Comparative analysis of controllers for a smooth switching between battery and ultracapacitor applied to E-vehicle

Generally, all EVs/HEVs batteries are charged from by conventional plugin type only. This type of arrangement all time will not correct the response of the vehicle. The solar panel is integrated with the vehicle itself to charge the battery during the solar power available time and also we can avoid load burden on the local grid. If sunlight not available throughout the day then the vehicle can be charged from a local renewable grid (Sadagopan et al., 2014). A solar power station has proposed for electric vehicle charging, generally many electric vehicles are fed with self-charge mechanism containing with its own PV panel. The designed solar power station can be useful if a vehicle own power is not sufficient to charge the battery (Bhavnani, 1994). Designing of the finest control method to battery/UC based HESS is one of the difficult tasks associated with EVs. This works mainly focusing on developing the supervisory control method which is able to give the optimal power sharing of energy sources. Including both conduction as well as switching losses a DC-DC converter is considered. Numerically a minimized solution is formulated based on the neural network which decides the proper power splitting. Along with the neural network method, a rule-based optimized control technique also implemented and compared both results during vehicle running conditions (Shen et al., 2015). A real-time non-linear control technique is developed and implemented to battery-supercapacitor powered EV. Several non-linearity is considered during changing of the source from on to another according to the driver requirements, with all the considerations predictive control model is obtained and implemented (Wu et al., 2015). Energy management is one of the finest problem present in HESS powered EV/HEVs. In order to obtain the superior energy management of HESS, this includes battery and UC a supervisory energy management strategy is developed. The proposed control technique is purely embedded with different intelligent controllers (Emadi et al., 2008). EVs are extensively used to achieve the green transportation system, for that generally fixed charging as well as mobile charging stations are utilized to charge the energy sources present in the EV. In case of fixed charging stations all charging ports and the other equipment are in stable condition, which means EV itself should reach the particular station to charge the battery/UC. On the other hand, mobile charging stations are available with the movable vehicle, which enhances the flexibility of charging to the EV. In this work proposed the method to charge the battery/UC in a quick way in case of both mobile as well as fixed charging stations (Atmaja, 2015). To obtain the effective energy management of HEV different DC-DC converter topologies are adopted. The supercapacitor (SCAP) bank is integrated into the already existed test bench system. The test bench system mainly containing with two motors and those are separately connected to the alternators to generate the required power to the load. The main intention of this work is to provide peak power with SCAP with a short interval of the time period that may be 20 sec. The original system consisting with main power source only and in this additionally, supper capacitor is connected to enhance the overall performance of the system (Camara et al., 2008).

An effective energy management scheme is designed with a neural network system. The HEV fed with various primary sources like a fuel cell, a battery which is unable to grab the energy during the regenerating period of the vehicle. In order to store energy during regenerating breaking effectively, UC banks are used. In order to store and measure the instantaneous current, vehicle speed and main source voltage values, the digital signal processor is used. Initially, energy management can be done in a conventional way thereafter used a neural network approach to obtain better performance (Moreno et al., 2006). To fulfill the unexpected driver behavior and different load condition, HESS has been designed with various artificial intelligence techniques. Recent preachers mostly concentrating on the optimal usage of fuels which used to drive the electric vehicles mainly on batteries and different algorithms have been proposed to find the optimal way of utilizing the energy sources (Sánchez, 2015; De Castro et al., 2012). During the transforming of traditional vehicles as EVs, several challenges are taking place, in that high power density energy source is the main one and which should be given better results during acceleration and regenerative breaking periods. An FLC based control method is implemented to EV for proper power splitting of energy sources. With the suggested control technique, battery and UC able to give the base as well as peak powers depending upon the vehicle road conditions (Dusmez et al., 2014). For BDC, zero voltage transition (ZVT) circuit is applied in order to create an interface between UC and some other main power sources like a battery which is implemented to HEVs/EVs. The designed auxiliary circuit provides smooth switching of energy sources. The proposed ZVT acts a buck during charging of UC, on the other hand, it will act as a boost during power supply to the load. In this mainly concentrates on the buck operation of BDC, for these different configurations are analysed (Mirzaei et al., 2011). This paper mainly addressed the problems present in control technique which are already existed implemented to EVs and these are powered by HESS. Here the BDC is used to control the current in the battery as well as SCAP actively. Total two control objectives are stated in that first one is used to identify the major source current and load voltage. The second one is used for smooth switching during load changing time (Jung et al., 2014). For multi-source EVs, a real-time energy management system is proposed especially for small urban areas. The energy management system is incorporated the rule-based technique with proper equations will come for proper sharing of energy among multiple sources depending upon the vehicle dynamics. The suggested method gives the optimized energy sharing during all conditions of EV (Trovao et al., 2015). In this, a novel HESS is designed with battery and UC. The traditional HESS has integrated the source through high rating DC-DC converters to the vehicle, whereas in proposed HESS; sources are connected through low rating DC-DC converters to the vehicle. An interface is created between battery and UC in order to meet the peak power requirement of the vehicle. The relative load profile is created to the battery for supplying the power directly to the vehicle even UC voltage value drops to a low value. The battery is not used here to charge directly from the regenerative braking which will reduce the charging and discharging burden (Cao et al., 2012). This work mainly concentrated on designing of a smooth control method of switching for UDC as well as BDC. Here extra core winding is inserted on the existed inductor coil in order to avoid the commutation problem present during working period. Zero voltage switching or hysteresis control is used to perform the transition of converters up to a maximum extent in load. In order to enhance the output power as well as reduce the filtering equipment make sure that main inductor current should be in continues mode (Zhang et al., 2003).

