069060 Secondary-side-only Siultaneous Power and Efficiency Control in Dynaic Wireless Power Transfer Syste 6 Giorgio ovison ) Daita Kobayashi ) Takehiro Iura ) Yoichi Hori ) ) The University of Tokyo, Graduate School of Frontier Sciences 5--5 Kashiwanoha, Kashiwa, Chiba, 77-856, Japan ) The University of Tokyo, Graduate School of Engineering, 7-3- Hongo, Bunkyo-ku, Tokyo, 3-8656, Japan 4 Presented at the EVTeC and APE Japan on May 6, 06 ABSTRACT: Electric vehicles and wireless power transfer are a convenient cobination. Past research considered either power control or efficiency control in wireless secondary side, but a ore advanced control is desirable. Therefore, the authors propose a control for power and efficiency with two converters entirely perfored on the secondary side of the wireless syste, independently fro the priary side. In this paper, a dynaic charging scenario is considered: the control ust provide the required power at axiu efficiency conditions while the utual inductance is changing. The experiental results verify that the proposed control effectively works during dynaic charging. KEY WORDS: dynaic wireless charging, power control, efficiency control. INTRODUCTION In the last few years, electric vehicles (EVs) have becoe a key point in saving energy, achieving high controllability by use of electric otors and reducing pollution. However, currently ost of EVs ust be charged via plug-in ethod: this process is tie consuing as well as carrying soe safety issues for the consuer such as the high voltage used during charging. Wireless power transfer (WPT) technology recently is becoing suitable for autootive applications. WPT by agnetic resonant coupling ) can achieve high power and high efficiency when the transission distance is lower than one eter because it allows robustness against isalignent, increased air gap and higher transission efficiency when copared to conventional induction ethod. There are different types of WPT by agnetic resonant coupling, depending on the copensation carried by the capacitances: one of the types used in autootive application is the series-series (SS) copensation because it allows high power to flow on the receiver side. Currently, static WPT is available on the arket; however, developing in-otion wireless power Fig. : Equivalent circuit of SS copensated WPT syste. transfer is necessary because it would ean extree reduction of battery weight and range anxiety; recently, soe trial units are been experiented upon and achieved proising results. A good converter control is necessary to send the necessary power to the vehicle with high efficiency ). In WPT, converter control is generally divided into power control and efficiency control; they can be perfored on priary side 3) 4), on secondary side 5) 6) or on both sides at the sae tie 7) 8). However, ost of the ties, in past research power and efficiency control have been perfored in different sides of the systes (e.g. efficiency control on Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved
Fig. : Reference circuit of WPT using SS copensation and equivalent coil configuration. priary side and power control on secondary side). A novel control able on the secondary side to transfer at the sae tie the desired power with high efficiency without any kind of counication with the priary side is desirable. In a WPT syste for EV battery charging, the secondary side usually include a rectifier and a DC/DC converter connected to the battery. Therefore, it is possible to control the efficiency and the power flow only by secondary side. The otivation to control only the secondary side consists in keeping the priary side siple, thus allowing a standardization for future coercial WPT installations. A paper dealing with this concept in a static charging scenario has been recently published 9). Therefore, in this paper, the authors propose a siultaneous power and efficiency control for the secondary side of a WPT syste with SS copensation for dynaic charging.. CASE OF STUDY This paper deals with the secondary side control of a WPT syste with SS copensation: it includes a half active rectifier (HAR), a DC/DC converter and a battery as load. The frequency in the WPT syste is the resonant frequency extracted fro the coil paraeters and is independent fro the load and the distance between the coils. The resonant angular