Review of Solid-State Transformer Applications on Electric Vehicle DC Ultra-Fast Charging Station

Abstract

The emergence of DC fast chargers for electric vehicle batteries (EVBs) has prompted the design of ad-hoc microgrids (MGs), in which the use of a solid-state transformer (SST) instead of a low-frequency service transformer can increase the efficiency and reduce the volume and weight of the MG electrical architecture. Mimicking a conventional gasoline station in terms of service duration and service simultaneity to several customers has led to the notion of ultra-fast chargers, in which the charging time is less than 10 min and the MG power is higher than 350 kW. This survey reviews the state-of-the-art of DC ultra-fast charging stations, SST transformers, and DC ultra-fast charging stations based on SST. Ultra-fast charging definition and its requirements are analyzed, and SST characteristics and applications together with the configuration of power electronic converters in SST-based ultra-fast charging stations are described. A new classification of topologies for DC SST-based ultra-fast charging stations is proposed considering input power, delta/wye connections, number of output ports, and power electronic converters. More than 250 published papers from the recent literature have been reviewed to identify the common understandings, practical implementation challenges, and research opportunities in the application of DC ultra-fast charging in EVs. In particular, the works published over the last three years about SST-based DC ultra-fast charging have been reviewed.

1. Introduction

The technical requirements of battery fast charging cannot be fulfilled by the electric vehicle (EV) onboard charger, which is connected to a single-phase or three-phase AC domestic supply [1]. To decrease the charging time, a huge amount of DC power is required [2], so it is mandatory to supply the charging station microgrid (MG) from a medium voltage (MV) AC source. This connection could be carried out through a low frequency transformer (LFT), a hybrid transformer (HT), or a solid-state transformer (SST) [3]. It is worth mentioning that the notion of SST is used here in a broad sense, which means that it can be considered as any conversion structure implemented with switching converters that is capable of transforming MV-AC into DC in the region between 200 V and 500 V.

This paper is motivated by the identification of the state of the art of ultra-fast charging stations, which can be based either on LFT or SST as illustrated in Figure 1a,b, respectively. It is apparent in the figure that the DC SST-based ultra-fast charging station allows the integration of renewable energies, energy storage, and AC/DC loads, so the resulting electrical architecture can be considered as an AC/DC/Hybrid MG. In a clear-cut contrast, a LFT-based ultra-fast charging station does not facilitate the integration of renewable energies but offers the advantage of high reliability at the expense of a relatively big size (

Table 1. A comparative analysis of DC LFT-based and SST-based ultra-fast charging stations.

Ultra-Fast Charging Station	Size	Controlability	Integrarion of Large Power Plants	Fault Protection	Modularity	Reliability
LFT-based	Big	Limited	No	No	No	High
SST-based	Small	Unlimited	Yes	Yes	Yes	Medium

).

1.1. State of the Art

Carbon dioxide emissions have a direct impact on climate changing and global warming. Currently, a significant portion of carbon dioxide emissions is issued by fossil fuel vehicles (FFV) [4]. According to statistics, the portion of transportation in carbon production is more than 20% [5]. Therefore, FFVs are being replaced by EVs. The EVs have higher efficiency and less maintenance in comparison with FFVs [6], which can be refueled within two minutes. The charging time of EVs is much longer since they are normally charged in the residential distribution network with AC on-board chargers (OBCs) (120–240 V) [1]. The provided power through these slow OBCs is 1.44–22 kW [7]. Therefore, for a 100 kWh battery, it takes more than 5 h to be fully charged, which is not satisfiable for long trips and increases the anxiety of the drivers. The advancement of OBCs is still continuing, but there are some barriers to high-power AC OBCs, such as charger size, cost, weight, maintenance, and EV safety requirements [8]. The next solution is to utilize an integrated charging structure by means of a traction motor, which was previously used in the railway traction industry [9]. Three-phase electrical motor winding can achieve high power and charge the battery through the inverse operation of the drive inverter [10]. However, this technique has some adverse effects on an EV, such as high stress on the traction system, torque production, acoustic noise, and mechanical vibrations, which reduce the lifespan of the whole EV [11]. Another alternative solution is to build a DC ultra-fast charging station like traditional fuel stations to directly communicate and charge the EV battery [2]. In a DC ultra-fast charging station, the charging time of EVs should significantly decrease to be competitive with traditional fuel stations. In order to decrease an EV’s charging time, the amount of the delivered power to the EV’s battery charger should increase. Until now, the maximum DC power delivered to the EV battery in the market is 250 kW. The limit is due to the high temperature [12] and state of health (SOH) issues of the battery [13]. However, it is foreseen to reach 1200 kW maximum delivered DC power [14] as a result of advancements in battery structure design, vehicle electrical architecture, and material in the near future [15].

In order to build a DC ultra-fast charging station, this should be fed by MV transmission lines. In order to decrease the voltage level to the one required for the EV’s battery chargers, LFT, HT [16], and SST [17] can be used in the charging stations. The first suffers from low efficiency, bulky size, high maintenance cost, slow dynamics, limited controllability, and lack of plug-and-play functionality [18]. These shortcomings make LFT less functional in DC ultra-fast charging stations. The HT is the combination of partially-rated power converters, normally between 5–20% of the rated power, in the input or output of a LFT. The HT also suffers from a bulky size, but it is more efficient and controllable than LFT. A 2 kVA small-scale HT is experimentally validated in [19]. The last promising solution to change the MV level into the EV battery charger voltage level is to utilize an SST. It has a higher degree of control freedom, storage integration capability, harmonic filtering, smaller size, fault current limitation, bidirectional power flow capability, and higher operational frequency [20]. In addition, the advancement of high-voltage high-power silicon carbide (SiC) and gallium nitride (GaN) devices has paved the way for the extensive use of SSTs in high-power applications [21]. The performances of SST, HT with integrated energy storage, and LFT have been technically and economically compared in [3] for distribution grid applications. Since the use of SST in power systems is still new, there exist many issues and challenges, especially in the DC ultra-fast charging applications, which should be addressed.

1.2. Objectives of the Paper

The aim of this paper is to provide a step-by-step comprehensive review about DC ultra-fast charging stations where SST is used. This paper provides information about ultra-fast charging definitions and requirements, SST definition, structure, and challenges, as well as DC SST-based ultra-fast charging station configurations, control, and challenges. A new topology configuration for a DC SST-based ultra-fast charging station is classified based on the input type, number of the output ports, and the converter type. The DC SST-based ultra-fast charging station is considered as an MG, and possible challenges of MGs are discussed.

1.3. Organization of the Paper

The rest of this paper is organized as follows. In Section 3, a complete description of the battery charging is surveyed. Section 3 describes the state of the art of DC ultra-fast charging stations. In Section 4, a comprehensive review of SST topologies is presented. Section 5 investigates all the technical and research works performed in DC SST-based ultra-fast charging stations. Finally, Section 6 concludes the paper.

2. Battery Charging Definition and Barriers

The battery charging procedure has different definitions and requirements. One of the most important definitions is the battery C rating. The battery capacity is usually rated at 1C (1C current) [22]. As an example, a 400 V battery with 100 kWh capacity can be charged or discharged with 250 A within one hour. This means that with the use of 1 kA of current, the battery can be recharged in 15 min. Therefore, an increase in C rating yields a faster charging or discharging time. However, C rate increase has technically some limitations, including increases in internal energy losses, battery thermal tolerance, and reducing battery’s life-cycle [23]. The 6C rate of charging is considered as an ultra-fast charging method according to the US Department of Energy (DOE) [24]. As a result, a trade-off exists between ultra-fast charging and battery health [25].

