Comprehensive Analysis of Microgrids Configurations and Topologies

Abstract

Microgrids have been proposed as a solution to the growing deterioration of traditional electrical power systems and the energy transition towards renewable sources. One of the most important aspects of the efficient operation of a microgrid is its topology, that is, how the components are connected. Some papers have studied microgrid topologies; however, these studies do not perform an exhaustive analysis of the types of topologies, their applications, characteristics, or technical advantages and disadvantages. The contribution of this paper is the integration of the most important functional properties of microgrid topologies in terms of reliability, efficiency, structure, costs, and control methods. The study analyzes 21 topologies divided into six classifications with their respective sub-classifications. The analysis was based on the characteristics of the current (AC or DC), the control mechanisms, the transition between the operating modes, and the operating costs. As a result of the evaluation, it was evidenced that SST-based completely isolated coupled AC topologies, completely isolated two-stage AC decoupled, and multiple microgrids show the best performances. In contrast, the use of two-stage and three-stage partially isolated AC decoupled topologies is not recommended because of their high operating cost and low efficiency and reliability.

1. Introduction

Traditional electric power systems (EPS) are characterized by supplying energy to users from centralized generation systems. Currently, these systems are in crisis due to the predominant use of fossil fuels that cause environmental problems. Another problem in these systems is that the power supplies are located far from the demand centers, causing energy and economic losses [1]. Additionally, the long distances between the generation and consumption centers, the obsolescence of EPS elements, and the growth of electricity demand have increased energy quality problems [2,3]. As a solution to these problems, EPS have recently been divided into a small distributed network, known as a microgrid (MG) [4].

MGs have been defined in various ways by specialized literature. The IEEE Std 2030.8-2018 standard defines an MG as the interconnection of a set of distributed energy resources (DER) and loads that act as a particular controllable entity concerning the EPS [5]. According to the IEC TS 62898-1:2017 standard, an MG is an electrical system with energy resources and loads that act as a controllable entity, able of operating in an island or EPS-connected mode [6]. In [7], the authors defined an MG as a small-scale controlled energy system that can operate in island mode or be connected to the EPS in a defined area. Despite the different descriptions, definitions agree on the characteristics of their modes of operation [8].

MGs have become an option to reduce dependency between consumption centers and EPS [1]. According to [9,10,11], MGs can export and import energy from the EPS and to the EPS using renewable energy sources.

It is expected that in the short- and medium-term the number of MGs may increase due to benefits such as improving power quality and supplying local power when EPS power outages occur [12,13,14]. However, challenges are posed for MGs due to their bi-directional power flows, EPS structure, configuration type, classification, location, and the varying characteristics of some distributed generation units [15,16].

An MG may change according to their topology and configuration [17]. Connection or disconnection of the DER produces different topologies that may cause variations in the current directions and limits [18,19]. Additionally, improper installation and intermittent behavior of DERs may produce some problems related to frequency variation, voltage instability, power losses increasing, and active and reactive power imbalance. Therefore, a control method that guarantees efficient and safe power transfer is essential [20].

Due to the importance of MGs in the current context of changes in EPS, several researchers have conducted studies on their evolution and challenges. In [2], hybrid MGs based on the interconnection of the current (DC or AC) networks and the EPS were reviewed. However, AC and DC topologies were not considered, nor were selection suggestions. In [21], a system of multiple interconnected hybrid MGs was studied; however, it was not detailed in the characteristics of the topology but in the structure used.

In [22], the authors focused on the obstacles to implementing DC MGs, such as standardization and protection schemes. The study argued that before moving towards protection challenges, it is necessary to understand the architecture of the DC MG. The authors briefly described some DC MG topologies with their advantages and disadvantages. In the study presented in [12], three topologies of MGs were studied that intend to adapt to the marine environment, selecting the most suitable one in a land–sea relay fishing net. In both studies, a limited number of topologies were analyzed.

In [23], MG models and strategies based on four dimensions were assessed: goals and modeling metrics, resilience scenarios, modeling approaches, and strategies and topologies. The network topologies used in each dimension were: (a) MGs as virtual feeders for global resilience, (b) dynamic formation of MGs for global resilience, (c) MGs in island mode for local resilience, and (d) MGs for local resilience. These dimensions only represent the centralized control method; therefore, it did not delve into other types of topologies.

