Description

Abstract

Design and implementation of an inductive power transfer system for wireless charging of future electric transportation

Kunwar Aditya, Ph.D.

 

The University of Ontario Institute of Technology, 2016

 

The motivation of this thesis was to formulate clear design guidelines for fabrication and control of an efficient series-series resonant inductive power transfer (SS-RIPT) system for electric vehicle battery charging application. In meeting this objective, several critical deficiencies about the field of RIPT based EV chargers specific to stationary charging have been solved. Firstly, to increase the tolerance to misalignments, use of an unsymmetrical coil pair for the charger has been proposed. An unsymmetrical coil pair, in which the outer diameter of the primary and the secondary coils are kept equal, whereas the inner diameter of the secondary is kept larger compared to the primary counterpart gives the best performance in misalignment conditions. By employing this unsymmetrical coil-pair, a charging pad which shows the horizontal tolerance to misalignment equal to 71% of the pad diameter has been presented. Secondly, a very simple yet novel analytical design procedure has been submitted, adopting which, eliminates the bifurcation issue for the entire range of load and coupling variation and therefore requires no sophisticated control. Finally, a simplified mathematical model of SS-RIPT system has been proposed for primary side control of output voltage and current. All the proposed theories and analysis have been verified by a 3.6 kW prototype of the SS-RIPT based charger fabricated in the lab. A DC-DC efficiency of 91% for rated load condition is achieved for the designed charger. For partial load conditions (less than 50% of the rated load), the efficiency is 87%.

 

 

Acknowledgements

I express my deepest sense of gratitude to my supervisor Dr. Sheldon S. Williamson for accepting me as his Research Assistant and for his guidance during the course of this research. I am highly indebted to him for the financial support during my tenure as his Research Assistant. I would like to thank him for the supervision, trust and time that I received during this research that proved very useful for increasing my research capabilities and increasing my knowledge of power electronics.

I owe the successful realization of this work to the financial support from Dr. Williamson through NSERC and Transport Canada, as well as the teaching assistantship provided by the University Of Ontario Institute Of Technology, Oshawa, Ontario.

Many thanks to Dr. Najath Abdul Azeez, Post-doctoral Fellow at UOIT for his timely support on the procurement of components and pieces of equipment requested by me. I also acknowledge the help and encouragement from my colleagues in the STEER group.

I would also like to acknowledge the utilization of facilities available at the Department of Electrical and Computer Engineering, Concordia University, Montreal, Quebec, where I pursued the first year of my Ph.D. research under the supervision of Dr. Sheldon Williamson.

Last, but not least, I am grateful to my parents for allowing me to study abroad and carry out the research, as well as for their love and moral support.

 

 

Table of Contents

Abstract……………………………………………………………………………………………………………………………………………….. iii

Acknowledgements……………………………………………………………………………………………………………………………… iv

List of Figures……………………………………………………………………………………………………………………………………… ix

List of Tables…………………………………………………………………………………………………………………………………….. xiii

List of Symbols………………………………………………………………………………………………………………………………….. xiv

Introduction………………………………………………………………………………………………………………….. 17

• Rationale for Adopting Wireless Charging system for EVs……………………………………… 18
• A Brief History of Development of RIPT System for EV Charging…………………………. 20
• Working Principle and Components of RIPT System……………………………………………….. 23
• Wireless Charging Standards for Electric Vehicle……………………………………………………. 27
• Research Goals and Objectives…………………………………………………………………………………… 29

Design Considerations for Resonant Inductive Link…………………………………………………. 33

• Brief Overview of Different Coil Shapes Employed In Wireless Charger………………. 33
• Electrical Equivalent Circuit for Series-Series Compensated RIPT System……………. 36
  • Quality Factor of an SS-RIPT System…………………………………………………………….. 38
  • Bifurcation Phenomena In an SS-RIPT System……………………………………………… 41
• Calculation of Electrical Parameters for Bifurcation Free Operation……………………. 47
• Analytical Design of Archimedean Spiral Coils………………………………………………………… 49
  • Analytical Model of Self-Inductance of Coils…………………………………………………. 50
  • Analytical Model of Mutual Inductance between Coils………………………………….. 52
• Finding Coil-Pair Least Sensitive to Misalignment…………………………………………………… 54
  • Calculation of Electrical Parameters for 500 W setup………………………………….. 55
  • Calculation of Geometric Parameters for 500 W setup…………………………………. 60
  • Verification of Analytical Expressions…………………………………………………………… 63
  • Mutual Inductance Profile of Coil-Pairs…………………………………………………………. 67
• Summary of Chapter 2………………………………………………………………………………………………… 72

Design of 3.6 kW Wireless Charger……………………………………………………………………………. 73

• Calculation of Electrical Parameters…………………………………………………………………………. 73
• Design of Litz Wire for the Coils………………………………………………………………………………… 74
• Fabrication of Coils for 6 kW Charger…………………………………………………………………… 76
• Addition of Ferrites to the Fabricated Coils………………………………………………………………. 77
• Verification of Designed Pads…………………………………………………………………………………….. 80
  • Verification of Magnetic Saturation in Ferrites…………………………………………….. 80
  • Verification of Bifurcation Free Design…………………………………………………………. 82
• Sensitivity Analysis of Designed Pads…………………………………………………………………………. 84
• Summary of Chapter 3………………………………………………………………………………………………… 88

