Desenvolvimento de ferramenta computacional baseado em técnicas de inteligência artificial para avaliação de estabilidade de tensão.

Abstract

The development of a computational tool capable of evaluating the security of a power system in relation to the voltage instability, using techniques of Artificial Intelligence (AI) is presented. An AI–based method for the assessment of voltage safety of power systems, composed of a hybrid arrangement of a Resilient Propagation neural network (RPROP) in series with a Levenberg-Marquardt, is proposed. Databases were constructed using the Area Least of Northeast Subsystem, belonging to the CHESF’s core network, and the software ANAREDE®, to the training, validation and testing techniques of AI analyzed. The performance of this system was verified through computer resources Neural Network Toolbox and Fuzzy Logic Toolbox of MATLAB®. The software was developed using the programming language C + +.