Adaptive Control of PV-Integrated Power Grids Using KNN-Smote-GCN And Mpc Techniques

Abstract

As the global energy crisis intensifies, the integration of renewable energy—particularly photovoltaic (PV) systems—has become vital for achieving a sustainable and resilient power infrastructure. This study focuses on dynamic modeling and efficient control of grid-connected PV systems to enhance power quality and system reliability. An adaptive PI controller is employed for voltage regulation, with a maximum power point tracking (MPPT) method ensuring optimal energy harvesting. A DC-DC boost converter and a three-phase PWM inverter are incorporated, with MATLAB used for simulation. The proposed approach integrates Model Predictive Control (MPC) with Graph Convolutional Networks (GCN) to manage grid instability and improve energy efficiency. A novel KNN-SMOTE-GCN algorithm is developed to mitigate voltage distortion, harmonic currents, and power fluctuations. The system replicates the behavior of traditional generators under disturbances, promoting renewable integration without compromising stability. Key performance metrics such as voltage deviation, reactive power fluctuation, power loss, and total harmonic distortion (THD) are analyzed.