The battery discharging procedure is related to the demanded power and is not controllable. However, the battery charging task can be controlled with optimal charging strategies. Monitoring and estimating the state of charge (SOC) and the state of health (SOH) play a significant role in battery charging methods [26,27]. Different charging strategies have been proposed in the literature for lithium-ion (Li-ion) batteries [28,29]. Some of them are independent from the battery chemical model and dynamic, e.g., constant-current (CC) [30], constant-current constant-voltage (CCCV) [31], multi-stage CCCV [32], and pulse charging techniques [33]. The typical CCCV charging method is illustrated in Figure 2 [34]. Some others are based on empirical models [35], including circuit-based models [36] and neural network models [37]. The pseudo-two-dimensional (P2D) [38] and single particle model (SPM) [39,40] are the most usable circuit-based models for battery charging optimization, battery SOC prediction [41], and control. The last charging approach is based on the battery electrochemical model, in which all the physical parameters are observed. This approach is the most complicated and accurate among others. However, due to the high computational burden, its real-time charging controller implementation is not practical yet.

To summarize, the optimal charging strategies should be as fast as possible while the temperature limitation and battery SOH are maintained within the desirable range [42]. In EV application, the responsible part for the optimal operation of battery charging is the battery management system (BMS), which analyses and monitors all the information taken from the EV battery pack [43]. The BMS cooperates with the battery charging circuit through the exchange of data in order to control the injected current and voltage to the battery cells [44]. Battery charging circuits can be placed on-board inside EVs [45], off-board in the charging stations, or wireless (inductive) with unidirectional or bidirectional power flow [46]. The OBCs connect to the residential single-phase or three-phase AC voltage, and the off-board chargers connect to the DC voltages with different kinds of cables, some of which are illustrated in Figure 3. The connectors are categorized into four levels according to their maximum power transfer capability. The levels 1 and 2 AC chargers are supplied with 120 V and 240 V AC, respectively [7]. In the market, the maximum supplied power for the battery is 1.92 kW in level 1 and 19.2 kW in level 2. The International Electrotechnical Commission (IEC) has published standard requirements for EV plugs, socket-outlets, connectors, and inlets (IEC 62196) [47]. According to the IEC 62196-2:2016 [48], the maximum power for the level 2 on-board charging mode is 33.6 kW (480 V AC, 70 A). In the pre-release version of IEC 62196-1:2022 PRV [49], the AC charger maximum power was determined as 172.5 kW (690 V AC, 250 A) and the DC ultra-fast charging maximum power was 1200 kW (1500 V DC, 800 A).

Table 2. Characteristics of battery charging connectors.

Model	Charging Level	Voltage	Maximum Power	100 kWh Battery Charging Time	Charger Location	Maximum Current
SAE J1772	Level 1 (Slow)	120 $V_{AC}$ ($1\varphi$)	1.44–1.92 kW	52–69 h	on-board	12–16 A
	Level 2 (Slow)	208–240 $V_{AC}$ ($1\varphi$)	5.0–19.2 kW	5–20 h	on-board	24–80 A
	Level 3 (Fast)	50–1000 $V_{DC}$	80 kW	75 m	off-board	80 A
	Level 4 (Ultra-fast)	50–1000 $V_{DC}$	400 kW	15 m	off-board	400 A
Mennekes/EU	Level 1	250 $V_{AC}$ ($1\varphi$)	4–8 kW	12.5–25 h	on-board	16–32 A
	Level 2	480 $V_{AC}$ ($3\varphi$)	13.3–22 kW	4.5–7.5 h	on-board	27–45 A
	Level 3	500 $V_{DC}$	70 kW	85 m	off-board	140 A
	Level 4	500–1000 $V_{DC}$	200 kW	30 m	off-board	200 A
GB/T	Level 1	250 $V_{AC}$ ($1\varphi$)	7 kW	14 h	on-board	28 A
	Level 2	400 $V_{AC}$ ($3\varphi$)	12.8 kW	7.8 h	on-board	32 A
	Level 3	250–950 $V_{DC}$	60 kW	100 m	off-board	250–400 A
	Level 4	250-950 $V_{DC}$	237.5 kW	25 m	off-board	250–400 A
Tesla	Level 1	120/240 $V_{AC}$ ($1\varphi$)	1.9–7.7 kW	13–52 h	on-board	16–32 A
	Level 2	208/250 $V_{AC}$ ($1\varphi$)	2.8–11.5 kW	9–35 h	on-board	48 A
	Level 3	300–480 $V_{DC}$	250 kW	24 m	off-board	800 A
	Level 4	300–480 $V_{DC}$	350 kW	17 m	off-board	800 A

shows some of the charging connectors’ voltage and power ratings existing in the market categorized by their charging speed. It should be noted that, due to the weight and size limitation in EVs, the utilization of OBCs is limited to the low-voltage low-power applications [50]. Therefore, they cannot be charged sufficiently quickly to mimic the same refueling experience in conventional gasoline stations. Commercial on-board charging including dedicated and integrated as well as off-board charging infrastructures are surveyed in [51]. Different charging topologies such as bidirectional active front end (AFE) [52], interleaved unidirectional charger topology [53], and bridgeless power factor correction (PFC) stages [54] are reported in detail in [55].

To increase the distance traveled by the EVs, the first solution is to raise the EV’s battery capacity, which yields a bigger EV size, cost, and weight [56]. Different types of batteries have been used in EVs such as lead–acid (Pb–acid), nickel–cadmium (NiCd), nickel–metal–hydrid (NiMH), Li-ion, Li-ion polymer, and sodium–nickel–chloride (NaNiCl) [57]. Among these, Li-ion batteries have been extensively commercialized in the EV industry thanks to their higher energy density, longer life cycle, and lower maintenance [58]. The EV’s battery characteristics for some vehicle manufacturers are shown in

Table 3. Battery characteristics for some of EV manufacturers.

Car Model	Battery Type	Capacity	Voltage	Range	Fast-Charging Time	Supercharger Maximum Power
Tesla Model Y	Li-ion	75 kWh	360 V	487 km	31 min	250 kW
Tesla Model X	Li-ion	100 kWh	350–375 V	536 km	28 min	250 kW
Tesla Model 3	Li-ion	50–75 kWh	350–400 V	507 km	31 min	250 kW
Tesla Model S	Li-ion	103 kWh	400 V	637 km	27 min	250 kW
Volkswagen ID.3	Li-ion	62 kWh	408 V	415 km	38 min	125 kW
Volkswagen ID.4	Li-ion	77 kWh	400 V	514 km	38 min	125 kW
Volkswagen ID.5 GTX	Li-ion	82 kWh	400 V	490 km	33 min	135 kW
Renault Zoe E-Tech	Li-ion	52 kWh	375 V	390 km	78 min	50 kW
Renault Megane	Li-ion	60 kWh	400 V	360 km	54 min	130 kW

[59,60,61]. As can be observed from this table, the battery capacity and supercharger maximum power reach 103 kWh and 250 kW, respectively, which reduces the full-charging time to within 30 min. From a material point of view, the fast charging of Li-ion batteries is reviewed in [62]. The research is still ongoing in the context of material engineering for the advancement of energy storage systems and to improve their energy density. Their charging time should be reduced to the range 5–10 min to be competitive with conventional fuel stations.

