Intelligent Modelling Techniques of Power Load Forecasting for the Western Area of Saudi Arabia

Load forecasting has become in recent years one of themajor areas of research in electrical engineering. Most traditionalforecasting models and artificial intelligence neural networktechniques have been tried out in this task. Artificial neural networks(ANN) have lately received much attention, and a great number ofpapers have reported successful experiments and practical tests. Thispaper presents the development of an ANN-based short-term loadforecasting model with improved accuracy for the Regional PowerControl Centre of Saudi Electricity Company, Western OperationArea (SEC-WOA). The proposed ANN is trained with weather-relateddata, special events indexes and historical electric load-related datausing the data from the calendar years 2003, to 2007 for training.Different neural networks topologies have been trained and tested forachieve the optimal topology and ranking the input variables in termsof their importance. Based on the optimal NN topology, the networkhas been trained to predict the ahead load at different time intervals.