UMA CONTRIBUIÇÃO PARA O CÁLCULO DO TEMPO CRÍTICO DA ESTABILIDADE TRANSITÓRIA DE SISTEMAS ELÉTRICOS DE POTÊNCIA USANDO REDES NEURAIS ARTIFICIAIS

Abstract

The objective of this work is the use of artificial neural networks (ANN) to estimate the critical clearing time of a power system submitted to severe disturbances concerning to electrical power system transient stability. The critical time is difficult to determinate, because it is direct or indirect affected by a series of factors, such as: type of fault, fault localization, configuration (state) of the electrical power system at the moment of the occurrence of the fault, and so on. The neural net must observe the disturbance caused in the angle wave form of one of the generators caused by a fault in a transmission line, thus determining, based in this observation, the critical time for that type of fault, in that same line, and also foreseeing the influence that changes in the generated powers can cause in the transient stability limit. Two proposals based on this main idea, applied in two different test systems, the WSCC3 and the New England. The first proposal was applied in the WSCC3 system, and represents the embryo of the second method. The second proposal was an advanced attempt of the first, about the applicability of the method in larger electrical systems.