Another solution is to build a DC ultra-fast charging station that is like traditional gasoline stations [63]. This can give the opportunity to the drivers to frequently charge their vehicles in urban or interurban DC ultra-fast charging stations. From an economical point of view, building DC ultra-fast charging stations is more appropriate [8], as detailed in the next section.

3. DC Ultra-Fast Charging Station

Different possible configurations of DC ultra-fast charging stations are AC-coupled, DC-coupled, and hybrid-coupled. These are shown in Figure 4. In this figure, all the power flows could be bidirectional to satisfy vehicle-to-grid (V2G), vehicle-to-home (V2H), and vehicle-to-vehicle (V2V) applications in the future smart grids [64]. To deliver the power to or from an EV, bidirectional power electronic converters are used, which are comprehensively reported in [65]. An example of an AC-coupled DC ultra-fast charging station is in Mountain View, California, with six superchargers and a 400 kWh integrated energy storage for load shaving within peak hours. The AC-coupled DC ultra-fast charging station has some advantages such as appropriate converter technology, switchgear, protection devices, and well-established standards. The only disadvantages of AC-coupled stations is the huge amount of used power converters in order to integrate with the DC systems, which leads to more complex and less efficient systems. In addition, some control challenges appear in the island mode of operation such as reactive power control, inverter synchronization, and voltage/frequency control. The DC-coupled configuration, shown in Figure 4b, has less conversion stages, higher efficiency, and a simpler control structure, although it has a complex protection scheme and non-standardized metering [66]. Moreover, the integration of renewable energies and energy storage is more simple. In addition, the proposed structure of a DC ultra-fast charging station with a bipolar DC bus in [67] can solve the grounding issue of the DC network [68,69]. In the hybrid-coupled charging station, shown in Figure 4c, which is a combination of a DC-fast charger and single-phase and three-phase AC chargers, the system suffers from load nonlinearity problems due to the use of power electronic converters as well as unbalance issues due to the connection to single-phase chargers or different fast charging pilot power requirements.

As reported previously in Table 3, the supercharger maximum power existing in the EV market is 250 kW. Consider this value as the maximum peak charging power; therefore, the maximum peak power required in the eight-slot DC-coupled DC ultra-fast charging station (Figure 4b) would be 2 MW whenever all the pilots are simultaneously charging the EVs with the lowest SOC. This is the worst-case scenario that could happen in the charging station. It should be considered that, due to the differences in the SOCs of EV batteries, the charging power delivered to each battery could be different (see Figure 2).

As shown in Figure 4, the DC ultra-fast charging station has three power stages, which are listed below. The power stages and utilized power electronic converters inside the EVs can be found in [70].

3.1. Power Stage 1

The first power stage is the voltage level change. The huge amount of power required in the DC ultra-fast charging station should directly be provided from the MV distribution line. Since the battery chargers work at less than 1000 V, a transformer is required to step-down the voltage level. Different types of transformers are used in the literature to step-down the voltage level in power systems, including LFT, SST, and HT [71]. These are shown in Figure 5 [3].

The LFTs are bulky in size, weighty, and suffer from a high installation cost. There is no control freedom in the input/output voltage/current of the LFTs. Although with mechanical tap changers, the voltages can be controlled to a limited extent, this is not sufficient in the fast voltage fluctuations of modern power systems in which the penetration of intermittent renewable energies, EVs, and DC ultra-fast charging stations is high. An alternative solution is to use HT, which is the combination of a power electronic converter in the input/output of the LFTs (Figure 4c). Thanks to the power converter installation, the integration of energy storage in the HT DC link is possible. The converter power rating is usually between 5–20%, which gives controllability to some extent. However, HT is also bulky and weighty. Another alternative solution is to utilize SST, which can provide full control freedom in the input/output voltage/current such as reactive power control, voltage regulation, and harmonic control. Furthermore, the integration of energy storage and renewable energies is very simple. The SST has advantages including a reduced overall cost and compact size, which are explained in detail in the next section. However, due to the switching and conduction losses in SSTs, their efficiency is lower than the LFTs [72]. The research is still open to improve the SST efficiency specifically in high-power applications.

3.2. Power Stage 2

The second power stage in the DC ultra-fast charging station is AC to DC conversion, which is responsible for the PFC [73]. In this power stage, the converter rectifies the three-phase input AC voltage to a fixed output voltage in the DC link provided that the power quality is desirable on the grid side. Different kinds of power converters are used for this power stage. The AFE rectifier, which is also called an active pulse width modulated (PWM) rectifier, is one of the most used three-phase power converter for this power stage thanks to its high reliability and simplicity. It is used in boost [74], buck [75], and buck–boost [76] modes of operation. Other used simple topologies that can keep the system modularity are half bridge (HB), full bridge (FB), and four-switch buck–boost converter [77] separately used in each phase. Other proposed topologies are multilevel [78], such as the cascaded H-bridge (CHB) converter, modular multilevel converter (MMC), three-level boost (TLB) converter, three-level T-type (TLT) rectifier, multi-cell boost (MCB) converter, neutral point clamped (NPC) converter [52], multi-pulse AC–DC converter [79], and Vienna rectifier (VR) [80,81]. A comprehensive state-of-the-art of AC–DC converters can be found in [82]. The L, LC, and LCL input filters are normally added to the converter input side to desirably satisfy the grid-code requirements as well as to reduce electromagnetic interference (EMI). In both modes of bidirectional AC–DC converter operation, PFC and IEEE-519 grid requirements should be satisfied.

3.3. Power Stage 3

The last power stage in the DC ultra-fast charging station is DC–DC conversion with multi-functional applications, such as stepping-down the input voltage to the battery voltage level, galvanic isolation to separate the EV from the grid, and cooperation with BMS to satisfy battery optimal charging [83]. Similar to power stage 2, the used converters in this stage can be unidirectional and bidirectional. From another point of view, they can be galvanically isolated or nonisolated DC–DC converters. On the one hand, the EV’s battery is not grounded; therefore, its direct connection with the isolated DC–DC converter secondary side will isolate it from the grid [66]. As a result, its protection scheme can work separately from the grid. On the other hand, if the isolation is performed by LFT, nonisolated DC–DC converters could be a promising solution with less complexity to operate in bidirectional power flow mode. Examples of unidirectional isolated DC–DC converters are the phase-shifted full-bridge (PSFB) converter [84], interleaved PSFB converter [85], LLC resonant converter [86], and buck–boost three-level semi-dual-bridge [87]. Examples of bidirectional isolated DC–DC converters are the dual active bridge (DAB) [88], dual half bridge (DHB) [89,90], LLC series-resonant converter (LLC-SRC) [91], quad active bridge (QAB) [92], DAB-based active NPC (DAB-ANPC) [93], and CLLC converter [94]. A comprehensive review of isolated DC–DC converters is reported in [95]. Since the battery voltage level is normally lower than the DC link voltage of AC–DC converters, nonisolated DC–DC converters can be categorized as a buck converter [96], interleaved buck converter [97], unidirectional TLB converter [98], bidirectional TLB converter [99], and three-level flying capacitor (FC) converter [100,101].

3.4. DC Ultra-Fast Charging Station Challenges and Research Gaps

In the vast adaptation of DC ultra-fast charging stations in the power grid, the important challenges are the increase in the daily peak load and shift, which may cause transformer and feeder overload, the power system devices’ deterioration due to aging, and the increase in power losses [102]. Therefore, investigating the impact of high-power DC ultra-fast charging station load profiles in terms of power grid stability and proposing novel optimal charging strategies for the EV batteries are necessary. Moreover, from the power system operation point of view, the DC ultra-fast charging stations should be a controllable and predictable load. For that, different solutions have been proposed [103]. One is integrating energy storage and renewable energies illustrated in the schematic shown in Figure 4 with either AC-coupled or DC-coupled approaches proposed in [66]. Indeed, such DC ultra-fast charging stations can be considered as AC/DC/Hybrid MGs with all their challenges [104,105,106]. In order to increase the charging station third-party interests, the optimal sizing of the energy storage system is performed in [107] through the station energy and storage cost reduction.

