Improved instantaneous reactive power (PQ) Theory Based control of DVR for Compensating Extreme sag and swell

In today’s power system, power quality is a critical topic having several impacts on customers and utilities. In the current electric power system, the integration of renewable energy sources, smart grid technologies, and significant usage of power electronics equipment has generated a slew of issues. The sensitive equipment might be damaged by harmonics, voltage sag, and swell. These devices are vulnerable to Interference with other elements of the system resulting in input voltage changes. As a result, in the contemporary period, Power quality is becoming more important as the number of sensitive and costly electronic devices grows. To overcome the challenges of non-standard voltage, the Dynamic Voltage restorer (DVR) device has been extensively utilized to keep the load voltage stable. To have a dynamic and fast response of the DVR a modified instantaneous reactive power (PQ) theory is proposed to control DVR under extreme transient voltage circumstances. The proposed technique is based on the extraction of the positive sequence component of grid voltage and the negative sequence component of load current for generating a voltage reference signal. The power system network with the proposed PQ control scheme is investigated and assessed under various scenarios to compensate for severe balanced, unbalanced (voltage sags and swells), and load change. MATLAB/Simulink is used to verify the mathematical models of the conventional PQ and proposed PQ control system for DVR. The complete system is implemented experimentally using a dSPACE 1104 based laboratory system to validate the presented control scheme. The simulation and experimental results are correlated, demonstrating the efficacy of the suggested modified PQ control technique.


