Single-Phase Wireless Electric Vehicle Charger Using EF 2 Inverter

,


Introduction
Wireless charging for electric vehicle (EV) is one of the spellbinding research topics for power electronics engineers in recent times. Te lack of public charging infrastructure in many countries leads the government to mandate manufacturers to include the charging circuit with the vehicle. Tis charging circuit is known as an on-board charger that is capable of charging the battery pack by tapping power from the utility grid without any additional infrastructure [1][2][3]. Te on-board charger generally uses a wired connection, which is often messy, and this adds to the total weight of the vehicle, which is undesirable [4,5]. Te wireless EV charging system reduces the weight burden on the vehicle because it only contains a receiving coil and a rectifer unit as most of the power electronics interfaces are ousted from the vehicle to the transmitting side. A pictorial representation of a typical single-phase domestic/commercial wireless EV charging system is represented in Figure 1. Tus, wireless EV charging ofers contactless power transfer and lighter onboard charging system [6][7][8]. An overview of wireless power transfer technology for EV charging is presented in [9].
Te basic concept of wireless power transfer (WPT) lies in transmitting a high-frequency sinusoidal AC wave from the transmitting side to the receiving side without any wired connection [10,11]. Instead of using a conventional step-up cycloconverter for producing high-frequency AC (HFAC) from the grid supply, an inverter is preferred to generate HFAC by operating this inverter at high switching frequency. So, a high-frequency inverter (HF inverter) is considered as the heart of the WPT system, which requires a DC source at its input to produce HFAC. Conventional WPT system uses an H-bridge inverter with fltering elements, operated at high frequency to generate HFAC [12][13][14][15][16]. Higher number of switches and hard switching of MOSFETs in case of H-bridge inverter are the two major concerns, which led to the introduction of resonant inverters for WPT applications.
Many resonant inverters using resonance network with H-bridge inverter are discussed in [17]. Moreover, resonant inverters with minimum switches such as class D and class E inverters are reported in the literature [18][19][20], which are mostly preferred due to their soft switching property. Class E inverters are widely used in WPT application as these use only a single switch to produce HFAC which is also capable to achieve zero voltage switching (ZVS) for reducing the switching losses, which is well documented in [20]. Major concern in case of a class E inverter is the higher voltage stress during the of period across the switch caused by the resonating elements that are responsible for producing sinusoidal current at the output. Tis extra voltage stress on the switch reduces the power processing capability of the inverter. To overcome this, class EF 2 inverter is reported in [21], where an additional series L-C series network is used in parallel with the switch to reduce the voltage stress. Te class EF 2 inverter is capable to achieve zero voltage switching (ZVS) as well as zero derivative switching (ZDS). But this topology is only applicable for a constant load resistance at the output. When the load resistance is changed to a value other than the designed value, the soft switching operation (e.g., ZVS and ZDS) is lost.
During the charging process of a battery, as the internal resistances of the battery continuously increase with the increase in the state of charge (SoC), the inverter used in the wireless charger should be capable to handle variable load resistance. To accommodate variable load resistance, a class EF 2 inverter with load independent criteria is discussed in [22]. Te reported class EF 2 inverter behaves as a constant AC current source for variable load resistance at the output. However, to verify the load independent criteria, researchers have used a fxed load resistance of 50 Ω at the secondary side of the WPT and the distance between the transmitting coil and receiving coil was varied to demonstrate the variable load resistance refected across the output of EF 2 inverter as an outcome of change in distance.
To implement the CC-CV charging technique, the charger should behave as a source of constant current during CC mode and a source of constant voltage during CV mode. Te earlier reported class EF 2 inverter can be implemented for CC mode, but it fails to deliver power to the battery during CV mode. Additionally, the EF 2 inverter uses a DC source as its input to provide AC power at high frequency. Drawing power from the AC grid is more convenient as the grid is more reliable and easily available everywhere as compared to any DC power source.
In order to address all these issues, this paper proposes a wireless EV charger that taps power from the single-phase AC supply and charges the EV battery pack wirelessly satisfying both constant current and constant voltage operation. Te transmitting side power converters with the transmitting coil are placed beneath the ground level, and the receiving coil with rectifer unit is ftted with the vehicle. Te idea is to charge the EV battery pack mostly during night time, when the vehicle is parked in home garage.
Overall contribution of this paper is listed as follows: (i) Te proposed charger is designed to be implemented at any charging station specifcally for home garage as it draws power from single-phase grid. (ii) Te EF 2 inverter used in the proposed charger is capable to deliver power in both CC and CV modes throughout the designed range of load variation. (iii) Switching scheme with fxed duty ratio is implemented for EF 2 inverter to avoid fast sensing requirement of variable AC voltage and current that also needs high-speed controller response as the inverter is operated at very high frequency.  (iv) Te proposed charger ensures CC-CV charging of the battery as well as power factor correction (PFC) operation at the input side with only one controlled signal.
1.1. Paper Organization. Section 2 of this paper describes the system confguration with complete circuit diagram. Design and operation of the complete charger are described in Section 3. Necessary mathematical modelling and analysis are also included in this section. Section 4 describes the control strategy of the charger with proper schematics. Te proposed charger is verifed experimentally as well as using simulation platforms, which are discussed in Section 5. Te paper is fnally concluded in Section 6.