Due to the extensive use of power electronic converters, the main grid suffers hugely from power quality problems. Satisfying grid-code requirements could be another challenge in the high-power application for both flows of power. Proposing new converter topologies for the AC–DC and DC–DC power conversion stages with advanced control methods could solve this problem. In addition, it could add some advantages to the DC ultra-fast charging station, including modularity, bidirectional power flow capability, and higher efficiency in high-power applications. The optimum design of an input filter to reduce the size of the system and EMI effect is another important challenge. It should be noted that the different power converter topologies with various control methods could have different small-signal behaviors; therefore, the transient performance analysis of DC ultra-fast charging stations could be another open research area and needs more attraction.

In terms of DC-coupled ultra-fast charging stations, the safety issues need new definitions and structures [108]. Furthermore, in the range of high-power transfer, the power loss during EV charging and discharging would increase due to the increase in the current value. Therefore, the measurement of power losses needs more attention [109]. The challenges in the DC ultra-fast charging station are summarized in Figure 6.

4. SST

Traditional LFTs have been extensively used in power systems to step-down or step-up the voltage level. However, due to the low operation frequency, their weight and size are huge [110]. Therefore, their installation cost and insulation techniques are expensive. Moreover, the control freedom in the input/output voltage/current is limited, and it can only be done by mechanical tap changers [111]. These barriers force the researchers to think about an alternative solution for the voltage level changes in power systems [112]. Thanks to the advancement in power electronic switches based on wide-bandgap materials such as SiC and GaN for high-power high-voltage applications, the SST, which is also known as a power electronic transformer, was first introduced in [113] and solves all the issues of LFTs [114]. In the SST, voltage-level changes are performed by means of a medium-to-high frequency transformer, which significantly reduces its size and weight. This is the reason why SSTs are vastly adapted in railway applications for power electronic traction transformers [115] as well as in shipboard power systems [116]. In addition, thanks to power electronic switches in the input/output of SST, capabilities of power flow control, voltage sag/swell compensation [117], fault current limitation, and satisfactory grid-code requirement are added to the system [118]. In addition, the distribution feeders have less power loss in SST-based power systems [18]. However, these options add complexity to the SST in terms of control, reliability, and protection, requiring more careful design than LFT to guarantee their smooth and stable operation in the power systems. Moreover, the standard efficiency of distribution transformers should be higher than 97% according to the U.S. Department of Energy [119], which is relatively high for the converter-based SSTs [112].

4.1. SSTs Classification

SST topologies can be classified from different points of view [120]. The first one is based on the power conversion stages. Since SSTs work at medium-to-high frequency, a power conversion stage is needed to increase the SST input voltage frequency. Therefore, the simplest configuration of an SST consists of an isolated AC–AC conversion stage, stepping-down from HVAC to LVAC, with high-frequency transformer as shown in Figure 7a [121]. This SST topology is called single-stage or type A. The advantages of this configuration are high-efficiency, simple control, high power density, and high reliability. However, the lack of a DC-link and the huge size of the input/output passive filters to reduce the switching ripple are some of its disadvantages. Different types of AC–AC power converters have been used in single-stage SST, such as DAB converters [122], FB converters [17], flyback converters [123], matrix converters [124], and Dyna-C [125]. However, this suffers from a lack of PFC and active filtering in the input side. In order to have a bidirectional power flow in the AC–AC DAB converter, four-quadrant switch cells are used with the phase-shift modulation technique [126]. The second possible topology for the SST is shown in Figure 7b and is called two-stage or type B. In the secondary side of the high frequency transformer (HFT), an AC–DC conversion stage is used to provide the low-voltage DC (LVDC). The AC–DC isolated boost and the AC–DC DAB are examples of power converters used in type B. The disadvantages are the different control methods in each direction of power flow as well as the non-existence of an HVDC link, which can yield a larger second-order frequency ripple in the LVDC link [127]. Similar to this topology, the two-stage or type C configuration illustrated in Figure 7c provides a high-voltage DC (HVDC) in the primary side of the HFT. The AC–DC conversion could be performed through a DAB converter for a wide range of load change. The integration of large renewable energy power plants is available in the HVDC link. The last possible topology is a three-stage conversion or type D with high-frequency isolation, illustrated in Figure 7d. In this topology, both the LVDC link and HVDC link exist, which provides more control freedom than in other topologies [128]. The integration of small and large sources of renewable energies in both DC links is easier. This is the reason why this topology has attracted more attraction in the industry. It is proposed in several papers that the type D topology with the DAB converter has the best performance and controllability [129]. A comprehensive review of high-frequency DC–DC converters in high-voltage applications can be found in [130,131]. They can be used in the DC–DC section of the type D SST topology. The disadvantage of type D topology is its lower efficiency than in other configurations due to the use of more power conversion stages. A detailed classification of SST topologies can be found in [132].

From another point of view, SST topologies can be classified based on the connection of power electronic converters in the input/output. The maximum blocking voltage of power electronic switches in previous generations was 6.5 kV. In order to connect the aforementioned possible SST topologies to the MV grids, series or parallel connections in the input/output are proposed for high-voltage and high-power applications. Four possible solutions are input series output parallel (ISOP), input parallel output series (IPOS), input parallel output parallel (IPOP), and input series output series (ISOS), in which all possible topologies from a power conversion stages point of view could be used. Furthermore, these kinds of topologies add a modularity feature to the SST, leading to higher reliability but also to control complexity. Meanwhile, thermal and insulation design in the MV side is a challenging task [133]; therefore, simple configuration with high-voltage high-power devices is more preferable, such as a two-level bridge. However, due to the poor harmonic performance, a huge passive filter is needed. As a result, the use of multi-level converters has more advantages in the MV side such as modularity, a smaller passive filter, higher reliability, and fault-tolerant capability. Examples of multi-level converters used in the MV side of SST are NPC [134], FC [135], CHB [136], and MMC [137,138]. Figure 8 shows the different possible topologies of SST according to the input/output connection of power electronic converters. According to the aforementioned SST classifications and existence of different types of power electronic converters, a wide variety of SST topologies with various power converters can be achieved.

SSTs can also be classified based on their winding configuration. The SST windings can be the same as single-phase [139], three-phase [140], single-phase multi-winding [141], three-phase multi-winding [142], split winding [132], and Scott-T transformers [143]. In [144], the HF transformers are replaced with loosely coupled inductive power transfer (IPT) coils leading to simple high-voltage insulation, low parasitic capacitance, reduction in the number of output power electronic converters, easy packaging, and scalability although lower efficiency. There are still other classification of SSTs based on voltage level, control of isolation stage, and modularity structure, which are reported in [20].

4.2. SST Power Devices

Power devices in SSTs can be classified as high-voltage switches in the primary side, high-frequency transformer (magnetic and winding structure), and low-voltage switches in the secondary side. The selection of these devices is related to the used power converter topology, voltage, and power levels [145]. Furthermore, the switching and transformer losses, which are directly related to the switching frequency, play important roles in the efficiency, power density, lifetime, and junction temperature [146,147]. Therefore, it is important to optimally select the SST power devices in each specific application leading to a more efficient SST with reduced size and cost.

