Using Aggregated Electrical Loads for the Multinodal Load Forecasting

Abstract

Forecasting electrical loads is essential from a practical and economic point of view. With this forecast, it is possible to plan the supply of energy safely and continuously, and without interruption. In the literature, most of the works that perform electric load forecasting consider the global demand, that is, the sum of the total energy consumption. This work proposes to carry out the load forecasting along with the buses of a distribution system (multinodal forecasting) based on the use of the load aggregation concept. The proposed method uses a Fuzzy-ARTMAP neural network to forecast electrical loads in substations (multinodal forecasting) 24 h ahead, with the main objective of studying and identifying possible aggregations of multinodal loads, aiming at improving the multinodal load forecasting. The database used was from an electricity distribution subsystem, consisting of nine substations.