Single-Phase Microgrid Power Quality Enhancement Strategies: A Comprehensive Review

Abstract

Renewable distributed generators (RDGs) have made inroads in recent power systems owing to the environmental effect of traditional generators and their high consumption of electric energy. The widespread use of RDGs has been a recent trend in numerous nations. The integration complexity and the intermittent nature of RDGs can undermine the security and stability of microgrids (µGs). In order to guarantee the effectiveness, dependability, and quality of the electricity delivered, appropriate control methods are necessary. RDGs are being included in single-phase microgrids (1Ø-µGs) to generate energy closer to the user. The creation of low-voltage µGs allows for increased energy efficiency and improved electrical supply dependability. Nevertheless, the combined power pumped by DGs might create power quality (PQ) difficulties, especially during off-grid operations. The three biggest problems with PQ are reactive-power swapping, voltage and frequency (VαF) variations, and current and voltage (IαV) harmonic falsification associated with 1Ø-µGs; these conditions may affect the operation of µGs. The designed and implemented (primary–secondary control systems) in RDGs are the prevalent strategy discussed in the literature for mitigating these PQ difficulties. Furthermore, emerging grid innovations like the electrical spring offer viable alternatives that might reduce some problems through decentralized operation. Although several research studies have addressed PQ concerns in 3Ø-µGs, not all of these solutions are immediately applicable to their 1Ø equivalents. In this paper, the state of the art and a performance comparison of several PQ enhancement strategies of µGs is discussed. Additionally, the primary difficulties and several PQ approach tactics are highlighted. All vital features from high-quality published articles and new dimensions in this field are presented for mitigating PQ difficulties in 1Ø-µGs.

1. Introduction

Single-phase microgrids (1Ø-µGs) have recently received significant consideration as an alternative solution to provide a reliable and sustainable power supply to remote and isolated communities [1,2,3,4,5,6]. These µGs are typically designed to operate autonomously from the main power grid, making them an ideal solution for areas with limited access to conventional grid infrastructure. However, power quality (PQ) issues have emerged as the primary challenge that undermines the efficiency and reliability of 1Ø-µGs. These issues occur due to the intermittent nature of renewable energy sources (RESs), voltage and frequency (VαF) fluctuations, and harmonic distortions (HDs) [7,8,9,10,11,12].

The PQ issues in 1Ø-µGs can be classified into three categories: voltage stability (VS), frequency stability (FS), and HD [13]. VS refers to the ability of the µGs to maintain a stable voltage level under varying load conditions [14,15]. It is a critical factor that influences the efficiency and reliability of a µG. Voltage instability can cause equipment damage, power interruptions, and system instability [16,17,18]. Similarly, FS refers to the ability of a µG to maintain a stable frequency level under varying load conditions. An underachieving FS can cause equipment damage and system instability. HD is another significant PQ issue in 1Ø-µGs. It refers to the distortion of sinusoidal waveforms due to the presence of stochastic/nonlinear loads, like electronic devices and converters. HD can cause equipment damage, electromagnetic interference, and system instability. Therefore, mitigating PQ issues is critical to ensuring the proper functioning of 1Ø-µGs [19,20,21,22].

Phasor measurement units (PMUs) offer time-synchronized signals from numerous µG positions; these signals are especially important when there is increased installation of DGs [23]. By monitoring the power system’s variables with PMUs, a µG can promptly identify any grid instabilities. Only SCADA and EMS information were previously used to track the P–V and Q–V curves, but earlier studies created a model-free approach to detect the voltage amplitude. Based on [24], PQ can also be handled by PMUs. To examine any variations in V and F at the structure, real-time information from the PMU can be recorded for a set amount of time [24,25]. PMUs can gather V and F for the PQ analysis of information [24,25]. A minimization model is required to choose PMUs at tactical bus positions before conducting any investigation with PMUs [26,27,28]. In order to find the best location for the PMU, Ref. [28] presented recursive quadratic programming (RQP) with fewer limitations on equality than the design variables specified on an intimate set. According to the findings, RQP identifies the ideal PMU positions with the smallest number necessary for rendering the power system completely visible. A notable enhancement in the visibility over the estimation of redundancies produced by the RQP optimization approach, originally reported in [27], was demonstrated through numerical experiments on common benchmark power networks. Furthermore, study directions and applications of PMUs were presented in [25,26,29]. The setup and analysis of PQ using a PMU situated at a sensitive load were covered in Ref. [24]. The data from the PMU were assessed and identified anomalies in the V and F readings from the sensitive loads’ PQ limitations.

Several mitigation strategies (energy storage systems (ESSs), VαF regulation, and active power (P) filtering) can be implemented to address PQ issues in 1Ø-µGs [30]. ESSs can be integrated into a µG to store excess power generated during peak hours and supply it during low-demand periods. This reduces the impact of power fluctuations on the VαF stability of the µG [31,32,33,34]. Voltage regulation (VR) techniques, such as tap-changing transformers, can be employed to maintain a stable voltage level under varying load conditions. VR techniques can effectively mitigate the voltage fluctuations in a µG [35,36]. Frequency regulation (FR) techniques, such as load shedding and using a dedicated frequency control system, can be employed to maintain a stable frequency level under varying load conditions. These techniques can effectively mitigate frequency instability in a µG [37]. Active power filters can be employed to reduce the HD caused by nonlinear loads. These filters act as a current source and inject a pay-off current into the system to offset harmonic distortions [38].

