Abstract
The main objective of this work is to develop an efficient reactive power compensated control technique for a fast-charging scheme for electric vehicle(s) (i.e., level-3 charging). The developed charging technique has the four-quadrant power flow operation with the simultaneous assurance of the compensated reactive power control. The developed charging infrastructure scheme involves a solar panel and 3-phase grid to charge the E-mobility. A DC-DC boost converter is used to achieve maximum power tracking (MPPT) of a solar PV array, and a 3-phase grid-tied bidirectional voltage source converter (VSC) is utilized to provide the bulk of power to charge the EV. The 3-phase VSC has multiple functionalities including grid side power quality (PQ) improvement with reactive power compensation, seamless flow of power from the grid to EV, and solar to grid or battery to grid (V2G) operation. This scheme also facilitates tariffs earned by discharging solar energy to the grid with additional benefits of reactive power compensation. An arrangement is also made to tackle grid failure conditions during battery discharging mode by connecting a load for ancillary purposes. The effectiveness of this charging scheme is first examined in MATLAB/Simulink environment and then validated on developed hardware.
1. Introduction
The growing concerns over environmental protection and energy conservation over the past few decades have reinforced stringent regulations on fuel consumption. One of the main sectors that witnessed significant change due to tougher environmental regulations is the transportation sector. Over the past couple of decades, advancements in EV technology have grown significantly [1]. As reported in [2], the deployment of electric vehicles has been growing rapidly over the past few years with the global stock of electric passenger cars reaching over 5 million in 2018 with a huge increase of 63% from the previous years. Extensive research throughout the world is going on to shift our reliance on renewable energy sources, and new technologies are coming up in the field of battery charging [3, 4]. Solar and wind-energy-based systems are quite popular these days. However, these initiatives require more capable power transmission lines with reduced transmission losses. Regarding this, a few easy solutions have been discussed in [5, 6], which emphasize installing a PV array in the vicinity of the charging station. This is considered a local solution with power generation and consumption locally [7]. This does not require up-gradation of power transmission line and saves extra cost incurred in it. This local solution for the charging scheme is independent of the tariff as the power is locally generated for this purpose [8]. Therefore, various solar parks have been developed near the vehicle parking area as reported in the literature [9]. So, the utilization of solar power in charging infrastructure eases the burden on the utility and also reduces the charging cost/miles.
The charging stations for E-mobility with solar power have been extensively researched [3, 4], which provides an alternate source for charging stations. However, its intermittency has always been a matter of concern. The grid integration eliminates the complete reliance of the system on solar PV arrays. The storage capacity of the battery can be manifested during the parking hours to feed the utility for additional services for many purposes such as V2G and V2H [10]. In addition, solar power can be fed back to the grid if EV charging is not required, as discussed in [11].
The reactive power balance comprises both lagging (inductive) and leading (capacitive) power requirements of the grid. The reactive power balance in V2G mode does not draw any real power from the battery and thereby is independent of battery SOC. Many attempts of research have been made to suggest the merits of reactive power generation and distribution near the load as it reduces the overall losses in the transmission line for long distances [12]. Nowadays, due to the nonlinear nature of most of the loads, distorted phase currents are drawn from the utility. The consequences of these undesirable grid indices on the system performance have been reported in the literature [13] with also the probable solutions to overcome it.
Several techniques are reported in the literature based on power quality enhancement of the grid. Among them, synchronous reference frame theory (SRF) [14] and instantaneous reactive power theory (IRPT) [15] are extensively used algorithms. The IRPT technique is based on the transformation of three-phase quantities into two-phase quantities (i.e., voltage and currents) to find active and reactive power. However, the transient response is a bit sluggish due to the utilization of the proportional PI regulator and low-pass filter to control DC link voltage. In addition, it becomes worse when it deals with unbalanced and distorted voltages. Similarly, the SRFT algorithm also suffers from these problems during unbalancing and distortion of voltages. Enhanced PLL (EPLL) [16] overcomes these difficulties and displays excellent tracking capability under these abnormal grid conditions but is computationally more intricate and hence more expensive. The generalized integrator-based algorithms like the second-order generalized integrator (SOGI) [17] based algorithm offer a fast convergence rate and good tracking efficiency in a narrow band of frequency, and the performance deteriorates when the frequency changes. Therefore, it is assisted by a feedback frequency loop to make it frequency adaptive but at the cost of increased computational burden on the processor. The SOGI can be replaced by a notch filter [18]. However, the problem associated with the selection of notch frequency for variable frequency application is very difficult. This problem has been resolved by an adaptive notch filter (ANF) [19], which updates the center frequency by using the feedback signal with the variable frequency. The SOGI FLL offers superior performance to other algorithms but still, the complex computation is its drawback [20].