frequency is given as follows: () C C with and C as the priary coil inductance and capacitance, respectively; siilarly, and C are the secondary coil inductance and capacitance. The coil internal resistance has no effect on the resonant frequency but affects the losses and the dynaic response of the syste. The utual inductance between the priary and secondary coil depends on factors such as the distance between coils, their geoetry, the edia between the, the presence of etallic parts in the coils vicinity, etc.. The reference circuit is shown in Fig.. Using a HAR and a DC/DC converter to control the efficiency and the power flow in the secondary side is proposed with respect to this circuit. 3.. Control concept 3. PROPOSED CONTRO In WPT theory, the forulae of the load power P and transitting efficiency, presented in previous papers 0) ) ), are given respectively by: Z Z R,0 P V () R Z R Z R Z R (3) where Z is the load ipedance seen fro the secondary coil, R is the secondary coil resistance, R is the priary coil resistance, is the utual inductance between the coils and V,0 is the rs value of the fundaental wave of priary side voltage. Fig. 3: HAR odes. In this paper, the power factor of the priary side voltage is assued to be unity, therefore the power factor of the fundaental wave of the secondary side voltage V is considered to be unity, too. The HAR is a full wave diode bridge whose low side diodes are replaced by active devices like Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved
Fig. 4: Transission efficiency and load power with respect to load ipedance for V 0 V and 39. μh. MOSFETs or IGBTs: by turning on the both the devices on the low side, the coil terinals are shorted (short ode) and no power is transitted to the load. Since the SS copensation of agnetic resonant coupling akes the secondary side coil behave like an equivalent current source, short-circuiting the coil terinals is allowed. On the other hand, if both the low side devices are off, the converter works like a single-phase rectifier coposed by diodes (rectification ode), as shown in Fig. 3, and power is sent to the load. The influence of load ipedance has been exained in 0) ) 3). In particular, the load ipedance associated with axiu transission efficiency and the one related to axiu deliverable power are different. The forer one is equal to: R Z (4), ax R R The secondary coil voltage that axiizes the transission efficiency can be expressed by: V R ax V,0 (5) R RR RR On the other hand, the load ipedance and the secondary coil voltage related to the axiu available power are given by: Z V, P ax R R (6) P ax V,0 (7) R Voltages in (5) and (7) are rs values of the fundaental wave coponent. Finally, the total input ipedance seen fro the secondary coil ust be Z, ax during rectification ode. On the other hand, when short ode happens the secondary input ipedance is only a few ohs. The control of the input Fig. 5: Control concept with paraeters wavefors (top to botto: HAR input voltagev, HAR input current I, HAR input power P, total efficiency ). ipedance allows to use the HAR to send the power desired by the load to the DC/DC converter, which in turn is used to guarantee axiu transitting efficiency. Fig. 4 shows the relationship between power and efficiency with respect to the load ipedance. The power corresponding to axiu transitting efficiency is fixed. However, by the proposed secondary side control, it is possible to send a different power value while retaining the sae transitting efficiency. This is because of the switching between short ode and rectification ode by HAR. The length of the short ode during the HAR switching period deterines the aount of power received by the load. Fig. 5 shows how the control works and how the secondary side input paraeters are shaped. In rectification ode, the rs value of secondary side voltage is unchanged, therefore the rs value of the secondary side input power is unchanged. The secondary side coil behaves like an equivalent current source so the current does not change as well. The power is calculated by averaging, as well as the efficiency; consequently, the averaged efficiency coputed fro the ratio between secondary power and priary power will be lower because of the resistive losses happening during short ode. In fact, during rectification ode the efficiency is axiu while in short ode it is zero since the load is not supplied. Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved
