Comparative Study of Machine Learning Models for Load Prediction and Energy Management

Accurate energy load prediction is crucial for optimizing energy management in smart grid systems. This study evaluates the performance of four machine learning models, which are random forest, gradient boosting, support vector regression (SVR), and linear regression, for load prediction using a dataset from Universiti Teknikal Malaysia Melaka (UTeM). The dataset consists of 21 months of hourly energy consumption data, including photovoltaic (PV) generation, battery storage, and grid meter readings. Among the models tested, gradient boosting model achieved the highest accuracy with an R2 of 0.72, demonstrating its effectiveness in forecasting energy demand. Random forest model exhibited strong training performance but suffered from overfitting, while SVR and linear regression models showed lower predictive accuracy. The predicted load values were integrated into an if-then rule-based control strategy for managing energy distribution among PV, battery, and grid sources. The findings highlight the potential of machine learning in enhancing energy efficiency by improving demand forecasting and optimizing resource allocation.