Enhanced Power Quality in Single-Phase Grid-Connected Photovoltaic Systems: An Experimental Study

Abstract

The main aim of the research work presented in this paper consists of proposing an effective control scheme for a grid-connected single-phase photovoltaic (PV) system to enhance not only the power quality at the point of common coupling (PCC) but also to operate with a maximum power point tracking (MPPT) controller. Moreover, an orthogonal signal generator (OSG) module for effective grid synchronization, a current reference generation controller, and a PWM generating block have also been designed and included in this paper. The proposed control strategy allows the MPPT controller to switch to faulty mode and maintains the voltage according to network requirements using an adaptive neuro-fuzzy inference system (ANFIS)-based control whenever a fault occurs at the PCC. The performance of the analyzed control strategy, which is based on the static compensation of the DC-link voltage fluctuations in a grid-connected inverter powered by PV, is further explored through simulations in MATLAB, and the results are included in this paper. Moreover, the control scheme is implemented experimentally using a dSPACE DS 1104 control board and then assessed on a small laboratory-scale single-phase PV system that is subjected to some fault scenarios. The simulation and experimental results have shown improved power quality and robustness against grid fluctuations, resulting in better dynamic performance.

1. Introduction

In recent years, power quality (PQ) has become a real problem for both utility providers and consumers due to the widespread interconnection of power-electronic-based nonlinear loads to the electricity distribution grid. Voltage sags and swells are among the most common PQ issues affecting consumers’ critical loads [1,2,3,4]. A voltage sag is defined as a drop in the AC voltage’s RMS value at a power frequency of 0.5 cycles per minute, according to IEEE 1346 [5] and IEEE 1159 [5,6,7] standards. It can also be defined as a rapid drop and simultaneous fast recovery in the voltage at any point of the electrical system according to the standard IEC 61000-2-1 [8]. Voltage sags are often caused by short-circuit faults such as single line-to-ground faults due to the sudden interconnection of a large load such as an industrial induction motor. A voltage swell is described by the IEEE standards as a rise in the RMS value of the supply voltage from 1.1 p.u. to 1.8 p.u. of the nominal voltage during a period from 0.5 s to one minute. The interconnection of large capacitors or the disconnection of large loads are the two primary sources of voltage surges. As a result, some of the industrial equipment, such as motor drives, surge arresters, and relays, may be overheated, tripped, or even damaged by voltage surges [8,9], whereas, for sensitive loads such as electronic equipment, harmonic pollution is of major concern because its regulation is based on the crossings of the zero or maximum value point of the supply voltage [3,4,5,6,7].

Renewable energy generation technology based on power electronics is developing rapidly. New energy generation technology is based on photovoltaic (PV) systems, which are generally located in mountainous areas and deserts to receive more sunlight. In addition, these remote areas are conducive to the large-scale installation of PV inverters [10,11]. Recently, an increase in the deployment of solar PV systems, from rooftop installations to large-scale solar farms in the megawatt range, has been driven not only by the continuous fall in the price of PV modules but also by the consumers’ rising awareness of and wish to adopt eco-friendly energy solutions [9]. Future PV systems are expected to meet power generation needs from a point closer to the point of consumption [12,13,14] to avoid transmission losses and enhance the efficiency of the distribution network. Although there is an extensive body of literature on PV systems, this area is still an active area of ongoing research and development. Solar energy plays a leading role in reducing greenhouse gas emissions and has received strong support through various incentive programs, although the price is still expensive.

In previous research [15,16], the authors describe a modeling technique and fuzzy logic-based MPPT controller for efficient maximum power extraction. The authors of [17] presented an accurate method to adjust the duty cycle under different climatic conditions. The switching pulses of the power converter semiconductor devices, which regulate the active and reactive forms of power, are generated using a peak current controller [18,19,20]. Future PV systems should be capable of managing reactive power, islanding detection, power quality enhancement, and fault ride-through (FRT), as described in earlier studies [21,22,23]. In grid-connected inverter systems, some of the voltage amplitudes and frequencies are fundamentally critical parameters; accurate knowledge of these parameters is vital for obtaining the appropriate reference signals and meeting the requirements of future grids [24,25].

