Proposal of a Probabilistic Electricity Demand Forecasting Method using Bayesian Inference

This paper proposes a new forecasting method for electricity demand in energy management systems of regional microgrids, based on Bayesian inference. The proposed method addresses multiple challenges, including improving prediction accuracy, ensuring interpretability, and quantifying uncertainty, while maintaining simplicity and real-time applicability. By considering meteorological factors, seasonal variations, and daily patterns, and applying Bayesian inference to historical data collected at the same time of day on previous days, the method provides useful information for system operators. Furthermore, comparative analyses demonstrate that the proposed method outperforms existing forecasting approaches in terms of prediction accuracy, interpretability, and uncertainty quantification.