1. Introduction
Transformers are one of the essential components in electricity transmission and distribution systems, enabling transmission of power with the features of being relatively inexpensive, highly reliable, and highly efficient. While power transformers step up the voltage of the generator for transmission, they also step down the transmission voltages for industrial and residential uses [
1]. However, while they are frequently used in transmission/distribution systems, they possess such drawbacks as: (i) bulky structure, because of the in-line operation frequency, (ii) sensitivity to harmonics, and (iii) voltage drops [
2]. The emerging concepts for the conventional power transformers offer new magnetic/insulation materials [
3,
4], optimization in the core [
5], and manufacturing processes [
6].
Nowadays, the tendency to integrate technological advancements in power electronics in mid/high-voltage and high-power levels to empower future grids has grown dramatically. The capabilities and features of power electronics technology have led to improved power quality, increased stability, increased reliability, and more flexibility of power systems. In this regard, solid-state transformers (SST) [
7] which are also known as power electronic transformers (PET) [
8] and power electronic traction transformers (PETT) [
9] are one of the emerging and promising power electronics technologies. SSTs have replaced the traditional bulky line transformers by performing many features, such as reactive power compensation, galvanic isolation, power factor correction, size reduction, voltage regulation, and DC integration.
Recently, several SST topologies which are formed as (i) a single-stage including matrix or cycle converters [
10], (ii) two-stage including AC–DC front end rectifier and inverter [
11], and (iii) three-stage including AC–DC/DC–DC/DC–AC converters [
12] have been proposed in the literature. Among them, although single-stage topology provides superior efficiency and power density, the three-stage SST is the most widely studied and used because of the smart features it can perform. The most outstanding feature of the three-stage SST is to provide low-voltage (LV) DC links for the integration of distributed generation (photovoltaic panels, fuel cells) and battery storage systems [
13]. The integration of the aforementioned energy units and medium-voltage distribution grid can be realized with a DC–DC converter which is able to perform galvanic isolation and voltage gain, and an inverter which makes it possible to reach medium-voltage levels in AC from stages of SSTs. Numerous DC–DC converter topologies have been studied in the literature.
Dual active bridge (DAB) DC–DC converter topology is frequently used in SST applications because of the features of bidirectional power transfer, isolation, zero-voltage switching, flexible voltage range, and lower number of passive components [
14,
15]. In [
16], a DAB-based SST study was conducted in order to interface photovoltaic arrays with 400 V and a utility grid with 10 kV. In addition to DAB topology, variations of it, such as LLC resonant and half-bridge topologies, are employed in SSTs adopted for medium-voltage applications. The front bridge arm of the DAB was replaced with half-bridge submodules in order to reduce the converter volume and the cost in [
17]. Authors in [
18] proposed an SST with a unidirectional LLC resonant converter. Some authors have addressed a topology that is integrated with a series-half-bridge LLC resonant DC–DC converter and a neutral-point clamped full-bridge inverter to reduce the series cell number in comparison with traditional two-level SSTs [
19]. The aforementioned SST topologies adopted for medium-voltage applications are employed with cascaded/modular multilevel schemes. The related concepts consist of an extendable structure with converter cells including bidirectional DC–DC converters and inverters connected in series on the high-voltage AC side and in parallel on the low-voltage DC side, considering the voltage levels of the systems [
7]. A current-fed series resonant converter-based SST topology presented in [
20] reaches MV level using high turns-ratio HFT. The available typical SST concepts which are endowed with half-bridges, full-bridges, DABs, and 3-phase full bridges were categorized in [
21]. The common aspect of the DAB-based SST topologies, and of the derivations of it, is the (requirement of modular structure) use of cascaded/modular multilevel structure in order to reach medium-voltage level. Also, the common drawback of the aforementioned DC–DC converter topologies is that they are operated at low-voltage gain for high efficiency. Therefore, a large number of series-connected converter cells and switching elements are required to reach the medium-voltage level. According to the existing topologies in the literature, a comparative investigation of the proposed topology is shown in
Table 1. The proposed converter is compared with some DC–DC converter topologies used in SSTs in terms of five parameters: SST voltage, DC–DC converter voltage range, frequency, number of switches/diodes, and efficiency. As illustrated in
Table 1, the proposed topology uses less switches. Furthermore, the efficiency of the proposed topology is higher than the compared existing topologies.
