Optimal Dispatch Control of Hybrid AC Microgrid Using Reinforcement Learning over rule-based approach for Enhanced Renewable Energy Utilization

Abstract

The current power systems need intelligent control methods which help them operate their microgrids safely and efficiently with minimal need for traditional diesel generators. This research paper describes the development and execution of an intelligent energy management system (EMS) which operates a hybrid AC microgrid that combines solar photovoltaic (PV) and wind energy conversion system (WECS) and battery energy storage system (BESS) and diesel generator. The primary objective of the proposed system is to reduce diesel generator operating hours by prioritizing renewable energy utilization and optimal battery scheduling while maintaining continuous load supply. A reinforcement learning (RL)–based supervisory control strategy is developed and integrated into the microgrid framework using MATLAB/Simulink. The RL controller establishes optimal operational procedures which include battery charging and discharging and renewable energy dispatch prioritization and DG activation based on system conditions which include load demand and renewable power availability and battery state of charge (SOC). The proposed controller is compared with a conventional rule-based energy management strategy to evaluate improvements in system performance. The simulation results show that the RL-based EMS system performs better than the rule-based system because it increases renewable energy usage while decreasing diesel generator usage during testing with different irradiance and wind speed and load demand conditions. The results also confirm enhanced battery utilization efficiency and stable power balance across the common AC bus while ensuring uninterrupted load supply throughout the operating period. The intelligent control system design establishes an effective system for control operations.