55　Predictive Modelling of Electrical Load: Analyzing the Impact of Weather and Time on Energy consumption

Abstract

This paper proposes a simple multiple linear regression model to predict the electrical load of Shizuoka University’s Hamamatsu campus. This is done by splitting the load profile into two distinct patterns and using meteorological data such as temperature, wind speed, humidity and GHI as predictors. Findings indicate that the most significant contributor to the load increase is the temperature, primarily due to air conditioning usage. The model offers a reduced complexity approach, while remaining flexible for future adaptability.