THE serious environmental pollution caused by internal

Transcription

1 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH Design and Implementation of Shaped Magnetic-Resonance-Based Wireless Power Transfer System for Roadway-Powered Moving Electric Vehicles Jaegue Shin, Member, IEEE, Seungyong Shin, Yangsu Kim, Seungyoung Ahn, Member, IEEE, Seokhwan Lee, Guho Jung, Seong-Jeub Jeon, Member, IEEE, and Dong-Ho Cho, Senior Member, IEEE Abstract In this paper, the design and implementation of a wireless power transfer system for moving electric vehicles along with an example of an online electric vehicle system are presented. Electric vehicles are charged on roadway by wireless power transfer technology. Electrical and practical designs of the inverter, power lines, pickup, rectifier, and regulator as well as an optimized core structure design for a large air gap are described. Also, electromotive force shielding for the electric vehicle is suggested. The overall system was implemented and tested. The experimental results showed that 100-kW power with 80% power transfer efficiency under 26-cm air gap was acquired. Index Terms Core structure, online electric vehicles (OLEVs), pickup, roadway-powered electric vehicles, wireless power transfer. I. INTRODUCTION THE serious environmental pollution caused by internal combustion engines, together with the depletion of fossil fuels, has motivated global interest in eco-friendly energy. Notably, electric vehicle technology has been developed to reduce the use of fossil fuels in vehicles, which are the main fossil fuel consumers. As a result, hybrid electric vehicles that use both a combustion engine and an electric engine have already been widely commercialized. However, all-electric vehicles, such as plug-in electric vehicles and battery electric vehicles, are distributed narrowly at present owing to some battery-related Manuscript received May 6, 2012; revised October 21, 2012, December 13, 2012, and February 26, 2013; accepted March 21, Date of publication April 16, 2013; date of current version August 23, This work was supported by the Transportation System Innovation Program of the Ministry of Land, Infrastructure, and Transport of the Republic of Korea. J. Shin, S. Shin, Y. Kim, S. Lee, and G. Jung are with the Wireless Power Transfer Research Center, Korea Advanced Institute of Science and Technology, Daejeon , Korea ( jkshin@kaist.ac.kr; jengsaebulan@kaist.ac.kr; soyangpa@kaist.ac.kr; shlee0322@kaist.ac.kr; ghjung9595@kaist.ac.kr). S. Ahn is with the Cho Chun Shik Graduate School for Green Transportation, Daejeon , Korea ( sahn@kaist.ac.kr). S.-J. Jeon is with the Department of Electronics Engineering, Pukyong National University, Busan , Korea ( jeub@pknu.ac.kr). D.-H. Cho is with the Department of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon , Korea ( dhcho@kaist.ac.kr). Color versions of one or more of the figures in this paper are available online at Digital Object Identifier /TIE drawbacks such as large size, heavy weight, high price, long charging time, and short driving range. These problems are not easily solved by current battery technology. In an effort to address battery problems, the concept of roadway-powered electric vehicles has been proposed. With this system, the electric vehicle is charged on the road by wireless power charging, and the battery can hence be downsized and no waiting time for charging is needed. Much research on wireless power transfer for electric vehicles has been performed over the past few decades. The Partner for Advanced Transit and Highways project of the University of California, Berkeley, developed a roadway-powered electric vehicle system with 60% power efficiency at an 8-cm air gap [1]. In this project, a powering roadway track was constructed and was experimentally validated. Design methods of loosely coupled inductive power transfer systems have been proposed to overcome the large air gap for practical operation on roadways [2] [8]. To achieve high efficiency of power transfer, many techniques, including resonant inverters for wireless power transfer [9] [12], efficient pickup modules [13] [15], effective pickup tuning methods [16] [18], and pickup voltage control methods [19], [20], have been proposed. The online electric vehicle (OLEV) center of the Korea Advanced Institute of Science and Technology has developed a high-efficiency roadway-powered electric vehicle system [21] [27]. The OLEV system achieved 100-kW output power with 80% power efficiency at a 26-cm air gap. The first generation of the OLEV system has already been commercialized in Seoul Grand Park, and some more pilot projects are being executed in the U.S. and Korea. This paper describes the design and implementation of a wireless power transfer system for moving electric vehicles with examples involving the practical OLEV system. Section II explains the basic design and introduces the system specifications and system operation. In Section III, the core structure design is presented. The designs of the inverter and the power line module are described in Section IV. In Section V, the pickup module, the rectifier, and the regulator are designed, and Section VI describes the electromotive force (EMF) shielding method. In Section VII, the implementation and experimental results of the system are explained and discussed, and conclusions are provided in Section VIII IEEE