I. INTRODUCTION
Power quality is a criterion of a pure and regularized power supply in terms of load. Many factors, including sensitive, non-linear loads, the integration of distributed generation (DG), and advancements in power electronics equipment affect the power quality of the grid [1] [2]. Electrical power quality has massive consideration in the electrical distribution system. The major source of concern is power quality issues relating to distribution system voltage stability [3]. Voltage sag and swell are the two most common power quality issues that impact sensitive loads [4]. Voltage sag is defined as an abrupt fall in the amplitude of supply voltage to a level of 10-90% [5] [6]. Short circuit faults in power systems lead to voltage sags. Many disadvantages have been explored in recent years resulting from voltage sags, such as electrical equipment malfunctioning, loss in the manufacturing line, and complete equipment failure [7] [8]. Voltage swell is defined as the rise in voltage level to 110-180% of its rated value. Voltage swell is the result of sudden disconnection of the large load, open circuit faults. This issue will lead to insulation breakdown, overheating of electrical equipment, and damage to electronic equipment.
Essam et al. state that Voltage sag is a serious power quality issue that plays havoc on sensitive loads in the distribution system [9]. To compensate for power quality issues, power electronics-based devices like power filters, unified power flow controller (UPFC), static compensator (STATCOM), distribution static compensator (DSTATCOM), and dynamic voltage restorer (DVR) are used [10] [11]. DVR requires a complex control mechanism to safeguard important distribution system loads from power quality issues [12].
DVR is used on distribution feeders to safeguard loads from problems caused by voltage sags and swells. DVR is linked in series with the load, and a battery energy storage system (BESS) is coupled with a transformer and inverter, which adjust the active and reactive power requirements for voltage sags and voltage swells [13]. For voltage stability, the DVR injects voltage into the distribution system, which is connected to the system through the transformer [14]. DVR is a FACTS device that adjusts for disturbances caused by loads such as voltage sags, swells, and voltage harmonics.
In normal settings, DVR injects voltages in series with the transmission lines and injects a modest amount of voltage. When a disturbance occurs, however, DVR calculates the voltages needed to safeguard the load using sinusoidal pulse width modulation (SPWM) [15]. The voltages are then fed back into the system to keep the condition stable. DVR either absorbs or delivers active or reactive power in the steady-state, but when a disturbance occurs, DVR either delivers or absorbs active or reactive power from the dclink. Martiningsih et al. have advocated installing a DVR in a PT DSS power plant, where the DVR functions as a compensator and is linked in series with the distribution line. The suggested PI-based DVR is capable of recovering power quality constraints [16] [17]. Eltamaly et al. developed a DVR-based technique for reducing voltage sag using DVR to improve the quality of power systems. To deterioration in electrical equipment performance. The findings show that DVR properly compensates for sag/swell and implements suitable voltage adjustment [18]. To alleviate symmetrical and asymmetrical sags and swells, J. Han et al. presented a unique DVR with a power electronic transformer (PET) [19]. The findings show that the unique design efficiently alleviates the symmetrical and asymmetrical problems.
The DVR control strategy can protect the load from power quality issues related to voltage [5]. To have proper control, a perfect reference generation technique must be implemented. Various approaches for reference generation have been suggested, including, Clark's and Park's transformations, Phasor parameter and, Symmetrical components, Instantaneous real and reactive power [20]. Park's transformation is a mathematical transformation approach used to simplify the analysis of three-phase circuits is direct-quadrature-zero (dq0) in electrical engineering. The application of the dq0 transform on threephase circuits reduces the three AC values to two DC quantities [8] [21]. The inverse transform is used to reconstruct the real three-phase AC results using simplified computations on these imaginary DC variables. It is often used to ease the study of three-phase synchronous machines as well as calculations for three-phase inverter control [22]. When applied to three-phase voltages and currents, the dq0 transform is The Phase Locked Loop (PLL) must create a signal with the same fundamental frequency and phase angle as the reference signal generation for the two approaches Clarke and Park's transformation [23].
The Phasor parameter or Phase Locked Loop (PLL) is a control system that attempts to create an output signal whose phase is allied to the phase of the input "reference" signal. It is an electrical circuit that consists of a phase detector and a variable frequency oscillator [24]. This circuit checks the phase of the input signal to the phase of the signal obtained from its output, then changes the frequency of its output signal to maintain the phases in synchronism [25].
The "Generalized Theory of Instantaneous Active and Reactive Power, "or "Theory of Instantaneous Power," or simply "PQ Theory System with a neutral wire was briefly mentioned in the original development of the theory, which was intended for three-phase, three-wire systems [7] [26] [27]. Afterward, it was expanded to three-phase four-wire systems (systems having phases a, b, and c, in addition to a neutral wire) After performing an algebraic translation (Clarke transformation) of the three-phase voltages and currents in the abc coordinates to the αβ coordinates, the PQ theory instantaneous power components are calculated [6] [28] [29].
In this article, an improved PQ method is proposed for the generation of reference signals related to the positive sequence component of the grid voltage and negative sequence component of load current. The appearance of load current negative sequence component arises due to power quality issues such as voltage sag, swell, load change, harmonic effect, balance and unbalanced load [30]. The main advantage of the modified technique is the nonutilization of low-pass filters, because of which the disadvantages as phase shifting and insufficient compensation. Comparison analysis between the performances of traditional and proposed PQ methods is presented through different scenarios of power quality issues. The proposed method is superior in detecting and compensating the power quality issues.
The paper is organized in the sections as followed. Section II provides a discussion on dynamic voltage restorer (DVR). Section III describes the proposed PQ algorithm. The experimental setup is described in Section IV. The result illustration, as well as a discussion, is presented in section V. The performance of the proposed DVR control system is summarized in Section VI.

II. PHILOSOPHY DYNAMIC VOLTAGE RESTORER (DVR):
Injection of compensation voltage with the precise magnitude and frequency is required to restore the load side voltage to the proper amplitude and waveform. The system may inject up to 50% of the normal voltage for a brief period of time (up to 0.1 seconds). Most voltage sags, on the other hand, are much below 50%. Dynamic voltage restoration or control is the term used to describe this (DVR) [31]. A dynamic voltage restorer (DVR) is described as the regulating device [12] [16] [32]. End-users who are experiencing power quality problems may benefit from DVRs [33]. Figure 1   In restoring the load side voltage to the required level , a compensating voltage of the required magnitude and frequency must be injected [36][37][38][39]. The system can inject up to 50% of the rated voltage, but only for a brief period (up to 0.1 seconds). Normally voltage sags, are far smaller than 50%. A dynamic voltage restorer (DVR) is the regulating device. DVRs may be useful for end-users who are prone to undesirable power quality issues [40]. DVRs are often mounted on the main feeder, delivering active power through DC energy storage and generating the requisite reactive power internally [30] [41].

