Load Forecasting Technology for Overloaded Microgrids

The rapid growth of distributed energy resources (DERs) and electric vehicles (EVs) has introduced significant challenges for load forecasting in overloaded microgrids, particularly under volatile and stochastic operating conditions. This paper proposes an integrated framework combining time-series analysis, deep learning models, and probabilistic approaches to enhance forecasting accuracy and adaptability. The proposed method captures multi-scale temporal patterns, reduces computational overhead, and effectively manages high penetration levels of renewable energy and EV charging stations. Extensive evaluations under diverse scenarios demonstrate the framework’s robustness and scalability, providing actionable insights for energy management and grid stability. This study contributes to advancing load forecasting technologies aligned with carbon neutrality objectives, offering a scalable solution for modern power systems.