Although the selection of the MG topology is one of the most significant aspects for the efficient incorporation of DER in EPS, studies on the subject are based on diverse and limited criteria that do not allow a comprehensive analysis of the types of topologies, their applications, characteristics, or technical advantages and disadvantages.

Due to the limited and scattered information reported on the main characteristics of MG topologies, this article aims to analyze and compare the main topologies presented in various studies. The evaluation is based on the characteristics of the current (AC or DC), the control mechanisms, the transition between the operating modes, and the operating costs, allowing for the assessment of the technical advantages and disadvantages of each topology. This study intended to contribute to establishing criteria that facilitate the design and selection of the appropriate topologies for EPS expansion projects that include the incorporation of MGs with DER.

The contribution of this paper is the integration of the most important functional properties of MG topologies in terms of reliability, efficiency, structure, costs, and control methods. The research carried out is relevant because MGs are part of possible solutions for the energy transition towards renewable sources and to reverse the growing deterioration of EPS.

This paper is organized as follows. Section 2 describes the classifications of MGs and operation modes. Section 3 discusses the topologies of MGs, classifications, advantages, disadvantages, and finally, the conclusion is presented according to the results.

2. Classification of MGs

An MG is classified as a controllable entity made up of loads and DER that can operate connected to the electricity EPS or in island mode under defined electrical limits [24]. DERs can be integrated into wind generators, photovoltaic (PV) panels, microturbines, and other low-power generators, which are located close to the users [25]. Due to the reduction in prices and the increase in the energy conversion efficiency of DERs, developed countries have reached a certain level of mass use of these resources. However, in many less developed countries, difficulties as high electricity prices, obsolescence of EPS elements, and the lack of attractive economic models have prevented the massive penetration of DERs into electricity systems [26].

The general structure of an MG is represented in Figure 1. This network is connected to the EPS through a point of common coupling (PCC) to exchange energy or operate in island mode in case of maintenance or unintentional island mode scenarios [27].

Figure 1 shows that an MG can be represented as a system that integrates a set of loads, DERs units, and energy storage systems (ESS) that allow storing and delivering power. In an MG, the generation units can be selected according to the primary energy available. The main types of generation are PV, wind, hydro, diesel, and hybrid energy [28].

Electricity production based on renewable energies increased from 10% in 2010 to 20% in 2020. This is due to the reduction in the costs of solar and wind technologies and the development of government policies aimed at encouraging the use of these forms of generation [29,30].

Wind generation has not only been proven to be highly profitable, with low operating costs, but it is also adaptable to various places because of abundant and free resources of wind.

Solar generation represents an unlimited resource, which does not generate noise to the population that is near the power plants. It is also highly profitable, and PV cells have also shown a strong increase in efficiency, which allows for greater transformer capacity. Another advantage is that these systems do not have moving parts, nor do they require high maintenance costs [31,32].

MGs have three operation modes. The first mode is connected to the EPS, which handles better stability issues and electricity costs. The second mode is disconnected from the EPS, which converts the EPS into a small grid that supplies the load in emergency or island mode. The third mode is MG shutdown, which is a network security measure to avoid damages to elements [33]. These modes of operation are represented in Figure 2.

According to Figure 2, three control mechanisms are used to change from one operation mode to another. The first control mechanism is disconnection from the EPS, in which the MG can turn the operation to island mode. The second control mechanism is connection to the EPS, which allows the MG to operate with the EPS. Finally, the third mechanism is shutdown control, which turns off the MG and stops operation [33,34].

MGs can be classified according to the criteria shown in Figure 3 considering the electricity demand, the capacity of the system, and the type of circuit (AC/DC) [1,33,35].

According to Figure 3, by considering the electricity demand, MGs can be classified into three types. The first type is the simple MG, which has a single type of distributed generation (DG). The second type is the multi-DG MG composed of several simple MGs. The third type is the utility MG, where loads are prioritized based on user reliability requirements.