Mathematical Model and Controller Design……………………………………………………………… 89

• A Reduced Dynamic Model for an SS-RIPT system…………………………………………………… 90
• Derivation of Small-Signal Model from Reduced Dynamic Model………………………….. 93
• Piecewise-Linear Model of Li-Ion Battery Pack……………………………………………………… 100
• Selecting the Compensation Capacitors for ZVS Tuning………………………………………… 103
• Design of Voltage Control Loop……………………………………………………………………………….. 107
  • Bode Plot of Open-Loop System……………………………………………………………………. 107
  • Derivation of Closed-Loop Controller………………………………………………………….. 110
• Design of Current Control Loop……………………………………………………………………………….. 113
• Summary of Chapter 4………………………………………………………………………………………………. 115

Simulation and Experimental Validation………………………………………………………………… 117

• Description of Hardware Setup…………………………………………………………………………………. 117
• Simulation and Experimental Results for Open Loop…………………………………………….. 120
• Simulation and Experimental Results for Current Control……………………………………. 122
  • Case I: Change in Load at Fixed reference Current…………………………………….. 122
  • Case II: Change in Reference Current at Fixed Load………………………………….. 124
  • Case III: Fluctuation in Input DC supply………………………………………………………. 125
  • Case IV: Change in Mutual Coupling at Fixed Load……………………………………. 126
• Simulation and Experimental Results for Voltage Control…………………………………….. 128
  • Case I – Change in Load at Fixed Reference Voltage………………………………….. 128
  • Case II – Change in Reference Voltage at Fixed Load…………………………………. 130
  • Case III – Tracking Performance of Controller for Variation in DC Input Voltage 131
  • Case IV – Tracking Performance of the Controller for Variation in Mutual Coupling 131
• Verification of ZVS in Inverter Switches………………………………………………………………….. 134
• DC-DC Efficiency of the Designed Charger……………………………………………………………… 136
• Summary of Chapter 5………………………………………………………………………………………………. 138

Safety and Shielding Issues……………………………………………………………………………………….. 139

• Safety Considerations of Wireless Charger for Electric Vehicles………………………….. 139
• Effect of shielding on leakage flux…………………………………………………………………………… 141
• Design of Shielding for Coils…………………………………………………………………………………….. 143
• Performance Analysis of Final Charging Pad…………………………………………………………. 146
• Summary of Chapter 6………………………………………………………………………………………………. 149

Comparison with SAE J2954…………………………………………………………………………………….. 151

• WPT Classification……………………………………………………………………………………………………. 151
  • WPT Power Classes with Efficiency Targets……………………………………………….. 151
  • WPT Z Classes………………………………………………………………………………………………… 152
• General System Requirements and Interface…………………………………………………………… 153
  • Functional/Physical Requirement…………………………………………………………………. 153
  • Frequency Range and Tuning………………………………………………………………………… 153
  • Nominal Position and Offset………………………………………………………………………….. 153
• Interoperability………………………………………………………………………………………………………….. 154
• EMF Exposure to Human and Implanted Medical Devices…………………………………….. 154
• Safety Testing…………………………………………………………………………………………………………….. 155

Conclusion………………………………………………………………………………………………………………….. 156

• Achieved results…………………………………………………………………………………………………………. 156
• Future Developments…………………………………………………………………………………………………. 158

References………………………………………………………………………………………………………………………………………… 160

Appendix A……………………………………………………………………………………………………………………………………….. 180

Output Power of an SS-RIPT Link Fed from a Voltage Source……………………………… 180

Efficiency of an SS-RIPT Link Fed from a Voltage Source……………………………………. 181

Appendix B……………………………………………………………………………………………………………………………………….. 183

Explanation for Negative Mutual Inductance and Null in Coupling Profile…………. 183

Appendix C……………………………………………………………………………………………………………………………………….. 187

PLECS Simulation Circuits………………………………………………………………………………………. 187

Appendix D……………………………………………………………………………………………………………………………………….. 188

FEA Models for Iterations Mentioned in Table 6.1…………………………………………………. 188

Curriculum Vitae……………………………………………………………………………………………………………………………… 190

Introduction

 

Electric Vehicles (EVs), which are not a new concept, were developed in the mid- 19th century. However, they became obsolete throughout almost entirely the 20th century mainly because of their limited driving range and high cost as compared to gasoline powered vehicles. In past the few decades, due to growing concern over increasing environmental pollution, depleting energy resources and increasing oil prices, electric propulsion has been considered as the ultimate solution since electricity can be produced by using non-polluting and continuously renewable energy sources [1]. Renewed interest in EVs has also led to increased interest in the development of charging methods for EVs. Traditionally, EVs usually use a simple plug-in charging method also known as conductive charging or plug-in charging in which a copper connected cable forms the power link. Plug-in charging is a widely accepted method on the market, and it is available for most EVs: Chevrolet Volt, Tesla Roadster, Nissan Leaf, and Mitsubishi i-MiEV. However, there are disadvantages to plug-in charging, such as safety issues caused by exposed plugs and damaged cables. To avoid the drawbacks of plug-in charging, wireless charging methods have been widely studied in recent years [2]–[5].

Wireless power transfer (WPT) has been demonstrated using various WPT systems, such as: Acoustic [6], [7]; Light [8]; Microwave [9]; Laser [10]; Capacitive [11]; and Inductive [12]. The basic layout of all WPT systems is similar. They all consist of a transmitter connected to a primary electronic circuit and a receiver connected to a secondary electronics circuit. The ‘medium of power transfer’ between receiver and transmitter makes them different from each other. It has been established in the literature that only an inductive WPT system has the potential to be applied for medium and high power applications, and particularly for the charging of EV batteries. In the literature, this method of power transfer has also been referred, contactless power transfer (CPT), contactless energy transfer (CET), inductively coupled power transfer

 

 

 

(ICPT), resonant inductive power transfer (RIPT) and inductive power transfer (IPT). For this thesis, the term RIPT will be used hereafter for consistency.