2 1180 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 Fig. 2. Conceptual diagram of OLEV system. Fig. 1. Basic circuit model of wireless power transfer system. II. OVERALL SYSTEM A. Basic Design A wireless power transfer system uses inductive coupling. One of the most important factors that must be considered in designing an inductive coupling system is the target power of the system. Voltage and current ranges, usable devices, and operating frequency of the system depend on the target power. Because the wireless power transfer system for moving electric vehicles is a public service system that is installed in a road, the use of the resonance frequency must be permitted by the government. Generally, wireless power transfer systems for electric vehicles use kHz frequency. In the OLEV system, the target power is 100 kw, and the resonance frequency is 20 khz. The basic circuit of the wireless power transfer system is shown in Fig. 1. This circuit is fundamentally the same as the circuit model of transformers. In the circuit, a larger mutual inductance M facilitates more effective power transfer. The mutual inductance M is determined by L 1, L 2, and the coupling coefficient k, as follows: M = k L 1 L 2 (1) where k indicates the degree of coupling strength and is between zero and one. However, k of a wireless power transfer system for moving electric vehicles is very small due to the large air gap distance between the bottom of a vehicle and the road, which is necessary for safe driving. Therefore, the OLEV system uses large 72-cm core width. To compensate the reactive power and increase the power efficiency, compensation capacitors are used in the OLEV system. These capacitors make the circuit resonate at the operating frequency and minimize the circuit impedance. There are four basic compensation topologies for primary and secondary sides: SS compensation, SP compensation, PP compensation, and PS compensation, where S denotes series compensation and P denotes parallel compensation. Series compensation has voltage-source characteristics, and parallel compensation has current-source characteristics. The four compensation topologies have four different reflected impedances. The required primary compensation capacitors for the reflected impedances were well analyzed in [4]. According to the analysis, for a wireless power transfer system for moving vehicles, the use of SS compensation is recommended, because the required primary compensation capacitor depends on the mutual inductance M and load resistance R in SP, PP, and PS compensations. To charge the moving vehicles, the system must be tolerant of unavoidable lateral displacement, which leads to a change of mutual inductance, and thus, SS compensation is selected for the OLEV system. B. System Operation Fig. 2 shows a conceptual diagram of the OLEV system. A wireless power transfer system consists of a power transmitter part and a power receiver part. The power transmitter part is composed of an inverter and power lines. The inverter provides power, and the power lines carry current and generate magnetic flux. The power receiver part is composed of pickup modules, rectifiers, and regulators. The pickup modules generate power from induced voltage and current, the rectifiers convert ac power to dc, and the regulators control the output voltage, which is input to batteries and motors. The inverter receives power from an electric power company and converts 60-Hz operating frequency into 20-kHz resonance frequency. Although the inverter can be controlled to provide constant voltage, constant current control is more advantageous in dealing with changes in the load resistance or multipickup charging [11]. Therefore, in the OLEV system, the inverter converts 60-Hz power to 260-A constant current at 20-kHz resonance frequency. The power line modules are installed underneath the road and along the road. For economic construction and operation, the power line is installed at the start point, destination point, stopping area, and slope area. Power line cables in the module generate magnetic flux, and the magnetic cores make the flux propagate upward. In the pickup modules, the magnetic cores capture the magnetic flux from power line modules, which induces voltage along the coils. Compensation capacitors are connected to the coils to compensate the impedance of the inductance. The regulators are dc dc converters, which control the effective load resistance to control the voltage and current of the pickup modules. One among the buck converter, the boost converter, and the buck boost converter may be used to increase or decrease the voltage. In the OLEV system, the boost converter is used in consideration of the battery input voltage. Some of the transferred power is used to drive the motors, and the remainder is used to charge the batteries. When the vehicle stops, all of the power is used to charge the batteries. C. Segment Operation The power line of the OLEV system is divided into several power line segments, as shown in Fig. 3. To mitigate inefficient power supply, the inverter turns on only the segment on which

3 SHIN et al.: DESIGN AND IMPLEMENTATION OF WIRELESS POWER TRANSFER SYSTEM 1181 Fig. 3. Concept of power line segmentation method. a vehicle is located [24]. A sensor in a segment senses the approach of the vehicle and sends a message to the inverter, and then, the inverter turns the segment on. If the vehicle moves to the next segment, the inverter turns the previous segment off and the next segment on. This segment operation method reduces power loss in the cables. It also blocks EMF exposure to people on and near the powered road. III. CORE STRUCTURE Wireless power transfer systems use magnetic cores to maximize magnetic flux density. The basic core structure can be categorized as two types: EE type and UU type. The EE type is composed of a pair of E-shaped cores, and the UU type is composed of a pair of U-shaped cores, as shown in Fig. 4. In the EE-type structure, power line cables are wound around the center magnetic core pole, and two target magnetic loops and two main leakage magnetic loops are formed. In the UU-type structure, power line cables are wound around a pole or a horizontal bar of the primary magnetic core, and a target magnetic loop and two main leakage magnetic loops are formed. In the EE-type example in Fig. 4(a), R target and R leak,the equivalent reluctances of the target loop and the leakage loop, respectively, can be expressed as follows: R target = R air R air R C (2) R leak = R l R l = 1 2 R l. (3) To transfer the maximum amount of power, the coupling coefficient k should be maximized by adjusting the reluctance, which is the ratio of the effective magnetic flux through the loop to the total magnetic flux k = = Φ target Φ target +Φ leak = R leak R target + R leak R l 2R air2 + R air1 + R C + R l. (4) Therefore, 2R air2 + R air1 + R C should be minimized, and R l should be maximized. The reluctance can then be calculated by R = l μa = l μ o μ r A. (5) Reluctance is proportional to the length l of the path and inversely proportional to the permeability μ and cross-sectional area A. Therefore, larger permeability of the core is more advantageous for reducing the core reluctance R C. The relative permeability value of normal ferrite is over hundreds, and the ferrite used in the OLEV system has a relative permeability higher than If the permeability of the ferrite is sufficiently high, the core reluctance can be ignored, and the reluctance of the target loop is approximately R air2 +1/2R air1. To reduce R air1 and R air2, it is better to increase the pole widths w p1 and w p2 to enlarge the cross-sectional area of the core poles. However, if w p1 and w p2 are lengthened, the distance d p between the poles will be shortened by a given pickup width constraint, and the leakage magnetic flux will be increased. This is the reason why it is necessary to find the optimal pole widths w p1 and w p2 under the given pickup width constraint. The saturation of the core should also be considered. The saturation magnetic flux density of ferrite is about 0.4 T. As the magnetic flux density in the core becomes higher and closer to saturation, the rate of increment shrinks, and the core loss rises. Heat generated due to the core loss increases the temperature of the core, which may degrade the characteristics of the core such as core loss or permeability. Accordingly, if more magnetic flux is needed, the thickness of the core should be increased to reduce the magnetic flux loss, and this factor has to be considered with the pickup width constraint. Fig. 5 shows two types of core structures that are optimally designed for the OLEV system. In these designs, the cross sections of the core poles in the secondary side are maximized to decrease the magnetic reluctance. The structure in Fig. 5(a) is called a dual type because two magnetic flux loops are formed, and that in Fig. 5(b) is called a monotype because one magnetic flux loop is formed. In designing these core structures, the primary module width is minimized to reduce the cost of road construction, and the thickness is minimized to reduce the weight of the pickup modules. The designs were supported by theory, finite-element method computer simulations, and practical experiments. In the power line module, the ferrite cores are not continuous but separated at regular distances. This reduces cost and improves the solidity of the power line module underneath the road to support the weight of vehicles. The decrement of the induced voltage due to the space between the core blocks is negligibly small, as shown in Fig. 6. These two types of core structures have merits and flaws. First, under the same pickup width constraint, the dual type has an advantage in generating magnetic flux considering the saturation because the magnetic flux is divided into two separated paths. In the monotype, the thickness of the horizontal core bar must be two times greater for the same amount of magnetic flux compared to the dual type. Second, the monotype is more tolerant of lateral displacement than the dual type, because the pole distance d p is longer. Fig. 7 shows the magnetic flux loop shapes of the dual type and the monotype when a lateral displacement of 300 mm is given. In the case of the dual type, the center core pole of the primary side is closer to the left core pole of the secondary side than the center core pole of the secondary side, and this increases the leakage magnetic flux. However, in the monotype, the magnetic flux loop distortion is much smaller. Not only changing the core structure but also changing the coil windings was proposed to improve the lateral displacement tolerance in [13].