Acknowledging the limitations of previous systems, the proposed system provides an efficient reactive power compensated control method for fast charging of EVs. The schematic of the developed charging scheme is shown in Figure 1. This developed charging system provides a unique scheme for the four-quadrant operation of the charger with an improved control algorithm. Solar power is used to feed power into the grid; however, the grid is the main source of charging the EV. Therefore, the power coming from the grid offers a fast-charging infrastructure to the station. The developed system has also the capability to work in V2G mode. In this case, the fully charged battery can feed power to the grid when propulsion mode is not needed. This mode is very effective in peak load hour operations. All four quadrant of P-Q power control is achieved by the proposed control technique. The system acts in a two-stage manner when it operates in G2V or V2G modes. The battery charging is accomplished until it achieves the extreme limit; that is, 80% SOC specifically decided for the present system. After that, the charger starts feeding power to the grid. In case of grid failure or absence condition, the fully charged battery is detached from the charging station and starts supporting ancillary services such as water pumping and lighting.
A grid-tied VSC performs as a controlled rectifier in the G2V mode, and the converter attached to the battery performs the buck operation to offer energy storage in a battery whereas, in the V2G operation, the same converter on the battery side performs the boost function to match the DC link voltage. The bidirectional power transfer is achieved by the DC link capacitor during different modes of EV charging/discharging modes. Moreover, the bidirectional converter inductor (Lb) suppresses the ripple current into the battery storage. The interfacing inductors (Ls) at the grid side are meant for the mitigation of the harmonics in the grid current caused due to AC-DC conversion.
2. Design Configuration
The developed system consists of a solar panel of 4.3 kW and an EV of 4.2 kW power rating. A deviation-free P&O (perturb and observe) MPPT technique is executed using a boost converter for PV panel power optimization. The parameters of the boost and bidirectional DC-DC converter are so selected that it always works in CCM (continuous conduction mode) irrespective of the working mode and battery and PV power. The design configurations of various elements present in the developed system are discussed here.
2.1. Estimation of Solar Parameters
The selection of solar panel indices primarily depends on the highest power necessity of the charging station. Here, to charge an EV with a battery bank of voltage and current ratings 240 V and 10 Ah, respectively, a 4.3 kW solar panel is designed with 19 and 13 modules in series and parallel, respectively. The complete electrical specification of the solar panel implemented in the developed system is given in Appendix.
2.2. Inductor Calculation for MPPT Converter
The duty ratio (D1) for the MPPT converter is calculated as
Therefore, the calculation of the inductor is done by the governing expression
2.3. Design of Low-Pass Filter for Grid
A series arrangement of RC is attached at the PCC (point of common coupling) to mitigate the switching voltage harmonics. The RC arrangement acts as an LPF (low-pass filter), which attenuates the high-frequency switching harmonics from the main supply and passes only the desired frequency signal to the system [21]. The parameters of the ripple filter are thus designed to achieve the condition, RfCf << Tsw2, where Rf, Cf, and Tsw2 are resistance, capacitance, and switching time, respectively, of the designed LPF.
Assuming the value of RfCf = Tsw2/4, where Tsw2 = 10−4 s and Rf = 5 Ω, the size of the required Cf is estimated as
2.4. Battery Converter Design
The duty ratio (Dbat) for the DC-DC battery converter in boosting operation is expected as
Therefore, the calculation of the inductor is done by the governing expressionwhere fs (switching frequency) = 10 kHz and ΔIbat (permitted ripple current) = 0.25 A. Similarly, for buck mode, the duty ratio would be 0.6.
3. DPC for Grid Voltage Source Inverter
The power transfer in both directions is executed as per the control technique shown in Figure 2. The active and reactive power signals are regulated as per the command signal given by the user as needed. The four-quadrant operation is achieved viz. charging/discharging modes with inductive and capacitive and reactive power compensation. The instant value of active and reactive powers is estimated from the grid parameters.
The expression for the instantaneous magnitude of complex power s(t) is expressed as
The magnitude of grid current changes by the variation in the VSC voltage, which changes the active and reactive power drawn from the AC mains, and it has to be controlled. For a zero voltage vector, the change observed in active and reactive power is given as
Neglecting resistive power loss, the equation becomes
For an instant VSC voltage, the variation in power, P, and Q can be defined by
In this control scheme, the error signals are made null to attain the optimal voltage. The resultant signals are given as
Thus, equations for and are used in the aforementioned expression to achieve the subsequent equations as
The resultant error indices are managed by the hysteresis-based controller, resulting in the generation of the switching pulses for VSC named G1–G6, as shown in Table 1..
4. Charging Control: E-Mobility Battery
Figure 3 shows the fast-charging control mechanism of the EV. The battery must discharge through the grid if it is fully charged. However, in this case, if there occurs a grid failure condition, then it must be disconnected and discharged through auxiliary applications. This mode is achieved via. the logic represented by the red line in Figure 3, where the battery current (Ibat) is regulated by a current regulator.