Along with the widespread applications of renewable energy sources (RESs) with low-output voltage, high-gain DC–DC converter topologies have also gained importance. Several high-gain DC–DC converter topologies have been improved, such as (i) z-source [
22], (ii) quasi z-source [
23], (iii) switched capacitor boost [
24], (iv) stacked boost [
25], (v) quadratic boost [
26], (vi) three-level boost [
27], and (vii) cascaded boost [
28] converters in the literature. Although RES-powered systems are equipped with high-gain DC–DC converter topologies, they are not widely used in SST applications that allow the integration of RESs and energy storage units. As it is understood from the literature review, SST topologies are frequently formed with modular/cascaded multilevel converter cells that require large number of active switching components with high costs to reach medium-voltage levels. In addition, the conventional DC–DC converter topologies are able to provide limited voltage gain because of the efficiency concern.
This paper proposes a novel SST structure with an improved flyback converter equipped with a quasi z-source converter for a medium-voltage utility grid. Also, the mathematical model of the proposed qZ_iFC and the corresponding control schemes are presented in the study. The proposed converter is composed of three cascaded converter cells and each converter cell consists of a quasi z-source converter, a flyback converter, a full-wave rectifier, and an inverter. The qZ_iFC is able to provide high-voltage gain with low-input voltage. Compared with the DC–DC converters employed in (conventional) SSTs presented in the literature, the proposed converter uses less switches to reach medium-voltage level. The performance of the proposed system is verified through different case studies.
The rest of the paper is organized as follows:
Section 2 introduces the proposed improved SST topology and describes its mathematical model operation limits.
Section 3 presents the control scheme of the overall system. Furthermore, the operations stages of the proposed novel qZ_iFC were analyzed. The performance results of the system are presented in
Section 4.
Section 5 presents the conclusions.
2. Power Circuit of the Proposed qZ_iFC-Based SST
In this section, a new, highly energy-efficient DC–DC/DC–AC stage of SST configuration, which is proposed to manage power flow from a low-voltage DC link to medium-voltage AC grid, is explained. The main motivation of the proposed converter is to provide high-voltage gain in order to makes it possible to reduce the number of series converter cells. The AC–DC stage of the SST is not investigated as it will be low-voltage rectifier topologies. The proposed system is capable of feeding from conventional and new generation low-voltage rectifiers, as well as directly feeding from renewable energy sources or DC microgrids. As illustrated in
Figure 1, the proposed converter is composed of three quasi z-source converters, three flyback converters, and a full-bridge modular multilevel converter (FB-MMC) in order to interface 100 V input side and 3.3 kV (peak) medium-voltage grid. The quasi z-source converter is employed to reach high-voltage gain with reduced switch number in comparison with traditional SST topologies. The switches of the quasi z-source converter is triggered considering the gain requirement of the overall system. From a topological perspective, the improved converter integrating the flyback converter switch (S
f) and the S
qz switch form a complementary operation to generate a high-frequency AC voltage on the HFT. The flyback converter is equipped with an IGBT without body diode to prevent the shoot-through current which prevents saturation from reaching the HFT. The proposed qZ_iFC is endowed with an HFT with 1:1:1 turns ratio. Although increasing the turns ratio reduces the core losses due to the low magnetic flux density, it has a drawback in the overall efficiency of the system because it increases the winding losses [
29,
30]. For this reason, the high-gain capability is provided with the quasi-z-source converter. Thus, the proposed converter is able to achieve high gain while providing galvanic isolation.
The system can be assumed to be composed of a two-stage circuit and can be analyzed by dividing it into two parts. The first analysis is performed for quasi z-source converter and the second analysis is performed for the flyback converter. The quasi z-source converter has two types of operational state: the non-shoot-through state and the shoot-through state. Assuming that during one switching cycle Ts, the interval of the shoot-through state is DTs, and the interval of the non-shoot-through state is (1 − D)Ts. Since the closing of the switch Sqz represents the shoot-through state, Dqz represents the duty cycle of the Sqz switch. While the Sqz is triggered during the shoot-through state, the diode of the quasi z-source converter is forward biased during the non-shoot-through state.
During the time interval of non-shoot-through and shoot-through states, the voltages of the quasi z-source converter can be obtained as
where V
s and V
pn represent the supply voltage and output voltage of the quasi z-source converter, respectively.