There are three types of power losses in semiconductor devices including conduction losses, switching losses, and blocking losses [148]. Among them, the blocking losses can be neglected [149]. Since the semiconductor switches are not ideal, there is a power loss during the switch conduction period which is called the conduction loss. Therefore, semiconductor switches practically have on-state resistance or an along-voltage drop [150]. The conduction v-i characteristics of switches and anti-parallel free-wheeling diodes can normally be found on the switches’ datasheets, which are dependent on the operating temperature. Therefore, the different thermal models and thermal insulations for the switches lead to various power losses, which could be exclusive in each application. The voltage and current in the semiconductor switches cannot immediately change from zero to maximum or vice versa. Therefore, during turn-on and turn-off switching events, the voltage and current intersect each other, yielding switching power losses [151]. In order to reduce switching power losses, the most usable method is to manipulate voltage and current waveform during turn-on/turn-off transition through the gate driver circuit and the external circuit. An alternative solution is to use soft switching techniques in which some passive components such as an extra capacitor, inductor, and diode are added to the circuit. Zero voltage switching (ZVS) and zero current switching (ZCS) are examples of soft switching techniques [152]. It is shown in [153] that the modulation method of a PWM converter has a great impact on its conduction and switching losses. Therefore, proposing different modulation methods can yield reductions in conduction and switching losses. Device manufacturers are still optimizing switches to reduce the switching losses.

4.2.1. High-Voltage Side Switches

In the high-voltage high-power application, the most usable semiconductor device is Si IGBT—for instance, the 6.5 kV–25 A Si IGBT. However, the MOSFET has better switching characteristics in comparison with the IGBT [154]. The parallel connection of IGBT and MOSFET is proposed in [155] and shows higher efficiency and more desirable switching characteristics. A new generation of wide-bandgap semiconductor devices such as SiC MOSFET and SiC IGBT has emerged recently. They have a higher blocking voltage, less switching loss, high thermal conductivity, and high operation frequency [156]. Therefore, the SiC-based power electronic converters could be more efficient and could work in high operation temperatures [21]. The disadvantages of these switches are their high cost and non-commercialization. In addition, there is an oscillatory behavior in the SiC-based semiconductor devices during the turn-off switching period, which can be damped through a parallel connection of Si IGBT and SiC MOSFET [157].

In the high-voltage high-power application, another problem of power electronic switches is their high output voltage slopes (dv/dt) due to the fast switching [158]. This phenomenon can cause voltage amplifications, which may shorten the transformer winding lifetime [159]. Furthermore, the gate-driver EMI is a result of high dv/dt [160,161]. In the high-voltage side, maybe a series connection of switches should be used, which can worsen the dv/dt issue [162]. Different solutions have been proposed to solve the dv/dt issue. The most known one is to use a passive filter in the input/output [163]. The optimal selection of CHB cells in the MV side is another alternative solution to make a trade-off among efficiency, power density, and dv/dt issue [164,165]. Another interesting solution is proposed in [166], where a new topology (S4T) was suggested for the soft-switching of three-phase SST, which has higher efficiency, a lower dv/dt rate, higher power density, longer lifetime, and less EMI.

Since the primary side of an SST is connected to the MV grid, the power electronic switches should be high-voltage low-current. The new wide-bandgap semiconductor devices such as the 10 kV–10 A SiC MOSFET, 15 kV–10 A SiC MOSFET [167], and 15–20 kV A SiC IGBTs could be promising solutions in the SST primary side since a simpler structure can be used in the MV grid [168]. The blocking voltage of the switches is increasing—as an example, a 4H-SiC n-IGBTs with ratings of 27 kV–20 A is built in [169]. However, in a simple structure such as two-level converters, the dv/dt issue and switches’ stress are worse than in multi-level converters, reaching up to 100 kV/$\mathsf{\mu }$s. Therefore, a careful design of modules, gate drivers, busbars, and passive filters for switches is necessary at such a stress level [170].

4.2.2. High-Frequency Transformer

The second power device of the SST is the HFT, which should satisfy the requirements of high-voltage, high-power, and high-frequency operation. The HFT consists of a magnet core, primary winding, and secondary winding. A comprehensive explanation of HFT design is presented in [171]. The magnetic material and its geometry has a direct impact on the transformer power density and losses [172]. Different kinds of magnetic material such as powdered iron, silicon steel, ferrite, amorphous, and noncrystalline have been used in the transformers, with the last three being the best for the HFT [173]. They can be compared with each other in terms of core losses, saturation flux density, operating temperature, and relative permeability [174]. In terms of core geometry, different kinds, such as core type, shell type, matrix type, and co-axial winding type, have been used in the past [175]. These parameters have conflicting objectives; therefore, the optimal design of HFT magnetic material, geometry, and frequency needs a trade-off among all of them, and this would be a multi-objective optimization problem [176,177].

At high frequencies, the efficiency of HFTs also depends on the material and configuration of the primary and the secondary winding because of skin and proximity effects [178]. Copper and aluminum are the common LFT winding materials. The transformer short circuit current has an impact on melting points, which plays a critical role in the selection of winding material [179]. In addition, due to the skin and proximity effects, the circular-shaped conductors are not a good choice for the HFT. Therefore, Litz wire, foil conductors, and primary and secondary interleaving have been used for the HFT [180], in which factors such as window dimensions, window utilization, window fill factor, and the transformer turn ratio should be optimally designed [181].

In the HFT, the cooling material cannot be oil-like LFTs. Moreover, the insulation material should tolerate partial discharges to provide a good insulation. Examples of insulation materials are Nomex paper and Mica. Therefore, thermal and electrical insulation design is a challenging task in high-voltage high-power application in the compact SST due to the excess of power losses [182].

4.2.3. Low-Voltage Side Switches

The last power device of an SST is the secondary side low-voltage switch. Different types of these switches are commercially available and are analyzed comprehensively in a wide range of applications.

4.3. SST Applications

SST has been used in different kinds of applications [183] such as wind energy [184,185], locomotives, and traction systems [186,187], interfacing asynchronous grids and loads [188], distribution networks [112], smart grid development [143,189], house power systems [190], and EV charging stations [191]. Details of some of the implemented SSTs are shown in

Table 4. Experimentally implemented SSTs in the literature for different applications.