The classification according to capacity refers to the type of loads, the power demand, and the area that the MG must supply. According to its extension, the size of the MG defines the availability of the equipment, the operation with the EPS, and the installation and maintenance costs [8].

Table 1. Classification of MGs according to capacity [33].

Type	Capacity (MW)
Simple MG	Less than 2
Corporate MG	Between 2 and 5
Feeder area MG	Between 5 and 20
Substation area MG Independent MG	Greater than 20 Depending on the loads in remote areas (island, mountainous area, or a village).

shows the types of MGs according to their capacity.

According to the type of circuit, the MGs can be classified as AC, DC, and hybrid MGs [36]. This is the most used classification because it considers characteristics of the electric current that is generated, distributed, and consumed [37]. A DC MG has the advantages of storage system integration, higher efficiency because of elimination of DG synchronization, and fewer AC-DC-AC conversions. AC MGs have had a predominance over DC MGs because of the easy transformation of voltage levels with low-frequency transformers, protection, and fault handling. However, AC MGs face challenges with DG timing and reactive power control [38]. On the other hand, studies indicate that 30% of the energy produced in AC is transferred to the DC supply or passes through at least one converter before it is used, a situation that, together with advances in semiconductor technology, allows us to reconsider the implementation of DC MGs [38].

3. MGs Topologies

During the design of an MG, the components and physical arrangement must be considered to achieve a proper transition between the different modes of operation. The connection of the loads, the microgenerators, and the storage elements, require rigorous analysis to obtain the operation and the desired efficiency by the network operator and the user. The way to interconnect all the elements of the network is known as MG topology [39,40].

Topologies can be selected considering the following characteristics [34,41,42]:

• Control mechanisms of the dynamic characteristics of the DGs resources;
• Voltage regulation and frequency for power balance both in island mode and connected to the EPS;
• The transition between operation modes to detect situations that cause changes;
• Economic dispatch to share the load between different DGs;
• Renewable sources are available;
• Minimum impact on the distribution network;
• Coordination between DERs.

The principal classifications of MG topologies are shown in Figure 4. Depending on the type of power supplied, MG topologies are divided into DC, AC, hybrid, and 3-NET [21,43,44]. According to its configuration, MGs are classified into cascade-type and parallel-type MGs.

AC MG systems use the same operating mechanisms as traditional AC power systems, such as frequency, voltage levels, and protection features [45]. DC MGs have been implemented in recent times because of the development of power electronics technology that has increased DC loads and power converters for DC voltage transformation at different levels for different applications [46]. At present, DC MG systems are implemented in special applications such as aviation, automotive, marine, and manufacturing industries; however, an expansion in residential distribution systems is planned [47,48].

Hybrid MG topologies, as seen in Figure 5, can be classified as AC coupling, AC decoupling, and multiple MGs [44].

In the coupled AC topology, the connection to the EPS is made through an AC MG. In the decoupled AC topology, there is not a direct connection between the utility EPS and the AC MG [2]; therefore, it is more expensive. A decoupled MG, in general, is more expensive than coupled MG because higher capacity AC-DC converters are needed [2,49,50]. The multiple MGs’ topology corresponds to a network of several MGs AC or DC that are connected to the high-voltage network and other MGs.

The 3-Net MG topology consists of the union of three different types of networks: a high-quality DC network, a low-quality DC network, and an AC network. This topology makes it possible to supply energy in a single MG to elements of different levels of sensitivity concerning changes in the power quality parameters. An example is rural networks that may have loads such as computers, which are very sensitive to voltage disturbances, along with loads such as water heaters, which are not sensitive to these disturbances. Although 3-Net MGs have high reliability, they represent a higher cost in their development and implementation [39].

According to the configuration, MG topologies can be divided into two categories: parallel type and cascade type. The parallel type has been the subject of numerous investigations and is used in applications that include state-of-charge (SOC) balance for storage systems and obtaining optimal economic distribution schemes for DGs [43]. They also include voltage drop control with a maximum power point tracking (MPPT) regulator of PV systems and an optimal DG distribution scheme.

The cascade type is recent and has a relevant function in applications that require the reliable use of DER at a high-voltage level [51]. Furthermore, it is practical in MGs that operate in island mode, where the energy balance between all modules is fundamental [51,52,53].

