S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 171 Article An effective power quality enhancement system for integrated photovoltaic cells utilizing cascaded ANFIS in a unified power quality conditioner Saritha Kandukuri1*, Ramesh Guguloth2, A. Sivakumar 3, I. Shivasankkar4, Ananthan Nagarajan5, N. Janaki6 1Department of Electrical and Electronics Engineering, Eklavya University, Damoh, Madhya Pradesh, 470661 2Department of Electrical and Electronics Engineering, Sreenidhi Institute of Science and Technology, Hyderabad, Telangana- 501301, India 3Department of Electrical and Electronics Engineering, Panimalar Engineering College, Nazerthpattai, Poonamalle, Chennai, 600 123, India 4Department of Electrical and Electronics Engineering, Sri Manakula Vinayagar Engineering College, Puducherry, 605107, India 5Department of Electrical and Electronics Engineering, Vel Tech Multi Tech, Dr. Rangarajan Dr. Sakunthala Engineering College, Chennai-62, India 6Department of Electrical and Electronics Engineering, Vels Institute of Science, Technology and Advanced Studies, Chennai, India A R T I C L E I N F O Article history: Received 22 April 2025 Received in revised form 04 June 2025 Accepted 14 June 2025 Keywords: Unified Power Quality Conditioner, Artificial Neural Network controller, PV system, Coupled quadratic SEPIC converter Cascaded ANFIS-MPPT *Corresponding author Email address: mugatha.saritha@gmail.com DOI: 10.55670/fpll.futech.4.3.16 A B S T R A C T The arrival of power electronic devices for the control of loads has an effect on the Power Quality (PQ) at the utility grid’s distribution side. Meanwhile, PQ problems cause malfunctioning equipment, lost production time, loss of money for industry, inconvenience, and possible damage to household electrical appliances. Thus, the requirement for increased system efficiency is essential. Hence, this study proposes the control of a Unified Power Quality Conditioner (UPQC) in conjunction with a Photovoltaic (PV) system. Shunt and series converters attached back-to-back via a shared DC-link make up the PV-UPQC system. Subsequently, the Artificial Neural Network (ANN) controller reduces PQ problems and simplifies the control complexity. A Coupled quadratic Single Ended Primary Inductor Converter (SEPIC) connects the PV system to UPQC, and the Cascaded Adaptive Neuro Fuzzy Inference System- Maximum Power Point Tracking (ANFIS-MPPT) technique enables the optimization of power extraction from PV sources. The developed approach is implemented using the MATLAB/Simulink platform, and its performance is evaluated for Total Harmonic Distortion (THD), sag, and swell. The results show that the control maintains THD within the B-phase THD of 3.97% and R and Y phase THDs of 4.82% and 4.86%, and also obtained a voltage gain ratio of 1:15; the output levels increase substantially with reduced voltage stresses on the switching devices. 1. Introduction The usage of non-linear loads and unbalanced loads has increased in the modern era due to the expansion of the distribution system and the expansion of industry. PQ problems get inferior during the non-linear load enhancements, and the s distribution grid's structure becomes more intricate [1]. This resulted in issues with PQ, such as distortion and imbalance in the current, sag/swell, and the production of harmonics and imbalance in the system’s supply voltage. Voltage quality issues, in particular, have the potential to impair the regular functioning of sensitive loads that are heavily linked to the distribution grid, resulting in financial losses and other consequences [2]. The essential industrial load is affected by grid voltage disruptions, which result in frequent tripping. In modern years, a number of methods and tools have been established to address PQ problems in distribution networks. Flexible AC Transmission System (FACTS) devices are appealing instruments for improving reactive power control and reliability in transmission systems. These gadgets react swiftly to any disruptions and provide more system flexibility [3]. Future Technology Open Access Journal https://doi.org/10.55670/fpll.futech.4.3.16 Journal homepage: https://fupubco.com/futech ISSN 2832-0379 August 2025| Volume 04 | Issue 03 | Pages 171-181 mailto:mugatha.saritha@gmail.com https://doi.org/10.55670/fpll.futech.4.3.16 https://fupubco.com/futech S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 172 The need for Passive Power Filters (PPF), Active Power Filters (APF), and hybrid power filters has increased due to these limitations, which include fixed compensation, massive size, difficulty in adjusting dependency filter settings, and resonance with source impedance [4-5]. These filters, which are often connected in parallel with the load, are developed to remove current harmonics and adjust for reactive power in the power system. Despite being more affordable and widely accessible, these filters must be retuned to a specific harmonic in order to produce the desired effect, which can lead to overvoltage situations when power demand is lower [6]. The STATCOM is a power electronics device that works by injecting reactive current into the power network’s point of common