In another study [26], the authors have shown that power electronic equipment consumes large quantities of reactive power and can lead to the introduction of harmonics into the network. The authors of [27] have proposed a conventional PI controller to enhance the power quality of a multilayer single-phase inverter. The authors of [28,29] have stressed that voltage sag is a frequent problem that often results in unnecessary tripping when applied to a single-phase system operating in both ON and OFF grid modes. In addition, in an earlier study [30], the authors use a PV system to solve a problem related to the efficient control of the reactive power caused by using loads that influence the current supplied from the source. In islanded mode, the voltage controller maintains the sinusoidal waveforms of the load voltage. In various other studies [31,32,33], the authors stated that the synchronization controller ensures the transition between the bidirectional islanded and grid-connected modes, depending on the grid’s accessibility.

The work presented in this paper intends to carry out experimental analysis on a laboratory prototype to assess the dynamic performance of single-phase grid-connected PV systems that are subjected to various faults. One of the merits of this research is to implement a PV system controller to regulate the active and reactive powers required by the load. A perturb and observe (P&O) and maximum power point tracking (MPPT) algorithm is introduced to allow the PV system to operate at its maximum power. In addition, when a fault occurs at the point of common coupling (PCC) under any voltage sag circumstances, the controller will inject a reactive current, and the system may operate in islanded mode.

The rest of the paper is organized as follows. Section 2 gives a description of the proposed system considered in this work. Section 3 describes the power control concepts and Section 4 presents the proposed ANFIS-based power controller. Section 5 is devoted to the various grid synchronization techniques. In Section 6, a series of simulation results are inserted, and in Section 7, the hardware test bed of the grid-connected PV system developed in this paper is described, together with the introduction of a series of experimental test results. Finally, Section 8 summarizes the main conclusions of the paper.

2. Description of the System

For PV systems to operate effectively, it is necessary to control various parameters, such as the PV side for maximum power extraction, the inverter side for injecting high-quality current, and the grid side for any additional tasks. The ambient temperature and solar irradiance must be considered during a PV system’s design and planning phases. These factors greatly influence the PV system’s output power [34,35]. Figure 1 shows the schematic diagram of a system comprising a dual stage with a PV system connected to the PCC through a DC-DC boost converter, grid inverter, and control system as proposed in the literature [7,22]. The system may operate in both grid-connected and islanded modes.

3. PV Array Modeling

The PV cell is the basic unit of a PV module; it transforms the photons of the sun’s rays directly into electric power. Figure 2 shows the equivalent circuit of a PV system, which consists of N strings of PV modules in parallel with a line consisting of series M modules to achieve the appropriate power rating.

The dynamic model of the PV array is described by the following equations [36].

$$\mathrm{I}={\mathrm{N}}_{\mathrm{p}}{\mathrm{I}}_{\mathrm{ph}}-{\mathrm{N}}_{\mathrm{p}}{\mathrm{I}}_0\left[\exp \left(\frac{\mathrm{q}\left({\mathrm{V}}_{\mathrm{pv}}+{\mathrm{I}}_{\mathrm{pv}}{\mathrm{R}}_{\mathrm{sh}}\right)}{{\mathrm{N}}_{\mathrm{s}}\mathrm{AKT}}\right)-1\right]-\frac{\mathrm{V}+{\mathrm{IR}}_{\mathrm{sh}}}{{\mathrm{R}}_{\mathrm{sh}}}$$

(1)

$${\mathrm{I}}_{\mathrm{s}}={\mathrm{I}}_{\mathrm{rs}}{\left(\frac{{\mathrm{T}}_{\mathrm{c}}}{{\mathrm{T}}_{\mathrm{ref}}}\right)}^3\left[\exp \left(\frac{{\mathrm{qE}}_{\mathrm{g}}\left(\frac{1}{{\mathrm{T}}_{\mathrm{ref}}}-\frac{1}{{\mathrm{T}}_{\mathrm{c}}}\right)}{\mathrm{KA}}\right)\right]$$