3.1. Control Structures of DC and AC MGs

According to the control structures, the AC and DC MGs are divided into two groups as can be seen in Figure 6. The first group is the control methods, and the second group is the load-sharing techniques [54,55].

In the control structures, the power electronic converters are essential components because they must ensure the adequate supply and distribution of the electrical load without affecting the correct operation of the system [55,56].

3.1.1. Control Method

PV Systems

In PV systems the control methods depend on the type of configuration (i.e., centralized inverter, inverter chain, and microinverter configuration).

Centralized Inverter Configuration

The centralized inverter configuration, presented in Figure 7, has one of the highest efficiencies of all PV systems (over 98%). This configuration is estimated to be used in 44% of PV systems installed in the commercial and utility sectors.

This configuration is used mainly in residential as well as in small and medium commercial applications offering reliability [57,58,59]. Additionally, both the implementation and maintenance costs are the lowest of the PV configuration types [57,58,59]. However, damage to the inverter stops the entire system, so constant maintenance is necessary, and available inverters are required for replacement [57,58,59]. Some applications using this topology for AC and DC MGs were assessed in [60,61,62].

Inverter Chain Configuration

The inverter chain represented in Figure 8 has high efficiency (approx. 98%), making it highly profitable. It has power ranges of around 150 kW peak and is estimated to have the highest percentage of use in the market [57,58,59].

This configuration is reliable, which allows it to be used in residential and some commercial applications. Because there are different inverter cell groupings, the damage of one does not prevent the system from generating power continually [57,58,59]. However, this configuration is expensive because it requires an inverter for each group of cells, increasing maintenance costs [57,58,59]. Applications with this configuration were studied in [63,64,65].

Microinverter Configuration

The microinverter configuration, presented in Figure 9, is the least used. In this configuration, the inverter is connected directly to a battery and converts DC to AC.

The batteries are charged by PV cells and are optimized to be used with a single PV panel [57,58,59]. In this configuration, installation costs are low, and parts are easy to replace, ensuring continuous use and fast maintenance. In addition, the panels are independent, which means that the failure of one does not affect the total system [57,58,59]. However, it does not have galvanic isolation between the AC-DC connection, affecting reliability between users [57,58,59]. In [66,67] various implementations of this architecture were assessed.

Wind Systems

In wind systems, the control methods depend on the technology of the generator used (i.e., induction generator, double-winding induction generator, synchronous generator, and permanent-magnet synchronous generator).

Induction Generator

The induction generator configuration (see Figure 10) can be divided into cage rotor and wound rotor. In addition, it can present multiple configurations according to the use of the loads, such as capacitor banks, frequency converters, and starters. This configuration generates electricity, converting the mechanical energy of the movement of the blades into a magnetic field [31,58].

This configuration is built robustly, ensuring its prolonged use. Generally, it does not require additional elements and can easily be used in parallel configurations in spaces already used, such as farms and vacant lots [31,58]. However, this configuration has moving parts that require constant maintenance, take up a lot of space, and generally require high starting torque. Additionally, the power factor of this generator can be reduced to low values [31,58]. This topology was used in [68,69].

Double-Winding Induction Generator

The double-winding induction generator is characterized by having a winding that covers the rotor. In this configuration, represented in Figure 11, the stator fulfills the function of controlling the power flow, while the power is controlled from the connection to the rotor [56,70].

This configuration allows greater power generation without overheating and allows the stator to be connected directly to the EPS [56,70]. However, they require additional equipment to control the frequency of the network [56,70].

Synchronous Generator

This configuration, as shown in Figure 12, comprises a fixed stator with a three-phase wound and a rotor with a magnetic field. They also have multiple subdivisions based on construction, excitation mode, and the parts used [70].

The synchronous generator has a wide range of configurations, making it suitable for multiple situations and locations. An example is the excitation mode, which can be used as power from the EPS or through a capacitor bank. In [70], a comparative study was performed of each type of synchronous generator according to its subdivisions. The disadvantage of this configuration depends on the selected construction. For example, if a variable speed induction generator with a partial power converter is chosen, an energy loss is produced due to the heat in its gears. In [70], an analysis was performed according to its subdivision.

