Efficiency Improvement of High Frequency Inverter for Wireless Power Transfer System Using a Series Reactive Power Compensator

IEEE PEDS 27, Honolulu, USA 2-5 December 27 Efficiency Improvement of High Frequency Inverter for Wireless Power Transfer System Using a Series Reactive Power Compensator Jun Osawa Graduate School of Pure and Applied Sciences University of Tsukuba Ibaraki 35 8573, Japan Takanori Isobe, Hiroshi Tadano Faculty of Pure and Applied Sciences University of Tsukuba Ibaraki 35 8573, Japan Abstract This paper proposes a wireless power transfer system using a series compensator as a primary side capacitor. The is a circuit module that equivalently functions as a series variable capacitor by controlling semiconductor switches. The advantage of applying the to a primary side capacitor is giving a controllability of power factor for a high frequency inverter. Therefore, the optimum operation of the high frequency inverter can be achieved regardless of the coil parameters. Experimental results with an kw laboratory prototype confirmed that the proposed system can achieve an optimum operation and high efficiencies of the high frequency inverter. I. INTRODUCTION In recent years, wireless power transfer (WPT) systems are activity studied. Especially, WPT systems for Electric Vehicles (EVs) or Plug-in Hybrid Electrical Vehicles (PHEVs) are highly paid attention since it can improve safety and convenience of charge operation. A magnetic resonant coupling is the leading circuit topology of WPT for EVs or PHEVs since it enables high-power and high-efficiency power transfer with a large air gap. In this topology, inductive reactance of the coils are compensated by using resonant capacitors to reduce current or voltage applied to the coils and the power source. However, there is one challenge that the reactance of the coil is not fixed because the air gap between the primary and secondary coils is expected to change according to the parking position. So the optimum capacitance for the resonant capacitors can vary, and fixed resonant capacitors cannot achieve the optimum operation constantly in that situation. To address this problem, this paper proposes a WPT system using an active series reactive power compensator named (Gate Controlled Series )[][2] instead of the fixed resonant capacitors. The is a circuit module connected in series that equivalently functions as a variable capacitor. In the proposed system, the optimum operation can be realized by controlling the equivalent capacitance of the according to the reactance of the coil. The coil efficiency improvement using the on the secondary side has been proposed in [3]. This paper proposes an inverter Fig.. Circuit configuration of the gate controlled series capacitor (). optimum operation by using the on the primary side. This paper reports experimental verifications of the proposed system with an kw laboratory prototype and discusses the efficiencies of the high frequency inverter and the. II. SERIES COMPENSATOR A. Circuit Configuration and Features Gate controlled series capacitor () is one of FACTS (flexible ac transmission system) devices proposed as a series reactive compensator applied for ac power transmission lines. Fig. shows the configuration of the. The is consists of two full-controlled reverse conductive semiconductor switches and a capacitor. The is controlled by line frequency switching, and the two switches share their source potential therefore the gate drive circuit can be simple. The can achieve soft-switching in all the operating range and there is no conduction loss when current is conducting the capacitor. Therefore, loss in the semiconductor switches is comparatively low. B. Operation Principles 978--59-2364-6/7/$3. c 27 IEEE Switch S and S2 are complementary turned on with the line frequency and a half duty ratio. The phase angle difference between the line current phase and the gate signals, δ, is given for control the equivalent capacitance of the, C. Fig. 2 shows the schematic waveforms of the and Fig. 3 shows possible current paths. The has two types of current paths, one is conducting the capacitor as shown in (d) (e), and the other is conducting semiconductor switches as shown in (f). The capacitor voltage waveform has a zero voltage period which is achieved by modes and (f), and its 992