(2)

$${\mathrm{I}}_{\mathrm{rs}}=\frac{{\mathrm{I}}_{\mathrm{s}}}{\left[\exp \left(\frac{{\mathrm{qV}}_{\mathrm{oc}}}{{\mathrm{N}}_{\mathrm{s}}\mathrm{KAT}}\right)-1\right]}$$

(3)

$${\mathrm{I}}_{\mathrm{ph}}=\left[{\mathrm{K}}_{\mathrm{i}}\left(\mathrm{T}-{\mathrm{T}}_{\mathrm{ref}}\right)\right]\left(\frac{\mathrm{G}}{1000}\right)$$

(4)

To be able to produce maximum power, the PV module must work within a peak value ${\mathrm{P}}_{\max }$ of the output power. The MPPT algorithm used in this study is based on the well-known P&O technique, and a 1 kW PV system is considered.

4. Analysis of Power Control

This section discusses the control scheme of a single-stage single-phase grid-connected PV system. In normal cases, the PV solar system would be operated in MPPT mode, which may give maximum power to the grid. However, when a fault occurs at the PCC, the grid will require the PV inverter to supply a reactive current to boost it [7]. During islanding, this mode will switch to fault mode, as illustrated by the flowchart diagram in Figure 3.

The control block diagram of the PV system is shown in Figure 4. It consists of the MPPT, the orthogonal signal generator (OSG) unit, the PWM block, and the PI controllers for current reference generation. A low pass (LP) filter is introduced to smooth the inverter output voltage waveform. In the $EF$ frame, the grid current and voltage components are represented by $I_{gE}$, $I_{gF}$, and $V_{gE}$, $V_{gF}$, respectively. The real ($P$) and reactive ($Q$) powers are defined by the following equations [23] and their values may be calculated using OSG systems in the same manner as in the dq-rotating reference frame:

$$P=\frac{1}{2}\left(V_{g\alpha } I_{g\alpha }+I_{g\beta } V_{g\beta }\right)$$

(5)

$$Q=\frac{1}{2}\left(V_{g\alpha } I_{g\alpha }-I_{g\beta } V_{g\beta }\right)$$

(6)

in which the currents in the 𝛼𝛽-stationary reference frame are given by:

$$\left[\begin{array}{c}{i}_{g\propto }\\ i_{g\beta }\end{array}\right]=\frac{2}{v_{g\beta }^2+v_{g\beta }^2}\left[\begin{array}{cc}{v}_{g\alpha }& v_{g\beta }\\ v_{g\beta }& -v_{g\alpha }\end{array}\right]\left[\begin{array}{c}P\\ Q\end{array}\right].$$

(7)

A controller such as a PI may be used to regulate the power since both powers are constant current quantities. The current reference is given by:

$$\left[\begin{array}{c}{i}_{g\propto }^{*}\\ i_{g\beta }^{*}\end{array}\right]=\frac{2}{v_{g\beta }^2+v_{g\beta }^2}\left[\begin{array}{cc}{v}_{g\alpha }& v_{g\beta }\\ v_{g\beta }& -v_{g\alpha }\end{array}\right]\left[\begin{array}{c}{G}_P\left(s\right) \left(P-P^{*}\right)\\ G_q\left(s\right) \left(Q-Q^{*}\right)\end{array}\right].$$

(8)

The transfer functions of the PI controllers for both these powers may be expressed as:

$$\begin{array}{c}{G}_P\left(s\right)=K_{pp}+K_{pi}\frac{1}{s}\\ G_q\left(s\right)=K_{qp}+K_{qi}\frac{1}{s}\end{array}.$$

(9)

5. ANFIS-Based Controller Development

ANFIS belongs to the class of neuro-fuzzy systems, which combine the learning capability and the parallel computation of neural networks with fuzzy systems. Neuro-fuzzy systems can be trained to develop fuzzy rules and determine the membership functions for the variables of the system. The structure of a neuro-fuzzy system is similar to that of a multi-layer neural network and includes an input and an output layer, along with three hidden layers that represent the membership functions and fuzzy rules. ANFIS training uses a hybrid learning algorithm combining the least-squares estimator (forward pass) and the gradient descent method (backward pass) [37]. These features provide ANFIS with superior performance and robustness [38]. In this section, an ANFIS-based controller is proposed for the control of the inverter DC-link voltage. Details of the design steps and a list of the parameters of the ANFIS training are given in Appendix A. Figure 5 depicts the triangular membership functions of the error, $\epsilon$, and error change, $\Delta \epsilon$. Figure 6 illustrates the variations of the output against the two inputs $\epsilon$ and $\Delta \epsilon$. ANFIS was originally designed with bell-shaped MFs.