This thesis focuses on the design and control of an efficient RIPT system for charging of EV battery. In the sections to follow, a brief history of RIPT system for EV, followed by an overview of the working principle and components of an RIPT system, and commercialization activities being carried out world-wide in the field of wireless EV charging, has been presented. The final sections in this chapter will outline the objectives and contribution of this thesis and gives a brief overview of the work completed by the author.

 

1.1       Rationale for Adopting Wireless Charging system for EVs

The transport sector alone accounts for approximately 23% of the total energy- related emissions [13]. EVs are seen as a viable solution to the growing pollution problem. However, the major hurdle in the broad acceptance of EVs is attributed to their high battery cost and limited driving range.

The Li-Ion battery is widely used as the primary power source for the EV’s drivetrain due to their high specific energy (100-265 Wh/kg) and specific power density (250-340 W/kg) compared to other battery technology. Despite its superior characteristics, it still adds considerable weight and size to the vehicle. For example, Nissan Leaf’s 24 kWh battery pack weighs around 200 kg. In addition to high weight and size, the estimated cost of the battery pack is about US$700/kWh. Therefore, the price of an EV is almost double that of a gasoline counterpart with nearly half of the cost for the battery itself [14].

The limited driving range is an even greater obstacle to the market penetration of EVs than their higher cost. For example, gasoline vehicles can go over 500 km before refuelling which takes about 2-3 minutes at a filling station which are located every few kilometers. On the other hand, most EVs can only go about 100-200 km before recharging, and take a long charging time [15]. For example, Level 2 charging circuit of 9.6 kW power takes about 1.5 hrs, to fully charge the battery pack of the Chevrolet Volt [16]. Besides this, charging stations are not as readily available as fuel stations. These limitations of EVs have been termed as ‘range anxiety’ issue for obvious reasons.

 

 

 

The aforementioned limitations of EVs can be overcome by adopting wireless charging technology for EV battery charging which provides the following advantages over conventional wired charging:

• Range Extension: Wireless charging has the scope for ‘opportunity charging’ e. charging the vehicle little and often during the day when the EV is not in use [17]. For example, in addition for residential garage, public and private parking areas, the wireless charger can also be installed at: traffic lights; bus stops; high traffic congestion area where vehicles are slow moving; and taxis ranks that move forward as taxis are hired [18]. These opportunity charging are possible since wireless charging does not require human intervention and therefore charging can be carried out automatically. This in turn leads to significant improvements in range compared to that available from a single overnight plug-in charge.
• Safety and Convenience: Wireless charging provides galvanic isolation between load and source. Therefore it eliminates the disadvantages of plug-in charging technology such as: risk of electrocution, especially in wet and hostile environment from aging wiring and bad connections; failure to plug in; trip hazard from a long connecting wire; poor visual appeal due to hanging cords; contactor wear caused by excessive use and thermal cycling; and, most importantly, discomfort in handling a plug-in charger in a harsh climate that commonly has snow and where the charge point may become frozen onto the vehicle [14], [19], [20].
• Battery Volume Reduction: Due to the scope of opportunity charging, charging can take place more Therefore, EVs can travel the same distance with a reduced battery pack [13], [20]. This, in turn, can lower the price of EVs and make them more efficient due to the reduced weight. Frequent charging also extends the battery life by reducing the depth of discharge in the battery.
• Weather Proof: In a wireless charger, power transfer takes place due to an electromagnetic link, therefore charging is not affected by the presence of snow, rain, or dust storms. Besides, a transmitter is embedded underground, therefore, is safe from extreme weather condition and requires less frequent maintenance or replacement than a plug-in charger would require.

 

 

 

• A Brief History of Development of RIPT System for EV Charging Resonant inductive power transfer (IPT) is not a new concept, and many attempts have been made in the past to transmit power wirelessly, most notably by Nikola Tesla (1856 –1943) in the late 1800s and early He was inspired by the work of Heinrich Hertz (1857–1984) who first confirmed the existence of electromagnetic radiation in his experiments in 1888. Tesla reported several experimental setups of his WPT study using a high-frequency oscillator for medical/therapeutic applications [21]. Figure 1.1 shows a simplified schematic of one of his experimental setup to power a light bulb

wirelessly using RIPT system.

 

Figure 1.1 An example of RIPT setup suggested by Tesla[21], [22]

 

In Figure 1.1 one can see the circuit contains two loosely coupled and tuned resonant circuits: a primary (P) and a secondary (S). An external capacitor C is used to tune the primary while self-capacitance of the solenoid coil is applied to tune the secondary coil. The operating frequency used by Tesla was in the range between 20-100 kHz. Periodic spark gap discharges were used to control the power in the resonant circuit, as the modern resonant converter do today by using power electronic switches. Also, these discharges converts the mains frequency to high frequency of the resonant circuit. It is worth noting that Tesla’s experiments were first to demonstrate power transfer using a resonant inductive link and forms the basis of the majority of today’s modern wireless power transfer system.