coupling. The primary benefit of the STATCOM is that it does not rely on the Point of Common Coupling’s (PCC) voltage level, hence the compensating current is not reduced as the voltage drops. Nevertheless, it has harmonics, high initial costs, and limited steady-state operating modes [7]. By reducing major PQ problems, including sags/swell, harmonics, flickers, and interruptions, the DVR protects the load from failure or tripping; nonetheless, they are ineffective at balancing large-scale voltage sags [8]. Through the regulation of voltage, power factor, and harmonics, the Static Var Compensator (SVC) enhances PQ. However, in order to compensate for surge impedance, SVCs need extra equipment [9]. The PCC provides reactive power to the Distribution Static Compensator (DSTATCOM), which regulates voltage. Nevertheless, its use is restricted by the issue of reactive power injection and power losses [10]. Therefore, this research proposes a UPQC for enhancing the PQ. The UPQC protects the vital loads connected to the distribution system by addressing issues such as neutral and negative sequence currents, harmonic isolation, flow of reactive power at harmonic distortions, voltage disturbances, and harmonic and fundamental frequencies. A PV system is exploited in a UPQC system to leverage the clean, renewable energy developed by solar panels to alleviate PQ issues, which have the highest annual growth curve among the available renewable sources because of their easy installation and limitless supply capacity. However, many PQ problems are also brought on by the extensive integration of PV into the power grid [11-12]. Thus, the design of solar PV integrated UPQC has many advantages, including enhancing grid PQ and shielding vital loads from grid-side disruptions. Furthermore, the current approaches ignore the problem of voltage instability brought on by PV system intermittency in favour of concentrating solely on the compensating capability and design of UPQC [13]. The conventional converters like Boost [14], Cuk [15], and SEPIC [16] are employed for boosting the voltage of the PV system. However, these conventional converters have a complex structure, high ripple current, and lower efficiency. Therefore, this research develops a coupled quadratic SEPIC converter in the PV-based UPQC system. To enhance the efficacy of the PV system, the MPPT approach is utilized that tracks the highest power from the PV system [17]. The conventional MPPT algorithms like ANN [18], Fuzzy logic [19], and ANFIS [20] have oscillations, undesirable performance, and excessive complexity. Also, the Perturb and Observe (P&O) MPPT [21] method has limitations in terms of oscillations around the Maximum Power Point (MPP), causing a loss in power, its inability to track rapidly changing irradiance conditions, and a lower efficiency with dynamic conditions. Similarly, Incremental Conductance MPPT [22] has a high computation burden, slower tracking with rapidly varying irradiance, and it is also sensitive to noise that causes small oscillations around the MPP. As a consequence, this paper develops a cascaded ANFIS MPPT algorithm for tracking the peak power from the PV system. 1.1 Problem statement PQ issues such as voltage sags, swells, and harmonic distortion are a growing problem in modern power systems with increasing nonlinear loads and distributed energy resources. Poor PQ causes equipment failure, production shutdowns, and financial losses. While UPQC is commonly employed to mitigate these issues, its implementation leads to complex control requirements and wasted energy extraction when combined with renewable sources. This research presents a novel PV-UPQC system, integrated using a Coupled Quadratic SEPIC converter, an ANN controller, and a Cascaded ANFIS-MPPT design, making it possible to improve efficiency and PQ. The key contributions are: • Integrating the UPQC for mitigating the PQ issues like voltage sag and swell. • Implementing the Coupled quadratic SEPIC converter for enhancing the low voltage of the PV system to a higher voltage. • The cascaded ANFIS MPPT is exploited for tracking maximum power from the PV system, which effectively enhances the PV system’s efficacy. • ANN controller approach to minimize control generalization and expand PQ mitigation operations. 2. Proposed methodology The developed PV-based UPQC system is indicated in Figure 1. The three-phase AC supply is connected with a linear/nonlinear load through a UPQC, which has a series and shunt converter with a DC link capacitor. The series converter is exploited for compensating voltage distortions and maintaining voltage stability at the load end. It ensures a seamless power supply for linear or nonlinear loads. Then, the shunt converter is exploited to mitigate current distortions. It ensures that the current drawn by the load remains sinusoidal and balanced, even under nonlinear conditions. Abbreviations APF Active Power Filters ANFIS-MPPT Adaptive Neuro Fuzzy Inference System- Maximum Power Point Tracking ANN Artificial Neural Network DSTATCOM Distribution Static Compensator FACTS Flexible AC Transmission System PCC Point of Common Coupling PLL Phase-Locked Loop PPF Passive