Power system fault diagnosis modeling techniques based on encoded Petri nets

In this paper, power system fault diagnosis based on Petri nets and coding theory is further studied. Previous research work is briefly reviewed. Characteristics of Petri nets model in power system fault diagnosis and identification are demonstrated in detail, also a fast model revision algorithm of power components is proposed, which makes the scheme more applicable to large-scale power network. The method is tested in the IEEE 118-bus power system and simulation results show that the suggested approach is accurate by using of error correction theory, model revision is easy, fast when power network is expanded or topology is changed, which makes the encoded Petri nets method more applicable in real power system.

I. Introduction

ON-LINE FDI (fault diagnosis and identification) in power system is of great importance to the restoration of power system. A number of techniques have been developed, including artificial intelligence techniques, expert systems, rule-based systems and neural networks. Fault diagnosis of Large-scale power system, however, is still an unresolved problem. It mainly because of the large amount of information received after fault and the need for high accuracy, speed, especially the existence of hidden failures (which include misoperation of circuit breakers and false alarms from relays) and model's un-adaptability to topological changes.