In 1894, Hutin and LeBlanc submitted a patent [23] that describes a transformer system for electric railways. Although they named their system a transformer, it was basically an RIPT system for street cars. It contains a single-wire (unipolar) elongated primary track with 2 kHz alternating current and multiple resonant pick-ups at the secondary side. Ferromagnetic material and a suspension system that lowers the pick-

 

 

 

ups were proposed to increase mutual coupling between the track and the pick-up. Although the proposed topology has some similarities to modern solutions, its practical application was not successful, so the myth was propagated that wireless transfer of “power” over large distances was impossible. However it was possible to send “signal” and there has been significant development of the wireless communications system and inductive antennas, therefore this “Signals-yes”, “power-no” categorization persisted for more than 100 years [12].

In the 1970s, due to the energy crisis and advances in materials science, semiconductors and power electronics, the temporarily suspended research in the area of EVs and inductive power supply was reinstated through academic studies. An extensive, long-term project called PATH (Partners for Advance Transit and Highways) was conducted at the University of California, Berkeley through the 1980s [24].This system was referred to as ‘Dual Mode Electric Transportation’ (DMET) as it considered both batteries as well as powered highways as the energy source of the EVs. For the test vehicle, a 60kW, 35-passenger bus was selected. The complete infrastructure was built for a 213 m long test track with two powered sections with a total length of 120 m. A bipolar primary track was used, and it was supplied with 1200A, 400 Hz AC. The pick- up had an area of 4.3 m², and a 7.6 cm distance from the primary track was used. Power control on the bus was achieved by capacitive detuning the pickup system, thereby placing a large reactive load on the generator, which the generator could easily supply at reduced efficiency. The air gap was controlled to be 50-100 mm when coupling power and 150–200 mm when not coupling power. The attained efficiency was around 60%. The results of the project have proved the substantial potential of roadway powered vehicles, but due to low efficiency and lack of economic viability project was abandoned.

In 1986, Kelly and Owens proposed powering aircraft entertainment systems using wires under the carpet in the passenger bay of a plane. It was a 38 kHz system supplying a load of 8 W for each passenger and total of 1 kW load without any controller, but the system was highly inefficient [25]. This innovation was followed by Turner and Roth (U.S. Patent 4 914 539) in 1990 using much the same infrastructure but with a controller on each parallel-tuned pickup circuit so that the VAR load on the pickup was varied to supply each entertainment system with a constant voltage. The system operated with

 

 

 

constant resonant voltages on all the pickups, which conserved real power but placed a large VAR load on the generator under light loading conditions.

In 1991, researchers at Auckland University, NZ led by Boys and Green focused their attention on high power applications, particularly on the inductive power supply of movable objects. Their IPT system for material handling (US Patent 5 293 308) became the cornerstone of much of the work in IPT systems over the past 20 years. This work was funded by Japanese company Daifuku Co. Ltd., so the development was compatible with Japanese regulations. Therefore, the operating frequency was constrained to be below 10 kHz. The complete system included a resonant power supply driving an elongate inductor, parallel tuned with a capacitor, and some parallel-tuned pickups, each with its decoupling controller, supplying power at nominally constant voltage to their particular load [26], [27]. Since then work in Auckland has been led by Covic and Boys. They have investigated various aspects of the IPT, such as: topologies of the primary resonant converter; compensation of the primary and pick-up circuits; the optimal control of the pick-up; the multi-phase design of the elongated primary track; and the bifurcation phenomenon. It is especially worth noting their recent achievement in designing an optimal pad for the stationary charging of EVs [28].

In 2007, a group of physicists from the Massachusetts Institute of Technology caught the world’s attention by powering a 60 W of lightbulb suspended in space, 2 m away from the transmitting coil [29]. Operating frequency was 9.9 MHz and self-capacitance of the coils were used for the resonant tuning. Although the reported end-to-end efficiency of the system was only 40% [30], the media publicity and growing public interest led to the development of the spin-off company named WiTricity. WiTricity is presently offering wireless charging solutions for household electronics and EV charging systems.

In addition to stationary charging applications, in-motion charging of EVs has also been undertaken at various academic and industry research groups. Notable among them is On-Line Electric Vehicle (OLEV) project conducted at the KAIST (Korea Advanced Institute of Science and Technology) in South Korea in 2009 [31]. The high price, high weight, and limited range of electric batteries for EVs motivated researchers to develop a powered roadway system that can reduce the required amount of cells by

 

 

 

80%. Three generations of OLEV systems have been developed, and three different vehicles have been tested: the first generation is a light golf cart which takes power of 3 kW at an air gap of 1 cm and total efficiency of 80%; the second generation is a bus for which 6 kW power transfer at an air gap of 17cm and total efficiency of 72% has been achieved; and the third generation is an SUV for which 17 kW power transfer takes place at 17 cm air gap and efficiency of 71% [13].

 

1.3       Working Principle and Components of RIPT System

Although it took more than century to develop and accept the RIPT system for wireless charging of EVs and other consumer products, the basic idea is based on well- established Ampere’s circuital law and Faraday’s law of induction. How power transfer takes place with these two laws, can be explained with the help of Figure 1.2.

 

Figure 1.2 Power transfer via two mutually coupled air cored coils

 

Figure 1.2 shows two coils linked by inductive coupling. Here, the subscript P and S refers to primary and secondary coil, respectively. The terms ϕM, ϕlP, and ϕlS are mutual flux, primary leakage flux, and secondary leakage flux, respectively. Let M, LP, LS be mutual inductance, self-inductance of primary and self-inductance of secondary coil respectively. When a time varying current is applied to the primary coil, a time varying flux of the same frequency is produced in the region surrounding the primary. The strength of the magnetic field around a closed path is directly proportional to the current carried by the coil and is given by Ampere’s law (for the case when the displacement current is neglected) by (1.1) [32].

