A Research Of Fuzzy Petri Net Fault Diagnosis Method Based On Alarm Information

With the rapid development of modern information technology and power system, the structures of grid have become more and more complicated. For the short-circuit fault in local power grid, if effective measures could not be taken timely, the fault area may expand and cause heavy economic losses. In order to analyze the fault cause accurately and provide the aid decision making for dispatchers timely, it is very necessary to carry out the fast and efficient fault diagnosis algorithm research. The fuzzy Petri net has a certain application prospect in the field of power system fault diagnosis, because of simple reasoning, clear structure, fast diagnosis and strong descriptive ability for alarm information. Therefore, the basic theory of fuzzy Petri net is mainly adopted, and the logical relevance and time constraint characteristics of alarm information are also fully utilized in this paper, then a series of power system fault diagnosis methods are proposed. The main achievements are as follows:Firstly, the logical relevance of alarm information is deeply analyzed, and a method of power system fault diagnosis based on improved fuzzy Petri net considering the logical relevance of alarm information is proposed. The corresponding relation between component failure, relay protective action and circuit breaker tripping is fully considered, then the fuzzy reasoning process and matrix manipulation scale are simplified, and the improved fuzzy Petri net of grid component is built in this paper. On this basis the detailed process of power system fault diagnosis method is put forward. The simulation results on partial power system indicate the validity of the proposed method.Secondly, to utilize the time constraint characteristics of alarm information and improve the accuracy and rapidity of power system fault diagnosis, a method of power system fault diagnosis based on time constraint fuzzy Petri net is proposed in this paper. By using the virtual directed arcs and inference rules of Petri net, the time constraint fuzzy Petri net model of grid component is built. The time intervals that all alarm information should meet can be acquired by the forward and backward temporal reasoning analysis, and the time sequence inconsistent alarm information can be effectively identified based on the state truth value matrix. Then, the identification rules of relay protection and circuit breaker are formulated. The simulation results and a real power system fault case indicate that the proposed method can effectively diagnose power system faults, and also has the broad application prospect.Finally, for the fault phase identification is inaccurate in existing fault diagnosis models based on fuzzy Petri net, a method of power system fault diagnosis based on colored split-phase fuzzy Petri net is proposed in this paper. A temporal fault tree network is built, and a time constraint checking method is also put forward to identify the time sequence inconsistent alarm information. On the basis of colored Petri net, the color attributes are used to distinguish fault types and fault phase, and the colored split-phase fuzzy Petri net of grid component is built. The simulation results of 220 kV partial power system and a part of Zhejiang power network indicate that the proposed method can accurately describe power system actual faults, and also has much higher fault tolerance.