A New Multi-Output DC-DC Converter for Electric Vehicle Application

Multiport converters play a significant role in portable electronic and electric vehicle (EV) applications. In literature, different configurations of single-input multi-output (SIMO) converters are presented. Most of the SIMO converters generate the outputs with operating constraints on the duty ratio and charging of inductors. The cross-regulation problem is still a challenge in SIMO converters design. A SIMO topology is proposed in this study to overcome the limitations mentioned earlier. It can generate three different output voltages without constraint on the duty cycle and inductor currents (like <inline-formula> <tex-math notation="LaTeX">$i_{L1} > i_{L2} > i_{L3} $ </tex-math></inline-formula> or <inline-formula> <tex-math notation="LaTeX">$i_{L1} < i_{L2} < i_{L3}$ </tex-math></inline-formula>). Cross regulation problems do not exist in the proposed topology, so the load voltage <inline-formula> <tex-math notation="LaTeX">$V_{01}$ </tex-math></inline-formula> (<inline-formula> <tex-math notation="LaTeX">$V_{02}$ </tex-math></inline-formula>) (<inline-formula> <tex-math notation="LaTeX">$V_{03}$ </tex-math></inline-formula>) is not affected by the variation of output current <inline-formula> <tex-math notation="LaTeX">$i_{03}$ </tex-math></inline-formula> (<inline-formula> <tex-math notation="LaTeX">$i_{02}$ </tex-math></inline-formula>) (<inline-formula> <tex-math notation="LaTeX">$i_{01}$ </tex-math></inline-formula>). The loads are isolated from each other during control. In the laboratory, a 200 W prototype circuit is developed; simulation and experimental results are validated.


I. INTRODUCTION
In the past decade, there has been an increase in demand for renewable energy sources utilization in electric vehicles (EVs), auxiliary power, and grid-connected applications [1]- [5]. In these applications, multiport DC-DC converters are essential for Hybridizing energy sources which lead to, reduce the components count, complexity, and cost of the system compared to several separate single input DC-DC converters [6], [7].
Over the past decade, MPC converters have been presented. A new SIMO converter is proposed in [8]. This structure simultaneously generates boost, buck, and inverted outputs controlled independently. However, producing 'n' voltage levels requires n + 2 switches, which increases the overall size and cost of the converter. Unexpected mistakes in calculating state-space equations and output voltages for a SIMO converter given in [8] are addressed and rectified in [9]. The single coupled inductor-based SIMO buck is presented in [10] with lesser output inductor current ripple than single The associate editor coordinating the review of this manuscript and approving it for publication was Zhilei Yao .
inductor SIMO converters. Nayak and Nath [11] elaborately presented the comparative performance of SIDO converters based on the coupled inductor and single inductor (SI) in terms of cross-coupling issues. Furthermore, they proposed that the coupled inductor SIDO converter has a better steadystate and transient performance. Nevertheless, in a SI SIMO configuration inductor is switched between the loads, which causes high ripples and cross-regulation problems.
Different control approaches are proposed in the literature to overcome the cross-regulation issue in a single inductor-based SIMO converter; the current predictor controller is presented in [12] instead of the conventional chargebalance approach. However, generating the duty ratios for active switches has been somewhat complicated. Similarly, the deadbeat-based control approach is presented in [13]. It is based on output current observer, and hence it is sensitive to the noise and significant parametric variations. In [14], a multivariable digital controller-based SIMO converter is proposed to minimize the voltage ripples, suppress the cross-regulation problems, and regulate the output voltages. However, controller design may lead to an increase in complexity. A non-isolated and single switch SIMO converter topology is presented in [15]. It has fewer components and reduces the cost of the system. However, it may be challenging to regulate the outputs independently.
To alleviate the problems in a single inductor SIMO converter, a non-isolated SIMO converter is proposed in [16]- [25], which are independently regulated the output voltages and does not require an additional control circuit. In [16], a new SIDO converter topology is proposed to integrate buck and super lift converter for generating the step-up and step-down output voltages for electrical vehicle applications. It has a constraint on-duty ratio viz. D 2 < D 1 , which limits the operation range of D 1 by increasing D 2 . The topologies proposed in [17] and [18] have fewer semiconductor switches. However, the operation of the converter is based on the charging time of inductors (i.e., i L1 > i L2 ). So this keeps the constraint on-duty ratio.
The combination of high gain step-up and SEPIC converter-based SIMO is suggested for PV applications in [19]. In this configuration, both the outputs are higher than the supply voltage and improve the output voltage by adding the capacitors and diodes. Nevertheless, the number of capacitors and diodes affects cost and conduction losses. A new SIDO buck-boost topology is developed in [20] to generate positive and negative outputs. A multi-output converter is suggested in [21] with the reduced part count. However, it has more diodes, which increases conduction losses. A structure of SIMO configuration is introduced in [22] with the advantages of reducing the passive filter size and low voltage stress. High-density multi-output converter is proposed in [23] for portable electronic applications based on the front-end switched-capacitor technique with improved power density and reduced switching losses.
Modified SEPIC and interleaved-based high step-up SIMO converter are introduced in [24]. It consists of a voltage multiplier, coupled inductor, and switched capacitors to boost the output voltage in sustainable energy applications. However, it has complexity due to more components. The SEPIC-Cuk converter-based four-phase interleaved converter is suggested for SIMO applications in [25]. It has the advantages of low ripple voltage, compact size and is suitable for high power applications with a dynamic response.
In the conventional approach, EVs' auxiliary power supply system to handle the load requirements is shown in Figure 1. It looks simple, but the main drawback of this approach is a cross-regulation problem, and the loads are not isolated from each other during their operation. There is also the chance of grounding issues while charging the battery with simultaneously turn-on loads and if the ground is involved. Further, the circuit complexity will increase to convert one of the negative output voltages into buck-boost operation mode.
In the proposed work, the onboard power converter is the main subject of the study. The configuration of the circuit shown in Figure 2(a) is such that energy stored in the inductor is confined to one output only and is not shared with the other outputs during the control, which allows regulating the output voltages with independent duty-cycles. More importantly, the loads are isolated from each other during control, and the cross-regulation problem is successfully eliminated. Also, there are no problems associated with grounding as it is an onboard power converter even if charging of battery and ground is involved.
The remaining sections of the article are organized as follows: The developed SIMO configuration and modes of operation are presented in Section II. Small-signal modeling is presented in Section III. The controller design, parameter design, power loss analysis, and comparative assessment are discussed in Section IV. The simulation and experiment and results are shown in Section V. Summarized in Section VI.