Power Filters PQ Power Quality PV Photovoltaic PWM Pulse Width Modulation RMSE Root Mean Square Error SVC Static Var Compensator SEPIC Single Ended Primary Inductor Converter THD Total Harmonic Distortion UPQC Unified Power Quality Conditioner S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 173 Figure 1. Proposed block diagram Then, the Pulse Width Modulation (PWM) generator generates PWM pulses for better functioning of the series and shunt converters. To give the power supply to the DC link, the PV system is exploited. Because of the environmental changes, low voltage is generated from the PV system that is improved by utilizing the coupled quadratic SEPIC converter, and its output is supplied to the DC link capacitor. For tracking the peak power from the PV system, the cascaded ANFIS MPPT controller is utilized. Consequently, the ANN controller is exploited to control the function of UPQC, and the PWM generator produces necessary pulses for UPQC. Accordingly, the PQ of the overall system is enhanced with reduced THD. 2.1 UPQC The UPQC is a power conditioning system with shunt and series compensation capabilities that effectively enhances the overall PQ of the system. Figure 2 shows the UPQC’s structural diagram. An unbalanced three-phase system’s source grid voltage 𝑉𝑔𝑟𝑖𝑑(𝑡) has fundamental and harmonics in its zero, negative, and positive sequence components. Equation (1) provides the system voltage for the equivalent circuit. 𝑉𝑔𝑟𝑖𝑑(𝑡) = 𝑉𝑔𝑟𝑖𝑑+(𝑡) + 𝑉𝑔𝑟𝑖𝑑−(𝑡) + 𝑉𝑔𝑟𝑖𝑑0(𝑡) + ∑𝑉𝑠ℎ (1) Where 𝑉𝑠ℎ is the shunt converter’s voltage and 𝑉𝑔𝑟𝑖𝑑−(𝑡), 𝑉𝑔𝑟𝑖𝑑+(𝑡) and 𝑉𝑔𝑟𝑖𝑑0(𝑡) are the negative, positive and zero sequence components. Equation (2) provides the inserted voltage of the series converter. 𝑉𝑠𝑒_𝑐𝑜𝑚𝑝(𝑡) = 𝑉𝐿𝑜𝑎𝑑(𝑡) − 𝑉𝑔𝑟𝑖𝑑(𝑡) (2) Where 𝑉𝑠𝑒_𝑐𝑜𝑚𝑝(𝑡) is the voltage of the series compensator, 𝑉𝑔𝑟𝑖𝑑(𝑡)is the source voltage, and 𝑉𝐿𝑜𝑎𝑑(𝑡)is the load voltage. The current of shunt compensator is, 𝐼𝑠ℎ_𝑐𝑜𝑚𝑝(𝑡) = 𝐼𝐿𝑜𝑎𝑑(𝑡) − 𝐼𝑔𝑟𝑖𝑑 (𝑡) (3) Figure 2. Structure of UPQC system Where 𝐼𝑠ℎ−𝑐𝑜𝑚𝑝(𝑡) is the compensating current, 𝐼𝐿𝑜𝑎𝑑(𝑡) is the current at the load and 𝐼𝑔𝑟𝑖𝑑 (𝑡) is the current passing over the grid. The current is injected into the grid by the shunt converter. 𝐼𝐿𝑜𝑎𝑑(𝑡) = 𝐼𝐿𝑜𝑎𝑑+(𝑡) + 𝐼𝐿𝑜𝑎𝑑−(𝑡) + 𝐼𝐿𝑜𝑎𝑑0(𝑡) + ∑ 𝐼𝑠ℎ−𝑐𝑜𝑚𝑝(𝑡) (4) Equation (4) provides the distorted load current. Where the load current’s positive sequence is denoted by 𝐼𝐿𝑜𝑎𝑑+(𝑡), its negative sequence by 𝐼𝐿𝑜𝑎𝑑−(𝑡), and its zero-sequence component by 𝐼𝐿𝑜𝑎𝑑0(𝑡). The current passing through the shunt compensator is denoted by 𝐼𝑠ℎ−𝑐𝑜𝑚𝑝(𝑡). This system is a three-phase system with a non-linear inductive load. Figure 3 depicts a circuit of the UPQC system. 2.1.1 Series converter By lowering voltage-related disturbances, including voltage swell and sag, the series converter enhances PQ. The series converter preserves the voltage control, as seen in Figure 4. S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 174 Figure 3. Equivalent circuit of UPQC Figure 4. Control structure for a series converter The series converter is in charge of using a series injection transformer to inject the voltage at PCC. The DC-link element is charged concurrently with the AC quantity being converted to DC by the series converter. The actual power exchange is also made possible by the series converter. The voltage sensors detect the basic and distorted voltage components at the PCC. The input voltage’s peak value is divided by the sensed distorted voltage. 𝑉𝑝𝑒𝑎𝑘 = √ 2 3 (𝑉𝑎_𝑠 + 𝑉𝑏_𝑠 + 𝑉𝑐_𝑠) (5) The three-phase frequency is synchronized using the Phase- Locked Loop (PLL) circuit. In the PLL circuit, the distorted voltage is separated by the peak voltage. Equation (6) provides the phase angle difference, 𝑉𝑃𝐿𝐿_𝑎 = 𝑠𝑖𝑛(𝜔𝑡) (6) 𝑉𝑃𝐿𝐿_𝑏 = 𝑠𝑖𝑛 (𝜔𝑡 − 2𝜋 3 ) (7) 𝑉𝑃𝐿𝐿_𝑐 = 𝑠𝑖𝑛 (𝜔𝑡 + 2𝜋 3 ) (8) 𝑉𝐿𝑜𝑎𝑑_𝑎𝑏𝑐 ∗ = 𝑉𝑝𝑒𝑎𝑘 ∗ 𝑉𝑃𝐿𝐿_𝑎𝑏𝑐 (9) An error signal is developed by comparing the generated reference signal with the load signal. The series converter’s gate pulse is developed by feeding the resultant error signal into a PWM signal generator. 2.1.2 Shunt converter In addition to compensating for current harmonics, the shunt converter also compensates for reactive power. The actual power required by the series converter at the DC link capacitor is provided by the shunt converter. The shunt converter transforms the series converter’s DC-link power demand back into an AC quantity. The shunt converter uses the shunt inductor to compensate for the power consumption on the load side. The shunt converter employs the 𝑝 − 𝑞 theory as its control system, in which Clark’s Transformation transforms 𝑎 − 𝑏 − 𝑐 coordinates into 𝛼 − 𝛽 coordinates, as represented in Figure 5. Equations (8) and (9) provide the electrical quantities in 𝛼 − 𝛽 coordinates. The reactive and real power based on the current and voltage at any given time are: ( 𝑣𝛼_𝑙𝑜𝑎𝑑 𝑣𝛽_𝑙𝑜𝑎𝑑 ) = √ 2 3 ( 1 − 1 2 − 1 2 0 √ 3 2 −√ 3 2 )( 𝑣𝑎_𝑙𝑜𝑎𝑑 𝑣𝑏_𝑙𝑜𝑎𝑑 𝑣𝑐_𝑙𝑜𝑎𝑑 ) (10) ( 𝑖𝛼_𝑙𝑜𝑎𝑑 𝑖𝛽_𝑙𝑜𝑎𝑑 ) = √ 2 3 ( 1 − 1 2 − 1 2 0 √ 3 2 −√ 3 2 )( 𝑖𝑎_𝑙𝑜𝑎𝑑 𝑖𝑏_𝑙𝑜𝑎𝑑 𝑖𝑐_𝑙𝑜𝑎𝑑 ) (11) Figure 5. Control structure for the Shunt converter 𝑝𝑙𝑜𝑎𝑑(𝑡) = 𝑣𝛼_𝑙𝑜𝑎𝑑(𝑡)𝑖𝛼_𝑙𝑜𝑎𝑑(𝑡) + 𝑣𝛽_𝑙𝑜𝑎𝑑(𝑡)𝑖𝛽_𝑙𝑜𝑎𝑑(𝑡) (12) 𝑞𝑙𝑜𝑎𝑑(𝑡) = −𝑣𝛼_𝑙𝑜𝑎𝑑(𝑡)𝑖𝛼_𝑙𝑜𝑎𝑑(𝑡) + 𝑣𝛽_𝑙𝑜𝑎𝑑(𝑡)𝑖𝛽_𝑙𝑜𝑎𝑑(𝑡) (13) Equations (12) and (13), which relate to real and reactive power, 𝑝𝑙𝑜𝑎𝑑 = 𝑝𝑎𝑐_𝑙𝑜𝑎𝑑̃ +𝑝𝑑𝑐_𝑙𝑜𝑎𝑑̅̅ ̅̅ ̅̅ ̅̅ ̅̅ (14) 𝑞𝑙𝑜𝑎𝑑 = 𝑞𝑎𝑐_𝑙𝑜𝑎𝑑̃ +𝑞𝑑𝑐_𝑙𝑜𝑎𝑑̅̅ ̅̅ ̅̅ ̅̅ ̅̅ (15) ( 𝑖𝑎_𝑙𝑜𝑎𝑑 ∗ 𝑖𝑏_𝑙𝑜𝑎𝑑 ∗ 𝑖𝑐_𝑙𝑜𝑎𝑑 ∗ ) = √ 2 3 ( 1 √2 1 0 1 √2 − 1 2 √3 2 1 √2 − 1 2 − √3 2 ) ( −𝑖𝑜_𝑙𝑜𝑎𝑑 𝑖𝛼_𝑙𝑜𝑎𝑑 ∗ 𝑖𝛽−𝑙𝑜𝑎𝑑 ∗ ) (16) The constant DC link voltage is the responsibility of the shunt converter. The phase angle δ decides the variation of reactive and real power control. The shunt voltage source converter receives the gating pulses from the PWM generator. The function of UPQC is managed with the aid of an ANN controller. 