Weekdays load forecasting in mexico using artif1cial neural networks

A wide area network (WAN) is the set of components interconnected in a power system. A WAN is operated in the control center, where operators are taking manual action on real time and engineers are attending offline tasks for obtaining necessary guidelines. The Automatic Generation Control (AGC) makes possible the balance between the generation and the load [1], based on a load demand. By this reason an accurate load forecast is a key element in many decision-making activities at the control center. This paper describes a Neural Network Load forecasting application to obtain an accurate prediction of WAN real power, with lead times ranging from one to seven days ahead. � 2006 IEEE