II. PROPOSED SIMO CONFIGURATION AND MODES OF OPERATION
The proposed single input three-output DC-DC configuration is depicted in Figure 2(a). In this configuration the components are as follows, input voltage V DC , switches (S 1 -S 3 ), diodes (D 1 -D 3 ), and passive elements (L 1 -C 1 , L 2 -C 2, and L 3 -C 3 ). It can generate three different output voltages, i.e., boost (V 01 ), buck-boost (V 02 ) with positive voltage polarity, and buck (V 03 ). The proposed converter is suitable for independently regulating the output voltages by the duty cycles D 1 , D 2, and D 3 , respectively. The theoretical waveforms of circuit elements are depicted in Figure 2 The proposed configuration is different from the conventional parallel combination of buck, boost, and buck-boost configuration. In the proposed circuit configuration, the loads are isolated during the simultaneous control. From the following figures, one may observe that during mode-1operation, load R 3 alone through S 3 is connected to the input power supply, but the other loads are isolated, as shown in Figure 3(a). Similarly, during mode-2 only load R1 alone through D 1 is connected to the input supply, but other loads are isolated, as depicted in Figure 3(b). In the proposed control strategy, all the loads are isolated from each other during their control in any mode of operation. However, this feature is impossible VOLUME 10, 2022 in the conventional parallel combination of buck, boost, and buck-boost converters.
This circuit configuration looks very simple, but it is novel and valuable. A comparison in terms of the number of components, modes of operation, and working conditions between the conventional and proposed SIMO converter is presented in Table 1, as given below In the conventional approach shown in Figure1, the main drawback is the cross-regulation problem, and the loads are not isolated from each other during their operation. Further, the circuit complexity will increase to convert the negative polarity of output voltages in the buck-boost mode of operation.
The proposed structure has the following advantages: a) It is a simple structure and no assumptions on operating duty ratio ( Switches S 1 , S 2 , and S 3 are turned ON. The current flow path is depicted in Figure 3(a) , and the energy port V DC magnetizes L1, L2, and L3. Consequently, the C 1 and C 2 are discharged to the loads (R 1 ) and (R 2 ), respectively, whereas (C 3 ) is charged. The inductor currents and capacitor voltages are represented in Eq. (1)-(4).
2) SWITCHING STATE 2 In this state, L 1 , L 2, and L 3 are de-magnetized and deliver their energy to the load through D 1 , D 2, and D 3 , respectively. It is illustrated in Figure 3(b). The inductor currents and capacitor voltages are in Eq. (5)-(11) as follows, where c 1 , c 2 , c 3 , c 4 , c 5 , and c 6 areinitial values. Output voltages of the proposed configuration are as follows D 1 , D 2 , and D 3 are duty ratios of the S 1 , S 2 , and S 3 respectively. It is observed that during switching state-1operation, load (R 3 ) alone through S 4 is connected to the ground but the other loads are isolated even when the ground is involved during charging the battery, as shown in Figure 3(a). Similarly, during switching state-2 only load (R 1 ) alone through D 1 is connected to the ground, but other loads are isolated from the ground and the load (R 1 ) as well as depicted in Figure 3(b). In the proposed control strategy, all the loads are isolated from each other during their control during any mode of operation. Moreover, the configuration of the circuit is such that energy stored in the inductor is confined to one output only and is not shared with the other outputs during the control and also, which allows controlling the output voltages with independent duty-cycles. As a result, the load voltage V 01 (V 02 ) (V 03 ) is not influenced by the variation of load current i 03 (i 02 ) (i 01 ). Hence the proposed configuration with this control approach avoids all the issues about crossregulation problems even when the ground is involved during battery charging. More importantly, the configuration is simple and it can generate three independent outputs without any assumptions on inductor currents (i L1 > i L2 > i L3 or i L1 < i L2 < i L3 or i L1 = i L2 = i L3 ) and/or operating duty cycle.