2.2 ANN controller The ANN controller’s response needs to be precise and quick for UPQC compensation. The ability of the ANN controller to learn, evaluate the mean square error, and forecast the uncertainty is needed to reduce the input-output disparity. Furthermore, the ANN controller trains the shunt and series compensators using the same method. This requires the controller to react quickly and accurately in order to correct for UPQC. In addition to processing the reference signal efficiently, the ANN controller demonstrates quick and accurate identification of a perturbed signal. Figure 6 illustrates the structure of the ANN controller. S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 175 Figure 6. ANN controller The ANN-based controller reacts quickly and dynamically under a wide range of operating conditions. All of an ANN’s inputs are received by the input layer, after which they are processed and stored in the hidden layer. Prior to further processing in the hidden layer, the input weights are multiplied by the bias. Following the completion of specific computations, the results are processed and transmitted to the output layer. ANNs handle data concurrently, which leads to quicker processing speeds than traditional systems. ANN generates reference currents and voltages by combining different learning architectures and principles. This design diagnoses the mean square error and makes both forward and backward weight adjustments until the intended output is attained, and the error is removed if the needed output is not produced. The ANN controller manages both the series and shunt converters in the PV-UPQC system. Particularly, the ANN controller creates the reference signals necessary for both converters to successfully adjust for voltage and current deviations. For the series converter, the ANN aids in adjusting the input voltage to reduce sags, swells, and harmonics, thus stabilising the load-side voltage. The ANN allows the delivery of compensatory currents into the shunt converter, thereby eliminating current harmonics and maintaining a balanced sinusoidal supply. This dual-control feature improves the general efficiency of the UPQC in handling PQ issues. To give the supply to UPQC, the PV system is equipped with a DC-DC converter. 2.3 PV System The PV system is made up of several PV cells coupled in parallel and series to produce the necessary output voltage and current. Figure 7 displays the circuit of the PV system. The solar temperature and intensity decide the supplied power of the PV system. The expression (1) is the output current generated by the solar cell: 𝐼 = 𝐼𝑃ℎ − 𝐼𝐷 − 𝐼𝑠ℎ (17) Where 𝐼 stands for the PV cell’s output current, 𝐼𝑃ℎ is photo- generated current, 𝐼𝐷 is the current of the diode and 𝐼𝑠ℎ is the shunt current. The current that is redirected through the diode is described using the Shockley diode equation as: 𝐼𝐷 = 𝐼𝑜 (𝑒𝑥𝑝 [ 𝑞(𝑉+𝐼𝑅𝑠) 𝑚𝑘𝑇𝑐 ] − 1) (18) The current in a PV cell is: 𝑰 = 𝑰𝑷𝒉 − 𝑰𝒐 (𝒆𝒙𝒑 [ 𝒒(𝑽+𝑰𝑹𝒔) 𝒎𝒌𝑻𝒄 ] − 𝟏) − ( (𝑽 + 𝑰𝑹𝒔) 𝑹𝒔𝒉 ⁄ ) (19) Where 𝑇𝑐 is the absolute temperature, 𝑅𝑠 is the series resistance, 𝑅𝑠ℎ is shunt resistance, 𝐼𝑜 is diode saturation current, 𝑞 is elementary charge, 𝐾 is the Boltzmann constant, 𝑚 is quality factor of diode and 𝑉 is the output voltage. Here, the low voltage of PV system is enhanced by a coupled quadratic SEPIC converter. Figure 7. Circuit of the PV system 2.4 Coupled quadratic SEPIC converter The coupled quadratic SEPIC converter transforms low and intermittent input voltage from the PV system to a higher voltage. Figure 8 reveals the coupled quadratic SEPIC converter. The following presumptions are taken into consideration in order to summarize the converter's principle: all of the components are ideal, the resistance of the capacitors and inductors is minimal; the ON resistance of the 𝑆, the diodes and parasitic capacitances’ voltage drop are all very small. The developed converter is operated in 3 modes, as shown in Figure 9 and Figure 10, which represent the functional waveform of the developed converter. Figure 8. Coupled quadratic SEPIC converter Figure 9. Stages of the developed converter Mode I The diode 𝐷1 and switch 𝑆 are conducting in this state. Capacitors 𝐶3and 𝐶𝑜are reverse biases 𝐷2 and𝐷𝑜, which are not conducting. With the current route 𝑉𝑑𝑐 − 𝐿1 − 𝐷1 − 𝑆 − 𝑉𝑑𝑐 , the input source (𝑉𝑑𝑐)energizes inductor𝐿1. Through the current path, windings 𝑁1and 𝑁2become magnetized as 𝑁2 − 𝑁1 − 𝐶2 − 𝑆 − 𝐶3 − 𝑁2. The output capacitor𝐶𝑜, which is separated from the DC source, powers the resistive load. The current of 𝐿1 and 𝐿𝑀increases from 𝑡0 𝑡𝑜 𝑡1. Mode II When the power switch 𝑆 is turned off and diode 𝐷1 is reverse-biased by capacitor𝐶2, in operating mode II. 