Abstract

Long-term Electrical Load Forecasting (ELF) is essential for infrastructure planning and the proper functioning of substations. ELF reduces the overall planning uncertainty added by the intermittent production of renewable energy sources. It helps to minimize the hydrothermal electricity production costs in a power grid. Although there is some research in the field and even several research applications, there is a continual need to improve forecasts. The use of Machine Learning Algorithms (MLAs) for prediction purposes is increasing in recent times. The paper presents the results of electrical load forecasting using various MLAs and their comparative analysis. The electrical load data for training the models are obtained from the 110/33/11kV substation of Haliyal, District: Uttara Kannada, Karnataka, India. Other features such as temperature and salary are included for enhancing the prediction. The MLAs are implemented in Python using Scikit-Learn. The performance of various MLAs is measured in terms of Root Mean Square Error (RMSE). The model validation is done using cross-validation. The comparative analysis shows that the Decision Tree Algorithm gives better results for the prediction of electrical load as compared to others. It is further concluded that MLAs prove to be an effective tool for substation planning, expansion, and proper functioning.