3.. Control operation In Fig. 6 and Fig. 7 are represented the control blocks for the HAR and DC/DC converter, respectively. The feedback part of both the controllers is designed with the pole placeent ethod. C PID ( s) k p ki kd s (8) s s The HAR can be thought as a non-linear resistance; however, in this paper dynaic charging is considered, therefore transient ust be included. Then, its plant includes only one pole due to secondary DC soothing capacitor P HAR C DC: PDC ( s) P ( s) (9) d( s) R C s DC with P as the HAR input power and R as the equivalent circuit load resistance. However, instead of a PID controller, a PI one is considered sufficient. In the proposed control, the following requireents ust be et: ) Priary side voltage source and syste frequency are given and fixed. The duty cycle of the priary converter is fixed, consequently the priary coil voltage is fixed, too. ) The HAR switching frequency is at least one order of agnitude lower than the DC/DC converter one to avoid operation conflict and potential instability. 3.. HAR control The HAR odulates the power by adjusting the duty cycle to switch between rectification ode and short ode. Therefore, a voltage forula related to the desired power * P and the axiu power is necessary. Fro (7), the above entioned voltage is coputed as follows 4) : V * RR P * VP ax V P P ax (0) R By rearranging (0), the value of the desired power as a function of the voltage is obtained as follows: * P P ax R V V * V P P ax () R R Since the ai is to achieve high transission efficiency, the upper liit of the desired power is reached when V * is equal to P V ax ; therefore, () is adjusted and becoes: P, ax P ax R V V ax VP ax () R R The feedforward part of the control in Fig. 6 is decided by the ratio between the HAR input power P and (). 3.. DC/DC converter control In this paper, a buck converter is considered. The control akes the DC link voltage equal to the axiu efficiency voltage, but the utual inductance ust be known. Since this paper deals with a dynaic charging scenario, the coils are oving and therefore the utual inductance changes. Therefore, a ethod to estiate the utual inductance is necessary. In 5), a real-tie estiation of the coupling coefficient by using secondary side DC link inforation and recursive least square (RS) filter has been proposed. In this paper, the DC/DC converter control concept is the sae. The estiation is iportant to the control because it affects also the HAR control: in fact, in () there are any factors related to it. Poor estiation leads to noisy, unstable references that severely reduce the efficiency. In particular, the operation of HAR is the ain cause of noisy estiation. This is because during short ode the DC link current is equal to zero: since the secondary DC current inforation is used as input in the RS filter, when its value becoes zero so does the estiated utual inductance and the voltage reference for high efficiency control. The RS filter output y [ and input [ are expressed as follows 5) : [ DC DC DC i y V V 4R I ( V [ R I [ ]) (3) [ I [ (4) DC Fig. 6: HAR power control block. Fig. 7: DC/DC converter DC link efficiency control block. with i as the sapling counter. Fro (3) and (4), the updating paraeters are coputed in discrete tie as: [ y[ ( [ [ i ]) (5) Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved
Table : Circuit paraeters. Paraeter Value oad battery voltage [V] 6.34 Priary coil capacitance C [nf] 6.03 Secondary coil capacitance C [nf].5 DC link capacitor C DC [µf] 000 Priary coil inductance [µh] 47.8 Secondary coil inductance [µh] 08.5 Mutual inductance [µh] (best alignent) 39. Coil gap [] (best alignent) 00 Resonant frequency [khz] 00 Priary coil resistance R [Ω].83 Secondary coil resistance R [Ω].8 Fig. 8: Experiental setup. (a) Mutual inductance (b) DC link voltage (c) Power (d) DC to DC efficiency Fig. 9: Siulation result with the DC/DC converter control proposed in 5) for a speed of 0 k/h. (a) Mutual inductance (b) DC link voltage (c) Secondary side power (d) DC to DC efficiency Fig. 0: Siulation result with the proposed control: reference power is 5 W and the speed is 0 k/h. T [ i ] [ [ [ i ] [ (6) [ T [ i ] T [ i ] T [ i ] [ T [ [ [ T [ i ] (7) However, in this paper,the application of conditional updating to the RS filter is considered: when short ode happens, the values are held at their last value; when rectification ode happens, the filter coefficients updating restarts. These paraeters are eployed in the PID feedback and in the feedforward equilibriu point as described in 5). Finally, the transient of the current I ust end uch earlier than the vehicle speed one because the filter can be applied only in steady state. 