 

 

∮𝑙 𝐻. 𝑑𝑙 = ∫𝑠 𝐽. 𝑑𝑠

(1.1)

 

 

 

If the currents are carried by wires in a coil with N turns, then (1.1) can be simplified as:

∮𝑙 𝐻. 𝑑𝑙 = 𝑁_𝑃𝐼_𝑃                                                                                        (1.2)

Here, ‘H’ is the magnetic field strength, NP is number of turns while IP is the current flowing in the primary coil, and l is the length of the circumference of the closed path. This time varying magnetic flux links the secondary coil and emf is induced in the secondary coil by the principle of Faraday’s law of electromagnetic induction given by (1.3):

 

 

𝑒_𝑆

= 𝑁_𝑆

𝑑𝜙𝑚

𝑑𝑡

(1.3)

 

 

Here, 𝑒_𝑠 is the emf induced in the secondary coil and 𝜙_𝑚 is the flux linking the secondary coil. This emf is capable of driving a current to the load if the circuit is closed. From (1.2) one can understand that the greater the magnitude of current in the primary, the stronger will be the magnetic field strength. From (1.3) one can understand that the higher the rate of change of mutual flux, the greater is the magnitude of emf induced in the secondary. This rate of change of mutual flux which is equal to supply frequency signifies that a high-frequency current in the primary is needed to establish high emf in the secondary. However, all the flux does not link the secondary; coupling coefficient k relates common flux to total flux.

 

 

𝑘 =    𝜙𝑚       

𝜙𝑚+𝜙𝑙𝑃

(1.4)

 

 

In (1.4) 𝜙_𝑃 = 𝜙_𝑚 + 𝜙_𝑙𝑃 is the total flux produced by the primary coil. Induced emf in terms of the coupling coefficient is given by (1.5)

 

 

𝑒_𝑆

= 𝑘𝑁_𝑆

𝑑𝜙𝑃

𝑑𝑡

(1.5)

 

 

If all the flux link the secondary, then k is 1 and if none of the flux links the secondary k is 0 i.e. 0 ≤ k ≤ 1. Based on the values of k the magnetically coupled system can be classified into two categories, namely tightly coupled systems and loosely coupled systems. In tightly coupled systems, such as transformer and induction motor, the primary is placed in proximity of secondary and flux is shaped by placing windings on the core of high magnetic permeability. Therefore, they have mutual inductance greater

 

 

 

than the leakage inductance. Because of tight coupling, k usually lies between 95% to 98% for the transformer and approximately 92% for the induction motor[33]. In the case of using magnetically coupled system for powering EVs, a large air gap is required to allow for inconsistency in the road surface and better clearance between the road and vehicle. Because of this large air gap, the leakage flux is very high and the coefficient of coupling is from 1% to 3% only. Such applications are classified under loosely coupled systems. Poor coupling in loosely coupled systems leads to poor transfer of power. To improve coupling and compensate leakage inductance, capacitive compensation in primary and secondary windings is required [34].

In light of the above discussion, a generalised block diagram of the RIPT system for EV battery charging can be drawn and has been shown in Figure 1.3.

Figure 1.3 Generalized block diagram of IPT system for EV battery charging Figure 1.3 shows the global energy chain for a typical RIPT based EV battery

charger. On the way-side, power is usually provided by a utility supply that may be DC from a battery or low frequency (LF) 60 Hz AC from a grid. If the utility is supplied from a grid then a power factor correction stage may also be included in this block to reduce the harmonic pollution of the grid. A high-frequency (HF) voltage at a few tens of kHz is then generated by a high-frequency converter which is simply an inverter if the utility is DC or a power factor correction rectifier followed by an inverter if the utility is 60 Hz AC [35].

This high-frequency voltage generates energy in the form of a high-frequency current through a compensation network and primary coil. A primary compensation circuit is added so as to have the primary input voltage and the current in phase to minimize the VA-rating and thus the size of the high-frequency power converter [36]. Moreover, primary compensation also acts as a band pass filter blocking undesirable

 

 

 

frequency components generated from the power electronic converter feeding the primary. Therefore, an almost sinusoidal current flows in the primary coil and this enables soft switching operation of the converter feeding the primary [37].

Energy is then transferred to the vehicle side through the secondary coil which is mutually coupled to the primary coil through the flux generated in the air-gap by the primary coil current. The energy received by the secondary coil is then processed by the secondary compensation circuit which is added to improve the power transfer capability of the system [38]. Finally, the voltage thus received is rectified so as to make it exploitable by the load (batteries). Depending upon the control, an additional DC-DC converter is sometimes included between the rectifier and the load [39]. Apparently, the DC-DC converter brings in more components and corresponding losses.

Compensating networks, which are capacitors, are made to resonate with coil inductance, thus forming a resonant inductive link. Depending on the connection of the compensating capacitor in the primary and secondary coils, four types of resonant inductive links can be defined: series-series (SS), series-parallel (SP), parallel-series (PS) and parallel-parallel (PP) [40]. Primary parallel compensation such as PP and PS resonant inductive links allows using a higher primary current as only a small part of the current flows through the semiconductor. However, PP and PS have several fundamental drawbacks. First, they require an additional series inductor to regulate the inverter current flowing into the primary resonant tank. This series inductor, in turn, increases the converter size and, therefore, the total cost of the RIPT system [41], [42]. Secondly, due to the circulating current in the primary resonance tank, the partial load efficiency of parallel compensated primary system is lower [43]. Thirdly, in PP and PS resonant inductive links, the value of the primary compensation capacitor is not constant but varies with varying mutual coupling and load. Therefore, PP and PS resonant inductive links will require sophisticated control strategies to maintain unity power factor operation in the primary power supply irrespective of load and mutual coupling variation.

