Sugestion of a neuro-genetic hybrid system as an alternative for the electric power consumption curves evaluation

Abstract

This paper presents a model based on hybridism between Artificial Neural Network (ANN) and Genetic Algorithm (GA), that can identify on load curve, acquired any point of distribution system, the portions of consumption relative to each one of main consumer sectors in that point. The initial result indicates a best performance of conventional ANN, however when the complexity training patterns increase, the hybrid system starts to present best results, showing its power. Made the necessary simulations and checked the validity of association between ANN with GA, the developed system can be consider a powerful tool to the distribution electrical network management.