltage curve
displays a high value for open circuit resonance while the power possesses two optimal points
of equal magnitude. It should be noted that the optimal power is again lower than that of the
ACIM case. In high coupling conditions, the resistance gets more complex. There are several
cases. It is important to note that the number of optimal loads as well as their magnitude varies
according to the damping and coupling. It is also possible to determine whether the coupling of
a system is high or low through the rule of thumb [102].
On one hand, if the coupling of the circuit is low, the maximum harvested power can be
expressed as a function of the normalized excitation frequency::
It means that for a low 𝑘𝑒2 ⁄𝜁 ratio, any growth of the coupling will cause a power increase. This
can be explained by the that when the coupling is very low, the sequential solution of the
equations of motion can be used. Nonetheless, although a relatively small increase in coupling
will allow more energy to be harvested per cycle, the mechanical domain parameters will not
be affected. the global mechanical behaviors.
On the other hand, if the coupling of the circuit is high, the maximum power, at the optimal
resistive load and excitation frequency can be expressed as:
89
1 1
𝑃′𝑚𝑎𝑥 = (2.78)
8+2√16+𝜋2 𝜁
The maximum harvested power remains identical for the short circuit and open circuit
resonance peaks. It is important to note that in case the electromechanical coupling is too strong,
the displacement will be hindered too much, resulting in a limitation of the energy input from
the excitation.
The Synchronized Switch Harvesting on Inductor (SSHI) circuit is an active circuit that works
very similarly to the DCIM circuit. The only difference is that DCIM circuits change the piezo
voltage using mechanical deformation of the patch while the SSHI circuit does it electronically.
The SSHI circuit processes the voltage of the piezoelectric element through a switched
electrical oscillator. This operation requires a capacitor and inductor. For that end, a switched
inductor is added on the AC side, in parallel with the piezo capacitance, as depicted in figure
2.68. This allows to considerably increase the piezo voltage resulting in a growth of the power.
The second phase starts when the piezo voltage 𝑉𝑝 starts dropping. The voltage drop is detected
by the switch which closes, resulting in a connection of the inductor to the piezo element. This
set forms an oscillator. The piezo voltage oscillates between positive and negative at a
frequency of 𝜔𝑒𝑙 . It is stopped after only half a cycle due to the switch opening again.
In the third phase, the piezo element operates in open circuit mode. The voltage is lower than
before the inversion due to the finite resistance of the RLC loop. After the piezo element has
recovered this voltage loss, the opposite conductive phase begins.
The waveforms of the associated displacement, current and voltage are represented in Figure
2.69, Figure 2.70 and Figure 2.71, respectively.
Like for the DCIM circuit, it is not possible to find an analytical solution. Besides the rectifier
which is non-linear, the switching action of the circuit complicates more the solution. The
circuitry which is necessary to make the switch itself is too complex to solve analytically.
92
Nevertheless, an approximate solution can be found by ignoring the inner components of the
switch as long as the voltage inversion terms are taken into account [103]. Like in the DCIM
circuit, the solution considers energy balance, current balance and a half cycle of the motion. It
should be noted that the natural frequency of an electrical oscillator can be written as:
𝐿1 is the inductance.
Owing to the finite resistance of the coil, there is some damping which causes a decrease in the
amplitude of the oscillations similar to viscous damping in the mechanical domain. The symbol
of damping in the electrical domain is 𝑄𝑖 which is also labeled as a quality factor or inductor
quality and the higher the quality factor, the less the damping. In the mechanical domain, the
viscous damping symbol is 𝜁. The expression of the quality factor is:
The voltage after inversion 𝑉𝐴𝐼 can be expressed as a function of the one before inversion:
The normalized displacement, voltage and power are expressed in the following equations
[103]:
𝑢0 1
𝑢0′ = = 1 (2.82)
𝐹0 ⁄𝑘𝑠 2 2
1−𝑞𝐼 2 𝑟Ω
𝑘𝑒 𝑟Ω 2[1+ (1−𝑞𝐼 2 )]𝑘2𝑒𝑟
[(1−Ω2 + 2 +Ω2 (2𝜁+ 2𝜋
1−𝑞𝐼 ) 1−𝑞𝐼 )]
𝑟Ω+𝜋⁄2 ( 𝑟Ω+ 𝜋⁄2)2
2 2
𝑉0 𝑟Ω
𝑉0′ = = 1−𝑞𝐼 𝑘𝑒2 𝑢0′ (2.83)
𝐹0 ⁄𝜃 + 𝜋⁄2
2
93
The governing equations allow representing normalized normalized response graph of the
displacement, voltage and power as shown in Figure 2.73, Figure 2.74 and Figure 2.75,
respectively. It was assumed the inductor quality 𝑄𝑖 is equal to 10 (𝑄𝑖 = 10). In low coupling
conditions, the most remarkable difference in the displacement plot is the absence of an open
circuit resonance peak at r ≫ 1. The reason for that is that although no power is harvested, the
circuit dissipates the energy internally in the form of resistive losses in the inductor, resulting
in motion suppression in the mechanical domain. There is a smooth maximum on the voltage
plot. A big percentage of the energy extracted from the piezo element is lost during voltage
inversion if the resistance r ≫ 1 due to the finite inductor quality.
There is an optimal operating point in the normalized power response graph, at short circuit
resonance. It should be noted that the power output is considerably higher than that of the DCIM
circuit.
94
Figure 2.72: The normalized response graph of the displacement for a low coupling factor 𝑘2𝑒 = 0.01
and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit.
Figure 2.73: The normalized response graph of the voltage for a low coupling factor 𝑘2𝑒 = 0.01 and a
dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit
95
Figure 2.74: The normalized response graph of the power for a low coupling factor 𝑘2𝑒 = 0.01 and a
dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit.
In the high coupling conditions, there is considerable damping for all r ≫ 1, which means that
almost all energy is lost in the voltage inversion. In comparison to the DCIM open-circuit case,
the voltage amplitude is very small. This can be explained by the lack of an open circuit
resonance. There is also a single power optimum.
Figure 2.75: The normalized response graph of the displacement for a high coupling factor 𝑘2𝑒 = 0.5
and a dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit.
The optimal resistance can be expressed as:
𝜋2 𝜁
𝑟𝑜𝑝𝑡 = (2.85)
4𝑘𝑒2
96
Figure 2.76: The normalized response graph of the voltage for a high coupling factor 𝑘2𝑒 = 0.5 and a
dimensionless damping ratio 𝜉 = 0.05 - SSHI circuit.
Figure 2.77: The normalized response graph of the average power for a high coupling factor 𝑘2𝑒 = 0.5
and a dimensionless damping ratio 𝜉 = 0.05 - DCIM circuit.
By substituting the optimal resistance with its expression in equation 2.85, the maximum power
in low coupling conditions can be expressed as:
97
2
′ 𝑘𝑒2 Ω𝑢0′
𝑃𝑚𝑎𝑥 = (1−𝑞𝐼 )𝜋
(2.86)
It can be noticed that when the SSHI circuit is operated in a displacement driven fashion, it
provides a significant increase in output over other circuits.
̅̅̅2 = 2
𝑘 𝑒 𝑘𝑒2 (2.87)
1−𝑞𝐼
The Synchronous Electric Charge Extraction circuit (SECE) [17] is an active circuit for power
harvesting. Its main analytical advantage is the absence of optimal electrical load R. The SECE
is a nonlinear technique for vibration energy harvesting. It consists in switching the
piezoelectric element on a coil for a short time. This technique is very suitable for
electromagnetic generators as well [104]. The main benefit provided by SECE circuits is to
provide a high electromechanical conversion rate as well as to solve the impedance matching
issue.
The Synchronous Electric Charge Extraction circuit is represented in Figure 2.79. It is a DCIM
circuit to which a coupled inductor and a switch are added. The symbol of the positive contact
of coils is a dot. In this particular case, the coils are connected in opposition in order to ensure
that the generated current flows in the desired direction.
The open phase is the main one in the circuit operation. During this phase, the piezo element is
in an open circuit configuration. The maximization of the voltage leads to the primary discharge
phase in which the switch closes. This allows the element to conduct through the rectifier and
primary inductor. The voltage on the piezo element is zero when the current through the rectifier
and primary inductor is maximum. Then, the switch opens back resulting in a secondary
discharge phase. The magnetic field of the coupled inductor is maximum in this state. Due to
the fact that the switch is open, the inductor discharges its energy through current generation in
98
the secondary combination of rectifier and inductor. This generated current flows through the
diode, charging 𝐶𝑆 . Upon completion of the secondary discharge phase, the cycle starts back
from the open phase.
Figure 2.80: Waveforms of the voltage of a SECE circuit. The full line represents 𝑉𝑝 while the dotted
line represents 𝑉𝑑𝑐 .
Figure 2.81: Waveforms of the current of a SECE circuit. The full line represents 𝐼1 while the dotted
line represents 𝐼2 .
In low coupling conditions, the power can be expressed using the equation of the energy stored
(Equation 2.88) in a capacitor because the SECE circuit operates mostly in open circuit mode.
100
1 2
𝐸𝐶 = 2 𝐶𝑝 𝑉𝑝,0 (2.88)
Let’s assume a harmonic displacement with an amplitude 𝑢0 , the maximum piezo voltage 𝑉𝑝,0
can be expressed as:
2𝜃𝑢0
𝑉𝑝,0 = (2.89)
𝐶𝑝
The normalized maximum power in low coupling conditions can be expressed as:
2
′ 𝑘𝑒2 Ω𝑢0′
𝑃𝑚𝑎𝑥 = (2.90)
𝜋
′ 1
𝑃𝑚𝑎𝑥 = 16𝜁 (2.91)
The suspension system mainly consists of shock absorbers, springs, and linkages which create
a connection between a vehicle and its wheel groups allowing relative motion between the two.
It is one of the key components of the chassis of the vehicle. It supports the body of the vehicle
and provides stability, comfortability, and good road handling by absorbing vibrations caused
by rough road profiles.
The hydro-pneumatic suspension systems, also known as oleo-pneumatic suspensions, are the
most common type of suspension on the latest mining dump trucks. They consist of a gaseous
medium that plays the role of the spring and a liquid one that plays the role of the damper. In
addition, a gas accumulator keeps a certain amount of nitrogen at a specified pressure.
Two other types of suspension are used on dump trucks apart from the hydro-pneumatic
suspensions. They are the mechanical and the pneumatic suspension systems [105]. An analysis
of the force vs displacement curves of all three different suspension system types allows
101
characterizing each one of them. As depicted in Figure 2.84, the spring stiffness rate of the
hydro-pneumatic and pneumatic suspension mechanism is variable and progressive. This can
be explained by the gas poly-tropic state change phenomenon. However, the spring stiffness
rate of the mechanical spring remains throughout the whole operation.
During the ride, the truck vibrations generate some energy which is dissipated in the
suspensions. The piston is pushed into the drum leading to a compression of the hydraulic oil
inside the system. The significant sectional variation causes the dissipation of a big amount of
that energy by the means of friction. The rest of it will act on the compressible gas chamber
using the gas accumulator. This allows the conversion of the damping hydraulic oil kinetic
energy into the potential energy of the spring gas. The reciprocal conversion occurs when the
gas (accumulator) expands resulting in a reverse displacement of the hydraulic oil through the
orifices. This phenomenon repeats until all generated energy is completely dissipated. The
necessary damping to stop the oscillation if provided by the hydraulic oil.
The pneumatic suspensions inspire and expire the gas, which is generally air. This means that
the gas volume remains constant after the energy dissipation resulting in a linear relationship
between the pressure variation and the load.
In the case of the hydropneumatic suspension mechanisms, it is the volume of the hydraulic oil
that is changed during the energy dissipation operation. In other words, there is no variation
throughout the whole process. However, the volume of this gas mass varies inversely with the
applied loads (the higher the load is the less the gas mass volume resulting in a higher spring
stiffness rate). This explains the progressive behavior exhibited by the hydro-pneumatic
suspensions in Figure 2.84.
It should be noted that mechanical suspensions provide low natural frequencies at high loads
and high natural frequencies at low loads. For all load conditions, pneumatic suspensions
deliver low natural frequencies. The natural frequencies of the hydropneumatic systems
increase slightly with increasing loads. For all three types of suspensions, the damping stop
occurs through the viscous friction of a fluid, which is usually oil.
Suspension systems components such as the guiding elements, the bushings, and seals cause
the negative impact of additional damping by boundary friction. So, it is very important to
103
eliminate the friction in seals and guiding elements and to keep the friction in the bushings at a
minimum level.
In mechanical suspensions, the mechanical spring undergoes elastic deformation and has no
friction. This means that the boundary friction in such systems can be generated only by the
guiding and sealing elements in the viscous damping fluid. The friction force in dynamic seals
is directly proportional to the length of the sealing edge and to differential pressure at the seal.
The pneumatic suspension systems also use the oil damping technique. The same types of
friction that occurs in mechanical suspensions are present in pneumatic suspensions too. In
addition, there is friction originating from the rolling bellow. The advent of the latest axial ply
bellow allows for decreasing the bellow friction. The additional bellow friction in a passenger
car with standard rolling bellow is approximately 20N. Using the new technology of axial ply
bellow leads to a 50% decrease in friction. In other words, the boundary friction in pneumatic
suspension is slightly higher than that in mechanical suspensions.
When designing hydro-pneumatic suspensions, it is very important to take the boundary friction
into account. Due to its single component integration, the addition of extra spring damping
elements is quite challenging.
The use of rubber bushings in such systems is possible and can help decrease friction levels.
Exposure to high loads makes it difficult to use soft rubber elements. However, it can also
enhance the noise transfer. It is vital to keep the allowable loads as high as possible and to keep
the rubber elements as soft as possible.
Mechanical spring suspension level control and adjustments are quite unusual owing to the
limited effect of this process. The most widespread level adjustment method for mechanical
spring suspension systems is the manual one. However, level control has had very little impact
on mechanical spring suspensions so far.
Hydro-pneumatic and pneumatic suspension level adjustments can be easily achieved through
the addition or drainage of oil or gas from the system. It should be noted that although both
types of suspension systems can go through that process, hydro-pneumatic suspension systems
104
are more accurately leveled. The leveling speed in hydro-pneumatic suspension systems is
higher than that of the pneumatic ones.
Figure 2.84: Forces equilibrium at the piston of a mono suspension system [105].
𝐹𝐹 = 𝐴𝐾 . 𝑃1 (2.92)
𝐹𝐹∗ = 𝐴𝐾 . 𝑃2 (2.93)
105
𝐹𝐹 − 𝐹𝐹∗
𝑐= (2.94)
∆𝑆
Moreover, the implementation of a flow resistor between the accumulator and the cylinder
allows provides the system with additional damping. This is achieved through the conversion
of a portion of the hydraulic fluid kinetic energy into heat. Thanks to this additional mechanism,
the desired damping is obtained alongside undesirable boundary friction caused by the cylinder
guiding elements and seals.
In state 0, the spring force 𝐹𝐹0 is equal to 0. The gas occupies the complete volume of the
accumulator 𝑉0 and the accumulator pressure is equal to the pre-charge pressure 𝑃0 defined
during the design of the hydro-pneumatic suspension.
In state 1, the static suspension force 𝐹𝐹1 loads the suspension the suspension strut. It should be
noted that 𝐹𝐹1 > 𝐹𝐹0 . The application of the static suspension force 𝐹𝐹1 leads to isothermal
accumulator volume change from 𝑉0 to 𝑉1 and pressure change from 𝑃0 to 𝑃1 .
In state 2, the dynamic suspension force 𝐹𝐹2 oscillates around 𝐹𝐹1 . Thus, the accumulator gas
experiences alternate and poly-tropic compressions and expansions. The new pressure and
volume are 𝑃2 and 𝑉2, respectively.
Growth of the external force acting on the cylinder of hydro-pneumatic suspension systems
causes simultaneously an increase in the pressure and a position change of the piston rod. This
is justified by the compression of the hydraulic fluid, the widening of the fittings and fluid lines,
and the compression of the gas in the accumulator. After removing the flow resistor from a
hydro-pneumatic suspension system, its spring stiffness rate can be found using the pure spring-
force displacement curve of the suspension cylinder. The spring stiffness rate consists of three
106
components that are related to hydraulic fluid compression, gas compression, and fluid lines
and fittings widening. The application of general laws of physics leads to equation (2.95)::
𝐶𝐺 𝐶𝐿 𝐶𝐹
𝐶𝑆 = 𝐶 (2.95)
𝐺 𝐾𝐿 +𝐶𝐺 𝐶𝐹 +𝐶𝐹 𝐶𝐿
The hydraulic fluid; the lines and fittings do not have a significant impact on the overall spring
stiffness rate 𝐶𝑆 owing to the high stiffness of the fluid fittings and lines as well as the high
compression modulus of the hydraulic fluid. In other words, the hydro-pneumatic suspension
system’s spring characteristics are mainly affected by the accumulators’ gas properties.
The advantage of the latter configuration over the first one is that it provides more rebound
damping. The cross-sectional area of the piston is considered the externally effective area of
the piston rod. The force-displacement curve for compression and rebound. For a mechanical
coil spring, this force-displacement curve is linear. Its linear nature can be changed using the
parallel connection of coil springs and winding techniques. The force-displacement curve of
hydro-pneumatic suspensions is disproportionately progressive.
The relationship between the pressure in the piston chamber and the external force applied to
the surface of the piston can be written as:
𝐹𝐾 = 𝑃𝐾 𝐴𝐾 (2.96)
𝑉2 = 𝑉1 − 𝐴𝐾 𝑆 (2.98)
108
𝑃1 𝑉1𝑛 𝑃1 𝑉1𝑛
𝑃2 = = (2.99)
𝑉2𝑛 (𝑉1 − 𝐴𝐾 𝑠)
The change of state from state 0 to state 1 is isothermal. It can be deduced that:
𝑃1 𝑉1 = 𝑃0 𝑉0 (2.100)
𝑃0 𝑉0
𝑉1 = (2.101)
𝑃1
𝐹𝐹1 = 𝑃1 𝐴𝐾 (2.102)
and therefore:
𝐹𝐹1
𝑃1 = (2.103)
𝐴𝐾
𝑃𝐾 (𝑠) = 𝑃2 (2.104)
𝑛
𝐹𝐹1 𝑃 𝑉
𝑋 ( 𝐹0 0 )
𝐴𝐾 𝐹1
𝐴𝐾
𝐹𝐾 (𝑠) = 𝑛 𝐴𝐾 (2.105)
𝑃 𝑉
( 𝐹0 0 − 𝐴𝐾 𝑆)
𝐹1
𝐴𝐾
and thus:
𝑛
𝑃 𝑉
( 0 0)
𝐹𝐹1
𝐹𝐾 (𝑆) = 𝐹𝐹1 𝑋 𝑃0 𝑉0
𝑛 (2.106)
( − 𝑆)
𝐹𝐹1
𝑃0 𝑉0
= ℎ0𝐹 (2.107)
𝐹𝐹1
(ℎ0𝐹 )𝑛
𝐹𝐾 (𝑆) = 𝐹𝐹1 𝑋 (ℎ0𝐹 − 𝑆)𝑛
(2.108)
ℎ0𝐹 represents the height of a column of gas whose pressure is equal to the pre-charge pressure
𝑃0 , whose volume is equal to 𝑉0. The external force 𝐹𝐹1 is applied to the top of the gas column
as represented in Figure 2.88.
A major characteristic of hydro-pneumatic suspension systems is that the higher the static
spring load is, the smaller the height of the gas column ℎ0𝐹 is. It can be deduced that at any
given displacement 𝑆, the higher the variation in gas pressure forces is, the higher the spring
(ℎ0𝐹 − 𝑆)
stiffness rate is. The relative change of the column height can be expressed as ⁄ℎ .
0𝐹
The smaller ℎ0𝐹 , the more significant the relative change of column height, the relative decrease
of volume and the relative increase of pressure. The spring stiffness rate depends on the static
spring load:
𝑑𝐹 𝑑(𝑃𝐴𝐾 ) 𝑑𝑃
𝑐= = = 𝐴𝐾 𝑑𝑠 (2.109)
𝑑𝑠 𝑑𝑠
110
Figure 2.87: height of gas column ℎ0𝐹 for different static spring loads 𝐹𝐹1 [105].
By differentiating Equation (2.99) and applying the rule of chains.
𝑑𝑃
= 𝑃1 𝑉1𝑛 (−𝑛)(𝑉1 − 𝐴𝐾 𝑆)𝑛−1 (−𝐴𝐾 ) (2.110)
𝑑𝑠
By taking into account the isothermal change from 0 to 1 and the equilibrium of forces at the
piston according to Equation (2.103) and Equation (2.101), Equation (2.111) can be deduced:
𝑑𝑃 𝐹𝐹1 𝑃0 𝐴𝐾 𝑉0 𝑛 𝑃0 𝐴𝐾 𝑉0 −𝑛−1
𝑐 = 𝐴𝐾 𝑑𝑠 = 𝐴𝐾 ( ) (−𝑛) ( − 𝐴𝐾 𝑆) (−𝐴𝐾 ) (2.111)
𝐴𝐾 𝐹𝐹1 𝐹𝐹1
By dissolving the expression and substituting ℎ0𝐹 with its expression in Equation (2.107):
𝑛
ℎ0𝐹
𝑐(𝑠) = 𝐹𝐹1 𝑛 (ℎ0𝐹 −𝑆)𝑛+1
(2.112)
Equation (2.112) demonstrates that ℎ0𝐹 is one of the variables of the spring stiffness rate. ℎ0𝐹
is a function of 𝐹𝐹1 .
In the particular case 𝑆 = 0 and by dissolving ℎ0𝐹 , the spring stiffness equation can be
expressed as:
111
2
𝐹𝐹1
𝑐= 𝑛 (2.112)
𝑃0 𝑉0
It should be noted that the geometry of the suspension cylinders does not affect any of these
equations. The suspended load and the properties of the gas in the accumulator define the shape
of the force-displacement curve on which the spring stiffness rate depends. The main
characteristics of the gas in the accumulator are its pre-charge pressure 𝑃0 and the accumulator
volume 𝑉0. The gas can also be described by its ideal gas state equation: 𝑚𝐺 𝑅𝑇. This shows
that the spring stiffness rate also depends on the mass of the gas fill 𝑚𝐺 as well as its temperature
𝑇. The lower the accumulator gas temperature is, the stiffer the spring is.
In a suspension system, the damping elements allow dissipation of the external energy
transmitted by the wheel groups, chassis, and skip. Thanks to the damping phenomenon,
oscillation amplitudes increase due to resonance are avoided. The damping elements play a role
in the conversion of mechanical energy (mostly kinetic and a bit potential) into heat through
the application of retarding forces during suspension systems operation. The damping forces in
a suspension are based on two main principles which are the fluid, viscous or hydrodynamic
friction, and the boundary, dry or solid body friction. On one hand, in the fluid, viscous or
hydrodynamic frictions are obtained through the placement of a flow resistor in the flow
channel resulting in an increased fluid pressure that acts on the moving parts of the cylinder
leading to the generation of the retarding force. On the other hand, Boundary, dry or solid body
frictions are obtained through the sliding of bodies against each other along their interface
generating resistance due to adhesion and surface roughness.
Moreover, there are other principles that are not quite usual such as the eddy-current principle
which is implemented in wear-free retarder applications for vehicles' speed reduction on
particular road profiles. The eddy-current principle uses the current induction of an electric
conductor during its movement in a magnetic field. In addition, it is important to note the so-
called gas-spring-damper-elements which play simultaneously the role of both a damper and a
spring using their internal gas fill.
112
In a hydro-pneumatic suspension system, the pressure is transferred through the hydraulic fluid
on the mobile parts of the piston to the accumulators. The piston displacement allows the
hydraulic fluid to flow between the accumulator and the cylinder with flow direction variations
based on the frequency. The shear flows inside the fluid convert the kinetic energy of the
hydraulic fluid into heat thanks to the installation of the flow resistors. A pressure loss is
generated by the flow resistor resulting in a counteracting force to the piston motion. Hence this
resulting force is a damping force as it will drain the energy out of the oscillation as depicted
in Figure 2.89.
𝐹𝐷,ℎ𝑦𝑑 represents the rod side force inserted from the upper part of the suspension connected to
the truck’s chassis. This force is mainly driven by the surface bumps (in motion) and high-
impact shovel loading operations (at idle).
𝐹𝐷,ℎ𝑦𝑑 is equal to the resulting pressure drop divided by the effective cross section of the piston.
𝑃𝐷,ℎ𝑦𝑑
𝐹𝐷,ℎ𝑦𝑑 = (2.114)
𝑣
It is very important to state that the fluid friction damping force is greatly related to the speed
of the suspension.