Each PV cell is capable of generates the voltage of less than one volt which means each Si photovoltaic cell develops the output voltage of around 0.7 V during open circuit time and 0.5 V under working condition. No. of cells are allied in series and parallel to assemble a PV module and a number of modules are allied in series and parallel to produce the required output. Using Si-based photovoltaic modules the PV system converts only 15% of solar energy into electricity. Perfect solar PV cell is demonstrated by a current source and an inverted diode coupled in parallel to it as shown in Figure 3.

Two key parameters are considered from the ideal circuit representation, as short-circuit current (I_sc) and open circuit voltage which are often used to illustrate the PV cell. By short-circuiting the terminals of the cell the photon generated current as shown in Figure 3(b), flows out of the cell called as a short circuit current (I_sc). Thus, we can say that I_cell=I_sc as the current I_cell is flowing in a single series circuit. As the terminals are short-circuited then the voltage across the circuit is equal to zero i.e V_oc=0 and the short-circuit current is the PV cell load current (or the output current which is very maximum as equal to that of the current source or photovoltaic photon generated current i.e. I_cell=I_sc=I_m Similarly, when the terminals are open circuited i.e. no load and nothing is connected as represented in 3 (c), the load current of a PV cell becomes zero. And the load voltage of a PV cell is equal to the maximum applied source voltage or open circuit voltage i.e. (V_m=V_oc)

3.png

Figure 3. (a) PV cell equivalent circuit, (b) PV cell at short circuit condition (c) PV cell at open circuit condition

The PV cell output current can be found by applying KVL to circuit 3(a)

$\mathrm { I } _ { \mathrm { m } } = \mathrm { I } _ { \mathrm { cell } } - \mathrm { I } _ { \mathrm { d } }$     (1)

The current through diode can be represented with bellow equation

$I _ { d } = I _ { r s c e l l } \left[ e ^ { \left( \frac { Q V } { K T a p } \right) - 1 } \right]$      (2)

By replacing Id in Equation 2, it gives the current-voltage association of the PV cell as shown below

$I _ { m } = I _ { c e l l } - I _ { r s c e l l } \left[ e ^ { \left( \frac { Q V } { K T a p } \right) - 1 } \right]$      (3)

The diode reverse saturation current (I_rscell) is calculated by the open circuit condition of PV cell as illustrated in Figure 3(b). From the Equation (3) it is observed that I_m=0 and solve for I_rscell

$I _ { \text {rscell} } = \frac { I _ { \text {cell} } } { \left[ e ^ { \left( \frac { Q V } { K T a p } \right) - 1 } \right] }$      (4)

The photon generated current is directly relative to the irradiance and temperature, on the other hand, voltage is directly relative to the irradiance and inversely proportional to the temperature. The value of ISCR is provided by the manufacturer datasheet at STC (standard test condition). At STC, the working temperature and irradiance are 250C and 1000 W/m respectively.