Application	SST Type	Power Rating (KVA)	MV (kV)	AC-DC Rectifier	DC-DC Converter	SST Efficiency	DC-link (HVDC LVDC)	High-Voltage Switches
Distribution network [192]	D	1.67	1.15	FB (75 kHz)	Switched-capacitor LLC-SRC (75 kHz)	93–97%	1.7 kV 600 V	3 kV–12 A IGBT/1.2 kV–20 A SiC MOSFET
Fast charging [193]	D	4	0.22	CHB (1.05 kHz)	QAB (10 kHz)	N/A	0.36 kV 220, 120, 48 V	N/A
Smart grid [194]	D	10	3.6	CHB (1.2 kHz)	DAB (3.6 kHz)	84–92%	5.7 kV 200 V	6.5 kV–25 A Si IGBT
Distribution network [195]	D	10	3.6	FB (6 kHz)	DHB (15 kHz)	N/A	6.1 kV 400 V	13 kV–10 A SiC MOSFET/JBS diode
Distribution network [196]	D	10	13.2	3-level PFC (20 kHz)	LLC (40 kHz)	N/A	25 kV 500 V	1.7 kV–5 A SiC MOSFET
Fast Charging [197]	D	16	3.8	TLB (20 kHz)	HB-LLC (98 kHz)	98%	6.2 kV 400 V	1.2 kV–31 A SiC MOSFET
Wind energy [198]	D	20	7.2	CHB (1.08 kHz)	DAB (3 kHz)	88–95%	11.4 kV 400 V	6.5 kV–25 A IGBT
Utility network [199]	D	25	7.2/8	TLB (93 kHz)	HB-LLC (93 kHz)	97.5%	14 kV 400 V	1.2 kV–36 A SiC MOSFET
Data centers [200]	D	25	3.8	FB (48 kHz)	LLC-SRC (48 kHz)	98–99.6%	7 kV 400 V	10 kV–10 A SiC MOSFET
Fast charging [201]	D	25	2.4	TLB (25 kHz)	NPC + diode rectifier (50 kHz)	96.6%	4.8 kV 400 V	1.2 kV–55 A SiC MOSFET
Mobile utility support [202]	D	100	4.16	AFE (20 kHz)	DAB (10 kHz)	N/A	7.2 kV 800 V	10 kV–90 A Gen3 SiC MOSFET
Fast charging [203]	D	400	13.2/4.8	NPC (5 kHz)	LLC-SRC (100 kHz)	96.5%	6.4 kV 1000 V	1.2 kV–50 A SiC MOSFET

. Moreover, different converters, input/output configurations, and topologies have been used for each power stage of type D SSTs, such as AC–DC and DC–DC power stages. As an example, by considering 31 kinds of three-phase AC–DC converters in [52] and 8 types of isolated DC–DC converters in [131], a total number of 248 SST type D topologies could be implemented. It should be noted that each converter type and configuration could have different numbers of switches with different blocking voltage levels, diodes, and passive elements, which have a direct impact on the cost, size, and reliability of the SST. Therefore, topology, switches, and passive elements selections could be an optimization problem in the SST design procedure. It can be perceivable from the Table 4 that in the high-voltage high-power application, there is still a gap between LFT efficiency (higher than 99.5%) and SST efficiency (96.5%) that needs more attention.

4.4. SST Transient Performance

In the future renewable electric energy and management (FREEDM) distribution network, the SST will play a critical role as an energy management unit or energy router. In this network, renewable energy sources and energy storage integration are possible with plug-and-play functionality. Control freedom, intelligence, and possible communication features of SSTs in the future smart grids can facilitate the high penetration of intermittent renewable energies, leading to improvements in the system’s reliability. Therefore, a small-signal analysis is required to check the operation of SSTs during sudden load/generation changes. In order to conduct a small-signal analysis, the state-space model of the system is needed. Since the SST topology could be different due to the use of various types of power converters, there exist a wide variety of SST state-space models. The simplified state-space average model of an SST type D is presented in [204], in which the AC–DC and DC–DC stages are performed through CHB and DAB converters, respectively. The time-scales of SST different modes are within ${10}^{-3}$ ms and 0.1 s [205]. Moreover, a comprehensive mathematical model of an SST including power stages and closed loop controllers is reported in [206]. This model can also be used for the general power system studies and implementation in real-time digital simulators [207]. A 70th-order simplified state-space average model of a FREEDM network including single SST, energy storage, DC loads and sources, and AC loads and sources is derived in [205]. Therefore, in the distribution networks in which the LFTs are replaced by SSTs, interactive dynamics may appear among SSTs. As an example, second-order harmonic oscillation can appear in the AC side of the AC–DC power stage leading to DC-link voltage variation, harmonic resonance among SSTs, and system instability [208]. Therefore, an optimal design of SST controllers based on eigenvalue analysis can improve the transient performance of the whole network.

The state-space average model of SST can also be used in the protection studies to anticipate the SST performance under fault conditions. Through the state-space model of SST, over-voltage and over-current protection schemes and voltage ride through capabilities can be proposed [209].

4.5. SST Challenges and Research Gaps

There exist a great number of challenges in the application of SSTs. The first one is the design of an SST high-frequency core and windings, in which material advancement is still ongoing. A proper thermal design of the SST core and windings can improve its efficiency, which is currently lower than that of LFT in high-power applications.

In the vast adaptation of SSTs in power distribution networks, feasible and dynamic stability analysis, power system analysis, and reliability concern sdue to the existence of power electronic switches are significantly important factors. In SST-based power distribution networks, extra communication and control layers are necessary. Moreover, it should be noted that, since the conventional mechanical MV circuit breakers are too slow, a fast-acting protection scheme is needed for the protection of MV power electronics switches during fault conditions. Therefore, the implementation of fast circuit breakers with solid-state materials could be another interesting research area in this field.

Another challenge is the unbalance problem, which can appear in the modular configuration of SST between the phases. Short-circuit faults and uneven distribution of loads can cause unbalanced operation in the power system. Furthermore, the three-phase grid voltages are not always balanced. In unbalanced operation mode, a negative sequence current will be drawn or injected from or to the grid. These may cause double-frequency ripples and unbalanced voltages in the SST DC-links of each phase, which can damage power electronic components due to the increase in the semiconductor device stress, over-voltage, and over-current [210]. Therefore, SSTs should properly work in unbalanced three-phase grid voltages and have the compensation capability of the three-phase grid currents in a cost-effective and reliable manner [211]. The unbalanced operation of SSTs is investigated in [210,212]. The most widely used unbalance compensation technique in the literature is zero sequence voltage injection [213], which leads to the choice of power devices with a higher rating and power losses [214]. The challenges in the application of SSTs still under research are illustrated in Figure 9.

5. DC Ultra-Fast Charging Station with SST

An interesting solution for power stage 1, voltage level changes, in the DC ultra-fast charging station shown in Figure 4 is to utilize SST instead of LFT, which can perform rectification, voltage step-down, and the isolation function. Thanks to the high-frequency operation of SSTs, their size and weight would be much lower than conventional service transformers [215]. Therefore, the DC ultra-fast charging station installation cost and footprint would hugely decrease [7]. Furthermore, the employed power electronics switches in the input and output of SSTs give full controllability of the input/output voltage/current, which can provide reactive power control, satisfactory grid-code requirement, grid ancillary services, and higher efficiency than the conventional charging stations [216]. Moreover, the presence of DC-links in the three-stage SST architecture could facilitate the integration of renewable energy sources and energy storage, as well as possibly decreasing power conversion stages, increasing the efficiency [217]. This integration possibility in the DC-links could maximize interest for the owners of DC ultra-fast charging stations and could provide ancillary services for the grid [218]. Another advantage of the DC SST-based ultra-fast charging station is its capability for bidirectional power flow, which could provide ancillary services for the grid, shave the grid peak load, and facilitate the V2G, V2H, and V2V applications [219]. Furthermore, thanks to the modularity of SSTs in the input side, any increase in the DC SST-based ultra-fast charging station power capacity could be easily performed, while the replacement of LFT is necessary in the DC LFT-based ultra-fast charging station. In order to implement a DC SST-based ultra-fast charging station, different aspects should be investigated, as explained below.