A. DVR OPERATING MODES
A DVR's switching states are classified into three categories based on operating states: protective, standby, and voltage compensation [42]. In the protective state, if the load current exceeds the allowable value owing to a short circuit or a significant overcurrent current, the DVR will be disconnected from the grid [43]. The inverter shorts the secondary winding of the injection transformer. In the standby state, allowing full load current to flow through the primary winding. In this operation mode, the DVR will not inject any correction voltage into the power grid. In the voltage compensation state, the DVR injects the appropriate compensatory voltage via injection transformer to the grid [44]. This mode of operation begins when a load voltage disturbance is detected and terminates when the voltage returns to normal operating conditions.

B. CONTROL STRATEGIES AND ALGORITHMS OF DVR
The detection of voltage disturbances is the major emphasis of the DVR's control system. Specifically, with sensitive loads, the detecting system should be fast enough to identify the voltage disturbance accurately for the assessment of DVR performance [20] [45]. As shown in  Table  I.

C. PROPOSED SYSTEM DESCRIPTION
The proposed configuration shown in Figure 3 includes a supply (grid) voltage with grid impedance, a three-phase load, an injection transformer, and the DVR system. The DVR system comprises a Voltage Source Inverter (VSI) powered by a DC power source with a dc link capacitor and a harmonic passive filter. A three-phase balanced, and unbalanced load is considered in this system [16] [47] [48].

III. INSTANTANEOUS REACTIVE POWER CONTROL TECHNIQUE
Generalized Theory of the Instantaneous Reactive Power in Three-Phase Circuits," also known as "Theory of Instantaneous Real Power and Imaginary Power," or "Theory of Instantaneous Active Power and Reactive Power," or "Theory of Instantaneous Power," or simply "PQ Theory," was proposed by Akagi et al. in 1983 for the control of active filters in three-phase power systems. The idea was first formulated for three-phase, three-wire systems, with a brief mention of neutral wire systems. Later, it was expanded to include three-phase four-wire systems (phases a, b, c, and neutral wire) [6] [29].

A. PQ THEORY
PQ theory is a time-domain representation of instantaneous power. This theory is based on Clarke's Transformation.
The voltage and current are transformed from abc coordinates to αβ0 coordinates. This method consists of a real matrix that transforms three-phase voltages and currents into the αβ0 stationary reference frames [5], [51]. With the help of Clarke's transformation, the voltages and current can be related in terms of abc and αβ0 as follows: And the three phase abc currents to αβ0 The instantaneous active (P), reactive(Q), and zero sequence instantaneous power (P0), can be represented in αβ0 of instantaneous phase voltage and current values using the following matrix: The instantaneous three phase active power can be represented by using αβ0 or abc components of voltage, current as shown in following equation In balanced three phase system zero sequence components of voltage and current V0 and I0 can be neglected then the instantaneous three phase active power is: Similarly, the instantaneous reactive power (Q) in αβ system can be represented as From above two equations both P and Q can be represented As PQ control technique is intrinsically a 3Φ system, used in with or without a neutral line and balanced or unbalanced. it can be implemented in both steady-state and transient state conditions. It concedes two control strategies, one constant instantaneous power and another is sinusoidal supply current [5] [21]. The power components of the PQ scheme are shown in Figure 4.

B. THE TRADITIONAL PQ CONTROL TECHNIQUE
This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.  Figure 5 shows the control model for generating Vref using Traditional PQ. Where the three-phase voltage and Load current are measured. The three-phase grid voltage is sensed and processed through an antialiasing filter and then αβ positive components of grid voltage ( αβ + ) is obtained as given in equation (1). Simultaneously the load current is measured and αβ negative component of load current ( αβ  − ) is obtained as shown in equation (2). Both the components are processed to calculate the real (P) and reactive power (Q) using equations (5) and (6), later P, Q are processed through a low pass filter for the generation of real and reactive power in αβ components where with reference voltage calculation block and inverse Clark's transformation the voltage reference Vref is generated.