4 1182 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 Fig. 4. Magnetic circuits and equivalent electrical circuits of (a) EE-type structure and (b) UU-type structure. Fig. 5. Optimal core structure designs for 20-kW pickup module in OLEV system: (a) Dual type and (b) monotype. Fig. 6. Normalized induced voltage along pickup cables as a function of distance between core blocks in power line module. Fig. 7. Magnetic flux loop shapes of (a) dual type and (b) monotype when a lateral displacement of 300 mm is given. Fig. 8. Power circuits for proposed inverter system. IV. DESIGN OF POWER TRANSMITTER PART A. Road-Embedded Power Systems The power supply systems for the proposed system are shown in Fig. 8. The road-embedded power circuit consists of two parts: a three-phase power inverter and a road-embedded rail. The power inverter converts three-phase 60-Hz ac voltage to a constant single-phase 20-kHz ac current. In the internal circuit of the power inverter, three-phase ac voltage is rectified

5 SHIN et al.: DESIGN AND IMPLEMENTATION OF WIRELESS POWER TRANSFER SYSTEM 1183 Fig. 10. Output current controller of proposed inverter system. component exists in this equivalent circuit. From this, the related equations in the phasor domain can be written as follows: V inv = R 1 I 1 jωmi 2 (8) V Ld = jωmi 1 R 2 I 2. (9) Fig. 9. Equivalent circuits of proposed power transmitter system: (a) Actual equivalent circuit and (b) ideal equivalent circuit in perfect resonance. to dc. This dc voltage is converted to an isolated single-phase ac voltage source by a single-phase inverter. The inductance of the power cable in the embedded road rail L 1 has a value that varies with the length of the coil and is about 20 μh for 5-m coil. The coil is divided into two segments. Meanwhile, based on L 1, the resonant capacitance C r1 is adjusted to meet the resonant frequency of 19.8 khz which is little below the inverter switching frequency. The equivalent impedance of the embedded coil including C r1 operates as a small inductance at a noload state, and its current has a 90 lagging phase compared to the inverter output voltage. Its equivalent impedance becomes almost a resistor at a full-load state, and its current has the same phase as the inverter voltage. B. Analysis of Proposed Power Transmitter Circuit AsshowninFig.9(a),L 1 and R 1 at the primary side of the high-frequency transformer are the inductance and resistance created by an embedded road rail and transformer, respectively. C r1 is the resonant capacitance that makes the resonant frequency based on L 1 and C r1 equal to the inverter switching frequency f inv. Thus, the impedance made by C r1 and L 1 becomes almost zero. In this case, the obtained circuit equations in the phasor domain are as follows: V inv = 1 I 1 + jωl 1 I 1 + R 1 I 1 jωmi 2 (6) jωc r1 V Ld = jωmi 1 jωl 2 I 2 1 I 2 R 2 I 2 (7) jωc r2 where M is the mutual inductance between the embedded rail and the pickup module and has a low value in the presented system due to the large air gap distance. The equivalent circuit at a resonant frequency which is the same as the inverter switching frequency is shown in Fig. 9(b) in the phasor domain. All the inductances are canceled by the additional resonant capacitances, and thus, only the resistance With the use of the resonant capacitors, the large impedances due to L 1 and L 2 are removed, and the total impedances are reduced dramatically. C. Inverter and Electrical Segmentation Design The power line current is changed according to various load conditions, for example, the drive pattern of a car or the stateof-charge level of batteries. However, it is necessary to supply constant current regardless of whether the transfer power level is high or not, because we want a uniform magnetic flux density in order to support steady induction voltages in the pickup modules. Therefore, an output current regulator is designed by using a pulsewidth-modulation (PWM) inverter. In Fig. 10, the controller consists of a proportional integral (PI) control loop. The PWM method is used for steady-state switching of a fullbridge inverter for one segment. Segmentation is accomplished using an electrical switching scheme, as shown in Fig. 8. Without using mechanical switches, which have poor performance and a short lifetime, we can switch from one segment to another by using another insulatedgate bipolar transistor (IGBT) arm. D. Power Line Module The power line cables are located on the primary ferrite, and 260-A/20-kHz currents are supplied from the inverter. These currents consistently generate magnetic flux for the pickup. The litz wire structure is used for the power line cables to avoid the skin effect. The cross-sectional area of the power line cables is nearly 95 mm 2. The area issue is important when a high current is supplied to the cable. Our analysis result proved that an area of 1 mm 2 can endure the current of 3 A. Fiber-reinforced plastic (FRP) tubes are used to prevent the power cables from being broken. FRP is durable enough to tolerate the weight of vehicles and is uninfluential electrically and magnetically. The FRP tubes physically protect the power line cables from underground pressure, heat, and water. To operate the power line segmentation method, the magnetic sensors are located at the entry of the segment to notice the arrival of electric vehicles. When an electric vehicle is sensed by the sensor of the first segment, the first power line segment turns on. Moreover, when the second sensor senses an electric vehicle, the first power line segment turns off, and the second

