Short term load forecasting using artificial neural networks

Short term load forecasting (STLF) for a lead time of one hour to 24 hours is essential for planning the start-up and shut-down schedules of generating units, reserve planning and load management. An accurate forecast eases the problem of generation and load management to a great extent. Methods ranging from statistical to artificial neural networks (ANN) have been applied to STLF. This paper presents an application of ANN to short term load forecasting. The proposed method works in two stages. In the first stage a load forecast with a lead time of 24 hours is done for unit commitment, generation planning, etc. In the second stage, the next hour forecast is refined using the latest load values and the error in their prediction.