Two different types of simple resistors exist depending on their main characteristics (pressure
loss and volume loss) which are throttles and orifices. The first type of simple resistor is the
throttles in which the flow is decelerated using a slow transition of the flow cross-section from
wide to narrow and back to wide as depicted in Figure).
∆𝑃 = 𝑉̇ 𝑣𝜌𝐾𝐷 (2.116)
128𝑙𝐷
𝐾𝐷 = 4 (2.117)
𝜋𝑑𝐷
114
The second type of simple resistor are the orifices in which the fluid flow is subjected to single
or multiple sudden transitions from wide to narrow and vice versa flow path as depicted in
Figure 2.91.
∆𝑃 = 𝑉 2̇ 𝐾𝐵 (2.118)
𝐾𝐵 is a constant which depends on the geometry, dimension of the throttle and the hydraulic
fluid density. For an orifice resistor 𝐾𝐵 can be calculated as:
8𝜌
𝐾𝐵 = 2 𝜋2 𝑑 4 (2.119)
𝛼𝐷 𝐵
Where 𝛼𝐷 represents the flow coefficient and depends on the inlet edge geometry and the
Reynolds number.
One of the characteristics of the suspension system is the limited stroke available to isolate the
excitations received from the input side. The amount of displacement between the top and
115
bottom sides of the suspension system depends on the system's hardness. In case the suspension
system is soft, the amount of displacement is huge owing to the acceleration decrease on the
isolated side. If excitations reach a certain magnitude, the required displacement soars to levels
greater than the available suspension stroke affecting negatively the suspension system
components. This situation leads to a decrease in ride quality and comfort.
A common solution to prevent this issue is to upgrade the suspension spring and damper by
making them harder. This provides sufficient suspension stroke for the worst and most extreme
operating conditions. However, one of the downsides of this upgrade is the reduction of comfort
and ride quality in normal operating conditions. In order to limit the severity of a bottoming-
out event at the end of the stroke, a reduction of the velocity of the piston relative to the cylinder
is necessary. In case the piston approaches the end positions too much, it might be necessary to
apply some extra decelerating force. This provides the system with a constant or slightly
increasing deceleration as the piston gets closer to the end positions.
Another alternative is the use of elastomer elements to mitigate the effects of end-of-stroke
damping. These elastomers elements are set to provide a deceleration to the system right before
the piston hits the end positions. This configuration allows the system to take both minor
impacts and extreme bumps without affecting steel elements. In light vehicles, riders can notice
this kind of end-of-stroke damping, and the components are well protected from overload. The
elastomer elements extend back to their original dimensions and shape at the end of the rebound
step. Not all the energy is dissipated as a little amount of it remains in the elastomer material
resulting in its loss angle. It is important to note that the elastomer elements present
disadvantages such as early aging due to settlement and operating conditions. Therefore, it is
necessary to replace them either on a regular frequency or depending on their conditions. To
avoid overloading, the system design includes mechanical wedges which limit the stroke in
order to prevent extreme strain.
The flow resistor creates a pressure drop which in turn leads to a pressure increase in the
cylinder chamber. As a result, some additional damping force is applied to the moving parts of
the suspension system. Sometimes the cross-sectional area of the additional is variable and
changes according to the cylinder stroke. In this case, a mechanism is necessary to define the
effect of the flow resistor on the basis of the piston displacements. The advantage of such
116
variable flow resistors over the ones with constant cross-section is that they can provide a more
stable end-of-stroke damping force.
In order to analyze vehicle dynamics, numerous mathematical vehicle models have been
developed over years. These models use lumped parameters and are based on the fundamental
laws of physics such as the conservation principle and Newton’s laws. That is why they are
called lumped parameter models and they can be utilized to assess the performance and features
of vehicle suspension systems. There are three main models which are commonly used. They
are known as the full car model (FCM), half-car model (HCM), and quarter car model (QCM).
Thanks to its simplicity, the quarter car model is often given priority for vehicle suspension
system analysis. Indeed, it provides the paramount parameters of the full vehicle model. Florin
et al. in their research work [106] compared three different techniques of quarter car model
analysis which are the state-space model, transfer function, and the simulation in the MATLAB
Simulink environment. The outcomes of this study revealed that all three techniques provided
the same suspension parameter values. In their research work, Sharma et al. [107] defined two
degrees of freedom for a quarter car model through the state-space representation in the
MATLAB Simulink environment. Step inputs excitations of 0.1m were supplied to the car. As
a result, a high overshoot of the unsprung mass (30%) and the sprung mass (80%) has been
observed. The obtained results were below the ride quality and comfort expectations. To solve
these shortcomings, a controller was designed to maintain the car movements within acceptable
comfort values. Previous works carried out by Vaishnav et al. [108] allowed the establishment
of the quarter car mathematical model using the transfer function alongside Cramer’s rule. A
study performed by Jamali et al. [109] focused on developing a quarter car model for an
electrically powered car for the purpose of assessing its isolation from road disturbance.
Comparisons between the acceleration and displacement amplitudes of the sprung and unsprung
masses at different damping coefficients were carried out. The minimum frequency response
was recorded at the damping constant 𝑏𝑜𝑝𝑡 in the Simulink MATLAB program. The expression
of 𝑏𝑜𝑝𝑡 was:
117
𝐾𝑠 𝑀𝑆 𝐾𝑢𝑠 +2𝐾𝑆
𝑏𝑜𝑝𝑡 = √ =√ (2.120)
2 𝐾𝑢𝑠
Hassaan [110] carried out a study in which he observed the behaviors of a quarter car half model
driving on a speed hump with conservation of ride quality and comfort. The state model was
used to determine the mathematical models. Then these models were analyzed on MATLAB.
118
The study revealed that for the ride quality and comfort to be conserved, the vehicle speed
should not exceed 7 km/h in the vicinity of speed humps.
Gao et al. [111] conducted research work on the impact of the randomness of a quarter car
model parameters on the dynamic analysis. The study revealed that the randomness of the mode
shapes does not depend at all on the tire stiffness. Nevertheless, it depends on the unsprung and
sprung masses. It is also important to note that while the first natural frequency is influenced
by the uncertainty sprung mass, the second natural frequency is affected by the uncertainty
unsprung mass system.
In order to perform experiments, quarter car models can be designed by connecting two moving
plates to the spring and the damper. The two moving plates represent the unsprung and sprung
masses. The major difference between the experimental models resides in the nature of the
unsprung mass which in some models is substituted by real car tires. Another varying important
factor is the intensity of the excitation input force.
On one hand, there are experimental models with real car tires. Salah [113] conducted an
experiment to study the quarter car model. His design consisted of real tire which was connected
to a moving plate with shock absorbers in between as depicted in Figure 2.93. Four stainless
steel rods were added to the design as guides to keep torsion at minimum levels.
The quarter of the sprung mass is represented by the moving plate while the spring and damper
are represented by the shock absorbers. The advantage of incorporating the model with a real
tire is the procurement of very realistic results. The maximum capacity of the designed model
300 kg and it could sustain a motor speed of 162 km/h.
On the other hand, several experimental models have been developed with a moving plate as
an unsprung mass over time. For instance, Jugulkar et al. [114] built their model using two
moving plates connected to four springs of equal stiffness (3.33 N/mm) and adjustable dampers.
119
Figure 2.93 shows that the springs were connected in parallel. An additional spring of stiffness
32.77 N/mm was linked to the damper in series at the center of the plate. A central bar with
roller bearings was implemented for guidance and friction reduction. The excitation input was
provided by an electric motor. In order to convert the rotational movement of the motor into
linear motion of the system a Scotch Yoke mechanism was incorporated within the setup.
Figure 2.92: Experimental model of the quarter car model with a real tire as unsprung mass [113].
120
Figure 2.93: Experimental model of the quarter car model with a moving plate as unsprung mass
[114].
The half-car model is a four degree of freedom model which includes 2 tires. One of the tires
is from the front axle and the other one is from the rear axle. In this model, an assumption is
made about the symmetry of the car about its longitudinal axis. The half-car model is
represented in Figure 2.95. Parameters such as the gravity center, spring, and damper
characteristics which are used in these models tally with those of actual cars. In a research
work conducted by Patel et al. [115], an algorithm was developed in MATLAB Simulink to
assess the road profile at various speeds using the half-car model. An input road profile was
121
used for the excitation of the model and a comparison was made between that input road profile
and the ones measured at different speeds. The outcomes of the simulation in MATLAB
Simulink illustrated the concordance between the input and measured road profile in time,
distance, and frequency domain. Gao et al. [116] in their research work demonstrated that the
uncertainty of the natural frequency of the cars is dependent on the uncertainty of the car
parameters. The two first natural frequencies depend on the distance between the rear and front
axle. The third and fourth natural frequencies depend on the front and rear unsprung masses
(tires).
Comparisons between the quarter and half car models were carried out in previous works. In
order to find the optimum damping coefficient values, Faheem et al. [117] carried out a study
to analyze both a quarter and half-car model's displacement and frequency responses. The
results of these analyses revealed that for optimum ride quality and comfort, the value of the
damping coefficient should not exceed 3000 N.s/m for both models. Moreover, Mehmood et
al. [118] analyzed the dynamics of a quarter and half car model using state-space representation
in the MATLAB program alongside Laplace transformation. The optimum damping coefficient
values for maximum ride quality and comfort were recorded. An optimum damping range
between 3000 and 4000 N.s/m was identified for the quarter car model (2 degrees of freedom).
Nonetheless, the half-car model exhibited an optimum damping interval from 1500 N.s/m to
2000 N.s/m.
Previous works were performed to design experimental half-car models. In order to confirm the
results of simulations with experimental ones, Chavan and Sawant [119] designed an
experimental setup made up of two suspension sets connected by a beam. As depicted in Figure
2.96, each suspension consists of a spring and shock absorber. The input excitation is provided
to each suspension by an eccentric cam. The results obtained from the experiments tallied the
simulation results at various frequencies by taking into account the non-linearity of the
suspension system.
Figure 2.97 shows the design of an experimental half-car model developed by Sapiński and
Martynowicz [120] in order to analyze the vibration control with considering the bounce and
pitch motions. Their design consisted of two identical suspension systems made of spring and
magnetorheological shock absorbers (MRA). A steel frame of a weight representing the half-
sprung mass connects both suspensions. The electro-hydraulic cylinder shaker provided the
input excitation to one of them and the electric motor which is connected to a cam crank
mechanism excited the other one. A central roller was added to the design to play the role of
123
guide and stabilizer. The results of the study revealed that the use of an appropriate
magnetorheological shock absorber (MRA) allows for minimizing the vibrations by reducing
the displacement and acceleration amplitudes.
Figure 2.95: Experimental setup of a half car model with two suspensions connected by a beam [119].
Figure 2.96: Experimental setup of a half car model with the MRAs and steel frame [120].
124
The full car model has seven degrees of freedom owing to the rolling motion. A typical full car
model is made of four identical quarter car models which are connected by rods as shown in
figure 2.98. The seven degrees of freedom are the four vertical motions of each unsprung mass,
the roll, pitch, and bounce motions of the car body.
Full car models are generally used to thoroughly analyze and study car ride quality, comfort,
handling, and other performances. In order to study the operator or passenger ride comfort and
quality, such as the speed over a bump, and suspension travel distance, the root-mean-square
(RMS) of the vertical acceleration and road handling must be considered. There are
internationally standardized benchmarks of RMS of vertical acceleration which is presented in
Table 2.11 [121].
0.8-1.6 Uncomfortable
No experimental full car model has ever been designed in any laboratory. Thus, the very few
studies on energy harvesting systems using the full and half car models are generally compared
with the quarter car model whose experimental setups are widely available. Mahala et al. [122]
carried out research work in order to assess the responses of the quarter, half, and quarter car
models using simulation techniques. The conclusions of the study were that increasing the
number of degrees of freedom increases the simulation time and improves the analysis results’
accuracy.
Iyer et al. [123] conducted a study on a full car model to determine the optimum values of the
suspension spring and damper coefficient to provide maximum ride quality and comfort. The
outcomes of these analyses revealed that the optimum stiffness coefficient of 8300 N/m and
damping coefficient of 700 N.s/m were the optimum parameters to achieve maximum ride
quality and car handling. Moreover; in a research work conducted by Mitra et al [124] on the
full car model, some specific values of the parameters were recorded for optimum ride quality
and comfort. The car speed must not exceed 10 km/h when passing over a speed bump, with
minimum suspension travel of 0.127 m and approximately 0.0508 m for the relative distance
between the road and the tires. In an analysis performed by Raju and Venkatachalam [125], the
equation of motions of the full car model was derived in order to observe and study the impact
of various disturbances on the suspension system. The right front wheel was subject to firstly
impulse disturbances and then pulse disturbances. For the case of the pulse input, the maximum
oscillations affected the front axle with little effect on the rear axle. It was found that the front
126
axle maximum oscillations when subject to impulse disturbances was much higher than those
of the pulse disturbances.
A massive amount of energy is lost in the suspension system through heat. It can be converted
and reused using numerous types of energy harvesting mechanisms. The piezoelectric material
can be incorporated with the suspension systems in various components such as springs,
dampers, and even the tires. Xiao et al. [126] carried out both theoretical and experimental
studies on piezoelectric energy harvesting. Figure 2.99 represents the designed quarter car
model where the springs and piezoelectric stacks were linked in parallel.
Figure 2.98: Quarter car modeled designed with piezoelectric material connected in parallel [126].
In the simulation, the two degrees of freedom model (quarter car model) was excited using a
sinusoidal input with a 9.8 𝑚/𝑠 2 (1 𝑔) amplitude. The maximum recorded power and voltage
harvested from the piezoelectric elements were 2.84 W and 274.62 V, respectively. The main
parameters which were influencing the output power and voltage were the damping
127
coefficients, tire and suspension stiffness. The experimental study outcomes validated the
results of the theoretical one. The experimental setup underwent a sine wave with an amplitude
of 1.5 𝑚/𝑠 2 . In another research work conducted by Zuo and Zhang [127], it was reported that
the potential harvestable power was in the interval from 46 to 7500 W.
Previous works were carried out on the energy harvesting potential of vehicle suspension
systems using the half-car model too. Al-Yafeai et al. [128] modeled a half-car model with
piezoelectric materials connected in parallel with both the front and rear suspensions as depicted
in Figure 2.100.
Figure 2.99: half-car model with piezoelectric elements connected in parallel [128].
The results from the simulations revealed that the power and voltage harvested from the quarter-
car model were lower than those harvested from the half-car model. The power and voltage
harvested from the half-car model were higher by 57% and 77%, respectively. Nevertheless,
the calculations were made without taking into account the stiffness of the piezoelectric material
for both the quarter and half car models. The harvested power and voltage will be greatly
affected if the piezoelectric stiffness is considered. An equivalent piezoelectric stiffness
128
coefficient of 6250 kN/m was assumed when coupling the piezoelectric materials to the
suspension systems. The overall system stiffness coefficient of the quarter car model was
calculated using the following expression:
𝐾𝑒𝑞 = 𝐾𝑠 + 𝐾𝑝 (2.121)
When taking into account the piezoelectric material stiffness, the overall stiffness increases.
Indeed, when the system’s stiffness coefficient increases, it experiences less deformation.
Therefore, the harvested voltage and power decrease as the stiffness of the system increases.
When the piezoelectric material is connected in series as shown in Figure 2.101, the stiffness
of the piezoelectric material does not affect the harvested voltage or power because the overall
equivalent stiffness coefficient is approximately equal to the suspension stiffness as described
by the following relationship:
𝐾𝑠 + 𝐾𝑝𝑖𝑒𝑧𝑜
𝐾𝑒𝑞 = ≈ 𝐾𝑠 (2.122)
𝐾𝑠 × 𝐾𝑝𝑖𝑒𝑧𝑜
Some research works were conducted on energy harvesting from vehicle tire deformation. The
deformation of the vehicle tires will lead to the stretching of the piezoelectric material which
will generate electricity. Xie and Wang [28] coupled a dual-mass piezoelectric harvester with
the quarter car model as shown in Figure 2.102 and Figure 2.103. The piezoelectric harvesting
system was mainly made of a lead zirconate titanate PZT4 bar with dimensions of 1.5 cm in
width and 10 cm in height. The model was exposed to random rough road excitations with three
different road profiles ranging from very smooth road to very rough road. By applying the
iteration method and by installing more than four identical piezoelectric bars in the system, a
maximum power value of 738 W could be harvested from the D road profile at a car speed of
35 m/s. The height and width of the bars are 10 cm and 1.5 cm, respectively. Parameters such
as the car speed and the piezoelectric bar width had a great impact on the harvested power and
voltage.
129
Figure 2.100: Quarter car model designed with piezoelectric material connected in series.
Figure 2.101: Dual mass piezoelectric energy harvesting system installed on a tire.
130
Figure 2.102: Quarter car model of car tire piezoelectric energy harvesting system.
Moreover, Zhao et al. [129] modeled a piezoelectric harvester implemented in the suspension
system. The model was exposed to random and pulse excitations. At a car speed of 30 km/h,
harvested power values of 18.83 and 102.24 W were recorded for the random and pulse
excitation, respectively. The conclusions were that the harvested power value is mainly affected
by parameters such as the vehicle speed and the piezoelectric material cross‐sectional area.
Behera [119] in his research work combined three lead zirconate titanate PZT5A piezoelectric
materials in series into the tire’s inner surface. At a car speed of 40 km/h, the model could
deliver a power of 14 mW. In addition, Lafarge et al. [131] conducted an analysis on suspension
wheels on which a lead zirconate titanate piezoelectric cantilever beam of type PZT 27. In the
simulation, the quarter car model was excited to the first natural frequency with different load
resistances. The power harvested using two resistors of 222 kΩ and 22 kΩ was 1.4 mW and 1
mW, respectively. Then the piezoelectric beam was embedded into the unsprung mass in order
to perform a real case study. At a car speed of 10 km/h, the harvested power ranged from 0.001
mW to 0.021 mW. The range moved to 0.01 mW and 0.07 mW as the car speed was increased
to 30 km/h.
131
The piezoelectric energy harvesting system can also be placed in the center of the tire. The
rotational movement of the tire provides the system with some energy. In order to optimize the
harvested power amount, it is necessary to match the frequency of the tire revolutions with the
natural frequency of the piezoelectric energy harvester. Hu et al. [132] stated that for an
acceleration of 30 m/s2 and a travel distance of 12 mm, an output voltage of 1.5 V can be
obtained by using a nanogenerator of 1.5 cm × 0.5 cm dimensions. The role of the
nanogenerator is to convert the tire rotation mechanical energy into electrical energy using
piezoelectric films bender. The experimental model used a bicycle tire instead of real tire in
order to simplify the equation. The harvested power could be stored in a battery or capacitor
for various purposes. The analysis revealed that parameters such as the car speed and the travel
distance influence greatly the output voltage and power.
Zhang et al. [133] simplified a ceramic piezoelectric harvester as a uni-morph cantilever beam
and installed it in the center of the rotating car tire. Experimental analyses were carried out on
a piezoelectric layer with dimensions of 7 cm ×1.5 cm ×1.5 cm. In this experiment, the
excitations were provided by a shaker whose frequency ranged from 0 – 2000 Hz. Power of
about 0.032 mW was harvested by the model using a 252 kΩ resistor. The test was performed
at different car speeds of 10 km/h, 20 km/h, 30 km/h, and 40 km/h. The study revealed that the
harvested power increased as the car speed rose. Nevertheless, when a stochastic resonance
input was provided to the system, the highest average power was obtained at a car speed of 20
km/h.
Makki and Pop-Iliev [134] conducted their analysis with three different piezoelectric materials.
These are PVDF thin sheet, PVDF ribbon on the tire bead, and PZT unimorph installed on the
inner surface of the tire. The outcomes of the study revealed that the PZT attached to the inner
tire surface provided the greatest harvested power value (4.6 mW) at a speed of 8 km/h.
Furthermore, PZT is cheap as compared to PVDF and exhibits high resistance to heat. An
energy harvesting model was designed by the Apollo project [135] with a piezoelectric PVDF
film size of 8 cm × 8 cm inserted into the inner surface of the vehicle tire. The model could
generate a power of 0.9 mW at a car speed of 80 km/h. In light of the findings of the study, it
was revealed that the harvested power is linearly dependent on the car speed, PVDF film area,
and wheel load. The piezoelectric material can be shaped as a shock absorber for car suspension
systems. For instance. Ali and Adhhikari [136] made a single degree of freedom piezoelectric
132
shock absorber coupled to the electric circuit made of electrical resistance and capacitance. The
excitation used in the study was sinusoidal wave input with constant frequency. Madhav and
Ali [137] designed a similar model but exposed it to the random input excitation instead. Lee
et al. [138] designed a regenerative shock absorber made of two piezoelectric layers. Electric
energy was generated when the suspension was subject to vibrations. Arizti [139] used a
multilayer piezoelectric stack as a shock absorber. The piezoelectric material was installed on
top of the piston as represented in Figure 2.104. The voltage harvested from the system when
the car was passing over a bump was 17.69 mV.
Figure 2.103: Multilayer piezoelectric stack installed inside the shock absorber [139].
In addition, Hendrowati et al. [140], in their quarter car model, utilized a multilayer (ML PZT
VEH) piezoelectric stack made of 15 layers connected in series with the suspension spring
(Figure 2.105). The purpose of the mechanism was to convert the vertical vibrations of the
suspension system into horizontal displacement to increase the spring force applied on the
133
piezoelectric stack. This model allowed to achieve a harvested power value that is 7.17 times
higher than the direct installation configuration. The stiffness coefficient of the piezoelectric
stack was taken into account for the calculations.
Figure 2.104: Quarter car model with the multilayer (ML PZT VEH) piezoelectric stack connected in
series with the spring [140].
A comparison of the main piezolecric energy harvesting systems was summarized in Table
2.12.
two PZT 27 layers Quarter car Random rough road, simulation Simulation (2V;
model (30 km/h), experiment (91 Hz; 0.02W) Experiment
10 km/h & 30 km/h) (13V; 1.4mW)
PZT stack located in the Quarter car Harmonic excitation, 43.2 km/h 6.23V; 1.6 mW
shock absorber model
135
When dump truck tires are properly inflated, a lot of parameters are improved such as the
reduction of the braking distance, increased fuel economy, and vehicle handling by the operator.
Another edge of the proper inflation of tires is their extended life as compared to overinflated
or underinflated tires. According to Wingert [141], approximately 3/4 of vehicles around the
world are driven or operated with at least an underinflated tire. Underinflation is mainly due to
road profile and hazards, huge temperature variations, and natural leakages [142].
Underinflation may sometimes not be visually identified and the layman does not have the
reflex to inspect the tires. Recent research in the field of piezoelectric energy allowed us to
foresee its future diverse applications in several domains. One of these applications is supplying
electrical power to remote tire pressure monitoring systems in order to obtain an autonomous
self-powered mechanism and to avoid issues related to battery replacement.
The main objective of the Tire Pressure Monitoring System is to notify the machine operator
when tire pressure is out of the recommended range resulting in unsafe driving conditions. It's
a very important system that is fundamental for proper tire inflation. Both overinflation and
underinflation can cause premature tire failure and lead to tread wear. On one hand, when the
tires are overinflated, the traction decrease and t, and the inability to absorb road impact.
Overinflated tires will show premature wear in the center the tire wear rates are above normal.
On the other hand, when the tires are underinflated, much fuel is consumed alongside the
excessive heat buildup in the tire material. An underinflated tire is also subject to overloading.
There are two main types of tire pressure monitoring systems.
In the direct configuration, the pressure drop is calculated directly by using the data transmitted
by the pressure sensors. A typical and simple TPMS consists of a sensor fitted on the tire valve
stem, which experiences a color change when the pressure decreases. This type of sensor is not
very accurate and cannot convey the data to the dashboard instrument clusters. Another
alternative of direct TPMS includes radio frequency transmitters and receivers as well as a
warning system as depicted in Figure 2.93 [143].
136
The size of the system must be taken into account for easy application. To that end, the sensor
module should not be bulky or heavy to mitigate the effects of centrifugal forces. Most TPMS
modules operating temperatures range from −40 ◦C to 120 ◦C. A typical sensor module consists
of a transmitter, control unit, antenna, sensor, and long-life battery for power energy as
represented in Figure 2.93. The sensor module possesses a valve stem that is fitted from inside
the tire through the holes on the wheel.
137
Figure 2.107: Direct TPMS sensor module fitting in the wheel station [143].