𝐿1 uses S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 176 𝑉𝑑𝑐 − 𝐿1 − 𝐶1 − 𝑉𝑑𝑐 to discharge the energy it has stored in𝐶1. Through the 𝐶3 −𝑁2 − 𝑁1 − 𝑉𝑂 − 𝐶3route, the energy saved in the coupled-inductors is released to the 𝐶3 and load. The current of 𝐿1and 𝐿𝑀 reduces from 𝑡1 to 𝑡2. Mode III In mode III, 𝐷𝑜is revered as biased while the power switch is off as a consequence of the leaking inductance effect. The consequence of turning off the diode 𝐷𝑜 is disregarded in converter operation by using the appropriate magnetizing inductance, high coupling coefficient, and low leakage inductance. The voltage relation among windings 𝑁1and 𝑁2 is: 𝑛 = 𝑉𝑁1 𝑉𝑁2 (20) By applying KVL in state 1, 𝑉𝐿1 = 𝑉𝑑𝑐 (21) 𝑉𝐿𝑚 = 𝑉𝐶3−𝑉𝐶2 𝑛−1 (22) For mode 2, 𝑉𝐿1 = 𝑉𝑑𝑐 − 𝑉𝐶1 (23) 𝑉𝐿𝑚 = − 𝑉𝐶2 𝑛 (24) Figure 10. Functional waveform of the developed converter By utilizing the volt-second balance law for the inductors and magnetizing 𝐿𝑀, ∫ 𝑉𝐿1𝑑𝑡 + ∫ 𝑉𝐿1𝑑𝑡 = 0 𝑇𝑆 𝐷𝑇𝑆 𝐷𝑇𝑆 0 (25) ∫ 𝑉𝐿𝑚 𝑑𝑡 + ∫ 𝑉𝐿𝑚 𝑑𝑡 = 0 𝑇𝑆 𝐷𝑇𝑆 𝐷𝑇𝑆 0 (26) Where 𝑇𝑆 and 𝐷 denote the switching period and duty cycle of the proposed converter. 𝑉𝐶1 = 1 1−𝐷 𝑉𝑑𝑐 (27) 𝑉𝐶2 = 𝑛𝐷 (1−𝐷)2(𝑛−1) 𝑉𝑑𝑐 (28) 𝑉𝐶3 = 𝑛−1+𝐷 (1−𝐷)2(𝑛−1) 𝑉𝑑𝑐 (29) The output dc voltage is: 𝑉𝑜 = 𝑛−1+𝑛𝐷 (1−𝐷)2(𝑛−1) 𝑉𝑑𝑐 (30) The voltage gain is: 𝐺 = 𝑉𝑜 𝑉𝑑𝑐 = 𝑛−1+𝑛𝐷 (1−𝐷)2(𝑛−1) (31) The cascaded ANFIS MPPT controller is exploited for tracking the peak power from the PV system. 2.5 Cascaded ANFIS MPPT controller By continuously modifying the operational parameters, the proposed work uses a cascaded ANFIS MPPT controller to optimize the power output from the PV system. Figure 11 shows a flow chart of the developed controller. Reference voltage and current are produced by the ANFIS controller based on the PV system’s operating state at the time. Important parameters, including temperature, pressure, and the PV system’s output current and voltage, are used as inputs. The membership function fuzzifies the input variables, which are in charge of converting clear input into fuzzy sets so that the ANFIS manages the inherent uncertainty in the behavior of PV systems. Finally, a set of fuzzy rules is developed according to the past data. The relationship between input variables and output is defined by these rules. To make sure the PV system runs at its MPP, the secondary ANFIS controller modifies the developed converter’s duty cycle. The secondary ANFIS receives real-time voltage and current measurements as well as reference values produced by the primary ANFIS controller. The secondary controller controls the duty cycle changes of the converter using a rule- based and fuzzification. Figure 11. Flowchart of cascaded ANFIS-MPPT controller S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 177 2.5.1 Pair selection module By choosing the optimal input variable pairings, the goal is to increase the ANFIS model’s accuracy. Each pair of input variables is assessed according to its capacity to reduce the Root Mean Square Error (RMSE) among the actual and expected outputs in a sequential feature selection procedure. An ANFIS model with two inputs is trained and evaluated for every pair. The pairing with the lowest RMSE is chosen. 2.5.2 Training module Each data pair’s RMSE is determined by comparing the predicted and actual outputs. Iterative training is applied to the cascaded ANFIS model till the RMSE is less than a predetermined goal error. The accuracy of the model is improved by using the outputs from each iteration as inputs for the next one. Assume that the four input variables 𝑍1, 𝑍2, 𝑍3and 𝑍4. The defined optimization problem is as follows. 𝑖𝑛𝑝𝑢𝑡 = {𝑍1, 𝑍2, 𝑍3, 𝑍4} (32) 𝑖𝑛𝑝𝑢𝑡𝑝𝑎𝑖𝑟𝑠 = {𝑍1, 𝑍3}, {𝑍2, 𝑍1}, {𝑍3, 𝑍4}, {𝑍4, 𝑍1} (33) The two outputs are the result of 𝑅𝑀𝑆𝐸𝑖 and predicted output 𝑌𝑖 . 𝑅𝑀𝑆𝐸 = √(𝐴 − 𝑃)2̅̅ ̅̅ ̅̅ ̅̅ ̅̅ ̅̅ (34) 𝑅𝑀𝑆𝐸𝐴,𝑃 = [∑ (𝑂𝐴𝑖−𝑂𝑝𝑖) 2 𝑁 𝑁 𝑖=1 ] 1 2⁄ (35) 𝑓 = 𝜔1 𝜔1+𝜔2 𝑓1 + 𝜔2 𝜔1+𝜔2 𝑓2 + 𝜔3 𝜔2+𝜔3 𝑓3 + 𝜔4 𝜔3+𝜔4 𝑓4 (36) The predicted and actual results are denoted as 𝑃 𝑎𝑛𝑑 𝐴 while 𝑁 is the size and sample. The obtained outputs from 𝑌 and RMSE at the end of the first iteration. After comparing the RMSE and goal error, the subsequent iteration is selected appropriately. The distinctive aspect of this approach is that the outcomes from iterations 𝑌1, 𝑌2, 𝑌3and 𝑌4are exploited as inputs for later iterations. To extract the peak power from the PV panel, the same process is used. The trained ANFIS modules continuously monitor the PV system variables while it is operating. Based on inputs, these modules forecast the required current and voltage, which is utilized to optimize the PV power output. The controller modifies the operating conditions to keep the system running at maximum efficiency. 