Current - Current - δ Current - δ S2 signal voltage - S2 signal voltage - S2 signal voltage - Fig. 2. Three waveform modes of the with different control phase angle conditions of δ =, < δ < 9, and δ = 9. and then, the equivalent reactance, X, can be derived as X = V /I ( 2δ = X c π sin2δ ). (4) π The current conducting semiconductor switches, i sw, within a half fundamental cycle( < θ < π) can be derived as { ( π i sw = 2 δ<θ< π 2 +δ) (5) 2Icosθ (else), Fig. 3. Possible current paths of the. duration can be controlled byδ. Whenδ =, flowing current conducts only the semiconductor switches and the capacitor is not charged, that meansc =. By increasingδ, flowing current conducts the capacitor and generated voltage in the capacitor increases. Finally, when δ = 9, flowing current conducts only the capacitor, that means C = C, where C is the capacitance of the equipped capacitor. Therefore, the can vary its equivalent capacitance from C to by controlling δ. The equivalent reactance of the can be derived by the similar way discussed in [4]. The injecting voltage, v gcsc, within a half fundamental cycle ( < θ < π) can be derived as { ( 2Xc I(sinθ cosδ) π v gcsc = 2 δ<θ< π 2 +δ) () (else), where the pure sinusoidal current as i = 2Icosθ, (2) is assumed. The rms value of the fundamental component, V, can be derived as 2 π V = v gcsc (θ)sinθdθ 2π ( 2δ = IX c π sin2δ ), (3) π The loss of the, P, is derived as P = π π r on i 2 sw(θ)+v F i sw (θ)dθ = π (r oni 2 (π 2δ +sin(π 2δ)) + 2V F Isin( π δ))), (6) 2 where r on is the on resistance of the MOSFETs, V F is the forward voltage of the free-wheeling diodes. III. INVERTER LOSS REDUCTION BY APPLYING A. Optimum Operation of High Frequency Inverter Generally, the inverter of WPT systems for EV is operated at high switching frequency, for instance 85 khz; therefore, the switching loss reduction is important to improve the system efficiency. A voltage-source type high frequency inverter with fundamental frequency switching, which means the output current frequency is same as the switching frequency, is used for the purpose. A lagging power factor at the inverter output, as shown in Fig. 4, is attractive for this topology, in which the turnoff is performed with mitigated dv/dt achieved by a snubber capacitor and/or device parasitic capacitance, and turn-on is performed with conducting free-wheeling diode, so that turnon is complete soft-switching. On the other hand, the leading power factor, as shown in Fig. 4, is not attractive since it results in shorting the capacitors and turn-off of the freewheeling diode that causes reverse recovery, and the switching 993

C C 2 M L L 2 R L Fig. 5. An equivalent circuit diagram of the WPT system of the series-series topology in the fundamental frequency. C =.68 nf C =2.79 nf power factor.5 C =3.49 nf 5 Mutual inductance [μh] Fig. 6. Example of the resulting power factor with varying mutual inductance, M, and fixed primary capacitance, C. Fig. 4. Schematic waveforms of a switching device of the high frequency inverter with Lagging power factor, Leading power factor. loss increases. And same phenomena occurs if the power factor is too high as the current crosses the zero during the dead time. Therefore, the high frequency inverter should be operated at lagging power factor that delays current beyond the dead time. At the same time, in order to minimize the voltage and current at turn-off and to reduce the turn-off losses, the power factor should be as high as possible. Therefore, operating the high frequency inverter at slightly lagging power factor which the current crosses zero at the end of dead time is optimum and can minimize the switching loss. B. Selection of Resonant In the WPT system in series-series circuit topology [5] which is shown in fig. 5, the secondary side capacitor can improve the efficiency of the coils by compensating for the secondary side self inductance and minimizing the primary current to achieve the same power transfer. Efficiency of the coils is maximized when the secondary side capacitor fully compensates for the self inductance of the secondary coil. The primary side capacitor improves the output power factor of the high frequency inverter. The capacitance of the primary capacitor is selected to operate the high frequency inverter at the slightly lagging power factor. When the secondary side capacitor fully compensates for the self inductance of the secondary coil, the power factor at Primary