Abstract

In this paper, two (series and parallel versions) quasi Z source inverter (qZSI)-based multioutput series-parallel topologies capable of supplying multiple three-phase AC and single boost DC outputs simultaneously are presented. The proposed parallel version topology can supply n number of AC outputs along with one boost DC with constant voltage and variable current. Similarly, the series version topology can give n number of AC including one boost DC output with constant current and variable voltage. The outputs of the proposed topologies can feed directly to DC/AC microgrids and multiple residential loads simultaneously without using any extra adapter/regulator and thereby avoiding local power conversion to meet the load demand. A hybrid pulse width modulation technique with constant frequency is used to operate the proposed topologies. The performance of the proposed topologies is verified by developing the prototypes of 2.18 kW (for parallel version) and 2.02 kW (for series version) for two three-phase AC outputs and one boost DC output. The developed prototypes show measured efficiency of 90.01% for the parallel and 89.95% for the series version topology.

1. Introduction

In recent times, the idea of the AC/DC hybrid microgrid systems (HMGS), containing many electronically controlled strategies of parallel-connected distributed resources that are capable of functioning in both islanded and grid-connected modes, evolved rapidly. It is a systematic approach to interconnect various converter outputs, renewable energy sources, energy storage systems, and local loads within certain areas. The HMGS also provides reliable, low-cost, renewable energy, improves local resilience, and improves the regional electric grid's operation and stability [1, 2]. Moreover, the required prerequisite for futuristic houses with AC/DC loads is a hybrid microgrid. For modern electrical applications such as hybrid microgrids, hybrid electric vehicles, standby power supplies, etc., conventional single-output converters cannot fully meet the simultaneous requirements of different types of voltage outputs [3]. As a result, increasing hybrid microgrid applications and other residential loads need power converters that provide multiple outputs simultaneously. Hybrid multioutput converters (MOCs) are becoming prominent to achieve these requirements due to their ability to generate multiple AC and DC outputs simultaneously. Besides, MOCs are a cost-effective alternative to solve the power supply issue in remote areas far from the grid. There are several advantages of these converters, such as higher power density, compact size, lower costs, and better reliability [4].

Many studies are ongoing due to the demand and benefits of hybrid MOCs and the literature lists different hybrid MOCs. The MOCs discussed in [58] are primarily DC/DC converters able to supply dual DC output using single or dual DC input. However, it does not have any AC at the output. Literature [9] presents a Z source-based bidirectional multiport DC/DC converter that can produce only DC. It is unable to produce multiple AC at the output and requires an additional input filter as the proposed converter has a discontinuous input current. Paper [10] deliberates a Z source-based converter capable of supplying dual AC using two DC inputs and nine switches. However, it is unable to produce DC at the output and needs an additional input filter to make the input current continuous. Topology [11] presents a DC-DC converter with a single input and multiple DC outputs having buck-boost capability with reduced switching losses, but no AC at the output. Article [12] presents a Z source-based dual-input single-output three-phase inverter with a high voltage gain. However, it is not equipped to produce multiple AC at the output. Paper [13] presents a DC-DC buck-boost converter topology with high voltage gain and dual outputs. However, it does not have any AC at the output. A high-efficiency coupled inductor-based DC-DC converter with single DC input multiple DC outputs is proposed in [14]. However, it does not have any AC at the output. Paper [15] proposed a novel nonisolated single-input dual output three-level DC-DC converter for medium and high voltage applications. However, it does not have any AC at the output. Manuscript [16] presents a novel integrated DC/DC topology with step-up and step-down outputs. Their output voltages are regulated simultaneously with no AC at the output. Articles [17, 18] presents single-inductor multi-input multioutput (SIMIMO) DC/DC converters with a high power density and low cost. However, it does not have any provision of AC at the output. Literature [19] discusses multi-input multioutput DC-DC converter with multiple DC outputs for hybrid renewable energy applications. However, it is unable to produce multiple AC. Papers [20, 21] propose DC-DC converters that can produce multiple high step-up DC outputs using a single input. However, they also do not have any provision of AC output. In paper [22], single-stage dual-input single-output high-gain three-phase inverter able to give only single three-phase AC output is proposed. However, it is unable to produce multiple AC at the output. Article [23] proposes a single-stage nonisolated bidirectional multiport DC-DC converter able to produce only DC output. However, it is unable to produce multiple AC at the output. Paper [24] proposes a high peak output power and high power conversion efficiency single-inductor multi-input multioutput (SIMIMO) converter for energy harvesting systems. However, it uses multiple inputs to produce multiple outputs. It is for low power applications and is unable to produce multiple three-phase AC outputs. Article [25] is a multicell multi-input multioutput reconfigurable converter that is unable to produce multiple AC at the output. In [26] a single-phase single-stage single-inductor multioutput DC-AC buck hybrid converter for battery charging application is proposed. It is can perform only buck operations at low power. It is unable to produce multiple three-phase AC at the output. Topology discussed in [27] is three-phase MOCs able to produce unregulated three-phase multiple AC outputs with no DC at the output. Hybrid MOCs discussed in [28] can produce a single DC and three-phase AC output simultaneously from a single DC input. Both the DC and three-phase AC outputs can be independently regulated. However, it is unable to produce multiple AC at the output. MOCs presented in [29] can supply single-phase regulated AC along with one DC output but not multiple AC. Paper [30] proposed a boost-derived converter that gives unregulated DC and AC outputs but not multiple AC. The topologies proposed in [31, 32] are an interleaved high-gain converter that gives simultaneous unregulated DC and AC outputs but not multiple AC. In [33], a minimum phase dual output hybrid converter with simultaneously regulated AC and DC outputs is presented. However, it is not equipped to offer multiple AC outputs. Hybrid MOC deliberated in [34] is a qZSI-based buck-boost MOC with dual DC and single AC outputs. With only one AC, it can supply two DC outputs in a wide range of output voltages.

