Short Term Load Forecasting using Machine Learning Techniques

Abstract

With recent technological and scientific advancements in the power systems, there has been a tandem need for load forecasting. This paper mainly discusses short-term load forecasting, which refers to the prediction of the system load demand over an interval ranging between minutes ahead to one week ahead. With advent of Machine Learning, the process of demand prediction has become easier and cost effective. The challenge of predicting the future demand can be characterized as a regression problem, hence the method of Support Vector Regression is used, as it has proved to be a robust method in the recent research. Different Neural Networks are also being used in several domains; hence Deep Neural Network has also been used to test the accuracy, The paper discusses the results obtained by two different methods. The comparison between the outcomes of the different algorithms has been discussed, in order to get a thorough understanding. The methods are explained vastly. The paper also discusses the factors affecting load forecasting directly.