Advanced Genetic Algorithm for Optimal Microgrid Scheduling Considering Solar and Load Forecasting, Battery Degradation, and Demand Response Dynamics

Abstract

Microgrids driven by distributed energy resources are gaining prominence as decentralized power systems offering advantages in energy sustainability and resilience. However, optimizing microgrid operation faces challenges from the intermittent nature of renewable sources, dynamic energy demand, and varying grid electricity prices. This paper presents an AI-driven day-ahead optimal scheduling approach for a grid-connected AC microgrid with a solar panel and a battery energy storage system. Genetic Algorithm generates demand response strategies and optimizes battery dispatch, while LightGBM forecasts solar power generation and building load consumption. The approach aims to minimize operational costs and ensure microgrid sustainability, using a battery degradation cost function to extend its lifespan. Simulation results conducted in the University of Moratuwa microgrid show a significant 14.22% decrease in electricity costs under Sri Lanka’s current tariff structure, attributed to intelligent energy dispatch scheduling. Proactive demand response management has the potential to minimize costs further. This research contributes to microgrid optimization knowledge, promoting the adoption of intelligent and sustainable energy systems.

Keywords

• Microgrid
• optimizing
• genetic algorithm
• machine learning
• decision trees
• demand response strategies