In general, existing MOCs have various limitations such as the following: (1) some can produce only multi-DC outputs and cannot produce AC output, (2) some of them can produce both AC and DC having single DC or multiple DC with one AC output and cannot supply multiple AC outputs, (3) the existing hybrid MOCs are having only one AC output, that is single-phase, and thus, they are suitable only for low power application, and (4) no MOC provides multiple three-phase AC outputs and thus, they are not suitable for high power applications.

Therefore, to fulfil the abovementioned gap/limitations the qZSI network-based multioutput series-parallel topologies are proposed in this paper which have the following advantages:(1)The proposed topologies can produce regulated multiple three-phase AC and single DC outputs simultaneously with different voltage/current(2)The proposed topologies inherit all the features of the qZSI(3)The suggested topologies have inherent shoot-through (ST) protection with buck and boost, AC as well as DC voltage capability(4)Due to single-stage conversion, the proposed topologies have fewer losses and are compact, which results in higher power density and efficiency compared to the traditional multioutput converter(5)It is less susceptible to electromagnetic interference (EMI)(6)Like Z source inverter-based multioutput converters, the proposed topology does not require any input filter since it has continuous input current(7)It has fewer voltage/current stresses on passive components compared to the Z source inverter(8)For DC-DC and DC-AC power conversion in the hybrid microgrid, renewables, uninterrupted power supplies, and residential AC/DC loads, the proposed topologies outputs can be used(9)Since it can supply multiple AC along with single DC outputs, the electricity bill can be saved and surplus power can be sold to the government(10)The proposed topologies can operate at different voltages and frequencies, i.e., 50 and 60 Hz

The configuration of the proposed topologies and principles of operation are explained in Section 2. The hybrid PWM control strategy is explained in Section 3. The theoretical validation is done with hardware results in Section 4. Finally, Section 5 concludes the paper.

2. Proposed Series-Parallel Topologies

Figures 1(a) and 1(b) show the circuit diagrams of the proposed topologies in parallel and series versions. By replacing the inverter switch of qZSI with n number of parallel or series connected three-phase inverters, the proposed parallel or series version converter topologies are formed. The Z source network of the proposed topologies of both versions (i.e., parallel and series) has used the same number of components (three capacitors, two inductors, and two diodes). Figure 1(a) depicts the proposed parallel version converter, capable of supplying n number of three-phase AC outputs with the constant voltage and variable currents along with one boost DC simultaneously.

Similarly, the series version can give n number of three-phase AC outputs with different voltage and constant load current along with one boost DC. In addition, a circuit modification is done by inserting a parallel branch with a filter capacitor (Cdc) in series with a diode (Dy) across the dc-link voltage of the quasi Z source network to obtain the DC output. The suggested topologies are validated for (n = 2) two inverter units (i.e., for two three-phase AC) and one boost DC in this work. All the outputs can be independently regulated as they are working in voltage mode control.

2.1. Operation Principle, Derivation of Boost Actor, of the Proposed Topologies

The proposed topologies operate in the same way as traditional ZSIs, with shoot-through (ST) and non-shoot-through (NST) states [35].