Proposed Wireless EV Charger
Te proposed charger taps power from the single-phase utility supply and wirelessly charges the battery set of EV. Te circuit diagram of the proposed topology is shown in Figure 2. Te proposed circuit is divided into four major subsections: (i) AC-DC boost converter, (ii) DC-HFAC resonant inverter, (iii) wireless power transfer scheme, and (iv) rectifer unit. As discussed earlier, a HF inverter is the heart of WPT scheme. A load independent EF 2 inverter is used to produce high-frequency sinusoidal AC. Te HF inverter produces constant AC current when excited with constant DC voltage and produces constant AC voltage when the input of the inverter is a constant DC current source. To accommodate either a constant DC voltage source or a constant DC current source at the input of the inverter, an AC-DC front-end converter is selected as stage-1. Te front-end converter is responsible to deliver power to the EF 2 inverter at constant voltage or constant current while maintaining the input power factor to near unity. Te WPT scheme uses the magnetic resonance coupling (MRC) technique to transfer power from the transmitting side to receiving side wirelessly. Tis consists of two coils (namely, primary and secondary) with their corresponding matching network. Te primary coil draws power from the inverter and transmits to the secondary coil. Te HFAC received at the secondary coil is converted to DC by a full-bridge diode rectifer that charges the battery.

Stage 1: AC-DC Boost Converter.
A bridge-less PFC structure with two switches on the lower side, two diodes on the upper side, and an inductor at the input is used as an AC-DC boost converter. Tis stage is responsible to deliver DC power to the HF inverter and to maintain near unity power factor at the grid side. Te inductor at the input side acts as a boost inductor for this stage. Te operation of this stage is described with four modes of operations as shown in Figure 3.
During positive half cycle of input AC supply, diode D A , switch S A , and body diode of switch S B (D SB ) are forward biased. When the switch S A is turned ON, current gets a shortest path through S A and fows through the input inductor, node A, switch S A , body diode D SB , and node B as shown in Figure 3(a). Te boost inductor is charged during this mode. When the switch S A is turned OFF, current will fow through inductor, node A, diode D A , load, body diode D SB , and node B as shown in Figure 3(b) ensuring the smooth discharge of energy stored in the inductor L in during previous mode. Tough, due to the positive nature of supply voltage, diode D 1 should be forward biased, turning ON of switch S A makes this reverse biased during mode 1, but it is forward biased during mode 2.
Similarly, for negative half cycle, while the switch S B is ON, the inductor is charged with the current, fowing through node B, switch S B , body diode D SA , and node A as shown in Figure 3(c). In mode 4, the stored energy in the inductor will be discharged to the load as shown in Figure 3(d).
In both the positive and negative half cycles of input supply, the charging of input inductor during the switch ON period and discharging of inductor during the switch OFF period confrm the boost operation. 2 Inverter. Te proposed schematic uses an EF 2 inverter to produce high-frequency AC voltage or current according to the requirement of CC-CV charging algorithm. Figure 4 shows the circuit representation of the class EF 2 inverter that supports both CC and CV modes of operation by delivering constant AC current and constant AC voltage during CC mode and CV mode, respectively, which is analyzed in this subsection. Te equivalent circuit diagrams during ON state and OFF state of the switch S inv are presented in Figures 5(a) and 5(b), respectively.

