Peak Load Forecasting of Electric Utilities for West Province of IRAN by Using Neural Network without Weather Information

Accurate peak load forecasting plays a key role in economical use from energy. Artificial Neural Networks (ANN) has recently applied on short term load forecasting in electrical utilities. The ANN is used to Predicting the relationship between past, current and future peak loads. Conventional systems require various variables from the past factors that can affect on peak load such as: load and weather information. Too many input variables cause some problems in prediction for the future operation of the system. However, we use just past load values for peak load forecasting. In this paper two operative algorithms used, Multi Layer Perceptron (MLP) and Radial Basis Function (RBF), for predicting peak load. Then, comparison has been made between these methods to show error in peak load forecasting. The result shows that in this case Multi layer perceptron has more accuracy than Radial basis function i.e., better mean relative error (MRE).