Power Quality Improvement in Renewable-Energy-Based Microgrid Clusters Using Fuzzy Space Vector PWM Controlled Inverter

Abstract

An increased electricity demand and dynamic load changes are creating a huge burden on the modern utility grid, thereby affecting supply reliability and quality. It is thus crucial for modern power system researchers to focus on these aspects to reduce grid outages. High-quality power is always desired to run various businesses smoothly, but power-electronic-converter-based renewable energy integrated into the utility grid is the major source of power quality issues. Many solutions are constantly being invented, yet a continuous effort and new optimized solutions are encouraged to address these issues by adhering to various global standards (IEC, IEEE, EN, etc.). This paper therefore proposes a concept of establishing a renewable-energy-based microgrid cluster by integrating various buildings located in an urban community. This enhances power supply reliability by managing the available energy in the cluster without depending on the utility grid. Further, a “fuzzy space vector pulse width modulation” (FSV-PWM) technique is proposed to control the inverter, which improves the power supply quality. This work uniquely optimized the dq reference currents using fuzzy logic theory, which were used to plot the space vectors with effective sector selection to generate accurate PWM signals for inverter control. The modeling/simulation of the microgrid cluster involving the FSV-PWM-based inverter was carried out using MATLAB/Simulink^®. The efficacy of the proposed FSV-PWM over the conventional ST-PWM was verified by plotting voltage, frequency, real/reactive power, and harmonic distortion characteristics. Various power quality indices were calculated under different disturbance conditions. The results showed that the use of the proposed FSV-PWM-based inverter adhered to all the key standard requirements, while the conventional system failed in most of the indices.

1. Introduction

The electricity demand of any country is a reflection of its economic growth. The increasing quality of life worldwide has led to more sophisticated energy demands. The world’s urban population is expected to increase from 55% (4.2 billion) to 69% (6.7 billion) of the 9.7 billion people on the planet by 2050, which will result in huge consumption of existing energy sources [1,2]. It will accordingly be critical to meet the worldwide consumer load demand with the desired quality of power supply. Thus, the topic of power quality is considered to be a significant perspective based on the current position of renewable energy resources and the frequent connection of these resources to distribution systems [3]. Thus, work on distributed-generation-based microgrids has been ongoing for several years. Initially, microgrids were introduced to deliver the power to the areas where there is less accessibility to conventional power. In such places, the microgrids are installed to work in both autonomous and grid-connected modes. The aggregation of low-power sources in a microgrid or virtual power plant is the actual direction of boosting the efficiency of distributed generation and renewable energy resources. These systems have numerous characteristics including a huge number of semiconductor devices, the stochastic nature of renewable generation, and bi-directional power flow. As a result, ensuring the required power quality indicators such as voltage fluctuations, frequency deviation, non-sinusoidality, and overall harmonic distortions is a critical responsibility [4,5].

For the past few years, researchers have developed various optimization techniques, filters, controllers, FACTS devices, compensators, and battery storage units to address the power quality concerns of the microgrids. These studies are primarily focused on three aspects: (1) inverter topologies controlled by modern modulation techniques, (2) alternative conversion methods to replace power electronic inverters, and (3) taking uncertainty into account when designing storage units for energy management. A review of these literature works on improving the power quality of distribution power systems is presented as follows.