The decentralization of energy production is a contemporary phenomenon resulting from the widespread interaction of distributed generators (DGs). RESs and ESSs are typically part of low-voltage (LV) systems since they generate electricity closer to the user. The construction of 3Ø and 1Ø µGs that exploit energy utilization and enhance the resiliency of the community power grid has been made possible by the coordinated operation of generation and load. Nevertheless, problems are caused by the increase in converters inside the LV systems, causing numerous PQ problems during on-grid and off-grid operations [5,39,40]. In 1Ø-µGs, VαF variations, reactive power (Q) exchange, and voltage/current (V/I) HD are PQ concerns. Recent attention has been drawn to the PQ difficulties in off-grid operations since the consequences of these phenomena are amplified owing to the absence of grid rigidity [41,42,43].

The VαF deviations ($\Delta \mathsf{{\rm E}}$ and $\Delta \mathsf{\omega }$, respectively) in to the droop control (DC) algorithm are required for demand-side management (DCM) [44,45]. The deviations’ steady-state records are determined by a variety of variables, together with the DC gains, the impedance’s load, and the injected P by the grid-following and -forming inverters linked to the µG. VαF variations are thus inevitable under regular events and are exacerbated by changes in loads. Additionally, the local voltage amplitudes regulate the DG units’ Q production. The effects of unbalanced line impedances and inadequate DGs output impedances result in Q exchange between DGs [46], which lowers the maximum output P and increases the loss of power in the system [47].

In 1Ø-µGs, V/I harmonics are an essential PQ problem. HD increases power losses and may result in µG stability issues, especially in off-grid µGs. Due to poorly constructed control loops, DG units may introduce current harmonics. In addition, the harmonic I that DGs pump into the grid is known to rise with V distortion at the point of common coupling (POCC). Moreover, DGs create V-HD at the PCC as a result of harmonic currents needed by local nonlinear loads [46]. µGs with 1Ø may also use traditional passive/active filters for selective harmonic frequency adjustment [48,49,50,51]. Unfortunately, the mentioned filters raise the system’s price and provide no further benefits to the µG except harmonic compensation. Moreover, these filters provide new resonant modes that might affect the µGs’ stability by amplifying certain harmonic frequencies.

The hierarchical control architecture (HCA) also coordinates the activities of the grid-following and -forming units in 1Ø-µGs. The µG-HCA is separated into 3 sections, with diminishing frequencies as one moves up the hierarchy, as depicted in Figure 1 [52,53,54,55,56,57,58]. Typically, control algorithms that address particular PQ issues are implemented in one or a mix of these levels. Since the tertiary level is often not linked with the functioning of 1Ø-µGs, only the primary control (PC) and secondary control (SC) mechanisms for DGs are analyzed in this review.

The primary DC and inside control loops (CLs) of the DGs comprise the first level. Using ($P$-$F$) and ($Q$–$V$), the DC method enables truly decentralized operation [26]. Traditional DC has been extensively discussed in [53,59,60,61,62,63,64]. The secondary level can be constructed as either a centralized HCA [65,66] or a distributed HCA [67,68] to optimize various operational characteristics of the µG. In the centralized architecture, SCLs are implemented in a central µG controller, whereas in the dispersed construction, the DG is involved in statistics to attain a shared aim. In contrast, power-based control (PBC) with a buckle-down accommodating strategy has been discussed [69,70,71,72] for addressing PQ problems in µGs. The PBC is situated between the classic DC and optimal control methods, which require comprehensive grid models. Nevertheless, the PBC method has primarily been explored for systems in 3Ø settings.

The 1Ø-µGs could benefit from an alternative solution provided by emerging grid technologies. Electric springs (ESs) were proposed as a substitute to traditional load voltage controllers (LVCs), which are fully described in [73,74,75,76,77,78]. Alternatively, in LV systems with RESs, ESs were coupled in series with nonessential loads to deliver V stabilization via essential loads [77]. In addition, ESs reduce the need for communication tools in DCM [77,79,80]. The configuration created by the ES and the nonessential load is usually denoted as a smart load (SL) [81,82,83]. In [84], the operating principle of ESs built on Hooke’s law was presented. Simply put, through the absorption/supply of any excess/deficit of P in the local µG, the function of the ESs minimizes any V variations diagonally the crucial load. In [81], a comparison was made between the operation of several simultaneous ESs and a STATCOM. This comparison showed that a group of ESs manages VR better than a STATCOM with less Q. In [85], the authors demonstrated that one can reduce the size of the battery by integrating ESs into µGs. However, this is only possible if the ESs themselves lack ESS. Otherwise, the ESs could be used to decentralize battery rather than reduce ESS needs. Recent literature reviews of µG control and management algorithms can be found in [86,87,88,89,90,91,92,93,94,95]. Nevertheless, solutions designed for 3Ø-µGs are not always applicable to 1Ø-µGs. In this paper, we present an analysis of the current difficulties associated with PQ in 1Ø-µGs and propose mitigation strategies to address the problem. Hence, the following literature review focuses on methodologies and algorithms suitable for 1Ø-µG