C. THE MODIFIED PQ TECHNIQUE
This improved method refers to the generation of a reference signal for compensating the power quality issues. This reference depends on the grid voltage and negative sequence component of load current. Power quality concerns like voltage sag, swell, load change, harmonic impact, balance, and unbalanced load induces the load current negative sequence component. To obtain the compensation, there are two issues firstly the magnitude of the compensation, and secondly the phase of the reference signal which is locked with the grid voltage. For the magnitude of the compensation power, the load current negative sequence component has to be evaluated. The magnitude of the compensating power is corresponding to the level of the negative sequence component of the load current which is given by the following equation.
The phasing of the reference signal is obtained by using the source side voltage by computing it from the positive sequence component of the grid voltage. It is evaluated with the following equation To obtain the phasing, amplitudes of positive sequence components are normalized at the desired value using the maximum amplitude detection as following: The grid voltages ( ) and the load currents ( ) are measured in per unit (pu). Then Clark's transformation is used to transform the three-phase positive sequence components of and negative sequence components of the currents signals into two signals ( α + , β + ) and ( α − , β − ) using: The instantaneous active and reactive power (p and q ); in terms of positive sequence components of grid voltage and negative sequence components of the load currents; are calculated as: The above equation in the matrix form can be expressed as: The inverter of the DVR is supplied from a constant DC power source. consequently, there is no need to control dc voltage. The reference voltages used to control the inverter can be obtained by the inverse of the above matrix as follow: These reference signals are then transformed into abc coordination using inverse Clark's transformation as: Figure 6 explains the proposed method for voltage reference generation.

D. HYSTERSIS VOLTAGE CONTROL
The switching signals for the voltage source inverter are generated using a hysteresis voltage controller. The load voltage references are compared with the measured load voltages. The error signals are then processed through the hysteresis band. This scheme can be seen in Figure 7. A hysteresis voltage controller has the benefits of effective dynamic response, superior precision, cheap cost, and ease of implementation, the hysteresis controller is favored over standard controllers such as PWM and SVPWM [52]. The hysteresis control approach overcomes the problems of classic systems such as switching losses with the high switching frequency, electromagnetic interference difficulties owing to higher-order harmonics, and a reduction in available voltage [2] [53]. Actual load voltages are compared to the produced voltage references. Figure 8 shows how the hysteresis band is used to process the difference between reference and actual voltage values to create the firing signals of inverter IGBT switches.

FIGURE 7. HYSTERESIS VOLTAGE CONTROLLER
Asymmetrical components (positive and negative sequence components) appear in the system grid voltages and load currents through abnormal operating. The proposed control system is designed to extract the positive sequence component from the grid voltage and the negative sequence component from the load current. The active and reactive power are calculated using the positive grid voltage and negative sequences of the load current using the equation (11), (12). To generate the reference of Compensating voltages, equation (15) is used to produce the reference voltage in (αβ) coordinate system, these (αβ) are transformed to abc coordinate system using inverse Clarke transformation by equation (16). Hysteresis voltage control is used to generate firing signals for VSI switches by comparing reference voltages to real load voltages. This reference generation scheme enables the control system to compensate for the load voltage by identifying the negative sequence components under any condition of power quality issues like voltage sag, swell, load unbalanced harmonics, phase failure for a short time, and load change.
This improved PQ control method does not include the phase-locked loop (PLL), conventional PI controller, and filters for the generation of voltage references, leading to the instantaneous, continuous and fast dynamic response for compensating the load voltage.  Figure 5 shows the positive, negative, and zero components under abnormal situations. The approach pursues to extract the positive sequence components of grid voltage which are three phasors with identical magnitudes that are displaced by 120° and rotate anticlockwise [54].

A. POSITIVE SEQUENCE COMPONENTS CALULATIONS
Original Signal +V sequence -V sequence Zero sequence where tph is the phase shift (Φ) in time between two phases, f is the fundamental frequency. The symmetrical components of three-phase ungrounded voltages system can be expressed as: The symmetrical components equations are converted to time-domain to separate the positive sequence components as follow: wherever V 1 (t) is the positive sequence component in time domain interpretation of the 3-phase grid voltages, T α1 and T α2 are the time phase angle shift of the symmetrical component. Wherever, T α1 = t + t ph − t α1 and T α2 = t + t ph + t α2 . Figure 6a shows the removal of grid voltage negative sequence component.

IV EXPERIMENTAL SETUP
The experimental setup for the proposed DVR system is shown in Figure 8. It is comprised of two circuits: one for the power and the other for control. The full experimental circuit consists of a three-phase AC voltage source, a transformer, VSI powered by a DC power source, and a 1kVA load. The control circuit is based on the digital signal processing type dSPACE (DS1104) to execute the proposed system, which was employed using Matlab/Simulink. LV25-P voltage transducer circuit is used to measure the supply and load voltages. LA25 current sensor circuits are used to sense the load currents. Measured signals are scaled down to 10V before being given as inputs to the DS1104 board to fulfil the control board's requirements.