6. Grid Synchronization Techniques

Following grid failures due to the occurrence of a fault, the synchronization of single-phase PV systems is critical. A system with good synchronization capabilities would have to be designed to respond quickly to any voltage drop. In the recent literature, several synchronization methods have been proposed [38], which may be broadly categorized into two groups:

• Mathematical methods, e.g., synchronized Fourier- and PLL-based methods. Currently, PLL-based synchronization techniques are the most popular choices since the design of a phase detector, a simple sinusoidal multiplier [38,39]. In a single-phase system, however, this procedure will result in a double-frequency term. Another approach to removing phase error for a PLL-based method is to apply the Park transform to an OSG system. A PLL based on T/4 delay [37] and inverse Park transform-based PLL (IPT-PLL) [40] are examples of such PLLs.
• Another approach is to employ adaptive filters that can automatically modify the output. The OSG is utilized to help in the calculations of active and reactive powers and provides appropriate grid synchronization. The OSG system is used to calculate $P$ and $Q$. Figure 7 depicts the OSG for a single-phase system [7].

The reactive power is subjected to:

$$I_p^2+I_q^2=I_N^2\le I_{max}.$$

(10)

The minimum value of the reactive current ($I_q$) to be injected into the grid system is given by Equation (11). The active power is controlled to keep the frequency constant.

$$I_q=\{\begin{array}{c} 0.9\le \frac{V_g}{V_N}\le 1 \\ 2\times \frac{V_g}{V_N}\times I_N, 0.5\le \frac{V_g}{V_N}\le 0.9\\ I_N, 0.9\le \frac{V_g}{V_N}\le 1 \end{array}$$

(11)

$$I_p=\{\begin{array}{c}Im, 0.9\le \frac{V_g}{V_N}\le 1\\ \sqrt{I_N-\left(2\times \frac{V_g}{V_N}\times I_N\right)}, 0.5\le \frac{V_g}{V_N}\le 0.9\\ 0, 0.1\le \frac{V_g}{V_N}\le 0.5\end{array}$$

(12)

Several recent papers have dealt with control methods to suppress the injected current harmonics. In this case, the harmonic compensation method is used, as shown in Figure 8.

Thus, the closed-loop transfer function is obtained as:

$$\frac{i_g\left(s\right)}{i_g^{*}\left(s\right)}=\frac{G_{PR+HC}\left(s\right) G_d\left(s\right)}{LS+G_{PR+HC}\left(s\right) G_d\left(s\right)}$$

(13)

where G_PR+HC(s) represents the controller’s transfer function and G_d(s) is the processing time delay.

The relationship of G_PR+HC(s) and G_d(s) can also be written as:

$$\begin{array}{c}{G}_{PR+HC}\left(s\right)=K_p+\frac{K_{i}s}{s^2+{\omega }_0^2}+{\displaystyle {\displaystyle \sum }_{h=3,5,3,7\dots }}\frac{K_{ih}s}{s^2+{\left(h{\omega }_0\right)}^2}\\ G_d\left(s\right)=\frac{1}{1+1.5T_{s}s}\end{array}$$

(14)

$$\left(G_{HC}\left(s\right)=\frac{K_{ih}s}{s^2+{\left(h{\omega }_0\right)}^2}\right)$$

(15)

where k_p is the value of the proportional gain and k_ih represents the resonant and harmonic compensator gains, respectively; h represents the value of the harmonic order, ω₀ denotes the value of the fundamental frequency, and T_s represents the value of the sampling period.

The current control strategy will decide on the quality of the power injected into the grid and the precise harmonics compensators [7] that are required to achieve this. In the $BC$ reference frame, the gains of the PR and HC controllers are defined by Equations (14) and (15) respectively.

The PV system parameters are listed in

Table 1. PV panel, Boost converter, and load specifications.