Numerous TPMS sensors are manufactured using the complementary metal-oxide-
semiconductor (CMOS) technology, which allows for achieving very low energy consumption
levels. Moreover, an analog-to-digital converter and nonvolatile memory are incorporated with
the module to store tire pressure data. In upgraded versions, a temperature sensor is coupled to
the pressure sensor. To control the functions of the sensor and transmitter, A microcontroller is
fitted in the module to control the sensor and the.
A typical receiver module consists of a control unit and antenna. Its role is to receive the data
and transfer it to a central processor which is connected to the instrument cluster and the
warning system. An alternative could be to integrate the receiver with the vehicle remote
keyless entry system by properly setting up the transmission frequency and protocols. A basic
portable receiver display unit consists of color symbols with a LED specific to each tire to signal
the tire under inflation and over inflation events (Figure 2.95).
2.9.1.1.Active TPMS
In the active TPMS configuration, the tire pressure sensors receive their power from short-life
lithium batteries. The signal is then transmitted to the central receiver module through a wireless
transmitter, to the dashboard cluster instruments on which a visual and audible alarm are
triggered in case of abnormal levels. The disadvantages of the active TPMS are related to its
lithium battery's limited life and cost [144].
2.9.1.2.Passive TPMS
There are three main types of passive tire pressure monitoring systems.
Micro-vibration power generation TPMS: The design of this technique was made
based on the patent “a kind of micro-vibration generator” applied by He Cho Ngzhong
in China. The module is fitted on the tire rim. The vibrations of the vehicle cause the
coil inside the module to cut the magnetic induction line resulting in the generation of
electric energy that is more than electromagnetic induction. This technique features a
large power generation capacity, but the vibration amplitude is small due to the lack of
tire inner space. This situation leads to low energy generation.
Piezoelectric Ceramic Power Generation TPMS: piezoelectric ceramics generate
piezoelectricity after undergoing mechanical stress. Microstructurally, the inner
negative and positive charges displacement leads to polarization. The piezoelectric
material is compact and small in size, light in weight and high in sensitivity (Figure
2.98).
RF passive TPMS: This technique is an alternative to the central receiver in a standard
direct TPMS with a central transceiver. The great advantage provided by such a system
is the handling of large weight and volume brought by the active battery. Moreover, it
has very high sensitivity and wide working. Nevertheless, the circuit design of such
system is complex and requires more engineering to implement.
In addition to the aforementioned passive TPMS techniques, there are other methods such as
temperature difference, and acoustic TPMS power generation. Research has been made to study
these alternatives in order to lower their relatively high production costs and increase their
power generation potential. It is important to note that most of the vehicles use active TPMS.
140
Figure 2.111: Direct TPMS Piezoelectric ceramic power supply schematic diagram [144].
2.9.2. Indirect Tire pressure Monitoring System
An indirect TPMS does not require any sensor to work. By using an observer and software
codes, it can forecast tire pressure variations. Most of the indirect TPMSs currently available
on the market monitor the wheel speed. In case the tire pressure decreases, the weight of the
vehicle decreases the diameter of the tire. The main parameter to be taken into account is the
tire angular speed ω which can be deduced by observing the deflection in Figure 2.99.
141
𝜔𝐿𝐹 , 𝜔𝑅𝑅 , 𝜔𝑅𝐹 , 𝜔𝐿𝑅 represent the left front, right rear, right front and left rear wheel angular
velocities whilst 𝜔𝑎 is the average angular speed. When the tire is properly inflated, β should
be close to zero. In other words, only software is needed to implement an indirect TPMS. A
disadvantage of the indirect TPMS is that it does not provide the actual pressure of each tire
Moreover, it is functional only when the vehicle is being operated.
Hybrid TPMS are systems that combine both direct and indirect TPMS methods. It was
designed to fill the gaps of indirect TPMSs by adding sensors. This method is based on an
indirect TPMS with pressure sensors fitted in two diagonal tire positions. A typical hybrid
TPMS consists mainly of an indirect TPMS algorithm coupled with two pressure sensors fitted
on diagonally located tires (Figure 2.100) [143].
142
Thanks to the presence of the pressure sensors fix the real-time pressure value monitoring issue.
Therefore, the detection capability and reliability are improved as compared to the indirect
TPMS. Nevertheless, its application is not widespread due to the fact that when both monitored
diagonal tires get underinflated at the same time, only one of them will probably be identified
by the system. Thus, the reliability of the hybrid system is not very high under certain
circumstances.
Mathematical models of piezoelectric energy harvesters and the estimation of tire parameters
are necessary to optimize the cost of direct TPMSs and overcome the shortcomings of hybrid
ones [145].
144
The energy harvested from the suspension system can be used to power mini devices.
Nevertheless, this energy needs also to be accumulated and stored in order to be reused. Before
storing the energy, a rectifier is required to convert the AC waveform voltage from the piezo
elements into DC voltage which can be easily stored using capacitors and batteries. An example
of a typical and simple rectifier circuit is the diode bridge which is described in the [146-149].
The schematic diagram of the circuitry is represented in Figure 3.1. The circuit mainly consists
of a diode bridge, a smoothing accumulator, and a battery. The harvested voltage will be first
rectified before being accumulated through the large capacitor. The capacitor is connected in
parallel with the battery which is supposed to store the energy. Several research works were
carried out on the implementation of voltage regulators into the circuit after rectification. The
main types of generators used in the energy harvesting circuit are the step-down and step-up
regulators. Voltage regulators could be a step-up or step-down converter.
frequency and increase the current flow to the battery. Moreover, the control helps to adjust the
duty cycle. Consequently, Hofmann et al. [151] designed their circuit with a fixed duty cycle
to maximize the harvested power. The results revealed that the harvested power soared to 30.66
mW which is approximately three times the value of the direct charging power of the battery.
The capacitors charge and discharge very rapidly owing to their high-power density. Hence,
they are suitable for rapid energy transfer systems. Nevertheless, electrochemical batteries are
preferred for storing energy. Sodano et al [152] carried out tests on a piezoelectric harvester
made of plates using two different storage devices which were a capacitor and a battery. In light
of the study findings, it was stated the most efficient storage option is the. Indeed, a battery can
power a larger variety of devices in comparison with capacitors due to their rapid discharge.
One of the disadvantages of the batteries is that they have a limited number of charging cycles
as mentioned by Chen et al. [153]. Indeed, Lead-acid, Nickel-Cadmium, Sodium Sulphur, and
Lithium-Ion batteries’ number of cycles are 500 – 1000, 2000 – 2500, 2500 cycles, and 10000
cycles, respectively.
The supercapacitors, also known as ultra-capacitors, are alternative storage devices to the
conventional capacitor or battery [154-156]. Their special property is that they have
significantly higher capacitance and energy storage ability than conventional capacitors.
Furthermore, they can provide a cycle life of about 106 cycles which is higher than that of
electrochemical batteries [157]. In order to achieve higher power outputs, it is possible to
combine the supercapacitors with rechargeable batteries as stated by Tolentino and Talampas
[158]. In their experiment, they placed a Lithium-Ion battery as a primary storage unit and the
146
supercapacitor as a secondary one. This allowed them to achieve greater performances than the
exclusive use of only one of them.
The following parts cover the purely mechanical system mathematical models. The mechanical
system modeling will be discussed.
The quarter car model is the simplest and most basic way to model a car. It represents the quarter
of the car and discusses exclusively the vertical displacement (vibrations) of its car body, also
known as bouncing for the purpose of early design. The model is mainly made of two masses
known as the unsprung and sprung masses and the suspension as depicted in Figure 3.2. The
unsprung mass Mus represents a tire assembly while the sprung mass Ms represents the quarter
of the car body. The suspension strut consists of a spring of stiffness coefficient Ks and a linear
viscous damper of damping coefficient bs. The suspension struts is connected to the bottom of
the sprung mass while the unsprung mass representing the wheel assembly is in direct contact
with the road surface through spring and damper which illustrate the tire stiffness Kus and
absorptivity bus, respectively. The excitation is provided to the system car by the road excitation
yR which leads to the vertical displacement of the tire yus as well as the car body as depicted in
147
Figure 3.2. In order to model the quarter car Model, Newton’s 2nd law can be utilized. Two
equations can be deducted from the two degrees of freedom:
𝑀𝑢𝑠 𝑦̈𝑢𝑠 = 𝐾𝑠 (𝑦𝑠 − 𝑦𝑢𝑠 ) + 𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 ) − 𝐾𝑢𝑠 (𝑦𝑢𝑠 − 𝑦𝑅 ) − 𝑏𝑢𝑠 (𝑦̇𝑢𝑠 − 𝑦̇ 𝑅 ) (3.2)
The half-car model is a representation of the vehicle with a front and a rear tire assembly. This
model consists of four degrees of freedom which are the vertical motions of the front (yuf) and rear
axles (yur) the pitch θs and bounce ys and pitch motions of the car body. As represented in Figure
3.3, the half-car body is connected to the front and rear tires. The front and rear unsprung masses
Muf and Mur, respectively, are free to bounce vertically about the half car body. Whereas, the sprung
mass Ms is free to bounce and pitch. The front and rear suspension systems are made of linear
springs of stiffness coefficients Ksf and Ksr as well as linear viscous dampers bsf and bsr. The tires
are represented as linear springs of stiffness coefficients Kuf and Kur as well as linear dampers of
damping coefficients buf and bur.
149
Four equations of motions can be derived from the four degrees of freedom of the half car
model:
𝑀𝑢𝑓 𝑦̈ 𝑢𝑓 = 𝐾𝑢𝑓 (𝑦𝑢𝑓 − 𝑦𝑅𝑓 ) − 𝑏𝑢𝑓 (𝑦̇ 𝑢𝑓 − 𝑦̇ 𝑅𝑓 ) + 𝐾𝑠𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑢𝑓 ) + 𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 −
𝑦̇ 𝑢𝑓 ) (3.3)
𝑀𝑢𝑟 𝑦̈𝑢𝑟 = 𝐾𝑢𝑟 (𝑦𝑢𝑟 − 𝑦𝑅𝑟 ) − 𝑏𝑢𝑟 (𝑦̇𝑢𝑟 − 𝑦̇ 𝑅𝑟 ) + 𝐾𝑠𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑢𝑟 ) + 𝑏𝑠𝑟 (𝑦̇𝑠 − 𝐿𝑟 𝜃̇𝑠 −
𝑦̇𝑢𝑟 ) (3.4)
𝑀𝑠 𝑦̈𝑠 = −𝐾𝑠𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑢𝑓 ) − 𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) − 𝐾𝑠𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑢𝑟 ) −
𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 )
(3.6)
𝑦̈ 𝑢𝑓 represents the vertical acceleration of the front unsprung mass (front tire),
150
𝑦̈𝑢𝑟 represents the vertical acceleration of the rear usprung mass (rear tire),𝑦̈𝑠 represents the
acceleration of the sprung mass (car body),
𝑦̇ 𝑅𝑓 represents the vertical velocity of the road at the front of the car,
𝑦̇ 𝑅𝑟 represents the vertical velocity of the road at the back of the car.
An approximation of the harmonic excitation of the front and rear tires of the car are given by
the following delayed function:
𝐿
𝜏 is the time delay which is equal to where 𝐿 is the wheelbase and 𝑢 is the velocity of the car.
𝑢
The suspension system is subject to numerous types of non-linearity such as friction damping,
clearance gaps, backlash of the components, quadratic, cubic and quartic stiffness. The quartic
151
non-linear stiffness will be considered for this study. The spring force in this case can be
described in this case by the following expression:
Where,
𝐾4 = 136150000 𝑁/𝑚4
𝐾3 = −24800000 𝑁/𝑚3
𝐾2 = −520000 𝑁/𝑚2
𝐾1 = 2320000 𝑁/𝑚
In the following sections, the connection between the springs and the piezoelectric stacks will
be considered as in series. The electromechanical system’s governing equations will be derived
for both the half and quarter car models using Laplace transformation.
One of the factors which significantly impacts the harvested power voltage is the location and
connection type of the piezoelectric stack in the suspension system. The series connection
152
between the suspension spring and the piezoelectric material has been preferred in this study
for it will not affect the mechanical properties and performances of the system. The equivalent
system stiffness is increased and the ride comfort is negatively affected in case of a parallel
connection between the suspension spring and the piezoelectric stack as shown by Equation
(3.12) and Equation (3.13).
Figure 3.4: Series connection between the suspension spring and the piezoelectric stack.
𝐾𝑠 + 𝐾𝑝𝑖𝑒𝑧𝑜
𝐾𝑒𝑞 = ≈ 𝐾𝑠 (3.12)
𝐾𝑠 × 𝐾𝑝𝑖𝑒𝑧𝑜
Figure 3.5: Parallel connection between the suspension spring and the piezoelectric stack.
𝐾𝑒𝑞 = 𝐾𝑠 + 𝐾𝑝 (3.13)
The piezoelectric element can be modeled as the resorting force in the piezoelectric material
and its output current 𝑖𝑝 by neglecting the effects of piezoelectric damping. This resorting force
is made of a mechanical component and an electrical one. The electric circuit plays the role of
collection and storage of the harvested energy. In the circuit, the piezoelectric stack can be
153
modeled as a resistance Rp and a capacitance Cp. The output AC voltage Vp is provided to the
electric circuit as depicted in Figure 3.6.
The relationships between the mechanical and electrical variables of the piezoelectric stack are
given by the following expressions:
𝐸𝑝 𝐴𝑝
𝐾𝑝 = (3.16)
𝑙𝑝
𝐴𝑝
𝛼 = 𝑑33 𝐸𝑝 ℎ (3.17)
𝑝
𝜀33 𝐴𝑝
𝐶𝑝 = 𝑁 (3.18)
ℎ𝑝
𝑙𝑝 = 𝑁ℎ𝑝 (3.19)
𝑣
𝐶𝑒 𝑣̇ + 𝑖𝑓 𝑣𝑝 = 𝑣
𝑅𝑒
𝑣
𝑖𝑝 = − 𝐶𝑒 𝑣̇ − 𝑖𝑓 𝑣𝑝 = −𝑣 (3.20)
𝑅𝑒
{ 0 𝑖𝑓 𝑣𝑝 = 0
As shown in Figure 3.6, the connection type between the resistors and capacitors is parallel.
The external load resistance is negligible in comparison with piezoelectric resistance (Rp >>
Re). In a parallel connection of resistors, the equivalent resistance is smaller than any individual
resistance. Therefore, the overall equivalent resistance R is almost equal to the external load
resistance (Req ≈ Re). Moreover, the piezoelectric capacitance Cp is negligible as compared to
the smoothing capacitor Ce (Cp << Ce). Unlike resistors, capacitors when placed in parallel
provide an equivalent capacity which is the sum of all the capacitors present in the parallel
circuit. The equivalent capacitance is almost equal to the smoothing capacitance (Ceq ≈ Ce). So,
it can be assumed that the DC rectified voltage is constant over a cycle [159, 160]. As additional
assumptions, the four diodes exhibit the ideal behavior (the diode drop is negligible) and the
155
external load is modeled as a constant current source. The equivalent electric circuit is
represented in Figure 3.7:
𝑉
= 𝛼𝑦̇ (𝑡) − 𝐶𝑝 𝑣̇ (𝑡) (3.21)
𝑅
Figure 3.7: Equivalent electric circuit to the piezoelectric stack and electric circuit set.
E33 53 (GPa)
156
The quarter car model with the piezoelectric stack represents quarterof the car, a tire alongside
the piezoelectric harvesting system which is connected in series with the suspension spring as
depicted in Figure 3.8.
𝑌𝑠 = 𝐵1 𝑌𝑢𝑠 + 𝐵2 𝑌𝑝 + 𝐵3 𝑉𝑝 (3.24)
Where:
𝑏 𝑠
𝐵1 = [𝑀 𝑠2 +𝑏𝑠 𝑠+𝐾 ] (3.25)
𝑠 𝑠 𝑝
𝐾𝑝
𝐵2 = [𝑀 𝑠2 +𝑏 𝑠+𝐾 ] (3.26)
𝑠 𝑠 𝑝
𝛼
𝐵3 = [𝑀 𝑠2 +𝑏 𝑠+𝐾 ] (3.27)
𝑠 𝑠 𝑝
𝑀𝑢𝑠 𝑦̈𝑢𝑠 = − 𝐾𝑢𝑠 (𝑦𝑢𝑠 − 𝑦𝑅 ) − 𝑏𝑢𝑠 (𝑦̇𝑢𝑠 − 𝑦̇ 𝑅 ) + 𝐾𝑠 (𝑦𝑝 − 𝑦𝑢𝑠 ) + 𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 ) (3.28)
[𝑀𝑢𝑠 𝑠 2 + (𝑏𝑢𝑠 + 𝑏𝑠 )𝑠 + (𝐾𝑢𝑠 + 𝐾𝑠 )]𝑌𝑢𝑠 = (𝑏𝑢𝑠 𝑠 + 𝐾𝑢𝑠 )𝑌𝑅 + (𝑏𝑠 𝑠)𝑌𝑠 + (𝐾𝑠 )𝑌𝑝 (3.29)
𝑌𝑢𝑠 = 𝐵4 𝑌𝑅 + 𝐵5 𝑌𝑠 + 𝐵6 𝑌𝑝 (3.30)
Where:
𝑏𝑢𝑠 𝑠+𝐾𝑢𝑠
𝐵4 = [𝑀 2 ] (3.31)
𝑢𝑠 𝑠 + 𝑢𝑠 +𝑏𝑠 )𝑠+(𝐾𝑢𝑠 +𝐾𝑠 )
(𝑏
𝑏𝑠 𝑠
𝐵5 = [𝑀 2 ] (3.32)
𝑢𝑠 𝑠 + (𝑏𝑢𝑠 +𝑏𝑠 )𝑠+(𝐾𝑢𝑠 +𝐾𝑠 )
𝐾𝑠
𝐵6 = [𝑀 2
] (3.33)
𝑢𝑠 𝑠 + 𝑢𝑠 +𝑏𝑠 )𝑠+(𝐾𝑢𝑠 +𝐾𝑠 )
(𝑏
𝑌𝑝 = 𝐵7 𝑌𝑢𝑠 + 𝐵8 𝑌𝑠 + 𝐵9 𝑉𝑝 (3.36)
Where:
𝐾𝑠
𝐵7 = 𝑀 2
(3.37)
𝑝 𝑠 +𝐾𝑠 +𝐾𝑝
𝐾𝑝
𝐵8 = (3.38)
𝑀𝑝 𝑠2 +𝐾𝑠 +𝐾𝑝
𝛼
𝐵9 = 𝑀 2
(3.39)
𝑝 𝑠 +𝐾𝑠 +𝐾𝑝
The equivalent electric system’s governing equation (3.21) can be rewritten as:
Where:
𝛼𝑅𝑠
𝐵10 = [𝐶𝑅𝑠+1] (3.43)
𝑉2
𝑃= (3.44)
𝑅
The half-car model with piezoelectric stack represents piezoelectric stacks incorporated in the
half car, with a front tire and a rear tire as depicted in Figure 3.10.