4. SIMUATION AND EXPERIMENTA RESUT In order to verify the effectiveness of the proposed ethod, experients have been perfored. The ai is to prove that even if utual inductance changes, the control is stable and the desired power is effectively sent to the load. The circuit paraeters are reported in table. The DC/DC converter resistance its inductance R conv and conv are 0. Ω and 000 µh, respectively; as for the DC/DC converter filter capacitance C conv, its value is 000 µf. The HAR switching frequency is 500 Hz, while the DC/DC converter one is 0 khz. In order to verify the proposed syste, a ini odel is being used. The priary voltage source is 8 V. The experiental setup is shown in Fig. 8. In the experient, the Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved
(a) DC link voltage (b) Power (c) DC to DC efficiency Fig. : Experient result with the proposed control: reference power is 3 W and the speed is 0 k/h. (a) DC link voltage (b) Power (c) DC to DC efficiency Fig. : Experient result with the proposed control: reference power is 5 W and the speed is 0 k/h. receiver coil is oved by a linear actuator. The position of the coil is recorded by an encoder, which transits the data to the controller. The speed by which the receiver coil oves is 0 k/h. The experiental results on DC to DC efficiency are copared with the experiental results obtained by using the control and the converters presented in 5). The experients for a desired power of 3 W and 5 W have been perfored. In Fig. 9, the siulation results using the ethod and the converters proposed in 5) are shown. In these siulation, the full bridge diode rectifier is considered instead of the HAR in order to to provide a reference for the transitted power and efficiency. In fact, if using a diode rectifier, the power is not controlled. The efficiency in Fig. 9(d) is assued to be the axiu DC to DC efficiency the syste can reach in ideal conditions. In Fig. 0, the siulation result of the proposed control are presented. The influence of short ode is very noticeable. The desired power is sent to the load and high efficiency is achieved. However, in (c), the filtered average value is higher than the actual value. This is due to the filter delay and the HAR low switching frequency. Unfortunately, the utual inductance estiation results in the experient are not as good as the ones described in 5). Nevertheless, with those results the control can still achieve the expected desired power and high efficiency, as it can be seen in Fig. and Fig.. In Fig., the experiental results for a required power of 3 W are shown. In (a), the DC link voltage reference and its actual value are shown. The voltage reference is coputed fro the utual inductance value and the actual voltage can atch its reference. As it can be seen fro (b), the DC link power is around 3 W, the reference value. The values of the power are filtered with weighted oving average, therefore the graphic shows the averaged value. When the coupling coefficient becoes lower due to the coil oveent, there is a power surge typical of dynaic charging. The transitted power atches the reference when the utual inductance is at its axiu value. In Fig. (c), the DC to DC efficiency is shown. The blue line represents its true value, while the red line is the filtered value. In fact, during short ode, the efficiency falls drastically to extreely low values; actually, in the experiental setup there is an offset on the sensors, therefore the efficiency is not shown to becoe zero. On the other hand, when rectification ode happens, the current transient causes the values of the actual efficiency to be higher than the theoretical values: this is only a coputational error and has no relevance. The averaged efficiency value is very close to the axiu efficiency value, represented by the black line. The difference in efficiency is due to the losses generated during short ode. These losses are sall because of the low power experient. Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved
The experiental results with a desired power of 5 W are presented in Fig.. In (b) and (c), the graphic shows filtered averaged values, too. As expected, in both the experient efficiency easureents the filtered value of the efficiency is higher than the actual value, just like in 0(d). The reason is the sae as explained before. The results in both cases show that desired power and high efficiency are achieved, thereby proving the effectiveness of the control. 5. CONCUSION The authors proposed a siultaneous power control and efficiency control to be perfored only on the secondary side of a wireless power transfer syste, independently fro the priary side, by using a HAR and a DC/DC converter. In this paper, a dynaic wireless power transfer scenario is considered. By adopting an already proposed utual inductance estiation concept, experients have been carried out. Experiental results verified the effectiveness of the proposed ethod since the power reference is atched and high efficiency is achieved. Future works include high power experient and utual inductance estiation iproveent. REFERENCES () A. Kurs, A. Karalis, R. Moffatt, J.D. Jonnopoulos, P. Fisher, M. Soljacic : Wireless power transfer via strongly coupled agnetic resonances, Science Expressions on 7 June 007, Vol. 37, No. 5834, pp. 83-86 (007). () S. i, C.C. Mi: Wireless power transfer for electric vehicle applications, IEEE Journal of Eerging and Selected Topics in Power Electronics, pp. -4 (03). (3) H.. i, A. P. Hu, G. A. Covic, T. Chunsen: A new priary power regulation ethod for contactless power transfer, IEEE International Conference on Industrial Technology (ICIT), pp. -5 (009). (4) J. M. Miller, C. P. White, O. C. Onar, P. M. Ryan: Grid side regulation of wireless power charging of plug-in electric vehicles, IEEE Energy Conversion Congress and Exposition (ECCE), pp. 6-68 (0). (5) W. Chwei-Sen, O. H. Stielau, G. A. Covic: Design considerations for a contactless electric vehicle battery charger, IEEE Transactions on Industrial Electronics, vol. 5, pp. 308-34 (005). (6) M. Fu, C. Ma, X. Zhu: A cascaded boost-buck converter for high efficiency wireless power transfer systes, IEEE Transactions on Industrial Inforatics, Vol. 0, No. 3, pp. 97-980 (04). (7) H.H. Wu, A. Gilchrist, K. D. Sealy, D. Bronson: A high efficiency 5 kw inductive charger for EVs using dual side control, IEEE Transactions on Industrial Inforatics, vol. 8, pp. 585-595 (0). (8) T. Diekhans, R. W. De Doncker: A dual-side controlled inductive power transfer syste optiized for large coupling factor variations, IEEE Energy Conversion Conference and Exposition (ECCE) 04, pp. 65-659 (04). (9) G. ovison, M. Sato, T. Iura, Y. Hori: Secondary-side-only Control for Maxiu Efficiency and Desired Power in Wireless Power Transfer Syste, Proceedings of 4st Annual Conference of the IEEE Industrial Electronics Society (IECON) 05, pp. 4965-4970 (05). (0) K. Hata, T. Iura and Y. Hori: Maxiu Efficiency Control of Wireless Power Transfer Considering Dynaics of DC-DC Converter for Moving Electric Vehicles, The Applied Power Electronics Conference and Exposition, pp. 330-3306, (05). () T. Hiraatsu, X. Huang, M. Kato, T. Iura and Y. Hori: Wireless Charging Power Control for HESS Through Receiver Side Voltage Control, The Applied Power Electronics Conference and Exposition (APEC), pp. 64-69 (05). () M. Kato, T. Iura, Y. Hori: New characteristics analysis considering transission distance and load variation in wireless power transfer via agnetic resosnat coupling, IEEE Proceedings INTEEC (0). (3) M. Kato, T. Iura, Y. Hori: Study on axiizing efficiency by secondary side using DC-DC converter in wireless power transfer via agnetic resosnat coupling, EVS7, pp. -5 (03). (4) K. Hata, T. Iura, Y. Hori: Dynaic Wireless Power Transfer Syste for Electric Vehicle to Siplify Ground Facilities - Power Control Based on Vehicle-side Inforation, The 8th International Electric Vehicle Syposiu and Exhibition (EVS), pp. (05) (5) D. Kobayashi, T. Iura, Y. Hori: Real-tie coupling coefficient estiation and axiu efficiency control on dynaic wireless power transfer for electric vehicles, 05 IEEE PES Workshop on Eerging Technologies: Wireless Power, pp. -6 (05). Copyright 06 Society of Autootive Engineers of Japan, Inc. All rights reserved