The adaption-based control strategy is proposed to improve the power quality parameters at the point of common coupling (PCC) of a three-phase distribution system. The control strategy is designed to balance three-phase currents and compensate for the reactive power of the system [6]. Microgrid power quality is managed using a model predictive control methodology, which regulates the microgrid’s power converters to meet the requirements. The control algorithm is designed to function with the microgrid when it is connected to the utility grid mode, or in standalone mode, or in interconnected mode [7]. Different control strategies for improving the quality of power in a microgrid at the distribution side are discussed in [8]. Moreover, in [9], the authors presented a comprehensive survey of power quality enhancement devices, with a focus on the auxiliary services provided by multi-functional generators. A review of the microgrid concept, testbeds, and related control approaches was also conducted. Later in [10], in the presence of a shunt active power filter (SAPF), the impact of power quality difficulties in an adopted standalone microgrid system (comprising solar, wind, and fuel-cell-based distributed generation) is explored. In the realization of this SAPF, a traditional synchronous reference frame and modified synchronous reference frame techniques are used for reference current generation, proportional–integral controller and fuzzy logic controller are used for DC-link capacitor voltage regulation, and a basic hysteresis band current controller technique is used for inverter switching pulse generation. Similarly, in [11], the authors proposed a decision-making approach for determining the power quality of a single-phase AC microgrid along with the development of a power quality monitoring index to evaluate the fuzziness in a random variation of power quality. Further, the authors in [12] proposed an upgraded custom power device termed a “distributed power condition controller” to improve the power quality in a multi-microgrid system with a high penetration of varied distributed generators in the island and interconnected modes. An improved triple-action controller design as a combination of a proportional–resonant controller, selective harmonic compensator (SHC), and feed-forward current controller is proposed [13]. This work examines a standalone distributed generating inverter’s controller, where the SHC produces a continuous sinusoidal voltage.

For a multilevel converter, an optimal current control mechanism is proposed [14]. In online harmonic compensation applications in microgrids, the control approach will overcome the inefficiencies of earlier control methods in harmonic current tracking. Grid-connected inverter-based multi-functional distributed generation converters can use this control mechanism. Later in [15], the authors tried to solve the fixed-frequency operation and fast dynamics problems by proposing a unique robust voltage control technique using an angle droop scheme. The voltage controller is designed with rotating reference frames in mind to reduce the detrimental effects of harmonics and negative-sequence disruption generated by nonlinear and unbalanced loads. The negative consequences of parameter variation, model uncertainty, and unmodeled dynamics are calculated in the required form and employed as a constraint in the optimization problem to increase the overall system’s robustness. An adaptive droop controller based on a fuzzy logic method is proposed in [16]. The droop coefficient is adaptively modified to lessen the influence of the line impedance and achieve appropriate power splitting among hybrid energy by detecting unbalanced power and voltage deviation. Further, a small-signal model was created to investigate the impact of droop coefficient modification on DC microgrid stability. In the work reported in [17], self-tuning filters along with fuzzy logic were used to extract the harmonic currents for controlling the DC voltage required for a multilevel inverter. Similarly, D-Statcom was modeled and controlled with the use of fuzzy and PI current controllers for voltage quality improvement [18].

A brief review was given of the FACTS technologies used for stabilization and power filtering in distribution systems [19]. Along with custom power devices, some different evolutionary algorithms are also proposed for improved power quality [20]. In [21], a robust, quick, and dynamic proportional–integral resonance controller with a harmonic and lead compensator feedback (PIR + HC + LC) is presented to manage the current of a 15-level neutral-point-clamped multilevel inverter. In [22], a unified power quality conditioner proposed as a standardized power quality conditioner for a nine-level structure is investigated. A solar system and a smart grid are both connected to the conditioner. For the switches of both series and parallel converters, a novel modulation technique to generate switching signals is discussed. This modulation technique uses an adaptive hysteresis band calculated by a fuzzy logic controller to produce the appropriate adjusted output voltage with minimal distortion.

The above literature addresses the power quality issues in a single microgrid that is operated in either autonomous or grid-connected mode, and all these methods are limited for multi-microgrid systems. Moreover, most of these methods have not considered all the typical characteristics of the power quality issue. A steady evolution of regulations on how the energy is to be generated and used demands the continuous exploration of new and effective solutions. Hence, in view of all these points, this paper,

• Proposes the concept of establishing a renewable-energy-based microgrid cluster by integrating multiple adjacent microgrids located in an urban community. This enhances power supply reliability by managing the available energy in the cluster without depending on the utility grid. An energy management system (EMS) is designed to manage the available energy in the cluster, which helps to reduce the dependency on the central utility grid.
• Proposes a “fuzzy space vector pulse width modulation” (FSV-PWM) technique to control the inverter, which improves the power supply quality. The use of the proposed fuzzy logic increases the capabilities of the control loop in regulating the system output under different operating conditions.

The rest of the paper is organized as follows: Section 2 describes the microgrid cluster system and its constituents. Section 3 discusses the proposed FSV-PWM control scheme for the inverter. Section 4 presents the results and discussion, followed by the conclusions in Section 5.