[𝑀𝑢𝑓 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑢𝑓 )𝑠 + (𝐾𝑠𝑓 + 𝐾𝑢𝑓 )]𝑌𝑢𝑓 = (𝑏𝑢𝑓 𝑠 + 𝐾𝑢𝑓 )𝑌𝑅𝑓 + (𝑏𝑠𝑓 𝑠)𝑌𝑠 −
(𝑏𝑠𝑓 𝐿𝑓 𝑠)Θ𝑠 + (𝐾𝑠𝑓 )𝑌𝑝𝑓 (3.46)
160
The vertical displacement of the front unsprung mass can be written as:
Where:
𝑏𝑢𝑓 𝑠+𝐾𝑢𝑓
𝐵11 = [𝑀 2
] (3.48)
𝑢𝑓 𝑠 +(𝑏𝑠𝑓 +𝑏𝑢𝑓 )𝑠+(𝐾𝑠𝑓 +𝐾𝑢𝑓 )
𝑏𝑠𝑓 𝑠
𝐵12 = [𝑀 ] (3.49)
𝑢𝑓 𝑠2 +(𝑏 𝑠𝑓 +𝑏𝑢𝑓 )𝑠+(𝐾𝑠𝑓 +𝐾𝑢𝑓 )
𝑏𝑠𝑓 𝐿𝑓 𝑠
𝐵13 = [𝑀 ] (3.50)
𝑢𝑓 𝑠2 +(𝑏 𝑠𝑓 +𝑏𝑢𝑓 )𝑠+(𝐾𝑠𝑓 +𝐾𝑢𝑓 )
𝐾𝑠𝑓
𝐵13 = [𝑀 2
] (3.51)
𝑢𝑓 𝑠 +(𝑏𝑠𝑓 +𝑏𝑢𝑓 )𝑠+(𝐾𝑠𝑓 +𝐾𝑢𝑓 )
𝑀𝑢𝑟 𝑦̈𝑢𝑟 = −𝑏𝑢𝑟 (𝑦̇𝑢𝑟 − 𝑦̇ 𝑅𝑟 ) − 𝐾𝑢𝑟 (𝑦𝑢𝑟 − 𝑦𝑅𝑟 ) + 𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) + 𝐾𝑠𝑟 (𝑦𝑝𝑟 − 𝑦𝑢𝑟 )
(3.52)
[𝑀𝑢𝑟 𝑠 2 + (𝑏𝑠𝑟 + 𝑏𝑢𝑟 )𝑠 + (𝐾𝑠𝑟 + 𝐾𝑢𝑟 )]𝑌𝑢𝑟 = (𝑏𝑢𝑟 𝑠 + 𝐾𝑢𝑟 )𝑌𝑅𝑟 + (𝑏𝑠𝑟 𝑠)𝑌𝑠 + (𝑏𝑠𝑟 𝐿𝑟 𝑠)Θ𝑠 +
(𝐾𝑠𝑟 )𝑌𝑝𝑟 (3.53)
The vertical displacement of the rear unsprung mass can be written as:
Where:
𝑏𝑢𝑟 𝑠+𝐾𝑢𝑟
𝐵15 = [𝑀 2
] (3.55)
𝑢𝑟 𝑠 +(𝑏𝑠𝑟 +𝑏𝑢𝑟 )𝑠+(𝐾𝑠𝑟 +𝐾𝑢𝑟
161
𝑏𝑠𝑟 𝑠
𝐵16 = [𝑀 2
] (3.56)
𝑢𝑟 𝑠 +(𝑏𝑠𝑟 +𝑏𝑢𝑟 )𝑠+(𝐾𝑠𝑟 +𝐾𝑢𝑟
𝑏𝑠𝑟 𝐿𝑟 𝑠
𝐵17 = [𝑀 2
] (3.57)
𝑢𝑟 𝑠 +(𝑏𝑠𝑟 +𝑏𝑢𝑟 )𝑠+(𝐾𝑠𝑟 +𝐾𝑢𝑟
𝐾𝑠𝑟
𝐵18 = [𝑀 2 ] (3.58)
𝑢𝑟 𝑠 +(𝑏𝑠𝑟 +𝑏𝑢𝑟 )𝑠+(𝐾𝑠𝑟 +𝐾𝑢𝑟
𝑀𝑠 𝑦̈𝑠 = −𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) − 𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) − 𝐾𝑝𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑝𝑓 ) −
𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) − (𝛼)𝑣𝑝𝑓 − (𝛼)𝑣𝑝𝑟 (3.59)
[𝑀𝑠 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑠𝑟 )𝑠 + (𝐾𝑝𝑓 + 𝐾𝑝𝑟 )]𝑌𝑠 = [(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 + (𝑘𝑝𝑓 𝐿𝑓 − 𝐾𝑝𝑟 𝐿𝑟 )]Θ𝑠 +
(𝑏𝑠𝑓 𝑠)𝑌𝑢𝑓 + (𝑏𝑠𝑟 𝑠)𝑌𝑢𝑟 + (𝐾𝑝𝑓 𝑠)𝑌𝑝𝑓 + (𝐾𝑝𝑟 𝑠)𝑌𝑝𝑟 − (𝛼)𝑉𝑝𝑓 − (𝛼)𝑉𝑝𝑟 (3.60)
𝑌𝑠 = 𝐵19 Θ𝑠 + 𝐵20 𝑌𝑢𝑓 + 𝐵21 𝑌𝑢𝑟 + 𝐵22 𝑌𝑝𝑓 + 𝐵23 𝑌𝑝𝑟 − 𝐵24 𝑉𝑝𝑓 − 𝐵24 𝑉𝑝𝑟 (3.61)
Where:
𝑏𝑠𝑓 𝑠
𝐵20 = [𝑀 𝑠2 +(𝑏 ] (3.63)
𝑠 𝑠𝑓 +𝑏𝑠𝑟 )𝑠+(𝐾𝑝𝑓 +𝐾𝑝𝑟 )
𝑏𝑠𝑟 𝑠
𝐵21 = [𝑀 𝑠2 +(𝑏 ] (3.64)
𝑠 𝑠𝑓 +𝑏𝑠𝑟 )𝑠+(𝐾𝑝𝑓 +𝐾𝑝𝑟 )
𝐾𝑝𝑓 𝑠
𝐵22 = [𝑀 𝑠2 +(𝑏 ] (3.65)
𝑠 𝑠𝑓 +𝑏𝑠𝑟 )𝑠+(𝐾𝑝𝑓 +𝐾𝑝𝑟 )
𝐾𝑝𝑟 𝑠
𝐵23 = [𝑀 𝑠2 +(𝑏 ] (3.66)
𝑠 𝑠𝑓 +𝑏𝑠𝑟 )𝑠+(𝐾𝑝𝑓 +𝐾𝑝𝑟 )
𝛼
𝐵24 = [𝑀 𝑠2 +(𝑏 ] (3.67)
𝑠 𝑠𝑓 +𝑏𝑠𝑟 )𝑠+(𝐾𝑝𝑓 +𝐾𝑝𝑟 )
162
𝐼𝑠 𝜃̈𝑠 = 𝐿𝑓 (𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) + 𝐾𝑝𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑝𝑓 ) + 𝛼𝑣𝑝𝑓 ) − 𝐿𝑟 (𝑏𝑠𝑟 (𝑦̇𝑠 +
𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) + 𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) + 𝛼𝑣𝑝𝑟 ) (3.68)
[𝐼𝑠 𝑠 2 + (𝑏𝑠𝑓 𝐿2𝑓 + 𝑏𝑠𝑟 𝐿2𝑟 )𝑠 + (𝐾𝑝𝑓 𝐿2𝑓 + 𝐾𝑝𝑟 𝐿2𝑟 )]Θ𝑠 = [(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 + (𝑘𝑝𝑓 𝐿𝑓 −
𝐾𝑝𝑟 𝐿𝑟 )]𝑌𝑠 − (𝑏𝑠𝑓 𝐿𝑓 𝑠)𝑌𝑢𝑓 + (𝑏𝑠𝑟 𝐿𝑟 𝑠)𝑌𝑢𝑟 − (𝐾𝑝𝑓 𝐿𝑓 𝑠)𝑌𝑝𝑓 + (𝐾𝑝𝑟 𝐿𝑟 𝑠)𝑌𝑝𝑟 + (𝛼𝐿𝑓 )𝑉𝑝𝑓 −
(𝛼𝐿𝑟 )𝑉𝑝𝑟 (3.69)
Θ𝑠 = 𝐵25 𝑌𝑠 − 𝐵26 𝑌𝑢𝑓 + 𝐵27 𝑌𝑢𝑟 − 𝐵28 𝑌𝑝𝑓 + 𝐵29 𝑌𝑝𝑟 + 𝐵30 𝑉𝑝𝑓 − 𝐵31 𝑉𝑝𝑟 (3.70)
Where:
𝑏𝑠𝑓 𝐿𝑓 𝑠
𝐵26 = [ 2 2 2 2
] (3.72)
𝐼𝑠 𝑠2 +(𝑏 𝑠𝑓 𝐿𝑓 +𝑏𝑠𝑟 𝐿𝑟 )𝑠+(𝐾𝑝𝑓 𝐿𝑓 +𝐾𝑝𝑟 𝐿𝑟 )
𝑏𝑠𝑟 𝐿𝑟 𝑠
𝐵27 = [ 2 2 2 2
] (3.73)
𝐼𝑠 𝑠2 +(𝑏 𝑠𝑓 𝐿𝑓 +𝑏𝑠𝑟 𝐿𝑟 )𝑠+(𝐾𝑝𝑓 𝐿𝑓 +𝐾𝑝𝑟 𝐿𝑟 )
𝐾𝑝𝑓 𝐿𝑓 𝑠
𝐵28 = [ ] (3.74)
𝐼𝑠 𝑠2 +(𝑏𝑠𝑓 𝐿2𝑓 +𝑏𝑠𝑟 𝐿2𝑟 )𝑠+(𝐾𝑝𝑓 𝐿2𝑓 +𝐾𝑝𝑟 𝐿2𝑟 )
𝐾𝑝𝑟 𝐿𝑟 𝑠
𝐵29 = [ ] (3.75)
𝐼𝑠 𝑠2 +(𝑏𝑠𝑓 𝐿𝑓 +𝑏𝑠𝑟 𝐿2𝑟 )𝑠+(𝐾𝑝𝑓 𝐿2𝑓 +𝐾𝑝𝑟 𝐿2𝑟 )
2
𝛼𝐿𝑓
𝐵30 = [ 2 2 2 2
] (3.76)
𝐼𝑠 𝑠2 +(𝑏 𝑠𝑓 𝐿𝑓 +𝑏𝑠𝑟 𝐿𝑟 )𝑠+(𝐾𝑝𝑓 𝐿𝑓 +𝐾𝑝𝑟 𝐿𝑟 )
𝛼𝐿𝑟
𝐵31 = [ 2 2 2 2
] (3.77)
𝐼𝑠 𝑠2 +(𝑏 𝑠𝑓 𝐿𝑓 +𝑏𝑠𝑟 𝐿𝑟 )𝑠+(𝐾𝑝𝑓 𝐿𝑓 +𝐾𝑝𝑟 𝐿𝑟 )
163
𝑀𝑝𝑟 𝑦̈𝑝𝑟 = −𝐾𝑠𝑟 (𝑦𝑝𝑟 − 𝑦𝑢𝑟 ) + 𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) + (𝛼)𝑣𝑝𝑟 (3.78)
[𝑀𝑝𝑟 𝑠 2 + (𝐾𝑠𝑟 + 𝐾𝑝𝑟 )]𝑌𝑝𝑟 = (𝐾𝑠𝑟 )𝑌𝑢𝑟 + (𝐾𝑝𝑟 )𝑌𝑠 + (𝑘𝑝𝑟 𝐿𝑟 )Θ𝑠 + (𝛼)𝑉𝑝𝑟 (3.79)
The vertical displacement of the rear piezoelectric stack can be written as:
Where:
𝐾𝑠𝑟
𝐵32 = [𝑀 2 ] (3.81)
𝑝𝑟 𝑠 +(𝐾𝑠𝑟 +𝐾𝑝𝑟 )
𝐾𝑝𝑟
𝐵33 = [𝑀 2 ] (3.82)
𝑝𝑟 𝑠 +(𝐾𝑠𝑟 +𝐾𝑝𝑟)
𝑘𝑝𝑟 𝐿𝑟
𝐵34 = [ ] (3.83)
𝑀𝑝𝑟 𝑠2 +(𝐾𝑠𝑟 +𝐾𝑝𝑟)
𝛼
𝐵35 = [𝑀 2
] (3.84)
𝑝𝑟 𝑠 +(𝐾𝑠𝑟 +𝐾𝑝𝑟)
[𝑀𝑝𝑓 𝑠 2 + (𝐾𝑠𝑓 + 𝐾𝑝𝑓 )]𝑌𝑝𝑓 = (𝐾𝑠𝑓 )𝑌𝑢𝑓 + (𝐾𝑝𝑓 )𝑌𝑠 + (𝑘𝑝𝑓 𝐿𝑓 )Θ𝑠 + (𝛼)𝑉𝑝𝑓 (3.86)
The vertical displacement of the front piezoelectric stack can be written as:
Where:
164
𝐾𝑠𝑓
𝐵36 = [𝑀 2
] (3.88)
𝑝𝑓 𝑠 +(𝐾𝑠𝑓 +𝐾𝑝𝑓
𝐾𝑝𝑓
𝐵37 = [𝑀 2
] (3.89)
𝑝𝑓 𝑠 +(𝐾𝑠𝑓 +𝐾𝑝𝑓
𝑘𝑝𝑓 𝐿𝑓
𝐵38 = [𝑀 2
] (3.90)
𝑝𝑓 𝑠 +(𝐾𝑠𝑓 +𝐾𝑝𝑓
𝛼
𝐵39 = [𝑀 2 ] (3.91)
𝑝𝑓 𝑠 +(𝐾𝑠𝑓 +𝐾𝑝𝑓
The constitutive equation of the rear suspension’s equivalent electric system can be deduced
from Equation (3.21):
Similarly, the constitutive equation of the front suspension equivalent electric system can be
deduced from Equation (3.21):
After the application of Laplace transformation to both rear and front equations:
𝑉𝑟2
𝑃𝑟 = (3.95)
𝑅
𝑉𝑓2
𝑃𝑓 = (3.97)
𝑅
Where:
𝛼𝑅𝑠
𝐵40 = [𝐶𝑅𝑠+1] (3.98)
𝑟 𝐿 𝛼𝑅𝑠
𝐵41 = [𝐶𝑅𝑠+1] (3.99)
165
𝑓𝐿 𝛼𝑅𝑠
𝐵42 = [𝐶𝑅𝑠+1] (3.100)
In the following section, simulation will be performed using the equation of motions and their
Laplace equivalents in order to study the behaviours of both the quarter and half truck models
4 3
𝑀𝑢𝑠 𝑦̈𝑢𝑠 = − 𝐾𝑢𝑠 (𝑦𝑢𝑠 − 𝑦𝑅 ) − 𝑏𝑢𝑠 (𝑦̇𝑢𝑠 − 𝑦̇ 𝑅 ) + 𝐾4 (𝑦𝑝 − 𝑦𝑢𝑠 ) + 𝐾3 (𝑦𝑝 − 𝑦𝑢𝑠 ) + 𝐾2 (𝑦𝑝 −
2
𝑦𝑢𝑠 ) + 𝐾1 (𝑦𝑝 − 𝑦𝑢𝑠 ) + 𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 ) (3.102)
Where,
𝐾4 = 136150000 𝑁/𝑚4
𝐾3 = −24800000 𝑁/𝑚3
𝐾2 = −520000 𝑁/𝑚2
𝐾1 = 2320000 𝑁/𝑚
Equation of the piezoelectric element:
𝑀𝑝 𝑦̈𝑝 = −𝐾4 (𝑦𝑠 − 𝑦𝑢𝑠 )4 − 𝐾3 (𝑦𝑠 − 𝑦𝑢𝑠 )3 − 𝐾2 (𝑦𝑠 − 𝑦𝑢𝑠 )2 − 𝐾1 (𝑦𝑠 − 𝑦𝑢𝑠 ) +𝐾𝑝 (𝑦𝑠 −
𝑦𝑝 ) + 𝛼𝑣𝑝 (3.103)
The equivalent electric system’s governing equation (3.21) can be rewritten as:
166
𝑣2
𝑃= (3.105)
𝑅
167
4. RESULTS
The piezoelectric ceramics can be used in several applications such as sensors (direct effect)
and actuators (converse effect). Their operation at the system resonant frequencies generates
the maximum amount of strain. Lead Zirconate Titanate (PZT) is one of the most common
piezo ceramics. For this simulation, PZT-5H was used due to its availability, economic cost,
and the high piezoelectric constant d33 value, availability, and low cost.
In the quarter car model, the sprung mass represents a quarter of the truck car body whereas, in
the half-car model, it represents half of the 777F car body weight. The unsprung mass represents
the mass of a wheel group.
Mass 1300 g Mp
Capacitance 4,5 𝜇𝐹 Cp
Stiffness 51 MN/m Kp
Internal resistance 20 kΩ Rp
168
Table 4.2: Quarter car model parameters (777F Off-highway Mining Dump Truck).
Parameter Value Unit Symbol
Capacitance 4,5 𝜇𝐹 Cp
Internal resistance 20 kΩ Rp
169
Figure 4.1: QCM Simulink representation using transfer functions of the motion equations.
170
Figure 4.2: MATLAB Simulink model representation of HCM with piezoelectric stack
A step function (or staircase function) represents constant in different intervals (pieces). It is
one of the most useful functions in the simulation field. It can simply be defined as the change
in the input from an initial value (generally zero) to a finite value at time t = a (where a is a
constant).
In this simulation, the following step function (constant 0.05m after 1s) is going to be used:
0 𝑓𝑜𝑟 0 ≤ 𝑡 < 1𝑠
Step input: 𝑦𝑟 = { (4.1)
0.05(𝑚) 𝑓𝑜𝑟 1 ≤ 𝑡 ≤ 20𝑠
171
The car traveling velocity u over a road is affected by the excitation frequency ƒ and the road
(texture wavelength) wavelength λ. The relationship between the car velocity and excitation
frequency can be expressed as:
𝑢=𝜆𝑓 (4.2)
Where 𝑢 is the car velocity[m/s]; 𝜆 is the road wave length[m] and 𝑓 is the frequency[Hz].
0 𝑓𝑜𝑟 0 ≤ 𝑡 < 1𝑠
Front tire step input: 𝑦𝑅𝑓 = { (4.3)
0.05(𝑚) 𝑓𝑜𝑟 1𝑠 ≤ 𝑡 ≤ 20𝑠
Rear tire step input: 𝑦𝑅𝑟 = 𝑦𝑅𝑓 (𝑡 − 𝜏) where 𝜏 is the time delay on the rear tire. 𝜏 =
𝐿 (𝑊ℎ𝑒𝑒𝑙𝑏𝑎𝑠𝑒)
(4.4)
𝑢 (𝑐𝑎𝑟 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦)
In this case, the truck velocity is assumed to be u= 7.145 m/s (25.722 km/h) which means that
𝐿 (𝑊ℎ𝑒𝑒𝑙𝑏𝑎𝑠𝑒) 4.56𝑚
𝜏 = 𝑢 (𝑐𝑎𝑟 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦) = 7.145 𝑚/𝑠 = 0.64𝑠. (4.5)
The harmonic response of a system at a specific frequency by means of sinusoidal loads that
vary with time helps to analyze its resistance to resonance, fatigue, and other effects under
forced vibration. In other words, it is when a system is exposed to a sinusoidal external force
of a certain frequency. Resonance is a phenomenon that happens when then the excitation
frequency equals the system's natural frequency(ies). The harmonic excitation analysis is the
perfect tool to determine the behaviors of the cars on more or less harmoniously wavy roads.
176
In harmnic exitationn, the road profile texture can be classsified according to the ISO norm
13473-1 [161]:
For this study, an uneven road texture (λ = 5m will be used for the harmonic simulations)
The bode plot represents the frequency response of a system in several disciplines such as
control theory and stability analysis. A bode plot generally consists of a magnitude and a phase
plot. The magnitude plot represents the magnitude (in absolute values or decibels) of the
frequency response while the phase plot expresses the frequency response phase shift. The bode
plot provides important information such as the gain margin and phase margin. The magnitude
plot provides relevant information such as the natural frequencies of the system.
The magnitudes represent the ratio of the unsprung mass (tire) displacement to input road
excitation and that of the sprung mass(quarter car body) to the input road excitation yr. The
QCM has 2 DoFs, consequently, it exhibits 2 resonant frequencies 1.454 Hz (suspension spring)
and 9.529 Hz (tire). These natural frequencies can be identified at the peaks of the plot in Figure
4.9.
177
The magnitudes represent the ratio of the unsprung masses (tires), sprung mass (half car body),
and sprung mass pitch displacement to the input road excitation yr. The half-car model has 4
DoF, consequently, it exhibits 4 resonant frequencies: 1.429 Hz (sprung mass), 1.004 Hz (pitch
resonant frequency), and 9.66 Hz (front and rear tire resonant frequencies). These natural
frequencies can be identified at the peaks of the plot in Figure 4.10.
178
The QCM was harmonically excited at the sprung resonant frequency with an amplitude of
0.05 m. The road input yr can be expressed as:
For excitation at the sprung natural frequency, the recorded peak value of the QCM car body
displacement is 0.3 m.
For excitation at the sprung natural frequency, the QCM, the maximum recorded harvested
voltage is 20.63V while the maximum harvested power is 0.02129 W. The root-mean-square
value of the voltage was obtained by dividing the peak value of the voltage by √2. The RMS
voltage and power are 14.59V and 0.0007294, respectively.
The unsprung masses of the HCM were harmonically excited at the sprung resonant frequency
with an amplitude of 0.05 m.
𝐿 (𝑊ℎ𝑒𝑒𝑙𝑏𝑎𝑠𝑒)
where 𝜏 is the time delay on the rear tire. 𝜏 = 𝑢 (𝑐𝑎𝑟 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦)
By assuming the road wavelength to be 5m, the sprung frequency (1.429Hz) corresponds to the
car velocity 𝑢=𝜆𝑓=5m x 1.429Hz =7.145 m/s (25.722 km/h)
𝐿 (𝑊ℎ𝑒𝑒𝑙𝑏𝑎𝑠𝑒) 4.56𝑚
𝜏 = 𝑢 (𝑐𝑎𝑟 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦) = 7.145 𝑚/𝑠 = 0.64𝑠
182
Figure 4.19: HCM harmonic excitation front stack harvested voltage plot.
183
Figure 4.20: HCM harmonic excitation front stack harvested voltage plot.
By applying a harmonic excitation to the HCM at the sprung natural frequency the maximum
harvested voltage is 24.83 V while the maximum harvested power is 0.03082. The root-mean-
square value of the voltage was calculated by dividing it by √2. The RMS voltage and power
are 17.56 V and 0.0008779 W, respectively. There is an increase of 20.36% in the RMS voltage and
power values in comparison with the QCM.
sprung natural
1.429 24.83 17.56 0.03082 0.0008779
frequency
sprung pitch
1.004 3.138 2.22 0.004562 0.0001109
natural frequency
Front unsprung
9.66 9.564 6.76 0.0004923 0.0003381
natural frequency
Rear unsprung
9.66 9.564 6.76 0.004562 0.0003381
natural frequency
The most common types of non-linearity in the suspension system are quadratic, cubic and
quartic stiffness, friction damping, clearance gaps, and backlash of the components. The quartic
non-linearity of the suspension spring is discussed in this study. Theoretically, the behavior of
spring is supposed to be linear. In practice, suspension springs display a quasi-linear behavior
for relatively small and medium loads. When the loads are big enough the displacement rate
drops leading to the non-linear behavior in suspension springs as depicted in the example in
Figure 4.21.
Figure 4.21: Comparison between linear and non-linear springs force-displacement curve.
Consequently, the piezoelectric stack is subject to less stress and generates less power. The car
body and piezoelectric stack in the non-linear model decrease as depicted in Figure 4.22. This
leads to an electric output lower than that of the linear model. Despite decreasing the output
185
harvesting performances, the non-linear model improves the ride quality and comfort (the car
body vibration amplitude is lower for the non-linear model than the linear one).
Figure 4.22: Bode amplitude plot of the proposed non-linear model (cardody dysplacement to
harmonic excitation ratio).
The non-linear model was proposed using experimental data from the literature and MATLAB
polynomial fitting iterative tools. The spring force in the proposed non-linear model can be
expressed as:
Where,
𝐾4 = 136150000 𝑁/𝑚4
𝐾3 = −24800000 𝑁/𝑚3
𝐾2 = −520000 𝑁/𝑚2
𝐾1 = 2320000 𝑁/𝑚
186
Several signal characteristics may not be observed or visible using time-domain analysis.
Thanks to the frequency domain, these hidden characteristics can be easily retrieved. For
example, the frequency domain analysis allows us to easily get the cyclic behavior of a signal.
Furthermore, in the frequency domain, it is very easy to design and adjust the system. In
addition to that, even if the system is nonlinear, the frequency domain analysis can still be
carried out. It is simple, accurate, and very suitable for measuring the system's sensitivity to
noise and parameter variation. Last but not least, the analytical study of higher-order differential
equations is complex. Hence the frequency domain represents a better alternative.
1
𝑦̈𝑠 = 𝑀 [−𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 ) −𝐾𝑝 (𝑦𝑠 − 𝑦𝑝 ) − 𝛼𝑣𝑝 ] (4.9)
𝑠
188
1
𝑦̈𝑢𝑠 = 𝑀 [− 𝐾𝑢𝑠 (𝑦𝑢𝑠 − 𝑦𝑅 ) − 𝑏𝑢𝑠 (𝑦̇𝑢𝑠 − 𝑦̇ 𝑅 ) + 𝐾𝑠 (𝑦𝑝 − 𝑦𝑢𝑠 ) + 𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 )] (4.10)
𝑢𝑠
1
𝑦̈𝑝 = 𝑀 [−𝐾𝑠 (𝑦𝑝 − 𝑦𝑢𝑠 ) +𝐾𝑝 (𝑦𝑠 − 𝑦𝑝 ) + 𝛼𝑣𝑝 ] (4.11)
𝑝
𝛼 1
𝑣 = 𝛼𝑅(𝑦̇𝑠 − 𝑦̇𝑝 ) − 𝐶𝑅𝑣̇ ⟺ 𝑣̇ = (𝑦̇𝑠 − 𝑦̇𝑝 ) − 𝑣 (4.12)
𝐶 𝐶𝑅
𝑀𝑢𝑓 𝑦̈ 𝑢𝑓 = −𝑏𝑢𝑓 (𝑦̇ 𝑢𝑓 − 𝑦̇ 𝑅𝑓 ) − 𝐾𝑢𝑓 (𝑦𝑢𝑓 − 𝑦𝑅𝑓 ) + 𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) + 𝐾𝑠𝑓 (𝑦𝑝𝑓 − 𝑦𝑢𝑓 )
(4.13)
𝑀𝑢𝑟 𝑦̈𝑢𝑟 = −𝑏𝑢𝑟 (𝑦̇𝑢𝑟 − 𝑦̇ 𝑅𝑟 ) − 𝐾𝑢𝑟 (𝑦𝑢𝑟 − 𝑦𝑅𝑟 ) + 𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) + 𝐾𝑠𝑟 (𝑦𝑝𝑟 − 𝑦𝑢𝑟 )
(4.14)
𝑀𝑝𝑟 𝑦̈𝑝𝑟 = −𝐾𝑠𝑟 (𝑦𝑝𝑟 − 𝑦𝑢𝑟 ) + 𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) + (𝛼)𝑣𝑝𝑟 (4.18)
𝑦𝑠 represents the vertical displacement of the half sprung mass (half carbody),
𝑦𝑢𝑓 represents the vertical displacement of the front unsprung mass (front tire),
𝑦𝑢𝑟 represents the vertical displacement of the rear unsprung mass (rear tire),
𝑦𝑅𝑓 represents the vertical excitation of the road at the front of the car,
𝑦𝑅𝑟 represents the vertical excitation of the road at the back of the car,
𝑀𝑢𝑠 𝑦̈𝑢𝑠 = 𝐾𝑠 (𝑦𝑠 − 𝑦𝑢𝑠 ) + 𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 ) − 𝐾𝑢𝑠 (𝑦𝑢𝑠 − 𝑦𝑅 ) − 𝑏𝑢𝑠 (𝑦̇𝑢𝑠 − 𝑦̇ 𝑅 ) (4.19)
By applying Laplace transform, we can swap the variables from the time domain to the s-
domain (also known as the frequency domain).
[𝑀𝑢𝑠 𝑠 2 + (𝑏𝑢𝑠 + 𝑏𝑠 )𝑠 + (𝐾𝑢𝑠 + 𝐾𝑠 )]𝑌𝑢𝑠 = (𝑏𝑢𝑠 𝑠 + 𝐾𝑢𝑠 )𝑌𝑅 + (𝑏𝑠 𝑠 + 𝐾𝑠 )𝑌𝑠 (4.20)
190
𝑌 𝑀 0 𝑏 + 𝑏𝑠 −𝑏𝑠 𝐾 + 𝐾𝑠 −𝐾𝑠
𝑌 = ( 𝑢𝑠 ) ; 𝑀 = ( 𝑢𝑠 ) ; 𝐵 = ( 𝑢𝑠 ) ; 𝐾 = ( 𝑢𝑠 );𝐹 =
𝑌𝑠 0 𝑀𝑠 −𝑏𝑠 𝑏𝑠 −𝐾𝑠 𝐾𝑠
(𝑏𝑢𝑠 𝑠 + 𝐾𝑢𝑠 )𝑌𝑅
( ) (4.23)
0
Where:
𝑀𝑢𝑓 𝑦̈ 𝑢𝑓 = 𝐾𝑢𝑓 (𝑦𝑢𝑓 − 𝑦𝑅𝑓 ) − 𝑏𝑢𝑓 (𝑦̇ 𝑢𝑓 − 𝑦̇ 𝑅𝑓 ) + 𝐾𝑠𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑢𝑓 ) + 𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 −
𝑦̇ 𝑢𝑓 ) (4.25)
By applying Laplace transform, we can swap the variables from the time domain to the s-
domain (also known as the frequency domain).