Permanent-Magnet Synchronous Generator

Figure 13 shows the permanent-magnet synchronous generator configuration.

This configuration is generally used in offshore wind turbines [71]. Its operating principle is like the synchronous generators, in terms of the dependence of the speed on the energy consumption of the EPS. Due to their construction characteristics, they generate a large amount of power [72,73]. However, its use is limited because it has more complex control systems and a high maintenance cost [72,73].

3.1.2. Load Sharing Techniques

Load sharing techniques are divided according to the type of control (i.e., centralized control, master and slave control, average load sharing control, and ring control).

Centralized control

In centralized control, the loads and DG units connect via a centralized connection or centralized DC bus. In this configuration presented in Figure 14, the main or centralized control has two functions: to link the MG with the user or with the EPS [74,75].

This configuration presents high efficiency at low costs and only requires a converter for connection to the EPS. Additionally, the electronic elements used are simple, and wiring is cheap and simple [74,75]. However, as the principal connection axis is the centralized control, a failure could isolate the MG and convert it to an inoperable network. Additionally, it is very susceptible to a bad network design [74,75].

In [76], a model to a mini-grid project in sub-Saharan Africa was implemented using a centralized control architecture considering multiple connected nodes to optimize the resource, size, nodal location, and generation dispatch. On the other hand, in [77], the centralized control strategy for AC MG connected to the EPS, taking into account the different DGs, was developed. Other applications that use centralized control were discussed in [78,79].

Master and Slave Control

In the configuration shown in Figure 15, a voltage inverter works as master control and adopts the reactive power control method (PQ). When the system works in island mode, the inverter adopts the voltage–frequency control method (V/F) [54,80,81]. In this configuration, the master unit can be classified into three subcategories: battery energy storage system (BESS), DG, and a combination of BESS-DG [54,80,81]. This type of control allows excellent energy performance and voltage recovery at the MG/EPS PCC [80,81].

According to Figure 15, the complete system depends on the master control. When this mater control fails, the configuration stops working, requiring constant monitoring and maintenance [80,81]. In [82], master–slave wave farm systems and control methods were investigated.

Average Load Sharing Control

In the average load sharing control (see Figure 16), each inverter shares the load regulation process. A common bus related to current is used to calculate the average of the current. Each time a current cycle occurs, the system performs a new calculation to average the load. It has two main configurations: single-inverter and parallel-inverter system [83].

This configuration makes the system reliable due to its ability to distribute the control actions among all inverters. Additionally, it also presents an optimal performance in the current and voltage variables [83]. However, it can be affected by the line impedance effect, producing some power losses [83].

In [84], a distributed control scheme, with current sharing and average voltage regulation in DC MGs, was proposed. The contribution is that the proposed control scheme achieves average voltage regulation without the need for voltage measurements. On the other hand, in [85,86] a hierarchical control strategy was used for DC MGs with DG.

Ring Control

As can be seen in Figure 17, in this strategy control, a loop is created where a module with its inverter is connected in series to another inverter, using a bus that controls the voltage output. Additionally, a connection is made between the last and the first modules to complete the loop [83].

This configuration presents a good response to changes in the system, and it allows for the obtainment of a stable output voltage [83]. However, if an inverter fails, the performance of the system is compromised, leading to the possibility of completely disabling the network. It also presents limitations in the electrical capacity [83,87]. A ring DC MG control architecture was used to manage load balancing and power distribution in [88,89,90].

3.2. Hybrid MG Topologies

Hybrid MG topologies are divided into AC coupled, AC decoupled, and multiple MGs with their respective subcategories. Each of these topologies is discussed below.

3.2.1. Coupled AC

AC coupled MGs can have partially isolated or completely isolated configurations. The characteristics of each configuration are described below.

Partially Isolated Configuration

This configuration is generally used to interconnect several asynchronous AC networks. The most typical applications of these topologies are large-scale PV and wind farms. In this configuration (see Figure 18) the AC MG is connected to the EPS in normal operation mode, and low-capacity AC-DC converters are used to handle the energy flow between the EPS and the DC network [2,49,50].

