Topological Advances in Isolated DC–DC Converters: High-Efficiency Design for Renewable Energy Integration

Abstract

The increasing penetration of renewable energy sources (RESs) into medium-voltage (MV) and low-voltage (LV) power systems presents significant challenges in ensuring power grid stability and energy sustainability. Advanced power conversion technologies are essential to mitigate voltage and frequency fluctuations while meeting stringent power quality standards. RES-based generation systems typically employ multistage power electronics to achieve: (i) maximum power point tracking; (ii) galvanic isolation and voltage transformation; (iii) high-quality power injection into the power grid. In this context, this paper provides a comprehensive review of up-to-date isolated DC–DC converter topologies tailored for the integration of RES. As a contribution to support this topic, recent advancements in solid-state transformers (SSTs) are explored, with particular emphasis on the adoption of wide bandgap (WBG) semiconductors technologies, such as silicon carbide (SiC) and gallium nitride (GaN). These devices have revolutionized modern power systems by enabling operation at a higher switching frequency, enhanced efficiency, and increased power density. By consolidating state-of-the-art advancements and identifying technical challenges, this review offers insights into the suitability of power converter topologies in light of future trends, serving as a valuable resource for optimizing grid-connected RES-based sustainable power systems.

1. Introduction

Environmental sustainability is widely recognized as a critical concern across diverse technological domains. Achieving carbon neutrality demands innovative solutions that effectively reduce greenhouse gas emissions while simultaneously catalyzing a paradigm shift within the energy sector, encompassing key areas such as generation, transportation, and industrial processes. As reported by the International Energy Agency, the energy sector is responsible for around 85% of total CO₂ emissions. In 2023, energy-related activities contributed to a record high of 37.7 gigatons of CO₂ [1]. Notably, the production of electricity and heat is identified as a major contributor to the above-mentioned emissions [2].

Over the last decades, partly due to growing environmental awareness, electricity production on the basis of renewable energy sources (RESs) has gained greater preponderance in the global energy mix. As reported in [3], the increase in power generation is supported by a surge in wind and solar production, helping to decarbonize the energy sector. Despite the impossibility of completely canceling the emissions of greenhouse gasses, an affirmative response to the inclusion of RES-based systems would significantly contribute to achieving the objectives proposed in the Paris Agreement, celebrated in 2015 and adopted by 196 countries [4].

In addition to transitioning towards clean energy production, it is equally important to prioritize the advancement of electric mobility [5,6], as well as energy storage, e.g., battery production, charge, and recycling [7]. Increasing the focus on such aspects significantly contributes to improving environmental quality and serves as the foundation for an inevitable and anticipated energy transition.

The field of power electronics plays an important and facilitating role in the integration of energy storage systems ESSs, electric vehicles (EVs), and RES-based systems into the utility grid. The growing propagation of such emerging technologies, most of them natively operating in DC, translates, on the other hand, into a series of challenges for the utility grid since, in the actual mold, its infrastructure and organization do not meet the expected premises [8,9]. Future smart grids settle on the principle of power bidirectionality between energy producers and consumers, which does not occur nowadays. It is vital to implement and develop more intelligent, reliable, and sustainable power grid architectures, abandoning centralized structures and migrating to decentralized ideologies [10].

In view of the increasing spread of emerging power devices, the different types of power quality problems have also increased. In the case of strong integration of RES-based energy generation systems, the main concerns relate to harmonic content and voltage and frequency fluctuations [11]. As expected, considering that the energy production profile is dependent on weather conditions, the stronger the integration of RESs, the greater the variability. Moreover, such a scenario also provokes repercussions on the interface between DC devices and AC distribution lines since voltage transients may occur, the vast majority of which are critical to the operation of the system as a whole.

In line with these challenges, coupled with the increasing difficulty of controlling voltage fluctuations at the point of common coupling (PCC), it is crucial to adopt new fault suppression mechanisms, strategies aimed at enhancing power quality and, fundamentally, suitable power electronics systems tailored to each scenario and RES [12]. While wind and hydropower generation have already reached an advanced stage of technological maturity, solar photovoltaic (PV) energy stands out as the most promising path for the development of new power electronics devices [13]. However, these challenges and corresponding solutions extend to various RESs and should be considered for developing interfacing systems with higher efficiency, flexibility, and added functionalities. Recent research has been conducted to address these needs, focusing on the exploration of new architectures, paradigms, power converter topologies, and modulation techniques.

Besides the conventional functionalities of galvanic isolation and maximum power point tracking (MPPT) control algorithms, innovative RES interfacing systems must also efficiently manage the energy needs of producers and consumers. Among these, in alignment with the concept of the smart grid, the mitigation of power quality problems, fault tolerance mechanisms, and the inclusion of ancillary services, a special feature concerning the dissemination of microgrids in decentralized architectures, are highlighted [14].

In line with the current trends, solid-state transformers (SSTs) stand out for their versatility and flexibility, enabling a galvanically isolated interface between two or more power devices through a high-frequency transformer (HFT) [15]. However, the use of medium-frequency transformers (MFTs) is equally valid, especially in high-power scenarios. Connecting RES-based generation systems to the utility grid is solely one of the several SST applications [16,17]. As with some existing multistage systems, three power stages are considered, each of them correspondingly guaranteeing (i) MPPT; (ii) galvanic isolation and voltage rise; and (iii) power injection into the utility grid. Nonetheless, while acknowledging the importance of each power stage, the intermediate one, traditionally associated with the isolated DC–DC converter, plays a significant role in the system’s operability. Beyond the benefits of incorporating a galvanically isolated converter, e.g., the elimination circulating current paths and improved safety in high-power scenarios, it is pertinent to analyze these features with high-frequency (HF) switching. This approach leads to a significant reduction in the volume and weight of power converters and passive elements, i.e., transformers, inductors, and capacitors.

Given the growing proliferation of RES-based energy generation systems, this review paper presents an overview of the most critical and up-to-date topologies for unidirectional isolated DC–DC power converters. In the literature, some review papers are commonly found about this subject; however, none are as comprehensive as this one [18,19,20,21,22,23]. The main and distinguished contributions of this paper include the following: (i) a comprehensive review of possible architectures and connection schemes for RES-based systems interfacing with the utility grid; (ii) an extended analysis of the most relevant and recent topologies, focusing on practical implementations rather than a broad, theoretical presentation of designs that are no longer suitable for modern applications; (iii) an in-depth study of phase-shift modulation techniques for bidirectional isolated DC–DC power converters, emphasizing their impact on efficiency and dynamic performance; (iv) a comparative discussion of DC–DC topologies, outlining the key strengths of each topology while associating their classification (conventional low-power, bridge-based, and multilevel converters) with their most conventional application cases; (v) an overview of wide-bandgap (WBG) semiconductors and their role in next-generation power converters, highlighting their advantages in terms of switching speed, thermal performance, and efficiency improvements. Furthermore, the study explores future directions in the field, focusing on the control algorithms, modulation techniques, and fault-tolerant mechanisms that enhance converters’ flexibility and resilience.

Thus, this review paper is organized as follows: Section 2 depicts the main architecture schemes for the interfacing system. Section 3 extensively analyzes and describes isolated DC–DC topologies to interface with RES-based generation systems, distinguishing between conventional structures, bridge-based converters, and multilevel topologies. Section 4 provides a comparative study between the previously analyzed DC–DC topologies, highlighting their relevance to the proposed topic. A concise overview of WBG devices is conducted in Section 5, while the main conclusions are outlined in Section 6.

2. Power Electronics Systems for RES-Based Generation Technologies

The integration of RES-based power generation units, which is considered inevitable in future smart grids, should be approached with a focus on favorable economic and social premises [24,25,26]. Simultaneously, from a technical perspective, it is essential to ensure the energy needs of consumers at all times by implementing smoothing mechanisms to address the intermittent generation profile of RESs. While ESSs are often linked with renewables, their widespread adoption is hindered by challenges such as costs and technological maturity [27]. To achieve this objective, the power converters that interface the utility grid and a specific RES will play a significant role in energy management. Similarly to SSTs, these power converters will incorporate added functionalities. Hence, it is essential to select appropriate power electronics solutions, ensuring that the chosen devices are well suited for the specific application requirements.

The main RES-based energy generation elements can be categorized into solar PV and turbine generator technologies, which encompass, e.g., hydro, wind, and geothermal energy. In rotational systems, the speed of the turbine’s central axis can be either fixed or variable. Figure 1 illustrates the four configurations commonly considered: (i) fixed speed; (ii) partial variable speed; (iii) variable speed with a partial-scale frequency converter; and (iv) variable speed with a full-scale power converter. In variable-speed systems, the use of voltage source converters in a back-to-back arrangement enables the incorporation of MPPT algorithms, reactive power compensation, increased robustness, and an enhanced rotation speed of the generator [28,29].

As previously stated, the use of bulky components and devices, such as low-frequency transformers (LFTs), leads to reduced efficiency ratings, lower reliability, and higher costs. Therefore, the adoption of converters with higher power density, i.e., based on HFTs or MFTs, is seen as the most viable solution. This situation is justified in [30,31,32,33], where various isolated interface configurations were introduced for variable-speed wind turbines and high-voltage direct current (HVDC) systems in offshore wind farms. In addition, [34] proposed replacing the common LFT with a SST structure.

Figure 1. Possible architectures for interfacing turbine generator technologies with the utility grid: (a) fixed speed; (b) partial variable speed; (c) variable speed with partial-scale frequency converter; and (d) variable speed with full-scale power converter.

Figure 1. Possible architectures for interfacing turbine generator technologies with the utility grid: (a) fixed speed; (b) partial variable speed; (c) variable speed with partial-scale frequency converter; and (d) variable speed with full-scale power converter.

Upon analyzing the literature and recent energy data, it becomes evident that both fuel cells and solar PV technology offer the greatest potential for innovation and evolution. Regarding solar PV, such technological maturation may involve modifications in the characteristics of the modules, but more notably, it necessitates improvements in the interface system between this RES and the utility grid. As observed in Figure 2, depending on the architecture and connection between a certain number of modules, a different number of conversion stages must also be picked [35].