III. SMALL-SIGNAL MODELING
The transfer function of the proposed topology is derived from small signal analysis as [26]. The state-space equations (16)-(25) are as follows where state-space coefficients are A, B, C, D and E X(t) = state vector, U(t) = input vector, and y(t) = output vector Where, State vector = x(t), Input vector = u(t) and Output vector = y(t). (18)- (21), as shown at the bottom of the next page.
The output voltagesV 01V02 andV 03 are determined byd 1 The proposed configuration control transfer function is given in Eq. (23-25) as followŝ v 01 (s) The bode plot of the proposed configuration is illustrated in Figure 4 for verifying the stability. It is observed that the gain margin is 6.65 dB, 1.54 dB, and −1.55 dB, whereas the phase margin is 90 • and 90 • and 0.393 • respectively for transfer functions of the proposed converter given in (23)- (25).

IV. CONTROLLER DESIGN, PARAMETER DESIGN, SMALL-SIGNAL MODELING, POWER LOSSES CALCULATIONS, AND COMPARATIVE ASSESSMENT A. THE CONTROL METHOD OF THE PROPOSED CONVERTER
A suitable control scheme is essential for good voltage regulation. A control transfer function has been derived for each output by using small-signal modeling. It is cascaded with a controller, as illustrated in Figure 5, where the PI-controller is chosen as given (26) to reduce the undamped behavior of the system and improve the low-frequency performance, i.e., it reduces the steady-state error [18].

B. PARAMETERS DESIGN CONSIDERATIONS
The converter parameters design can be calculated using equations (27)-(29) as given in [27].
R L max f s , Calculation of filter capacitance value is where V 01,02,03 = Output voltage, D max = Maximum duty ratio, f s = Switching frequency, R L1,2max = Maximum load resistance, r c = Maximum ESR of thefilter capacitor and V cpp = Peak-to-peak value of the capacitor.
The ripple voltage (V r ) is 1% of V 0

C. POWER LOSSES CALCULATIONS
Power losses are essential for calculating efficiency as follows [28], [29], equations are presented in Eq. [30]-[35] P loss_IGBT = P con + P sw (30) The IGBT conduction losses are R on = Switch ON-state resistance, V Fo = Threshold voltage, i F = Forward current, and T = Switching period.
The switching losses are calculated as, where E ON and E OFF are and is the energy delivered in ON and OFF time of the power switches, respectively, and f is the switching frequency   The power loss of the capacitor (P C ) is calculated as where I Lrms is the RMS value of the inductor current and I Crms RMMMS values of the capacitor current. r C, and r L are the ESR of the capacitor and inductor, respectively. The efficiency of the proposed converter is η = P out P out + P sw + P con + P L + P C (35)

D. COMPARATIVE ASSESSMENT
The comparative assessment is presented in this section in terms of components, passive elements, and stresses on active switches for recently developed SIMO DC-DC converters in the literature.

1) THE NUMBER OF COMPONENTS
The comparative assessment based on the number of components has been done with recently reported single input multi output topologies as depicted in Table. 3. A single switch SIMO converter is presented in [15]; it reduces the control complexity of the system. Nevertheless, it may not be easy to regulate the outputs independently. A SIDO configuration is developed in [16] using a super-lift Luo-converter. It generates both step-up and step-down outputs. However, it has more components count. Reference [20] observed that the presented SIMO generates positive and negative output voltages. However, it increases the number of components that result in big size, high cost, and more power losses. The proposed converter in [18] has reduced part count and is suitable for EV auxiliary power supply applications.
Nevertheless, it has such as i L1 > i L2 for generating output voltages. In [21], a multi-output converter is developed with reduced components. Nevertheless, it may have high conduction losses due to more diodes. A new SIDO topology is presented in [22] has the advantages of low semiconductor stress and the size of the filter elements. However, it has more device count, which may affect the size of the power converter. A high-density multioutput converter is suggested in [23] for portable electronic applications, has more active switches, which may decrease the converter efficiency. The comparison presented in Table 3 depicts that the proposed configuration is simple, and there are no assumptions on the inductor currents and operating duty ratio. It can generate three independent outputs and loads are isolated from each other during control and the cross-regulation problem is successfully eliminated.

