Adaptive ANN-based UPQC control for a meshed hybrid AC-DC microgrid integrating renewable energy, battery storage, and electric vehicles

Abstract

The increasing integration of renewable energy sources (RESs), battery energy storage systems (BESSs), and electric vehicle (EV) charging infrastructure into modern power systems has introduced significant challenges, including power-quality degradation, poor voltage regulation, harmonic distortion, and complex dynamic power management. This paper presents an adaptive artificial neural network (ANN)-based optimized unified power quality conditioner (UPQC) for a meshed hybrid alternating-current/direct-current (AC/DC) microgrid. The system integrates a photovoltaic (PV) generation system, a wind energy conversion system (WECS) employing a permanent-magnet synchronous machine, a BESS, the utility grid, and EV loads operating in both grid-to-vehicle (G2V) and vehicle-to-grid (V2G) modes. The proposed ANN-based controller generates adaptive reference signals for the series and shunt converters of the UPQC, thereby enabling effective voltage regulation, harmonic mitigation, and dynamic power-flow control under varying operating conditions. A detailed hybrid AC/DC microgrid model is developed in MATLAB/Simulink, and its performance is evaluated under variations in renewable energy availability, load conditions, and EV operating modes. The simulation results demonstrate that the proposed controller maintains stable AC and DC bus voltages, improves dynamic performance, reduces total harmonic distortion (THD), and enhances power quality compared with conventional control schemes. Furthermore, the learning capability of the ANN enables the controller to identify and mitigate disturbances, thereby improving the performance and reliability of the hybrid microgrid.