Primary series compensation allows canceling the significant voltage drop of a primary coil, therefore the required voltage rating of the power supply is reduced. In SS and SP resonant inductive links, no extra inductor is needed. Moreover, primary

 

 

 

compensation is independent of load [44]. However, in an SP resonant inductive link, primary compensation depends upon mutual coupling and, therefore, needs consideration in dynamic charging applications. The SP resonant inductive link requires a higher value of capacitance for stronger magnetic coupling, and its peak efficiency is inferior to an SS resonant inductive link [45]. Therefore, an SS resonant inductive link is theoretically the best regarding efficiency, component count, the complexity of control, and cost, and hence is the focus of this thesis.

 

1.4       Wireless Charging Standards for Electric Vehicle

In 2010 the Society of Automotive Engineers (SAE) assembled an international committee, known as SAEJ2954, to develop a working industry standard that establishes the interoperability, frequency band, electromagnetic compatibility, minimum performance, safety, testing criteria as well as coils definitions for wireless charging of light duty electric and plug-in electric vehicles. The international committee includes automotive equipment manufacturers (OEMs), such as GM, BMW, Ford, Nissan and Toyota; Tier 1 suppliers Delphi, Panasonic and Magna; WPT suppliers, such as Qualcomm and LG, and a collection of other organizations, such as the Argonne National Laboratory, the EPA, the DOT, UL and the University of Tennessee.

The current version of SAEJ2954 technical information report (TIR) was made available for purchase from the SAE website on May 31, 2016. The latest version addresses unidirectional charging from grid to vehicle and is intended to be used for static charging applications. Dynamic charging (charging while the vehicle is in- motion) and bidirectional power may be considered in the future based on industry feedback. This TIR specifies a common frequency band using 85 kHz band (81.39 kHz– 90 kHz) for all light duty vehicle systems. In addition, it defines the power levels in four classes as shown in Table 1.1.

Table 1.1 SAE J2954 Light Duty Vehicle WPT charging classes

 

Classification	WPT1	WPT2	WPT3	WPT4
Power Levels	3.7 kW	7.7 kW	11 kW	22 kW
Status	Specified in TIR J2954	Specified in TIR J2954	To be Specified in TIR J2954	To be Specified in TIR J2954

 

 

 

 

 

SAE TIR J2954 compatible systems have been built by automakers and suppliers and are currently under test with a cross-industry team with the US Department of Energy, Idaho, and Argonne National Labs [46]. The experimental data, first in the bench and later in the vehicle, will be used to finalize the Standard by 2018 to support the roll-out of this technology.

The author would like to mention at this point that SAE TIR J2954 was made available to download only after 31st of May 2016 and the author of this thesis started working on the hardware setup in the middle of 2015. Therefore, it was not possible to follow the frequency guidelines in the TIR as it was unknown at that time. Before the release of TIR J2954, 20 kHz to 200 kHz was a very common frequency of interest in RIPT. Therefore, 40 kHz frequency was used for this thesis based on the availability of components and signal processing unit available in the lab at the time. For power level, charging standards defined by SAE J1772 for plug-in electric vehicle conductive charging [46] were followed. Therefore, for power ratings, 3.6 kW for level 2 charging was found appropriate depending on the rating of the load and power supply available in the lab. Table 1.2 gives the charging standards defined by SAE J1772 used in this thesis. However, the work presented in this thesis is scalable to all the power levels and frequency band defined by TIR J2954.

Table 1.2 Charging standard defined by SAE J1772 [46]

 

Charging Level		Power Level	Supply Voltage/Current	Setting
AC Level 1		1.7 kW	Single Phase 120 V/20 A AC (16 A continuous)	Residential/Parking lot
AC Level 2	Minimum	3.4 kW	Split Phase 208/240 V/20 A AC (16A continuous)	Residential/Commercial
	Maximum	19.2 kW	208/240 V/20 A AC (80 A continuous)	Commercial

 

 

 

DC Level	Level 1	40 kW (up to 500 V at 80 A DC)	3-Phase 208 V/480 V AC ~20 A-200 A AC	Commercial
	Level 2	100 kW (up to 500 V at 200 A DC)	500 V/200 A DC ~20A-400 A AC	Commercial

 

1.5       Research Goals and Objectives

The primary goal of this thesis is to design and implement a prototype of 3.6 kW of the wireless charger for static charging of Li-Ion batteries used in EVs. It is evident that the design of the entire inductive charging system includes more than one academic area such as power electronics, control system, magnetic circuit design, automation, communication system, mechanics, economics, and EMI regulations. Since it is impossible to cover all mentioned aspects, this thesis will focus on the control and magnetics design aspect of the charger. To meet this goal, several objectives are formulated:

1)   To investigate the magnetic characteristics of an unsymmetrical coil pair employing Archimedean spiral, with the aim of finding the coil pair least sensitive to coupling variations

A fundamental challenge of implementing an RIPT system for wireless power transfer is a coupling variation between the primary and secondary coils due to misalignment while parking a car over the primary coil. It is well known that deviation from an optimal coupling condition degrades the power transfer efficiency. It is thus desirable to have a coil pair which is least sensitive to misalignments to ensure efficient power transfer over a wide operating area.

