Electricity Load/demand Forecasting in Sri Lanka using Deep Learning Techniques

Abstract

Most of the time electricity cannot be stored, it should be generated as soon as it is demanded. Therefore, electricity demand forecasting is a vital process in the planning of electricity industry and the operation of electric power systems. Two major scenarios should be considered when forecasting the electricity demand. They are short term and long term forecasting scenarios. The short term scenario is more critical since many features have to be considered. In this research study, deep learning techniques such as Recurrent Neural Network(RNN), Long Short Term Memory (LSTM) and Convolutional Neural Network (CNN) were considered for electricity demand forecasting of Sri Lankan demand profile. Further, the results of deep learning approaches were compared with traditional techniques such as Linear Regression, Lasso Regression, Light Gradient Boosting Model (LGBM) and Random Forest Regressor. It was found from our studies that LSTM based approach performs better than other approaches.