3. Results and discussion This section discusses the outcomes of the PV-based UPQC system for voltage swell and sag conditions. The developed research is executed in the MATLAB/Simulink tool, and a performance comparison is included to reveal the efficacy of the developed research. Table 1 depicts the parameter values of the proposed research. Figure 12 reveals the waveform of the AC source. The voltage of the AC source is 400 V in the starting period, and it is reduced to 280 V. Then, it changed back to 400 V (voltage sag is 120 V). Likewise, the current of an AC source does not maintain a stable value and experiences continuous variations. Both the voltage waveform and the current waveform from the AC source are continuously changing throughout the analysed time frame. Figure 13 represents the waveform of the developed converter. The input voltage of the developed converter is maintained at 72 V in the entire system. In the initial stage, the input current is varied and then sustained at 2500 A throughout the system. The output voltage of the converter is gradually raised and settled at 740 V. Likewise, the output current is randomly changed and maintains a value of 12A. Table 1. Specification of parameters Parameter Specification AC Source Load Resistance 100Ω Load Inductance 10𝑚𝐻 PV system Total Power 10K W Voltage (Open circuit) 22.6 𝑉 Current (Short circuit) 8.95 A Maximum Peak Current 8.35 𝐴 Number of panels in series connection 2 Number of panels in the shunt connection 17 Coupled quadratic SEPIC converter 𝐿1 4.7 𝑚𝐻 𝐶1, 𝐶2 𝑎𝑛𝑑𝐶3 22 𝜇𝐹 𝐶0 2200 𝜇𝐹 Switching Frequency 10 𝑘𝐻𝑧 Case 1: Voltage sag condition Figure 12. Waveform of the AC source Figure 13. Waveform of the developed converter S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 178 Figure 14 shows the output voltage waveform of the Coupled Quadratic SEPIC Converter controlled by a Cascaded ANFIS MPPT Controller. The output voltage stabilizes at 320 V, suggesting a fast response and proper regulation of voltage at the output by the controller. The waveform of UPQC for the voltage sag condition is represented in Figure 15. The 3𝜙 reference voltage for the series converter is randomly changed, and it increased to 80 V with small fluctuations. Then, the 3𝜙 reference current is initially altered and is enhanced to a value of 40 A. Consequently, the power factor improved during the time frame that considering, stabilizing at 1, indicating the voltage and current are in sync. The waveform of the load is displayed in Figure 16. The constant voltage of 450 V and stable load current of 35 A is sustained in the entire system stability throughout the testing. Thus, the developed PV-UPQC is effective in increasing the PQ on the source and load sides. Figure 14. Waveform of developed converter output voltage using cascaded ANFIS MPPT controller Figure 15. Waveform of UPQC Figure 16. Waveform of load Case 2: Voltage swell condition Figure 17 illustrates the waveform of the AC source under a voltage swell condition. The source voltage is 400 V, and it increased to 450 V. Finally, it sustained at 400V, and it is influenced by the PQ issue with a voltage swell of 50 V. The source current continues to vary due to PQ issues, and because both the source voltage and current are continuously changing, the source current maintains these continual variations. Figure 18 indicates the waveform of the developed converter under a voltage swell condition. The input voltage is sustained at a stable value of 72 A in the entire system. Similarly, the input current is sustained at a value of 2400 A throughout the system. Then, the converter’s output voltage is randomly varied and settled at a value of 880 V with little variation. Finally, the output current is varied arbitrarily in the whole system. The waveform of UPQC for the voltage swell condition is seen in Figure 19. Initially, the reference voltage of the series converter is 10 V, and it increased to 80 V (that is, the voltage swell is 70 V). Also, the reference current of the shunt converter is 10 A, and it increased to 40 A. Then, the power factor exhibits changes at the beginning due to system changes, but as the time frame transitions, and the power factor stabilizes at the value of 1. The waveform of load voltage and current is depicted in Figure 20. The load voltage is sustained at a value of 450 V, and the load current is settled at 35A for stable operation of the system. The stabilization of the load current and voltage makes a substantial difference in the reliability and operation of the power system. Figure 17. Waveform of the AC source Figure 18. Waveform of the developed converter S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 179 Figure 19. Waveform of UPQC Figure 20. Waveform of load THD waveforms presented on the three R, Y, and B phases are illustrated in Figure 21. The B phase had the lowest THD of 3.97%, followed by the R and Y phases at 4.82% and 4.86% respectively, which shows that the presented harmonics were adequately reduced, resulting in an improvement in PQ. The analysis of voltage gain, improved high gain [23], Non-isolated buck-boost [24], and developed converter is depicted in Figure 22. The developed converter attains the highest voltage gain compared to other approaches, ensuring the overall performance of the system is enhanced. Figure 23 displays the comparison of voltage stress for the developed, switched LC-based high-gain [25] and improved high-gain [23] converter. The developed converter has the lowest voltage stress compared to other approaches, indicating that the efficacy of the system is enhanced. The technical benefits of this converter, such as the increased