2. Microgrid Cluster and Its Constituents

The general architecture of the microgrid cluster system considered for the study is shown in Figure 1. It is formed by interconnecting four different buildings, viz. residential, software industry, academic institute, and manufacturing industry buildings, with different load profiles and is completely based on the AC power park. Each building is referred to as a single microgrid and is associated with locally available sources such as photovoltaic (PV), wind units (WTs), and fuel cells (FCs) along with storage batteries, AC/DC converters (rectifiers), DC/DC converters, and DC/AC inverters for local building loads. The modeling of all constituents is discussed in brief [23,24,25,26]. Each microgrid is linked to renewable sources in combination with paid sources and a storage system and DC/DC converters along with building loads. The voltage produced from the renewable sources is very low, so the voltage is increased to the value of 500 V (DC) which is the equivalent to an AC voltage of 415 V. Before being connected to neighboring microgrids or the utility grid, each microgrid in the cluster is equipped with a circuit breaker, which is controlled by the control signals provided by the EMS. The EMS manages the power flow by controlling and monitoring the entire process of each building in the microgrid cluster. It is connected to all of the individual buildings and the utility grid. The EMS generates a control signal based on the power excess/deficit situations and delivers it to the circuit breaker for switching ON/OFF, resulting in energy transfer. All individual energy resources in each building are connected to their own DC bus and then connected to a DC/AC converter, which powers a variety of loads.

Table 1. Generation and load profile of the microgrid cluster.

Type of Microgrid	Resource Used	DC Capacity in kW (Generation)	Type of Building Loads	Average AC Load in kW and KVAR
Microgrid-1	Solar PV, Battery	4.5	Residential building (B1)	38.23 + j11.81
Microgrid-2	Wind, Fuel Cell	2.5	Software industry building (B2)	35.93 + j13.73
Microgrid-3	Solar PV, Wind, Battery	4.25	Academic institute building (B3)	54.03 + j18.69
Microgrid-4	Solar PV, Fuel Cell, Battery	4	Manufacturing industry building (B4)	58.31 + j14.97

shows the details of the generation and load profiles of the proposed microgrid cluster. The information about all the typical parameters used for modeling the cluster is mentioned in Appendix A. The integration of more distributed generations, storage systems, and necessary loads makes microgrid operation more difficult, especially when microgrids are networked and interoperable. Figure 2 shows the EMS operations developed to share the power among the neighbor microgrids connected in the cluster in island/grid-connected conditions [26].

3. Description of the Proposed Control Strategy

When compared to sinusoidal pulse-width modulation, space-vector pulse-width modulation has become one of the most common PWM approaches due to its quicker digital implementation and greater DC bus utilization. The block diagram representation of the control circuit for producing switching gate pulses applied to the inverter of an ith microgrid is shown in Figure 3. It consists of two stages: (1) fuzzy-logic-based abc-dq0 transformation for calculation of I_ref and (2) switching states and switching times of the current source inverter (CSI) of ith microgrid. The sequence of steps to produce switching pulses from the proposed control strategy to the inverter is shown in Figure 4.

3.1. Calculation Reference Current (I_ref) Using Fuzzy Logic

The set of voltages and currents are transformed to the $\alpha -\beta -o$ frame using Equations (1) and (2) [10].

$$\left[\begin{array}{c}{\nu }_o^{MG,i}\\ {\nu }_{\alpha }^{MG,i}\\ {\nu }_{\beta }^{MG,i}\end{array}\right]=\sqrt{\frac{2}{3}}\left[\begin{array}{ccc}\frac{1}{\sqrt{2}}& \frac{1}{\sqrt{2}}& \frac{1}{\sqrt{2}}\\ 1& -\frac{1}{2}& -\frac{1}{2}\\ 0& \frac{\sqrt{3}}{2}& -\frac{\sqrt{3}}{2}\end{array}\right]\left[\begin{array}{c}{\nu }_a^{MG,i}\\ {\nu }_b^{MG,i}\\ {\nu }_c^{MG,i}\end{array}\right]$$