Stage 2: High-Frequency EF
While the inverter is excited with a constant DC voltage source, assuming a sinusoidal output current of the inverter, (1) During ON state of the switch S inv ( Figure 5(a)) (for 0 ≤ ωt < 2πD), voltage across the switching MOSFET can be written as Current through the capacitor C 1 is i C1 � 0. As the voltage across the switch is zero, it is rewritten as Diferentiating the above equation and solving the second-order diferential equation, solution for i L2 is obtained. Again, normalizing i L2 with respect to I IN will be in the form of International Transactions on Electrical Energy Systems AC/DC Boost converter EF 2 Inverter as HF Inverter WPT Scheme To Inverter

International Transactions on Electrical Energy Systems
During OFF state of the switch ( Figure 5 As solved for the ON state, V DS � 0. Solving (7) for i L2 and normalizing the expression of i L2 with respect to I IN , it can be written as where where A 1 , B 1 , A 2 , and B 2 are the arbitrary constants considered for solving the diferential equations during both ON and OFF modes. Tese arbitrary constants are determined based on boundary conditions of inductor currents and capacitor voltages. Using KCL at drain node, Normalizing i C1 (ωt) with respect to I IN , Using boundary conditions, During switch OFF period, voltage across the switch is the same as voltage across the capacitorC 1 : where V DS can be written as follows using equations (13) and (19) and the ZVS conditions.
Using all the above conditions and considering maximum power output capability (q 1 � 1.66 and k � 1.2706) as discussed in [22], the expressions for the circuit parameters at 200 kHz are found and given in Table 1. Te same expressions are also valid for constant voltage operation.

Stage 3: Wireless Power Transfer Scheme.
Te transmitting and receiving coils are designed according to the required current rating. Series-series topology is considered as the compensating network for MRC. A combination of capacitor in series with the coil at both transmitting and receiving sides forms the WPT section as shown in Figure 2. Values of these capacitors (C pmat and C smat ) are designed to resonate with the corresponding coil inductances at the operating frequency using f s � (1/2π ��� LC √ ). Same number of turns are used in both transmitting and receiving side coils, which facilitate to use same values of capacitances in both sides, i.e., C pmat � C smat .

Stage 4: Rectifer Unit.
Te rectifer unit uses a simple full-bridge diode rectifer to convert the HFAC to DC as shown in Figure 2. Te selection of diodes is of major concern, as the high-frequency operation requires less

Control Scheme for the Proposed Wireless EV Charger
Te most important section of the proposed transmitting side topology is the EF 2 inverter. Closed-loop operation of such high-frequency inverter is difcult, as this requires the sensing of HF inverter output parameters. Te sensing of high-frequency parameter requires even higher frequency of controller operation. To avoid these issues, the EF 2 inverter is operated with constant duty ratio and constant switching frequency. Tis HF inverter requires a regulated voltage or current source at its input for CC-CV operation. To achieve this, the front-end converter (stage-1) needs to be operated accordingly. For the front-end converter to supply regulated voltage/current at its output terminal, its output voltage and current need to be sensed. But as the EF 2 inverter is operated at constant duty cycle, the required voltage and current at the battery terminal can be multiplied by a gain factor to fnd the new reference values of voltage or current at the output of stage-1. Terefore, instead of sensing voltage and current at the output of stage-1, these can be sensed at the battery terminal and corresponding gain factor can be multiplied. Tese signals are then compared with their corresponding reference signals, and using proper controllers, the PWM signals are generated for the front-end converter. Te PWM pulses are used to switch the MOSFETs S A and S B in order to achieve regulated voltage or current at the output of the front-end converter as well as the PFC operation at the input grid side.

CC Mode.
In CC mode, a constant amount of current is dumped into the battery. When the HF inverter is operated with a suitable fxed duty ratio and the input excitation is a constant DC voltage source, it will deliver sinusoidal AC current with constant magnitude throughout the range of load variation, for which it is designed. Te battery terminal current is sensed and compared with its reference value and corresponding PWM signal is generated to maintain constant voltage at the input of HF inverter. Te corresponding control scheme is shown in Figure 6(a).

CV Mode.
During CV mode, the battery is charged with a constant terminal voltage, where the charging current decreases gradually with the increase in state of charge (SoC) of the battery. To support this, the charger should deliver power while maintaining the output voltage constant. For maintaining constant voltage at the battery terminal, the HF inverter should be capable to provide constant voltage at its output. As discussed earlier, the front-end converter is operated to act as a source of constant DC current for the inverter.
In this mode, the battery terminal voltage (V bat ) is compared with its reference value (V ref ) and fed to the PI controller. In a similar way to CC mode, the PI controller generates the required PWM signal in order to make the front-end converter as a source of constant current to feed the HF inverter as shown in Figure 6(b).
During both the modes, the HF inverter operates at a constant duty ratio. CC-CV logic is implemented to decide the mode of operation according to the SoC of the battery. During low SoC, the battery is charged with constant current, and at high SoC, the battery is charged with a constant voltage. Te transition from CC mode to CV mode is decided based on the battery terminal voltage. Te charger remains in CC mode until the battery voltage reaches the full voltage as mentioned in the manufacturer's data sheet. Once the battery voltage reaches the full voltage level, then the controller forces the charger to enter into CV mode of operation. Accordingly, the front-end converter is operated to behave as a regulated source of current or voltage. A single PI is sufcient to deliver power to the inverter at regulated voltage or current during both CC and CV modes.