On the other side, if the grid is available, the control follows the blue line loop, which follows the reference current input. In the first case, the duty ratio is generated by the current controller to be transformed into the high-frequency switching pulses for the DC-DC converter attached to the battery.where Kp and Ki are termed as proportional and integral gains of PI controller, respectively. In another situation, the desired current for charging purpose relies on the percentage of SOC of the storage system and is generated according to it.
5. Simulated Results
The developed charging scheme is simulated and examined under a wide range of operating conditions. The battery terminal voltage is 240 V, and the value of the DC link voltage is set to 400 V. The complete performance is analyzed under all possible situations.
5.1. Harmonic Distortion Analysis of Utility Currents
The distortion in grid side current in two working modes viz. G2V and V2G, respectively, is analyzed here. The irradiance level is set at 300 W/m2. The nature of grid current is shown in Figures 4(a)-4(b). In each situation, the current THD is 3.5%, and 3.8%, respectively, which successfully follows the IEEE-519 std.
(a)
(b)
5.2. Dynamic Behavior under Varying Insolation Level
Figures 5(a) and 5(b) show the sudden drop in the solar insolation from 1000 W/m2 to 300 W/m2 at 1.0 s. The system operates in constant charging mode, which draws extra power from the grid, increasing the in-phase grid current (). As soon as the battery charge attains its limiting value at 1.5 s, the battery starts discharging through the grid with the PV array at 300 W/m2. The phase displacement of grid current and grid voltage () is π rad. The positive sign of active grid power (P) indicates the G2V operation while the magnitude of “P” below the zero line illustrates the power flow from solar to the grid or V2G operation. Under each situation, Vdc is regulated at 400 V.
(a)
(b)
5.3. Battery Discharging during Grid Failure Condition
Figures 6(a) and 6(c) show the condition when the battery is discharging and suddenly encounters the grid outage. In this case, a propulsion mode has started, and the battery starts discharging. Figure 6(a) shows this condition and the solar-battery behavior as the storage battery gets discharged at 0.4 s with the rated irradiance level.
(a)
(b)
(c)
Both start feeding power to the grid at this moment, as demonstrated in Figure 6(b). At 1.0 s, the utility is cut out from the system. Therefore, as in Figure 6(c), the battery initiates feeding a propulsion motor. Moreover, Figure 7 clearly depicts that the proposed DPC approach has effective performance in terms of “% overshoot” and has a minimum settling time as compared to the other control schemes reported in the literature.
6. Experimental Validation
The experimental validation is carried out with a solar PV array (ETS600 × 17DPVF TerraSAS) of 4.3 kW, 334 V, and 12.4 A. The specification of the battery is 240 V and 10 Ah. A DSP microcontroller dSPACE (1006) is used to execute the control algorithm with the current and voltage signals acquired from voltage and current transducers (LEM LV-25P and LEM LA-55P), respectively. Figure 8 shows the photograph of the hardware setup.
6.1. Tracking Performance of PV Array
The tracking efficiency of the PV array is almost 100% using a well-known P&O MPPT algorithm at the maximum insolation level, as can be seen from Figure 9. The specification of the PV array for experimental work is 4.3 kW.
6.2. Steady-State Response of System
The steady-state response of the system in grid-connected mode is exhibited in Figure 10. The power is generated by a solar PV array, which partly feeds it to the grid and partly charges the battery. Figure 10(a) shows three-phase balanced grid voltages and currents with their values given in Figure 10(b).
(a)
(b)
6.2.1. Harmonics Analysis of Grid Fed System
Figures 11(a) and 11(b) illustrate the behavior of the developed charging system while the grid is being fed by a solar PV array. These results show that the THD of grid currents and grid voltages (and ) of all three phases is less than 5%, which justifies the suitability of the VSC control for not only ensuring MPPT but also compensating the nonlinearity of EV current.
(a)
(b)
6.2.2. Power Flow Management of PV Fed System
Figure 12 shows the solar PV power fed to the grid and the battery. Out of 4.2 kW of power generated by the PV array, 2.6 kW of power is absorbed by the grid, and the rest of the power is used to charge the battery. The negative sign in Figure 12(b) depicts that the battery is being charged.
(a)
(b)
(c)
6.3. Dynamic Performance of System
The dynamic performance of the complete system in terms of insolation change to be followed correspondingly by a change in power flow to the battery and grid to maintain the constancy of power generation and consumption is discussed in this section.
