Detect and classify faults using neural nets

The analysis of transmission line faults is essential to the proper performance of a power system. It is required if protective relays are to take appropriate action and in monitoring the performance of relays, circuit breakers and other protective and control elements. The detection and classification of transmission line faults is a fundamental component of such fault analysis. Here, the authors describe how a neural network, trained to recognize patterns of transmission line faults, has been incorporated in a PC-based system that analyzes data files from substation digital fault recorders.