6 1184 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 Fig. 11. Proposed power line module structure. Fig. 13. Peak voltages (a) in conventional capacitance connection and (b) in proposed distributed capacitance connection. Capacitors are connected to the cables to compensate the inductance of the winding cables. To make the impedance zero at the operating frequency, the capacitance is selected by Fig. 12. Proposed pickup module structure. segment turns on. The signal cable delivers the notice message from the sensor of the power line module to the inverter controller. The power line module consists of ferrite cores, segment power cables, signal cable, common power cables, FRP tubes, and aluminum tubes, as shown in Fig. 11. The FRP tubes are used to protect the power cables and sensor signal cables. The common power cables are used to connect each segment to the inverter and need to be shielded so as not to affect the segment power cables magnetically. Thus, aluminum tubes are used for the common power cables. Aluminum has low permeability and high conductivity, which are good shielding characteristics against magnetic flux. V. D ESIGN OF POWER RECEIVER PART A. Pickup Module The pickup module is composed of ferrite core blocks, pickup cables, compensation capacitors, and an FRP case. The case protects the inside physically and insulates the outside from electrocution. Similar to the power line cables, pickup cables are litz wires. Fig. 12 shows a pickup module without the cover. In Fig. 12, the ferrite core blocks are arranged as shown in Fig. 5(a). The pickup cables are wound around the ferrite cores. In the OLEV system, the center winding has 64 turns of the pickup cables, and the left and right windings have 28 turns, respectively. To add up the induced voltages, the center winding is clockwise, and the other windings are counterclockwise. A half of the center winding is connected to the left winding, and the other half is connected to the right winding. C r2 = 1 4π 2 f 2 L 2. (10) In the OLEV pickup, the reactive voltage due to a very large L is near 10 kv. This high voltage is difficult to insulate even if compensated. Hence, the capacitance is not connected all at one point but divided and distributed among the winding cables to reduce the peak voltage. This connection method drastically reduces the peak voltage caused by the inductances, as shown in Fig. 13. B. Rectifier and Regulator The rectifiers convert ac current to dc current. The regulator boosts the voltages of the rectifiers up to a reference voltage suitable for battery charging. Five 20-kW pickup modules are used to obtain 100-kW power capacity in the OLEV system, and each pickup module has two separated windings. Therefore, the OLEV system has ten rectifiers and one regulator for ten windings. As shown in Fig. 14, the inputs of ten boost converters are connected to the outputs of rectifiers, respectively, and the outputs of ten boost converters are connected to a battery. The ten boost converters output voltages are controlled by adjusting the ten PWM signals of the IGBTs. The OLEV regulator controls the ten phases of the PWM signals so that each difference among ten phases uniformly becomes 36.Asa result, output voltage ripples are minimized by minimizing the sum of the input current variations. Fig. 15 represents the input currents of ten boost converter modules when the duty ratio is 0.5. Fig. 16 shows output voltage ripples of ten boost converter modules according to the number of IGBT PWM signal phases. As the number of phases becomes larger, the output voltage ripples of the regulator become smaller under the duty control.

7 SHIN et al.: DESIGN AND IMPLEMENTATION OF WIRELESS POWER TRANSFER SYSTEM 1185 Fig. 17. Functional block diagram of proposed regulator controller. calculated by the following equations of the Ziegler Nichols method: K p = K u 2.2 (11) Fig. 14. Functional diagram of OLEV power receiver system. K i = 1.2K p. (12) T u Here, T u is the oscillation period. Furthermore, the gains are slightly adjusted by trial and error method through experiments for a moving vehicle. In Fig. 17, I ref1 and I ref2, the reference values of the currents, are controlled to balance the voltage of the two separated windings in one pickup module by using the following equations: Fig. 15. Input ripple currents of ten boost converters when regulator controls the phases of ten boost converters to make each phase difference to be 36 (control period = 360 ). Fig. 16. Output voltage ripples of regulator for the numbers of phases of boost converters. The control algorithm used in the regulator is shown in Fig. 17. The regulator uses dual PI controllers. Two PI controllers are used to control the input currents of two boost converters, and also, one PI controller is used to control the output voltage of the regulator. The IGBTs gates in Fig. 14 receive PWM signals constructed by the algorithm in Fig. 17. To tune the PI controllers, the Ziegler Nichols method is used. When the bus is arranged on the power line, ultimate gain K u is obtained. Proportional gain K p and integral gain K i are I offset =(V 1 V 2 ) Scaling factor (13) I ref1 = I cmd + I offset (14) I ref2 = I cmd I offset (15) where V 1, V 2, and V out are the input voltage of the first boost converter module, the input voltage of the second boost converter module, and the output voltage of the regulator, respectively. I cmd is the output value of the first PI controller. I offset, which indicates the difference between V 1 and V 2,is added to and subtracted from I ref1 and I ref2, respectively. By this procedure, the regulator absorbs more current from the winding that has higher voltage, and the winding voltage is decreased by the increased current and parasitic resistance. Inversely, the voltage of the other winding is increased, and finally, the two voltages become equal. In this balance, the sum of two duty ratios is minimized, which maximizes the efficiency of the whole regulator. VI. DESIGN OF EMF SHIELDING The design of the electromagnetic field distribution is also a significant factor in a wireless power transfer system, particularly in a high-power transfer system. As a magnetic flux of more than hundreds of thousands of milligauss from the current of thousands of amperes is generated between the power lines and pickup coils, even 0.1% of leakage from the main flux can be hundreds of milligauss, which is several times larger than the magnetic flux regulation suggested by the International Commission on Non-Ionizing Radiation Protection [30]. Therefore,

