Modeling and simulation of substation fault diagnosis based on Petri nets

In order to improve the efficiency of the substation fault diagnosis with multiple protections, the proposed method describes protection grades by varied weight of the transition vector in the Petri net models. According to relays and circuit breakers, special markings are produced in the diagnosis results that enhance the identifying capability of error information. The process of fault diagnosis is accomplished by algebra matrix calculation. A simulation on actual substation fault diagnosis process indicates that the improved method has fast calculation speed and high veracity.

I. Introduction

There have been many techniques presented for finding the exact location of fault in substation. In most of these approaches, substation fault have been diagnosed by expert knowledge or artificial neural networks (ANN), and other artificial intelligence technology. Since these models do not represent the change of the substation configuration, or could not identify the error of the input signal, or needing much of training, significant errors are generally encountered in their fault diagnosis process [1], [2].