capacitance [nf] 4 3 2 pf=.8 pf=.9 pf= 5 Mutual inductance [μh] Fig. 7. Required primary capacitance, C, to achieve constant power factors for varying mutual inductance, M. the inverter output can be derived as ( pf = cos arctan (ωl ) /ωc )R L (ωm) 2, (7) where L is the self inductance of the primary coil, C is the capacitance of the primary side series capacitor, R L is the load equivalent resistance, and M is the mutual inductance between the coils. To operate the high frequency inverter at a lagging power factor, C should be selected appropriately to compensate for L partially. However, the power factor is influenced by the possible mutual inductance changing in WPT applications. Therefore, fixed capacitors can not achieve the optimum operation for varying mutual inductance. In conventional WPT systems using a fixed capacitor, the 994

Fig. 8. 2 2 6 L L Coils used in the experiment at primary side secondary side. 4 TABLE I S PECIFICATIONS OF C OILS Winding type Size (W / D / H) Wire specification Number of Parallels Number of Turns Core Resistance Primary Secondary Spiral type 3 mm / 3 mm / 3 mm Litz wire (φ.8 mm 3) 3 4 24 45 PC95 (ferrite).2 Ω.59 Ω primary capacitor should be designed to achieve a slightly lagging power factor when mutual inductance is the largest value, to achieve the ZVS turn-on for all the range of varying mutual inductance. For example, the resulting power factor with fixed L (3 µh), fixed RL ( Ω), and varying M is shown in Fig. 6. As can be seen from the figure, one capacitance can achieve unity power factor regardless of varying M ; however, the other values of capacitance can not achieve constant power factor as function of M. Therefore, fixed capacitance can not achieve a lagging constant power factor for varying M. Fig. 7 shows required C to achieve constant power factors for varying M. As can be seen from the figure, varying C is needed to maintain the power factor at a certain value except for unity power factor. This is the reason for the need of the active compensator for the WPT applications with possible miss alignment. In the proposed system, the power factor can be kept constant by controlling the equivalent capacitance of the connected in the primary side. IV. E XPERIMENTAL V ERIFICATIONS A. Experimental Setup To confirm the above discussion, experiments were conducted. Fig. 8 shows the picture of coils used in the experiments. The coil shape is a spiral winding type and the diameter is 3 mm. Litz wires were used to reduce the loss increased by skin effect and proximity effect, and ferrite cores are placed on the back side of the winding to increase the coupling coefficient and quality factor. In order to suppress the increase of the primary copper loss, which is the dominant loss component of the coils, the primary coil was designed to reduce resistance by increasing the number of parallels of the winding. L2 [μh] L,M [μh] 4 8 2 Fig. 9. Inductance characteristics of the coils used in the experiments as function of the horizontal gap. M Vout P P2 P3 P4 Fig.. The proposed circuit configuration for the WPT using the as a primary side capacitor. TABLE II E XPERIMENTAL C ONDITIONS Fixed cap. A Frequency Rated power Dead time Vout C C C2.28 nf Fixed cap. B 85 khz kw 8 ns 35 V 2.29 nf 3.9 nf with 32.9 nf 7.42 nf Table I lists specifications of the coils. In the experiments, the vertical gap between the primary coil and secondary coil was set constant at mm and the horizontal gap was changed from mm to 2 mm. Fig. 9 shows varying L, L2 and M of the coils as function of the varying horizontal gap. It can be seen from the figure that the self inductance L, L2 are almost constant; on the other hand, the mutual inductance M highly varies with the gap change. Fig. shows the circuit diagram of the proposed system using the as a primary side capacitor, which were evaluated in the experiments. Table II shows the circuit parameters of the experiments. The power supply voltage was adjusted to achieve a constant output power of kw for varying condition of the horizontal gap. The secondary side capacitor was selected to achieve full compensation for L2. Three cases with different primary side capacitors were evaluated as shown in the table. The first condition refereed as fixed capacitor A was given to achieve full compensation for L. The second condition referred as fixed capacitor B was set 995