(1)

$$\left[\begin{array}{c}{I}_o^{MG,i}\\ I_{\alpha }^{MG,i}\\ I_{\beta }^{MG,i}\end{array}\right]=\sqrt{\frac{2}{3}}\left[\begin{array}{ccc}\frac{1}{\sqrt{2}}& \frac{1}{\sqrt{2}}& \frac{1}{\sqrt{2}}\\ 1& -\frac{1}{2}& -\frac{1}{2}\\ 0& \frac{\sqrt{3}}{2}& -\frac{\sqrt{3}}{2}\end{array}\right]\left[\begin{array}{c}{I}_a^{MG,i}\\ I_b^{MG,i}\\ I_c^{MG,i}\end{array}\right]$$

(2)

Here, $I_{abc}^{MG,i}$ and $V_{abc}^{MG,i}$ are the source currents and voltages with respect to the $a-b-c$ frame, and $I_{\alpha \beta o}^{MG,i}$, $V_{\alpha \beta o}^{MG,i}$ are the source currents and voltages with respect to the $\alpha -\beta -o$ frame. The transformation of the $\alpha -\beta -o$ frame to the $d-q-o$ frame is given in Equation (3). The reference current with respect to the d-axis ($I_d^{*}$) is obtained as given in Equation (4), where $I_d^{MG,i}$ is the d-axis current, and the loss component value at $r\mathrm{th}$ sampling instant ($I_{loss(r)}^{MG,i}$) for meeting the losses in the ith microgrid is obtained as given in Equation (5).

$$\left[\begin{array}{c}{I}_o^{MG,i}\\ I_d^{MG,i}\\ I_q^{MG,i}\end{array}\right]=\left[\begin{array}{ccc}1& 0& 0\\ 0& \cos \theta & \sin \theta \\ 0& -\sin \theta & -\cos \theta \end{array}\right]\left[\begin{array}{c}{I}_o^{MG,i}\\ I_{\alpha }^{MG,i}\\ I_{\beta }^{MG,i}\end{array}\right]$$

(3)

$$I_d^{*}=I_d^{MG,i}+I_{loss}^{MG,i}$$

(4)

$$I_{loss(r)}^{MG,i}=I_{loss(r-1)}^{MG,i}+k_{pd}\left({(V_{dc}^{*}-V_{dc})}_r-{(V_{dc}^{*}-V_{dc})}_{(r-1)}\right)+k_{id}{(V_{dc}^{*}-V_{dc})}_r$$

(5)

where $k_{pd},k_{id}$ are PI controller gains and are tuned by fuzzy logic rules. Similarly, the reference current with respect to q-axis ($I_q^{*}$) is obtained as given in Equation (6).

$$I_q^{*}=I_q^{MG,i}+I_{qref}^{MG,i}$$

(6)

Here, $I_q^{MG,i}$ is the q-axis current and $I_{qref}^{MG,i}$ is the output component produced by the PI controller. The magnitude of the terminal voltage of the ith microgrid at PCC is calculated as given in Equations (7) and (8).

$$V_t^{MG,i}={\left(\frac{2}{3}\left(V_{as}^2+V_{bs}^2+V_{cs}^2\right)\right)}^{1/2}$$

(7)

$$I_{qref(r)}^{MG,i}=I_{qref(r-1)}^{MG,i}+k_{pq}\left({(V_{te}^{MG,i})}_r-{(V_{te}^{MG,i})}_{(r-1)}\right)+k_{iq}{(V_{te}^{MG,i})}_r$$

(8)

where $I_{qref(r)}^{MG,i}$ is the q-axis current at the $r\mathrm{th}$ instant and $V_{te(r)}^{MG,i}=\left(V_t^{*MG,i}-V_t^{MG,i}\right)$ is the terminal voltage amplitude at the $r\mathrm{th}$ sampling instant, where $k_{pq},k_{iq}$ are PI controller gains. The reference current can be obtained as given in Equation (9). The reference current vector can be obtained as given in Equation (10).

$$\left[\begin{array}{c}{I}_{os}^{*MG,i}\\ I_{\alpha s}^{*MG,i}\\ I_{\beta s}^{*MG,i}\end{array}\right]=\left[\begin{array}{ccc}1& 0& 0\\ 0& \cos \theta & -\sin \theta \\ 0& \sin \theta & \cos \theta \end{array}\right]\left[\begin{array}{c}0\\ I_d^{*MG,i}\\ I_q^{*MG,i}\end{array}\right]$$

