Optimal Scheduling of a Solar-Powered Microgrid Using ML-Based Solar and Load Forecasting

Le résumé fourni par la source

Microgrids, powered by distributed energy resources, are gaining traction as decentralized power systems. However, optimizing microgrid operation poses challenges due to intermittent renewable energy sources and dynamic load patterns. To tackle this, we propose an AI-driven day-ahead optimal scheduling approach for a grid-connected AC microgrid equipped with a solar panel and a battery energy storage system. Our approach leverages Genetic Algorithm, a popular optimization algorithm, to generate demand response strategies and optimal battery dispatch schedule. Additionally, we utilize LightGBM, a decision tree-based machine learning method, for solar and load forecasting prior to scheduling. Our objective is to minimize operational costs while ensuring the sustainability of the microgrid. Our simulation results showcase the effectiveness of our approach in reducing costs, with a 13.86% decrease in electricity costs observed in the University of Moratuwa microgrid under the tariff structure in Sri Lanka. Our proposed demand response optimizing strategies further contribute to cost reduction. Our approach showcases the power of AI in addressing the challenges of microgrid operation and optimization, with promising results in reducing costs and ensuring sustainability.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.