Short-Term Load Forecasting Using Hybrid ARIMA and Artificial Neural Network Model

Load forecasting is basic for building up a power supply strategy to enhance the reliability of the power line and gives optimal load scheduling to numerous developing nations where the demand can be expanded with high development rate. Short-Term Electric Load Forecast (STLF) is very important because it can be used to preserve optimum behaviour in daily operations of electrical system. For this purpose, Autoregressive Integrated Moving Average Model (ARIMA) is utilised which is a linear prediction procedure. Neural networks have capability to model complex and nonlinear relationship. The aim of this paper is to explain how neural network is able to change linear ARIMA model to create short-term load forecasts. The hybrid methodology, combining ARIMA and ANN model, will purposely take advantages of the unique power of ARIMA and ANN models in linear and nonlinear domains, respectively.