Research Of Power Grid Fault Diagnosis Based On Multi-source Information Fusion

With the continuous development in power grid, interconnection between differentregions are more and more close, which makes the system fault more influential to thepower system. At present, power grid automation is developing at a high speed. Soinformation is becoming more and more convenient to obtain. Once complex fault occurs,lots of alarm and fault information swim into dispatching centre in a short time, which isbeyond the dispatching personals’ handling ability. This will usually lead to incorrectdiagnosis. So it is necessary to rely on fault diagnosis system for fast and correctidentification. It can help dispatching personals with decision to ensure the safety ofpower system.However the traditional researches of fault diagnosis are mostly based on theinformation of protection and circuit breaker action to recognize the fault elements usingkinds of intelligent algorithm. The degree of completeness and accuracy of alarminformation will have a great impact on fault diagnosis. In an actual situation, protectionrelays and circuit breakers may fail to operate or refuse to act, and some alarm may befalse or lost in the transmitting process, which have negative impact on the accuracy offault diagnosis result. Given this background, this paper fully considers thecharacteristics of electrical measurement after fault and adds it to fault diagnosis ofpower grid making use of its timeliness and accuracy. Comprehensive Diagnosis isconducted by multi-source information fusion.The characteristic of fault recorder information after fault is firstly analyzed, and theelectrical information fault diagnosis model based on Hilbert-Huang Transform isestablished. On the base of Fast Intrinsic Mode Decomposition and Hilbert Transform,Fault degree is obtained. So the fault can be diagnosed. Secondly considering thecharacteristic of WAMS information deeply, different element has its own fault feature. Afast fault diagnosis method based on WAMS measurement information is established bysetting up different criterion model. Acquisition of fault information though powernetwork topology results can recognize the fault element effectively. Making full use ofWAMS measurement information can make up for the shortcomings of traditional faultdiagnosis method in protecting information missing. Then, considering the reliability ofprotection and breakers, fuzzy degree in introduced to analytical model, separate faultdegree can be obtained by switching and electrical information, which are fused by D-Sevidence theory. Diagnosis results can be obtained though fuzzy K-means model. Amulti-source diagnosis method is realized, simulations of IEEE39system prove thefeasibility and superiority of the method. Finally, on the base of Yunnan power gridinformation platform, topology and fault diagnosis programming of Yunnan powersystem are conducted.