2.1.1. Shoot-Through (ST) State

To illustrate the circuit behavior during the ST state, the circuit is simplified by replacing inverters with short circuit switches, as shown in Figure 2. All the power switches of all the legs are turned ON at the same time, and the inverter bridge operates in the short circuit mode; consequently, the dc-link voltage is equal to zero. Diodes Dx and Dy, are reverse biased and currents through them are zero. The reverse-biased diode Dy prevents a short circuit across the capacitor Cdc and thus protects against circuit damage. The source voltage in and capacitors Ca and Cb discharge the energy to inductors La and Lb. The time interval of this mode is DsTs, where Ds is the ST duty ratio during one switching period Ts. By applying KVL and KCL to Figure 2, the equations are given in

2.1.2. Non-Shoot-Through (NST) State

Figure 3 depicts the proposed topologies corresponding circuit diagram in the NST state. An inverted current source () with a potential represents the three-phase inverter units. None of the switches is triggered at the same time. The voltages through the diodes Dx and Dy are zero since they are forward biased. In this state, the inductors La and Lb discharge energy and capacitors Ca and Cb store it, and the dc-link voltage is not zero. This operating mode's time interval is (1 − Ds)Ts. The following equations are given by expression (2) by applying KVL and KCL to Figure 3.

2.2. Derivation of the Voltage Boost Factor

Based on the flux balance property of La and Lb, the average inductors voltage in steady-state over one switching period Ts is zero. Thus, from (1) to (2), we have

Solving the above equation (3), the capacitor voltages Ca, Cb and the peak dc-link voltage of the proposed topology can be derived as

Similarly, according to the charge balance principle, the average capacitor current in Ca and Cb in steady-state over one switching period Ts is zero. Therefore, from (1) and (2),

Solving equation (5), the inductor currents iLa and iLb of the proposed topologies can be derived as

From (4), the voltage boost factor (B) of the proposed topologies can be obtained as

From equation (7), as the shoot-through duty ratio (Ds) increases the denominator term (1 − 2Ds) decreases; consequently the boost factor B increases in a rectangular hyperbola manner (Figure 4(a)) and finally becomes infinite at Ds = 0.5.

The peak-phase output voltage () of the three-phase inverter is is the input voltage of each converter unit and is equal to and /2 for parallel and series versions, respectively, and ma is the modulation index of the inverter.

The modulation index (ma) controls the inverter output voltage. The ratio of peak output voltage () to the DC input voltage () is known as gain (G) of the three-phase inverter and given by

From (10), it is clear that the AC gain (G) is directly proportional to modulation index (ma) and inversely proportional to Ds. It increases as the ma increases at constant Ds. The Ds must be less than 0.5 because, at Ds = 0.5, the G will become infinite. The gain of the proposed topologies is the same as the gain of conventional qZSI. Equation (11) represents the relation between G and ma, and as the ma increases, the G also increases. The value of ma cannot be equal to 0.5 because, at ma = 0.5, G will become infinite. For better performance, high efficiency, and ripple-free sinusoidal AC output of the proposed topologies, the value of Ds should be small and ma should be large.

Note that the hybrid PWM used for this proposed topology has the following limitations for the ma and duty ratio (Ds).

Figure 4(a) shows the variation of boost factor (B) with respect to DS. As the DS increases the denominator term, () decreases which is inversely proportional to the B. Therefore, as the DS increases the B increases in a rectangular hyperbola manner, which can be seen and verified from Figure 4(a) and equation (7). Figure 4(b) shows the variation of AC gain (G) with respect to modulation index (ma) since G is directly proportional to ma; therefore, as the ma increases G also increases in a straight-line manner, which can be justified from Figure 4(b) and equation (10). Figure 4(c) shows a graphical 3D representation of the variation of the gain G with Ds and ma, where G is the ratio of peak and input DC voltage . The operating region of the proposed topologies is shown in Figure 4(d). As Ds and ma increase, the operating region also increases of the proposed topologies; however, the sum of Ds and ma cannot be more than 1.

2.3. Mathematical Expression for AC and DC Power

The following sections give the AC and DC power expressions for the parallel and series versions of the proposed topologies with two inverter modules.

2.3.1. Proposed Parallel Version Topology

The input voltage for both the inverter units in the parallel version will be the same and equal to . For the same reference voltages (ref) and balanced AC loads, the peak AC voltages ( and ) will be equal, as for both inverter units will be equal and expressed by (8). The and will be different for different ref to the inverter modules. is the peak value of the inverter unit's required AC output voltage (). For both converters, the gain G will be the same and expressed by (10) and (11).