8 1186 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 Fig. 18. Passive shielding for the magnetic field from transmitter and receiver coils using a metal plate installed underground and soft contacts based on a metal brush. Fig. 20. Mechanism of the magnetic field cancellation using a reactive resonant current loop. be increased when larger lateral displacement is required. As the position of the metallic brush is controllable, the metallic brush does not touch the ground when the vehicle is moving, and it comes down to the ground and contacts to the ground only when the passive shielding is required. Fig. 19. Implemented passive shield structure consisting of vertical ground shield, metallic vehicle body, and metallic brushes connecting ground shield and vehicle body. it is essential to control the leakage magnetic flux for a wireless power transfer system in automotive applications because of the high current in the system. A. Passive Shielding Although the magnetic field is shaped by the coil design and ferromagnetic material, some form of shielding for a reduction of the EMF from the transmitter and receiver coils is required. A traditional passive shield using a metallic plate or a grid can be used to ensure low-frequency magnetic field reduction. A vertical ground metal plate is installed underground. Additionally, soft electrical contacts between the vertical ground plate and the bottom of the vehicle s metallic sidewall using metal brushes are used to construct a passive shield. Each brush consists of a bundle of metal wires attached beneath the bottom of the vehicle s sidewall. An example of passive shielding for an automotive application which is adjusted upward or downward is shown in Fig. 18. When the number of connections using metal brushes increases, the EMF level has significantly decreased from 144 to 35 mg. Fig. 19 shows the implemented passive shield structure which consists of buried vertical ground, metallic vehicle body, and metallic brushes which electrically connect the vertical ground and vehicle body. As the brush has 25-cm width and the exposed part of the vertical ground shield is 5 cm wide, the allowable lateral displacement of the passive shield is =30 cm (±15 cm). The width of the exposed part of the vertical ground shield and the width of metallic brushes can B. Reactive Shielding by Magnetic Field Cancellation Passive shielding is effective in blocking the magnetic field being emitted from the bottom of the vehicle to the side of the vehicle. However, the passive shield has a physical limitation when used in a wireless power transfer system because the shield should be placed between the transmitter and receiver coils. To maintain the air gap, active magnetic field cancellation methods have been suggested. The magnetic field from the pickup coil can be canceled by using an active shield; however, the size, weight, and power consumption of the additional power supply for an active shield are additional burdens for the OLEV system. We used a reactive resonant current loop instead of a traditional active shield which combines the advantages of a passive shield and an active shield. A reactive shield works as an active shield, but it does not require any additional power source to generate the intentional fields. Fig. 20 shows the mechanism of reactive shielding by magnetic field cancellation using the reactive resonant current loop method. When a reactive resonant current loop is placed between a magnetic field source and the measurement point, the original magnetic field induces voltage at the loop, and the voltage generates current which provides the canceling magnetic field. The induced magnetic field for cancellation can be controlled by the capacitance value as the current at the reactive resonant current loop is determined by the induced voltage and total impedance of the loop as follows: V induced = dφ dt = d B source ds (16) dt I = V induced = V induced Z loop jωl + 1. (17) jωc Fig. 21 shows a block diagram of the reactive shielding method for the OLEV system. The magnitude and phase are controlled by changing the capacitance of the reactive resonant loop to minimize the EMF. To control the capacitance optimally, the feedback system using a magnetic field sensing

9 SHIN et al.: DESIGN AND IMPLEMENTATION OF WIRELESS POWER TRANSFER SYSTEM 1187 Fig. 23. Implementation of OLEV inverter. Fig. 21. Block diagram of the reactive shielding method using a reactive resonant current loop. Fig. 22. Placement of the reactive resonant current loop under the vehicle as seen from the side lower position. TABLE I INVERTER SYSTEM PARAMETERS loop is implemented to determine the magnetic field intensity at the measurement position, and the processor in the controller block finds the optimal combination of capacitors to minimize the total EMF by controlling the switches. The application of a reactive resonant current loop to the OLEV system is depicted in Fig. 22. VII. IMPLEMENTATION AND DISCUSSION A. Inverter and Power Line We used a two-segment circuit topology and parallel connection of IGBT modules for margin. The output ports are connected with a blocking capacitor and an isolation transformer, as shown in the output part in Fig. 8. The resonance capacitances used to compensate the resonance of the power supply line were placed after the transformer. The system parameters and controller parameters are shown in Table I. We used a parallel snubber capacitor for each IGBT arm to protect the devices. In the control algorithm in Fig. 10, the PI loop sets the desired dc voltage by using the PWM method. The implementation of the OLEV inverter is shown in Fig. 23. Air-forced heat sinks were used for the inverter system. We used fin-type heat sinks and a blower that has a 1000-m 3 /h volume rate. The measured voltage and current waveforms of the inverter are shown in Fig. 24. Fig. 24(a) shows the no-load states, and Fig. 24(b) shows the full-load states. Note that, in the noload state, the inverter has an inductive load that gives 90 lag of the current to the voltage and, in the full-load states, the inverter has a resistive load that gives no lag. The efficiency of the implemented OLEV inverter measured by using a power analyzer is higher than 96% efficiency when the output power capacity is 116 kw. The transitions of the inverter current are shown in Fig. 25. The velocity of the vehicle was 20 km/h, and the output power was 30 kw. The distance from the sensor to the segment was 1 m. The time from sensing to turn-on of the inverter was 8.9 ms. It was much shorter than the time that is needed for the vehicle to reach the power line segment (about 180 ms). Fig. 25(b) shows the current behavior due to the variation of the load, and there was no significant fluctuation. As the bus enters the power line, each pickup module is positioned above the power line sequentially. Hence, the variation in load is not rapid. The power supply line was implemented, as shown in Fig. 26. We used E-shape dual-type cores, and the power line was placed on the side of the E-core, as noted in Section III. Its inductance was measured as 15 μh. Therefore, together with the resonance capacitors, the resonance frequency of the power supply module was khz. B. Pickup We implemented the pickup module as shown in Fig. 27. The line inductances of the pickup module were measured as 3.87 and 3.98 mh, in the left and right submodules, respectively. We compensated these inductances with resonance capacitors so that the pickup module has a 20-kHz resonance frequency using the series compensation method. As shown in Fig. 28, the induced voltages of the pickup were 470 V at a no-load state and 390 V at a full-load state (20-kW output).