In the literature, different types of charging pads have been proposed such as: circular pads, bipolar pads, DD-Q pads, and H-shaped pads. Among these, circular charging pads employing Archimedean spiral is the most well-known and widely adopted pad

 

 

 

shape in EV battery charging application, despite its lower coupling compared to other pads. This is mainly because circular pads (CPs) are non-polarised and have the same tolerance to misalignment in all directions which makes them easier to operate. Due to their non-directional characteristics, the vehicle can approach them from any direction which ensures simplicity of use by drivers.

Due to these features the author wishes to adopt CPs for the RIPT charger. However, CPs presented in the literature employ an identically sized Archimedean spiral coil in both the primary and secondary side. For such a symmetrical coil pair, a magnetic null occurs in their coupling profile for misalignment equal to 40% of the pad diameter irrespective of separation between them. This severely limits the operation area over which power transfer takes place.

Therefore, for this thesis instead of a symmetrical coil pair, the magnetic characteristics of a CP employing unsymmetrical coil pairs have been thoroughly analysed. From the analysis, an asymmetrical coil shape which gives magnetic null at 71% of pad diameter has been derived. It is also established that the position of null, shifts with changing air-gap for unsymmetrical coil pairs.

2)   To develop simple and accurate design guidelines for an SS-RIPT system considering bifurcation issue

The phenomena of existing more than one zero phase angle frequency in an RIPT system is known as the bifurcation phenomena. In an SS-RIPT system, it occurs due to an increase in the coupling coefficient above a certain critical value which is a function of load and secondary inductance. If not addressed carefully, bifurcation can cause a decreases in the voltage gain of the system, which therefore will affect the power transfer capability and overall efficiency of the system. Moreover, due to bifurcation, hard-switching of the primary side inverter semiconductor occurs if operating close to the resonant frequency. Some authors handle this issue by adopting a sophisticated variable frequency control strategy [47]–[50] which increases the overall complexity and the cost of the system. For these reasons, bifurcation is best avoided by the design process. Design steps to eliminate bifurcation have not yet been reported in the literature. In addition to this, missing from the literature is a complete step by step design procedure addressing issues, such as: calculation of electrical parameters for a

 

 

 

given load profile; selecting the number of turns in coils; and air-gap between coils. Some authors use three-dimensional (3-D) finite element simulation for coil design. However, finite element simulations are complicated and may require a significant number of iterations before achieving the final design goals and therefore are time- consuming.

In this thesis, a design guideline for creating an SS-RIPT system employing a circular Archimedean spiral has been developed. Formulated design guidelines use simple analytical expressions for calculating the electrical parameters (self-inductance and mutual inductance) to achieve the desired load profile and use FEA analysis only for the verification purposes. Electrical parameters of the system are calculated in such a way that bifurcation does not occur for the entire load and coupling variations. This thesis has also covered analytical expressions for calculating the physical parameters of the coils, such as: number of turns; turn spacing; and outer and inner diameter.

The literature, always emphasizes on having a coupling coefficient value as high as possible. Evolution of polarised pads is evidence of this mindset. However, through the research for this thesis, it has been established that a coupling value lower than a critical value is desirable and sufficient for the efficient operation of the charger for an entire range of load variations. Circular coil geometry can easily provide this critical value of coupling and thus complicated polarised pad designs are not required at least not for stationary charging application.

• To derive a simplified mathematical model for SS-RIPT system for primary side control

In most RIPT systems, power flow regulation is implemented on the secondary side, which requires extra switching circuits and controllers between the rectifier and the load. Introducing an additional power converter is advantageous in cases where multiple secondary systems exist, with varying loads. However, in the case of static charging where only one secondary is coupled to the primary, it is preferable to achieve power flow regulation from the primary side converter itself, by varying the voltage magnitude of the HF inverter. Eliminating an additional power converter stage is important for having an efficiency figure close to those of plug-in charger value.

 

 

 

For the above reasons, in this thesis primary side control has been preferred. To control the power flow from the primary side, a suitably designed closed-loop controller is consistently needed, to improve tolerance to misalignments and parameter variations. For the design of control loops, a small-signal model of the system is required for calculating the appropriate value of the phase and gain margins of the system to guarantee a robust control.

However, small-signal modelling for an RIPT system has not been reported extensively in the literature and mainly covers the steady state model which is not useful for designing the control loops. Only a few research papers have reported the dynamic model of an SS-RIPT system. The extended describing function (EDF) is a well-known method for modelling resonant converters, due to its high accuracy. However, it requires complex mathematical formulation effort, evident from the small-signal model presented by some authors. Presented models in the literature are 9th order system.

For this thesis, a reduced dynamic model of an SS-RIPT system has been derived considering the soft-switching in the primary side inverter. From this reduced dynamic model, a small-signal model of the SS-RIPT system has been derived. The derived model is a 5th order system and requires less computational effort than other presented models in the literature.

The derived model accurately predicts the frequency response of the system in low- frequency regions (up to one tenth of the resonant frequency). In the high-frequency region, the frequency response of derived model doesn’t match the actual system frequency response but maintains the similar trend of variations. Since the derived model accurately predicted low-frequency behaviour, therefore it was found useful for the design of the output voltage and current controller. The current and voltage controllers derived from the small-signal model were tested on experimental setup for different dynamic conditions and were found to give robust performance.