(9)

$$I_{ref}=\frac{2}{\sqrt{3}}\left(\sqrt{{\left(I_d^{*}\right)}^2+{\left(I_q^{*}\right)}^2}\right)$$

(10)

The fuzzy controller of the proposed inverter is implemented by considering and evaluating some linguistic rules. The fuzzy controller consists of two inputs and one output function, the input error value is calculated by taking the difference between the reference voltage (V*_dc) and sensed voltage (V_dc).

Here, both input signals, i.e., $e(n)$ and $\Delta e(n)$, are numerical variables, which are transformed into linguistic variables. The fuzzy controller surface plot is shown in Figure 5. The proposed fuzzy scheme is characterized by the following features or assumptions.

(1)

Consists of seven fuzzy sets, namely, negative big (NB₋₃), negative medium (NM₋₂), negative small (NS₋₁), extreme zero (EZ₀), positive small (PS_+1.), positive medium (PM₊₂), and positive big (PB₊₃),

(2)

Triangular membership function with the Mamdani inference mechanism is used.

(3)

Centroid method is used for defuzzification.

(4)

A fuzzy logic controller is designed by considering 49 rules as listed in

Table 2. Rules proposed for the implementation of the fuzzy logic.

	Δe(n)	NB−3	NM−2	NM−1	EZ0	PS+1	PM+2	PB+3
e(n)
NB−3		NB−3	NB−3	NB−3	NB−3	NM−2	NM−1	NB−3
NM−2		NM−2	NB−3	NB−3	NM−2	NM−1	EZ0	NM−2
NM−1		NB−3	NB−3	NM−2.	NM−1	EZ0	PS+1	NM−1
EZ0		NB−3	NM−2	NM−1	EZ0	PS+1	PM+2	EZ0
PS+1		NM−2	NM−1	EZ0	PS+1	PM+2	PB+3	PS+1
PM+2		NM−1	EZ0	PS+1	PM+2	PB+3	PB+3	PM+2
PB+3		EZ0	PS+1	PM+2	PB+3	PB+3	PB+3	PB+3

.

3.2. Determination of Switching State and Switching Times

The switching limitation for CSI [27,28] space vector PWM is that only two switches should be active at any given moment. One is from the CSI’s top commutation group, while the other is from the bottom commutation group. The switching time duration of any sector from 1 to 6 of the inverter is shown in

Table 3. Switching time duration of all sectors.

Sector	Upper Commutation Group (S1, S3, S5)	Lower Commutation Group (S4, S6, S2)
1	$S_1={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_3={\tau }_2+{\tau }_0/2$ $S_5={\tau }_0/2$	$S_4={\tau }_0/2$ $S_6={\tau }_1+{\tau }_0/2$ $S_2={\tau }_1+{\tau }_2+{\tau }_0/2$
2	$S_1={\tau }_1+{\tau }_0/2$ $S_3={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_5={\tau }_0/2$	$S_4={\tau }_2+{\tau }_0/2$ $S_6={\tau }_0/2$ $S_2={\tau }_1+{\tau }_2+{\tau }_0/2$
3	$S_1={\tau }_0/2$ $S_3={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_5={\tau }_2+{\tau }_0/2$	$S_4={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_6={\tau }_0/2$ $S_2={\tau }_1+{\tau }_0/2$
4	$S_1={\tau }_0/2$ $S_3={\tau }_1+{\tau }_0/2$ $S_5={\tau }_1+{\tau }_2+{\tau }_0/2$	$S_4={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_6={\tau }_2+{\tau }_0/2$ $S_2={\tau }_0/2$
5	$S_1={\tau }_2+{\tau }_0/2$ $S_3={\tau }_0/2$ $S_5={\tau }_1+{\tau }_2+{\tau }_0/2$	$S_4={\tau }_1+{\tau }_0/2$ $S_6={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_2={\tau }_0/2$
6	$S_1={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_3={\tau }_0/2$ $S_5={\tau }_1+{\tau }_0/2$	$S_4={\tau }_0/2$ $S_6={\tau }_1+{\tau }_2+{\tau }_0/2$ $S_2={\tau }_2+{\tau }_0/2$

for the space phasor diagram of the CSI shown in Figure 6.

