Redes neurais artificiais aplicadas para determinar a seção faltosa em uma linha de transmissão com compensação série na presença de geração eólica

Transmission lines are connected to several electrical power system devices and must be adequately protected to avoid damaging these elements. However, the protection task is more complex when compensators or alternative sources are connected to transmission lines once their protection devices are, generally, non-linear. In this context, this work proposes an algorithm for determining in which section of a series compensated transmission line with wind generation a single-phase fault occurred. The Discrete Stockwell Transform is a preprocessing tool for feeding an Artificial Neural Network (ANN) with resultant coefficients. For the moment, single-phase voltage and current signals from both transmission line terminals are considered in this research. Since the results show 97 % of the proper responses for a few training cases, the method is promising and offers many possibilities to be applied.