For low-power applications, a single inverting stage may be commonly applied, whereas for a multistring configuration, multistage architectures are always required [36]. This scenario is equally replicable in microinverter configurations, in which a dedicated power converter is associated with each solar PV module. On the other hand, for medium- and high-power scenarios, centralized and multistring configurations are normally considered. Nevertheless, the remaining architectures, albeit less commonly, may be equally used, as proven in [37], where a comparative study regarding power quality, costs, general characteristics, and power losses in large-scale solar PV power plants was carried out. As expected, in higher power scenarios, to reach the desired voltage and current values, several modules must be interconnected. As with microinverter configurations, each solar PV array has a dedicated DC–DC power converter, which consequently, may result in different arrangements for the power converters, i.e., cascade or in parallel. In a single-stage scenario, the converter of each module or array is, necessarily, an inverter, whilst the combination of topologies in a multistage scenario, as seen in Figure 3, provides greater controllability and flexibility to the solar PV system. For these cases, a DC–DC converter is dedicated to each module (with MPPT functionality), and one or more DC–AC converter is used to invert the power generated by all the arrays.

For any scenario involving solar PV modules, i.e., residential, commercial, portable, and, most importantly, large-scale power plants, there is always a need for isolation due to safety issues. As already mentioned in the introduction, the inclusion of a galvanically isolated transformer, although not mandatory in certain cases, allows, in addition, the prevention of equipment damage or malfunctioning due to electrical faults (e.g., short circuits). Likewise, the efficiency of solar PV systems becomes, consequently, higher, since possible paths for leakage currents are eliminated.

Figure 3. Possible multistage multiple input configurations to interface grid-connected PV systems: (a) parallel DC; (b) parallel AC; (c) cascade DC; and (d) cascade AC.

Figure 3. Possible multistage multiple input configurations to interface grid-connected PV systems: (a) parallel DC; (b) parallel AC; (c) cascade DC; and (d) cascade AC.

As with solar PV systems, the use of fuel cells also gathers strong consensus within the scientific community as a prominent technological trend with substantial developmental potential. These devices directly convert chemical energy into electrical energy, typically relying on hydrogen as the primary fuel source. This chemical reaction enables the generation of electricity in a clean and efficient manner, although power electronics converters are required to regulate the generated values from each individual cell.

In line with any renewable technology, based on the specific requirements of the system, such converters must be responsible for voltage, current, and, in the case of AC systems, frequency adjustments. However, as experienced with solar PV modules, the output values of a fuel cell will be influenced by temperature fluctuations, load changes, and variations in the fuel supply [38,39,40]. Therefore, power converters also bear the responsibility of providing stable and reliable operation to the RES-based generation system. This necessitates the implementation of appropriate modulation and control algorithms. Notably, MPPT control algorithms are once again mandatory, with numerous variants commonly documented in the literature, as depicted in [41,42,43,44].

Fuel cells, in a simplified manner, can be described as a DC energy source and, as such, are commonly associated with and integrated into multiple input architectures, as depicted in Figure 3. This approach aims to enhance power generation, efficiency, and flexibility. Depending on the application scenario, these DC–DC power converters can also be isolated or non-isolated, as mentioned in [45]. In this comparative study, all possible topologies for interfacing with fuel cells were enumerated, as well as a detailed section regarding technical challenges, power quality problems, and power control issues.

For multiple input systems, careful attention must be devoted to energy management, particularly in hybrid scenarios involving different RESs. Interface systems must adjust various voltage and current values to efficiently deliver energy to the utility grid, loads, or ESSs. To this end, mechanisms for phase and frequency synchronization, management and regulation of imbalances and fluctuations, as well as mitigation and smoothing of transients between different RESs, need to be ensured [46]. Among various examples, power sharing techniques are highlighted, as indicated in [47].

The need for isolation is equally highlighted in turbine generator technologies. Both in this case and in large-scale solar PV power plants, LFTs are used nowadays for interfacing with low-voltage feeders, which confer a static and passive behavior in the face of possible momentary changes in operating conditions. In this regard, for the above-mentioned reasons, migrating towards isolated power electronics solutions switching at medium- or high-frequencies, e.g., the SST (shown in its traditional structure in Figure 4), is seen as one of the facilitating trends for the continuous dissemination of smart grids and microgrids. Furthermore, for any of the cases illustrated in Figure 2, Figure 3 and Figure 4, the three-phase configuration may also be considered.

The isolated DC–DC conversion stage, depending on the objectives, may be employed at different points in the interface circuit. For instance, in cases where each module or array disposes of a dedicated converter, isolation can be achieved individually (using multiple converters) or globally, i.e., through a single isolated power converter when interfacing with the utility grid. As expected, the greater the number of isolated power converters used, the greater the cost of the entire installation, but the greater the safety and flexibility. In other words, each array can be sized independently, a particularly useful feature when solar PV modules are subject to different radiation levels or shading. On the other hand, a greater number of isolated power converters also leads to fluctuating efficiency values, since transformer losses are replicated in multiple points. With regard to reliability, higher values are achieved if each array is isolated from the others since possible localized faults will not represent a reprisal for the global solar PV system.

Due to the growing adoption of emerging DC power devices, the establishment and utilization of DC power grids have become more prevalent. As a key element, such architectures potentially allow for an effective reduction in the number of conversion stages, which consequently translates into enhanced performance and reduced losses. In DC systems, issues related to power quality are also effectively addressed, thereby facilitating the integration of RES-based systems. In the case of DC microgrids, frequency fluctuations are also easily mitigated, which presents a significant advantage during islanding mode.

On the other hand, the migrations towards DC power grids still lack technological maturity and would require a significant overhaul of a substantial portion of the existing infrastructure. Furthermore, the standards and regulatory frameworks for the implementation of such DC structures are still in an unstable phase, precluding their immediate adoption. In addition, DC faults are recognized as one of the major limitations to the ongoing proliferation of DC power grids. In most instances, power converter topologies are chosen based on their capability and characteristics to efficiently mitigate such faults.

According to [48], DC pole-to-pole faults originate from either a short-circuit occurrence or insulation breakdown. Since the impedance of a DC line is much lower when compared to traditional AC systems, this type of fault will result in an almost immediate capacitor discharge. While less common, DC pole-to-pole faults can cause severe damage to electronic equipment, the majority of which have limited capabilities in terms of overcurrent protection. Therefore, a faster response time is required for protection systems, which is considered the primary drawback compared to AC energy transmission and distribution systems [49].

As potential problems caused by DC faults can quickly escalate into a cascade effect, affecting several devices, it is imperative for modern power systems to incorporate advanced monitoring and control technologies. Additionally, robustness and redundancy should also be strengthened to enhance the stability of transmission and distribution systems, thereby isolating DC faults.

To mitigate these issues, DC circuit breakers should be integrated into the DC system. Alternatively, converter topologies with enhanced DC-side fault handling capabilities, such as modular multilevel converters (MMCs) [50], can also be employed. Nonetheless, it is always advisable to use multiple DC circuit breakers to prevent extensive propagation of the fault. As a result, other segments of the DC line can continue to operate redundantly, similar to current AC systems. In turn, when utilizing power converters as an active fault mitigation element, appropriate control algorithms should also be considered, as demonstrated in [51]. In this article, fault behaviors in the modal-domain, time-domain, and frequency-domain were analyzed, highlighting the need to implement systems with a high sampling frequency. Additionally, extensive studies of different protection methods in HVDC transmission systems were conducted in [52,53].

To better integrate emerging technologies, hybrid power grids have also been considered a versatile and flexible solution. These architectures can harness the advantages of both grid types, albeit with more intricate control structures [54,55].

The need for isolation and the features of a determined RES interfacing system will ultimately depend on the specific needs of each project. According to the proposed objectives, different architectures and configurations must be chosen, as well as the topologies for power converters. For isolated DC–DC topologies, the design and sizing are of particular importance since there will be greater losses and electromagnetic interferences will directly influence the operation of all devices. Thus, appropriate modulations and strategies aimed at increasing reliability, flexibility, and efficiency must be defined.

3. Isolated DC–DC Converter Properties, Topologies, and Modulations

The selection of topologies for interfacing RES-based generation technologies is always in consonance with the characteristics of the converter itself, namely, the power and voltage ratings, number of stages, added functionalities, and isolation requirements. As discussed in the previous section, the use of isolated topologies is reflected globally in a greater number of advantages, but to maximize efficiency, appropriate modulation techniques, as well as MPPT and utility grid connection algorithms, should also be chosen.

In medium- and high-voltage applications, renewable energy interface systems predominantly adopt a three-phase configuration, making the choice of topologies for each converter of utmost importance, with a particular emphasis on isolated DC–DC topologies. Nevertheless, considering the aforementioned advantages, isolation also plays a preponderant role in low-power single-phase systems. Given the high technological maturity of LFTs, they are widely used in numerous cases; however, in line with the technological assumptions for future smart grids, this conversion stage will necessarily have to present new and innovative functionalities.

These innovative features can only be achieved through the implementation of appropriate modulations and architectures in each power electronics system. In the case of isolated topologies, specifically DC–DC converters, the utilization of HFTs (or MFTs), supported by WBG semiconductors, is indispensable. By considering the transformer turn ratio (N_t = N₁/N₂), a broader range of voltage conversions is attainable, which endows these converters with increased flexibility, reliability, and compatibility with various devices. Additionally, these isolated stages offer advanced fault tolerance capabilities, effective mitigation of ground potential differences (e.g., ground loops), and enhanced immunity to noise and electromagnetic interference (EMI) [56]. Such attributes are particularly critical in circuits used for data transmission or analog signal applications.

Among the DC–DC isolated topologies, the dual active bridge (DAB) converter holds a prominent position and is widely utilized across a wider range of scenarios [57,58,59]. However, it is also important to note that various other topologies have been extensively studied in the literature. The selection of a particular topology depends on several factors, including the specific application scenario, control and implementation simplicity, power range conversion requirements, and the desired number of outputs.

3.1. Conventional Isolated Topologies

In the interface with RES-based generation technologies, among conventional isolated DC–DC converters, the flyback, forward, and push–pull topologies are commonly mentioned and utilized. These topologies are depicted in Figure 5, Figure 6 and Figure 7, showcasing both their classic architecture and occasional variations. Despite having distinct characteristics, especially in terms of operating power, they are widely employed due to their simplicity, ease of control, and relatively high efficiency across a wide range of voltage levels. Consequently, the abovementioned isolated DC–DC topologies provide increased flexibility, are considered excellent cost-effective power electronics solutions, and given their technological maturity, are frequently found in a broad range of scenarios. Nonetheless, besides differences in operating power, these topologies also vary in transformer type, number of possible voltage outputs, and immunity to interference caused by semiconductor switching. Once again, both component selection and converter design play crucial roles in achieving the desired performance and efficiency.