The mathematical expression for the active vector ${\widehat{I}}_n$ of the CSI space vector is given by Equation (11), where ‘n’ is the sector number of the current vector position,

$$\widehat{I_n}=1.15I_{dc}{e}^{j(\frac{\pi }{6}-(k-1)\frac{\pi }{3})}$$

(11)

The time that space vectors spend in a sector during ${\tau }_s$ is represented by the dwell time for stationary vectors. The so-called ampere-second balance principle is the foundation for dwell-time estimates. The key principle is that the total of the current vectors multiplied by the time interval of chosen space vectors equals the product of the sampling period ${\tau }_s$ and the reference vector $I_{ref}$, and it can be assumed to be constant throughout time if ${\tau }_s$ is small enough.

The reference current is approximated using this concept by using the nearby vectors corresponding to the real sector and the zero vectors to adjust the length. The ampere-second balance principle is given by Equations (12) and (13) to calculate $I_{ref}$ and ${\tau }_s$.

$$I_{ref}·{\tau }_s=I_1{\tau }_1+I_2{\tau }_2+I_0{\tau }_0$$

(12)

$${\tau }_s={\tau }_1+{\tau }_2+{\tau }_0$$

(13)

From Figure 6a, the reference current vector is projected onto the real and imaginary axis, and we get Equations (14) and (15). Upon solving Equations (12)–(15), Equation (16) can be obtained.

$$I_{ref}·\cos \delta ·{\tau }_s=I_{dc}({\tau }_1+{\tau }_2)$$

(14)

$$I_{ref}·\sin \delta ·{\tau }_s=\frac{I_{dc}}{1.73}(-{\tau }_1+{\tau }_2)$$

(15)

$$\begin{array}{c}\begin{array}{c}\begin{array}{l}{\tau }_1={\mu }_a\sin ({30}^0-\delta ){\tau }_s\\ \end{array}\\ {\tau }_2={\mu }_a\sin ({30}^0+\delta ){\tau }_s\\ \begin{array}{l}\\ {\tau }_0={\tau }_s-{\tau }_1-{\tau }_2\end{array}\end{array}\}\forall -{30}^0\le \delta \le {30}^0\\ where,{\mu }_a=\frac{I_{ref}}{I_{dc}}\end{array}\}$$

(16)

Here,${\tau }_s$, ${\tau }_1$, and ${\tau }_2$ are dwell times of the phasors $I_1$, $I_2$, $I_0$, and ${\mu }_a$ is the modulation index. By using Equation (11), we can determine the individual space vectors in sector 1. The generalized equation of the space vector in sector 1 is calculated as given in Equation (16). The typical value of space vector pulse width modulation is in the range [0.577–0.637], which is an improvement on the sinusoidal-PWM value range [0.5–0.577].

4. Results and Discussion

The usefulness of the proposed fuzzy space vector pulse width modulation (FSV-PWM)-based inverter for urban community microgrid cluster application was analyzed by considering parameters related to power quality, viz. voltage swell/sag [29,30], voltage imbalance [31,32], frequency deviations [33,34], power characteristics [35], and total harmonic distortion [35,36] with respect to various standard requirements. All these are explained in the following subsections. Various test conditions implemented to conduct the proposed study are presented in

Table 4. Test conditions and procedures for power quality study.