FIGURE 8. EXPERIMENTAL SETUP
The system parameters are presented in Table 4.

V. RESULTS AND DISCUSSIONS
To verify the proposed modified PQ control system for reference generation the complete system model is implemented in two different ways. Firstly, MATLAB simulation of the system based on "mathematical equations 1 to 16" is implemented and secondly experimental setup using "DSPACE DS1104 control board". Different scenarios of severe power quality conditions are extracted to verify and validate the efficacy of the proposed PQ scheme. The different scenarios of balanced, unbalanced, (sag and swell) and load reduction are considered and discussed in this section.

CASE 1: BALANCED LOAD CONDITION a) 20% SAG OF GRID VOLTAGE IN BALANCED LOAD CONDITION
In this case of balanced load condition, the three-phase grid voltage is affected by a 20% sag. Figure 9A shows the simulation results of the grid, load and compensating voltage for the balanced load voltage scenario using the conventional PQ technique. Here, the sag is starting at the instant of 0.2 seconds (s). With conventional PQ, it is observed that the load voltage waveform is distorted, and the associated compensating voltage is injected at almost the same instant of 0.2S. Figure 9B shows simulated waveforms of Gird voltage, load voltage, and appropriate compensating voltage with the proposed PQ control with 20% sag of grid voltage under balanced load conditions. Under this balanced condition, the compensating voltage is injected instantaneously at 0.2s. The load voltage waveform shows very low distortion and a good voltage profile when the proposed PQ control is used to compensate for 20 percent sag.
The comparison of the obtained total harmonic distortion (THD) in grid voltage and load voltage for the traditional and proposed PQ technique is shown in Figure 9C. The THD value is measured before, during, and after This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication. The amplitude of the grid voltage and the load voltages are also measured and compared before and during compensation. The comparison for amplitude with traditional and proposed PQ methods is shown in Figure  9D. It can be observed that the grid voltage with traditional and proposed PQ methods is the same. Whereas a decent load voltage is maintained nearer to 1 by the proposed PQ method. The traditional PQ method load voltages were not well compensated as the modified PQ method under conditions before and during compensation.  Figure 10A shows the experimental results using the traditional PQ control technique implemented for a balanced 20% sag scenario. It can be observed that the load voltage has distortions even after compensating voltage is injected at 0.075 s. The complete compensation is not obtained with the traditional PQ. On the other hand, the experimental results of the proposed PQ control for 20% sag under balanced conditions provide an instantaneous injection of compensating voltage at 0.075 s, leading to enhanced load voltage quality, as illustrated in Figure 10B.