The rest of this paper is organized as follows: Section 2 offers a quick overview of the PQ issues discussed in this work. Section 3 and Section 4 discuss the HCA approaches for improving PQ. Section 5 discusses the emerging technology of ESs for improving PQ in 1Ø-µGs, while Section 6 compares and contrasts the various algorithms and methods. Future research directions are presented in Section 7. The conclusion of this review is presented in Section 8.

2. Difficulties with PQ in 1Ø-µGs

Multiple DGs with inverter interfaces that provide PαQ to nearby loads make up a µG. PQ issues with off-grid 1Ø-µGs are made worse by the grid’s shortage of rigidity. Here, voltage/current HD, Q exchange, and VαF fluctuations are PQ issues. It is well known that electrical networks’ current harmonics are primarily caused by power-electronic systems. Due to the nonlinear voltage dips caused by current harmonics in the distribution network, voltage HD is created at the equipment’s PCC. The harmonics in VαI also interfere and increase losses [96,97].

The fragile equilibrium between production and demand is maintained by the DGs under control of the PαQ and VαF fluctuations; departures from nominal values are what cause fluctuations. VαF transitions happen when the dynamic balance changes as a result of changing loads. Depending on the DG’s ability to react to load fluctuations, oscillations may result. Off-grid operations experience considerable frequency variations; whole loads can be powered by inertia-free, converter-driven machines. Furthermore, an excessive P fed by DGs may set in motion overvoltage, which needs to be regulated to ensure the adjoining critical loads are met [98,99].

The Q generated by the DG units during off-grid service does not help with power transfer, despite the fact that certain loads and the PCC’s need for VR demand it. Q transmission among the converters is made possible by the differential converters’ output voltages and load-dependent voltage fluctuations. In order to improve the operation of a µG, the Q produced by the DGs need to be reduced [100,101].

Typically, distribution systems function in radial mode, which involves radially connecting most of them. Others might have feeders that close the loops, but, generally, open switches maintain the loops open, and the loops only close when other components of the loops become open because of failures. This maintains this structure, and the system’s safety features are built to work in a radial fashion. On the other hand, because of the DG connections at various sites, power flow within the µGs can be bidirectional. As a result, the µGs are not adequately protected by the standard protective measures [102,103,104].

The limited fault current capability of the converter devices within a µG is another issue. Except in situations when they are expressly engineered to generate a large fault current, the amount is often less than 50% of the rated current. When compared to utility DGs, the fault current provided by the µ sources is less in this situation. If enough µ sources have converter interfaces, switching from grid-tied to separate functioning would significantly reduce the amount of µG faults. As a result, the structure’s overcurrent relays’ sensitivity and functionality are impacted. If the relays are configured for larger fault currents in a grid-tied execution, the separate mode causes identical relays to work extremely slowly or not at all because of decreased fault currents. There are several different strategies for protecting µGs that have been developed and documented in the literature, such as adaptive protection, differential protection, distance protection, voltage-based protection, deployment of external systems, overcurrent, and symmetrical components [105,106,107].

3. HCA for 1Ø-µGs

The investigated µGs have also adopted a hierarchical architecture to increase the performance of droop control. These 1Ø-µGs networks can implement the following solutions to address power quality issues: elimination of Q exchange between DG units, regulation of VαF fluctuations, and alleviating VαI harmonics [108,109,110,111,112,113,114,115,116]. Various strategies are discussed in the next subsections to show the topic’s significance.

3.1. PC Loops (PCLs)

Traditional DC performs poorly in 1Ø-µGs because these networks are mostly resistive. The physical PαQ appearance of DGs with an output filter inductor–capacitor (LC) is connected to these types of µGs [117]. This coupling results in a finding of the middle ground between VR and power-sharing precision [118]. Hence, the voltage (V) variations, frequency (F) oscillations, and Q issues of DGs are present in µGs. In [119,120,121,122], inverse/reverse droops for primarily resistive µGs were proposed. For resistive networks, the DC laws are stated as:

$$F=F^{*}-mQ$$

(1)

and

$$\mathrm{V}=V^{*}-nP$$

(2)

where $m$ and $n$ represent the DC gains.

However, reverse DC gains also result in V/F fluctuations and high Q exchange between DGs. Moreover, the inverse droops prevent the straight linking of synchronous generators (SGs) to the µG. Following the installation of a supplementary L at the exit of the LC filter of the DGs, researchers [123,124] showed that standard droops may be employed in LV-Gs with acceptable performance after this. The large gains in traditional DC minimize the Q exchange between DGs, as proposed in [125,126,127]. The VαF oscillations are greatly exacerbated by high DC gain increases, particularly when the load is changed [117].