In the present work, the standard PV array has been taken and generated the power with different temperatures and irradiance values. Thereafter using DC-DC solar panel voltage has been changed according to the electric vehicle requirement. Here tree control switches have been connected to the solar panel, battery, and UDC. The SOC of the battery and the output voltage of the solar panel decide the control switches action. Finally, electrical vehicle requirement can be fulfilled using solar power based on the availability of sunlight.

The proposed work can be analyzed in four modes with different loads. Switches action always based on the load condition on the electric motor. All four mode condition with different loads and switches ON and OFF conditions are explained separately in each and every mode of operation.

5.1. Mode-I operation

In this mode, a heavy load is applied to the motor, which initiates the operation of BDC in boost mode. Due to that switch, S₃ gets the required pulse signals and other two switches S₁, S₂ are in disable state. Here the battery gets charged from the solar panel depending on control switches operation. Finally, during this mode of operation, the entire power required by the motor can supply with UC only.

4.png

Figure 4. Main circuit model with DC-DC converters during mode-I operation

5.2. Mode-II operation

Slightly more than rated load is related to the mode-2 operation. The battery and UC collectively meet the required power of the electric motor, which means UC is reducing the extra burden on the battery. So switches S₁ and S₃ are in the active state, another switch S₂ is in disable state. In this work, the battery gets charged from the solar panel during sunlight available timings and discharges the same amount of energy to the electric motor during no irradiance and temperature period. Finally, in this mode of operation, the power flows from battery as well as UC to the electric motor.

5.png

Figure 5. Main circuit model with DC-DC converters during mode-II operation

5.3. Mode-III operation

During this mode a rated load is applied to the electric motor, due to that switch S₁ is in an active state and other switches S₂, S₃ are in disable state. During rated load condition motor requires average power only, which can be supplied by the battery itself. In this work, the battery gets charged from the solar panel during sunlight available timings and discharges the same amount of energy to the electric motor during no irradiance and temperature period. Finally, in this mode of operation, power flows to the motor from battery only.

6.png

Figure 6. Main circuit model with DC-DC converters during mode-III operation

5.4. Mode-IV operation

A light load or no load is applied during mode-4 operation. During this period, motor needs little quantity of power that may be less than average power supply the battery. In this case, the battery can meet the power requirement of the electric motor as well as UC. Switches S₁, S₂ are in active state and another switch S₃ is in disable state. In this work, the battery gets charged from the solar panel during sunlight available timings and discharges the same amount of energy to the electric motor during no irradiance and temperature period. Finally, in this mode of operation, power flows to the motor as well as UC from the battery.

7.png

Figure 7. Main circuit model with DC-DC converters during mode-IV operation

7.1. Mode-I operation results (heavy load condition)

During this mode, a heavy load is applied to the electric motor. MFB with ANN as well as MFB with PID has taken 0.6 sec and 0.9 sec times at starting to reach the steady state. Thereafter at 1.5 sec, a heavy load is applied due to that both the controller's speed responses decreased drastically. MFB with ANN controller has taken 0.8 sec time to reach steady state, whereas MFB with PID doesn’t reach the steady state within a given time.

At 1.5 sec a heavy load is applied to the electric motor due to that huge current variations have been observed from above figure 10. MFB with ANN current response reached steady state within 0.8 sec, thereafter no current ripples presented. But MFB with PID current response does not reach the steady state within a stipulated time.

During starting, the motor can stats with UC so controlled pulse signals are produced to BDC (boost), after some time period again controlled signals are produced to UDC as boost and BDC (buck). Due to a heavy load applied at 1.5 sec controlled switching signals are produced to BDC (boost) until response reaches the steady state. After reaching steady state, again controller signals are produced to BDC, working under buck mode and UDC (boost).