From (8), the rms AC output voltage ( ) is given as

The three-phase AC power output (P3-∅) of both units at the same is

Similarly, the three-phase AC power output P3 − ∅ of both units at different iswhere ma1 and ma2 are the modulation indices of inverter units 1 and 2, respectively, and RAC is the AC load resistance.

The DC output power PDC of the proposed topology in parallel version iswhere RDC is the DC load resistance of the proposed topologies. It is clear from (15) and (16) that P3-∅ depends on ma and Ds both whereas PDC depends only on Ds.

2.3.2. Proposed Series Version Topology

In a series version with two units, the dc-link voltage () is equally divided across the inverter units for a balanced AC load. The peak AC voltages () will be equal for the same voltage reference () and given as (from (8)):

The rms AC output voltage () for both units is given as

The three-phase AC power output P3−∅ of a single unit is given as

The three-phase AC power output () of both units is given as

The DC output power PDC of the series version is the same as that of the parallel version expressed in (16). Since is directly proportional to ma of the inverter and equal to the peak value of ref in voltage control mode, as the ref is increased, the ma increases due to which output power increases and vice versa.

2.4. Voltage and Current Stresses of the Components of the Proposed Topologies

The voltage and current stresses of each component of the proposed topologies are shown in Table 1. In the case of parallel version, if the number of converter units increases the current stress on La, Lb, and Dx increases, since these elements are dependent on iin and . As the input voltage (i.e., switch node voltage ) depends only on constant ST duty Ds, the remains constant consequently; there will be no voltage stress. However, in the series version, as the number of inverter units increases, also increases to get the desired output voltage; therefore, the voltage stresses on capacitors and diodes increase, and current stresses remain constant.

With the increase in the stress of the front-end passive components (inductors, diodes, and capacitors of the impedance network), the rating will increase and so the cost of individual components increases. However, as the passive components counts are fixed irrespective of the number of units, the overall cost of the proposed topologies for more number units will be less in comparison with the same number of converters connected independently even for three-phase loads. Therefore, it is worthwhile and justified to use the proposed topologies with multiple units for practical applications.

Figures 5(a) and 5(b) show the variation of voltage and current stress with respect to DS. From (21) as the DS increases the voltage stress across the capacitors (Ca and Cb) and diodes (Dx and Dy) increases. Similarly, the current stress of inductors (La and Lb) increases (6).

Cost analysis of two single-unit individual converters and the proposed topologies with two units connected in parallel for the same power (total 2.18 kW) is carried out as shown in Tables 2 and 3, respectively. The total number of components (passive elements and switches) is 26 and the cost is 130.36 (USD) in the case of two single-unit individual converters. However, in the case of the proposed topology with two units, the number of components is 19 and the cost is 116.74 (USD). The number of components is less in the proposed topologies with two units, because the passive components are fixed. It is clear from Tables 2 and 3 that, for proposed topologies with two units, the price per component is high due to increased stress/rating. However, the overall cost is less as compared with its counterparts, i.e., two single-unit individual converters.

2.5. Comparative Analysis among Existing Topologies and Proposed Topologies

The proposed topologies are compared with the closely related topologies in Table 4. The table demonstrates some meaningful merits of the proposed topologies in comparison with the existing converter topologies.

3. Control Strategy for the Proposed Topologies

For the proposed topologies, the control strategy to control the DC/AC power flow is defined in the following sections.

3.1. Hybrid Pulse Width Modulation (PWM) Scheme for the Proposed Topologies

A hybrid sinusoidal pulse width modulation scheme with constant frequency shoot-through is implemented on the DSP board to control the proposed topologies. Using the hybrid PWM modulation method, the switching signals for NST and ST state for Z source inverters are obtained [36]. The complete modulation scheme of the proposed switching signal topology is shown in Figure 6. The modulation scheme's logic diagram is shown in Figure 6(a) and the corresponding switching signals are shown in Figure 6(b). For switches 1 to 6 of unit 1, sx to sz' are corresponding switching signals. By comparing the reference triangular waves with three-phase modulating sinusoidal signals (ma, mb, and mc) and their complementary signals, the switching signals are generated. The magnitudes of the modulating sinusoidal signals are as per the reference AC output voltages. The ST periods are decided by comparing the reference triangular waves with two constant dc signals (upper u(dc) and lower l(dc)), the magnitude of which is decided by the required DC output voltage. When the triangular carrier wave is higher than the upper constant dc signal u(dc) or less than the lower dc signal l(dc), ST signals are generated. In the ST condition, switches of all the legs are ON at the same time. The signal for the first leg of the inverter is produced by comparing ma and its complement with the triangular carrier wave. With the addition of the ST switching signals, signals sx and sx' are generated for upper and lower first leg switches, respectively. Comparing the carrier wave with mb and mc, respectively, as well as dc modulating signals, produces switching signals for the second and third legs (sy, sy') and (sz, sz').

