Power System Fault Diagnosis Based On The Fast Topology And The Improved Petri Net Model

With the development of our country, more and more energy isconsumed, which increases the scale and the complexity of the powersystem. So further requests have been advanced to the existing powersystem fault diagnosis method based on the Petri net. On one hand thepower system fault diagnosis method is need to handle the large number offault information to improve the fault diagnostic precision, on the otherhand the power system fault diagnosis method is need to be real-time andrapid. Thus, improving the existing power system fault diagnosis methodbased on the Petri net is urgently needed.This paper proposes a new rapid topological method of blocking powernetwork for the power network firstly, which intend to find an approach toovercome the shortcomings of the high dependence of the traditional powersystem fault diagnosis method based on the Petri net model on the complexstructure of power system. This new rapid topological method for the powernetwork based on the traditional depth-first search algorithm introduces theconcept called bay, abstracts out the equipment models into nodes,establishes the node adjacency list, and analyzes the power network usingthe block topological analysis method according to some certain rules.When the network topology changes, through the local network topologyreconstruction involving in changes, we can achieve the purpose of updatingthe whole network topology. This makes the speed of the topology analysismuch faster. Then the fault zones can be found based on the rapidtopological method of blocking power network which reduce the number ofthe faulty elements needed be established the fault diagnosis model as well as the scale of each fault diagnosis models. Above is to speed the powersystem fault diagnosis. The new rapid topological method of blockingpower network for the power network is demonstrated by many numericalexamples.Then, improvement has been made on the exiting power system faultdiagnosis method based on the Fussy Petri net and the new power systemfault diagnosis method based on the improved fussy Petri net modelestablished of the multi-path model structure has been proposed in this paper.The parameters of the definition have been improved and some calculationfunctions of the Fussy Petri net model have been defined, which improvedthe ability of diagnosing the power system fault with the incompletenessfault information. On the other hand, the elements in the power systembeing established the fault diagnosis models include the line, thetransformer and the bus according to the characteristics of the power system.The modeling direction is from the element being diagnosed to the adjacentelements around, which improving the adaptability of the model. A detailedexample illustrates the validity of the new power system fault diagnosismethod.Finally, this paper proposes to use the power grid partition method andthe information fusion technology to improve the new power system faultdiagnosis model with the problem of the large dimension of matrix and thecombination explosion as a kind of distributed fault diagnosis method. Thepower grid partition method used in this paper is the butterfly partition andthe information fusion technology used in this paper is based on theDempster-Shafer evidence theory. A large number of detailed examplesillustrate the validity of the new method.