DC Voltage or Current Source
International Transactions on Electrical Energy Systems It is necessary to maintain near unity power factor at the input side as the proposed wireless charger taps power from the single-phase supply for charging the EV battery pack. To implement PFC operation at the grid side, input AC voltage and current are sensed and fed to the PFC controller. Te combined efort of PI and PFC controller is then fed to the PWM block. Te fnal PWM signals are fed to the gate driver circuits of two lower switches of the front-end converter. Te complete control scheme for the proposed wireless charger is implemented using a TMS320F28335 DSP experimenters' kit from Texas Instruments and is shown in Figure 7. Te reference values in the CC-CV logic are decided based on the required charging current and charging voltage of the battery.

Experimental Verification
Te primary and secondary coils of the WPT system are simulated for demonstrating the strength of magnetic feld intensity (B) in and around the coils and are presented in Figure 8. Te color coding of its strength in Tesla is also shown in Figure 8. Considering the leakage and mutual inductances from Ansys, the complete topology is simulated using PSIM to verify the constant current and constant voltage property of the proposed charger. Te aim is to test the charger with dynamic load condition as the battery is a dynamic load during its charging operation. Terefore, the charger is tested with resistive load and the load resistance across the DC terminal is changed instantaneously to check the robustness of the charger. Te dynamics are captured for both constant current and constant voltage mode using simulation platform. Te corresponding results for CC and CV modes are given in Figures 9(a) and 9(b), respectively.
A scaled-down laboratory prototype of the proposed wireless charger is developed and verifed to charge a 24 V, 30 Ah battery set as shown in Figure 10. Te proposed wireless EV charger is also verifed in the laboratory with a resistive load to wirelessly transfer 191 W of power with a distance of 120 mm between the primary and secondary coils. Te required air gap between the transmitting coil and the receiving coil depends on the minimum ground clearance required by a vehicle as the transmitting (primary) coil is usually kept at the ground level and the receiving (secondary) coil is ftted just below the vehicle. Ground clearance for a light motor vehicle (LMV) is normally in the range of 120 mm to 150 mm. In order to test the efectiveness of wireless power transfer (WPT) system, the proposed charger is tested with a distance of 120 mm between the primary and secondary coils.
A 4.5 mH inductor is used at the input AC side to boost the input voltage as well as to smoothen the ripple of input AC current. A Semikron IGBT module is used to realize the AC-DC boost stage. Circuit parameters of EF 2 inverter are derived for a switching frequency of 200 kHz, and the corresponding values are given in Table 2. Te WPT coils are designed according to the Ansys design and connected at the output terminals of EF 2 inverter as shown in Figure 10. Figure 11 shows the clear view of the PCB implementing EF 2 inverter using a GaN MOSFET. Si or SiC MOSFET can also be used to implement a switching frequency of 200 kHz, but reason behind considering GaN MOSFET in this work is to increase the switching frequency in the order of MHz. Te WPT section with transmitting and receiving side coil is shown in Figure 12.
A four-layer FR4 PCB board is designed to implement the EF 2 inverter as shown in Figure 11. Special care is taken while fabricating the SMD components especially GaN  International Transactions on Electrical Energy Systems 7 MOSFET. Te GaN device is assembled on the board by properly following the refow soldering process. An enhancement-mode high-electron mobility transistor (E-HEMT) GS66508B is used as S inv for the HFAC inverter, which is a bottom-side cooled device. Te recommended gate voltage range is 0 V to +6 V for optimal performance. Si-8271-GB-IS with proper biasing is used as a gate driver circuit for operating the GaN MOSFET. Te bulky ferrite core-based inductors are replaced by lightweight air-core inductors to avoid saturation at high-frequency operation      International Transactions on Electrical Energy Systems except for L 1 as this is only responsible to reduce the current ripple at DC side of the inverter. A 684 uH SMD inductor is used for this purpose. Tese air-core inductors are designed using enamel painted single-strand copper wire of diameter of 0.85 mm. Metalized polypropylene capacitors are used to support high-frequency operation. Te rating of inductors and capacitors used is shown in Table 2.

