Artificial Neural Network Modelling for Short Term Load Forecasting

Load forecasting is an essential part of the deregulated market to determine the future power demand accurately. A situation with generation lesser than demand results in irregular supply, whereas the opposite results in losses for the service provider. This paper highlights the importance of short term load forecasting using Artificial Neural Networks. Multi-layer Perception (MPL) networks employing supervised learning methods like back-propagation are used for the load prediction 24, 48 and 72 hours in advance. The effectiveness and performance are analysed with MAPE as 1.87 %, 1.98 % and 1.78 % for 24, 48 and 72 hours, respectively.