Conclusion

8.1       Achieved results

• Finding Coil-Pair Least Sensitive to Misalignment: Compared to symmetrical circular coil-pairs, use of unsymmetrical circular coil-pair increases tolerance to horizontal Due to its unsymmetrical shape, the position of the magnetic null in the coupling profile shifts with the change in separation between the coils. By operating the primary and secondary pads at distinct separation, a position of null can be shifted further away from the centre of the coils. This, in turn, increases the effective area, over which a given amount of power can be transferred to the load. After investigating several unsymmetrical coil-pairs, it was found that the outer radius of the secondary coil should be kept equal to the outer radius of the primary coil, and inner radius of the secondary coil should be maintained greater than the inner radius of the primary coil to obtain the best coupling profile.

A 3.6 kW charging pad was fabricated based on the proposed unsymmetrical coil- pair. The designed charging pad shows magnetic null in its coupling profile at about 78% of the overall pad diameter. With the addition of shielding this magnetic null was found to occur at 71% of the overall diameter of the charger. This is a significant improvement over the symmetrically shaped circular charging pad presented in the literature in which magnetic null position occurs at about 38-40% of the overall diameter [12], [167]. Due to increased tolerance of circular charging pad, it can now be implemented in dynamic charging application as well.

• Development of Bifurcation Free Design Guidelines: A set of novel and easy to follow bifurcation free design guidelines for calculating the electrical parameter of SS- RIPT link has been presented. A SS-RIPT link designed using this analytical approach was found to operate without any bifurcation phenomena for the entire load range and coupling variations. Therefore, sophisticated control strategies are not required for the operation of the charger in varying load and misalignment conditions. To fabricate the coils based on the calculated electrical parameters, accurate and straightforward analytical expressions have been formulated based on wheelers formula. Proposed design guidelines reduces the complexity of the controller and therefore overall cost of the
• Derivation of a Simplified Small-Signal Model for SS-RIPT System: A simplified mathematical model was derived for designing the voltage and current control loops for the primary side control. A mathematical model is of fifth order as compared to the ninth order model existing in the literature. This model has been derived assuming a reduced dynamic model of the SS-RIPT system suitable for the control of power flow from the primary side. The model was verified using a frequency sweep performed in the PLECS simulation tool. The derived model very accurately predicts the low- frequency behaviour, and therefore was found useful for designing the output voltage and current control loops. Designed control loops were implemented on a lab built 3.6 kW prototype of the SS-RIPT based wireless Both the current control loop and the voltage control loop were found to give optimum performance for different dynamic case studies.

The experimental results are in good agreement with the analysis presented in this thesis. A DC-DC efficiency of 91 % is achieved for the designed charger for rated load conditions. Even at a partial load condition (less than 50% of rated load), the efficiency is 87%. For the experimental setup, shielding was not designed. Therefore, the impact of adding shielding to the efficiency was analysed using FEA analysis in JMAG. The inclusion of aluminium shielding generates eddy current loss in them. However, due to the decrease in the number of turns in shielded charger, the overall efficiency remains almost constant. The efficiency of the SS-RIPT link was measured using JMAG and is found to be 98.47% for the unshielded design and 98.25% for the shielded design. Therefore, the final charger with shielding is expected to give the same value of DC- DC efficiency under the same experimental setup conditions. Moreover, all the experiments were performed in a non-ideal environment. The table used for equipments, such as CRO and DC load, was made of steel and therefore acts as an undesirable load to the unshielded charger which was kept in the proximity to the measuring equipment due to limited lab space. Since the equipments themself contains metal, they acts as a passive load to the leakage flux. For the completely shielded charger, DC-DC efficiency is expected to be slightly better.

The efficiency figure does not include the front-end power factor and correction stage, but with recent publications [171], it can be shown that the efficiency of this stage

can reach as high as 98%. Factoring this component, the grid to battery efficiency of 89% is expected from the designed charger.

Further, under the ICNIRP 2000 and IEEE Std. C95.1-2005 guideline, the safety evaluation for the designed 3.6 kW SS-RIPT based charger was conducted in JMAG. The average magnetic flux density at 0.8 m from the center of the pad is 14.5 μT. This value is within the safety guidelines recommended by both ICNIRP 2010 as well as the IEEE.

 

 

8.2       Future Developments

For making the system more useful for practical implementation, the following improvements are suggested:

1. For the successful integration of the charger into the vehicle, shielding of the secondary pad is For this purpose, the final charging pad presented in Chapter 6 should be implemented for hardware testing to examine its usefulness in practical situations.
2. Asymmetric coil-pairs increases the tolerance to However, no systematic relationship between different dimensions of the asymmetric coils has been presented in this thesis. It would be very helpful for a designer, if a generally applicable rule of null point as function of coil dimension could be developed for the coil-pair P- S4.
3. As observed from the experimental results, the efficiency decreases with increasing Therefore, for efficient charging of EVs, it is necessary to park the secondary pad in close alignment with the primary pad. This requires the development of a coil positioning system to assist drivers in parking the car over the primary pad.
4. For the hardware setup, BNC cable was used to transfer the sensed voltage and current to the DSP module. However, for practical implementation, development of an infrared communication (as it is fast and immune to the electromagnetic interference) and associated signal conditioning would be required for sending the information from the load side to the primary
5. Bidirectional power transfer can be investigated by replacing the rectifier on the load side with the active rectifiers to increase the overall efficiency of the system. More importantly, it allows for the scope of dual side control, therefore can provide more controlling