3.1.1. Flyback

Ordering flyback, forward, and push–pull converters based on desired output power, the flyback converter presents the lowest nominal value. However, by increasing the voltage levels supported by WBG semiconductors and the power density of converters, and by including high-voltage gate drivers in control circuits, topologies that were previously limited to low-power applications are now being employed in a wider range of power solutions. Nevertheless, the selection of a specific topology also considers other significant factors, including associated costs, volume and weight, electrical stress, output noise, and input voltage range, among others.

Since the flyback converter generates a non-inverting output, its application in several devices is facilitated, eliminating the need for additional complementary circuits. Thus, this isolated DC–DC topology is regularly employed in low-power scenarios, such as power supplies and microinverter configurations for solar PV systems [60]. As shown in Figure 5a, it incorporates a single active switch (S₁), resulting in reduced costs, as well as switching losses. A coupled-inductor transformer is added to store energy during the on-time of S₁ and transfer it to the secondary side during the off-time [61]. It is important to note that current rectification is solely performed by the diode D₁, leading to an increased ripple in the output voltage waveform (V_out) and, consequently, lower efficiency. Moreover, the use of HFTs or MFTs may also be considered, thus providing not only galvanic isolation but also increasing the step-up ratio. Therefore, based on N_t and the duty cycle (D) applied to S₁, V_out is determined as follows [23]:

$$V_{out}={\displaystyle \frac{N_2 D}{N_1 (1-D)}}{V}_{dc}$$

(1)

With the use of a flyback converter, it is also feasible to obtain multiple DC outputs by adding extra windings to the HFT and employing suitable circuits and modulation techniques to regulate each individual output [62]. Depending on N_t and the implemented control circuit, these DC outputs can operate independently, which means the possibility of obtaining distinct voltage levels and connecting devices with differently rated power. However, as the number of DC outputs increases, the control complexity also escalates.

Regarding topology classification, the flyback converter is categorized as a single-ended configuration. In other words, only the primary winding of the transformer is connected to the power switches, whereas in double-ended topologies, the power switches are connected to both sides of the transformer. As a result, single-ended topologies do not support bidirectional power flow. In a flyback converter, the active semiconductor S₁ is connected to the primary winding, while the rectifying element, D₁, is linked to the secondary. Analyzing the transformer’s B–H curve, the flux only exists in one quadrant, specifically from zero to positive values, without reversing direction. This leads to enhanced power handling capabilities and efficient conversion since the absence of bidirectional power flow reduces losses associated with reverse current and voltage stress.

However, when the flyback converter operates in discontinuous conduction mode (DCM), the primary side current stress of the HFT increases, leading to higher conduction losses and elevated peak currents. Additionally, the operation in DCM contributes to greater THD due to abrupt switching transitions and increased current ripple. This effect is aggravated as voltage ripple on the DC bus augments, which negatively impacts the converter’s performance by increasing switching losses and reducing overall efficiency [63]. In DCM operation, switching and conduction losses tend to decrease at lower power levels given the reduced duty cycle and shorter conduction intervals. However, as the power increases, the current stress on the switching devices grows, leading to the need for thermal dissipation requirements. Additionally, core losses in the HFT remain significant or even increase due to the higher flux density and frequency components associated with DCM operation. To mitigate these drawbacks, as shown in Figure 5b, interleaved topologies operating in continuous conduction mode (CCM) are often employed [64,65]. These topologies are more robust and efficient and exhibit lower voltage stress, as well as increased immunity to EMI. At higher power levels, current ripple, peak stress, and core saturation effects are significantly reduced. In the case of the conventional flyback converter, CCM operation is unpopular. Nonetheless, its performance can be improved by employing zero-voltage switching (ZVS) [66] or soft-switching techniques, which, in turn, require the inclusion of clamping circuits to the basic topology.

In the case of the interleaved flyback topology, the inclusion of two large magnetizing inductances (L_m1 and L_m2) ensures that load variations do not impact the converter’s performance [67]. However, it is worth noting that, apart from DCM and CCM, the converter can also operate in boundary conduction mode (BCM). This mode enables higher power density and improved output voltage power quality. Nevertheless, it comes at the cost of increased switching and gate drive losses compared to DCM and CCM.

Figure 5. Isolated DC–DC flyback converter: (a) conventional; (b) interleaved flyback.

Figure 5. Isolated DC–DC flyback converter: (a) conventional; (b) interleaved flyback.

3.1.2. Forward

In industrial applications and motor drives, isolated DC–DC forward converters are commonly chosen due to their excellent voltage regulation and high efficiency. As illustrated in Figure 6a, similar to the flyback converter, the forward converter is classified as a single-ended topology, and a single active switch, S₁, is considered. Consequently, by adjusting D, a given input voltage, V_dc, may either be stepped up or stepped down. This versatility makes it suitable for various applications; however, it is important to note that this topology suffers from a poor transformer utilization ratio.

Figure 6. Isolated DC–DC forward converter: (a) conventional; (b) active clamp forward.

Figure 6. Isolated DC–DC forward converter: (a) conventional; (b) active clamp forward.

The improved efficiency exhibited by this topology is attributed to lower conduction losses and its ability to operate, commonly, both in DCM and CCM. Additionally, compared to flyback, the forward converter presents lower semiconductor voltage stress, as it only needs to handle the value of V_dc [68]. In contrast, the semiconductor S₁ of the flyback converter needs to withstand both V_dc and the reflected voltage of V_out.

Nonetheless, by recycling the energy stored in the transformer’s leakage inductance, a forward topology with active clamp circuitry, as depicted in Figure 6b, significantly enhances the overall efficiency of the power system [69]. This modification in the primary circuit enables further reductions in voltage stress and switching losses, as soft-switching is ensured. Despite operating in two quadrants, it is worth noting that an active clamp forward converter may experience high peak flux levels during startup and transient states. Additionally, since control complexity increases, this variant of the conventional forward converter is commonly used in devices and technologies that demand higher efficiency, reduced stress, and substantial power levels, such as the telecommunications sector and RES-based power elements.

3.1.3. Push–Pull

Unlike flyback and forward converters, a push–pull topology is classified as double-ended, offering several advantages in terms of transformer core utilization. As mentioned above, double-ended topologies are particularly well suited for high-power applications. Furthermore, the architecture of the push–pull converter allows for an extended D range, approaching values close to 100%. As demonstrated in Figure 7, the secondary stage of the converter includes two diodes, D₁ and D₂, connected to the secondary windings of the transformer [70,71]. Once again, these components are responsible for rectifying the current, which conveys a unidirectional nature to the power converter. However, in modern devices, these diodes are often replaced by active switches such as Si IGBTs, SiC MOSFETs, or GaN-based technology [72].

Figure 7. Conventional push–pull converter topology.

Figure 7. Conventional push–pull converter topology.

Therefore, similar to the forward converter, the push–pull topology is commonly employed in high-power scenarios that require an enhanced efficiency of conversion. Due to its symmetrical structure, the operation of the push–pull converter is balanced, which minimizes common-mode noise, EMI, and output current ripple. However, some drawbacks must be considered. The topology exhibits moderate complexity, and there is increased voltage stress on switches S₁ and S₂ during their off-state periods, approximately twice the value of V_dc. Additionally, the use of a center-tapped HFT adds to the cost and volume of the power converter, which can be a limitation in applications where it is vital to reduce the physical footprint, such as in RES-based generation systems [73].

3.1.4. Other Topologies

Based on conventional buck and boost structures, isolated versions have been analyzed in the literature, though less frequently. Their relevance is particularly evident in integrating RES-based systems in future microgrids and smart grids. As with non-isolated counterparts, their main function is to step-down or step-up a given V_dc, respectively. Isolated DC–DC converters rely on the principle of pulse width modulation (PWM), but in this case, V_out, is also determined by N_t. In terms of efficiency, isolated topologies significantly improve power dissipation and losses related to the leakage inductance (L_k), enhancing overall converter performance. However, the inclusion of galvanic isolation increases EMI levels, negatively impacting safety and the ability to effectively stabilize and regulate V_out.

On the one hand, isolated buck converters are commonly used in the telecommunications industry, automotive sector, and power supply devices. On the other hand, isolated boost converters are primarily employed in high-voltage DC bus regulation and ESSs. However, both of them are highly suitable for the renewable energy sector, especially in microgrid applications, where voltage step-up or step-down functionalities are essential for efficient power management. Additionally, they can be utilized as an intermediate DC–DC stage between the solar PV modules and the corresponding grid-tied inverter. Based on the literature analysis, these topologies may be interconnected and associated with full-bridge converters, both in the inversion stage and in the rectification stage. Furthermore, isolated versions of zeta, SEPIC, and cuk converters are also available, as illustrated in Figure 8, showcasing multi-output structures [74]. These configurations are mentioned in [75], where an in-depth study on multi-output DC–DC converters is provided.

The article not only discusses the fundamental topologies but also explores potential combinations between different structures and analyzes their applicability in various scenarios. However, these variations not only provide improved voltage regulation through N_t but also effectively mitigate the input and output current ripple. This capability is particularly important in RES-based generation elements where it is necessary to interface between power systems with different ground references, attenuate noise, and minimize the effects of switching transients. Additionally, these isolated converters facilitate stable V_out regulation even in the presence of load variations and fluctuations in V_dc.

3.2. Bridge-Based Isolated Topologies

Bridge-based isolated topologies, compared to the above-presented conventional structures, offer a broader range of advantageous features, making them the most commonly used converters in modern power electronics systems [76]. Similarly to the pushpull converter, bridge topologies are classified as double-ended, which allows for increased power handling capability. Among these properties, improved stability and voltage regulation are highlighted, as different modulation techniques can be easily implemented.

Depending on the specific topology, the voltage stress on semiconductor devices can be minimized, while simultaneously providing greater flexibility in voltage conversion. In other words, depending on the application scenario, these topologies enable both stepping down or stepping up a given V_dc. Moreover, excellent scalability and modularity are offered, i.e., there is great flexibility in interconnecting multiple isolated DC–DC bridge converters in cascaded or parallel arrangements [77]. Multilevel configurations are particularly adequate for high-power scenarios where a given voltage or current needs to be divided among several isolated modules. Moreover, this modular characteristic enables increased redundancy and efficient fault mitigation within individual modules [78].