EXPERIMENTAL RESULTS OF GRID VOLTAGES, LOAD VOLTAGE AND DVR VOLTAGES (A) TRADITIONAL PQ (B) MODIFIED PQ UNDER BALANCED 3Φ LOAD VOLTAGE SAG OF 20%
The per-unit (pu) load voltage and load frequency of the traditional and proposed PQ techniques are compared in Figure 11 under the scenario of a 20% balanced sag.  Figure 12A represents the simulation results of the grid, load, and compensating voltages with traditional PQ under balanced load conditions with 70% swell in grid voltage. It can be observed that the swell in the system is initiated at 1.1 s. The corresponding compensation voltage is also injected at 1.1 s promptly. Though with the instantaneous injection of the compensation voltage the load voltage is severely distorted with still the load voltage having a swell of around 20% from 1.12 s to 1.14 s. Figure 12B illustrates the simulation results of the modified PQ where the compensation voltage is injected instantaneously at 1.1 s. In this case, the load voltage is compensated to the maximum with a very low swell after compensation of around 10% between 1.12 s to 1.14 s. The load voltage has shown a fair improvement in the voltage profile and excellent compensation with the proposed PQ control technique. The comparison of total harmonic distortion (THD) in grid voltage and load voltage for traditional and proposed PQ techniques with 70% swell under balanced conditions is shown in Figure 12C. It is observed for grid voltage, both THDs with traditional and proposed methods are showing almost equal values for before and after compensation. Whereas for load voltage, the THD is much improved with the proposed PQ method for all the conditions of before, during, and after compensation. The comparison of the amplitude of grid voltage and load voltage before and during with traditional and Proposed PQ methods is shown in Figure 12D. It can be observed that the grid voltage measured with traditional and Proposed PQ methods are same. With the traditional PQ method load voltages are slightly greater than one in all the phases under conditions before and during compensation. Whereas a perfect load voltage of 1 pu is maintained by the proposed PQ method.  Figure 13A shows experimental results of the grid, load, and compensating voltages using the traditional PQ approach for the balanced 70 % swell. Under these conditions, the swell at 0.052 s is observed and instantaneous compensating voltage is injected resulting in a much-distorted load voltage with swells of around 30% still existing in the system from 0.052 s. Figure 13B explains the experimental results of the grid, load, and compensation voltage for the proposed PQ technique. It is observed that the load voltage from 0.052 s has an improved voltage profile with less distortions with a low swell of 10% when compared to the traditional PQ technique.  One of the phases is influenced by a 20% sag in this unbalanced scenario. Figure 15A shows the simulation results for this unbalance scenario with grid voltage, load voltage, as well as compensating voltage using traditional PQ. Here, the sag in a single-phase starts at 2 s. The associated compensating voltage is injected almost at the same instant of 2 s. Figure 15B shows simulated waveforms of Gird voltage, load voltage, and appropriate compensating voltage with the proposed PQ control under 20% sag in one phase. It is observed that the load voltage waveform has very low distortion with the initiation of compensation voltage at 2 s. The load voltage waveform shows a good voltage profile when the proposed PQ control is used to compensate for 20% sag in one of the phases.

FIGURE 11 [UPPER] PU LOAD VOLTAGE, [LOWER] LOAD FREQUENCY UNDER BALANCED 3Φ LOAD VOLTAGE SAG OF 20% b) 70% SWELL OF GRID VOLTAGE IN BALANCED LOAD CONDITION
The comparison of total harmonic distortion (THD) for grid voltage and load voltage with traditional and proposed PQ technique having 70% swell under unbalanced condition is shown in Figure 15C. It is observed for grid voltage, both THDs with traditional and proposed methods are showing nearly equal values for before and after compensation. And in during compensation of grid voltage THD with traditional PQ is 0.18% whereas with proposed PQ the THD value is improved to 0.03%. For load voltage, the THD is much improved with the proposed PQ method for all the conditions of before, during, and after compensation. Especially in "during compensation" with traditional PQ THD obtained is 18.42% whereas, with proposed PQ THD in source voltage is improved to 1.35%.
The comparison of the amplitude of grid voltage and load voltage before and during with traditional and Proposed PQ methods is shown in Figure 15D. It can be observed that the grid voltage measured with traditional and Proposed PQ methods are the same. With the proposed PQ the load voltage had shown a more satisfactory value of nearer to 1pu when compared to the traditional PQ control.  Figure 16A shows the experimental results using traditional PQ control for an unbalanced 20% sag scenario. Figure 16A explains that one of the phases is having sag at 0.053 s and an instantaneous injection of compensating voltage is performed. The resultant load voltage has still some distortion and unbalanced conditions. Figure 16B illustrates the grid, load, and compensation voltage for experimental results of the proposed PQ technique. It can be observed with the instantaneous injection of compensation voltage at 0.053 s the load voltage has improved voltage profile with the proposed PQ method when compared to the traditional PQ. The per-unit load voltage and load frequency of the traditional and proposed PQ techniques are compared in Figure 17 under the scenario of a 20% unbalanced sag.   Figure 18A shows the simulation results of the grid, load, and compensating voltages with traditional PQ for an unbalanced condition during a single phase 70% swell. In this case, the load voltage is severely distorted, and the swelled phase still has a slight swell without total compensation. On the other side, by using the proposed PQ control technique, excellent compensation is observed with perfect compensation of voltage. Figure 18B shows simulation results of the grid, load, and compensating voltages using the new PQ approach for an unbalanced single phase 70 % swell.

