Comparing peak electricity load forecasting models for an industrial and a residential building

Electrical load forecasting models are becoming more and more essential for energy savings and effective energy management and the ability to forecast the load peak is a crucial feature for many applications. This paper presents a comparative analysis of various methods with low computational complexity, including Naïve Persistence, Statistical load forecasting, Seasonal Auto-Regressive Integrated Moving Average with eXogenous regressors, and Long Short-Term Memory enhanced by Empirical Mode Decomposition pre-processing. They are tested to forecast daily peak electricity load and peak hour in two distinct existing scenarios: residential and industrial. The study investigates traditional statistical models and artificial intelligence techniques to determine the most effective method to obtain accurate load prediction. Through performance evaluation and data analysis, insights are provided into their applicability and effectiveness in peak load forecasting.