Furthermore, bridge-based topologies may also be easily adapted to enable bidirectional power flow. While this feature may not be essential in RES-based generation solutions, it plays a significant role in enhancing flexibility and versatility in the context of smart grids (e.g., for interfacing ESSs), while also reducing losses. Additionally, by incorporating active switches instead of passive components (e.g., diodes), recent and innovative modulation techniques and control algorithms may be more easily implemented, thereby improving the overall performance of the interface system [79].

3.2.1. Conventional Half-Bridge, Single Active Half Bridge, and Dual Half-Bridge

In isolated half-bridge topologies, as illustrated in Figure 9a, the inversion of a given V_dc is accomplished using two active semiconductor devices, S₁ and S₂, and a split DC bus constituted by capacitors C₁ and C₂. The windings of a transformer, ideally an HFT or an MFT, are connected to the midpoint of each arm, allowing the subsequent rectification of the signal on the secondary side of the converter.

In comparison with the conventional isolated DC–DC converters, particularly the push–pull topology (both double-ended), half-bridge configurations offer lower voltage stress as the value of V_dc is not exceeded in the semiconductor devices. Therefore, it is vital to carefully select the appropriate switching elements that best suit the requirements of the power electronics system. In addition, as previously described, there is no need to use transformers with multiple windings, leading to improved utilization of the transformer’s core window and overall performance. Nevertheless, compared to other bridge topologies, the half-bridge converter provides simplicity and cost advantages due to its lower number of active switches. However, it is limited to voltage mode operation since maintaining voltage balance at the midpoint of the DC bus at ½ V_dc is crucial [80].

Therefore, in conjunction with a half-bridge topology based on active switches, the rectification block is equally important to ensure efficient energy conversion. As shown in Figure 9b, the conventional isolated asymmetrical DC–DC half-bridge employs a center-tapped HFT, typically resulting in a single DC voltage output. To obtain a higher number of levels and isolated outputs, extra circuits are required, which increases complexity and introduces noise. Thus, it is evident that the conventional half-bridge is not easily scalable, necessitating the use of additional HFTs to increase power levels. Regarding fault tolerance, ensuring continuous operation of the converter becomes extremely challenging if one of the semiconductors gets damaged, which indicates a lower level of redundancy. On the positive side, among the topologies depicted in Figure 9 and Figure 10, the conventional half-bridge exhibits the lowest associated costs and complexity. This is because, for unidirectional arrangements, it is only necessary to control the half-bridge inverter.

In a single half-bridge converter, as the name suggests, only one switching stage is adopted, i.e., one side of the isolated DC–DC power converter operates in a passive mode for rectification purposes, while the other, specifically the inverting primary side, incorporates active semiconductor devices [81]. As illustrated in Figure 10a, signal rectification is achieved by connecting the secondary winding of the HFT to an arm constituted by two diodes (D₁ and D₂) and to the midpoint of the output split DC bus. On the other hand, a dual half-bridge (DHB) configuration has a bidirectional nature. However, as already indicated, this requirement is not mandatory when interfacing with RES-based generation technologies, despite the significant advantages it offers. As observed in Figure 10b, the primary and secondary sides of the transformer are symmetrical, i.e., rectification is now performed using a half-bridge topology with active switches [82]. In total, four switching devices are used, with a split DC bus on both sides.

Once again, due to the characteristics of bridge-based topologies, a half-bridge rectification stage offers greater redundancy, enhanced fault tolerance, and the generation of multiple independent DC outputs, thus demonstrating its flexibility and scalability. Additionally, thanks to the high modularity of the power converter, an interleaved DHB topology can easily be considered for a wide range of applications. In general, interleaved topologies may be described as a certain number of power converters connected in parallel to share a certain current among them. This type of topology offers significant advantages, such as low ripple frequency and losses, but they come with higher costs. Similarly to multilevel topologies, interleaved converters are commonly utilized in high-power scenarios to increase conversion efficiency and reliability.

Figure 10. Half-bridge isolated DC–DC topologies: (a) single active half-bridge and (b) dual half-bridge.

Figure 10. Half-bridge isolated DC–DC topologies: (a) single active half-bridge and (b) dual half-bridge.

3.2.2. Conventional Full-Bridge and Half-Full-Bridge

Despite the numerous advantages offered by isolated DC–DC half-bridge-based topologies, particularly in terms of cost and complexity reduction, the voltage conversion process is not entirely efficient. The semiconductors in the single switching arm operate in a complementary manner, resulting in a voltage across the primary winding that is half the value of V_dc. In contrast, in a full-bridge structure, as represented in Figure 11a, when a diagonal pair of active semiconductors is in the one-state, the voltage across the primary winding assumes the value of V_dc without the need to split the DC bus. As a result, for a given power level, the current flowing through the primary side of the HFT is half compared to a half-bridge topology, thus leading to higher efficiency.

In isolated DC–DC full-bridge converters, the rectifying stage, among different possibilities, may employ different topologies such as the conventional structure (as shown in Figure 11b), a half-bridge structure, a diode full-bridge, or even a full-bridge converter using active semiconductor devices. Similarly to the conventional isolated DC–DC half-bridge converter, the incorporation of a center-tapped HFT is also required, which constitutes one of the limitations compared to alternative rectifier stage topologies.

Figure 11. Conventional full-bridge isolated DC–DC converter (a) structure and (b) topology.

Figure 11. Conventional full-bridge isolated DC–DC converter (a) structure and (b) topology.

Therefore, an isolated DC–DC full-bridge topology, despite its moderate complexity, significantly improves the efficiency of conversion. Unlike half-bridge topologies, where each active semiconductor handles half the value of V_dc, in a full-bridge structure, each switching device must support V_dc entirely, which, consequently, reduces the current stress. Simplistically, by distributing the current among a larger number of semiconductors, both conduction and switching losses are minimized, particularly in high-power scenarios. In terms of costs, although the number of active semiconductors is doubled (four instead of two in the inverting stage), the number of capacitors is halved.

Since full-bridge structures are highly suitable for a wider input voltage range, they offer superior versatility in various application scenarios. Examples include their use in grid-tied inverters, motor drives, uninterrupted power supplies (UPSs), and, as expected, in the renewable energy sector. Globally, compared to isolated DC–DC half-bridge topologies, a full-bridge-based structure provides enhanced voltage regulation, increased redundancy, minimized switching stress, and the ability to operate at higher power levels. However, even in the case of isolated DC–DC half-bridge topologies, the rectification stage can be implemented using full-bridge-based structures, utilizing either diodes or active semiconductors such as MOSFETs or IGBTs. These mixed configurations, commonly referred to as half-full-bridge configurations, are shown in Figure 12 and Figure 13 [83].

As previously mentioned, the converters used in the bidirectional topology shown in Figure 12b could be easily reversed [84]. In other words, the inversion of V_dc would be achieved using a full-bridge topology, which yields higher conversion efficiency, while rectification would be accomplished using a half-bridge structure [85]. However, as indicated in the section discussing half-bridge-based topologies, interleaved structures can also be implemented in these mixed configurations, as shown in Figure 13 [86]. The inverter stage is composed of an interleaved full-bridge topology, whereas rectification is accomplished using a half-bridge diode-based converter.

Figure 13. Unidirectional interleaved full–half-bridge isolated DC–DC converter.

Figure 13. Unidirectional interleaved full–half-bridge isolated DC–DC converter.

3.2.3. Single Active Bridge and Dual Active Bridge

The single active bridge (SAB) and DAB converters utilize full-bridge structures in both the inverting and rectifying stages, as depicted in Figure 14. An active full-bridge converter is connected to the primary side of the HFT. However, the main difference lies in the topology and behavior of the rectifying stage. In the SAB converter, a full-bridge diode-based AC–DC converter is connected to the secondary winding of the HFT, which confers a passive behavior [87]. Subsequently, as is common practice in all DC–DC converters, a DC bus is included to smooth and stabilize a specific V_out. As a result, the power flow in this converter is unidirectional, requiring modulation on the DC-AC converter connected to the primary side of the HFT, while the secondary remains passive.

The interface with power generation units involves the use of unidirectional structures. However, there are numerous advantages if the rectifying stage also operates actively and dynamically. Therefore, a DAB converter can be described as a symmetrical isolated DC–DC structure in which two full-bridge topologies, composed of active semiconductors devices, are connected to each winding of an HFT [88,89,90,91]. When interfacing with RES-based systems, bidirectional power flow is not the primary factor driving the adoption of a DAB converter. Instead, the focus is on the benefits of flexibility, scalability, and efficiency. In general, the independent control of each bridge aids in stabilizing the voltage on the primary or secondary side of the DAB converter, i.e., V_dc1 and V_dc2, respectively. This feature plays a critical role in the operation of modern electronic solutions.

Additionally, it is possible to achieve soft-switching throughout the entire power range, which is not feasible, e.g., in half-bridge isolated topologies. With effective controllability over power flow, DAB converters exhibit naturally higher power density. This characteristic is extremely important in power electronics systems for interfacing with RESs, as it is desirable to implement compact and lightweight technologies [92].

Similarly to all bridge-based structures, DAB converters offer superior modularity, scalability, flexibility, and redundancy. To enable the interface with high-power systems, it is particularly important to interconnect different bridge converters in cascade or parallel configurations, which form the common structures: (i) input-parallel/output-parallel (IPOP); (ii) input-series/output-series (ISOS); (iii) input-series/output-parallel (ISOP); and (iv) input-parallel/output-series (IPOS) [93,94]. By distributing the current and/or voltage through a larger number of active semiconductors, the stress on each device is significantly reduced, leading to improved overall performance and reliability.

In this respect, Figure 15 depicts the respective block diagrams that represent the abovementioned arrangements. However, the specific topology of each block should be selected based on the requirements of the power electronics solution [95]. If a full-bridge topology is chosen, these diagrams would indicate the connection of multiple DAB converters with different arrangements. Notwithstanding, half-bridge-based topologies could also be chosen for this purpose. As depicted in red in this figure, the complexity of control and modulation increases as the number of interconnected isolated DC–DC converters rises. However, as mentioned above, this results in lower semiconductor stress.

However, as a distinguishing feature, DAB also allows for the mitigation of harmonic content in V_out. By facilitating the implementation of loss-minimization techniques, such as ZVS, and appropriate modulation techniques, including phase-shift variations, EMI is significantly reduced. Moreover, the adoption of a DAB converter ensures, inclusively, compliance with grid interconnection standards, making them highly adequate for a broader range of operating scenarios. These key factors contribute to the widespread adoption of DAB converters in the most diverse power electronics systems.