Selected harmonics compensation approaches incorporated in the DGs’ PCLs have also been suggested as a means of reducing the VαI harmonic infusion by DGs in LV-µGs. In [128], an $h$th harmonic DC that dampens harmonic VsαIs was investigated. In essence, this technique raises the harmonic frequencies of the traditional DC. The level of complexity of the DC gain structure of the inverters is increased by the requirement to add several harmonic DCs according to the chosen harmonics.

Alternatively, linear harmonic controllers (HCs), proposed in [116,129,130,131], were utilized to alleviate VαI harmonics in µGs. The applied controllers were proportional resonant (PR) with low impedance at the harmonic F (HF) under consideration. Ref. [116] examined selective HC using PR controllers in 1Ø-µGs. The inner VαI loops are controlled by PR controllers with selective HC, as depicted in Figure 2. Each controller’s transfer functions (TFs) are stated as follows:

$$G_{I,V}\left(s\right)=k_{pI,V}+{\displaystyle \sum }_{h=1,3,5,7}\frac{k_{iI, Vh}s}{s^2+{\omega }_{cI, Vh}s+{\omega }_h^2}$$

(3)

where ($k_{pI,V}$, $k_{iI, Vh}$), ${\omega }_{cI, Vh}$, and ${\omega }_h$ are the proportional and resonant gains at the HF, which regulates the bandwidth at any HF, and the resonant F, respectively.

The low output impedance of the DGs as a result of the selective HC by the PR controllers reduces the VαI of the harmonics in µGs. Nevertheless, since external factors might affect the DGs’ terminal impedance, PR controller functionality may suffer.

3.2. Virtual Impedance (VI) for PQ Enhancement

Other PC approaches suggested for improving the PQ of 1Ø-µGs are built on the impression of VI loops (VILs). VILs can be implemented in both 1Ø and 3Ø-µGs. R or L, RL, and RC are the emulated VI in [116,132,133,134]. As shown in Figure 3, the VIL is unified into the PCLs of the DGs. Multiplying the VI’s $Z_V$ by the output I ($i_o$) and additional to the desired V from the DC loop $\left(V_{ref}\right)$ to attain.

$$V_c=V_{ref}-i_o{Z}_V$$

(4)

where $V_c$ is the wanted V output.

The R/L of VIL helps improve I allotment though minimizing steady-state VαF disparities. This suggests that the Q interchange between DGs has also been boosted. RL-VIL was proposed in [132,133,134], allowing for multiple output impedance options. This strategy was proposed to minimize the exchange of Q. Via a preservative, the virtual reactance, and a varying virtual resistance, the Q output was adjusted [134]. However, RL-VILs can diminish the efficacy of PR controllers by reducing resultant harmonic I and can increase V-HD, as reported in [116].

In [113,116,135], an RC-VIL that decreases the harmonic I by the DGs and the V-HD was suggested. As depicted in Figure 4, the C-VILs simulate a virtual C bank linked in series with the converter’s terminals [136,137]. In this conceptual diagram, each C is a bandpass filter and a C impedance tuned to a separate HF. For compensation of the third, fifth, and seventh harmonics, the TF Zv(s) for the simulated C bank is expressed as:

$$Z_d\left(s\right)=R_V-{\displaystyle \sum }_{h=1,3,5,7}\frac{{\omega }_{ch}(k_{ph}s+k_{ih})}{s^2+{\omega }_{ch}s+{\omega }_h^2}$$

(5)

where $k_{ph}$ represents proportional gains, and $k_{ih}$ represents integral gains.

As shown in Figure 5, an adaptive negative virtual harmonic impedance (NVH-Z) loop to reduce I’s harmonic in µGs was investigated [138]. A fast Fourier transform (FFT) function was applied to obtain the fundamental current and harmonic components$I_o$ and$I_h$. The harmonic component and fundamental component were multiplied by$Z_h$, and $Z_V$, respectively. The $Z_h$ value was adjusted by an algorithm for controlling harmonic droop, which is defined by the next equation [132,133]:

$$Z_h=b(H_o-H)$$

(6)

where b represents the $Z_h-H$ droop coefficient, $H_{o }$represents the harmonic variance capacity, and H represents the real harmonic produced power.

The resultant desired V is then fed into the VαI controllers.

4. Centralized and Distributed SCs

Since PCLs can only measure local variables, they cannot successfully enhance the PQ of µGs. Typically, VIs are designed with a known µG impedance [133]. However, in an off-grid µG, this impedance changes dynamically due to load switching. To achieve this advanced functionality, SC loops (SCLs) are required [118].

In [118,139], the authors proposed an HCA in which the µG central controller (µGCC) regulates the µG’s VαF. The µGCC restores VαF by cascading the CLs of the DGs in the µG with PI control. Since SCLs are not essential to the operation of the µG and serve only to optimize its operation, only a low-bandwidth communication (LBC) connection is essential. However, because the Q is not accounted for in the choice of V compensation factors, these SCLs may increase the Q exchange flanked by DGs. In [140], Q sharing in islanded µGs using adaptive V-DC [141] and SCLs that eliminate Q exchange between DGs in 1Ø-µGs and simultaneously regulate VαF were examined. In [136,142], a Q-sharing block (SB) was added to the µGCC to eliminate the Q exchange between the inverters. As shown in Figure 6, the Q-SB was used with the VR loop. The µGCC determined the Q demand for each inverter, enabling per-unit Q sharing. In this instance, LBCs were applied to spread the pertinent work. In [140], additionally, the process took the limited quantity of power that could be obtained from the power source into consideration.