Figure 6(b) shows the PWM signals with the modulation index ma = 0.3684 and Dst = 0.3158 for inverter unit 1 (Figure 7(d)) with ref 70 V and dcref 380 V. However, the PWM signals with ma = 0.2631 and Dst = 0.3158 for inverter unit 2 with ref = 50 V and dcref = 380 V are shown in Figures 6(c) and 8(a). From equation (8) is directly proportional to ma and in voltage mode control, the is equal to ref. Therefore, by varying the ref the m can vary; consequently the output power can be varied. Since the proposed topologies use two inverter units, it is possible to operate both units independently by giving them separate switching signals with a single DSP kit as shown in Figures 6(b) and 6(c).

3.2. Closed-Loop Control Strategy for the Proposed Topologies

The block diagram of the closed-loop control strategy for the suggested topologies is shown in Figure 6(d). The three-phase output voltages () of the inverter units are controlled by a PI compensator using the d − q control mode. Using the Park and Clark transformations, the sensed AC output voltages are transformed into their d − q components and compared with the d − q components of the reference voltage. Finally, from the controller, sinusoidal modulating signals (ma, mb, and mc) are obtained. The required DC voltage (dcref) is compared with the actual DC voltage and the error is passed through the PI controller. Consequently, DC modulating signals are obtained for ST cycles. By using these modulating signals with a triangular signal for PWM operation, the switching signals are produced. Figure 9 displays the block diagram of the overall implementation of the proposed topologies.

4. Verification of the Proposed Topologies

To check the effectiveness of the proposed topologies, they are validated experimentally for n = 2 (i.e., for two simultaneous AC outputs and one DC output). A hardware prototype is fabricated using the same parameters as given in Table 5. The hybrid PWM signals to control the switches of the proposed topologies are generated by a 32-bit TMS320F28335 DSP operating with a clock frequency of 150 MHz. The full steady-state and dynamic results of the proposed parallel and series version topologies are shown in Sections 4.1 and 4.2. Figure 10 provides a photograph of the experimental setup installed in the laboratory to test proposed topologies. The hardware system consists of an impedance network, DC network, two three-phase inverters, LC filter, and a DSP kit. The three-phase AC outputs are loaded with six 20 Ω loads in a Y-connection and DC output is loaded with a 100 Ω.

4.1. Validation of Parallel Version (Regulated Dual AC and Single DC Outputs)

In the parallel version, the proposed topology is tested for 2.18 kW with DC power PDC = 1444 W and AC power PAC = 735 W.

4.1.1. Steady-State Result of the Proposed Parallel Version

Figure 11 shows the simulation results and Figures 7, 12, and 13 show the hardware results for the steady-state response of the parallel version of the proposed topologies for AC and DC voltage reference ref = 70 V (peak AC for both units) and dcref = 380 V, respectively. Figure 11 shows the simulation results for the steady-state response of parallel version of the proposed topologies for AC and DC voltage references ref = 70 V (Peak AC for both units) and dcref = 380 V, respectively. With the input voltage in = 140 V, the DC load resistance is 100 Ω and the AC load resistance for both the inverter units is 20 Ω. Figure 12(a) shows the input voltage (in = 140 V), capacitors voltages (Ca = 125 V and Cb = 255 V), and inductor (La) current (iLa = 16 A). Figure 12(b) shows inductor (Lb) current (iLb = 16 A), dc-link voltage ( = 380 V), and diode voltages (Dx = 380 V and Dy = 380 V). Figure 11(c) shows DC output current and voltage (iDC = 3.8 A and DC = 380 V and) along with three-phase voltages (140 V peak-peak) of unit 1. Figure 12(d) shows three-phase currents (7 A peak-peak) of unit 1, three-phase voltages (140 V peak-peak), and currents (7 A peak-peak) of unit 2. All the values are equal to the respective theoretical values.