9.png

Figure 9. Speed responses of two hybrid controllers during mode-I operation

10.png

Figure 10. Current responses of two hybrid controllers during mode-I operation

11.png

Figure 11. Controlled pulse signals of UDC as well BDC produced by MFB plus ANN controller

12.png

Figure 12. Controlled pulse signals of UDC as well BDC produced by MFB plus PID controller

13.png

Figure 13. Voltage and current responses of UDC at the input

14.png

Figure 14. Voltage and current responses of UDC at output

15.png

Figure 15. Voltage and current responses of BDC at the input

16.png

Figure 16. Voltage and current responses of BDC at output

7.2. Mode-II operation results (slightly more than rated load condition)

17.png

Figure 17. Speed responses of two hybrid controllers during mode-II operation

18.png

Figure 18. Current responses of two hybrid controllers during mode-II operation

19.png

Figure 19. Controlled pulse signals of UDC as well BDC produced by MFB plus ANN controller

20.png

Figure 20. Controlled pulse signals of UDC as well BDC produced by MFB plus PID logic controller

21.png

Figure 21. Voltage and current responses of UDC at the input

22.png

Figure 22. Voltage and current responses of UDC at output

23.png

Figure 23. Voltage and current responses of BDC at the input

24.png

Figure 24. Voltage and current responses of BDC at output

7.3. Mode-II operation results (rated load condition)

25.png

Figure 25. Speed responses of two hybrid controllers during mode-III operation

26.png

Figure 26. Current responses of two hybrid controllers during mode-III operation

27.png

Figure 27. Controlled pulse signals of UDC as well BDC produced by MFB plus ANN controller

28.png

Figure 28. Controlled pulse signals of UDC as well BDC produced by MFB plus PID controller

29.png

Figure 29. Voltage and current responses of UDC at input

30.png

Figure 30. Voltage and current responses of UDC at output

31.png

Figure 31. Voltage and current responses of BDC at the input

32.png

Figure 32. Voltage and current responses of BDC at the output

7.4. Mode-IV operation results (no-load condition)

33.png

Figure 33. Speed responses of two hybrid controllers during mode-IV operation

34.png

Figure 34. Current responses of two hybrid controllers during mode-IV operation

35.png

Figure 35. Controlled pulse signals of UDC as well BDC produced by MFB plus ANN controller

36.png

Figure 36. Controlled pulse signals of UDC as well BDC produced by MFB plus PID controller

37.png

Figure 37. Voltage and current responses of UDC at input

38.png

Figure 38. Voltage and current responses of UDC at output

39.png

Figure 39. Voltage and current responses of BDC at input

40.png

Figure 40. Voltage and current responses of BDC at output

7.5. Continuous load condition

41.png

Figure 41. Speed responses of two hybrid controllers during a continuous load condition

42.png

Figure 42. Current responses of two hybrid controllers during a continuous load condition

43.png

Figure 43. Controlled pulse signals of UDC as well BDC produced by MFB plus ANN controller

44.png

Figure 44. Controlled pulse signals of UDC as well BDC produced by MFB plus PID controller

45.png

Figure 45. Voltage and current responses of UDC at the input

46.png

Figure 46. Voltage and current responses of UDC at the output

47.png

Figure 47. Voltage and current responses of BDC at the input

48.png

Figure 48. Voltage and current responses of BDC at output

Table 1. Comparative analysis of three hybrid controllers in four modes of operation based on a load applied

7.6. Battery parameters

49.png

Figure 49. The battery parameters during charging and discharging periods

In present work battery minimum SOC has taken as 20%, if battery SOC is bellowed 20% then it should get charged from the solar power directly, in the same way, discharges the same amount of power to the electric vehicle until its SOC reduces to 20%. From the above figure 49, it is clear that during discharging time of the battery current showing positive whereas shown negative value under the charging period.

7.7. Solar panel parameters

Figure 50 represents the solar panel input parameters and duty cycle value connected at UDC side. Here solar power can be generated based on the irradiance and temperature availability so, that is the reason why those two parameters have been changed and according to that duty cycle of the converter. The irradiance and temperature values are changed and obtained the different voltage levels corresponding to changed values of irradiance and temperature.

50.png

Figure 50. Solar panel parameters and duty cycle of the converter

Table 2. DC-DC converters ON/OFF states based on the mode of operation

Table 3. MFB controller outputs corresponding to speed