Component	Value
PV panel maximum power
LC filter values	2.3 mH, 100 KF
Linear load	450 W, 125 Var
Nonlinear load	850 W, 250 Var
Converter values	2.9 mH, 70 KF, 15 kHz

. Details of the Simulink implementation of a single-phase grid-connected PV system with an LCL filter is given in Appendix C.

7. Simulation Results

The performance of the proposed control strategy, applied to a grid-connected inverter powered by PV solar energy, has been analyzed through a simulation created using MATLAB 2016 b. The combined PV array and DC-DC boost converter is connected to the grid through a voltage source inverter. A filter is introduced to obtain a near-sinusoidal voltage waveform.

Figure 9a,b shows the various forms of power during islanding and sag voltage. Moreover, the DC link voltage and real and reactive powers’ waveforms for a linear load are depicted in Figure 10a,b, respectively. These results show the good performance of the proposed control system.

Figure 11a,b shows the various forms of power in the case of power supplied by the source and PV system.

The proposed system is assessed under fault conditions, in the presence of a sag applied at t = 0.1 and cleared at t = 0.2 s. The results are shown in Figure 12.

Figure 13 present the total harmonic distortion (THD), the grid current with active filtering has been reduced to 2.15%, which can be considered a significant enhancement.

8. Experimental Results

The experimental setup designed to validate the proposed control strategy for the grid-connected PV system is shown in Figure 14. It consists of:

• SEMIKRON IGBT based inverter;
• Current and voltage sensors;
• DC-DC boost converter;
• Linear load;
• dSPACE DS 1104 and power quality analyzer;
• Host PC and oscilloscope;
• Five PV panels.

Figure 15 shows the PV panel used in this work, with no load and in operating temperature conditions ranging from 25 °C to 35 °C; the solar irradiance is between 970 W/m² and 800 W/m².

8.1. Islanding Mode

Initially, the boost converter and MPPT controller were designed and tested with linear and non-linear loads for an input voltage of around 200 V and an output voltage of 400 V. Figure 16a shows the experimental results of the PV voltage (185 V) and output voltage (330 V) of the boost converter. As shown in Figure 16b, the duty cycle has a value of 0.7.

As can be seen in Figure 17 and Figure 18, the converter reacts to the load by decreasing the current from 12.3 A to 5.65 A. The power remains constant at 1030 W for the input and output of the boost converter. This demonstrates that the proposed control system presents a good dynamic response in both grid-connected and islanded modes.

At time $t=1 \mathrm{s}$, the grid is disconnected, and the PV system is subjected to a fault that occurs at the PCC point. At this stage, it is necessary to satisfy the load demand by using the PV system in the islanded mode. The prime concern would be to disconnect the grid and use the available PV system power. Figure 19 and Figure 20 are included to show, respectively, the voltage waveform during the fault and the active and reactive power supplied to both the linear and nonlinear loads.

Moreover, and as can be seen in Figure 21a, the signals given to the gates of IGBTs 1 and 3 are in opposition to the signals delivered to the gates of IGBTs 2 and 4. Between the pulses delivered to the same branch of the inverter bridge, a dead time of 4 µs between the pulses has been maintained to prevent short circuits, as clearly seen in Figure 21b.

8.2. Load Mode as Supplied Solely by the Grid

Figure 22 and Figure 23 show the responses of the active and reactive powers supplied to the linear and non-linear loads, respectively. It can be seen that the value of the active power supplied to the linear load is 450 W, and that of the reactive power is equal to 125 Var. For the nonlinear load, the values of (P) and (Q) are 850 W and 210 Var, respectively. This clearly shows that the control system presents good dynamics.

8.3. Load Mode as Supplied by the Grid and PV System

When the experimental system is connected to the grid, the source begins to supply the load with sufficient power, as can be seen at the start of Figure 24, where the values of the active and reactive powers are equal to 850 W and 125 Var, respectively. When a fault occurs at the PCC point, the source fails to meet the load demand; hence, the PV system starts to supply power to the load. The corresponding results are clearly shown in Figure 24, where the values of P and Q become 1030 W and 210 Var, respectively.