- - - voltage [V] 5-5 2 voltage [V] 5-5 2 voltage [V] 5-5 2 Fig.. Measured waveforms of the current flowing through the (top), and the capacitor voltage of the (bottom) when the horizontal gap was mm, 6 mm, 2 mm. Power factor A B A B 4 8 2 4 8 2 Fig. 2. Measured power factors at the output of the high frequency inverter with varying horizontal gap. Efficiency [%] 95 9 A B 4 8 2 Fig. 3. Measured efficiencies of the high frequency inverter with varying horizontal gap. to achieve partial compensation for L to achieve a slightly lagging power factor when the mutual inductance becomes maximum. The third condition is using the, which is the proposed one in this paper. In the proposed system using the, the and Fig. 4. Measured output current magnitude in rms of the high frequency inverter with varying horizontal gap. a fixed capacitor were connected in series and divided the voltage in order not to exceed the rated voltage of the. The capacitance of the and the fixed capacitor were designed to achieve full compensation for the self inductance of the primary coil with the maximum compensation degree of the with some margin. At the same time, the voltage sharing of the should be high as possible within an available voltage rating of the device in order to reduce the conduction loss of the semiconductor switches. For the experiments, the capacitor in the was designed to be applied up to 8 V with considering the use of 2 V SiC- MOSFET. Then, the equivalent reactance in the fundamental frequency of the was controlled to obtain a slightly lagging power factor that delays current beyond the dead time. In the experiment, the power flows between components, P, P 2, P 3, and P 4, shown in Fig., were measured and losses were evaluated. B. Experimental Results Fig. shows the waveforms of the voltage of the and line current when the horizontal gap was set at mm, 6 mm and 2 mm. As can be seen from the figure, the 996

4 2 2 2 5 Free-Wheeling Diode current -2-5 5 4 Free-Wheeling Diode Current - - 5 2 2 2 5 Free-Wheeling Diode current -2-5 5 4 Free-Wheeling Diode Current - - 5 2 2 2 5 Free-Wheeling Diode current -2-5 5 Fig. 5. Measured current and voltage waveforms of the semiconductor devices of the inverter when the horizontal gap was mm. With fixed capacitor A. With fixed capacitor B. was used as a primary side compensator. Free-Wheeling Diode Current - - 5 Fig. 6. Measured current and voltage waveforms of the semiconductor devices of the inverter when the horizontal gap was 2 mm. With fixed capacitor A. With fixed capacitor B. was used as a primary side compensator. worked as an equivalent variable capacitor since the capacitor conduction section varied according to the horizontal gap. The current conducted the semiconductor switches only in a short section and the semiconductor loss was generated only in this section. Fig. 2 shows the output power factor of the high frequency inverter. In the case of fixed capacitor A, output power factor of the high frequency inverter was kept at almost unity power factor. In the case of fixed capacitor B, a slightly lagging power factor was achieved when the horizontal gap was mm; however, the power factor was significantly decreased according to the increase of the horizontal gap. On the other hands, in the case of the, the slightly lagging power factor was achieved regardless of the horizontal gap. Fig. 3 shows the efficiencies of the high frequency inverter. The fixed capacitor A achieved lowest efficiencies than the others regardless of the horizontal gap. The fixed capacitor B achieved higher efficiencies than the capacitor A; however, the efficiency was remarkably decreased by the horizontal gap increase. In the case of the, inverter efficiencies were always high and the efficiency decrease in high horizontal gap conditions was mitigated. Fig. 4 shows the output current of the high frequency inverter in rms. The output current of the high frequency inverter, which is therefore same as the primary coil current, to achieve the same output power was almost same regardless of the primary compensation. Therefore, the conduction loss of the MOSFETs can be same for all three conditions and the sole difference for the inverter efficiency is switching losses. Fig. 5 and Fig. 6 show the waveforms of current and voltage applied to the MOSFETs and external free-wheeling diodes when the horizontal gap was mm and 2 mm, respectively. With the fixed capacitor A, turn-on was hardswitching and the switching loss can be though to be high since the parasitic capacitance was charged before the turnon, and it was shorted by the turn-on. With the fixed capacitor 997

