Enhancing Voltage and Frequency Control in Islanded VSC-Based Microgrids: A PSO-Driven Multilayer Perceptron Approach

Abstract

This paper proposes a novel design approach using the particle swarm optimization-trained multilayer perceptron neural network (PSO-MLPNN) for inner current and voltage controllers in voltage source converter (VSC)-based islanded AC microgrids. Unlike conventional controllers that depend on line parameters, our intelligent controller structure operates independently of these parameters. The PSO-MLPNN controller effectively replaces the traditional Proportional-Integral (PI) controller, achieving accurate power-sharing and excellent dynamic performance across various load scenarios. Simulation results validate the effectiveness of our proposed controller.