voltage gain and lower voltage stress, support its use in applications needing high DC-DC conversion, and also support overall output quality improvements, reduced voltage ripple, and increased energy efficiency, which all justify using this design in renewable energy systems, including PV applications. Figure 22. Analysis of voltage gain Figure 21. Waveform of THD S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 180 Figure 23. Analysis of voltage stress The analysis of Grid current THD (%) for R, Y, and B phases for NN [26] and the developed control approach is illustrated in Figure 24. The developed approach demonstrates a reduction in THD compared to NN control across R, Y, and B phases, indicating better harmonic suppression and enhanced PQ. Figure 25 compares the tracking efficiency of listed MPPT techniques. The proposed method has an efficiency of 98.90%, higher than both ANFIS (97.71%) [27] and Fuzzy (97%) [19], thus proving that it extracts more power under the same conditions than both these techniques. The proposed method demonstrates that it achieved improved tracking efficiency over ANFIS-based and Fuzzy-based MPPT methods, and superior performance. Figure 24. Analysis of grid current THD (%) for R, Y and B phases for unbalanced load conditions Figure 25. Comparison analysis of tracking efficiency 4. Conclusion This research presents a novel UPQC system with an ANN controller to diminish PQ problems and offset the load demand in PV systems. As a result, the PV-UPQC provides a superior solution for electrical distribution systems' PQ issues. By eliminating current harmonics, reducing voltage fluctuation, lowering the THD level in accordance with IEEE standards, and improving PQ, the ANN control system offers superior control. Since the PV is an intermediate power source, connecting it directly to the UPQC results in voltage instability, which is resolved by utilizing a coupled quadratic SEPIC converter with better efficacy. Consequently, the cascaded ANFIS MPPT is also exploited for tracking the peak power from the PV system with better tracking efficiency. The series converter compensates for voltage sags, swells, and harmonics, ensuring a stable voltage supply. Then, the shunt converter mitigates current harmonics, corrects power factor, and balances load currents. The results of the MATLAB simulation demonstrate that the developed PV-UPQC system effectively raises the PQ of the source voltage and load current with the lowest THD. Ethical issue The authors are aware of and comply with best practices in publication ethics, specifically with regard to authorship (avoidance of guest authorship), dual submission, manipulation of figures, competing interests, and compliance with policies on research ethics. The author adheres to publication requirements that the submitted work is original and has not been published elsewhere. Data availability statement Datasets analyzed during the current study are available and can be provided upon a reasonable request from the corresponding author. Conflict of interest The authors declare no potential conflict of interest. References [1] S. J. Alam and S. R. Arya, “Control of UPQC based on steady state linear Kalman filter for compensation of power quality problems,” in Chinese Journal of Electrical Engineering, vol. 6, no. 2, pp. 52-65, 2020. https://doi.org/10.23919/CJEE.2020.000011 [2] C. Jiang and S. Zhang, “Power Quality Compensation Strategy of MMC-UPQC Based on Passive Sliding Mode Control,” In IEEE Access, vol. 11, pp. 3662-3679, 2023. https://doi.org/10.1109/ACCESS.2022.3229893 [3] K. Sarita, S. Kumar, A. S. Vardhan, R. M. Elavarasan, R. K. Saket, G. M. Shafiullah and E. Hossain, “Power Enhancement With Grid Stabilization of Renewable Energy-Based Generation System Using UPQC-FLC- EVA Technique,” in IEEE Access, vol. 8, pp. 207443- 207464, 2020. https://doi.org/10.1109/ACCESS.2020.3038313 [4] S. R. Arya, S. J. Alam and P. Ray, “Optimized PI Gain in UPQC Control Based on Improved Zero Attracting Normalized LMS,” In CPSS Transactions on Power Electronics and Applications, vol. 9, no. 2, pp. 242-251, 2024. https://doi.org/10.24295/CPSSTPEA.2024.00007 [5] S. K. Yadav, A. Patel and H. D. Mathur, “PSO-Based Online PI Tuning of UPQC-DG in Real-Time,” in IEEE Open Journal of Power Electronics, vol. 5, pp. 1419- S. Kandukuri et al. /Future Technology August 2025| Volume 04 | Issue 03 | Pages 171-181 181 1431, 2024. https://ieeexplore.ieee.org/document/10643265 [6] J. Yu, Y. Xu, Y. Li and Q. Liu, “An Inductive Hybrid UPQC for Power Quality Management in Premium-Power- Supply-Required Applications,” In IEEE Access, vol. 8, pp. 113342-113354, 2020. https://doi.org/10.1109/ACCESS.2020.2999355 [7] A. Fouad, H. Kotb, K. M. AboRas, H. B. ElRefaie, M. Alqarni, A. M. Baqasah and A. H. Yakout, “Enhancing Grid-Connected DFIG’s LVRT Capability Using Dandelion Optimizer Based the Hybrid Fractional- Order PI and PI Controlled STATCOM,” in IEEE Access, vol. 12, pp. 120181-120197, 2024. https://doi.org/10.1109/ACCESS.2024.3427008 [8] N. Abas, S. Dilshad, A. Khalid, M. S. Saleem and N. Khan, “Power Quality Improvement Using Dynamic Voltage Restorer,” In IEEE Access, vol. 8, pp. 164325-164339, 