An adaptive DC loop (ADCL), which eliminates Q exchange among DGs and simultaneously controls VαF, was proposed in [141]. Figure 7 depicts the ADCL’s block diagram. A central/fundamental controller of the µG determines the Q* for each DG. Then, the initial DC gain n is supplemented with a small departure from the integrated controller’s DC gain. The magnitude of the output (E) is calculated using the suggested DC parameter n*. The ADCL’s operating is comparable to that described in [136]. In the latter, SCLs were fed a V deviation that was additional to the PCLs; whereas in the former, DC parameters were adjusted to realize identical V deviation.

In [46,133], the PCC’s voltage HC was also taken into account. The authors in [46] proposed the HCA depicted in Figure 8, which reduces Q exchange and regulates VαF. The value and polarities of the voltage harmonics are extracted by a large number of second-order generalized integrators and diffused to the µGCC. To calculate the required harmonic correction by the DGs at the µGCC, a controller is employed for each harmonic.

The local DG then produces the compensating voltage, which is also managed by the PCLs. In [133], combining VIs with SCs was investigated. The researchers suggested that instead of a proportionate gain, the voltage harmonics are altered by VI in the DGs before being fed into the PCLs.

5. ESs in 1Ø-µGs

By virtue of their DSM capabilities and possible additional services, the ESs deployed in 1Ø-µGs may offer numerous benefits. Recent applications of ESs include VαF variations mitigation via DSM (PαQ) [79,82,143], power factor (PF) correction [144,145,146], and harmonic compensation [145]. Furthermore, the cooperative activity of several ESs was investigated in [147,148,149], because the parallel working of several SLs without coordination may find the balance in a µG’s solidity.

5.1. Alleviating of VαF Variations

The primary ES configuration was covered in [150], and it has recently been referred to as ES-1 [151,152,153]. As illustrated in Figure 9a, ES-1 is linked in series with $Z_{nc}$ and is made up of a half-/full-bridge 1Ø converter with a capacitor. The ES-1 in [143] operates in three different ways that are reliant on the accessibility and need for electrical P in the area. $V_{pcc}$ is within the usual working range when the power production and load need are equal; hence, the ES-1 does not take part in DCM. It increases or decreases based on the relationship between power production and load request. An L/C voltage ($V_{es}$) is fed by ES-1 to keep the desired $V_{pcc}$ via the control of $V_{nc}$. The ES-1’s functioning causes the SL to provide restricted Q compensation, which controls the µG voltage, but it suffers from $V_{es}$ and should be 90° leading/lagging $i_{SL}$.

The ES-1’s PCL comprises two closed CLs that regulate the value and phase angle (PA) of $v_{ES}$ [143]. Figure 10 illustrates the PCL’s configuration of the ES-1. The value of |$v_{ES}$| directly controls the Q output of the ES-1, whereas the PA ($\angle v_{ES}$= $q_{ES}$) between $v_{ES}$ and $i_{SL}$ controls the used P. For minimizing the ES-1 P’s input/output, the 2CLs regulate the DC-link voltage. Ideally, it should not matter if some P flow is required to control the V’s DC link (DCL). Additionally, the ES-1’s Q compensation capabilities are constrained by the stored energy in the DCL capacitor.

For providing resident V support, ESs with DCM capabilities could also provide F regulation. Ref. [82] proposed an algorithm that enables an ES-1 to control the primary F. Figure 11 is a block diagram of this CL, which regulates the value and PA of the ES-1 $v_{ES}$ based on $\Delta F$. Consequently, the needed P consumed by the SL (P_SL) is calculated as:

$$P_{SL}=P_{SLo}+\Delta P_{SL}=P_{SLo}+m\Delta f$$

(7)

where $\Delta P_{SLo}$is the SL’s nominal P.

The $v_{ES}$ is dependent on I_SL, $\Delta f$, and $v_{PCC}$, as an ES-1 is only capable of exchanging Q. Fluctuations in V are limited to 10 percent as illustrated in IEEE1547 [1993]. Due to the limitations of the ES-1, however, P output managed via Q management is complicated [154].

5.2. Correction for PF

By adjusting for P/Q, an ES keeps the µG stable throughout variations in RES production/generation. Nevertheless, ESSs are necessary for sustaining P adjustment by ESs. ES-2 and ES-3 are two alternative ES structures that were designed in order to overcome around the P constraint of ES-1 [151,152]. As shown in Figure 9b, the battery bank in ES-2 takes the position of the DCL capacitor in ES-1. However, batteries have a number of known downsides, such as large expansion, a short lifespan, and a high price [155,156], which might restrict ES implementation in 1Ø-µGs. Figure 9c shows how ES-3 uses a 1Ø converter to substitute the batteries [157]. In general, converter B operates similarly to ES-2, while inverter A adjusts the DCL’s V. ES-2’s and ES-3’s output’s V ($v_{PCC }$) is adjusted at any angle, giving ES-2 6 more operating states than ES-1. By directing the ES-2’s and ES-3’s PαQ into a shared goal, PF adjustment is achieved.