5.1. Topology Configuration

As mentioned previously, there are four types of SST configurations that can be used in the DC ultra-fast charging stations. Among them, type D (Figure 7d) is the suitable choice thanks to its higher control freedom. The choice of power converters for the AC–DC and DC–DC power stages could play a vital role in the performance of DC SST-based ultra-fast charging stations. The single-phase and three-phase AC–DC power converters can be used in the first power stage such as HB, FB, AFE, three-level PFC, NPC, TLB, TLT, MCB, VR, CHB, MMC, four-wire bidirectional boost converter, four-legged bidirectional boost converter, matrix-converter bidirectional buck–boost, and so on. In isolated DC-DC power stage, power converters such as DAB, series-resonant DAB, PSFB, LLC-SRC Full-bridge, hybrid switched-capacitor LLC-SRC, DHB, QAB, NPC, DAB-ANPC, FC, CLLC, and any other novel topology can be utilized. Some of the possible power electronic converters that could be used in a DC SST-based ultra-fast charging station are shown in Figure 10.

Furthermore, the SST can be used in ISOP, ISOS, IPOP, and IPOS (Figure 8) configurations, which adds modularity to the whole circuit [220], facilitating its use in high-voltage high-power application and reducing the size of input passive filters. From a topological point of view, SSTs in the DC ultra-fast charging stations could be categorized as single-phase single-port (SPSP) [221], single-phase multi-port (SPMP), three-phase single-port with single-phase converters (TPSP-SPC) identically connected in delta or wye [222], three-phase multi-port with single-phase converters (TPMP-SPC) identically connected in delta or wye [193,223], and three-phase single-port with three-phase converter (TPSP-TPC) or three-phase multi-port with three-phase converter (TPMP-TPC) either connected in delta or wye configurations. Therefore, the optimal selection of the DC SST-based ultra-fast charging station configuration needs more attention. The aforementioned topology architectures are illustrated in Figure 11.

5.2. Transient Performance

A DC ultra-fast charging station could be similar to a FREEDM network with a single SST. Therefore, the state-space average model of the DC ultra-fast charging station is of a high order. This model should be considered for the dynamic stability and small-signal analysis in order to guarantee stable and feasible operation in the charging station. It should be noted that different configurations and topologies yield a wide variety of state-space average models.

Since the ultra-fast charging stations could have more than one charge ports for the EVs, immediate disturbances or load changes can take place specifically during the fast-charging mode of operation for the EV batteries. Therefore, the transient performance of the charging station should be deeply investigated. Methods should be proposed for the improvement of the transient performance through passive (adding passive elements to the circuit) or active (manipulating control loops or using different control strategies) techniques [224]. The dynamic stability issue could be worsen for the distribution network with the dense penetration of DC ultra-fast charging stations [225]. Smart and dynamic charging techniques could solve this problem. In addition, the smart charging techniques could have advantages such as optimization of the EV charging price in variable grid prices, grid load peak shaving, grid ancillary services, and lower reserve generation capacity [226].

5.3. Energy Management and Optimal Sizing

The size of a DC ultra-fast charging station is related to the urban traffic flow based on factors such as the road system, location, direction, intersections, and length [227]. In the DC SST-based ultra-fast charging station equipped with distribution generations and energy storage, an intelligent energy management system (IEMS) is required to increase the interest of the charging station’s owner. The IEMS can decide between the charging/discharging schedule of energy storage [228,229] and participation in grid ancillary services [216]. In addition, the required charging power of a DC ultra-fast charging station is uncertain due to different factors such as the arrival time and SOC of the EVs [230,231]. Therefore, the load profile of a DC ultra-fast charging station is not a constant power load from a distribution network point of view. In order to predict the charging station’s load profile, curve fitting through standard exponential load model [232], conventional optimization [233], and machine learning techniques have been used in the literature [234]. By means of these methods, the optimal sizing of the resources in DC ultra-fast charging stations is achievable [14], thereby leading to maximizing third-party interest [235].

5.4. Power Imbalance Problems

In order to keep the modularity of DC SST-based ultra-fast charging stations with three-phase AC grid voltage, power converters can be separately used in each phase connected in delta or wye. In these kinds of charging stations, each phase can have different power consumption according to the EV demand. In addition, there exist parameter mismatches in the real implementation of HFT (isolation stage) in each modules [236]. Therefore, a power imbalance can occur in the station leading to DC-link voltage deviation, drawing unbalanced currents from the main grid, and increasing the total harmonic distortion (THD) of the grid current. To solve this issue, the zero-sequence current is added to the reference currents of each phase [237]. However, this increases the magnitude of the total current [238]. Therefore, semiconductor devices with a higher rating current should be used. To mitigate this problem, in the DC–DC power stage of type D SST, power balance windings are added to the HFT and connected in parallel for all phases to reduce the zero-sequence required current [193].

5.5. Control Methods

In a DC ultra-fast charging station with type D SST, different control objectives can spread among AC–DC and DC–DC power converters. The objectives of the AC–DC power converter are to keep the HVDC link voltage fixed to a reference value, to keep the grid synchronization like a phase lock loop (PLL), to keep the injected grid current sinusoidal with a THD lower than 5% [239], to keep the injected three-phase grid current balanced, to keep robustness against grid voltage harmonics, to compensate grid voltage sag/swell, to limit the fault current or start-up current [240], to achieve a bidirectional power flow, and to keep unitary power factor in the grid-side or other power factors for injecting/consuming reactive power to/from the grid, providing ancillary services [241]. In multi-level structure at the input side of SST, the voltage balance among all the capacitors would be another control objective. In the modular structure of SST implemented separately in each phase with three-phase input, power balancing is another important objective in order to improve the system performance. In this structure, if energy storage systems are integrated with DC-links, making a balance in their SOC rates could be an interesting performance criteria.

The objectives of the DC–DC power converter are to keep the LVDC link voltage fixed to a reference value, to achieve ZVS/ZCS leading to lower stress on switches, to guarantee lower EMI and higher system efficiency, to provide galvanic isolation against the high-voltage side, and to give a bidirectional power flow through the leading/lagging of phase angle between the primary and secondary bridge.

Table 5. The desired performance of a SST-based DC ultra-fast charging station.

Power Stage	Performance Criteria
AC-DC power stage	• HVDC link voltage control
	• LVDC link voltage control
	• Grid current THD improvement
	• Grid synchronization
	• PFC
	• ZVS/ZCS
	• Unidirectional power flow
	• Bidirectional power flow
	• Power balancing control
	• Voltage balancing control
	• SOC balancing control
	• Grid ancillary services
	• Fault tolerant and limitation
DC-DC power stage	• LVDC link voltage control
	• ZVS/ZCS
	• Galvanic isolation
	• Unidirectional power flow
	• Bidirectional power flow
	• Power balancing control

shows the desired performances that the control approach has to achieve in DC SST-based ultra-fast charging stations.

5.5.1. AC–DC Power Converter Control Approaches

In order to control the single-phase and three-phase AC–DC power converter, different control strategies have been used in the literature. The first popular one in three-phase converters is the voltage oriented control (VOC), which performs the indirect control of active and reactive power through the current vector orientation [242]. The second popular control approach is the direct power control (DPC), which performs instantaneous active and reactive power control. Another popular control approach is the virtual flux (VF) technique [243], which mimics the control of electric motor drive systems. The proportional integral (PI) controller, proportional integral derivative (PID) controller, proportional resonant (PR) controller, sliding-mode controller (SMC), hysteresis controller, fuzzy logic controller (FLC), model predictive controller (MPC), adaptive controller, and neural-network-based controller are employed to satisfy the desired performance [52,82].

