Authors: Sharma Pankaj Kanhaiya, Professor Devendra Sharma, Professor Saurabh Gupta
Abstract: The integration of photovoltaic (PV) systems with battery and supercapacitor storage in DC microgrids demands efficient energy management to enhance system stability, reliability, and operational efficiency. This research presents an optimized neural network-based energy management approach tailored for a standalone DC microgrid incorporating PV panels, lithium-ion batteries, and supercapacitors. The neural network model is specifically designed to handle the nonlinear characteristics of the microgrid, optimize power flow, and maintain the state of charge (SoC) of energy storage devices within safe limits. By utilizing advanced training algorithms inspired by optimization techniques such as artificial rabbit optimization, the proposed system achieves improved prediction accuracy and load balancing. The approach also integrates a fuzzy logic control mechanism to facilitate real-time adaptive responses to dynamic load changes and renewable generation variability. Simulation results demonstrate enhanced voltage stability, reduced power fluctuations, and efficient energy distribution compared to conventional methods. This optimized neural network strategy effectively mitigates the challenges inherent in hybrid energy storage management, promoting longer battery life, quicker response times from supercapacitors, and overall system resilience. The study contributes significant insights toward the development of intelligent energy management systems for sustainable and autonomous DC microgrid applications ((PDF) Artificial Rabbits Optimized Neural Network-Based Energy …, 2024).