Figure 13(a) shows the input voltage (in = 140 V), dc-link voltage () 380 V which is equal to the theoretical value, and inductor currents iLa and iLb which are equal to the input current (iin = 16.5 A). Figure 13(b) shows input voltage, the voltage across the capacitors Ca (Ca = 125 V) and Cb (Cb = 255 V), and DC output voltage (DC = 380 V). Figure 12(c) shows the input voltage, capacitor voltage (Ca = 125 V), DC output voltage (DC = 380 V), and DC output current (iDC = 3.8 A) with DC load resistance of 100 Ω. Hence the DC power (PDC) output is equal to 1444 W. Figure 12(d) shows the dc-link voltage, DC output voltage, and current with in. As the and DC network are in parallel,  = DC = 380. Figure 12(e) shows the diodes (Dx and Dy) voltage waveforms Dx, Dy and capacitor voltage Ca along with input voltage in. The diodes Dx and Dy are forward biased during the NST state; therefore, the voltage across it (Dx and Dy) is zero. Further, during the ST state, both are reverse biased; consequently, the Dx and Dy are negative (-380 V). Figure 12(f) shows in = 140 V, DC output voltage (DC = 380 V), dc-link voltage ( = 380 V), and voltage waveform across diode Dy. The diode Dy is also forward biased during the NST state; therefore, the voltage across Dy is zero; at the same time the dc-link voltage = 380 V. Further, during the ST state, Dy is reverse biased and the Dy is negative (= -380 V).

Figure 7(a) shows the input voltage in, dc-link voltages of converter units 1 and 2, i.e., 1 = 2 = 380 V, which is equal to the output of the impedance source network (i.e. ) as the inverter units are connected in parallel. Figures 7(b) and 7(c) show input voltage in = 140 V and three-phase output voltages of converter units 1 and 2 which have a magnitude of 140 V (peak-peak) while the reference voltage ref is 70 V. Figure 7(d) shows in, output DC voltage (DC= 380 V with dcref = 380 V), and phase a voltages (140 V peak-peak) of converter units 1 and 2 with a reference voltage (ref) of 70 V for both units. Figure 7(e) shows output DC currents iDC = 3.8 A, phase a voltage (140 V peak-peak), phase b current ib1 = 7 A peak-peak of converter unit 1 along with in. Figure 7(f) shows the output DC voltage, current, and phase a voltage (140 V peak-peak) of inverter units 1 and 2. The output DC voltage and currents are 380 V and iDC = 3.8 A as the DC load resistance is 100  while dcref = 380 V and acref = 70 V. For AC reference voltage ref = 70 V, all the three phases have the same magnitude (i.e., 140 V peak-peak) and 120° phase difference showing the stable and balanced system.

Figure 12(a) shows output DC voltage DC = 380 V, output DC current iDC = 3.8 A, phase a voltage (140 V peak-peak), and phase a current (7 A peak-peak) as the AC load resistance is 20 Ω of the converter unit 1 while dcref = 380 V and acref = 70 V. Phase a voltage and current are 0° phase apart from each other. Output DC voltage, current, and phase a voltage of unit 2 with the input voltage (in = 140 V) are shown in Figure13(b)which is equal to the theoretical value.

4.1.2. Dynamic Response of the Proposed Parallel Version with the Same AC Reference Voltages

Figure 14 shows the dynamic response of the proposed topologies for DC and AC load transients with the same AC reference voltages. Figures 14(a) and 14(b) show step-up and step-down DC load change in DC network with phase a voltage (140 V peak-peak) of converter unit 1 while the ref is 70 V and dcref is 380 V. As the DC load current is changed from 3.8 A to 7.6 A, the dc-link voltage slightly decreases and finally, it restores its original position very quickly (within one cycle) and vice versa. Figures 14(c) and 14(d) show DC output voltage (dc), the step-up and step-down load change in DC network, and phase a voltage (140 V peak-peak) of converter units 1 and 2 while the ref is 70 V and dcref is 380 V. As the DC load current is changed from 3.8 A to 7.6 A, dc slightly decreases and finally it restores its original position very quickly (within one cycle) and vice versa. Figures 14(e) and 14(f) show step-up and step-down AC load change in converter unit 1. In Figure 14(e) the load current of converter unit 1 changes from 7 A to 14 A (peak-peak). As the load current increases the corresponding voltage of phase a (a1) slightly decreases and it restores its original position within very little time. The DC output voltage (DC) remains unaffected by AC load change. Similarly, in Figure 14(f) the load current of converter unit 1 changes from 14 A (peak-peak) to 7 A. As the load current decreases the corresponding voltage of phase a (a1) slightly increases and it restores its original position within very less time (in one cycle) and the DC output voltage (DC) remains unaffected. It indicates that the system with the proposed converter topology is stable and has a good dynamic response.