Loss [W] Rectifier Transformer Inverter Loss [W] Rectifier Transformer Inverter Loss [W] Rectifier Transformer Inverter 4 8 2 4 8 2 4 8 2 Fig. 7. Loss breakdown of the overall system with Fixed capacitor A, Fixed capacitor B, and. Efficiency [%] 95 9 A B 4 8 2 Fig. 8. Measured efficiencies of the overall system including the inverter, coils, rectifier, and compensators. B, a completely soft turn-on and a turn-off at low current were achieved when the horizontal gap was mm. However, the current at the turn-off was increased when the horizontal gap was 2 mm due to the lower power factor at the inverter output. That can increase the turn-off loss. On the other hands, in the case of the, the slightly lagging power factor achieved complete soft-switching turn-on and turn-off, and at the same time, the current at the turn-off could be maintained as low regardless of the horizontal gap. Fig. 7 shows the overall system losses. The system losses were divided into losses of the inverter, the coil part including compensators (capacitors and ) and the rectifier by taking differences between the measured power, P, P 2, P 3, and P 4, shown in Fig.. Then the losses of the were calculated by using equation 6 and separated from the coil losses. As can be seen from the figure, the loss of the coils was almost same regardless of the compensation at the primary side. The possible loss increase due to harmonic components generated by the is considered to be negligible. The loss of the was less than 3.5% of the overall loss and was much smaller than other loss components. The inverter loss using the was reduced by 8.4 W compared with the system using fixed capacitors A and B at a maximum. Fig. 8 shows the overall system efficiencies. The proposed system using the achieved high efficiency compare with the system using capacitor A in the whole area. When the horizontal gap was larger than 8 mm, the proposed system achieved high efficiency compare with the system using the capacitor B since the loss reduction of the inverter was remarkable and it exceeds the additional loss of the. V. CONCLUSION This paper proposed the WPT system using series compensator named as a primary side capacitor. The proposed system can control the output power factor of the high frequency inverter and minimize its switching loss. The main advantage of using the than the other topologies like a full-bridge converter as a series compensator is low semiconductor losses. Experiments were conducted with an kw laboratory prototype. The proposed system achieved the efficiency improvement of the high frequency inverter, and considerably low loss of the. The loss reduction of the inverter exceeded the loss increase by the introduce of the ; therefore, the system overall efficiency was improved when the variation degree of the coil parameters was large. The proposed system reduced maximum loss by 8.4 W compared with the system using fixed capacitors as a primary side capacitor. REFERENCES [] D. D. Karaday, B. R. Pilvelait and D. Maratukulam, Continuously regulated series capacitor, IEEE Transactions on Power Deliverly, Vol. 8, No. 3, pp. 348 355, 993 [2] E. H. Watanabe, L. F. W. de Souza, F. D. de Jesus, J. E. R. Alves and A. Bianco, -gate controlled series capacitor. a new facts device for series compensation of transmission lines, IEEE/PES Transmission and distribution conference and exposition: Latin America, pp. 98 986, 24 [3] T. Isobe, K. Kobayashi, K. Wakasugi, R. Shimada, Efficiency Improvement of Contactless Energy Transfer Systems Using Series Compensation Device Named MERS, 4th European Conference on Power Electronics and Applications (EPE 2), 2 [4] T. Isobe, A full-bridge AC power flow controller with reduced capacitance operated with both FFS (fundamental frequency switching) and PWM, in IEEE energy conversion congress and exposition (ECCE 24), Pittsburgh, PA, United States, 4 8 September 24. [5] Yeong H. Sohn, Bo H. Choi, Eun S. Lee, Gyu C. Lim and Gyu- Hyeong Cho, General Unified Analyses of Two- Inductive Power Transfer Systems: Equivalence of Current-Source SS and SP Compensations, IEEE Transactions on Power Electronics, Vol. 3, No., November 25. 998