2020. https://doi.org/10.1109/ACCESS.2020.3022477 [9] C. Gong, W. K. Sou and C. S. Lam, “H∞ optimal control design of static var compensator coupling hybrid active power filter based on harmonic state-space modeling,” CPSS Transactions on Power Electronics and Applications, vol. 6, no. 3, pp. 227-234, 2021. https://doi.org/10.24295/CPSSTPEA.2021.00021 [10] K. H. Tan, M. Y. Li and X. Y. Weng, “Droop Controlled Microgrid With DSTATCOM for Reactive Power Compensation and Power Quality Improvement,” In IEEE Access, vol. 10, pp. 121602-121614, 2022. https://doi.org/10.1109/ACCESS.2022.3223352 [11] R. Rudraram, S. Chinnathambi and M. Mani, “A Novel Intelligent Neural Network Techniques of UPQC with Integrated Solar PV System for Power Quality Enhancement,” International Journal of Electronics and Telecommunications, vol. 69, 2023. http://dx.doi.org/10.24425/ijet.2023.146514 [12] S. S. Dheeban and N. B. Muthu Selvan, “ANFIS-based power quality improvement by photovoltaic integrated UPQC at distribution system,” IETE Journal of Research, vol. 69, no. 5, pp. 2353-2371. 2023. https://doi.org/10.1080/03772063.2021.1888325 [13] K. S. Kavin, P. Subha Karuvelam, M. Devesh Raj and M. Sivasubramanian, “A Novel KSK Converter with Machine Learning MPPT for PV Applications,” Electric Power Components and Systems, pp. 1-19, 2024. https://doi.org/10.1080/15325008.2024.2346806 [14] K. Salim, M. Asif, F. Ali, A. Armghan, N. Ullah, A. S. Mohammad and A. A. Al Ahmadi, “Low-stress and optimum design of boost converter for renewable energy systems,” Micromachines, vol. 13, no. 7, pp. 1085, 2022. https://doi.org/10.3390/mi13071085 [15] Z. Haider, A. Ulasyar, A. Khattak, H. S. Zad, A. Mohammad, A. A. Alahmadi and N. Ullah, “Development and analysis of a novel high-gain CUK converter using voltage-multiplier units,” Electronics, vol. 11, no. 17, pp. 2766, 2022. https://doi.org/10.3390/electronics11172766 [16] A. Prakash and C. Naveen, “Combined strategy for tuning sensor-less brushless DC motor using SEPIC converter to reduce torque ripple,” ISA transactions, vol. 133, pp. 328-344, 2023. https://doi.org/10.1016/j.isatra.2022.06.045 [17] K. S. Kavin, P. S. Karuvelam, M. Matcha and S. Vendoti, “Improved BRBFNN-based MPPT algorithm for coupled inductor KSK converter for sustainable PV system applications,” Electrical Engineering, pp. 1-23, 2025. https://doi.org/10.1007/s00202-025-02952-9 [18] N. F. Ibrahim, M. M. Mahmoud, A. H. Al Thaiban, A. B. Barnawi, Z. S. Elbarbary, A. I. Omar and H. Abdelfattah, “Operation of grid-connected PV system with ANN- based MPPT and an optimized LCL filter using GRG algorithm for enhanced power quality,” IEEE Access, 2023. https://doi.org/10.1109/ACCESS.2023.3317980 [19] K. Ullah, M. Ishaq, F. Tchier, H. Ahmad and Z. Ahmad, “Fuzzy-based maximum power point tracking (MPPT) control system for photovoltaic power generation system,” Results in Engineering, vol. 20, pp. 101466, 2023. https://doi.org/10.1016/j.rineng.2023.101466 [20] S. A. Ibrahim, A. Nasr and M. A. Enany, “Maximum power point tracking using ANFIS for a reconfigurable PV-based battery charger under non-uniform operating conditions,” IEEE Access, vol. 9, pp. 114457- 114467, 2021. https://doi.org/10.1109/ACCESS.2021.3103039 [21] Rudraram, Ramesh, Sasi Chinnathambi, and Manikandan Mani. "PV Integrated UPQC with Intelligent Control Techniques for Power Quality Enhancement." International Journal of Electrical and Electronics Research 11, no. 1 (2023): 202-212. [22] Asoh, Derek Ajesam, Brice Damien Noumsi, and Edwin Nyuysever Mbinkar. "Maximum power point tracking using the incremental conductance algorithm for PV systems operating in rapidly changing environmental conditions." Smart Grid and Renewable Energy 13, no. 5 (2022): 89-108. [23] S. Ahmad, M. Nasir, J. Dąbrowski and J. M. Guerrero, “Improved topology of high voltage gain DC–DC converter with boost stages,” Int. J. Electron. Lett, vol. 9, no. 3, pp. 342–354, 2021. https://doi.org/10.1080/21681724.2020.1744040 [24] A. Sarikhani, B. Allahverdinejad and M. Hamzeh, “A nonisolated buck– boost DC–DC converter with continuous input current for photovoltaic applications,” IEEE J. Emerg. Sel. Topics Power Electron, vol. 9, no. 1, pp. 804–811, 2021. https://doi.org/10.1109/JESTPE.2020.2985844 [25] A. Kumar, Y. Wang, X. Pan, M. Raghuram, S. K. Singh, X. Xiong and A. K. Tripathi, “Switched-LC based high gain converter with lower component count,” IEEE Trans. Ind. Appl., vol. 56, no. 3, pp. 2816–2827, 2020. https://doi.org/10.1109/TIA.2020.2980215 [26] O. E Okwako, Z. H. Lin, M. Xin, K. Premkumar and A. J. Rodgers, “Neural network controlled solar PV battery powered unified power quality conditioner for grid connected operation,” Energies, vol. 15, no. 18, pp. 6825, 2022. https://doi.org/10.3390/en15186825 [27] E. Touti, M. Aoudia, C. H. Hussaian Basha, and I. M. Alrougy, "A Novel Design and Analysis Adaptive Hybrid ANFIS MPPT Controller for PEMFC‐Fed EV Systems," International Transactions on Electrical Energy Systems, vol. 2024, no. 1, pp. 5541124, 2024. This article is an open-access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/). https://doi.org/10.3390/en15186825 https://creativecommons.org/licenses/by/4.0/