Figure 15. Arrangements for modular converters: (a) IPOP; (b) IPOS; (c) ISOP; and (d) ISOS.

Figure 15. Arrangements for modular converters: (a) IPOP; (b) IPOS; (c) ISOP; and (d) ISOS.

The different arrangements depicted in Figure 15 are a consequence of the high modularity and scalability of bridge-based topologies, including DAB converters. However, multiport configurations can also be easily achieved by modifying the structure of the MFT [96]. As shown in Figure 16, the number of windings in the MFT corresponds to the number of independent input/output terminals [97]. Depending on the topology of each bridge and the power device connected to each terminal, the power flow can be either unidirectional or bidirectional. Each input/output terminal is galvanically isolated from the others, while efficient and flexible energy exchange is enabled between them. Since the power electronics system is totally integrated, higher power density is obtained, simplifying the implementation of control algorithms and modulation techniques. Thus, Figure 16a presents the block diagram of an isolated bidirectional multiport DC–DC converter with four input/output terminals, while Figure 16b illustrates the same converter based on a quad active bridge (QAB) configuration [98,99,100].

Given that DAB converters are used to enable dynamic behavior in power interface systems during transient states, the role of modulation and control algorithms becomes crucial in optimizing the functionalities provided by the topology itself. Among other aspects, one of the key objectives is to mitigate current offsets by attenuating the effects of the magnetizing current of the HFT and the current that flows through L_k (i_Lk). In general, to improve voltage conversion and minimize losses, it is essential to incorporate the latest WBG devices, implement techniques aimed at improving power quality, and eliminate circulating currents. Additionally, it is equally important to address issues related to semiconductor stress, transformer saturation, and current spikes, which require a fast dynamic response from the power system.

In addition to adopting various strategies aimed at mitigating the aforementioned issues, the implementation of phase-shift modulation is widely accepted as a common practice. The consensus surrounding the adoption of phase-shift modulation is mainly due to its relative simplicity and ability to achieve high conversion efficiency. However, it is worth noting that, depending on the specific application and the chosen phase-shift variation, considerable levels of reactive power and circulating currents may be generated. Therefore, it becomes critical to extend the ZVS range, especially in situations where the primary side voltage, V_dc1, deviates from the value imposed by N_t, i.e., N_t V_dc2, where V_dc2 represents the secondary side voltage of the DAB converter [101].

Among the possible variations in phase-shift modulation, single phase-shift (SPS), dual phase-shift (DPS), extended phase-shift (EPS), and triple phase-shift (TPS) are highlighted, as compared in [102,103]. The key distinction lies in the number of degrees of freedom, i.e., the value assigned to each phase-lag angle, both between the two full-bridge converters that compose the DAB (outer phase angle, D₀), and within each bridge arm (inner phase angles, D₁ and D₂). The selection of a specific modulation technique depends on the application scenario, but generally, more intricate techniques offer a greater range of advantages in terms of efficiency [104]. In essence, D₀ regulates the direction and magnitude of power transfer within the DAB converter, while D₁ and D₂ angles contribute to enhanced performance and improved power quality.

Nevertheless, for any variant, all eight-semiconductor devices are constantly switching with a fixed D of 50%. In this regard, the phenomenon that enables power transfer in this isolated DC–DC converter is the phase difference between the squared voltage waveforms across each winding of the HFT, i.e., V_pri and V_sec. When these waveforms are phase-shifted, a voltage is induced across L_k and, consequently, a current will flow through it (i_Lk). Depending on whether the phase difference is positive or negative, the direction of i_Lk will vary accordingly, as well as the power flow direction in the DAB converter. In other words, if V_pri leads V_sec, power will flow from the primary to the secondary side. Conversely, if V_pri lags behind V_sec, the opposite will occur. Therefore, adjusting this phase difference regulates power transfer, both in terms of direction and magnitude.

In power systems where V_dc1 closely approximates V_dc2, SPS modulation is commonly utilized. Among the four aforementioned variations, it is the least complex, despite providing satisfactory practical results. However, a single degree of freedom is considered, D₀, which may result in increased reactive power generation [105]. The signals applied to semiconductors S₁ and S₃ are 180 degrees out of phase, as well as S₅ and S₇. The value of D₀ is applied to S₅, which indicates the phase difference between S₁ and S₅.

In turn, for DPS modulation, an extra degree of freedom, D₁, is introduced along with D₀. D₁ is applied to S₃, considerably improving DAB’s efficiency, particularly in expanding the ZVS range. Moreover, this modification allows for the generation of a three-level waveform across V_pri and V_sec, which effectively contributes to reducing reactive power and total harmonic distortion (THD) produced by the converter [106].

In turn, EPS and TPS modulations are used in specific cases. Regarding the triggering gate signals applied to each switching device, the difference lies in the value applied to S₇. TPS modulation introduces an additional degree of freedom, D₂, which totals three phase-lag angles with respect to S₁. The control complexity substantially increases, but so does the conversion efficiency, as indicated in [107,108,109]. However, it should be noted that the number of applications that require the use of such modulation is more limited, whereas the use of the EPS technique is directed towards power solutions that require the transfer of higher power magnitudes within a shorter time period. Both techniques enhance the converter’s flexibility, leading to greater precision and reduced losses [110,111].

Therefore, the summary of gate signal values applied to each switching device is presented in

Table 1. Gate signals applied to the top arm semiconductors for SPS, DPS, EPS, and TPS modulation.

	S1	S3	S5	S7
SPS	Ref.	Ref. + 180°	Ref. + D0	S5 + 180°
DPS	Ref.	Ref. + D1	Ref. + D0	S5 + D1
EPS	Ref.	Ref. + D1	Ref. + D0	S5 + 180°
TPS	Ref.	Ref. + D1	Ref. + D0	S5 + D2

. Since all gate signals are phase-shifted from the value assigned to S₁, this signal is regarded as the reference (Ref.) to which the phase angles (degrees of freedom) will be added. Nonetheless, as expected, the signals applied to the bottom semiconductors of each arm are complementary to the ones depicted in the same table.

3.2.4. Series Resonant Full-Bridge-Based Topologies

Achieving ZVS across a wide range of output voltage and power is a major challenge in DAB converters. To overcome this issue, phase-shift modulation techniques with multiple degrees of freedom have proven to be an excellent solution. However, alternative approaches involving circuitry modifications can also be effective.

Among these changes, the inclusion of a series resonant tank into the conventional DAB topology (FB-LC-DAB) has shown promising results, as exemplified in Figure 17a, in which an LC resonant tank is connected to the windings of a single phase [112] and three-phase [113] HFTs. Additionally, the inclusion of an LC series resonant tank in a mixed full–half-bridge converter topology is studied in [85]. On the other hand, LLC resonant converters have demonstrated high efficiency in voltage conversion, providing increased power density, as shown in Figure 17b, where a diode bridge is integrated on the secondary side of the HFT for rectification [114]. Nonetheless, these LLC resonant tanks may even be included in multilevel topologies, as indicated in [115], or in hybrid configurations [116]. Furthermore, Figure 17c presents a bidirectional version, replacing the diode bridge with a full-bridge converter equipped with active switches and adding a resonant capacitor on the secondary side. This modified converter is referred to as a resonant CLLC converter [117,118]. Moreover, as mentioned in [119] and depicted in Figure 17d, the incorporation of a CLC resonant tank is considered. Figure 17e introduces a complementary approach by proposing an LCL variant [120,121].

In summary, isolated resonant DC–DC converters offer a broader range of advantages in terms of power density and ease of voltage conversion. Their operation under ZVS conditions minimizes switching losses, leading to improved thermal performance and increased reliability [122]. Additionally, as the name suggests, these converters operate at resonant frequencies, effectively attenuating HF noise. Thus, the need for additional filtering components is eliminated, significantly reducing EMI [123].

According to the aforementioned advantages, the utilization of resonant converters is particularly well suited for high-power applications [124]. A key feature of these topologies is their ability to enhance load regulation, as the control over the output voltage is more precise, stable, and adjustable in response to load changes. Furthermore, the attenuation of HF components reduces the voltage waveforms’ THD, thereby improving the power quality. However, achieving optimal performance requires the adoption of suitable modulation techniques, components, and control algorithms, as well as careful design considerations for the resonant tank and other crucial components.

Figure 17. Series resonant isolated DC–DC converters: (a) bidirectional LC DAB; (b) unidirectional LLC; (c) bidirectional CLLC; (d) bidirectional CLC; and (e) bidirectional LCL.

Figure 17. Series resonant isolated DC–DC converters: (a) bidirectional LC DAB; (b) unidirectional LLC; (c) bidirectional CLLC; (d) bidirectional CLC; and (e) bidirectional LCL.

As shown in Figure 18a and mentioned in [125,126], a center-tapped LC series resonant DAB (CT-LC-DAB) converter is proposed to eliminate backflow power by blocking reverse current. Moreover, in the same article, a comparative analysis is conducted with FB-LC-DAB and center-tapped L series resonant DAB (CT-L-DAB) converters, as illustrated in Figure 18b and mentioned in [127]. On one hand, the CT-L-DAB is unable to prevent backflow power, leading to higher conduction losses. On the other hand, the conventional FB-LC-DAB topology incurs significant switching losses due to hard switching. Therefore, the CT-LC-DAB topology emerges as a solution that addresses the limitations of the compared topologies in this article. As added functionality, it also enables operation in both DCM and BCM by implementing suitable modulation techniques.

3.2.5. Other Full-Bridge-Based Topologies

In the literature, conventional isolated DC–DC topologies are sporadically combined with bridge-based structures to obtain enhanced flexibility and efficiency. Therefore, in [128], the utilization of an integrated full-bridge forward converter for battery charging and discharging is proposed. Unlike DAB and conventional forward converters, the proposed configuration incorporates an HFT with three windings, one of which has a center-tapped structure, as shown in Figure 19a. As mentioned in this paper and reinforced in [128], this bidirectional topology exhibits low-input and -output current ripple and high voltage ratio. Another example in the literature is the utilization of a full-bridge/push–pull converter, as shown in Figure 19b [129,130].