6.3.1. Transient Change in PV Array Power with the Same Rate of Battery Charging
Figure 13(a) shows the dynamic behavior of insolation change. As it can be seen, the rated power is delivered by the PV array (), which is distributed for charging the battery (Pbat) and also feeding the grid (). As the insolation drops down to a lower value, the grid power () is seen to be positive, which demonstrates that the utility provides power for EV charging. Therefore, the charging of the battery is achieved by drawing power from the PV array and partial power from the grid. Figure 13(b) shows a similar trend in terms of , Ibat, Ipv, and ig. The zoomed-in version shows the PV current (Ipv) increases in correspondence to insolation increment. The charging rate of the battery is kept constant by maintaining the charging current (Ibat). Therefore, the increased PV power is fed to the grid. As a result, grid current (ig) increases with constant DC link voltage ().
(a)
(b)
6.3.2. Changeover in Charging Condition with Constant PV Array Power
Figures 14(a) and 14(b) show the dynamic characteristics in terms of , Ipv, , and Ibat. The system is operated with constant rated PV power, as shown by constant PV current (Ipv). The DC link voltage () is maintained at a constant value, as shown in both figures. In Figure 14(a), the PV array delivers power for battery charging (presented by negative Ibat) and feeding power to the grid (shown by grid current ()). The charging control for the battery is achieved through battery current. The battery is now discharged maintaining the same PV array power. Therefore, both the sources now feed power to the grid, which results in increased grid current (). Figure 14(b) shows similar performance.
(a)
(b)
6.3.3. Dynamic Insolation Condition with Constant Power Fed to Grid
Figures 15(a) and 15(b) demonstrate the change in solar PV array insolation well abided by a change in bidirectional power flow of battery to maintain constant grid current. In Figure 15(a), rated power is being fed by the array to the grid with battery in fully charged condition. As soon as PV insolation drops down, the PV power is decreased (shown by Ipv). Therefore, to maintain constant grid input, extra power is delivered by the battery (shown by +Ibat). In Figure 15(b), the reverse case is discussed. Here, the battery along with the PV array with reduced insolation is feeding power to the grid. As soon as the insolation is resumed to its rated value, the battery stops discharging, and the total power is fed by the PV array only.
(a)
(b)
6.3.4. Grid Failure Condition
Figure 16 shows the blackout condition when the grid is unavailable (shown in the zoomed-in portion). It is noticed that as PV power is available, it quickly starts charging the battery (shown by −Ibat), maintaining the DC link voltage constant.
6.4. Reactive Power Compensation
Figure 17 shows one more capability of the proposed system to be able to compensate for the reactive power. It is seen that as the capacitive reactive power () is changed over to inductive reactive power (), it is properly compensated by the proposed control algorithm by regulating the DC link voltage.
7. Conclusion
A 3-phase solar-powered fast-charging station for EV charging has been proposed here. The developed system has been operated for a varied range of working circumstances, and its suitability has been examined by both simulated and experimental results. The system is comprised of a three-phase bidirectional charger with the capability of VAR compensation. The operating conditions such as charging, discharging, inductive, and capacitive reactive power coon have been demonstrated here. The four-quadrant operation has been successfully achieved by the proposed control. The suitability of the system has been justified for all modes of operation. The simple control contributes to fast response and low settling time. The feasibility is enhanced in terms of its capability of a smooth transition in various modes. Since the charger control depends on the direct power measurement and control, therefore, the distortion Iist seen in the final charging/discharging modes as well as grid currents. This proves the robustness of the system. The THD of the grid current is within 5% in simulated performance as well as experimental observation.
Appendix
A. Specification of PV and Battery (for Experimental Purpose)
Maximum PV power () = 4.3 kW, maximum PV voltage () = 334 V, and maximum PV current () = 12.4 A.
B. Specifications of EV Battery (for Experimental Purpose)
240 V and 10 Ah.
Nomenclature
| : | PV array irradiance level |
| : | Current, voltage, and power output of PV array |
| : | Current and voltage of EV battery |
| P, Q: | Active and reactive power |
| Vdc: | DC link voltage |
| : | Grid voltage and current, respectively |
| D1, Db: | Duty cycles of MPPT and battery converters |
| L1, Lbat: | Inductors of boost and battery converters |
| fS: | Switching frequency of battery converter |
| S(t): | The instantaneous value of complex power |
| , : | Grid current components after transformation |
| , : | Grid voltage components after transformation |
| , : | The change observed in active and reactive power |
| , , : | Instantaneous rectifier voltage and its components |
| , : | Active and reactive components of error |
| Kp and Ki: | Proportional and integral gains of PI controller for DC link controller. |
Data Availability
The data supporting the findings of this study are available from the corresponding author upon reasonable request.
Conflicts of Interest
The authors declare that they have no conflicts of interest.
Acknowledgments
The authors are thankful to the Science and Engineering Research Board for supporting the work under National Science Chair Fellowship.