5.5.2. DC–DC Power Converter Control Approaches

Various isolated bidirectional DC–DC power converters have been proposed, such as resonant converters, dual flyback, dual-Cuk, dual push–pull, and DAB [72]. Among them, the DAB converter is the most usable one in the DC–DC power stage of DC SST-based ultra-fast charging stations. The advantages of the DAB converter can be listed as bidirectional power flow capability and fast power flow mode changing, wide voltage conversion range to interface different voltage levels, and ZVS capability to increase the efficiency. The feedback control, linearization control (PI/PID), feedforward plus feedback, disturbance-observer-based control, feedforward current control, MPC, SMC, and moving discretized control set MPC have been employed in the DAB converter to achieve control objectives. Different modulation schemes have been used for the DAB converter such as single-phase-shift (SPS), dual-phase-shift (DPS), triple-phase-shift (TPS) [244], and extended-phase-shift (EPS) [245]. The DAB converter transfer power P with SPS modulation scheme is defined as follows [246]:

$$P=\frac{N{v}_p{v}_s}{f_{s}L}\varphi (1-2|\varphi \left|\right)$$

(1)

where N, $v_p$, $v_s$, $\varphi$, $f_s$, and L are the DAB turn ration, DC input voltage, DC output voltage, phase shift, switching frequency, and inductance, respectively. The DAB reduced-order model is a first-order transfer function and can be obtained by neglecting inductor current dynamics. This transfer function $G\left(s\right)$ from $\varphi$ to $v_s$ can be expressed as follows:

$$G\left(s\right)=\frac{\widehat{v_s}}{\widehat{\varphi }}=\frac{N{V}_p(1-4\Phi )}{f_{s}L}\frac{R_L}{R_L{C}_{2}s+1}$$

(2)

where $V_p$ and $\Phi$ are the DC values of $v_p$ and $\varphi$. $R_L$ and $C_2$ are the resistance of output load and output capacitance respectively.

5.5.3. Analysis of Existing Control Approaches

Table 6. Literature control approaches of DC SST-based ultra-fast charging station.

References	AC–DC Power Converter	AC–DC Control Approach	DC–DC Power Converter	DC–DC Control Approach
[193]	CHB	Cascaded PI control: • PFC • Grid synchronization • Power balancing control • Voltage balancing control • HVDC link voltage control	QAB	PI control: • DPS bidirectional power flow • Galvanic isolation • Power balancing control • LVDC link voltage control
[197]	TLB	Digital PI control: • PFC • Grid synchronization • Voltage balancing control • HVDC link voltage control • Grid current THD improvement	HB-LLC	Digital open-loop control: • Unidirectional power flow • Galvanic isolation • LVDC link voltage control • ZVS
[223]	CHB	Cascaded PI control: • PFC • Grid synchronization • Power balancing control • Voltage balancing control • SOC balancing control • HVDC link voltage control	DHB	Digital PI+PR control: • Bidirectional power flow • Galvanic isolation • LVDC link voltage control • ZVS
[247]	CHB	PI control: • Grid ancillary services • HVDC link voltage control	DAB	PI control: • Bidirectional power flow • Galvanic isolation • LVDC link voltage control • ZVS
[248]	CHB	PI control: • PFC • Grid synchronization • Power balancing control • HVDC link voltage control	DAB	Cascaded PI control: • DPS bidirectional power flow • Galvanic isolation • LVDC link voltage control
[249]	VR	Multi-loop PI and hysteresis current control: • PFC • Grid synchronization • Voltage balancing control • HVDC link voltage control	DAB	Cascaded PI control: • SPS bidirectional power flow • Galvanic isolation • LVDC link voltage control
[250]	TLB	Multi-loop PI and predictive current control: • PFC • Grid synchronization • Voltage balancing control • HVDC link voltage control	DAB-ANPC	PR control: • TPS bidirectional power flow • Galvanic isolation • LVDC link voltage control • ZVS
[251]	TLT	Digital multi-loop PI control: • PFC • Grid synchronization • Voltage balancing control • HVDC link voltage control	N/A	N/A
[252]	CHB	Multi-loop PI and predictive power control: • Fault tolerant and limitation • Voltage balancing control • LVDC link voltage control	DAB	PI control: • SPS bidirectional power flow • Galvanic isolation • ZVS
[253]	MCB	PI control: • Voltage balancing control • HVDC link voltage control	NPC	PI control: • Power flow • Galvanic isolation • LVDC link voltage control

shows the existing control approaches of DC SST-based ultra-fast charging stations. For the control of the DAB converter, combining the LVDC-link voltage control with AC–DC rectifier control is a simple way to reduce the number of control loops, but the DAB output voltage would be sensitive to the load variation. Another method is to control the DAB converter independently through the PI control of LVDC and phase shift magnitude between primary and secondary windings. Grid voltage harmonics can cause ripples in the DC-links and can have adverse effects in conventional VOC for three-phase AC–DC rectifiers [247].

5.6. DC SST-Based Ultra-Fast Charging Station Social Repercussions

The DC SST-based ultra-fast charging station has different impacts on the community. One of the most positive impacts is that it will decrease the anxiety of the EV drivers since they can refuel their EVs in a short time, comparable to the time needed for refueling FFVs. Therefore, the drivers’ concerns about continuing their route on interurban roads would significantly decrease [254]. Moreover, the bidirectional power flow capability and V2G application of DC SST-based ultra-fast charging stations could increase people’s participation in electrical network planning. Therefore, both the society and the government can benefit from this participation, leading to an increase in social welfare [255]. Another impact of the DC SST-based ultra-fast charging station on the community is that the needs of carrying fuel with big tankers, which is a dangerous and time-consuming task, would be removed.

To summarize, the vast adaption of DC ultra-fast charging stations could hugely facilitate reaching a sustainable life in the coming future. As a result, environmental activists and the whole society would be more satisfied [2].

5.7. DC SST-Based Ultra-Fast Charging Station Challenges and Research Gaps

In addition to the previously mentioned challenges, there are still other gaps in the application of DC SST-based ultra-fast charging stations. These include cybersecurity and data protection schemes, which should be addressed in the online control schematic for the future smart grids. Another challenge is the requirement of advanced protection schemes against short circuits, over-voltages, and overloading fault conditions with fast solid-state circuit breakers (current interruption in several hundred microseconds) [256]. The advanced protection scheme can guarantee the safety of EV owners and the personnel of DC SST-based ultra-fast charging stations. Reliability concerns of replacing the passive transformer with a power electronic equivalent could be another research challenge in this field. Furthermore, in the direct connection of a DC ultra-fast charging station to the MV grid, the safety, protection, standardization, and certification of the EV charging equipment are necessary. To summarize this section, the research challenges in the DC SST-based ultra-fast charging station are presented in Figure 12.

6. Conclusions

The electrical architecture of an ultra-fast charging station offers numerous open aspects to investigate. In that context, the low-frequency service transformer could be substituted by an SST with the aim of increasing the control capability and reducing the size and weight. The SST is considered for any conversion structure implemented with switching power converters that is capable of transforming AC MV into DC in the range between 200 and 500 V.

The main features of existing SSTs in renewable energy applications have been reviewed in this paper, considering the rated power, the MV magnitude, the topology of both AC–DC and DC–DC stages, the voltage level in the DC link, the voltage stress supported by the high-voltage switches, and the resulting efficiency.

The penetration of SSTs in ultra-fast charging stations is in an incipient stage, but there are already important antecedents in fast charging. A classification of SST-based electrical architectures for charging stations has been presented in this work considering the type of input, number of output ports, and converter configuration. Four SST configurations have been identified with potential use in ultra-fast charging, with the type D being the most appropriate one. Other aspects to be considered in the design of the SST are the stability of the interconnection with the rest of the blocks in the ultra-fast charging station and the energy management in a hierarchical supervision system that should also handle the control of the different converters solving power imbalance problems.