4.1.3. Dynamic Response of the Proposed Parallel Version Converter with the Different AC Reference Voltages

Figure 8 shows the dynamic response of the proposed converter topologies for DC and AC load transients at different AC reference voltages. Figures8(a) and 8(b) show step-up and step-down DC load change in DC network with phase a voltages 140 V and 100 V (peak-peak) of converter units 1 and 2 while the ref is 70 V, 50 V, and dcref is 380 V. As the DC load current is changed from 3.8 A to 7.6 A, the DC output voltage slightly decreases and finally, it restores its original position within one cycle and vice versa.

During the DC load, the AC output voltage (AC) remains unaffected. The 380 V DC output voltage level [37] has been used in DC data centers. Because this is a rather high voltage, particularly effective grounding and protection procedures are required.

4.2. Validation of Proposed Series Version (Regulated Dual AC and Single DC Outputs)

The proposed series version is verified for 2.02 kW with DC power PDC = 1056 W and AC power Pac = 960 W power output.

4.2.1. Steady-State Response of the Proposed Topologies in the Series Version

Figures 15 and 16 show the steady-state simulation and hardware result of the proposed series version, respectively. The DC and AC reference voltages for both the inverter units are 325 V and 80 V.

Figures 15 shows the simulation result for the steady-state response of the series version of the proposed topology for AC and DC voltage reference ref = 80 V (peak AC for both units) and dcref = 325 V, respectively. With the input voltage in = 140 V, the DC load resistance is 100 Ω and the AC load resistance for both inverter units is 20 Ω.

Figure 15(a) shows the input voltage (in = 140 V), capacitors voltages (Ca = 97 V and Cb = 228 V), and inductor (La) current (iLa = 15.5 A). Figure 15(b) shows inductor (Lb) current (iLb = 15.5 A), dc-link voltage ( = 325 V), and diode voltages (Dx = 325 V and Dy = 325 V). Figure 15(c) shows DC output current and voltage (iDC = 3.25 A and vDC = 325 V and) along with three-phase voltages (160 V peak-peak) of unit 1. Figure 15(d) shows three-phase currents (8 A peak-peak) of unit 1, three-phase voltages (160 V peak-peak) of unit 2, and currents (8 A peak-peak) of unit 2. All the values are equal to the respective theoretical values.

It is important to mention that if one unit becomes unbalanced, the other unit will continue to supply power without any interference from unit 1 and vice versa in the case of the parallel version of the proposed topologies. However, in the case of the series version of the proposed topologies, with an unbalanced load, the output voltages and currents of units 1 and 2 will become unbalanced.

Figure 16(a) shows the hardware results with the input voltage in = 140 V, dc-link voltage  = 325 V, and inductor currents iLa = iLb = 16.1 A which is also input current (iin). Since iLa = iLb, charging discharging nature of both inductor currents is the same. Figure 16(b) shows in = 140 V, capacitor voltages Ca = 97 V and Cb = 228 V, and diode switching voltage Dx = 325 V. Figure 16(c) gives input voltage in, DC output current iDC = 3.25 A, DC output voltage DC = 325 V, and dc-link voltage  = 325 V. Here is equally divided between both units 1 and 2. Figure 16(d) shows the dc-link voltage ( = 325 V) and diode voltages (Dx = Dy = 325 V) along with input voltage (in = 140 V). It is noticed that diodes are reverse biased during ST state and forward biased during NST state. During ST state,  = 0 V and the diodes Dx and Dy are reverse biased; consequently the voltages across the diodes are negative (Dx = Dy = −325 V).

Further, during the NST state, the diodes are forward biased (Dx = Dy = 0 V) and  = 325 V. Figures 16(e) and 16(f) show three-phase voltages (a1, b1, c1) and (a2, b2, c2) of the inverter units 1 and 2 with DC and phase c currents ic2. All the phases of inverter units 1 and 2 have the same voltage magnitude, i.e., 160 V (peak-peak) when the acref is 80 V. All the three-phase voltages are 120° phase apart from each other and balanced.

4.3. Efficiency Analysis and Power Loss Distribution of the Proposed Topologies

Figures 17 and 18 show the experimental efficiency of both the parallel and series version topologies. Figure 17(a) shows the efficiency versus total power output curve during parallel version when AC power is constant and DC power variable. In Figure 17(b), AC power is variable and DC power is constant. From Figures 17(a) and 17(b), it can be observed that the measured efficiency of the proposed parallel version topology is 90.01% at 2.18 kW load power.

Similarly, Figure 18(a) shows the efficiency versus total power output curve during the series version when AC power is constant and DC power variable. In Figure 18(b) AC power is variable and DC power is constant. It can be seen from Figures 18(a) and 18(b) that the efficiency of the proposed series version topology is 89.95% at a load power of 2.02 kW. It is necessary to mention that the efficiency of the proposed topologies can be further improved by optimizing the passive components, printed circuit boards, and semiconductor devices.