The AC-DC converter that connects the DC network to the EPS is not behind the transformer; therefore, the nominal power of the transformer reduces because it must conduct the flow of energy from the AC network. Therefore, galvanic isolation does not exist for the DC network unless a second transformer is added [2]. In this configuration, the protection of devices for DC medium voltage (MV) networks is unusual, and their cost is expensive, so it is not often used because MV networks DC are rarely used in MGs [2].

Completely Isolated Configuration

This configuration presented in Figure 19 is divided into three main stages and is more common than the partially isolated configuration. The micro-source stage is the first one, where DG units, ESS, and the DC link are connected. The second one is the combined source, where the inverter and the AC link are located to connect them. Finally, the third stage is the MG, where the low-voltage interconnections and the EPS are carried out [2,49,50].

This configuration compared to conventional configurations strengthens the reliability and flexibility of the distribution network. In addition, its plug-and-play-based system benefits the incorporation of next-generation, charging, or storage devices [2].

A transformer between MGs and the EPS is used at the PCC. It supplies galvanic isolation to the MG and decreases the voltage to create AC and DC low-voltage networks [2]. Although this configuration is suitable for incorporating DG units in the network, many converters are required to connect the DG AC units to the DC network. However, connecting DG units to the network would improve the efficiency of the configuration [2].

3.2.2. Decouple AC

AC decoupled MGs are divided into the following configurations: two-stage completely isolated, two-stage partially isolated, and three-stage partially isolated configuration. The characteristics of each configuration are described below.

Two-Stage Completely Isolated Configuration

In this configuration, shown in Figure 20, a solid-state transformer (SST) is at the input, supplying galvanic isolation to the MG.

This configuration avoids the stability or timing problems that occur when DG units are integrated into the traditional EPS. Additionally, the system works securely with several DG units [2]. However, tight control of interface converters can affect system stability [2,49,50].

Two-Stage Partially Isolated Configuration

The two-stage partially isolated topology (see Figure 21) (as the AC coupled partially isolated topology), is less usual than the other topologies because the protection of the devices for the DC MV network is not as usual, and their cost is relatively high.

This topology presents a simple perspective on the generation of the DC MG regarding the conversion stages. The SST, which is placed in the AC LV network, uniquely guarantees the isolation of this network [2,49,50].

Three-Stage Partially Isolated Configuration

This configuration, represented in Figure 22, provides DC MV and LV networks as well as an AC LV network [2]. Additionally, it employs a DC MT network as the two-stage partially isolated configuration.

The usage of a medium-frequency (MF) transformer in the DC-DC stage supplies galvanic isolation of the LV side of the MG and the size of the devices is extremely reduced [2]. Moreover, the use of an MF transformer makes it suitable for small- or large-scale integration of DG units, ESS, or loads. Nevertheless, it must improve reliability and efficiency while it reduces price and size devices [2,49,50].

3.2.3. Multiple MGs

Multiple MGs are autonomous, independently managed, and operated systems that allow the use of DG units and loads efficiently. Each MG may have some functions or capacity such as excess renewable generation, which could benefit other MGs in the same area and at different times [91].

AC—DC

Figure 23 shows the structure of multiple MGs AC and DC.

Neighboring MGs connected can increase their performance in backup, reliability, economic dispatch, and power quality [43]. However, the energy management of multiple MGs system is a complex problem since the DG units within each multiple MG, the energy exchange between the multiple MGs, and the energy exchange between the EPS and the MGs must be coordinated. The development of multiple MG systems is limited by the current energy management capacity [54,91].

Based on SST

The configuration based on SST (see Figure 24) uses an SST that acts as a power router, which can perform all the functions of a conventional transformer [43].

This configuration can coordinate the power exchange between neighboring AC and DC MGs. It offers high-power quality and allows a reduction in the number of faults [43,92,93]. However, the three-phase unbalance problem can cause adverse impacts on the power quality of the SST [92,93]. Applications that integrated SSTs and DG were studied in [94,95,96].

3.3. 3-NET

This configuration is based on the union of three different networks, as seen in Figure 25. The first one is a high-quality DC network, the second one is a low-quality DC network, and the third one is an AC network connected to the EPS through a PCC. This configuration is essential for situations where there are constant interruptions and concerns about power quality [39,97].

