Deep-Fuzzy Logic Control for Optimal Energy Management: A Predictive and Adaptive Framework for Grid-Connected Microgrids

This paper introduces a novel energy management framework, Deep-Fuzzy Logic Control (Deep-FLC), which combines predictive modelling using Long Short-Term Memory (LSTM) networks with adaptive fuzzy logic to optimise energy allocation, minimise grid dependency, and preserve battery health in grid-connected microgrid (MG) systems. Integrating LSTM-based predictions provides foresight into system parameters such as state of charge, load demand, and battery health, while fuzzy logic ensures real-time adaptive control. Results demonstrate that Deep-FLC achieves a 25.7% reduction in operational costs compared to the conventional system and a 17.5% saving cost over the Fuzzy Logic Control (FLC) system. Additionally, Deep-FLC delivers the highest battery efficiency of 61% and constraints depth of discharge to below 2% per time step, resulting in a reduction of the state of health degradation to less than 0.2% over 300 h. By combining predictive analytics with adaptive control, this study addresses the limitations of standalone approaches and establishes Deep-FLC as a robust, efficient, and sustainable energy management solution. Key novel contributions include the integration of advanced prediction mechanisms with fuzzy control and its application to battery-integrated grid-connected MG systems.