Test Case		Load_1 (Base)	Load_2 (Test)	Procedure for Testing
1. Voltage characteristics
Voltage sag or swell: Effect of sag or swell on the profile of the voltage	1(a)—sag	250 kW	Fault	Applying SLG fault nearer to PCC from 0.1 to 0.2 s.
	1(b)—sag	250 kW	Arc furnace Load	Applying arc furnace load from 0.15 to 0.25 s.
	1(c)—swell	250 kW	275 kW	Disconnecting a portion of baseload, e.g., 75% from 0.3 to 0.4 s.
Voltage imbalance: Reactive loading effect on imbalances	1(d)—imbalance	275 kW + j50 kVAR	j50 kVAR	Injecting j50 kVAR (inductive load) in phase ‘a’ at 0.12 s
2. Frequency characteristics
Frequency deviations with constant impedance loads	2(a)	250 kW + j180 kVAR-j50 kVAR	----	Applying constant impedance load to observe the settling time for frequency stability and deviations.
Frequency deviations with large dynamic loads	2(b)	250 kW	j175 kVAR and −j175 kVAR	Applying large dynamic reactive loads as follows. (1) Switching the inductive load ON at 0.5 s and OFF at 0.8 s, (2) switching the capacitive load ON at 1.3 s and OFF at 1.6 s, and (3) observing deviation in case of dynamic loading.
3. Power characteristics
Deviations in real power and reactive power characteristics when subjected to a very large reactive load	3	275 kW + j50 kVAR	j100 kVAR and −j75 kVAR	Applying large dynamic reactive loads as follows. (1) Switching the inductive load ON at 0.1 s and OFF at 0.2 s, (2) switching the capacitive load ON at 0.3 s and OFF at 0.4 s, and (3) observing deviation in real and reactive powers.
4. Total harmonic distortion
Effect of large resistive loading	4(a)	250 kW + j25 kVAR	(0–300) kW	Loads are injected into the system as follows. 150 kW is applied from 0.04 to 0.1 s; 300 kW is applied from 0.08 to 0.16 s; 150 kW is applied from 0.12 to 0.2 s.
Effect of large reactive loads (long-term)	4(b)	300 kW + j50 kVAR	j150 kVAR and −j200 kVAR	Switching ON the j150 kVAR load at 0.12 s and switching OFF the pre-connected −j200 kVAR load at 0.12 s.

.

4.1. Analysis of Voltage Characteristics

The voltage waveform shape is to be maintained as pure sinusoidal, otherwise the distorted waveform leads to unequal voltage distribution in the cluster system. The main reason for the voltage sag is an abrupt change in the current flowing through the impedance (source), and sudden application of load or faults draws a high quantity of source currents, thus reducing the voltage. In the case of a voltage swell, sudden removal of load leads to less source current being drawn and causes an increase in the voltage. Typical voltage characteristics such as voltage sag/swell and imbalance are measured by considering test conditions given in Table 4. Test cases from 1(a) to 1(d) were considered and executed to observe the voltage sag, voltage swell, and voltage imbalance characteristics. The results obtained using simulations are shown in Figure 7, Figure 8, Figure 9 and Figure 10. The quantitative values are consolidated and presented in

Table 5. Comparison of power quality indices with conventional and proposed methods.

Power Quality Indices		Test Case	Control Technique Used for Inverter’s Gate Pulse Generation		Standard Requirements
			Conventional ST-PWM	Proposed FSV-PWM
Voltage Characteristics	Sag	1(a)	37.5%	17.5%	40% (IEC 61000-4-11 [30])
		1(c)	43.9% (violated)	19.5%
	Swell	1(b)	2.5%	3.75%
	Imbalance	1(d)	3.07% (violated)	1.233%	3% (IEEE 1159.3 [31], EN 50,160 [32])
Frequency Characteristics	Settling time	2(a)	0.48 s	0.3 s	2% (IEC 61,727 [33], IEC 61000-2-2 [34])
	Deviation (dynamic load)	2(b)	1	0.28
Power Characteristics	Real power disturbance settling time	3	2 ms	1.3 ms	3 ms (IEEE 1547.1 [35])
	Reactive power disturbance settling time		4 ms (violated)	1.3 ms
Total Harmonic Distortion		4(a)	1.53%	1.18%	5% (IEEE 1547.1 [35], IEEE 519 [36])
		4(b)	7.45% (violated)	4.73%

.