The reliability offered by this configuration is very high. In addition, many loads can be connected to any of the types of networks. For example, it would be feasible to connect a heater to the low-quality DC network and sensitive electronic devices to the high-quality DC network. However, the design and implementation costs of this network are high, and high power losses can be presented [39,97].

Currently, it is necessary to develop flexible MGs capable of operating both connected to the EPS and in island mode. Therefore, studies about MG topologies, architectures, planning, and configurations are necessary. A challenge of this configuration is the need to integrate new technologies of power electronics, telecommunications, generators, and ESS, among others [47].

3.4. Classification by Configuration

The topologies according to their configuration can be classified as cascade and parallel types. The characteristics of each type are discussed below.

3.4.1. Cascade Type

The cascade-type configuration [98,99], also known as series-type configuration, is observed in Figure 26 [100]. This configuration offers an efficient solution using DG in high voltage. It also allows a high-quality power supply due to frequency regulation [52,53]. Generally, most ESS are located in a localized area, where centralized control is in common use [81].

This configuration allows easy increasing the voltage without the use of expensive and huge transformers [51]. One stage for power conversion is required to integrate the low-voltage devices into the MG [53,101]. It is easy and flexible to implement, especially for connecting PV systems to the EPS, MGs, and battery management [101,102]. However, increasing the number of cascaded MGs makes it more difficult to install the ESS in a localized area. The cost of communication through a sophisticated and expensive centralized control depends on a large bandwidth which becomes a complex situation [51].

3.4.2. Parallel Type

The parallel-type configuration [103,104] (see Figure 27) uses droop control, does not require physical communication links, and is easy to implement [62]. However, all local controllers need to communicate with a central controller, which weakens the reliability and scalability of the system if adequate control strategies are not adopted [53,105].

When the MG is connected to the EPS, DG units are controlled as current sources to maintain operating with the EPS [53]. In island mode, the decentralized control methods, which mainly include droop control and its variants, are used to achieve frequency synchronization and power-sharing [53]. Additionally, it allows increasing the power capacity and efficiency of the system and reduces the ripple of the output current [105]. However, a two-stage DC/AC power conversion is required to integrate the low-voltage devices into the MG [53,101]. The accuracy of the power distribution is very sensitive to the output impedance of the inverters. The harmonic power in the case of non-linear loads is poorly compensated, and not connecting the inverters at the same time to the bus affects the power distribution [106].

3.5. Summary of the MG Characteristics

Table 2. Summary of the MG characteristics.