Operation of EF 2 Inverter.
Te EF 2 inverter is operated at a switching frequency of 200 kHz, and the results are shown in Figure 13. V GS shows the gate pulse to the GaN, and V DS shows the drain to source voltage of MOSFET, while I OUT shows the output current supplied by the inverter. Te output current is sinusoidal, and the frequency is same as the switching frequency. Te voltage across the switch V DS proves the ZVS operation of the inverter.

Proposed Charger with Resistive Load.
Te charger is verifed with resistive load to ensure smooth operation for all probable conditions before charging the battery. Steady-state operation showing output voltage and current (V 0 and I 0 ) for a 20 Ω load while maintaining ZVS is shown in Figure 14.
Te transmitting side HFAC voltage and current along with voltage and current at the output terminal are shown in Figure 15. Use of proper matching networks at both the transmitting and receiving sides results in perfect sinusoidal voltage and current at the transmitter side. Te charger is tested at nearly 200 W to transfer power wirelessly to a 20 Ω load. Te corresponding result is shown in Figure 16, where the ZVS condition is achieved with 61 V output voltage and 3.12 A of output current. Figures 13, 14, and 16 show the peak and valley in V DS waveform. Tis is due to the presence of the additional parallel network (L 2 in series with C 2 ) across the switch. Tis series combination across the switch    International Transactions on Electrical Energy Systems 9 reduces the peak of V DS as compared to class E inverter. Te complete operation of class EF 2 inverter with relevant mathematical analysis is well documented in the literature [21,22].  Figure 17 shows the CC mode, and Figure 18 shows the CV mode, where the charging current is decreasing gradually maintaining a constant voltage of 14.5 V.

Charging a 24 V Lead-Acid Battery.
Finally, the proposed wireless charger is validated to charge a 24 V, 30 Ah battery set to obtain the CC-CV profle. Waveforms at diferent nodes are captured and discussed in this subsection. Figure 19 shows that the battery is charged in CC mode with a current of 1 A, and the result is captured when the battery terminal voltage reached 26.3 V. Te transmitting side voltage and current during this period are also found to be sinusoidal, and the frequency is same as the switching frequency. Te proposed charger draws power from the grid utility while maintaining near unity power factor at the grid side as shown in Figure 20(a). Te zoomed-in result of Figure 20(a) is shown in Figure 20(b) to verify the sinusoidal nature of current delivered by the HF inverter as well as its smooth ZVS operation. Figures 21(a) and 21(b) show the PFC operation at the input side of the proposed charger and the sinusoidal voltage and current at the transmitting side of the WPT system. Figure 22 shows both the transmitting and receiving side voltage and current during charging of the 24 V, 30 Ah battery set. Tis confrms the perfect wireless power transfer and correct matching network parameters associated with this. Te battery terminal voltage and current with primary side current of EF 2 inverter satisfying ZVS operation are shown in Figure 23.
Te grid side input to the charger and its battery terminal output while charging the 24 V, 30 Ah battery are shown in Figure 24. Tis shows that a constant amount of current is dumped into the battery during low SoC while maintaining near unity power factor at the input side. Te CC-CV profle while charging the 24 V battery set is drawn and is shown in Figure 25, which confrms the implementation of CC-CV charging technique. Te efciency is measured during charging of both 12 V and 24 V batteries at diferent terminal voltages. Te efciency curves are prepared separately for both the batteries and are presented in Figure 26.

Conclusion
A wireless battery charger for electric vehicle (EV) applications is proposed in this paper. Te proposed topology is capable to charge the EV battery pack wirelessly by taking power from a single-phase wall outlet available in domestic premises. Te control scheme of the charger ensures power factor correction (PFC) operation at the input grid side along with CC-CV charging of the battery. During CC mode, the AC-DC stage supplies DC power at constant voltage to the EF 2 inverter, which produces constant AC current at high frequency. At the receiving side, this constant HFAC current is rectifed to charge the battery with constant DC current. Similarly, during CV mode, the EF 2 inverter is fed by the AC-DC stage with constant current for producing constant HFAC voltage at the output of inverter. Te constant AC voltage at high frequency is then rectifed by the rectifer unit to charge the battery with constant DC voltage. Te WPT coils are frst simulated using Ansys Maxwell, and the values of leakage inductances and mutual inductances are used to simulate the complete charger in PSIM simulation platform. A scaled-down experimental prototype is developed in the laboratory to verify the proposed charger and tested to transfer 200 W power wirelessly over a distance of 12 cm. Finally, a 24 V, 30 Ah battery set is charged wirelessly using the proposed charger.

Data Availability
Te data used to support the fndings of this study are included within the article.

Conflicts of Interest
Te authors declare that they have no conficts of interest.  International Transactions on Electrical Energy Systems 13