As shown in Figure 20, isolated current-fed topologies are also usually documented. Facing voltage-fed topologies, they exhibit shoot-through immunity and low-current stress, making them particularly suitable for low-voltage, high-current applications. In [131], the use of L-type full-bridge, full-bridge push–pull, and L-type half-bridge converters is mentioned. In these topologies, compared to voltage-fed converters, the input current ripple is reduced, as well as N_t. This characteristic is further enhanced in the converters depicted in Figure 20a,b as they employ interleaved variants.

3.3. Multilevel Topologies

The use of multilevel topologies, as extensively studied and widely adopted, is strongly recommendable in high-power scenarios. As the name suggests, in these types of converters, the voltage and/or current are distributed among multiple active semiconductors, which helps to minimize stress on each individual device. Consequently, the reliability of such power solutions is enhanced, while achieving improved conversion efficiency. Therefore, the application of multilevel topologies is highly recommended, not only in medium and high-power scenarios, but also in systems that prioritize power quality, such as grid-tied inverters, EV chargers, and RES interfacing systems.

In addition, multilevel converters are well suited for power electronics solutions demanding increased power density. The generation of multiple voltage levels assists in lowering stress on each active semiconductor and minimizes EMI [132]. This feature is particularly beneficial in EV chargers, as efficient voltage conversion is critical for battery charging and discharging operations. It is intended to extend the battery’s lifecycle by avoiding potential overcharging and deep discharges and ensuring precise voltage regulation. As an additional feature, multilevel topologies are also well suited for the implementation of EV fast-charging stations, which is becoming increasingly important and convenient to users due to the growing concern over range anxiety.

In summary, depending on the system requirements, multilevel topologies offer significant advantages over traditional topologies, including improved efficiency, power quality, reliability, and performance.

3.3.1. Multilevel DAB-Based NPC Converter

One of the most commonly used multilevel configurations is based on the incorporation of neutral point clamped (NPC) topologies. As indicated in [133], they are considered a variant of the DAB, commonly referred to as multilevel DAB (ML-DAB) converters. As illustrated in Figure 21, at least one of the full-bridge stages that compose the traditional DAB converter is replaced by an NPC converter, which generally allows for higher efficiency and the capability to operate higher power levels [134]. For all the documented cases, phase-shift modulation techniques are equally implemented; however, the major difference lies in the number of voltage levels that can be obtained in V_pri and V_sec.

Therefore, Figure 21a illustrates the operation of the ML-DAB converter based on a boost configuration, where the NPC topology is connected to the secondary side of the MFT. With this arrangement, V_pri commonly achieves two voltage levels, while V_sec can attain five levels [135]. Conversely, in a buck configuration, the opposite occurs, with five voltage levels obtained in V_pri as the NPC topology is connected to the primary windings [136]. As shown in Figure 21b, by utilizing a full-bridge converter, two voltage levels are once again obtained, this time regarding V_sec. However, depending on the adopted modulation and topology, it is also possible to obtain three voltage levels in the windings of the MFT.

Furthermore, as studied in [137,138], an NPC converter can potentially be connected on both sides of the MFT. As should be expected, the degrees of freedom increase, which reduces the simplicity of the modulation technique. Conversely, it enhances flexibility and further improves performance. In addition, active switches are not subject to high-voltage stress since they do not need to withstand the total DC bus voltage. As a result, costs are reduced, and the switching and conduction losses are minimized.

Nevertheless, the number of voltage levels may vary depending on the specific NPC topology. As suggested in [139], the ideal scenario is to obtain multiple voltage levels in V_dc1 and V_dc2; however, this would significantly increase the number of degrees of freedom. Obtaining optimal values for each phase-lag angle and ensuring voltage balancing across the DC bus capacitors pose significant challenges. As a result, implementing a ML-DAB converter with multiple degrees of freedom is impractical, despite the significant energy efficiency benefits it would offer. To obtain three (or more) voltage levels, NPC converters based on half- and full-bridge structures are considered, as depicted in Figure 22.

Notwithstanding, the selection of each topology depends on the specific requirements of the power electronics system, leading to varying performance levels. Among the ML-DAB converters based on NPC converters, the diode NPC (DNPC-DAB) and active NPC (ANPC-DAB) topologies are highlighted. These topologies can be implemented in both half-bridge and full-bridge configurations, and, once again, the complexity of the modulation will vary accordingly [140,141].

The main features of DNPC-DAB and ANPC-DAB configurations were compared in [142], highlighting the greater suitability of ANPC-DAB topologies for achieving ZVS and soft-switching. As mentioned, they provided reliable voltage clamping for all switches but also redundant zero switching states. It was also indicated that adopting ANPC-DAB instead of DNPC-DAB results in better loss distribution and balanced power devices.

3.3.2. Modular Multilevel Converter (MMC)

Similarly to multilevel topologies, MMCs are specifically designed for high- and very-high-power solutions [143]. This configuration is considered a distinct type of multilevel converter, where multiple submodules are cascaded to divide a high value of V_dc. Although MMC topologies are commonly used in three-phase systems, Figure 23, for the sake of simplicity, depicts simplified single-phase architectures, both in half-bridge and full-bridge configurations [144]. However, what sets apart each MMC configuration is the chosen topology for the submodules [145]. Considering the aforementioned points, MMCs are recognized as a highly complex configuration that, conversely, provides substantial flexibility, modularity, and scalability. As a result, they allow for multiple arrangements among submodules and even across multiple MMCs, offering numerous possibilities regardless of the power solution. Their scalability allows effortless adaptation to different voltage and power levels by simply adding or removing specific submodules.

According to these characteristics, MMCs are frequently employed in power electronics solutions for energy transmission and distribution, particularly in HVDC transmission, flexible AC transmission systems (FACTSs) [146], and large-scale grid-connected renewable energy systems. Accounting for these application scenarios, MMC structures are highly suitable for high-voltage and high-power solutions due to their enhanced reconfigurability, redundancy, and fault tolerance. Furthermore, they play a crucial role in achieving efficient voltage conversion and mitigating power quality issues, including THD [147]. However, issues related to EMI are also considered [148].

The number of voltage levels produced and applied to the windings of the MFT is determined by the modulation technique and the chosen number of submodules. However, the main challenge encountered in MMCs lies in regulating the voltage across the DC bus of each submodule. Thus, extensive research has been conducted to explore different techniques for effectively stabilizing the submodule voltage, as documented in [149,150,151,152,153]. By precisely controlling this voltage value, the MMC achieves the desired output with increased performance.

To ensure system balance, each submodule is typically constructed with equal topologies. As common practice, depending on the power solution, these submodule topologies can vary from half-bridge structures to full-bridge-based topologies or even isolated stages like the DAB converter, as shown in Figure 24 [154,155,156]. As another example, each submodule can also be based on multiple stages with independent outputs, allowing for the distribution of different feeders. Despite operating as an isolated DC–DC structure, MMCs can also be interconnected with three-phase AC power distribution and transmission systems, fulfilling the role of voltage rectification.

4. Comparative Discussion of Isolated DC–DC Topologies

To reinforce the merits of the previously discussed isolated DC–DC topologies, this section presents a detailed comparative study. This analysis, summarized in Table 2, mainly focuses on the key trade-offs and electrical characteristics of each converter, considering efficiency, cost, complexity, flexibility, and resilience.

The above-mentioned factors are critical for integrating DC–DC converters into RES-based systems and smart grids, guiding the selection of the most appropriate topology for each application. This choice largely depends on the converter’s operational power range, as well as the trade-offs between cost and complexity. Specific applications may require more efficient topologies, aiming to maximize power density, minimize switching losses, enhance waveforms, or meet given power grid integration requirements.

The control complexity of each isolated DC–DC converter is classified in Table 2 as low, medium, and high. Moreover, the table above highlights a potential bibliographic reference for each topology, indicating a relevant study or work where the respective converter has been analyzed or applied. It is also worth noting that the order in which the converters appear aligns with the sequence of illustrations presented throughout this review paper, ensuring consistency between the textual discussion and visual representations. The comparative analysis primarily concerns the physical characteristics of the converters, providing an overview of the number of switching devices (#S), diodes (#D), inductors (#L), capacitors (#C), and transformers (#T) used in each topology.

In this context, large-scale RES-based power systems, such as hydropower plants, wind farms, and utility-scale solar PV parks, demand efficient and scalable conversion solutions. Multilevel topologies, including MMC- and DAB-oriented NPC variants, have emerged as reliable and widely adopted solutions in the renewable energy sector. These topologies inherently offer high-quality waveform generation and enable seamless integration of RESs into power transmission and distribution grids.

In recent years, SSTs have gained prominence as a promising alternative for facilitating the integration of RESs into the power grid, providing advanced and additional interface functionalities. One of their key advantages is the ability to mitigate voltage and frequency fluctuations caused by the intermittent nature of RES generation. By implementing dynamic control mechanisms, SSTs effectively reduce the adverse effects of power variations in wind farms and solar parks, thereby enhancing power quality and improving overall grid stability. Thus, DAB converters and corresponding variants assume a central role in the development of SSTs, offering high efficiency, modularity, and flexibility. Their structure enhances the electrical interface between multiple DC buses, facilitating the direct integration of solar PV systems, ESSs, and DC-native loads into the power grid. Nevertheless, the implementation of advanced control algorithms and phase-shift modulation techniques is the feature that optimizes power transfer in such topologies, improving the performance under various operating conditions. These characteristics appoint DAB converters as a robust and versatile solution, not only for RES applications but also for EV charging systems and hybrid AC/DC power grids.

Table 2. Comparison of isolated DC–DC converters, with references, differentiating between conventional low-power, bridge-based, and multilevel topologies. It presents key design parameters, including the number of switching devices (#S), diodes (#D), inductors (#L), capacitors (#C), and transformers (#T), along with a classification regarding control complexity.