10 1188 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 Fig. 24. Voltage and current waveforms of inverter (a) in no-load state and (b) in full-load state. Fig. 25. Current waveforms of inverter at transition state (a) in the case of turning on and (b) in the case of there being variation of the load. Fig. 29. Implementations of (a) rectifier and (b) regulator. TABLE II RESULTS OF THE REGULATOR TEST Fig. 26. Fig. 27. Implementation of power supply lines. Implementation of pickup module. for each subpickup module. We used ten boost converter modules for each subpickup module. The input capacitor was 75 μf, the inductor was 520 μh, and the output capacitor was 150 μf for each boost converter. The first PI controller was operated at 750-Hz control frequency, and we set K P to 0.02 and K i to 0.8. The second PI controller was operated at 7.5-kHz control frequency, and we set K P to 1 and K i to 100. The scaling factor was set as The test results are shown in Table II. The test was conducted at a temperature of 40 C using a forced water cooling method with a nipple of 3/4 PT TAP (pipe taper thread). In a full-load state, the efficiency of the proposed regulator is over 98%. Fig. 28. current. Induced voltage of one pickup module as a function of pickup output C. Rectifier and Regulator Fig. 29 depicts the implementations of the rectifier and the regulator. The output capacitor of the rectifier is set to 360 μf D. EMF Shielding As the position and impedance of the loop are important factors, we have performed 3-D magnetic field simulation using ANSYS Maxwell. The ten independent reactive shields at the left and right sides are fabricated and attached at the side of the vehicle, and the capacitor arrays are tuned to protect the magnetic field from the five pickup modules. Fig. 30(c) shows one of the implemented reactive resonant current loops with a capacitor array and a loop cable. The total capacitor values are tuned to minimize the leakage magnetic field. We used an International Electrotechnical Commission measurement

11 SHIN et al.: DESIGN AND IMPLEMENTATION OF WIRELESS POWER TRANSFER SYSTEM 1189 TABLE III PICKUP COMPENSATION RESONANCE FREQUENCIES Fig. 32. Output powers of pickup modules versus output current of regulator. TABLE IV TEST RESULTS OF WHOLE SYSTEM IN LABORATORY ENVIRONMENT Fig. 30. Reduction of EMF through application of a reactive shield. (a) Measurement position. (b) Measured EMF level. (c) Photograph of an implemented reactive resonant current loop. Fig. 31. system. Construction and test of wireless power transmitter and receiver standard to evaluate the EMF characteristics. With the reactive shield, the magnetic field was decreased from 103 to 53 mg to satisfy the EMF regulation in Korea, as shown in Fig. 30. E. Testbed Implementation and Discussion We constructed all the modules in a laboratory environment, as shown in Fig. 31. We set the air gap to 26 cm (core to core) and placed pickups at the center of the power supply module, which creates a fully aligned pickup position. We compensated the resonance frequency of the entire pickup module system at 20 khz. Because of mutual inductances between pickups, we compensate each pickup, as shown in Table III. By this compensation, we obtained the maximum power transfer at 20-kHz resonance frequency. The results of the power output of the regulator are shown in Fig. 32. Using the proposed regulator controller and a tuning method, we can obtain almost the same power output for each pickup. The end-to-end efficiency of the entire OLEV system is shown in Table IV. The OLEV system has greater than 80% end-to-end efficiency, which is defined as the ratio of the output power of the regulator to the input power of the inverter in the case of over 100-kW output power. In the OLEV system, while the inverter is controlled to provide a constant current, the regulator may change the current of the pickup modules to adjust the output power. Hence, the power loss in the power transmitter part is almost constant, and the power loss in the power receiver part increases as the pickup current increases. When the pickup current is small, the constant power loss is dominant, and the power transfer efficiency increases as the pickup current and the output power increase. However, when the pickup current is large, the power receiver part loss rises drastically, and the efficiency decreases in the end. Fig. 33 shows the output power, the power loss, and the power transfer efficiency, which vary depending on the pickup current. In this experiment, the maximum efficiency was

12 1190 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 TABLE V WEIGHT COMPARISON OF PROPOSED SYSTEM WITH CONVENTIONAL SYSTEMS Fig. 33. Output powers, power loss, and power transfer efficiency as functions of pickup current. TABLE VI COST ESTIMATION OF POWER SUPPLY SYSTEM PER 1km Fig. 34. Power transfer efficiency of whole wireless power transfer system as a function of output power of regulator in field environment. 81.7% with 79.5-kW output power and 17.8-kW power loss at 19.3-A pickup current. According to the results, roughly 80-kW charging is recommended if high output power is not needed for fast charging or high-speed driving. F. Field Test and Discussion We tested our wireless power transfer system in field environment using the OLEV electric bus with the pickup and the regulator. In the field environment, the lateral displacement of the moving vehicle was considered. In the test, we charged our battery with 30 kw and used an additional 70-kW static resistive load bank to obtain 100-kW output power. The test results are shown in Fig. 34. Although over 100-kW output power capacity with over 80% efficiency was achieved at 0-cm lateral displacement, the output power and the efficiency gradually decrease as the lateral displacement increases. At a larger displacement of 15 cm, the efficiency is reduced to 70%, but there is no meaningful change within 10-cm displacement. Thus, solutions for the lateral displacement such as vehicle guidance methods and use of a core structure that is tolerant of lateral displacement are being researched at present. VIII. CONCLUSION This paper has presented the design and implementation of a wireless power transfer system for moving electric vehicles. To achieve high output power and power transfer efficiency, an inverter, power line modules, pickup modules, rectifiers, and regulators were optimally designed. A power line segmenta- tion method was also proposed. Considering EMF exposure to people, EMF shielding was designed to satisfy EMF level regulations. The entire system was implemented and tested. The system provided 100-kW power with over 80% power transfer efficiency at 26-cm air gap. The performance and actual operation were verified through simulations, experiments, and field tests. APPENDIX A WEIGHT OF PROPOSED SYSTEM A weight comparison of the proposed system with conventional systems is shown in Table V. PRIMUS and e-primus are bus models of the Hankuk Fiber company. APPENDIX B COST ESTIMATION OF PROPOSED SYSTEM Cost of Power Transmitter Part: The cost of the power transmitter part depends on the power requirement. We assumed that the maximum power requirement is 400 kw/km. Cost estimation of the power supply system per 1 km is given in Table VI under the assumption of mass production. Cost of Power Receiver Part: The total cost of an OLEV bus is approximately $73 900, as shown in Table VII under the

