ANN Based Short Term Load Forecasting for Karnataka State Demand

Abstract

Evolution of smart grid has impact on electric load forecasting. Short term load forecasting aims to predict the load for a day/week ahead. Accurate electrical load forecasting has become a challenging issue due to influence of non-linear load time series in modern power systems. This paper explores Artificial Neural Networks (ANN) approach to deal with this challenge. ANN is trained using back propagation technique. Karnataka State Demand-2019 obtained from Karnataka Power Transmission Corporation Limited (KPTCL) Website is considered for analysis and short-term load forecasting. Weather and non-weather variables are taken into account so as to minimize the forecasting error. Accurate load forecasting would have major positive impact in energy trading. Among several proposed approaches for short term load forecasting, ANN based models has its own significance due to its quicker prediction and accurateness. The efficacy of the proposed model with 8 input variables is investigated by calculating mean absolute percentage error (MAPE), mean absolute error (MAE) and daily peak mean absolute percentage error.