Distribution system and long term load forecasting using artificial neural network

This thesis will be a reference for organizations that are interested in electricitymarketing. In this thesis, using artifical neural network architecture in long termload forecasting, the better results are obtained.Long term load forecasting is achieved using regression analysis and artificialneural network methods in the scope of privatization of energy distributionsystem. In this thesis, a multilayer feed-forward neural network based long termload forecasting method is chosen.It is known that electricity load depends on many factors such as population, GrossNational Product (GNP), Development Velocity (DV), Industrial Production Index(IPI) and petroleum price. Artificial neural network is trained by using past loggeddata and results are compared with both test data of artificial neural network andregression analysis results. As a result of, an artificial neural network model whichsucceeds better than regression analysis model is used for 12th region.Science Code: 905Key Words : Energy Distribution System, Long Term Load Forecasting,Artificial Neural Networks, Regression.Page Number : 98Adviser : Prof. Dr. M. Cengiz TAPLAMACIOĞLU