191
[𝑀𝑢𝑓 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑢𝑓 )𝑠 + (𝐾𝑠𝑓 + 𝐾𝑢𝑓 )]𝑌𝑢𝑓 = (𝑏𝑢𝑓 𝑠 + 𝐾𝑢𝑓 )𝑌𝑅𝑓 + (𝑏𝑠𝑓 𝑠 + 𝐾𝑠𝑓 )𝑌𝑠 +
(−𝑏𝑠𝑓 𝐿𝑓 𝑠 − 𝐾𝑠𝑓 𝐿𝑓 )Θ𝑠 (4.26)
𝑀𝑢𝑟 𝑦̈𝑢𝑟 = 𝐾𝑢𝑟 (𝑦𝑢𝑟 − 𝑦𝑅𝑟 ) − 𝑏𝑢𝑟 (𝑦̇𝑢𝑟 − 𝑦̇ 𝑅𝑟 ) + 𝐾𝑠𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑢𝑟 ) + 𝑏𝑠𝑟 (𝑦̇𝑠 − 𝐿𝑟 𝜃̇𝑠 −
𝑦̇𝑢𝑟 ) (4.27)
[𝑀𝑢𝑟 𝑠 2 + (𝑏𝑠𝑟 + 𝑏𝑢𝑟 )𝑠 + (𝐾𝑠𝑟 + 𝐾𝑢𝑟 )]𝑌𝑢𝑟 = (𝑏𝑢𝑟 𝑠 + 𝐾𝑢𝑟 )𝑌𝑅𝑟 + (𝑏𝑠𝑟 𝑠 + 𝐾𝑠𝑟 )𝑌𝑠 +
(𝑏𝑠𝑟 𝐿𝑟 𝑠 + 𝐾𝑠𝑟 𝐿𝑟 )Θ𝑠 (4.28)
[𝐼𝑠 𝑠 2 + (𝑏𝑠𝑓 𝐿2𝑓 + 𝑏𝑠𝑟 𝐿2𝑟 )𝑠 + (𝐾𝑠𝑓 𝐿2𝑓 + 𝐾𝑠𝑟 𝐿2𝑟 )]Θ𝑠 = [(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 + (𝑘𝑠𝑓 𝐿𝑓 −
𝑀𝑠 𝑦̈𝑠 = −𝐾𝑠𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑢𝑓 ) − 𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) − 𝐾𝑠𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑢𝑟 ) −
𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) (4.31)
[𝑀𝑠 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑠𝑟 )𝑠 + (𝐾𝑠𝑓 + 𝐾𝑠𝑟 )]𝑌𝑠 = [(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 + (𝐾𝑠𝑓 𝐿𝑓 − 𝐾𝑠𝑟 𝐿𝑟 )]Θ𝑠 +
(𝑏𝑠𝑓 𝑠 + 𝐾𝑠𝑓 )𝑌𝑢𝑓 + (𝑏𝑠𝑟 𝑠 + 𝐾𝑠𝑟 )𝑌𝑢𝑟 (4.32)
192
Where:
𝑀𝑢𝑓 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑢𝑓 )𝑠 + (𝐾𝑠𝑓 + 𝐾𝑢𝑓 ) 0 𝑏𝑠𝑓 𝐿𝑓 𝑠 + 𝐾𝑠𝑓 𝐿𝑓 −(𝑏𝑠𝑓 𝑠 + 𝐾𝑠𝑓 ) 𝑌𝑢𝑓
0 𝑀𝑢𝑟 𝑠 2 + (𝑏𝑠𝑟 + 𝑏𝑢𝑟 )𝑠 + (𝐾𝑠𝑟 + 𝐾𝑢𝑟 ) −(𝑏𝑠𝑟 𝐿𝑟 𝑠 + 𝐾𝑠𝑟 𝐿𝑟 ) −(𝑏𝑠𝑟 𝑠 + 𝐾𝑠𝑟 ) 𝑌
( 𝑢𝑟 ) =
−(−𝑏𝑠𝑓 𝐿𝑓 𝑠 − 𝐾𝑠𝑓 𝐿𝑓 ) −(𝑏𝑠𝑟 𝐿𝑟 𝑠 + 𝐾𝑠𝑟 𝐿𝑟 ) 𝐼𝑠 𝑠 2 + (𝑏𝑠𝑓 𝐿2𝑓 + 𝑏𝑠𝑟 𝐿2𝑟 )𝑠 + (𝐾𝑠𝑓 𝐿2𝑓 + 𝐾𝑠𝑟 𝐿2𝑟 ) −(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 − (𝑘𝑠𝑓 𝐿𝑓 − 𝐾𝑠𝑟 𝐿𝑟 ) Θ𝑠
( −(𝑏𝑠𝑓 𝑠 + 𝐾𝑠𝑓 ) −(𝑏𝑠𝑟 𝑠 + 𝐾𝑠𝑟 ) −(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 − (𝐾𝑠𝑓 𝐿𝑓 − 𝐾𝑠𝑟 𝐿𝑟 ) 𝑀𝑠 𝑠 + (𝑏𝑠𝑓 + 𝑏𝑠𝑟 )𝑠 + (𝐾𝑠𝑓 + 𝐾𝑠𝑟 ) ) 𝑌𝑠
2
𝑀𝑢𝑠 𝑦̈𝑢𝑠 = − 𝐾𝑢𝑠 (𝑦𝑢𝑠 − 𝑦𝑅 ) − 𝑏𝑢𝑠 (𝑦̇𝑢𝑠 − 𝑦̇ 𝑅 ) + 𝐾𝑠 (𝑦𝑝 − 𝑦𝑢𝑠 ) + 𝑏𝑠 (𝑦̇𝑠 − 𝑦̇𝑢𝑠 ) (4.35)
[𝑀𝑢𝑠 𝑠 2 + (𝑏𝑢𝑠 + 𝑏𝑠 )𝑠 + (𝐾𝑢𝑠 + 𝐾𝑠 )]𝑌𝑢𝑠 = (𝑏𝑢𝑠 𝑠 + 𝐾𝑢𝑠 )𝑌𝑅 + (𝑏𝑠 𝑠)𝑌𝑠 + (𝐾𝑠 )𝑌𝑝 (4.36)
The equivalent electric system’s governing equation (3.18) can be rewritten as:
Where:
𝑀𝑢𝑓 𝑦̈ 𝑢𝑓 = −𝑏𝑢𝑓 (𝑦̇ 𝑢𝑓 − 𝑦̇ 𝑅𝑓 ) − 𝐾𝑢𝑓 (𝑦𝑢𝑓 − 𝑦𝑅𝑓 ) + 𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) + 𝐾𝑠𝑓 (𝑦𝑝𝑓 − 𝑦𝑢𝑓 )
(4.41)
[𝑀𝑢𝑓 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑢𝑓 )𝑠 + (𝐾𝑠𝑓 + 𝐾𝑢𝑓 )]𝑌𝑢𝑓 = (𝑏𝑢𝑓 𝑠 + 𝐾𝑢𝑓 )𝑌𝑅𝑓 + (𝑏𝑠𝑓 𝑠)𝑌𝑠 − (𝑏𝑠𝑓 𝐿𝑓 𝑠)Θ𝑠 +
(𝐾𝑠𝑓 )𝑌𝑝𝑓 (4.42)
𝑀𝑢𝑟 𝑦̈𝑢𝑟 = −𝑏𝑢𝑟 (𝑦̇𝑢𝑟 − 𝑦̇ 𝑅𝑟 ) − 𝐾𝑢𝑟 (𝑦𝑢𝑟 − 𝑦𝑅𝑟 ) + 𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) + 𝐾𝑠𝑟 (𝑦𝑝𝑟 − 𝑦𝑢𝑟 )
(4.43)
[𝑀𝑢𝑟 𝑠 2 + (𝑏𝑠𝑟 + 𝑏𝑢𝑟 )𝑠 + (𝐾𝑠𝑟 + 𝐾𝑢𝑟 )]𝑌𝑢𝑟 = (𝑏𝑢𝑟 𝑠 + 𝐾𝑢𝑟 )𝑌𝑅𝑟 + (𝑏𝑠𝑟 𝑠)𝑌𝑠 + (𝑏𝑠𝑟 𝐿𝑟 𝑠)Θ𝑠 +
(𝐾𝑠𝑟 )𝑌𝑝𝑟 (4.44)
𝑀𝑠 𝑦̈𝑠 = −𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) − 𝑏𝑠𝑟 (𝑦̇𝑠 + 𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) − 𝐾𝑝𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑝𝑓 ) −
𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) − (𝛼)𝑣𝑝𝑓 − (𝛼)𝑣𝑝𝑟 (4.45)
195
[𝑀𝑠 𝑠 2 + (𝑏𝑠𝑓 + 𝑏𝑠𝑟 )𝑠 + (𝐾𝑝𝑓 + 𝐾𝑝𝑟 )]𝑌𝑠 = [(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 + (𝑘𝑝𝑓 𝐿𝑓 − 𝐾𝑝𝑟 𝐿𝑟 )]Θ𝑠 +
(𝑏𝑠𝑓 𝑠)𝑌𝑢𝑓 + (𝑏𝑠𝑟 𝑠)𝑌𝑢𝑟 + (𝐾𝑝𝑓 𝑠)𝑌𝑝𝑓 + (𝐾𝑝𝑟 𝑠)𝑌𝑝𝑟 − (𝛼)𝑉𝑝𝑓 − (𝛼)𝑉𝑝𝑟 (4.46)
𝐼𝑠 𝜃̈𝑠 = 𝐿𝑓 (𝑏𝑠𝑓 (𝑦̇𝑠 − 𝐿𝑓 𝜃̇𝑠 − 𝑦̇ 𝑢𝑓 ) + 𝐾𝑝𝑓 (𝑦𝑠 − 𝐿𝑓 𝜃𝑠 − 𝑦𝑝𝑓 ) + 𝛼𝑣𝑝𝑓 ) − 𝐿𝑟 (𝑏𝑠𝑟 (𝑦̇𝑠 +
𝐿𝑟 𝜃̇𝑠 − 𝑦̇𝑢𝑟 ) + 𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) + 𝛼𝑣𝑝𝑟 ) (4.47)
[𝐼𝑠 𝑠 2 + (𝑏𝑠𝑓 𝐿2𝑓 + 𝑏𝑠𝑟 𝐿2𝑟 )𝑠 + (𝐾𝑝𝑓 𝐿2𝑓 + 𝐾𝑝𝑟 𝐿2𝑟 )]Θ𝑠 = [(𝑏𝑠𝑓 𝐿𝑓 − 𝑏𝑠𝑟 𝐿𝑟 )𝑠 + (𝑘𝑝𝑓 𝐿𝑓 −
𝐾𝑝𝑟 𝐿𝑟 )]𝑌𝑠 − (𝑏𝑠𝑓 𝐿𝑓 𝑠)𝑌𝑢𝑓 + (𝑏𝑠𝑟 𝐿𝑟 𝑠)𝑌𝑢𝑟 − (𝐾𝑝𝑓 𝐿𝑓 𝑠)𝑌𝑝𝑓 + (𝐾𝑝𝑟 𝐿𝑟 𝑠)𝑌𝑝𝑟 + (𝛼𝐿𝑓 )𝑉𝑝𝑓 −
(𝛼𝐿𝑟 )𝑉𝑝𝑟 (4.48)
𝑀𝑝𝑟 𝑦̈𝑝𝑟 = −𝐾𝑠𝑟 (𝑦𝑝𝑟 − 𝑦𝑢𝑟 ) + 𝐾𝑝𝑟 (𝑦𝑠 + 𝐿𝑟 𝜃𝑠 − 𝑦𝑝𝑟 ) + (𝛼)𝑣𝑝𝑟 (4.49)
[𝑀𝑝𝑟 𝑠 2 + (𝐾𝑠𝑟 + 𝐾𝑝𝑟 )]𝑌𝑝𝑟 = (𝐾𝑠𝑟 )𝑌𝑢𝑟 + (𝐾𝑝𝑟 )𝑌𝑠 + (𝑘𝑝𝑟 𝐿𝑟 )Θ𝑠 + (𝛼)𝑉𝑝𝑟 (4.50)
[𝑀𝑝𝑓 𝑠 2 + (𝐾𝑠𝑓 + 𝐾𝑝𝑓 )]𝑌𝑝𝑓 = (𝐾𝑠𝑓 )𝑌𝑢𝑓 + (𝐾𝑝𝑓 )𝑌𝑠 + (𝑘𝑝𝑓 𝐿𝑓 )Θ𝑠 + (𝛼)𝑉𝑝𝑓 (4.52)
The constitutive equation of the rear suspension’s equivalent electric system can be
deduced from Equation (3.18):
Similarly, the constitutive equation of the front suspension equivalent electric system can be
deduced from Equation (3.18):
After the application of Laplace transformation to both rear and front equations:
𝑌𝑢𝑓 𝑀𝑢𝑓 0 0 0 0 0 0 0
𝑌𝑢𝑟 0 𝑀𝑢𝑟 0 0 0 0 0 0
Θ𝑠 0 0 𝐼𝑠 0 0 0 0 0
𝑌𝑠 0 0 0 𝑀𝑠 0 0 0 0
𝑌= 𝑌 ;𝑀= ;
𝑝𝑓 0 0 0 0 𝑀𝑝𝑓 0 0 0
𝑌𝑝𝑟 0 0 0 0 0 𝑀𝑝𝑟 0 0
𝑉𝑓 0 0 0 0 0 0 0 0
( 𝑉𝑟 ) ( 0 0 0 0 0 0 0 0)
Where:
4.3.7. Power Dissipated from the QCM Suspension System in the Frequency
Domain.
The potential power that can be harvested from the suspension systems is the power dissipated
by the suspensions [16]. The power dissipated by the suspension system can be expressed as
the product of the relative velocity and their damping force:
𝑏𝑠 𝜔2 (𝑦𝑠− 𝑦𝑢𝑠 )2
𝑃𝑑𝑖𝑠𝑄𝐶𝑀 = (4.57)
2
At the sprung natural frequency (1.454 Hz) the dissipated power is 1.233 .105 W and at the
unsprung natural frequency (9.529 Hz) its value is 1.112 .106 W.
198
Figure 4.25: Dissipated power from the QCM suspension in the frequency domain.
Figure 4.26: Dissipated power from the QCM suspension according to the truck velocity.
199
4.3.8. Power Dissipated from the HCM Suspension System in the Frequency
Domain.
The power dissipated by the suspension system in the half car model can be expressed as the
product of the relative velocity and their damping force:
Figure 4.27: Dissipated power from the HCM suspension in the frequency domain.
Figure 4.:28: Dissipated power from the HCM suspension according to the truck velocity.
200
At the sprung pitch natural frequency (1.004 Hz), the dissipated power is 3695 W. At the srung
natural frequency (1.429Hz) its value is 2.495.105 W. At the unsprung natural frequencies, the
dissipated power is 2.274.105 W.
The harvested voltage and power behaviors have been also studied in the frequency domain for
both models (QCM and HCM). The excitation frequency value is within the range of [0; 20Hz]
and the other car parameters were kept constant as they were. The QCM and HCM harvested
voltage and power plots in Figure 4.29 and Figure 4.30 revealed that the maximum values were
reached at the sprung resonant frequency. The following matrix representation of the Laplace
transform of the motion equations of the QCM and HCM with piezoelectric stacks can be used:
Figure 4.29: Maximum harvested voltage and power from the QCM in frequency domain plots.
201
The maximum harvested power and voltage were harvested for the QCM at the sprung resonant
frequency (1.454 Hz) and were 0.02129 W and 20.63 V, respectively. The maximum harvested
power and voltage for the HCM were at the sprung resonant frequency (1.429 Hz) and were
0.03082 and 24.83V, respectively. An increase of 20.36% was recorded for the HCM in
comparison with the QCM.
Figure 4.30: Maximum harvested voltage and power from the HCM in frequency domain plot.
The maximum harvested and voltage were also represented according to the car velocity vy
using the relationship in Equation (4.2) (𝑢=𝜆𝑓) . For the simulations, an uneven road texture (λ
= 5m will be used for the harmonic simulations). λ represents the road wavelength. The
maximum harvested power and voltage were harvested for the QCM at 26.172 km/h and were
0.02129 W and 20.63 V, respectively.
202
Figure 4.31: Maximum harvested voltage and power from the QCM according to the truck’s velocity.
203
Figure 4. 32: Maximum harvested voltage and power from the QCM within the truck’s operating
velocities.
Figure 4.33: Maximum harvested voltage and power from the HCM according to the truck’s velocity.
The maximum harvested power and voltage for the HCM were at 25.722 km/h and were
0.03082 and 24.83V, respectively.
Figure 4.34: Maximum harvested voltage and power from the QCM within the truck’s operating
velocities.
204
4.4.RANDOM EXCITATIONS
When a car moves at a constant speed, the road roughness is considered as a stationary random
process that is subjected to Gauss distribution.
Random road profile can be expressed in the space domain in terms of power spectral density
PSD using the equation [162]:
𝑛 −𝑊
𝐺(𝑛) = 𝐺(𝑛0 ) (𝑛 ) (4.59)
0
Where:
𝑛 is spatial frequency
𝑛0 is spatial reference frequency (n0=0.1 m-1 in our case),
𝐺(𝑛0 )is the road roughness coefficient [m2/m-1] which is given by the ISO 8608 norm
[162] according spatial reference frequency and the road profile grade as depicted in
Table 4.10,
𝑊 (𝑊 = 2 in this case) represents the waviness, the shortness or length of the road
wavelength. For most simulations, the waviness must be in the following interval: 1.75
≤ W ≤ 2.25 [163].
The following equation allow to convert the spatial frequency n[m-1] into time frequency 𝑓 [s-
1
]:
𝑓 =𝑢×𝑛 (4.60)
Where:
1
𝐺(𝑓) = 𝑢 𝐺(𝑛) (4.61)
205
𝑛 −𝑊 𝑛 −2
𝐺(𝑛) = 𝐺(𝑛0 ) (𝑛 ) = 𝐺(𝑛0 ) (𝑛 ) (4.62)
0 0
1 1 𝑛 −2
𝐺(𝑓) = 𝑢 𝐺(𝑛) = 𝑢 𝐺(𝑛0 ) (𝑛 ) (4.63)
0
𝑓
𝑓 =𝑢×𝑛 ⟺𝑛=𝑢
𝑓 −2
1 𝑛 −2 1 𝑢
𝐺(𝑓) = 𝑢 𝐺(𝑛0 ) (𝑛 ) = 𝑢 𝐺(𝑛0 ) ( 𝑛 ) 𝑢
= 𝐺(𝑛0 )𝑛0 2 𝑓2 (4.64)
0 0
The time derivative of Equation (4.62) is the car vertical speed PSD (power spectral density)
In case the random excitation is stationary, the road roughness in the time domain is equivalent
to filtering white noise generator.
The road random excitation of the front tire can be expressed as follows [129]:
Where:
𝑤(𝑡) is the Gaussian white noise (mean 0),
𝑓0 is the cut-off frequency,
𝑢(𝑡) is the car velocity,
206
Figure 4.35: Front tire random excitation source MATLAB SIMULINK representation.
By taking into account the lag between the front and rear tire random road excitations, the
random road excitation of the rear tire is given by the following relationship [164]:
𝑦̇ 𝑅𝑟 (𝑡) = 𝑦̇ 𝑅𝑓 (𝑡 − 𝜏) (4.67)
𝐿 (𝑊ℎ𝑒𝑒𝑙𝑏𝑎𝑠𝑒)[𝑚]
Where 𝜏 is the time delay on the rear tire(𝜏 = 𝑢 (𝑐𝑎𝑟 𝑣𝑒𝑙𝑜𝑐𝑖𝑡𝑦)[𝑚/𝑠]).
Where 𝐿 is the wheelbase (distance between the front and rear axle)
207
Figure 4.36: Rear tire random excitation source MATLAB SIMULINK representation.
Table 4.10: ISO 8608 road roughness levels for spatial reference frequency 𝒏𝟎 = 𝟎. 𝟏𝒎−𝟏 [162].
The harvested voltage and power were investigated for four road classes at a truck speed of
40km/h:
For the QCM, it is visible on the random excitation input plots that the class A road provides
the best profile and ride comfort as depicted in Figure 4.37. The amount of vibration is
minimum and as a result, the harvested power and voltage peak values are very low (0.00019W
and 1.949V respectively). The class D road presents an average ride quality with non-significant
roughness. The peak values of the harvested power and voltage are average (0.01217W and
15.6V, respectively). The class G road input plots reveal that it is a very rough road with
amplitude peaks reaching 0.8445m. The peak values of the harvested power and voltage for
class G are 0.7815W and 125 V, respectively. The road class H has proven to be the roughest
profile with peak excitation amplitude reaching 1.689 m. The peak values of the harvested
power and voltage for class H are 3.126W and 250 V, respectively.
Figure 4.41: QCM class A road random excitation harvested voltage (V).
211
Figure 4.42: QCM class A road random excitation harvested power (W).
Figure 4.43: QCM class D road random excitation harvested voltage (V).
212
Figure 4.44: QCM class D road random excitation harvested power (W).
Figure 4.45: QCM class G road random excitation harvested voltage (V).
213
Figure 4.46: QCM class G road random excitation harvested power (W).
Figure 4.47: QCM class H road random excitation harvested voltage (V).
214
Figure 4.48: QCM class G road random excitation harvested power (W).
The figures are similar for the HCM since the class A road provides the best profile and ride
comfort. Due to the low level of vibrations, the harvested power and voltage peak values are
very low (0.0002506W and 2.239V, respectively). The class D road is the average ride quality
road whose harvested power and voltage peak values are 0.1607W and 17.93V, respectively.
The class G road’s harvested power and voltage peak values are 1.027W and 143.3V,
respectively. The class H road’s recorded maximum harvested power and voltage are 4.108W
and 286.6V, respectively. An increase of approximately 31% and 14% were recorded for the
215
peak power and voltage values, respectively for the HCM, in comparison with QCM peak
values.
Figure 4.49: HCM class A road random excitation harvested voltage (V).
Figure 4.50: HCM class A road random excitation harvested power (W).
216
Figure 4.51: HCM class D road random excitation harvested voltage (V).
Figure 4.52: HCM class D road random excitation harvested power (W).
217
Figure 4.53: HCM class G road random excitation harvested voltage (V).
Figure 4.54: HCM class G road random excitation harvested power (W).
218
Figure 4.55: HCM class H road random excitation harvested voltage (V).
Figure 4.56: HCM class H road random excitation harvested power (W).
219
4.4.3. QCM class D random Excitation Harvested Power and Voltage according to
the Truck’s Velocity
The harvested voltage and power of the CAT 777F truck on a Class D road was represented
according to its velocity for the QCM. Then the plot was narrowed down to the real operating
velocities of the truck [20; 65km/h].
Figure 4.57: QCM Class D random excitation harvested power and voltage according to the truck’s
velocity.
220
Figure 4.58: QCM Class D random excitation harvested power and voltage in the truck’s real
operating velocities [20; 65km/h].
4.4.4. HCM class D random Excitation Harvested Power and Voltage according to
the Truck’s velocity.
The harvested voltage and power of the CAT 777F truck on a Class D road were represented
according to its velocity for the HCM as well. The plot was narrowed down to the real operating
velocities of the truck [20; 65km/h].
221
Figure 4.59: HCM Class D random excitation harvested power and voltage according to the truck’s
velocity.
Figure 4.60: HCM Class D random excitation harvested power and voltage in the truck’s real
operating velocities [20; 65km/h].
222
In the next section of this study, the obtained results and the effects of the road and car
parameters on the harvested power and voltage will be discussed.