PF compensation using ES-2 was proposed in [74], while strategies applicable to ES-3 were introduced in [158]. In [145], CLs in the $dp$reference frame to control the total incoming I $i_{load},$as depicted in Figure 12, were designed. The $i_{load}$ in Figure 9b comprises $i_{sl}$ and $i_{sl}$. By individually adjusting the separated I components to produce an appropriate $v_{ES}$, PαQ compensation is carried out. In this instance, the limitation is that $v_{pcc}$ is not directly regulated by $v_{ES}$. Since the $i_{load}$ is being regulated, it presumes that all loads (critical/noncritical) are one load. In actual use, these two impedances may come from various equipment, and their combined I is measured outside of the SL.

A $dp$ frame-implemented input VαI control structure was presented in [74]. The PCL is depicted in Figure 13. PI controllers (PICs) regulate the components $v_{pcc, d}$, and $i_{in,q}$ in circuit 13, where the value is 13. In addition, the 1Ø phase−locked loop controls component $v_{s, q}$ to zero. In this instance, $v_{pcc}$ is regulated in a straight line by $v_{ES}$, whereas $i_{in}$ is regulated according to [145]. Ref. [159] proposed a $\delta$-control method that models the local network using an observer to generate the desired $v_{ES}$. Figure 14 depicts the PF correction using the $\delta$-control technique via ES-2. A PR controller helps to control $v_{ES}$, whereas an internal P controller control $i_L$. The viewer needs to know several parameters, including $Z_{CL}$, $Z_{NC}$, $Z_{grid}$, and voltages ($V_{grid}$ and $V_{pcc}$). The SL is unable to determine these readings regionally, and they are not subject to change as a result of connecting or disconnecting loads.

Ref. [160] suggested a radial–chordal decomposition (RCD) algorithm for controlling the resultant ES−2. The strategy shown in Figure 15 hinges on regional variables that the SL is able to measure. To perform PF adjustment with the ES−2, a $v_{ES}$ vector must be calculated, which is not an easy task.

Ref. [161] proposed a CL for ES-3 that adjusts the $v_{pcc}$and corrects the PF. As seen in Figure 16, converter A employs a PIC to regulate the DCL voltage ($V_{DC}$). The results of this loop are then used to create the reference $v_{ES}$ for converter B by applying it to a mathematical representation of the SL. This CL’s constraint is the unidentified equivalent R/X ratio of the system being used to weight the output of the PIC. This limits how the SL is deployed because an easy to set up capability is not available. In [158], an alternate solution dependent only on local parameters was introduced. Figure 17 depicts CLs in the dq frame to control $v_{ES}$. Converter B makes use of ES-2’s capabilities, which uses a PIC to control |$v_{ES}$| while maintaining a constant PA. Therefore, it is possible to obtain a sinusoidal V at the output of converter B ($v_{ES,B}$). By adapting the $V_{DC}$ and $i_{in,A}$ with PICs in the $dq$ frame, the CLs of converter A control the P sent to ES−3.

5.3. Harmonics Compensation (HC)

In [145], an HC process for ES-2 was introduced. As seen in Figure 12, the modified CLs are demonstrated in Figure 18. Using an FFT, the I harmonics are removed, and the CLs for each harmonic are introduced together. The method’s effectiveness due to the measured/tested I harmonics phases was not taken into account during regulation. Also, the decreases in harmonics were not enumerated, so the algorithm/method’s effectiveness is not objectively sound.

5.4. Cooperative Operating of Several ESs

Ref. [145] examined how ES−1 and ES−2 operate simultaneously with not serious resistive loads. The two ESs serve complementary purposes: ES−1 adjusts the $v_{pcc}$, and ES-2 optimizes the PF through I management. This scenario shows how different ES types can cooperate because their individual purposes are diverse. Nevertheless, problems arise when several concurrent ESs autonomously handle the V, F, and/or PF at the POCC. When examining the VR in µGs, for example, all ESs within the µG are unable to use the same voltage reference. The I passes; subsequently, V drops at the line impedances cause the Vs in the LV circuits to differ [87].

A coordinated action amongst identical-type spread ESs was contemplated in [147]. In [148], a method for tuning ES-1′s PI gains to guarantee solidity when paralleling several ESs was described. To ascertain the PI gains to attain stability, a stability model was developed. Unfortunately, the range of benefits is solely applicable to the case study in question. Due to the necessity to tailor the gains for each unique circumstance, this technique fails to result in plug-and-play possibilities. Application of a V droop technique was used to manage the work of many distributed ES-1s, as first reported in [147].

$$V_{PCC}=V_{PCC}^{*}-K{V}_{ES}$$

(8)

where K is the V’s DC parameter, $V_{PCC}^{*}$ is the desired V’s RMS, $V_{ES}$is the V’s output, and $V_{PCC}$ is the desired V obtained by the DC rule.