Topology	Advantage	Disadvantage	Application	Reference
Centralized inverter	Reliability. Widely used. Low cost.	Depends on the power inverter. Constant maintenance	Commercial: MV and HV.	[57,58,59,60,61,62,83]
Inverter chain	Efficiency. Peak power. Reliability.	High cost.	Residential. Commercial: LV, MV.	[57,58,59,63,64,65,83]
Microinverter	Maintenance. Cost. Independent panels.	Non-galvanic isolation in ac-dc connection	Residential. Commercial: LV.	[57,58,59,66,67,83]
Induction generator	Robust. Additional electronics. Easy parallel connection. Installation area.	Constant maintenance. Installation area. Starting torque is required. Power factor	High and low power. Variable speed generators.	[31,58,68,69]
Double-winding induction generator	Power. No overheating. Stator connection.	Load dependency. Network frequency	High and low power. Variable speed generators.	[56,70]
Synchronous generator	High flexibility. Wide range of configurations.	Subject to the generator selected	Industrial. Generating stations.	[56,70]
Permanent-magnet synchronous generator	Power.	Control system’s maintenance cost	Offshore wind turbines Generator stations	[71,72,73]
Centralized control	Efficient. Cost. Use one power converter. Simple electronics.	Depends on the converter Susceptible to design.	Residential. Commercial: LV and MV.	[74,75,76,77,78,79]
Master–slave control	Performance. Voltage in the PCC.	Constant monitoring. Constant maintenance. Depends on the master control.	Commercial: LV and HV. Power storage in capacitors.	[54,80,81,82]
Average load sharing	Reliability. Performance	Impedance effect. Power losses.	Parallel power systems. Power storage in capacitors.	[83,84,85,86]
Ring control (3C)	Response time. Robust. Stability.	Electrical capacity. A faulty inverter disables the network completely.	Parallel power systems. Power storage in capacitors. Uninterrupted systems.	[83,87,88,89,90,107]
Partially isolated coupled AC	Size of the power converter	Uncommon. Costs.	Interconnection of several AC networks. Remote or large-scale wind or PV farms.	[2,49,50]
Completely isolated coupled AC	Flexibility. Reliability. Plug-and-play system. Galvanic isolation.	Number of power converters	Integration of DG units into the EPS.	[2,49,50]
Two-stage completely isolated decouple AC	Galvanic isolation. Stability. Synchronization Reliability.	Regulation of power converters.	Integration of DG units into the ESPS	[2,49,50]
Two-stage partially isolated decouple AC	SST conversion stages	Uncommon. Costs.	Integration of DG units into the EPS.	[2,49,50]
Three-stage partially isolated decouple AC	MF transformer.	Efficiency. Reliability.	MV, LV networks. Small- or large-scale integration of DG, ESS, or loads.	[2,49,50]
Multiple AC—DC	Power backup. Reliability. Power quality.	Energy management. Energy exchange between MGs.	LV networks	[54,91]
SST	Power quality. The number of failures. The power is exchanged between neighboring AC-DC MGs.	Three-phase imbalance.	Integration of DG units into the EPS.	[43,54,91,92,93,94,95,96]
3-NET	Efficiency. Reliability. The number of loads.	Cost. Loss of power demand.	Residential and Commercial: LV and MV networks.	[39,47,97]
Cascade type	Electricity supply. Frequency control implementation.	The number of MGs. Storage units. Communication. Bandwidth.	High-voltage applications. Energy storage applications. PV applications. Battery management.	[51,52,53,81,98,99,100,101,102]
Parallel type	Power capacity. Efficiency. The ripple of the output current. Reliability.	Power share.	Frequency synchronization.	[53,101,103,104,105,106]

summarizes each topology assessed, references, advantages, disadvantages, and application.

During the selection of the topologies, the mechanisms to control the dynamic characteristics of the DG resources, the frequency and voltage regulation, the energy balance, the transition between the operation modes, and the economic dispatch must be considered. These characteristics allow the EPS to supply the energy demanded by the loads in a reliable and coordinated manner with minimal impact on the distribution networks.

Based on the analysis performed of the topologies, it can be inferred that hybrid MG topologies, such as SST-based completely isolated coupled AC, two-stage completely isolated decoupled AC, and multi-MGs, have proven to be feasible to operate in combination with the EPS. In contrast, two- and three-stage partially isolated AC decoupled topologies still have a challenge in terms of the need to reduce operating costs, and improve efficiency and reliability, for this reason, this topology is rarely used.

4. Conclusions

Currently, EPS face challenges due to the obsolescence of their components and the need to carry out an energy transition towards renewable sources. One way to face these problems has been the development of MGs that promise to be a good alternative solution due to their flexibility, efficiency, and capacity for dynamic operation in conjunction with EPS.

The form of connection of the energy sources and the loads, as well as the possible connection alternatives between the MGs themselves and the EPS, constitute one of the most important aspects in the operation of the MGs. These forms of connection are called topologies, and it is one of the most relevant research fields within the current conjuncture of the development of new forms of management of EPS.

In this study, 21 MG topologies were analyzed, taking as reference the characteristics of the current (AC or DC), the control mechanisms, the transition between the operating modes, and the operating costs.

As a result of the evaluation, it was evidenced that the use of hybrid MG topologies is recommended, specifically, the SST-based completely isolated coupled AC, the two-stage completely isolated decoupled AC, and multiple MGs. These topologies have the advantage that they are feasible for the incorporation of DG units in the EPS and for improving the flexibility and reliability of the distribution network.

It was also observed that two -and three-stage partially isolated AC decoupled topologies have high operating costs and have low efficiency and reliability; therefore, they should continue to be improved.