Classification	Topology	Ref.	#S	#D	#L	#C	#T	Control Complexity
Conventional Low-Power Converters	Flyback	[61]	1	1	1	2	1	Low
	Interleaved Flyback with nt = T	[65]	nt	nt	nt	2	nt	Low
	Forward	[68]	1	3	2	2	1	Low
	Active Clamp Forward	[69]	2	2	2	3	1	Low
	Push–Pull	[71]	2	2	1	2	1	Low
	Multi-Output Isolated Zeta	[75]	1	no	1 + no	1 + 2no	1	Medium
	Multi-Output Isolated SEPIC	[75]	1	no	1	2 + no	1	Medium
	Multi-Output Isolated Cuk	[75]	1	no	1 + no	2 + 2no	1	Medium
Bridge-Based Converters	Isolated Asymmetrical Half-Bridge	[80]	2	2	2	3	1	Low
	Single Active Half-Bridge	[81]	2	2	1	4	1	Low
	Dual Half Bridge (DHB)	[82]	4	0	1	4	1	Medium
	Bidirectional Half-Full Bridge	[84]	6	0	1	3	1	Medium
	Uni. Interleaved Full-Half Bridge with nt = T	[86]	4nt	2nt	nt	1 + 3nt	nt	Medium
	Single Active Bridge (SAB)	[87]	4	4	1	2	1	Low
	Dual Active Bridge (DAB)	[90]	8	0	1	2	1	High
	Quad Active Bridge (QAB)	[99]	16	0	4	4	1	High
	Series Resonant LC-DAB	[112]	8	0	1	3	1	High
	Series Resonant LLC-SAB	[114]	4	4	2	3	1	High
	Series Resonant CLLC-DAB	[117]	8	0	2	4	1	High
	Series Resonant CLC-DAB	[119]	8	0	1	4	1	High
	Series Resonant LCL-DAB	[121]	8	0	2	3	1	High
	Series Resonant CT-LC-DAB	[126]	8	0	2	5	1	High
	Series Resonant CT-L-DAB	[127]	8	0	2	3	1	High
	Hybrid Full Bridge-Forward	[128]	5	3	2	2	1	Medium
	Hybrid Full Bridge-Push-Pull	[129]	6	2	2	2	1	Medium
	Current-Fed L-type Full-Bridge	[131]	8	0	2	2	1	High
	Current-Fed Full-Bridge Push-Pull	[131]	8	0	1	2	1	High
	Current-Fed L-type Half-Bridge	[131]	4	0	1	4	1	High
Multilevel (ML) Converters	Boost ML-DAB	[135]	12	4	1	3	1	High
	Buck ML-DAB	[136]	12	4	1	3	1	High
	Three-Level Half-Bridge DNPC-DAB	[142]	8	4	1	4	1	High
	Three-Level Half-Bridge ANPC-DAB	[141]	12	0	1	4	1	High
	Three-Level Full-Bridge DNPC-DAB	[137]	16	8	1	4	1	High
	Three-Level Full-Bridge ANPC-DAB	[138]	24	0	1	4	1	High
	MMC-DHB with nm Half-Bridge Submodule	[155]	2nm	0	3	4 + nm	1	High
	MMC-DHB with nm Full-Bridge Submodule	[156]	4nm	0	3	4 + nm	1	High

Table 2. Comparison of isolated DC–DC converters, with references, differentiating between conventional low-power, bridge-based, and multilevel topologies. It presents key design parameters, including the number of switching devices (#S), diodes (#D), inductors (#L), capacitors (#C), and transformers (#T), along with a classification regarding control complexity.

Classification	Topology	Ref.	#S	#D	#L	#C	#T	Control Complexity
Conventional Low-Power Converters	Flyback	[61]	1	1	1	2	1	Low
	Interleaved Flyback with nt = T	[65]	nt	nt	nt	2	nt	Low
	Forward	[68]	1	3	2	2	1	Low
	Active Clamp Forward	[69]	2	2	2	3	1	Low
	Push–Pull	[71]	2	2	1	2	1	Low
	Multi-Output Isolated Zeta	[75]	1	no	1 + no	1 + 2no	1	Medium
	Multi-Output Isolated SEPIC	[75]	1	no	1	2 + no	1	Medium
	Multi-Output Isolated Cuk	[75]	1	no	1 + no	2 + 2no	1	Medium
Bridge-Based Converters	Isolated Asymmetrical Half-Bridge	[80]	2	2	2	3	1	Low
	Single Active Half-Bridge	[81]	2	2	1	4	1	Low
	Dual Half Bridge (DHB)	[82]	4	0	1	4	1	Medium
	Bidirectional Half-Full Bridge	[84]	6	0	1	3	1	Medium
	Uni. Interleaved Full-Half Bridge with nt = T	[86]	4nt	2nt	nt	1 + 3nt	nt	Medium
	Single Active Bridge (SAB)	[87]	4	4	1	2	1	Low
	Dual Active Bridge (DAB)	[90]	8	0	1	2	1	High
	Quad Active Bridge (QAB)	[99]	16	0	4	4	1	High
	Series Resonant LC-DAB	[112]	8	0	1	3	1	High
	Series Resonant LLC-SAB	[114]	4	4	2	3	1	High
	Series Resonant CLLC-DAB	[117]	8	0	2	4	1	High
	Series Resonant CLC-DAB	[119]	8	0	1	4	1	High
	Series Resonant LCL-DAB	[121]	8	0	2	3	1	High
	Series Resonant CT-LC-DAB	[126]	8	0	2	5	1	High
	Series Resonant CT-L-DAB	[127]	8	0	2	3	1	High
	Hybrid Full Bridge-Forward	[128]	5	3	2	2	1	Medium
	Hybrid Full Bridge-Push-Pull	[129]	6	2	2	2	1	Medium
	Current-Fed L-type Full-Bridge	[131]	8	0	2	2	1	High
	Current-Fed Full-Bridge Push-Pull	[131]	8	0	1	2	1	High
	Current-Fed L-type Half-Bridge	[131]	4	0	1	4	1	High
Multilevel (ML) Converters	Boost ML-DAB	[135]	12	4	1	3	1	High
	Buck ML-DAB	[136]	12	4	1	3	1	High
	Three-Level Half-Bridge DNPC-DAB	[142]	8	4	1	4	1	High
	Three-Level Half-Bridge ANPC-DAB	[141]	12	0	1	4	1	High
	Three-Level Full-Bridge DNPC-DAB	[137]	16	8	1	4	1	High
	Three-Level Full-Bridge ANPC-DAB	[138]	24	0	1	4	1	High
	MMC-DHB with nm Half-Bridge Submodule	[155]	2nm	0	3	4 + nm	1	High
	MMC-DHB with nm Full-Bridge Submodule	[156]	4nm	0	3	4 + nm	1	High

The need for integrating emerging technologies has driven the adoption of multiport isolated DC–DC converters in distributed generation systems. Variants such as the DHB, SAB, and DAB provide highly flexible solutions, enabling optimized control over the magnitude and direction of bidirectional power flow. Additionally, isolated multiport DC–DC converters, which may include the triple active bridge (TAB) and the QAB, are highlighted in the literature as widely employed in microgrid applications, particularly in the integration of hybrid solar PV-ESS.

For microinverters, often used in individual solar PV modules, the need for galvanic isolation is not always a fundamental requirement. However, due to safety concerns and regulatory compliance, conventional low-power topologies, such as flyback and forward active clamp converters, are frequently adopted due to their simplicity, reduced semiconductor count, and low cost. Simultaneously, interleaved topologies have emerged as a relevant solution for specific scenarios where compact design and high-power density are critical requirements. As a result, these topologies are also commonly used in residential cases and standalone solar PV systems, where the efficiency–cost ratio plays an important role in converter selection.

Regardless of the chosen topology and application, the performance of a power conversion stage is heavily influenced by the control algorithms and modulations employed. In bridge-based topologies, in addition to phase-shift techniques, efficiency can be further improved by adopting soft-switching techniques and zero-voltage/zero-current switching (ZVS/ZCS). Complementary to these strategies, advanced control techniques, such as model predictive control (MPC) and direct-quadrature-zero (DQ0) transformation, are widely explored to enhance stability and improve the dynamic response of RES-based systems. For NPC topologies, the use of space vector PWM (SVPWM) has proven to be particularly effective, notably in reducing THD and improving conversion efficiency, especially in high-power applications. Nevertheless, further efforts are required to augment the resilience of power electronics systems, particularly regarding fault tolerance, cybersecurity, switching redundancy, and self-healing algorithms. These advancements are vital to ensuring operational continuity in critical systems, including microgrids and power transmission and distribution grids.

5. Analysis of Power Semiconductor Technologies

In line with the ongoing technological advancements in the power semiconductor industry, particularly in WBG devices, isolated converters no longer necessitate the utilization of bulky components. In other words, as the switching frequency (f_sw) increases, the power density of the converters also rises. Thereby, power electronics solutions employing HFTs enable enhanced controllability and flexibility on the interfacing elements, while significantly reducing the overall volume and weight.

Owing to their superior material properties, WBG semiconductors doped with silicon carbide (SiC) or gallium nitride (GaN) have become increasingly prevalent in power electronics applications [157], particularly in high-frequency and high-power scenarios where traditional silicon (Si) devices are no longer viable. WBG devices exhibit a higher energy bandgap (E_g), improved stability, and reduced conduction losses compared to conventional Si semiconductors, as evidenced by their lower drain-source on-resistance (R_DS(on)) [158,159]. Nonetheless, the increase in E_g also results in a higher electric breakdown field (E_B) [160,161]. Consequently, a larger E_B magnitude enables WBG devices to achieve higher speed under elevated temperature and drain-source voltage (V_DS) conditions. Therefore, SiC and GaN semiconductors have become, respectively, attractive for high-power and HF applications [162,163,164].

As observed in Figure 25a, the use of SiC and GaN semiconductors leads to an extended voltage operating range and enhanced device speed, thereby mitigating the significant limitations of Si doping technology. Furthermore, due to the reduced required charge for switching between turn-on and turn-off states in WBG semiconductors, the switching losses, even at high frequencies, are notably diminished compared to Si devices. However, as highlighted in [165], the long-term reliability of WBG devices is intrinsically linked to the performance of a given power electronics system. Considering the relatively reduced technological maturity of WBG devices, as well as the operating conditions they are subjected to, this topic still demands study and characterization. Critical challenges include high thermal stress, encapsulation degradation, and bond wire failures. Additionally, EMI mitigation mechanisms and the need for robust gate drivers impose additional constraints on WBG device adoption. Addressing these challenges, e.g., the development of failure prevention mechanisms at the semiconductor level, requires the study of different packaging approaches and investigation of high-current avalanche robustness.