𝑀𝑠 𝑦̈𝑠 = (−𝐾𝑠𝑓𝑟 −𝐾𝑠𝑓𝑙 −𝐾𝑠𝑟𝑟 −𝐾𝑠𝑟𝑙 )𝑦𝑠 + (−𝐶𝑠𝑓𝑟 −𝐶𝑠𝑓𝑙 −𝐶𝑠𝑟𝑟 −𝐶𝑠𝑟𝑙 )𝑦̇𝑠 + (𝐿1 𝐾𝑠𝑓𝑟 + 𝐿1 𝐾𝑠𝑓𝑙 −
𝐿2 𝐾𝑠𝑟𝑟 − 𝐿2 𝐾𝑠𝑟𝑙 )𝜃 + (𝐿1 𝐶𝑠𝑓𝑟 + 𝐿1 𝐶𝑠𝑓𝑙 − 𝐿2 𝐶𝑠𝑟𝑟 − 𝐿2 𝐶𝑠𝑟𝑙 )𝜃̇ + (𝐿𝑙 𝐾𝑠𝑓𝑟 − 𝐿𝑟 𝐾𝑠𝑓𝑙 − 𝐿𝑙 𝐾𝑠𝑟𝑟 +
𝐿𝑟 𝐾𝑠𝑟𝑙 )Φ + (𝐿𝑙 𝐶𝑠𝑓𝑟 − 𝐿𝑟 𝐶𝑠𝑓𝑙 + 𝐿𝑙 𝐶𝑠𝑟𝑟 − 𝐿𝑟 𝐶𝑠𝑟𝑙 )Φ̇ + (𝐾𝑠𝑓𝑟 )𝑦𝑓𝑟 + (𝐾𝑠𝑓𝑙 )𝑦𝑓𝑙 + (𝐾𝑠𝑟𝑟 )𝑦𝑟𝑟 +
(𝐾𝑠𝑟𝑙 )𝑦𝑟𝑙 + (𝐶𝑠𝑓𝑟 )𝑦̇𝑓𝑟 + (𝐶𝑠𝑓𝑙 )𝑦̇𝑓𝑙 + (𝐶𝑠𝑟𝑟 )𝑦̇𝑟𝑟 + (𝐶𝑠𝑟𝑙 )𝑦̇ 𝑟𝑙 (4.69)
𝐼𝑦 𝜃̈ = 𝐾𝑠𝑓𝑟 (𝑦𝑠 − 𝐿1 𝜃 − 𝐿𝑙 Φ − 𝑦𝑓𝑟 )𝐿1 + 𝐾𝑠𝑓𝑙 (𝑦𝑠 − 𝐿1 𝜃 + 𝐿𝑟 Φ − 𝑦𝑓𝑙 )𝐿1 − 𝐾𝑠𝑟𝑟 (𝑦𝑠 + 𝐿2 𝜃 −
𝐿𝑙 Φ − 𝑦𝑟𝑟 )𝐿2 − 𝐾𝑠𝑟𝑙 (𝑦𝑠 + 𝐿2 𝜃 + 𝐿𝑟 Φ − 𝑦𝑟𝑙 )𝐿2 − 𝐶𝑠𝑓𝑟 (𝑦̇𝑠 − 𝐿1 𝜃̇ − 𝐿𝑙 Φ̇ − 𝑦̇𝑓𝑟 )𝐿1 +
𝐶𝑠𝑓𝑙 (𝑦̇𝑠 − 𝐿1 𝜃̇ + 𝐿𝑟 Φ̇ − 𝑦̇𝑓𝑙 )𝐿1 − 𝐶𝑠𝑟𝑟 (𝑦̇𝑠 + 𝐿2 𝜃̇ − 𝐿𝑙 Φ̇ − 𝑦̇𝑟𝑟 )𝐿2 − 𝐶𝑠𝑟𝑙 (𝑦̇𝑠 + 𝐿2 𝜃̇ + 𝐿𝑟 Φ̇ −
𝑦̇ 𝑟𝑙 )𝐿2 ⟺
𝐼𝑦 𝜃̈ = (𝐿1 𝐾𝑠𝑓𝑟 + 𝐿1 𝐾𝑠𝑓𝑙 − 𝐿2 𝐾𝑠𝑟𝑟 − 𝐿2 𝐾𝑠𝑟𝑙 )𝑦𝑠 + (𝐿1 𝐶𝑠𝑓𝑟 + 𝐿1 𝐶𝑠𝑓𝑙 − 𝐿2 𝐶𝑠𝑟𝑟 − 𝐿2 𝐶𝑠𝑟𝑙 )𝑦̇𝑠 +
(−𝐿21 𝐾𝑠𝑓𝑟 − 𝐿21 𝐾𝑠𝑓𝑙 − 𝐿22 𝐾𝑠𝑟𝑟 −𝐿22 𝐾𝑠𝑟𝑙 )𝜃 + (−𝐿21 𝐶𝑠𝑓𝑟 − 𝐿21 𝐶𝑠𝑓𝑙 − 𝐿22 𝐶𝑠𝑟𝑟 −𝐿22 𝐶𝑠𝑟𝑙 )𝜃̇ +
(−𝐿1 𝐿𝑙 𝐾𝑠𝑓𝑟 + 𝐿1 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿2 𝐿𝑙 𝐾𝑠𝑟𝑟 − 𝐿2 𝐿𝑟 𝐾𝑠𝑟𝑙 )Φ + (−𝐿1 𝐿𝑙 𝐶𝑠𝑓𝑟 + 𝐿1 𝐿𝑟 𝐶𝑠𝑓𝑙 + 𝐿2 𝐿𝑙 𝐶𝑠𝑟𝑟 −
𝐿2 𝐿𝑟 𝐶𝑠𝑟𝑙 )Φ̇ + (−𝐿1 𝐾𝑠𝑓𝑟 )𝑦𝑓𝑟 + (−𝐿1 𝐾𝑠𝑓𝑙 )𝑦𝑓𝑙 + (𝐿2 𝐾𝑠𝑟𝑟 )𝑦𝑓𝑟 + (𝐿2 𝐾𝑠𝑟𝑙 )𝑦𝑟𝑙 +
(−𝐿1 𝐶𝑠𝑓𝑟 )𝑦̇𝑓𝑟 + (−𝐿1 𝐶𝑠𝑓𝑙 )𝑦̇𝑓𝑙 + (𝐿2 𝐶𝑠𝑟𝑟 )𝑦̇𝑓𝑟 + (𝐿2 𝐶𝑠𝑟𝑙 )𝑦̇ 𝑟𝑙 (4.70)
𝐼𝑥 Φ̈ = 𝐾𝑠𝑓𝑟 (𝑦𝑠 − 𝐿1 𝜃 − 𝐿𝑙 Φ − 𝑦𝑓𝑟 )𝐿𝑙 − 𝐾𝑠𝑓𝑙 (𝑦𝑠 − 𝐿1 𝜃 + 𝐿𝑟 Φ − 𝑦𝑓𝑙 )𝐿𝑟 + 𝐾𝑠𝑟𝑟 (𝑦𝑠 + 𝐿2 𝜃 −
𝐿𝑙 Φ − 𝑦𝑟𝑟 )𝐿𝑙 + 𝐾𝑠𝑟𝑙 (𝑦𝑠 + 𝐿2 𝜃 + 𝐿𝑟 Φ − 𝑦𝑟𝑙 )𝐿𝑟 + 𝐶𝑠𝑓𝑟 (𝑦̇𝑠 − 𝐿1 𝜃̇ − 𝐿𝑙 Φ̇ − 𝑦̇𝑓𝑟 )𝐿𝑙 − 𝐶𝑠𝑓𝑙 (𝑦̇𝑠 −
𝐿1 𝜃̇ + 𝐿𝑟 Φ̇ − 𝑦̇𝑓𝑙 )𝐿𝑟 + 𝐶𝑠𝑟𝑟 (𝑦̇𝑠 + 𝐿2 𝜃̇ − 𝐿𝑙 Φ̇ − 𝑦̇𝑟𝑟 )𝐿𝑙 − 𝐶𝑠𝑟𝑙 (𝑦̇𝑠 + 𝐿2 𝜃̇ + 𝐿𝑟 Φ̇ − 𝑦̇ 𝑟𝑙 )𝐿𝑟 ⟺
𝐼𝑥 Φ̈ = (𝐿𝑙 𝐾𝑠𝑓𝑟 − 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿𝑙 𝐾𝑠𝑟𝑟 − 𝐿𝑟 𝐾𝑠𝑟𝑙 )𝑦𝑠 + (𝐿𝑙 𝐶𝑠𝑓𝑟 − 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿𝑙 𝐶𝑠𝑟𝑟 − 𝐿𝑟 𝐶𝑠𝑟𝑙 )𝑦̇𝑠 +
(−𝐿1 𝐿𝑙 𝐾𝑠𝑓𝑟 + 𝐿1 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿2 𝐿𝑙 𝐾𝑠𝑟𝑟 − 𝐿2 𝐿𝑟 𝐾𝑠𝑟𝑙 )θ + (−𝐿1 𝐿𝑙 𝐶𝑠𝑓𝑟 + 𝐿1 𝐿𝑟 𝐶𝑠𝑓𝑙 + 𝐿2 𝐿𝑙 𝐶𝑠𝑟𝑟 −
𝐿2 𝐿𝑟 𝐶𝑠𝑟𝑙 )θ̇ + (−𝐿2𝑙 𝐾𝑠𝑓𝑟 − 𝐿2𝑟 𝐾𝑠𝑓𝑙 − 𝐿2𝑙 𝐾𝑠𝑟𝑟 −𝐿2𝑟 𝐾𝑠𝑟𝑙 )Φ + (−𝐿2𝑙 𝐶𝑠𝑓𝑟 − 𝐿2𝑟 𝐶𝑠𝑓𝑙 −
224
𝐿2𝑙 𝐶𝑠𝑟𝑟 −𝐿2𝑟 𝐶𝑠𝑟𝑙 )Φ̇ + (−𝐿𝑙 𝐾𝑠𝑓𝑟 )𝑦𝑓𝑟 + (𝐿𝑟 𝐾𝑠𝑓𝑙 )𝑦𝑓𝑙 + (−𝐿𝑙 𝐾𝑠𝑟𝑟 )𝑦𝑓𝑟 + (𝐿𝑟 𝐾𝑠𝑟𝑙 )𝑦𝑟𝑙 +
(−𝐿𝑙 𝐶𝑠𝑓𝑟 )𝑦̇𝑓𝑟 + (𝐿𝑟 𝐶𝑠𝑓𝑙 )𝑦̇𝑓𝑙 + (−𝐿𝑙 𝐶𝑠𝑟𝑟 )𝑦̇𝑓𝑟 + (𝐿𝑟 𝐶𝑠𝑟𝑙 )𝑦̇ 𝑟𝑙 (4.71)
𝑀𝑓𝑟 𝑦̈𝑓𝑟 = 𝐾𝑠𝑓𝑟 𝑦𝑠 + 𝐶𝑠𝑓𝑟 𝑦̇𝑠 + (−𝐿1 𝐾𝑠𝑓𝑟 )𝜃 + (−𝐿1 𝐶𝑠𝑓𝑟 )𝜃̇ + (−𝐿𝑙 𝐾𝑠𝑓𝑟 )Φ + (−𝐿𝑙 𝐶𝑠𝑓𝑟 )Φ̇ +
(−𝐾𝑠𝑓𝑟 − 𝐾𝑢𝑓𝑟 )𝑦𝑓𝑟 + (𝐶𝑠𝑓𝑟 )𝑦̇𝑓𝑟 + 𝐾𝑢𝑓𝑟 𝑦𝑅𝑓𝑟 ⟺
𝑀𝑓𝑟 𝑦̈𝑓𝑟 = 𝐾𝑠𝑓𝑟 (𝑦𝑠 − 𝐿1 𝜃 − 𝐿𝑙 Φ − 𝑦𝑓𝑟 ) − 𝐾𝑢𝑓𝑟 (𝑦𝑓𝑟 − 𝑦𝑅𝑓𝑟 ) + 𝐶𝑠𝑓𝑟 (𝑦̇𝑠 − 𝐿1 𝜃̇ − 𝐿𝑙 Φ̇ −
𝑦̇𝑓𝑟 ) (4.72)
𝑀𝑓𝑙 𝑦̈𝑓𝑙 = 𝐾𝑠𝑓𝑙 𝑦𝑠 + 𝐶𝑠𝑓𝑟 𝑦̇𝑠 + (−𝐿1 𝐾𝑠𝑓𝑙 )𝜃 + (−𝐿1 𝐶𝑠𝑓𝑙 )𝜃̇ + (𝐿𝑟 𝐾𝑠𝑓𝑙 )Φ + (𝐿𝑟 𝐶𝑠𝑓𝑙 )Φ̇ +
(−𝐾𝑠𝑓𝑙 − 𝐾𝑢𝑓𝑙 )𝑦𝑓𝑙 + (−𝐶𝑠𝑓𝑙 )𝑦̇𝑓𝑙 + 𝐾𝑢𝑓𝑙 𝑦𝑅𝑓𝑙 ⟺
𝑀𝑓𝑙 𝑦̈𝑓𝑙 = 𝐾𝑠𝑓𝑙 (𝑦𝑠 − 𝐿1 𝜃 + 𝐿𝑟 Φ − 𝑦𝑓𝑙 ) − 𝐾𝑢𝑓𝑙 (𝑦𝑓𝑙 − 𝑦𝑅𝑓𝑙 ) + 𝐶𝑠𝑓𝑙 (𝑦̇𝑠 − 𝐿1 𝜃̇ + 𝐿𝑟 Φ̇ − 𝑦̇𝑓𝑙 )
(4.73)
𝑀𝑟𝑟 𝑦̈𝑟𝑟 = 𝐾𝑠𝑟𝑟 𝑦𝑠 + 𝐶𝑠𝑟𝑟 𝑦̇𝑠 + (𝐿2 𝐾𝑠𝑟𝑟 )𝜃 + (𝐿2 𝐶𝑠𝑟𝑟 )𝜃̇ + (−𝐿𝑙 𝐾𝑠𝑟𝑟 )Φ + (−𝐿𝑙 𝐶𝑠𝑟𝑟 )Φ̇ +
(−𝐾𝑠𝑟𝑟 − 𝐾𝑢𝑟𝑟 )𝑦𝑟𝑟 + (−𝐶𝑠𝑟𝑟 )𝑦̇𝑟𝑟 + 𝐾𝑢𝑟𝑟 𝑦𝑅𝑟𝑟 ⟺
𝑀𝑟𝑟 𝑦̈𝑟𝑟 = 𝐾𝑠𝑟𝑟 (𝑦𝑠 + 𝐿2 𝜃 − 𝐿𝑙 Φ − 𝑦𝑟𝑟 ) − 𝐾𝑢𝑟𝑟 (𝑦𝑟𝑟 − 𝑦𝑅𝑟𝑟 ) + 𝐶𝑠𝑟𝑟 (𝑦̇𝑠 + 𝐿2 𝜃̇ − 𝐿𝑙 Φ̇ − 𝑦̇𝑟𝑟 )
(4.74)
𝑀𝑟𝑙 𝑦̈ 𝑟𝑙 = 𝐾𝑠𝑟𝑙 𝑦𝑠 + 𝐶𝑠𝑟𝑙 𝑦̇𝑠 + (𝐿2 𝐾𝑠𝑟𝑙 )𝜃 + (𝐿2 𝐶𝑠𝑟𝑙 )𝜃̇ + (𝐿𝑟 𝐾𝑠𝑟𝑙 )Φ + (𝐿𝑟 𝐶𝑠𝑟𝑙 )Φ̇ + (−𝐾𝑠𝑟𝑙 −
𝐾𝑢𝑟𝑙 )𝑦𝑟𝑙 + (−𝐶𝑠𝑟𝑙 )𝑦̇ 𝑟𝑙 + 𝐾𝑢𝑟𝑙 𝑦𝑅𝑟𝑙 ⟺
Where,
225
𝐿𝑙 is the distance between the left side of the car and the C.G.
𝐿𝑟 is the distance between the right side of the car and the C.G.
𝐼𝑦 is the moment of inertia about the y-axis (passing through the C.G. and parallel to the truck
axles)
𝐼𝑥 is the moment of inertia about the x-axis (passing through the C.G. and perpendicular to the
truck axles)
By substituting the variables in the seven equations of motion, their state space matrix form can
be written.
Where,
1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 1 0 0 0 0 0 0 0 0 0
0 0 0 0 0 1 0 0 0 0 0 0 0 0
[𝐶] = 0 0 0 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 1 0
[0 0 0 0 0 0 0 0 0 0 0 0 0 1]
[𝐶(1)] = [1 0 0 0 0 0 0 0 0 0 0 0 0 0] to get 𝑦𝑠
[𝐶(3)] = [0 0 1 0 0 0 0 0 0 0 0 0 0 0] to get 𝜃
[𝐶(5)] = [0 0 0 0 1 0 0 0 0 0 0 0 0 0] to get Φ
[𝐷] = [0 0 0 0 0 0 0 0 0 0 0 0 0 0]
[𝐴1 ] = [0 1 0 0 0 0 0 0 0 0 0 0 0 0]
[𝐴3 ] = [0 0 0 1 0 0 0 0 0 0 0 0 0 0]
[𝐴5 ] = [0 0 0 0 0 1 0 0 0 0 0 0 0 0]
229
[𝐴7 ] = [0 0 0 0 0 0 0 1 0 0 0 0 0 0]
[𝐴9 ] = [0 0 0 0 0 0 0 0 0 1 0 0 0 0]
[𝐴11 ] = [0 0 0 0 0 0 0 0 0 0 0 1 0 0]
[𝐴13 ] = [0 0 0 0 0 0 0 0 0 0 0 0 0 1]
230
𝑇
−𝐾𝑠𝑓𝑟 − 𝐾𝑠𝑓𝑙 − 𝐾𝑠𝑟𝑟 − 𝐾𝑠𝑟𝑙
𝑀𝑠
−𝐶𝑠𝑓𝑟 − 𝐶𝑠𝑓𝑙 − 𝐶𝑠𝑟𝑟 − 𝐶𝑠𝑟𝑙
𝑀𝑠
𝐿1 𝐾𝑠𝑓𝑟 + 𝐿1 𝐾𝑠𝑓𝑙 − 𝐿2 𝐾𝑠𝑟𝑟 − 𝐿2 𝐾𝑠𝑟𝑙
𝑀𝑠
𝐿1 𝐶𝑠𝑓𝑟 + 𝐿1 𝐶𝑠𝑓𝑙 − 𝐿2 𝐶𝑠𝑟𝑟 − 𝐿2 𝐶𝑠𝑟𝑙
𝑀𝑠
𝐿𝑙 𝐾𝑠𝑓𝑟 − 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿𝑙 𝐾𝑠𝑟𝑟 − 𝐿𝑟 𝐾𝑠𝑟𝑙
𝑀𝑠
𝐿𝑙 𝐶𝑠𝑓𝑟 − 𝐿𝑟 𝐶𝑠𝑓𝑙 + 𝐿𝑙 𝐶𝑠𝑟𝑟 − 𝐿𝑟 𝐶𝑠𝑟𝑙
𝑀𝑠
𝐾𝑠𝑓𝑟
𝑀𝑠
[𝐴2 ] =
𝐶𝑠𝑓𝑟
𝑀𝑠
𝐾𝑠𝑓𝑙
𝑀𝑠
𝐶𝑠𝑓𝑙
𝑀𝑠
𝐾𝑠𝑟𝑟
𝑀𝑠
𝐶𝑠𝑟𝑟
𝑀𝑠
𝐾𝑠𝑟𝑙
𝑀𝑠
𝐶𝑠𝑟𝑙
( 𝑀𝑠 )
231
𝑇
𝐿1 𝐾𝑠𝑓𝑟 + 𝐿1 𝐾𝑠𝑓𝑙 − 𝐿2 𝐾𝑠𝑟𝑟 − 𝐿2 𝐾𝑠𝑟𝑙
𝐼𝑦
𝐿1 𝐶𝑠𝑓𝑟 + 𝐿1 𝐶𝑠𝑓𝑙 − 𝐿2 𝐶𝑠𝑟𝑟 − 𝐿2 𝐶𝑠𝑟𝑙
𝐼𝑦
−𝐿1 2 𝐾𝑠𝑓𝑟 − 𝐿1 2 𝐾𝑠𝑓𝑙 − 𝐿2 2 𝐾𝑠𝑟𝑟 − 𝐿2 2 𝐾𝑠𝑟𝑙
𝐼𝑦
−𝐿1 2 𝐶𝑠𝑓𝑟 − 𝐿1 2 𝐶𝑠𝑓𝑙 − 𝐿2 2 𝐶𝑠𝑟𝑟 − 𝐿2 2 𝐶𝑠𝑟𝑙
𝐼𝑦
−𝐿1 𝐾𝑠𝑓𝑟
𝐼𝑦
−𝐿1 𝐶𝑠𝑓𝑟
[𝐴4 ] = 𝐼𝑦
−𝐿1 𝐾𝑠𝑓𝑙
𝐼𝑦
−𝐿1 𝐶𝑠𝑓𝑙
𝐼𝑦
𝐿2 𝐾𝑠𝑟𝑟
𝐼𝑦
𝐿2 𝐶𝑠𝑟𝑟
𝐼𝑦
𝐿2 𝐾𝑠𝑟𝑙
𝐼𝑦
𝐿2 𝐶𝑠𝑟𝑙
( 𝐼𝑦 )
232
𝑇
𝐿𝑙 𝐾𝑠𝑓𝑟 − 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿𝑙 𝐾𝑠𝑟𝑟 − 𝐿𝑟 𝐾𝑠𝑟𝑙
𝐼𝑥
𝐿𝑙 𝐶𝑠𝑓𝑟 − 𝐿𝑟 𝐶𝑠𝑓𝑙 + 𝐿𝑙 𝐶𝑠𝑟𝑟 − 𝐿𝑟 𝐶𝑠𝑟𝑙
𝐼𝑥
−𝐿1 𝐿𝑙 𝐾𝑠𝑓𝑟 + 𝐿1 𝐿𝑟 𝐾𝑠𝑓𝑙 + 𝐿2 𝐿𝑙 𝐾𝑠𝑟𝑟 − 𝐿2 𝐿𝑟 𝐾𝑠𝑟𝑙
𝐼𝑥
−𝐿1 𝐿𝑙 𝐶𝑠𝑓𝑟 + 𝐿1 𝐿𝑟 𝐶𝑠𝑓𝑙 + 𝐿2 𝐿𝑙 𝐶𝑠𝑟𝑟 − 𝐿2 𝐿𝑟 𝐶𝑠𝑟𝑙
𝐼𝑥
−𝐿𝑙 𝐾𝑠𝑓𝑟 − 𝐿𝑟 𝐾𝑠𝑓𝑙 − 𝐿𝑙 2 𝐾𝑠𝑟𝑟 − 𝐿𝑟 2 𝐾𝑠𝑟𝑙
2 2
𝐼𝑥
−𝐿𝑙 𝐶𝑠𝑓𝑟 − 𝐿𝑟 𝐶𝑠𝑓𝑙 − 𝐿𝑙 2 𝐶𝑠𝑟𝑟 − 𝐿𝑟 2 𝐶𝑠𝑟𝑙
2 2
𝐼𝑥
−𝐿𝑙 𝐾𝑠𝑓𝑟
[𝐴6 ] = 𝐼𝑥
𝐿𝑙 𝐶𝑠𝑓𝑟
𝐼𝑥
𝐿𝑟 𝐾𝑠𝑓𝑙
𝐼𝑥
𝐿𝑟 𝐶𝑠𝑓𝑙
𝐼𝑥
−𝐿𝑙 𝐾𝑠𝑟𝑟
𝐼𝑥
−𝐿𝑙 𝐶𝑠𝑟𝑟
𝐼𝑥
𝐿𝑟 𝐾𝑠𝑟𝑙
𝐼𝑦
𝐿𝑟 𝐶𝑠𝑟𝑙
( 𝐼𝑥 )
233
𝐾𝑠𝑓𝑟 𝑇 𝐾𝑠𝑓𝑙 𝑇
𝐾𝑠𝑟𝑟 𝑇 𝐾𝑠𝑟𝑙 𝑇
𝑀𝑓𝑟 𝑀𝑓𝑙
𝐶𝑠𝑓𝑟 𝐶𝑠𝑓𝑙 𝑀𝑟𝑟 𝑀𝑟𝑙
𝐶𝑠𝑟𝑟 𝐶𝑠𝑟𝑙
𝑀𝑓𝑟 𝑀𝑓𝑙
𝑀𝑟𝑟 𝑀𝑟𝑙
−𝐿1 𝐾𝑠𝑓𝑟 −𝐿1 𝐾𝑠𝑓𝑙
𝐿2 𝐾𝑠𝑟𝑟 𝐿2 𝐾𝑠𝑟𝑙
𝑀𝑓𝑟 𝑀𝑓𝑙
𝑀𝑟𝑟 𝑀𝑟𝑙
−𝐿1 𝐶𝑠𝑓𝑟 −𝐿1 𝐶𝑠𝑓𝑙 𝐿2 𝐶𝑠𝑟𝑟 𝐿2 𝐶𝑠𝑟𝑙
𝑀𝑓𝑟 𝑀𝑓𝑙 𝑀𝑟𝑟 𝑀𝑟𝑙
−𝐿𝑙 𝐾𝑠𝑓𝑟 𝐿𝑟 𝐾𝑠𝑓𝑙 −𝐿𝑙 𝐾𝑠𝑟𝑟 𝐿𝑟 𝐾𝑠𝑟𝑙
𝑀𝑓𝑟 𝑀𝑓𝑙 𝑀𝑟𝑟 𝑀𝑟𝑙
−𝐿𝑙 𝐶𝑠𝑓𝑟 𝐿𝑟 𝐶𝑠𝑓𝑙 −𝐿𝑙 𝐶𝑠𝑟𝑟 𝐿𝑟 𝐶𝑠𝑟𝑙
[𝐴8 ] = 𝑀𝑓𝑟 ; [𝐴10 ] = 𝑀𝑓𝑙 ; [𝐴12 ] = 𝑀𝑟𝑟 ; [𝐴14 ] = 𝑀𝑟𝑙
−𝐾𝑠𝑓𝑟 −𝐾𝑢𝑓𝑟 0 0 0
( ) 0 0
𝑀𝑓𝑟 0
−𝐶𝑠𝑓𝑟 −𝐾𝑠𝑓𝑙 −𝐾𝑢𝑓𝑙 0 0
( ) 0 0
𝑀𝑓𝑟 𝑀𝑓𝑙
−𝐶𝑠𝑓𝑙 (
−𝐾𝑠𝑟𝑟 −𝐾𝑢𝑟𝑟
) 0
0 𝑀𝑟𝑟
𝑀𝑓𝑙 0
0 −𝐶𝑠𝑟𝑟 −𝐾𝑠𝑟𝑙 −𝐾𝑢𝑟𝑙
0 0 𝑀𝑟𝑟 ( )
𝑀𝑟𝑙
0 0 0 −𝐶𝑠𝑟𝑙
0 0 ( 0 ) ( 𝑀𝑟𝑙 )
( 0 ) ( 0 )
The Simulink representation of the FCM sprung mass using a transfer function is represented
in Figure 4.62. The sprung mass resonant frequency was identified at the peak magnitude. Its
value is 1.49 Hz as depicted in Figure 4.63. For good ride quality, the sprung mass resonant
frequency should be less than 1.5 Hz. Then, the ride comfort requirement is met..
