Diagnóstico inteligente de falhas em pára-raios

Abstract

This work proposes and implements a methodology to develop a fault diagnosis system applied to lightning arresters. The methodology applies methods based on historical data of measurements of lightning arresters, using techniques of computational intelligence. The systemmakes use of the termovision, leakeage current and radiointerference techniques to generate the diagnosis.The methodology was validated using a database supplied by the CEPEL, getting about 10% of error. This work also contributes to the management of lightning arresters information. The developed computational system implements an environment of work completely integrated, allowing that all the measurements information of lightning arrester to become available to itsusers by a transparent form. The new paradigm for fault diagnosis, management of lightning arresters implemented can be seen by the the improvement of quality and availability of the pertinent information to the lightning arresters.