The voltages of the quasi z-source converter stage capacitors can be derived as follows:
The V
zDC voltage can be derived by summing Equations (3) and (4). Thus, the V
zDC and V
s voltage relations can be derived as follows:
The second analysis is performed for the flyback stage. The secondary side of the relevant converter is equipped with a center-tapped HFT. The switches of S
qz and S
f operate in a complementary fashion. The steady-state relationship between input and output sides of the improved flyback converter is derived as
where D
f represents the duty cycle of the flyback converter switch.
For the designed qz_iFC, the voltage gain can be increased by regulating the duty cycle in order to reach the desired output voltage levels. The characteristics of qZ_iFC topology is examined and the voltage gain/efficiency curve according to the duty cycle is illustrated in
Figure 2. The proposed topology provides high-voltage gain values up to 20 and high-efficiency values up to 99.2%. The optimum duty cycle can be maintained between the range of 0.4 and 0.8 considering the most convenient area of efficiency and voltage gain for practical implementations.
The qZ_iFC makes it possible to eliminate the bulky line transformer from the inverter side using HFT. Thus, the MV power conversion system is only characterized by using H-bridge converters and an inductor (L)/capacitor (C) filter. The simplified circuit topology of the FB-MMC is shown in
Figure 1. The FB-MMC consists of one phase and three series-connected full-bridge sub-modules (FBSMs) and each FBSM is fed by the individual DC links that are provided by the proposed qZ_iFC. While the switch pairs of S
fb1_1-S
fb4_1 and S
fb2_1-S
fb3_1 are triggered in turn, FBSMs are triggered synchronously. Moreover, the diodes of the switches are used to provide the paths for the load current driven by the stored energy in the inductor.
3. Control Scheme and Operation Stages of Proposed System
The control scheme of the proposed SST is composed of two different parts including qZ_iFC and inverter controllers as illustrated in
Figure 3. The qZ_iFC controller is adopted to adjust the output voltage of the developed qZ_iFC stage by regulating the shoot-through duty cycle of S
qz and the duty cycle of Sf operating in reverse. The controller triggers the switches by complementary duty cycles considering the required output voltage rating. The related controller monitors the DC-link voltage (V
dc) continuously. It compares the difference between the actual and reference values of the DC link. Following the determination of the error value, the computed value is applied to a PI controller. The PI controller produces the duty cycle value for the Sf switch and its complement for the S
qz switch.
The steady-state operation waveforms of qZ_iFC in terms of switching states (S
qz, S
f), input current (I
i), voltages of the quasi z-source converter’s capacitors (C
qz1, C
qz2), and primary voltage of the HFT (V
pri) are presented in detail in
Figure 4. As shown in
Figure 4, the proposed converter has two operation stages during one switching cycle of T
s. Since the switches of S
qz and S
f operate in a complementary fashion, 1 − D
qz represents the D
f.
Stage 1 [t1 − t2]: At t1, Sqz switch is in ON state while Sf switch is in OFF state. Thus, quasi z-source converter operates in shoot-through state. Whenever it is in the related switching state, the current of the input inductor increases gradually. While the input inductor charges, the capacitors Cqz1 and Cqz2 discharge until the time point t2. During this state, the shoot-through current path is blocked on the flyback converter side for high currents that can adversely affect the transformer, since there is no body diode in the Sf switch.
Stage 2 [t2 − t3]: At t2, Sqz switch is in OFF state while Sf switch is in ON state. Thus, the quasi z-source converter operates in non-shoot-through state. During this operation, the capacitors Cqz1 and Cqz2 charge synchronously due to the energy stored in the inductors of the quasi z-source converter. The converter operates in continuous conduction mode. In parallel with the charging of the relevant capacitors, VzDC voltage is also increased to high levels in proportion to their sum. The flyback converter switch provides the accumulation of energy in the HFT's primary winding, and the energy is transferred to the secondary windings. Following these operation sequences, the proposed qZ_iFC provides the desired DC-link voltage level with low ripple.
The inverter is equipped with phase-shift pulse width modulation (PS-PWM) to control the output voltage by rapidly switching the DC rail voltages. While the H-bridge converters share the same modulating sinusoidal signal of Vout-ref, three phase-shifted high-frequency carrier signals are used for individual H-bridge converters. The carriers are phase-shifted by an angle of 60 degrees obtained by dividing 180 degrees by the number of bridges.