A comparison of voltage sag/swells of microgrid cluster with conventional ST-PWM and proposed FSV-PWM-based inverters as per test cases 1(a) and 1(c) is given in Figure 7. From this, it is observed that the voltage sag (drop) in the conventional system is 170 volts, while it is 70 volts in the proposed system. Similarly, the voltage swell (rise) in the conventional system is 10 volts, while it is 8 volts in the proposed system. Hence, both the sag and swell are reduced with the use of the proposed FSV-PWM-based inverter. Similarly, the current waveforms during the voltage sag and swells as per test cases 1(a) and 1(c) are shown in Figure 8. From the obtained results it is shown that when a line to ground fault is applied in phase ‘a’, the fault current is higher, i.e., 50 kA as in the case of the ST-PWM-based inverter used in the cluster, which is high when compared with the proposed inverter which gives 35 kA. A comparison of voltage sag of the microgrid cluster with conventional ST-PWM and proposed FSV-PWM-based inverters as per test case 1(b) is given in Figure 9. From this, it is observed that the voltage sag (drop) in the conventional system is 180 volts, while it is 60 volts in the proposed system. Hence, this test condition also suggests the importance of the proposed method to address the voltage sag issue.

Similarly, a comparison of voltage imbalance of the microgrid cluster with conventional ST-PWM and proposed FSV-PWM-based inverters as per test case 1(d) is given in Figure 10. The positive sequence voltage and negative sequence voltages are plotted. From this, it is observed that the conventional system produces more deviation, thereby creating an imbalance in the voltage characteristic compared to the proposed method. The imbalance issue is solved using the proposed FSV-PWM-based inverter.

4.2. Analysis of Frequency Characteristics

Frequency is the important parameter that is to be continuously monitored and controlled when the microgrid is either operated in grid-connected or autonomous mode. For analyzing the frequency characteristics of the proposed system, different test cases 2(a) and 2(b) from Table 4 were considered and executed. The simulated results are shown in Figure 11 and the consolidated quantitative results are listed in Table 5.

4.3. Analysis of Power Characteristics

The real and reactive power changes in the microgrid cluster are observed by injecting an inductive load of 100 kVAR from 0.1 to 0.2 s and a capacitive load of 75 kVAR from 0.3 to 0.4 s along with the baseload of 275 kW + j50 kVAR, as defined by test case 3 in Table 4. The corresponding simulation results are as shown in Figure 12; from which it is observed that the transient variations are smaller in the case of the proposed FSV-PWM-based inverter when compared to the conventional ST-PWM-based inverter during the applied large reactive load disturbances. These consolidated quantitative results are listed in Table 5.

4.4. Analysis of Total Harmonic Distortion

Total harmonic distortion is due to the nonlinear/high reactive loads in the microgrid cluster. For analyzing these distortions, consider the test cases 4(a) and 4(b) from Table 4 which are simulated as per the given procedure. The corresponding simulated results are shown in Figure 13. The consolidated results are listed in Table 5.

5. Conclusions

With the rising prevalence of renewable energy sources in power distribution systems, power electronic inverters are becoming increasingly crucial. Due to the rising importance to urbanization, microgrids are interconnected and run to provide stable, high-quality power to the consumers as well as to the grid. Thus, by considering factors such as the growing requirement for grid-independent systems, clean energy generation, and reliable and quality supply to feed consumer equipment, this paper proposes the concept of establishing renewable-energy-based microgrid clusters by integrating multiple adjacent microgrids located in an urban community. Further, to control the inverter, a “fuzzy space vector pulse width modulation” (FSV-PWM) technique is proposed. The salient advantages of this proposed system over the conventional system are as follows.

• The interoperable cluster microgrids can enhance the power supply reliability by managing the available energy in the cluster without depending on the utility grid. This further helps to reduce the dependency on the central utility grid, thereby solving the issue of frequent grid outages. However, the operation of an energy management system plays a crucial role in achieving effective energy sharing.
• Improved power quality through the proposed FSV-PWM technique by increasing the capabilities of the control loop in regulating the system output under different operating conditions. This can protect the power system during different critical uncertain conditions, such as large resistive/reactive loads, faults, arcing loads.
• Simple construction and reduced usage of clamping diodes and capacitors.
• Reduced $dv/dt$ stress with a low rating of the switching devices.
• The use of fuzzy logic reduces the usage of mathematical formulations.
• The modulation index is high.
• Reduced current harmonics when compared to the conventional mechanism.

Hence, from the comprehensive results obtained on various power quality indices, it is concluded that the proposed FSV-PWM controlled inverter is the best option to improve the power quality in urban community cluster microgrids.