Research On Power Grid Fault Diagnosis Based On Graph Theory And Cluster Analysis Fusion Algorithm

With the development of IT and smart grid, the construction of power grid advances, applying wide area backup protection system based on fault diagnosis is an inevitable trend. Wide area backup protection can make a reasonable complement to the traditional protection systems. Currently, wide area measurement system based on phase measurement unit matures, and has been used in parts of power system, opens up new ideas for fault diagnosis information acquisition. Taking these two points, the paper carries out the research of grid fault diagnosis based on WAMS.WAMS is a real-time dynamic monitoring system observing the power system running state. The paper takes the actual WAMS system as an example to study the structure and the PMU base unit, summarizes the advantages of fault diagnosis based on wide area information.Traditional fault diagnosis method used relay protection, breaker and other remote communication informations, it is difficult to meet the wide area backup protection requirements on time. In this paper, the system adopt the real-time telemetry informations provided by WAMS, on the basis of analysis and comparison the fault diagnosis results of the hierarchical clustering and K-means clustering method, propose to combine hierarchical clustering and K-means clustering method for fault diagnosis.Against missing and incorrect information problems in data transmission systems, a new method of fault diagnosis combining grid graph theory and clustering analysis theory is proposed. Under normal and abnormal conditions of information transmission, the fusion method can locate faulty components quickly and accurately, improving fault tolerance in fault diagnosis. By contrast with the hierarchical clustering and K-means clustering method, the method combinate fusion method and K-means clustering is better to complete fault diagnosis.By simulating on IEEE39node power system, the validity of the fusion method is verify.In this paper, power fault diagnosis is extensed to wide area backup protection system, the structure of wide area backup protection system is studied, and action strategies are proposed. Adding the fault type identification process into fault diagnosis algorithm, this method can effectively improve the accuracy of fault diagnosis.The research of this paper is funded by special fund of Key laboratory of Guangdong Regular Higher Education Institutions(ZDSYSZOO701).