Given the critical role of SiC and GaN devices on high-switching-frequency (f_sw) applications, it is crucial to subject these semiconductors to demanding testing under extreme operating conditions. Several reliability tests are explored in the literature, aiming to assess performance and longevity. The major focus lies on key stress factors that impact f_sw, thermal stability, and degradation. Among the methodologies, the high humidity, high temperature, reverse bias (H³TRB) test stands out, particularly in applications where isolation and packaging integrity are critical [166]. Moreover, avalanche capability testing is fundamental for assessing semiconductor robustness against transient overvoltages [167], while radiation reliability and neutron radiation hardness testing ensure performance consistency [168]. On the other hand, backside laser testing allows us to identify weak points at the junction level [169]. Lastly, double-pulse testing plays a pivotal role in analyzing switching behavior, particularly in response to threshold voltage shifts, dynamic R_DS(on) variations, and temperature deviations [170].

In [171], a characterization model for SiC MOSFETs was presented for study under various operating conditions. The main objective was to obtain concrete insights into nonlinear interelectrode capacitances, switching transitions, losses under both hard- and soft-switching conditions, and device-capacitance losses. To this end, a setup with multiple current-probe methods was developed, where three SiC MOSFETs (600 V, 900 V, and 1200 V) were subjected to different conditions. The results showed that semiconductors with higher V_DS have lower on-state resistance and smaller interelectrode capacitance compared to the 600 V devices. Consequently, both switching transients and losses are reduced.

To further enhance reliability, the packaging of WBG devices plays a pivotal role. Specifically, methods aimed at reducing on-state resistance need to be developed, as demonstrated by technologies like direct-cooled power modules with liquid [172] and double-sided direct cooling [173]. Additionally, modules with ultralow on-resistance can also be employed, as proposed in [174]. However, to effectively increase reliability and thermal response, suitable ceramic materials should also be selected for the substrate.

In the case of GaN-dopped WBG devices, this issue becomes even more prominent due to the adoption of higher f_sw. As crucial measures to enhance reliability, emphasis should be placed on reducing parasitic capacitances during layout design and utilizing magnetic materials with variable-width winding structures to mitigate core and winding losses resulting from skin effect [175]. Moreover, winding losses must also be duly considered in optimizing the performance of power electronics systems. To achieve this, over the past years, the use of planar transformers has gained significant importance, particularly in lower power and HF applications [176,177].

As highlighted in [178], to estimate the lifetime of each semiconductor module, the Coffin–Manson–Arrhenius formula is commonly employed:

$$N_f=b_1 \Delta T^{-\alpha } e^{{\displaystyle \frac{E_a}{k_B{T}_{jm}}}}$$

(2)

Here, the number of life cycles to failure, N_f, is determined based on the average operating temperature, T_jm, and the amplitude of thermal cycles, ΔT_j. Additionally, k_B corresponds to the Boltzmann constant, while the coefficients b₁, α, and E_a are dependent on the application scenario and operating conditions.

In this context, an extensive study regarding the reliability of WBG semiconductors under various operation scenarios was conducted in [179]. As an important feature, it mentioned the use of physics-of-failure models to accurately assess their reliability. These models encompass variations to evaluate time-dependent dielectric breakdown, stress–strain, and thermal cycling. The primary conclusion drawn from this article emphasizes the imperative need to transition to power electronics solutions based on WBG devices. Additionally, the repercussions stemming from the challenges of developing devices with enhanced characteristics in bond wires, solders, substrates, and encapsulants are highlighted. As a consequence, WBG devices are limited to an operating temperature of 175 °C, which is still far from sufficient to fully harness their potential.

For the aforementioned reasons, the selection of the semiconductor doping material is highly dependent on the specific power electronics application and its complementary characteristics. SiC technology, for instance, is notorious for its high breakdown voltage (V_BR(DSS)) and capability to operate at high temperatures and voltages, i.e., exhibiting superior thermal conductivity (k_th) compared to other commonly studied technologies [180,181,182]. However, as depicted in Figure 25b and detailed in Table 3, WBG SiC devices have lower electron mobility (µ_e), which may potentially result in slower semiconductor operation. However, this situation is primarily accentuated in low-voltage scenarios, in addition to other parameters that actively compensate for this negative effect. These parameters include the reduced reverse recovery time (t_rr) of the semiconductor’s internal diode, low R_DS(on), high electron saturation velocity (V_nsat), and enhanced E_g [183,184].

Figure 25. Comparison between different semiconductor technologies (Si, SiC, and GaN) regarding the following: (a) operating power and switching frequency (f_sw) range and (b) electrical properties.

Figure 25. Comparison between different semiconductor technologies (Si, SiC, and GaN) regarding the following: (a) operating power and switching frequency (f_sw) range and (b) electrical properties.

GaN doping processes, on the other hand, represent a relatively recent and less mature technology, which explains their limited suitability in high-power scenarios [185]. Notwithstanding, similar to SiC devices, GaN semiconductors exhibit low R_DS(on) values, but, on the contrary, lower k_th, thus indicating a reduced adequacy to dissipate heat. Therefore, in low-current and high-f_sw scenarios, GaN semiconductors demonstrate higher efficiency compared to SiC technology [186]. This improved performance can be attributed to their reduced switching losses and high values of E_B and µ_e, which, in turn, allow for a smaller-sized switching device [187]. Additionally, the higher values of µ_e, V_nsat, and E_g, coupled with lower internal capacitances, enable WBG GaN semiconductors to achieve faster switching transients, making them highly suitable for high-f_sw application [188].

Table 3. Comparison of key electrical and thermal properties of silicon (Si), silicon carbide (SiC), and gallium nitride (GaN) semiconductors.

Property	Si Devices	SiC Devices	GaN Devices
Energy Bandgap (Eg) [eV]	1.1	3.3	3.4
Electric Breakdown Field (EB) [MV/cm]	0.3	2.8	3.3
Thermal Conductivity (kth) [W/cm∙K]	1.5	3.9	1.3
Electron Mobility (µe) [×103 cm2/V∙s]	1.3	0.9	2.0
Electron Saturation Velocity (Vnsat) [×107 cm/s]	1.0	2.2	2.6
Drain-Source On-Resistance (RDS(on))	High	Low	Very Low
Switching Losses	High	Low	Very Low
Maximum Operating Temperature [°C]	150	175–200	175

Table 3. Comparison of key electrical and thermal properties of silicon (Si), silicon carbide (SiC), and gallium nitride (GaN) semiconductors.

Property	Si Devices	SiC Devices	GaN Devices
Energy Bandgap (Eg) [eV]	1.1	3.3	3.4
Electric Breakdown Field (EB) [MV/cm]	0.3	2.8	3.3
Thermal Conductivity (kth) [W/cm∙K]	1.5	3.9	1.3
Electron Mobility (µe) [×103 cm2/V∙s]	1.3	0.9	2.0
Electron Saturation Velocity (Vnsat) [×107 cm/s]	1.0	2.2	2.6
Drain-Source On-Resistance (RDS(on))	High	Low	Very Low
Switching Losses	High	Low	Very Low
Maximum Operating Temperature [°C]	150	175–200	175

Summarizing, based on the analysis of both the images in Figure 25, the use of GaN semiconductors is recommended over SiC when higher speed is required. On the other hand, SiC technology is considered more efficient for power electronics scenarios that require higher voltage and temperature operation. Both scenarios are of utmost importance regarding the implementation of isolated DC–DC converters. Consequently, in line with the premises associated with future smart grids and RES-based systems, these topologies must be endowed with higher power density (increased speed) and the ability to operate at elevated temperatures. More importantly, such behavior is only ensured with the selection of appropriate topologies, relevant modulation techniques, and, as expected, WBG switching devices.

As a practical example, in microinverter configurations, considering the mounting location of the interface system, it is pertinent to exhibit high power density and, at the same time, efficiently dissipate heat. From another perspective, in high-power scenarios, it is expected that the semiconductors will be subjected to higher stress, necessitating improved solar or wind energy conversion efficiency. Nevertheless, beyond their application in RES-based generation technologies, WBG devices also play an equally important role in various application scenarios, such as power systems interfacing with energy storage elements, EV charging stations, and power transmission and distribution, among others.

6. Conclusions

The primordial objective of this review paper is to establish an extensive but accurate study of different isolated DC–DC converters for power generation systems based on renewable energy sources (RESs). Alongside exploring various topologies and arrangements for the power converters, this review also encompasses the possible architectures and connection schemes for interfacing RES-based systems with the utility grid. Furthermore, the analysis has been extended to include emerging technologies and the current deployment scenarios of these converters.

Therefore, a concise yet pertinent analysis has been carried out to determine the most suitable switching power devices for the interface systems. To increase the power density of DC–DC converters utilizing high-frequency transformers (HFTs), there has been considerable interest in employing wide-bandgap (WBG) devices doped with silicon carbide (SiC) or gallium nitride (GaN), with particular prominence in applications such as solid-state transformers (SSTs). SiC semiconductors are strongly suitable for high-power applications due to their superior thermal conductivity, enabling efficient heat dissipation. However, although they are also applicable for high-frequency (HF) switching, they do not possess the same speed characteristics as GaN devices. On the downside, GaN technology is not yet fully suitable for high-power solutions and still demonstrates relatively lower technological maturity.

Although the selection of active semiconductors is a critical step in converter design, the converter’s topology is arguably the characteristic of utmost importance. Isolated DC–DC converters can be categorized not only based on their power handling capability but also considering their scalability, flexibility, modularity, efficiency, cost, and reliability. In addition, they enhance immunity to noise and interference, fault tolerance, and most importantly, ease of voltage conversion.

As expected, multilevel topologies offer enhanced conversion efficiency compared to conventional ones. However, both half-bridge and full-bridge configurations also exhibit highly satisfactory efficiency indicators, which justifies their frequent use in a broader scenario. In addition, these topologies provide considerable technological maturity, thus facilitating the implementation of suitable and innovative modulation techniques. Nevertheless, in high- and very-high-power solutions, multilevel converters, particularly modular multilevel converters (MMCs), are the preferred option due to their capability to distribute a certain voltage and/or current across a larger number of active semiconductors. This scenario is equally observed when multiple converters are interconnected in parallel or cascade configurations, highlighting the significance of modularity and scalability.

Therefore, the selection of each topology will primarily depend on the requirements of the power electronics solution and the specific application. When it comes to interfacing with RES-based generation systems, all the converter topologies discussed in this review paper are considered appropriate. However, certain topologies may be considered more suitable for specific generation technologies or different power levels, ranging from microinverters to large-scale power plants. Among all the isolated DC–DC converters, bridge-based topologies are widely considered as the most consensual due to their flexibility and technological maturity. Furthermore, when interconnected to form a dual active bridge (DAB) converter, the advantages are further accentuated, enabling enhanced scalability and modularity based on the power rating of the application.