13 SHIN et al.: DESIGN AND IMPLEMENTATION OF WIRELESS POWER TRANSFER SYSTEM 1191 TABLE VII COST ESTIMATION OF OLEV BUS assumption of mass production. The implementation cost of the power receiver unit per kilowatt is approximately $89/kW. REFERENCES [1] Roadway powered electric vehicle project track construction and testing program phase 3D, Partners Advanced Transit Highways (PATH), Berkeley, CA, USA, Res. Rep. [Online]. Available: berkeley.edu [2] O. H. Stielau and G. A. Covic, Design of loosely coupled inductive power transfer systems, in Proc. Int. Conf. Power Syst. Technol., Dec. 2000, vol. 1, pp [3] C. J. Chen, T. H. Chu, C. L. Lin, and Z. C. Jou, A study of loosely coupled coils for wireless power transfer, IEEE Trans. Circuits Syst., vol. 57, no. 71, pp , Jul [4] C. Wang, G. A. Covic, and O. H. Stielau, Power transfer capability and bifurcation phenomena of loosely coupled inductive power transfer systems, IEEE Trans. Ind. Electron., vol. 51, no. 1, pp , Feb [5] M. Budhia, G. A. Covic, and J. T. Boys, Design and optimization of circular magnetic structures for lumped inductive power transfer systems, IEEE Trans. Power Electron., vol. 26, no. 11, pp , Nov [6] C. Wang, O. H. Stielau, and G. A. Covic, Design considerations for a contactless electric vehicle battery charger, IEEE Trans. Ind. Electron., vol. 52, no. 5, pp , Oct [7] J. Sallan, J. L. Villa, A. Llombart, and J. F. Sanz, Optimal design of ICPT systems applied to electric vehicle battery charge, IEEE Trans. Ind. Electron., vol. 56, no. 6, pp , Jun [8] T. Imura and Y. Hory, Maximizing air gap and efficiency of magnetic resonant coupling for wireless power transfer using equivalent circuit and Neumann formula, IEEE Trans. Ind. Electron., vol. 58, no.10, pp , Oct [9] H. Abe, H. Sakamoto, and K. Harada, A noncontact charger using resonant converter with parallel capacitor of the secondary coil, in Proc. IEEE APEC, Feb. 1998, vol. 1, pp [10] C. Wang, G. A. Covic, and O. H. Stielau, Investigating an LCL load resonant inverter for inductive power transfer applications, IEEE Trans. Ind. Electron., vol. 19, no. 4, pp , Jul [11] J. Meins, G. Buhler, R. Czainski, and F. Turki, Contactless inductive power supply, in Proc. 19th Int. Conf. Magn. Levitated Syst. Linear Drives, Sep. 2006, pp [12] M. Borage, S. Tiwari, and S. Kotaiah, Analysis and design of an LCL-T resonant converter as a constant-current power supply, IEEE Trans. Ind. Electron., vol. 52, no. 6, pp , Dec [13] S. Raabe, G. A. J. Elliott, G. A. Covic, and J. T. Boys, A quadrature pickup for inductive power transfer systems, in Proc. IEEE Conf. Ind. Electron. Appl., May 2007, pp [14] M. L. G. Kissin, G. A. Covic, and J. T. Boys, Steady-state flat-pickup loading effects in polyphase inductive power transfer systems, IEEE Trans. Ind. Electron., vol. 58, no. 6, pp , Jun [15] G. A. J. Elliott, S. Raabe, G. A. Covic, and J. T. Boys, Multi-phase pickups for large lateral tolerance contactless power transfer systems, IEEE Trans. Ind. Electron., vol. 57, no. 5, pp , May [16] J. James, J. T. Boys, and G. A. Covic, A variable inductor based tuning method for ICPT pickups, in Proc. 7th Int. Conf. Power Eng., Dec. 2005, vol. 2, pp [17] M. Zaheer, N. Patel, and A. P. Hu, Parallel tuned contactless power pickup using saturable core reactor, in Proc. Int. Conf. Sustain. Energy Technol., Dec. 2010, pp [18] G. A. Covic, J. T. Boys, A. M. W. Tam, and J. C. H. Peng, Self tuning pick-ups for inductive power transfer, in Proc. IEEE Power Electron. Spec. Conf., Jun. 2008, pp [19] H. H. Wu, J. T. Boys, and G. A. Covic, An AC processing pickup for IPT systems, IEEE Trans. Power Electron., vol. 25, no. 5, pp , May [20] H. H. Wu, G. A. Covic, and J. T. Boys, A series-tuned inductive-powertransfer pickup with a controllable AC-voltage output, IEEE Trans. Power Electron., vol. 26, no. 1, pp , Jan [21] J. Shin, B. Song, S. Lee, S. Shin, Y. Kim, G. Jung, and S. Jeon, Contactless power transfer systems for on-line electric vehicle (OLEV), in Proc. IEEE Int. Electric Veh. Conf., Mar. 2012, pp [22] S. W. Lee, J. Huh, C. B. Park, N. S. Choi, G. H. Cho, and C. T. Rim, On-line electric vehicle using inductive power transfer system, in Proc. IEEE Energy Convers. Congr. Expo., Sep. 2010, pp [23] G. Jung, B. Song, S. Shin, S. Lee, J. Shin, Y. Kim, S. Jeon, and D. Cho, High efficient inductive power supply and pickup system for on-line electric bus, in Proc. IEEE Int. Elect. Veh. Conf., Mar. 2012, pp [24] S. Shin, J. Shin, Y. Kim, S. Lee, B. Song, G. Jung, and S. Jeon, Hybrid inverter segmentation control for online electric vehicle, in Proc. IEEE Int. Elect. Veh. Conf., Mar. 2012, pp [25] J. Huh, S. W. Lee, C. B. Park, G. H. Cho, and C. T. Rim, High performance inductive power transfer system with narrow rail width for on-line electric vehicles, in Proc. IEEE Energy Convers. Congr. Expo., Sep. 2010, pp [26] Y. Kim, Y. Son, S. Shin, J. Shin, B. Song, S. Lee, G. Jung, and S. Jeon, Design of a regulator for multi-pick-up systems through using current offsets, in Proc. IEEE Int. Elect. Veh. Conf., Mar. 2012, pp [27] S. Ahn, J. Park, T. Song, H. Lee, J. Byun, D. Kang, C. Choi, E. Kim, J. Ryu, M. Kim, Y. Cha, Y. Chun, C. Rim, J. Yim, D. Cho, and J. Kim, Low frequency electromagnetic field reduction techniques for the online electric vehicle (OLEV), in Proc. IEEE Int. Symp. Electromagn. Compat., Jul. 2010, pp [28] C. Kim, D. Seo, J. You, J. Park, and B. H. Cho, Design of a contactless battery charger for cellular phone, IEEE Trans. Ind. Electron., vol. 48, no. 6, pp , Dec [29] A. P. Sample, D. A. Meyer, and J. R. Smith, Analysis, experimental results, and range adaptation of magnetically coupled resonators for wireless power transfer, IEEE Trans. Ind. Electron., vol. 58, no. 2, pp , Feb [30] Guidelines for limiting exposure to time-varying electric, magnetic, and electromagnetic fields (up to 300 GHz), Health Phys., vol. 74, no. 4, pp , Apr Jaegue Shin (M 13) received the B.S. and M.S. degrees from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 2007 and 2009, respectively. He is currently with the Wireless Power Transfer Research Center, KAIST. His main research interests include roadway-powered electric vehicles, wireless power transfer, and circuit design. Seungyong Shin received the B.S. and M.S. degrees from the Korea Advanced Institute of Science of Technology (KAIST), Daejeon, Korea, in 2009 and 2011, respectively. He is currently a Researcher with the Wireless Power Transfer Research Center, KAIST. His research interests include wireless power transfer, power electronics, robust and optimal control, and artificial intelligence. Yangsu Kim received the B.S. degree from Dankook University, Yongin, Korea, in 2009 and the M.S. degree from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in He is currently with the Wireless Power Transfer Research Center, KAIST. His research interests include analysis of wireless power transfer and receiver systems, and control and modeling of converters.