2) VOLTAGE STRESS COMPARISON
The efficacy of the proposed configuration is also compared in terms of the voltage stress and is shown in Table 2. The maximum voltage stress of the proposed topology in [20] is the addition of input and output voltage. Similarly, topologies introduced in [18] and [22] have less voltage stress, i.e., half of the output voltage and supply voltage. The proposed configuration in [16] is the subtraction of output and supply voltage. The maximum voltage stress in the presented topology in [15] and proposed configuration is the output voltage. The suggested topology in [23] has low semiconductor stress. From Table 2, one can observe that the proposed topology has less semiconductor stress compared to suggested topologies in [16], [18], and [22]. The current stress on the switch is high in the presented topologies [18], and [20] is equal to the addition of inductor current. The proposed topology and converter proposed in [15], [16], [22], and [23] have less current stress, i.e., current flows through the one inductor (i L2 ) only.
The proposed converter's comparative analysis has also been done in terms of control complexity and power density, as depicted in Table 4. The control complexity and power density are mainly dependent on the number of active switches and the total number of components in the power converter. It is observed that the topologies proposed in [19], [20], and [21] have a lesser number of active power switches as compared with [18], [22], [23], and the proposed topology. Hence they had low complexity in control. Similarly, the power density of any DC-DC converter mainly depends on the total number of components, especially active power switches, and they occupy more space. Consequently, the proposed power converter and topologies presented in [18], [20], and [22] have higher power densities.
Moreover, with the comparison of different aspects of power converter such as component count, semiconductor stresses, from Table. 2 suggests that each converter has its own merits and demerits. The proposed converter structure has low semiconductor stresses and avoids cross-regulation problems if the ground is involved during the charging of the input battery. The configuration is suitable for EVs' auxiliary power system applications.

A. SIMULATION RESULTS
The model has been built in MATLAB environment to verify the proposed system with V DC = 50 V, frequency is 50 kHz, and the duty ratio is 50%. The parameter details are  Table. 2. The corresponding output voltages (V 01 , V 02 , and V 03 ) and inductor currents (i L1 , i L2 , and i L3 ) are illustrated in Figure 6(a-f), respectively. The output voltages in Figures 6(a), 6(c) 6(e) are close to the theoretical results. The closed-loop control is implemented for the proposed configuration, and the dynamic performance of the overall system is validated for a sudden change in the input voltage. Figure 7. shows the simulation result of closed-loop control for a sudden change in the input voltage (V DC ) from 50V to 70 V at 0.5 sec. The PI control gains are chosen as K p = 0.1 and K i = 15 for Buck output, similarly K p = 0.005 and K i = 0.5 for Boost and Buck-Boost voltages. The results show that the proposed configuration generates stiff independent output voltages and is not affected by the sudden change in supply. The efficiency of the proposed converter at different duty ratios and various power ratings is depicted in Figure 8.

B. EXPERIMENTAL VERIFICATION
The practical feasibility of the proposed configuration is tested on the laboratory prototype of the proposed converter. The parameter specifications of the designed prototype are given in Table 2. The control signals are generated using controller DSP 28335 for IGBT's (STGW30NC120HD). The test is conducted at V DC = 50 V and D 1 = D 2 = D 3 = 50%. The corresponding, Figure 9(a), 9(c), and 9(e) illustrates the output voltages (V 01 , V 02 , and V 03 ), inductor currents (i L1 , i L2 , and i L3 ), and load currents (I 01 , I 02 , and I 03 ) are depicted in Figure 9(b), 9(d) and 9(f), respectively. Output voltages match the theoretical simulation results, i.e., Eq. (12), respectively. The converter efficiency at different output powers is illustrated in Figure 10(a), and the laboratory hardware experiment on the developed prototype circuit is shown in Figure 10(b).

VI. CONCLUSION
The structure of the SIMO converter is proposed in this paper. The operating principle and modes of operation have been explained in detail. The proposed configuration is simple and without assumptions on the charging of inductors and operating duty cycle. It can generate the buck, boost, and buck-boost output voltages with independent regulated voltages. Cross regulation problems do not exist in the proposed topology, so the sudden change in inductor and load currents does not affect the output voltages. Finally, simulation and experimental results validate the proposed converter operation and performance.