Multiple ESs can operate simultaneously without affecting the reliability of the µG, according to tests. In [135], the researchers proposed a two-level control technique (2LCT) to allow ESs to operate simultaneously. Figure 19 shows a block schematic of the 2LCT. Each ES−1 has a PCL architecture at the initial level, identical to that mentioned earlier [147], where $G_V$, $Z_V$, K, and ($G_2$, $G_3$) are the PR regulator, virtual impedance, V droop gain, and plant’s TFs, which are represented as:

$$G_2=\frac{1+sX{Z}_o}{s^{2}L{Z}_o+sLC+Z_o}$$

(9)

and

$$G_3=\frac{1}{s^{2}L{Z}_o+sLC+Z_o}$$

(10)

The PCL’s outside loop implements the DC legislation mentioned in (8). An extra virtual resistance is added to the CL in order to uniformly set the final impedance of each ES.

$$V_{ES}^{*}=V_{PCC}-R_V{i}_L$$

(11)

where $V_{ES}^{*}$, $R_V$, $i_L$, are the intended V’s output, virtual resistance, and ES’s inductor current, respectively.

Nevertheless, it is unclear from the analysis in [162] what effect $R_V$ has on the SL’s resultant impedance and the reactive compensated transfer. To achieve the desired V, a higher level that employs agreement control was subsequently created.

6. Discussion of the Approaches under Consideration

The above sections discussed the various methods and procedures that could be applied to address the PQ problems in 1Ø−µGs.

Table 1. Strengths and restrictions of PCL approaches.

Method	Advantages	Limitations
Conventional DC [163,164,165,166]	Easily implemented. Resembles the functionality of SGs.	VαF fluctuations. A trade-off between VR and Q-sharing precision. There is still Q interaction between DGs. No harmonic I/V regulation (NH I/VR).
Traditional DC with additional GS inductor [57,123,124]	Easy to implement. Imitates how SGs work. Enhances isolation between both DGs’ PαQ signals.	High VαF fluctuations. A trade-off between VR and precise Q sharing. The Q exchange (QE) between DGs is not disregarded. NH I/VR.
Inverse/reverse droops [121,167,168]	Easy to use.	High VαF fluctuations. There is no way to directly interact with SGs. The P exchange among DGs. NH I/VR.
Large traditional droop parameters [110,125,169,170]	Easy to implement. Reduces the Q exchange between the DGs.	Meaningfully increases VαF oscillations. NH I/VR.
$h$th harmonic DC [171,172,173]	Provides selective harmonic compensation. Leeway of DC to higher harmonic Fs. Lessens VαI harmonics.	Oscillations of VαF. There is still a Q exchange between DGs. Complicated DG-CLs. It is difficult to tune harmonic droop.
PR controllers [131,174,175,176]	Limits Q exchange within DGs. Provides selective harmonic compensation. Suppresses both IαV harmonics.	VαF instabilities. Performance is affected by the impedance characteristics of the electrical network. Lowers the I harmonic flowing through the DGs (but does not eliminate it).
R/L or RL virtual impedance [132,133,177,178,179,180,181]	Easily designed and implemented. Enhances the I share. Minimizes Q transfer within DGs.	VαF oscillations cannot be suppressed. The Q transfer among DGs is not eradicated. Rises the V at the POCC. Predominantly beneficial for specified µG impedance.
RC virtual impedance [46,124,182,183]	Simple to implement Enhances I harmonic exchange. Minimizes Q transfer within DGs. Improves V-HD at the POCC.	VαF variations are inevitable. The Q transfer within DGs is not completely reduced. This method is primarily successful for known µG impedances.
NVH-Z [138]	Improves I fundamental and harmonic share. Diminishes the Q transfer flanked by the DGs.	VαF variations are inescapable. The Q transfer between DGs is not abolished. Regularly active for an identified µG impedance.

,

Table 2. Properties and limitations of the SCL techniques.

Method	Advantages	Limitations
VαF restoration loops (RLs) [139,184,185]	Low-bandwidth noncritical communication (LBNC) Eradicates VαF variations.	Potentially increases the Q transfer between the DGs. NH I/VR.
VαF RLs together with Q control [46,140,141,186,187]	LBNC. Eliminates VαF fluctuations Eliminates Q exchange between the DGs.	NH I/VR.
VαF RLs including Q control and V- HC [46,133,188]	LBNC. Eliminates VαF variations. Eliminates Q transfer between DGs. Risk mitigation of the THD of VαI.	Tuning of SCLs for harmonic recompense is difficult. Synchronization of harmonic injection is a serious element.

and

Table 3. Emerging grid technology strengths and limitations.