4. Performance Analysis
This section presents the performance evaluation of the proposed qZ_iFC SST. In order to test and evaluate the performance of the improved SST topology and controller, a proof-of-concept model was developed in MATLAB/Simulink environment. The qZ_iFC SST model was designed for 3.3 kV (peak) application with 25 kVA cascaded qZ_iFCs (each rating is 8.3 kVA) and 25 kVA MMC. The operating frequencies are determined as 10 kHz and 20 kHz for MMC and qZ_iFC stages, respectively. The switching components of the power circuit are IGBT because of the power rating and switching frequency of the system. The parameters of the developed system are shown in
Table 2. The presented system was tested with different case studies, including steady-state and transient conditions.
In this section, it is emphasized in some performance results of the proposed system for the 3.3 kV (peak) 25 kVA application. In the proposed system, a 100 V DC source is used as an input of the overall system. The presented system is tested under some case studies including transient and steady-state conditions of different loadings on the consumer-side. The dynamic load profile was created as 15, 25, and 19 kW. A 15 kW loading performance test was performed between t = 0 s and t = 3 s. Then, the load-side was increased to 25 kW at t = 3 s. Subsequently, the load was decreased to 19 kW at t = 6 s.
The main aim of the study is to increase the low-input voltage using high-gain qZ_iFC for the full-bridge MMC.
Figure 5 shows the waveforms of the DC side of the proposed system. Voltage/current of input and V
zDC/V
DC voltages are illustrated for a variety of case studies. It is obvious that the low-input voltage level is increased to a high-voltage level using the shoot-through specification of the proposed converter. While the load variations directly affect the input current, the output voltage of the qZ_iFC (V
DC) remains constant at the desired level (1100 V). In other words, the proposed qZ_iFC supplies uninterrupted power transfer by providing a robust input DC link to the MMC.
The voltage, current, and power waveforms of load (AC) side are introduced in
Figure 6. During the transient and steady-state operations, the output voltage of the proposed SST is kept constant at 3.3 kV peak value. In addition, the current waveform of the load varies in direct proportion to the load change. However, the state detection and reference extraction times cause instantaneous voltage fluctuations at point-of-load variations, the controller smoothly compensates the transient states. During these states, the switches of S
qz and S
f are triggered by the complementary duty cycle values computed by the controller in order to provide constant DC-link voltage for the inverter. As illustrated in
Figure 6, the adapted duty cycle values of S
qz switch are 0.698, 0.799, and 0.759 for states, respectively. As understood from the zoom-in view of
Figure 6, the total harmonic distortion (THD) of load voltage is about 2.8% and is always kept below voltage harmonic limits of the IEEE 519-1992 standard [
30].
In addition, the presented system was also tested with the case studies of input voltage variations and disturbances under constant load conditions. During the related case, the input voltage variation is designed as follows: V
s = 75 V between 0 and 3 s, V
s = 125 V between 3 and 6 s, and V
s = 100 V between 6 and 9 s. In order to demonstrate the operation performance of the proposed converter, the input/output voltage, current and power waveforms are presented in
Figure 7. The S
qz and S
f switches are triggered by the complementary duty cycle values computed by the controller in order to provide constant DC-link voltage for the inverter. As illustrated in
Figure 7, the adapted duty cycle values of the S
qz switch are 0.78, 0.64, and 0.70 for states, respectively. The resulting DC-link voltage highlights that the proposed converter is able to keep the DC-link voltage constant during the input voltage variations.
Also, the disturbance rejection capability analysis was performed. As the load increases or decreases, the DC link often experiences undesired sags or swells to accommodate the varying load current. Therefore, the rejection capability of DC-link control against load disturbance becomes critical. The disturbance rejection capability of the converter can be described by applying disturbance to the input of the converter. A 20% (i.e., 0.2 p.u.) step disturbance that represents a sudden load disturbance was applied at t = 9 s to the output of three controllers of the qZ_iFC simultaneously. The responses of the synchronous controllers to the given disturbance, the DC-link voltage and the duty cycles are shown in
Figure 7.
As illustrated in
Figure 7, the proposed control scheme achieves a high disturbance rejection capability with a smooth control signal. The DC-link voltage is not affected by the input disturbances. The disturbance rejection capability of the controller also verifies the robustness of the controller.
The efficiency of the system was investigated under dynamic loading, varying input voltages, and input disturbances states, and illustrated in performance results figures. In
Figure 6, the system efficiency values are 97.9%, 97.7%, and 98% for the 15, 25, and 19 kW loading conditions, respectively. In
Figure 7, the system efficiency values were obtained as a minimum of 96.5% and a maximum of 98.5% for varying input voltages and input disturbance conditions, respectively.