SIMULATION RESULTS OF GRID VOLTAGES, LOAD VOLTAGE AND DVR VOLTAGES (A) TRADITIONAL PQ (B) MODIFIED PQ (C)THD Of VOLTAGE (D) VOLTAGE AMPLITUDE UNDER UNBALANCED 3Φ LOAD VOLTAGE SAG OF 20%
THD for grid voltage and load voltage with traditional and proposed PQ technique having 70% swell in one phase is shown in Figure 18C. It is observed for grid voltage, both THDs with traditional and proposed methods are displaying similar weights for before, during, and after compensation. For load voltage, the THD is much improved with the proposed PQ method for all the conditions of before, during, and after compensation. The measured THD values with modified PQ are 1.33, 1.17, and 1.33 under before, during, and after compensation respectively.
The comparison of the amplitude of grid voltage and load voltage before and during with traditional and Proposed PQ methods is shown in Figure 18D. It can be observed that the grid voltage measured with traditional and Proposed PQ methods are identical. With the proposed PQ the load voltage had shown suitable voltage amplitudes equal to 1 pu when compared to the traditional PQ control.  Figure 19A illustrates the experimental results of a 70% swell unbalanced three-phase grid using the traditional PQ control strategy. Under these conditions, the load voltage has more distortion and still has swell. The experimental results of the grid, load, and compensating voltage for the proposed PQ technique are shown in Figure 19B. It can be observed that the 70% swell in one phase is observed at 0.05 s. The compensation voltage is injected instantaneously at 0.05 s. The load voltage profile has improved with a completely compensated swell under this proposed PQ control.   In this scenario, the system is subjected to a 50% drop in load. The simulation waveforms of the grid, load, and compensated voltages, using the traditional PQ and improved PQ are shown in Figures 21A and 21B respectively. It can be observed that the load voltage has been dropped and distorted from 3.3 s. The load decrease does not affect grid voltages. However, the impact is visible on load voltages, where the load decrease has a significant influence when utilizing the traditional PQ. The correction is substantially quicker and more efficient when using the proposed PQ.
THD for grid voltage and load voltage with traditional and proposed PQ technique with a 50% reduction in load is shown in Figure 21C. It can be observed for grid voltage, THDs with traditional and proposed methods are displaying the same THDs for before, during, and after compensation. For load voltage, the THD with traditional PQ measure at before, during, and after compensation is 1.7%, 18.23%, and 4.89% respectively. With the proposed PQ method for all the conditions of before, during, and after compensation. The measured THD values are 1.14%, 4.23%, and 1.41% under before, during, and after compensation respectively. These THD values suggest the effectiveness of the proposed PQ method.
The comparison of the amplitude of grid voltage and load voltage before and during with traditional and Proposed PQ methods for 50% load change is shown in Figure 21D. It can be observed that the grid voltage measured with traditional and Proposed PQ methods are identical with 1 pu for all scenarios of before, during, and after compensation. With the proposed PQ the load voltage had demonstrated an acceptable voltage quality of nearly 1 pu when compared to the traditional PQ control.  Figure 22A explains the experimental results of 50% load change using the traditional PQ control strategy. Under these conditions, the load voltage has more distortion from the instant of 50% load change from 0.05 s. Whereas, the grid voltage remains constant at 1 pu. The experimental results of the grid, load, and compensating voltage for the proposed PQ technique are shown in Figure 22B. It can be observed under the 50% load change load voltage has been well compensated with very few distortions as depicted in Figure 22C. From Figure 22D, the load voltage profile has improved completely under this proposed PQ control for 50% load change.

CONCLUSION
The paper provided a method for compensating for voltage disturbances in the DVR. The proposed approach improves the quality of load voltages by protecting them against grid voltage abnormalities. The proposed DVR control approach is based on a modified version of PQ theory that employs a detection method for the positive and negative sequence components. The detection technique is carried out in the time domain. The efficiency of the proposed approach is assessed using extensive simulations in MATLAB/Simulink and experiments under several special disturbance scenarios: severe unbalanced sag and swell, load change, and voltage harmonics. The proposed method has shown capability in improving the voltage quality as well as the voltage profile. The results have emphasized the applicability of the proposed DVR compensation method. To sum up, the following advantages summarize the performance of the proposed system: • Less computational effort.
• Faster response. This article has been accepted for publication in IEEE Access. This is the author's version which has not been fully edited and content may change prior to final publication.