Method	Advantages	Limitations
ES-1 [82,143,189]	No ESS. Mitigates V oscillations via Q compensation.	Only provides Q reimbursement. Cannot discourse F’s fluctuations. NH I/VR. The complexity of multiparallel ES-1 SL coordination.
ES-2 [74,145,190]	Alleviates VαF instabilities.	Requires ESS. No harmonic I/V regulation. It requires complicated coordination of numerous concurrent ES-3 SLs.
ES-3 [157,158,161,191]	No ESS. Mitigates VαF instabilities.	No harmonic I/V regulation. It is difficult to coordinate several concurrent ES-3 SLs.

summarize the characteristics and restrictions of the techniques outlined in the earlier parts of this paper. The capability of the strategy/system to lower/eliminate V/F changes, the Q between DGs, and VαI–HDs when applied to 1Ø−µGs are the bases for the comparisons.

Among the most well-established of the approaches presented in this work are primary DC techniques built on traditional DC. Despite the performance issues these methods have, the technology has developed enough to be used in practical scenarios. The most important compromise is between VR and Q-transfer precision in DGs. To reduce or eliminate the Q exchange across DGs, Q regulation must also be adopted. The PCLs of the DGs’ VI methods are subject to identical criticisms. Inadequate optimization of such techniques can reduce their efficiency because the grid’s impedance is frequently undetermined. The use of such strategies is particularly beneficial if the features of the regional network are specified.

In the literature, centralized and distributed techniques are distinguished as SC techniques. The methods used in traditional electricity generation are expanded upon by centralized methods, although both implement the same functionality. While the DGs in a µG are distributed over a greater geographic region, traditional electricity production is often placed close together in power plants. The main benefit of centralized SC is convenience because decisions are taken directly and sent regionally to the DGs in charge of providing energy to the µG. The reliability of the SCLs is dependent on the functionality of the central controller; therefore, this is a drawback. Decentralized SCLs provide more dependability at the cost of more sophisticated network structures and control method design. Despite the fact that decentralized methods have a lot of potential for practical µG executions, centralized control is now the most practical choice.

ES-based SLs are a developing grid solution that could offer dispersed PQ adjustment. The Q compensation needed to control local voltages can be provided by ES-1. But, because no ESSs exists, such gadgets are not very useful when there are significant local RESs. Improved abilities and potential PQ improvement functions are offered by ES-2 and ES-3. Before implementing such equipment, it is also necessary to look into how ESs may be integrated into actual loads like air conditioning.

7. Future Research Directions

There will be a variety of problems in future systems as a result of the implementation of µGs in contemporary power systems. The examined studies that were described here address a few concerns. However, further research is needed on a few of the previously listed issues. As a result, the following proposals are made for further work in µG control, operation, and management and mitigating PQ issues.

• Additional features, including the PQ index and equipment longevity, can be thought of as objectives in the context of µG management. The load control strategies ought to be examined more thoroughly than before on the control side.
• The increased use of µGs in recent systems creates a host of new problems, such as connections between µGs, multi µGs, multi agents, decentralized and centralized control procedures, and many others.
• Considering the advancement of technology, it is critical to understand how new machinery, particularly µGs, will affect power systems.
• Response to demand and load management in DGs have become crucial issues. With the growth in RESs and sophisticated metering systems in current decades, this topic may now be more important than ever.
• New solvers can be used to simplify and expedite the solving process because heuristic methods have improved.
• Proper uncertainty modeling can make the network functioning resistant to change. The uncertainties in µGs have been addressed in a number of studies, although a comprehensive approach needs to be offered, particularly if multiple uncertain factors exist simultaneously.
• Future systems will also need to address smartening. Every day, more and more systems will use information and communication technologies. So, it is important to take into account the connections between cyber and physical systems and their issues.
• Recent power systems are more open to incorporating µGs thanks to the use of innovative nonlinear and adaptive control techniques.

8. Conclusions

This study presented an overview of 1Ø−µGs PQ issues and potential solutions. VαF oscillations happen throughout off-grid mode as a result of the inherent limits of conventional DC. Some concerning additional issues are the Q exchange between DGs and HD. The HCA that is used to address PQ matters was also thoroughly analyzed and reported. The vast majority of available techniques involve changing the DG units’ primary CLs, which may include implementing droop gains like inverse/reverse droops (have limited applications because they cannot be directly connected to synchronous generators) or large droop gains (greatly increase VαF deviations as well as the flow of Q across inverters).

The performance restrictions of additional primary control techniques based on the idea of VI were also mentioned. Investigations have also been conducted into SCLs-based techniques for reducing PQ problems. These comprised the abolition of VαF variations, the abolition of the transfer of Q between DG units, and a reduction in HD. When aiming for identical objectives, it was found that secondary control approaches significantly outperform primary control methods in terms of effectiveness and performance.

Additionally, a review of the newly developed grid technology for ESs was given. ESs are capable of offering extra auxiliary services in addition to controlling voltage in weak grids. Further studies at the PCC have concentrated on reducing VαF variations and boosting PF. The DC input capacitor must be substituted by a battery or another converter (ES−2 and ES−3) in order for this capability to be possible. The reduction in the current HD was also described, but the supplementary services offered by such types of equipment are still in the early stages; thus, there is still space for major progress.