The power loss distribution of elements of the proposed parallel and series version is shown in Figures 19(a) and 19(b) and calculated as discussed in [38]. This is achieved by calculating the nonidealities, voltages, and currents of the elements of the proposed topologies. Power losses in the proposed topologies are contributed by conduction and switching states of semiconductor devices, as well as losses in passive components [39, 40].

For calculating the power losses, the following assumptions are taken:(1)The switches are modeled by an ideal switch having a forward resistance and a collector to emitter saturation voltage ()(2)The diodes are modeled by an ideal switch having a forward series resistance and a forward voltage drop ()(3)The winding dc resistance (DCR) of inductors and equivalent series resistance (ESR) of capacitors are considered(4)The collector to emitter current of switches (), anode to cathode current of diodes (), the current through inductors (), and the current through capacitors () are considered(5)The inductor current ripples are ignored

4.4. Losses in Switches (IGBTs)

The conduction loss of switches in the inverter bridge is calculated as

The switching loss is calculated as

Total power losses of the inverter switches are calculated as

4.5. Losses in Diodes

The conduction losses of diodes are calculated as

Generally, turn-on switching losses of diodes are negligible. The turn-off switching losses of diodes are calculated as

4.6. Losses in Inductors

The power losses of inductors include core losses and winding losses. Generally, the core losses are negligible for pulse width modulated converters. The winding losses depend on the resistances of the windings () of inductors and are calculated as

4.7. Losses in Capacitors

The power losses of the capacitors depend on the equivalent series resistances () and are calculated as

The overall loss is theoretically calculated by the addition of the above breakup of losses.

To compare the efficiencies of the proposed topologies with other topologies, a comparison table (Table 6) between the proposed topologies and closely related similar hybrid converter topologies is prepared. It is found that the efficiency of the proposed topologies is quite comparable with others. Table 6 shows that the efficiencies of BDHC [35], CFSI [29], and hybrid buck-boost converter [34] are less than that of the proposed topologies. Two converters IHC [32] and minimum phase dual output converter [33] show higher efficiency than that of the proposed topologies. However, there are some advantages of the proposed topologies over these converters. The converters of [32, 33] have only one DC and one single-phase AC output, whereas the proposed topologies are capable of providing simultaneous multiple series or parallel three-phase AC along with one DC output.

Figure 20 shows the efficiency comparison curve of the proposed parallel version topology with the other topologies [3134].

5. Conclusion

This paper presents two versions of diode-assisted switched LC series-parallel topologies with n number of AC outputs and one boost DC output simultaneously. The proposed parallel version topology is capable of giving n number of three-phase AC including one boost DC output with constant voltage and variable load current. Likewise, the series version gives n number of three-phase AC including one boost DC output with constant load current and variable voltage. All the outputs of the proposed topologies are independently regulated and can be fed directly to microgrids and multiple loads simultaneously without using an extra local adapter/regulator. The proposed topologies are operated with a hybrid pulse width modulation technique. Detailed mathematical modeling, steady-state, and efficiency analysis are carried out for the proposed topologies in the paper. Finally, 2.18 kW and 2.02 kW prototypes for parallel and series versions topology are developed, respectively, to validate the performance for two AC outputs and one DC output. The measured efficiency of the parallel and series version of the proposed topology is 90.01% and 89.95% respectively.

Abbreviations

Dx and Dy:Diodes
La and Lb:Inductors
DSP:Digital signal processor
PI:Proportional integral
SPWM:Sinusoidal pulse width modulation
HMGS:Hybrid microgrid systems
MOCs:Multioutput converters
qZSI:Quasi Z source inverter
ST:Shoot-through
NST:Non-shoot-through
RDC:DC load resistance
Cdc:Filter capacitor
:dc-link voltage
in:Input voltage
KVL:Kirchhoff's voltage law
KCL:Kirchhoff's current law
Ds:Duty cycle
Ts:Switching period
C:Capacitor voltage
D:Diode voltage
DC:DC voltage
iL:Inductor current
iD:Diode current
B:Boost factor; peak AC voltage
ma:Modulation index
G:Gain
3d:Three-dimensional
AC:AC voltage
RAC:AC load resistance
ref:AC reference voltage
P3−∅:Three-phase AC power
PDC:DC power
dcref:DC reference voltage
LC:Low pass filter
Y:Star connection
Z:Impedance.

Data Availability

No data were used to support this study.

Conflicts of Interest

The authors declare that they have no conflicts of interest.