Figure 4.62: FCM sprung mass simulink representation using a transfer function.
234
The step road input and car body displacement plot are depicted in Figure 4.64 and Figure 4.65,
respectively.
Figure 4.65: FCM step road input sprung mass displacement plot.
The harmonic road excitation and car body displacement plots are depicted in Figure 4.66 and
Figure 4.67, respectively.
Figure 4.67: FCM harmonic road excitation sprung mass displacement response.
237
5. DISCUSSION
The sine wave acceleration amplitude has a great impact on the performance of the piezoelectric
energy harvesting system. As depicted in Figure 5.1, the harvested voltage and power increase
with the amplitude of the sine wave excitation. The power increases through a quadratic relation
since the power is the voltage square over the resistance. For this simulation, the road amplitude
varied between 0 and 0.1m. The other parameters were kept constant.
Figure 5.1: Effect of the road sine wave amplitude on the harvested voltage and power of the HCM.
Moreover, roads with a harmonic profile provide more consistent and sustainable energy as
compared to the step and random excitation input types. The maximum power and energy are
harvested at the sprung (car body) resonant frequency. The frequency will depend on the road
wavelength and the car velocity.
238
5.2.1. Effect of the Force Factor of the Piezoelectric Stack on the Harvested Power
and Voltage.
The force factor of a piezoelectric stack can be defined as the necessary amount of force to
generate a volt. As depicted in Figure 5.2, the harvested power decreases inversely with the
force factor of the piezoelectric stack.
Figure 5.2: Effects of the piezoelectric stack force factor on the harvested voltage and power.
239
5.2.2. Effect of the Capacitance of the Piezoelectric Stack on the Harvested Power
and Voltage.
The capacitance is one of the key parameters of piezoelectric materials as many of them are
made of the same ceramic materials as capacitors. The piezoelectric effect can actually be
described as the vibration of a capacitor. Figure 5.3 revealed that within a range of [0.1x10 -5;
4x10-5F], the harvested voltage and power decrease inversely with the capacitance.
Figure 5.3: Effects of the piezoelectric stack capacitance on the harvested voltage and power.
It should also be noted that the harvested voltage increases linearly with the number of layers
of the stack). As stated by Cao and Li in their research work [26], the displacement of the
piezoelectric stack can be expressed as:
1
𝑦𝑠𝑡𝑎𝑐𝑘 = 𝑁𝑦𝑝 = 𝑘 𝐹𝑝 + 𝑁𝑑33 𝑉 (5.1)
𝑝
240
This relationship shows that the increase in the number of stacks leads to the growth of the
overall displacement. Moreover, the harvested charges equal the sum of charges collected from
all individual layers of the piezoelectric stack
Where:
5.3.1. Effect of the Suspension Damping Coefficients bsf and bsr on the Harvested
Voltage and Power.
In this simulation, both suspensions (front and rear) of the half truck model varied between
3.104 and 5. 104 N.s/m. The other car and road parameters were kept constant. As depicted in
Figure 5.4, the harvested voltage decreased from 37.5V to 22V. The harvested power decreased
from 0.07W to 0.02W. This figure demonstrated that the harvested power and voltage decrease
inversely with the suspension damping coefficients as the piezoelectric stacks undergo less
stress. The lower the suspension damping the more stress applied to the piezoelectric stack and
the more power harvested. However, it should be noted that suspensions with low damping
coefficients lead to stability and maneuverability issues.
241
5.3.2. Effect of the Unsprung Damping Coefficients buf and bur on the Harvested
Voltage and Power.
During this simulation, the unsprung (tire) damping coefficients varied between 1.5.104 And
3.104 Ns/m . The other parameters of the truck were kept constant. As depicted in Figure 5.5,
the voltage decreased from 24.51 to 24.13 V while the power decreased from 0.03 to 0.0291
W. Consequently, the harvested voltage and power decrease inversely with the unsprung (tire)
damping coefficients.
242
5.3.3. Effect of the Suspension Stiffness Coefficients Ksf and Ksr on the Harvested
Voltage and Power.
During this simulation, the stiffness coefficient of the suspensions (front and rear) varied
between 1.106 And 3.106 N/m . The other parameters of the truck were kept constant. As
depicted in Figure 5.6, the maximum voltage and power are harvested at 2.75.106 N/m. From
1.106 to 2.75.106 N/m, the voltage increases from 3 to 24 V and the power from 0.001 to
0.029W. From 2.75.106 to 3.106 N/m, the voltage decreases from 24V to 16V and the power
decreases from 0.029 to 0.012W. The optimum suspension stiffness in this study is 2.75.106
N/m for maximum power harvesting.
243
Figure 5.6: Effects of the suspension stiffness coefficients on the harvested power and voltage.
5.3.4. Effect of the Unsprung (tire) Stiffness Coefficients Kuf and Kur on the
Harvested Voltage and Power.
This simulation was carried out with the unsprung stiffness coefficients varying from 0.8.107
to 1.2 .107 N/m. The other parameters of the truck and road were kept constant. The peak values
of the harvested voltage and power were reached at 0.93 .107 N/m. Between 0.8.107 and 0.93
.107 N/m, the voltage increased from 24.3V to 24.75V. The harvested power rose from 0.0295
to 0.031 W. The harvested voltage decreased from 24.75V to 23.5V between 0.93 .107 and 1.2
.107 N/m. The harvested power decreased from 0.031 to 0.0275W. In addition to that, a high
tire stiffness is an indicator of the low harvested resonance power.
244
Figure 5.7: Effects of the unsprung stiffness coefficients on the harvested power and voltage.
5.3.5. Effect of the Sprung Mass MS on the Harvested Voltage and Power.
This simulation was carried out with the sprung mass varying from 3.104 to 6.104 kg. The other
parameters of the truck and road were kept constant. The peak values of the harvested voltage
and power were reached at Ms = 4.75. 104 kg. From 3.104 to 4.75 .104 kg, the voltage increased
from 7V to 25V. The harvested power rose from 0.004 to 0.032 W. The harvested voltage
decreased from 25V to 15V between 4.75.104 and 6.104 kg. The harvested power decreased
from 0.032 to 0.013W.
245
Figure 5.8: Effects of the sprung mass on the harvested power and voltage.
5.3.6. Effect of the Unsprung Masses Muf and Mus on the Harvested Voltage and
Power.
This simulation was carried out with the front and rear unsprung masses varying from 2000 to
10000 kg. The other parameters of the truck and road were kept constant. As depicted in Figure
5.9, the harvested voltage and power grow with the unsprung masses in the interval of [2000;
10000]. The harvested voltage increased from 23.9V to 25.7V while the harvested power
increased from 0.028 to 0.033W.
246
Figure 5.9: Effects of the unsprung masses on the harvested power and voltage.
All previous results prove that there is great potential in dump truck suspension piezoelectric
energy harvesting. The road unevenness creates excitations of the tires which leads to vibrations
in the suspension system. A considerable amount of the energy produced by the car is lost this
way. The car suspension piezoelectric energy harvesting is a clean and efficient alternative to
scavenge this energy and store it in batteries or use them to power the multiple sensors present
on the truck. The road, piezoelectric stack, and car parameters have a significant impact on the
performance of the harvesting system.
247
6.1.CONCLUSION
In order to evaluate the potential of off-highway dump truck suspension system piezoelectric
energy harvesting potential, the quarter and half car models were mathematically developed for
the CAT 777G dump truck. The piezoelectric stacks were placed in series with the suspensions.
The equations of motion were derived from the different models. By using the Laplace
transforms and transfer functions, the behaviors of these models were studied in both the time
and frequency domains. The simulation for the harmonic excitation revealed that the maximum
voltage and power were harvested at the sprung (car body) resonant frequency.
In the quarter car model, a harmonic excitation of amplitude 0.05m at the QCM sprung resonant
frequency (1.454) allowed to harvest voltage and power peak values of 20.63V and 0.02129 W,
respectively. The root-mean-square values of the voltage and power were found to be 14.59V
and 0.0007294W. In the half-car model which brings more accuracy to the results of the QCM,
the application of the harmonic excitation of amplitude 0.05m and at the sprung resonant
frequency generated voltage and power peak values of 24.83V and 0.03082W, respectively.
The root-mean-square harvested voltage was found to be 17.56V. The calculated root-mean-
square value of the power was 0.0008779W. An increase of approximately 20.36% was
recorded in the harvested power and voltages of the HCM in comparison with those of the
QCM. The main differences between both models are the existence of the pitch motion and a
second unsprung (tire) mass in the HCM. These characteristics allow us to obtain results that
are closer to the actual values of a full car. The behaviors of the energy harvesting system were
also observed in the frequency domain.
The non-linear quarter car model displayed lower car body displacement than the linear one
resulting in lower electrical output performance. However, in terms of stability, car handling
and passenger comfort, the non-linear suspension spring presents more advantages than the
linear one.
By using the Laplace transform, the time domain variable t was swapped with s which is the s-
domain (or frequency domain analysis). The frequency-domain analysis presented several
advantages among which the identification of the resonant frequencies of the systems. Being a
248
DoF system, the QCM exhibited two resonant frequencies while the HCM which is a 4 DoF
system had 4 resonant frequencies. In addition to that, the frequency domain analysis also
allowed us to observe the behavior of the harvested voltage and power according to the car
velocity. Indeed, the car velocity is proportional to the frequency of excitation. The coefficient
of proportionality is the road wavelength 𝜆 that was assumed to be 5m in this case. The
frequency analysis revealed that the harvesting system could generate a considerable amount
of power even at non-resonant frequencies. Some random excitations were also applied to both
models.
In order to carry out the random excitation simulations, the ISO 8608 norm [162] for road
roughness levels has been used as a benchmark. It provides road roughness coefficients
according to the spatial reference frequency (which is assumed to be 0.1m-1 in this study) and
the road class (classified from A to H, respectively very smooth and extremely rough as
depicted in Table 4.12). The simulations were carried out for four road classes (A, D, G, and
H). The results revealed that the best road quality which is the class A road generated the least
amount of voltage and power for both the QCM and the HCM. Indeed the class A road
generated peak values of 1.949V and 0.00019W for the QCM and 2.239V and 0.0002506W for
the HCM. The class D road generated peak values of 15.6 and 0.01217W for the QCM and
17.93V and 0.01607W. The class G road profile generated peak values of 125V and 0.7815W
for the QCM and 143.3V and 1.027W for the HCM. The results revealed that the extremely
rough road profile which is the class H road profile generated the maximum voltage and power.
Indeed the class H road profile generated 250V and 3.126W for the QCM and 286.6V and
4.108W for the HCM. An increase of 14% and 32% were recorded in the HCM values of the
voltage and power, respectively, as compared with those of the QCM.
Moreover, the road, car, and piezoelectric stack parameters’ effects on the harvested voltage
were investigated. The sine wave acceleration amplitude affects greatly the amount of energy
harvested. For that, the performances were computed for the acceleration amplitude between 0
and 0.1m. The results revealed that the harvested voltage and power increased with the
amplitude. The power increase had a quadratic nature.
In addition to that, the effects of the piezoelectric parameters such as the force factor and
capacitance were also investigated. The results of the investigation revealed that the harvested
voltage decreased inversely with the force factor in the range of [0.2 .104; 6.104 N/v]. The
249
harvested power and voltage increased inversely with the piezoelectric capacitance in the range
of [0.1.10-5; 4 .10-5F]. Besides, it was noticed that the harvested voltage increases linearly with
the number of layers of the stack.
The simulations revealed that the harvested voltage and power decreased inversely with the
suspension damping coefficients (bsf and bsr) and unsprung (tire) tire damping coefficients.
The increase of sprung and unsprung damping ratios will lead to lower harvested power due to
the reduction of vibration in the piezoelectric stack. Although low damping coefficients allow
reaching high harvested power and voltage, they impact negatively the handling and
maneuverability of the car.
It was also observed that the harvested voltage increased for values of the suspension stiffness
coefficients (Ksf and Ksr) in the interval [1.106; 2.75.106 N/m] and decreases in the interval of
[2.75.106; 3.106 N/m]. The study on the effects of the unsprung (tire) stiffness coefficients (Kuf
and Kur) revealed that the harvested power and voltage increased for the values within [0.8.107;
0.93 .107 N/m] and decreased for values within [0.93 .107; 1.2 .107 N/m]. It should be noted
that the stiffer the tire, the lower the maximum harvested resonant power.
The effects of the sprung and unsprung masses on the harvesting system performances were
also studied. The plots revealed that for the values of the sprung mass included in [3.104; 4.75
.104 kg] the harvested power voltage increased while it decreased for values included in
[4.75.104; 6.104 kg]. The harvested power and voltage increased with the unsprung (tire) masses
in a range of [2000; 10000].
Several methods could be used to enhance the amount of harvested power. The manufacturing
of piezoelectric materials with high piezoelectric constants and force factors could help increase
the amount of vibration converted into electrical energy. Another downside of commercial
piezoelectric materials is their premature degradation over time. Indeed in a study undertaken
by a railway company [166], it was revealed that a couple of weeks after installing the
piezoelectric stacks underneath the floor of a station the harvested power started decreasing.
Another important aspect of improvement is the shape and geometry of the piezoelectric stacks.
250
A geometry that facilitates the application of a force that is perpendicular to the planar surface
of the piezoelectric material will positively affect the harvested power.
A non-linear analysis of the half and full car model would bring significant findings. Besides,
it would be very useful to find the equilibrium point between the car ride comfort, good
handling, and the harvested power so that all requirements can be met in a balanced way. The
study could also be extended to the analysis of the output voltage and power for a full car model.
In this study, the piezoelectric stack was connected in series with a suspension system and
located between the car body (sprung mass) and the suspension. Further studies could be carried
out in order to investigate the effects that different locations will have on the harvesting system.
This study was theoretical, it would be very useful to implement it in a real truck for the
evaluation of the harvested energy in real-life condition at mines.
251
REFERENCES
[1]. Erturk, A. and Inman, D. J., 2011, Piezoelectric Energy Harvesting, John Wiley &
Sons, Ltd. ISBN: 978-0-470-68254-8
[2]. Jacobson, M.Z., Delucchi, M.A., Bauer, Z.A., Goodman, S.C., Chapman, W.E.,
Cameron, M.A., Bozonnat, C., Chobadi, L., Clonts, H.A., Enevoldsen, P. et al., 2017, 100%
clean and renewable wind, water, and sunlight all-sector energy roadmaps for 139 countries of
the World, Joule, 1, 108–121.
[3]. Williams, C.B. and Yates, R.B., 1996, Analysis of a micro-electric generator for
microsystems, Sensors and Actuators A, 52, 8–11.
[5]. Cao, S. and Li, J.A., Survey on ambient energy sources and harvesting methods for
structural health monitoring applications, Adv. Mech. Eng., 9, 1687814017696210.
[6]. Roundy, S., Wright, P.K. and Rabaey, J., 2003, A study of low level vibrations as a
power source for wireless sensor nodes. Computer Communications, 26, 1131–1144.
[7]. Wang, H., Jasim, A. and Chen, X., 2018, Energy harvesting Technologies in roadway
and bridge for different applications—a comprehensive review, Applied Energy, 212, 1083–
1094.
[8]. Patel, P., 2011, Powering your car with waste heat. MIT technology review.
[9]. Pugi, L., Pagliai, M., Nocentini, A., Lutzemberger, G., and Pretto, A., 2017, Design
of a hydraulic servo-actuation fed by a regenerative braking system, Applied Energy, 187, 96–
115.
[10]. Noaman, M. and Remon, P., 2011, International Workshop SMART MATERIALS,
STRUCTURES & NDT in AEROSPACE Conference NDT in Canada 2-4 November 2011,
Montreal, Quebec, Canada.
[13]. Abhijit, G., Jendrzejczyk, J., Mulcahy, M., 2006, Design of electromagnetic shock
absorbers, International Journal of Mechanics and Materials in Design, 3, 285-291.
[14]. Velinsky, S.A., and White, R.A., 1980, “Vehicle energy dissipation due to road
roughness,” Veh. Syst. Dyn., 6, 359–384.
[15]. Jacota, V.G., 2017, “Evaluation of the Dissipated Energy by the Automobile
Dampers,” Glob. J. Res. Eng., 1, 1-9.
[16] Doaa A., Tariq D. and Abdel-Hamid I.M., 2020, “A State-Of-The-Art Review of Car
Suspension-Based Piezoelectric Energy Harvesting Systems”, energies, 13, 1–39.
[17]. Wei C., and Taghavifar, H., 2017, “A novel approach to energy harvesting from
vehicle suspension system: Half-vehicle model,” Energy, 134, 279–288.
[18] Abdelkareem, M.A., Xu L., Ali, M.K.A., Hassan, M.A., Elagouz, A. and Zou, J.,
2018, On-field measurements of the dissipated vibrational power of an SUV car traditional
viscous shock absorber, ASME 2018 International Design Engineering Technical Conferences
and Computers and Information in Engineering Conference.
[19] Yildiz, F., 2009, “Potential Ambient Energy-Harvesting Sources and Techniques.,” J.
Technol. Stud., 35(1), 40–48.
[20] Knight, C., Davidson, J. and Behrens, S., 2008, “Energy options for wireless sensor
nodes,” Sensors, 8(12), 8037–8066,
[21] Chalasani, S. and Conrad J. M., 2008, A survey of energy harvesting sources for
embedded systems, IEEE SoutheastCon 2008, 442–447.
[22] Twiefel, J. and Westermann, H., 2013, Survey on broadband techniques for vibration
energy harvesting, J. Intell. Mater. Syst. Struct., 24(11), 1291–1302.
253
[23] Devi, P.K., 2014, “Survey on Energy Harvesting Techniques,” Int. J. Sci. Eng.
Comput. Technol., 4(10), 256–259.
[24] Roundy, S., Wright, P. K. and Rabaey, J., 2003, A study of low-level vibrations as a
power source for wireless sensor nodes, Comput. Commun., 26(11), 1131–1144.
[25] Radousky, H.B. and Liang, H., 2012, Energy harvesting: an integrated view of
materials, devices and applications,” Nanotechnology, 23(50), 1-66.
[26] Cao S. and Li, J., 2017, A survey on ambient energy sources and harvesting methods
for structural health monitoring applications, Adv. Mech. Eng., 9 (4), 1-14.
[27] El-Sayed, A.R., Tai, K., Biglarbegian, M. and Mahmud, S., 2016, A survey on recent
energy harvesting mechanisms, IEEE Canadian Conference 2016 on Electrical and Computer
Engineering, 1–5.
[28] Xie, X. D. and Wang, Q., 2015, “Energy harvesting from a vehicle suspension
system,” Energy, 86, 385–392.
[29] Priya, S. and Inman, D. J., 2009, Energy harvesting technologies, Springer, vol. 2.
[30] Despesse, G., Jager, T., Chaillout, J. and Leger, J., 2005, Fabrication and
characterization of high damping electrostatic micro devices for vibration energy scavenging,
Proceedings of Design, Test, Integration and Packaging of MEMS and MOEMS, 386-390.
[31] Saha, C.R., O’Donnell, T., Loder, H., Beeby, S. and Tudor, J., 2006, “Optimization
of an electromagnetic energy harvesting device,” IEEE Trans. Magn., 42(10), 3509–3511.
[32] Jeon,S Y. B., Sood, R., Jeong, J.-H. and Kim, S.-G., 2005, “MEMS power generator
with transverse mode thin film PZT,” Sens. Actuators Phys.,122(1), 16–22.
[33] Stevens, J., 1999, Optimized thermal design of small thermoelectric generators,
Proceedings of 34th Intersociety Energy Conversion Eng. Conference. Society of Automotive
Engineers, 1999-01-2564.
254
[34] Yeatman, E.M., 2004, Advances in power sources for wireless sensor nodes,
Proceedings of International Workshop on Wearable and Implantable Body Sensor Networks,
Imperial College, 20–21.
[35] Rabaey, J. M., Ammer, M. J., Da Silva Jr, J. L., Patel, D., and Roundy, S., 2000,
Picoradio supports ad hoc ultra-low power wireless networking, IEEE Computer, 42–48.
[36] Gonsalez, C.G., Franco, V.R., Brennan, M.J., Da Silva S. and Junior, V.L., 2010,
Energy harvesting using piezoelectric and electromagnetic transducers, 9th Brazilian
Conference on Dynamics, Control and Their Applications, Serra Negra, Brazil, 1166–1171.
[37] Cary, B., Daniel, J. A., Deepam, M., Shashank, P., and Walter V., 2014, Advances in
Piezoelectric Polymer Composites for Vibrational Energy Harvesting, Polymer Composites for
Energy Harvesting, American Chemical Society, Washington, DC, 1-27.
[38] Aljadiri, R.T., Taha, L.Y., and Ivey, P., 2017, Wind energy harvesting systems: A
better understanding of their sustainability, 3rd International Conference on Control,
Automation and Robotics (ICCAR), 582–587.
[39] Ruiz-Garcia, L., Lunadei, L., Barreiro, P. and Robla, I., 2009, A review of wireless
sensor technologies and applications in agriculture and food industry: state of the art and current
trends, sensors, 9(6), 4728–4750.
[40] Ipema, A.H., Goense, D., Hogewerf, P.H., Houwers, H.W.J. and Van-Roest, H., 2008,
Pilot study to monitor body temperature of dairy cows with a rumen bolus, Comput. Electron.
Agric., 64, 49-52.
[41] Nadimi, E.S., Sogaard, H.T., Bak, T., Oudshoorn F.W., 2008, ZigBee-based wireless
sensor networks for monitoring animal and pasture time in a strip of new grass, Comput.
Electron. Agric., 61, 79-87.
[42] Chen, J., Cao, X., Cheng, P., Xiao, Y. and Sun, Y., 2010, Distributed collaborative
control for industrial automation with wireless sensor and actuator networks, IEEE Trans. Ind.
Electron., 57(12), 4219–4230.
255
[43] Christin, D., Mogre, P.S. and Hollick, M., 2010, Survey on wireless sensor network
technologies for industrial automation: The security and quality of service perspectives, Future
Internet, 2(2), 96–125.
[44] Ko J., Lu C., Srivastava M.B., Stankovic J.A., Terzis A. and Welsh M., 2010, Wireless
sensor networks for healthcare,” Proc. IEEE, 98 (11), 1947–1960.
Park, G., Rosing, T., Todd, M.D., Farrar, C.R. and Hodgkiss, W., 2008, Energy harvesting
for structural health monitoring sensor networks, J. Infrastruct. Syst., 14(1), 64–79.
