Abstract

Power quality problems such as voltage sag, voltage swell, harmonics, and load imbalance have increased in modern power distribution systems because of the large-scale integration of renewable energy sources. Traditional PI and PID controllers, even when optimized, often struggle to adapt properly under nonlinear and rapidly changing operating conditions. Their performance becomes limited when the system experiences sudden disturbances or variations. To overcome these issues, this paper proposes an intelligent hybrid control method for a three-phase Unified Power Quality Conditioner (UPQC) based on D–Q theory. The proposed system combines an Adaptive Neuro-Fuzzy Inference System (ANFIS) with Model Predictive Control (MPC). In addition, Particle Swarm Optimization (PSO) is used to optimally tune the controller parameters and predictive weighting factors. In this approach, ANFIS works as a nonlinear disturbance estimator. It helps the system adapt more effectively to changes and improves compensation performance. MPC predicts the future behaviour of the system and minimizes a multi-objective cost function to ensure a fast-dynamic response and accurate voltage and current control. PSO is applied to optimize the membership functions of ANFIS and the weighting coefficients of MPC, which improves convergence and enhances harmonic reduction performance. The proposed system is modelled and simulated in MATLAB/Simulink under nonlinear, unbalanced load conditions and renewable energy disturbances. The simulation results show that the PSO-optimized ANFIS-MPC based UPQC provides faster transient response, better DC-link voltage stability, accurate voltage restoration during sag and swell conditions, and significant reduction in Total Harmonic Distortion (THD). The THD values are maintained well within the IEEE-519 standard limits. A comparative study shows that the proposed method performs better than conventional PID and metaheuristic-tuned PID controllers in terms of robustness, adaptability, and harmonic suppression. Therefore, the proposed intelligent predictive control framework offers a reliable and efficient solution for im