14 1192 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 61, NO. 3, MARCH 2014 Seungyoung Ahn (M 03) received the B.S., M.S., and Ph.D. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 1998, 2000, and 2005, respectively. In 2001, he was a Visiting Associate Research Fellow with the Singapore Institute of Manufacturing Technology, Singapore. In 2005, he joined Samsung Electronics, Suwon, Korea, as a Senior Engineer in charge of high-speed board design for laptop computer systems. From 2009 to 2011, he was an Associate Research Professor with the Department of Electrical Engineering, KAIST, with contributions on the electromagnetic field design for online electric vehicles. He is currently an Assistant Professor with the Cho Chun Shik Graduate School for Green Transportation, KAIST. His research interests include design of wireless power transfer system and electromagnetic compatibility designs for high-speed digital systems. Seokhwan Lee received the B.Eng. and M.Eng. degrees from Yonsei University, Seoul, Korea, in 2008 and 2010, respectively. He is currently with Mando, Seongnam, Korea. His main research interests include mechanical design and motor started generators. Seong-Jeub Jeon (M 82) was born in Busan, Korea, in He received the B.S. degree from Soong- Jun University, Seoul, Korea, in 1980 and the M.S. and Ph.D. degrees from the Korea Advanced Institute of Science and Technology, Daejeon, Korea, in 1982 and 2001, respectively. Since 1986, he has been with Pukyong National University, Busan, where he is currently a Professor with the Department of Electronic Engineering. From 2001 to 2004, he was also with Virginia Polytechnic Institute and State University, Blacksburg, VA, USA. He is interested in power theory, motor drive systems, dc dc converters, reactive power compensation, high-power-factor control of ac dc converters, and wireless power transfer. Dong-Ho Cho (M 85 SM 00) received the Ph.D. degree in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in From 1987 to 1997, he was a Professor with the Department of Computer Engineering, Kyunghee University, Seoul, Korea. Since 1998, he has been a Professor with the Department of Electrical Engineering, KAIST. His research interests include wired and mobile communication networks and protocols, online electric vehicles based on wireless power transfer, construction information technology (IT) convergence, and bio-it convergence. Guho Jung received the B.S. degree in electrical engineering from Hanyang University, Seoul, Korea, in 1992 and the M.S. and Ph.D. degrees in electrical engineering from the Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea, in 1994 and 1999, respectively. For ten years, he has been developing various high-voltage and high-current power supplies, inverters, and active filters for industrial applications. Since 2009, he has been a Senior Researcher with KAIST, where he has been researching wireless power transfer systems for electric bus and train, etc. His main research interests include design, analysis, control, and verifications for applying wireless power transfer methods to various systems.