[46] Liao, W.H., Wang, D.H. and Huang, S.L., 2001, Wireless monitoring of cable tension
of cable-stayed bridges using PVDF piezoelectric films, J. Intell. Mater. Syst. Struct., 12(5),
331–339.
[47] Chalard, L., Helal, D., Verbaere, L. and Wellig, A., 2007, Wireless sensor networks
devices: Overview, issues, state-of the art and promising technologies, ST Journal of Research,
4(1), 4–8.
[48] New Atlas, 2010, Piezoelectric generator creates power from shoes,
https://newatlas.com/piezoelectric-generator-shoes/14945/ [Visited on 08/04/21]
[50] Photodoto, 3 basic steps for photographing motion, moving objects and the art of
panning, https://photodoto.com/photographing-motion-moving-objects-and-the-art-of-
panning/ [Visited on 08/04/21]
[51] World economic forum, 2021, Scaling Climate Partnerships for Industry Transition,
https://www.weforum.org/events/scaling-climate-partnerships-for-industry-transition-2021
[Visited on 08/04/21]
256
[52] Care city, Londoners advised to know their pulse during Heart Rhythm Week,
https://www.carecity.london/articles/123-londoners-advised-to-know-their-pulse-during-
heart-rhythm-week [Visited on 08/04/21]
[57] The news Wheel, 2020, TPMS Explained: How Does the Tire Pressure Monitoring
System Work, https://thenewswheel.com/tpms-explained-how-does-the-tire-pressure-
monitoring-system-work/ [Visited on 08/04/21].
[58] Carazo A.V., 2000, Novel piezoelectric transducers for high voltage measurements,
PhD Thesis, Universitat Politecnica de Catalunya.
[60] IEEE power system relay committee, 1990, Gapped core current transformer
characteristics and performance, IEEE Trans., 5, 1732 – 1740.
[62] Somayyeh H., 2017, Energy Harvesting in Pneumatic Tires through Piezoelectric
Material and its Life Cycle Environmental Impact, Thesis (Master), Department of Building
Civil and Environmental Engineering, Concordia University.
[64] Jaffe, B., Roth, B.S. and Marzullo, S., 1954, Piezoelectric properties of lead zirconate
titanate solid solution ceramics, J. Appl. Phys., 25(6), 809-810.
[66] Meeker, T.R., 1996, Publication and proposed revision of ANSI/IEEE standard 176-
1987, IEEE Trans. Ultrason. Ferroelectr. Freq. Control, 43(5), 717–772, 1996.
[69] Gao, Z., Lu, C. and He H., 2016, Super stable ferroelectrics with high curie point, Sci.
Rep., 6, 1-6.
[70] Wenk, H.-R., Bulakh, A., 2004, Minerals: their constitution and origin, Cambridge
University Press, New York, ISBN 978-0-521-52958-7.
[71] Priya, S.J., Zhou, Y., Song H.C., Varghese, R., Chopra, A., Kim, S.G., Kanno, I., Wu
L., Ha, D.S., Ryu, J. and Polcawich, R.G., 2017, A review on piezoelectric energy harvesting:
Materials, Methods and circuits, Energy harvesting systems, 4 (1), 1-32.
[72] Boyanov, B., 2006, Synthesis and determination of curie temperature of ferrites from
the system BaO-CoO-ZrO2-Fe2O3, Journal of the University of Chemical Technology and
Mettalurgy, 41(1), 61-64.
[73] Chai, Y.L., Chang, Y.S., Teoh, L.J., Lin Y.J., Hsiao Y.J., 2008, Influence of ZrO2
addition on the structure, thermal stability and dielectric properties of ZnTiO3 ceramics, J.
Mater. Sci, 43, 6771-6776.
[74] Maznichenko, I.V., Ostanin, S., Ernst, A. and Mertig, I., 2009, First principles study
of manganese-stabilized hafnia, Journal of Magnetism and Magnetic Materials, 321, 913-916.
258
[75] Valasek, J., 1921, Piezoelectric and allied phenomena in Rochelle salt, Phys. Rev., 17
(4), 475-481.
[78] Saini, P.K., Biswas, A. and Bhanja, D., 2015, Performance Evaluation and Simulation
of Solar Panel, Wind Mill, Fuel Cell Hybrid System for Small Scale Energy Harvesting, Journal
of Clean Energy Technologies, 3(6), 417–421.
[79] Yang, B., 2010, Hybrid energy harvester based on piezoelectric and electromagnetic
mechanisms, Journal of Micro/Nanolithography, MEMS and MOEMS, 9(2), 023002.
[81] Baumgart, F., 2000, Stifness – an unknown world of mechanical science ?, Injury, 31,
14 – 84.
[82] De Jong, P., 2013, Power harvesting using piezoelectric materials - applications in
helicopter rotors, PhD thesis, University of Twente.
[83] Shu, Y.C., and Lien, I.C., 2006, Efficiency of energy conversion for a piezoelectric
power harvesting system, Journal of micromechanics and microengineering, 16, 2429–2438.
[84] Lefeuvre, E., Badel, A., Richard, C. and Guyomar, D., 2005, Piezoelectric energy
harvesting device optimization by synchronous charge extraction, Journal of Intelligent
Material Systems and Structures, 16, 865–876.
[85] Polla, D.L and Francis, F.L.,1998, Processing and characterization of piezoelectric
materials and integration into microelectromechanical systems, Annual Review of Materials
Science, 28, 563-597.
[87] Sencadas, V., Costa, C.M.and Senentxu, L.M., 2005, Poling of β-poly (vinylidene
fluoride): dielectric and IR spectroscopy studies, E-polymers, 5(1), 1-12.
[88] Nalwa, H.S. and Miyata, S., 1996, Non-linear optics of organic molecules and
polymers, CRC Press, 246 , ISBN 978-0-8493-8923-8.
[89] Kaura, T., Nath, R., Perlman R.R., 1991, Simultaneous stretching and corona poling
of PVDF films, J. Phy. D: Appl. Phys. 24(10), 1848.
[90] Marshall, J.M., Zhang, Q., and Whatmore R.W., 2008, Corona poling of highly (001)
/ (100) – oriented lead zirconate titanate thin films, Thin solid films, 516(15), 4679-4684.
[91] Lee, C. and Tarbutton J.A., 2019, Polyvinylidene fluoride direct printing for sensors
and actuators, The International Journal of advanced manufacturing Technology, 104 (2015),
3155-3163.
[92] Rose, C.A., Fish, K.A. and Rothenbert, C.A., 1954, Electromechanical transducer,
U.S. application, serial number 401, 812.
[93] Carazo A.V., 2011, 50 years of Piezoelectric Transformers. Trends I the technology,
MRSOnline Proceeding Library, 1-12.
[95] Uchino, K., 1997, Piezoelectric actuators and ultrasonic motors, Kluwer Academic
Publisher, Massachusets, USA.
[96] Goldfarb, M. and Jones, L.D., 1999, On the efficiency of electric power generation
with piezoelectric ceramic. Transaction of the ASME, 121, 566–571.
[97] Taylor, G.W., Burns, J.R., Kammann, S.M., Powers, W.B. and T.R. Welsh, 2001, The
energy harvesting eel: A small subsurface ocean/river power generator, IEEE Journal of
Oceanic Engineering.
[98] Jansma. S., 2010, The feasibility and design of piezoelectric energy harvester for civil
applications, Master thesis, University of Twente.
260
[99] Roundy. S., 2005, On the effectiveness of vibration-based energy harvesting, Journal
of Intelligent Material Systems and Structures, 16, 809–823.
[100] Kirchhhoff, S., 1845, Ueber den durchgang eins elektrischen stromes durch eine
ebene, insbesondere durch eine kreisförmige. Annalen der Physik, 140, 497–514.
[101] Horowitz, P. and Hill, W., 1989, The Art of Electronics. Cambridge University
Press, 2nd edition.
[102] Shu, Y.C. and Lien, I.C., 2006, Analysis of power output for piezoelectric energy
harvesting systems, Smart Materials and Structures, 15, 1499–1512.
[103] Shu, Y.C., Lien, I.C. and Wu, W.J., 2007, An improved analysis of the SSHI
interface in piezoelectric energy harvesting, Smart Materials and Structures, 16, 2253–2264.
[104] Arroyo, E. and Badel, A., 2011, Electromagnetic vibration energy harvesting device
optimization by synchronous energy extraction, Sensors and Actuators A Physical, 171 (2), 1-
23.
[105] Bauer, W., 2010, Hydro-pneumatic suspension systems, Suspension systems basics,
Springer, ISBN: 978-3-642-15146-0, 1-17.
[106] Florin, A., Ioan-Cozmin, MR. and Liliana, P., 2013, Passive suspension modeling
using MATLAB, quarter-car model, input signal step type, New Technol. Prod. Mach. Manuf.
Technol., 258–263.
[107] Sharma, P., Saluja, N., Saini, D. and Saini, P., 2013, Analysis of automotive passive
suspension system with Matlab program generation, Int. J. Adv. Technol., 4(2), 115-119.
[108] Vaishnav, A., Sarvaiya, M. and Dhabaliya, P., 2016, Mathematical Modelling and
Comparison of Two Degree of Freedom Suspension System of Quarter Car, Imp. J. Interdiscip.
Res., 2(11), 128-137.
261
[109] Jamali, M.S., Ismail, K.A., Taha, Z. and Aiman, M.F., 2017, Development of Matlab
Simulink model for dynamics analysis of passive suspension system for lightweight vehicle,
Journal of Physics: Conference Series, 908, 1-8.
[110] Hassaan, G.A., 2014, Car dynamics using quarter model and passive suspension,
part i: Effect of suspension damping and car speed, Int. J. Comput. Tech., 1(2), 1–9.
[111] Gao, W., Zhang, N. and Dai, J., 2008, A stochastic quarter-car model for dynamic
analysis of vehicles with uncertain parameters, Veh. Syst. Dyn., 46(12), 1159–1169.
[112] Prabhakar, S. and Arunachalam, K., 2015, Simulation and analysis of passive
suspension system for different road profiles with variable damping and stiffness parameters,
J. Chem. Pharm. Sci., 974, 32-36.
[113] Salah, M., 2017, A Laboratory Automotive Suspension Test Rig: Design,
Implementation and Integration., Jordan J. Mech. Ind. Eng., 11(2), 67-71.
[114] Jugulkar, L.M., Singh, S. and Sawant, S.M., 2016, Analysis of suspension with
variable stiffness and variable damping force for automotive applications, Adv. Mech. Eng.,
8(5), 1-19.
[115] Patel, C.B., Gohil, P.P. and Borhade, B., 2010, Modelling and vibration analysis of
a road profile measuring system, Int. J. Automot. Mech. Eng., 1(1), 13–28.
[116] Gao, W., Zhang, N. and Du, H.P., 2007, A half-car model for dynamic analysis of
vehicles with random parameters, 5th Australasian Congress on Applied Mechanics (ACAM
2007), 1, 595–600.
[117] Faheem, A., Alam, F. and Thomas, V., 2006, The suspension dynamic analysis for
a quarter car model and half car model, 3rd BSME-ASME International Conference on Thermal
Engineering, 20–22.
[118] Mehmood, A., Khan, A.A. and Khan, A.A., 2014, Vibration analysis of damping
suspension using car models, Int. J. Innov. Sci. Res., 9(2), 202–211.
262
[119] Chavan, S.P., Sawant, S.H. and Tamboli, J.A., 2013, Experimental Verification of
Passive Quarter Car Vehicle Dynamic System Subjected to Harmonic Road Excitation with
Nonlinear Parameters, IOSR J. Mech. Civ. Eng. IOSR-JMCE, 39–4.
[122] Mahala, K.M., Gadkari, P. and Deb, A., 2009, Mathematical models for designing
vehicles for ride comfort, ICORD 09: Proceedings of the 2nd International Conference on
Research into Design, Bangalore, India 07.-09.01.
[123] Iyer, A.R., Venkatachalam, R. and Balaraju, A., 2014, Analysis of Optimum
Suspension Parameters of a Semi-Independently Suspended Automobile, Int. J. Innov. Res. Sci.
Eng. Technol., 3(5), 12047-12053.
[124] Mitra, A. Benerjee, N. Khalane, H.A., Sonawane, M.A., Joshi, D.R. and Bagul,
G.R., 2013, Simulation and Analysis of Full Car Model for various Road profile on a
analytically validated MATLAB/SIMULINK model, IOSR J Mech Civ Eng IOSR-JMCE, 22–
33.
[125] Raju, A.B. and Venkatachalam, R., 2013, Analysis of Vibrations of Automobile
Suspension System Using Full-car Model, Int. J. Sci. Eng. Res., 4(9), 2105–2111.
[126] Xiao, H., Wang, X. and John, S., 2015, A dimensionless analysis of a 2DOF
piezoelectric vibration energy harvester, Mech. Syst. Signal Process., 58, 355–375.
[127] Zuo, L. and Zhang, P.S., 2013, “Energy harvesting, ride comfort, and road handling
of regenerative vehicle suspensions, J. Vib. Acoust., 135, no. 1, 295-302.
[128] Al-Yafeai, D., Darabseh, T. and Mourad, A.H. I., 2019, Quarter vs. Half Car Model
Energy Harvesting Systems, 2019 Advances in Science and Engineering Technology
International Conferences (ASET), IEEE, 1–5.
263
[129] Azizi, A. and Yazdi, P.G., 2019, White Noise: Applications and Mathematical
Modeling, in Computer-Based Analysis of the Stochastic Stability of Mechanical Structures
Driven, Springer, 25–36.
[130] Behera, M.M., 2015, Piezoelectric Energy Harvesting from Vehicle Wheels, Int. J. Eng.
Res. Technol., 4(5), 1-4.
[131] Lafarge, B., Grondel, S., Delebarre, C., and Cattan, E., 2018, A validated simulation
of energy harvesting with piezoelectric cantilever beams on a vehicle suspension using Bond
Graph approach, Mechatronics, 53, 202–214.
[132] Hu, Y., Xu, C., Zhang, Y., Lin, L., Snyder, R. L. and Wang, Z. L., 2011, A
Nanogenerator for Energy Harvesting from a Rotating Tire and its Application as a Self-
Powered Pressure/Speed Sensor, Adv. Mater., 23(35), 4068–4071.
[133] Zhang, Y., Zheng, R., Shimono, K., Kaizuka, T. and Nakano, K., 2016, Effectiveness
testing of a piezoelectric energy harvester for an automobile wheel using stochastic resonance,
Sensors, 16(10), 1-16.
[134] Makki, N. and Pop-Iliev, R., 2011, Piezoelectric power generation in automotive
tires, Proc. Smart Mater. Struct.
[135] Consortium, A., 2003, Intelligent tyre for accident-free traffic, Tech. Res. Cent. Finl.
VTT Tech Rep, 1-64.
[136] Ali S. F. and Adhikari, S., 2013, Energy harvesting dynamic vibration absorbers, J.
Appl. Mech., 80(4), 1-9.
[137] Madhav C. and Ali, S. F., Harvesting energy from vibration absorber under random
excitations, IFAC-Pap., 49(1), 807–812.
[138] Lee, H., Jang, H., Park, J., Jeong, S., Park, T. and Choi, S., 2013, Design of a
piezoelectric energy-harvesting shock absorber system for a vehicle, Integr. Ferroelectr.,
141(1), 32–44.
[139] Arizti, M., 2010, Harvesting energy from vehicle suspension, Master of Science
Thesis, Tampere University of Technology, Tampere.
264
[140] Hendrowati, W., Guntur, H.L. and Sutantra, I.N., 2012, Design, modeling and
analysis of implementing a multilayer piezoelectric vibration energy harvesting mechanism in
the vehicle suspension, Engineering, 4(11), 728-738.
[142] Normann, N., 2000, Tyre Pressure Monitoring System for all Vehicle Categories,
Crane Communications Inc.: ATZ Worldwide.
[143] Sankaranarayanan, V. and Levent, G., 2008, Tire Pressure Monitoring [Applications
of Control], IEEE control systems, 27(6), 22 – 25.
[144] Quan, X., Gong, J., Guan J., Bai R., You M., Zou P., 2019, Automobile tire pressure
monitoring technology and development trend, IOP Conf. Series: Journal of Physics: Conf.
Series, 1-7.
[145] Shraim, H., Rabhi, A., Ouladsine, M., M’Sirid, NK., and Fridman, L., 2006,
Estimation and analysis of the tire pressure effects on the comportment of the vehicle center of
gravity, Proc. 9th Int. Workshop on Variable Structure Systems, Italy, 268–273.
[146] Guan, M.J., and Liao, W.H., 2007, On the efficiencies of piezoelectric energy
harvesting circuits towards storage device voltages, Smart Mater. Struct., 16(2), 498–505.
[147] Sodano, H.A., Inman, D.J. and Park, G., 2005, Generation and storage of electricity
from power harvesting devices, J. Intell. Mater. Syst. Struct., 16(1), 67–75.
[148] Priya, S., 2007, Advances in energy harvesting using low profile piezoelectric
transducers, J. Electroceramics, 19(1), 167–184.
[149] Sodano, H.A., Inman, D.J. and Park, G., 2005, Comparison of piezoelectric energy
harvesting devices for recharging batteries, J. Intell. Mater. Syst. Struct., 16(10), 799–807.
[150] Ottman, G., Bhatt, A., Hofmann, H. and Lesieutre, G., 2002, Adaptive piezoelectric
energy harvesting circuit for wireless, remote power supply, IEEE Transactions on Power
Electronics, 17(5), 669-676.
265
[151] Ottman, G.K., Hofmann, H.F. and Lesieutre, G.A., 2003, Optimized piezoelectric
energy harvesting circuit using step-down converter in discontinuous conduction mode, IEEE
Transactions on Power Electronics, 18(2), 696-703.
[152] Sodano, H.A., Magliula, E.A., Park, G. and Inman, D.J., 2002, Electric power
generation from piezoelectric materials, Proceedings of the 13th International Conference on
Adaptive Structures and Technologies, 7–9.
[153] Chen, H., Cong, T.N., Yang, W., Tan, C., Li, Y. and Ding, Y., 2009, Progress in
electrical energy storage system: A critical review, Prog. Nat. Sci., 19(3), 291–312.
[154] Zhao, Z., Wang, S. and You, C., 2010, Piezoelectric micro-power generation to
charge supercapacitor with optimized duty cycle, J. Intell. Mater. Syst. Struct., 21(11), 1131–
1140.
[155] Casciati, F., Faravelli, L. and Rossi, R., 2005, Architecture optimization for wireless
sensor networks, Smart Structures and Materials 2005: Sensors and Smart Structures
Technologies for Civil, Mechanical, and Aerospace Systems, 5765, 253–258.
[156] Jiang, X., Polastre, J. and Culler, D., 2005, Perpetual environmentally powered
sensor networks, IPSN 2005. Fourth International Symposium on Information Processing in
Sensor Networks, Boise, ID, USA, 463-468.
[157] Li, H., Tian, C. and Deng, Z.D., 2014, Energy harvesting from low frequency
applications using piezoelectric materials, Appl. Phys. Rev., 1(4), 1-20.
[158] Tolentino, I.M., and Talampas, M.R., 2012, Design, development, and evaluation of
a self-powered GPS tracking system for vehicle security, Sensors, 1–4.
[159] Jiang, X., Li, Y., Li, J., Wang, J. and Yao, J., 2014, Piezoelectric energy harvesting
from traffic-induced pavement vibrations, J. Renew. Sustain. Energy, 6(4), 1-16.
[160] Ramadass Y.K. and Chandrakasan, A.P., 2010, An efficient piezoelectric energy
harvesting interface circuit using a bias-flip rectifier and shared inductor, IEEE J. Solid-State
Circuits, 45(1), 189–204.
266
[161] ISO 13473-1, 1997, Characterization of pavement texture by use of surface profiles–
Part 1: Determination of mean profile depth, Eur Stand ICS 1714030 Eur Comm Stand Bruss,
1-47.
[162] ISO 8608, 1995, Mechanical vibration– road surface profiles – reporting of measured
data”. Geneva, Switzerland: ISO, 1-36.
[163] He, L., Qin, G., Zhang, Y. and Chen, L., 2008, Non-stationary random vibration
analysis of vehicle with fractional damping, in 2008 international conference on intelligent
computation technology and automation (ICICTA), 2, 150–157.
[164] Chen, X. and Zhou, Y., 2018, Modelling and analysis of automobile vibration
system based on fuzzy theory under different road excitation information, Complexity, 2018, 1-
9.
[165] Hossain, M., Z. and Chowdhury, M., N., A., 2012, Ride Comfort of a 4 DOF NonLinear
Heavy VehicleSuspension,” ISESCO J Sci Technol, 8, 80–85.
[166] East Japan Railway Company, 2008, Demonstration experiment of the Power-
Generating Floor at Tokyo station, Report JR East, Japan, 1-3.
267
APPENDICES
Appendix 1. MATLAB code for QCM dump truck parameters – Step & Harmonic inputs.
clear all;
f= 1.47;%Hz (Frequency)
amp=0.05;%meter (Amplitude)
Appendix 2. MATLAB code for HCM simulation parameters – Step & Harmonic inputs
268
clear all;
f= 1.429;%Hz (Frequency)
amp=0.05;%meter (Amplitude)
tdelay= 0.64;%seconde (time delay between the front and rear axle = wheel base / car velocity
= (Lf+Lr)/u )
%First of all you need to import the system from simulink using the bode diagram block
%declare the mag, phase and radiant frequency of the variable sys
[mag,phase,w]=bode(sys.values);
[mag2,phase2,w2]=bode(sys2.values);
270
F=w/(2*pi);
F2=w2/(2*pi);
%added into the bracket after the variables were declared as for this example
xlim([0 20]);
xlabel('Frequency (Hz)');
ylabel('Magnitude');
legend('yus/yr','ys/yr')
%First of all you need to import the system from simulink using the bode diagram block
%declare the mag, phase and radiant frequency of the variable sys
[magsprung,phasesprung,wsprung]=bode(syscarbody_sprung.values);
[maguf,phaseuf,wuf]=bode(sysuf.values);
[magur,phaseur,wur]=bode(sysur.values);
[magpitch,phasepitch,wpitch]=bode(syspitch.values);
Fsprung=wsprung/(2*pi);
Fuf=wuf/(2*pi);
Fur=wur/(2*pi);
Fpitch=wpitch/(2*pi);
%added into the bracket after the variables were declared as for this example
plot(Fsprung, squeeze(magsprung),'b','linewidth',1);
hold on
272
plot(Fuf, squeeze(maguf),'r--','linewidth',1);
hold on
plot(Fur, squeeze(magur),'g:','linewidth',2);
hold on
plot(Fpitch, squeeze(magpitch),'m--','linewidth',2);
hold off
xlim([0 20]);
xlabel('Frequency (Hz)');
ylabel('Magnitude');
legend('ys/yr','yuf/yr','yur/yr','teta/yr')
%%QCM dissipated power in the frequency domain - AKA without piezoelectric stack
%% Mass matrix
273
M=[Mus 0; 0 Ms];
%% Stiffness matrix
%% Damping matrix
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Feq=0:0.01:20;
W=2*pi*Feq;
for N=1:length(W)
s=1i*W(N);
E=M.*s^2+B.*s+K;
Y=E\F;
end
274
plot(Feq,P,'b','LineWidth',1)
xlabel('Frequency (Hz)','color','k')
%% Title
set(gca,'FontSize',10,'FontName','times')
grid
275
CURRICULUM VITAE
Personal Information
Name Surname Drissa Mohamed Malo
Place of Birth Orodara
Date of Birth 31.05 1993
Nationality T.C. Other:
Phone Number 00226 54 16 17 27
Email madrmo1993@gmail.com
Web Page
Educational Information
B. Sc.
University Nazi Boni University
Faculty Engineering Institute
Department Mechanical Engineering
01.10.2014
M. Sc.
University Istanbul University
Institute Institute of Sciences
Department Mechanical Engineering
Programme Mechanical Engineering
Ph. D.
University Istanbul University-Cerrahpasa
Institute Institute of Graduate Studies
Department Mechanical Engineering Department
Programme Mechanical Engineering Programme
Publications
American Journal of Mechanics and Applications “Design and Finite Elements
Analysis of a Hydraulic Excavator’s Robot Arm System” doi:
10.11648/j.ajma.20190703.11
European Business and Management Journal “Mining Machinery Maintenance
Key Performance Indicators Improvement at the Nordgold Taparko Mine” doi:
10.11648/j.ebm.20190506.13
World Journal of Advanced Research and Reviews, “Piezoelectric energy
harvesting from a dump truck suspension system”
https://doi.org/10.30574/wjarr.2022.15.2.0769