Comparative evaluation and architectural enhancement of a genetic algorithm-tuned fuzzy logic battery control in microgrid energy management

Abstract

Abstract Efficient battery energy management remains a key challenge in renewable-integrated microgrids due to nonlinear battery dynamics, intermittent generation, and variable load demand. This paper investigates a fuzzy logic-based battery control strategy for a hybrid microgrid comprising photovoltaic generation, wind energy, battery energy storage, and a bidirectional grid interface. A two-phase control framework is adopted. In Phase 1, a genetic algorithm (GA) is used to tune an established fuzzy logic controller under identical operating conditions previously reported for particle swarm optimization (PSO) and backtracking search algorithm (BSA) approaches, enabling a fair comparative evaluation. In Phase 2, the controller architecture is enhanced by embedding battery charge-discharge and grid import-export decisions directly within the fuzzy inference system using real-time power imbalance and state-of-charge as inputs. Simulation results show that, while GA converges more slowly than BSA, all tuning methods achieve comparable battery current and state-of-charge regulation in Phase 1. In contrast, the enhanced control architecture in Phase 2 yields more coherent energy management behavior, improved utilization of renewable energy, reduced unnecessary battery cycling, and stable electrical performance. The results demonstrate that controller architecture has a greater impact on practical microgrid energy management performance than tuning strategy alone.