Figure 25 and Figure 26 show the load current before and after harmonics correction.

The waveform of the load current before correcting the harmonics is presented in Figure 26. The spectral analysis indicates that the THD is 40.8%, as illustrated in Figure 27, where the load is supplied solely by the source. However, after the fault, the THD has reduced drastically to 2.8%, as can clearly be seen in Figure 28.

8.4. Voltage Sag Mode

During the appearance of the voltage sag after the occurrence of a fault, the system is shifted to sag fault mode, and the current controller still supplies reactive current to restore the voltage according to the grid requirements, as shown in Figure 29.

It is clear from Figure 30 that upon the occurrence of the sag, the load voltage remains unchanged, and the supply to the load is maintained by the PV system, due to the compensating current.

Both Figure 31 and Figure 32 are included to show, respectively, the responses of the active and reactive powers during this sag disturbance and the DC link voltage across the capacitor, which is maintained to be constant at 330 V.

We can see the close similarity of the simulated results with those of the experimental tests, as summarized below:

• Waveforms of real and reactive power are depicted during islanding and during sag voltage, as shown by the simulated results in Figure 9 and those of the experimental test in Figure 20.
• Waveforms of the DC link voltage and real and reactive power for a linear load are shown in the simulated results in Figure 10 and those of the experimental test in Figure 22.
• Waveforms of the load supplied by the source and PV as well as the powers for linear and nonlinear loads are shown in the simulated results in Figure 11 and those of the experimental test in Figure 24.
• Waveforms of the source voltage in fault conditions and during sag voltage are shown in the simulated results in Figure 12 and those of the experimental test in Figure 19 and Figure 29.
• Finally, the waveforms of the grid current THD analysis are shown in the simulated results in Figure 13 and those of the experimental test in Figure 27 and Figure 28.

9. Conclusions

In this paper, a robust and effective control technique for a grid-connected single-phase photovoltaic (PV) system to improve the power quality at the PCC is proposed. The control strategy ensures that the PV inverter manages and performs its functions simultaneously (active power injection, reactive power compensation, and current harmonics filtering) without overrating by limiting its output current. Various scenarios, variable reactive power baselines, and nonlinear load power demands below and above PV power generation have been studied to assess the efficiency of the proposed strategy. Moreover, a performance analysis of the proposed control strategy when applied to a grid-connected inverter powered by PV has been carried out through simulations using MATLAB; some of the simulation results are included in this paper. The control scheme has been validated experimentally on a laboratory-scale PV system and tested under various grid fault and variable load conditions; the main results are included in this paper. The experimental results demonstrate that the proposed control strategy is robust, stable, and efficient. Moreover, in this research, the proposed controller was able to successfully deal with both linear and nonlinear load conditions and demonstrates that the use of the PV system significantly enhances the power quality.

Referencing several studies that have been reviewed,

Table 2. Summary of the research work related to grid-connected PV system control strategies.

Ref.	Year	Method	Merits/Disadvantages
[41]	2017	Conventional PQ control method	- Ease of implementing the PI controller - Response is within an acceptable range
[42]	2015	CMPN-based PI controller	- Quick response - Reduced oscillations in power waveforms - Optimum active current is injected during fault time
[43]	2016	Resonant controller	- Reduced oscillations in DC link voltage - High over current conditions
[34]	2014	Dynamic resistor braking	- Ease of implementation - No reactive current injection
[35]	2016	Quasi Z-source inverter	- Regulation of DC link voltage - Reactive power is not considered
[44]	2015	Energy storage system in supercapacitor	- Ease of design and implementation - Oscillations in power observed
[36]	2019	Conventional PI-based PQ control	- Easy to design and implement - Transfer of maximum active power into the grid during faults
[40]	2018	Model predictive control approach	- Does not support over current reduction in the grid - Quick response time
[45]	2017	Fault ride in control based on fuzzy logic	- Response is rapid and hardware implementation is easy - High over current conditions
[46]	2015	Probabilistic wavelet fuzzy-based neural network controller	- Requires complex calculation - High over current conditions

summarizes the previous research work investigating different control techniques for grid-connected PV systems, with a brief outline of their merits and disadvantages.