基于GA祖糙集和贝叶斯网络的高压断路器故障诊断-电力系统及其自动化专业论文.docxVIP

Abstract Abstract Title:SYNTHETIC FAULT DIAGNOSIS METHOD OF HIGH VOLTAGE CIRCUIT BREAKER BASED ON GA ROUGH SET THEORY AND BAYESIAN NETWORK Major: Electric Power System and its Automation Name: Guanxiong WANG Signa阳ture:W阳 叫价u恻俨叫矶瓜b斗fj s创u叩p阳e例阳圳阿川川i丽附臼s阳。r伫:阳Jia叫ng呻bin 川NG Signature牙吨以 Abstract High voltage circuit breaker is the key equipment in power system，related to the safe and rcliable operation of the power grid. I f the high voltagc circuit breaker failure wi ll bring losses，so the rapid and effective fault diagnosis for high voltage circuit breaker system can quick and accurate to find the fau11 posilions，for the time1y fault processing to provide a guarantee，this has the viωsigniftcance to rapid1y restore the norma1 operation of e1ectric power system. At present ，in the aspcct of fau1t diagnosis for high voltagc circuit breaker has achieved somc results，mai n1y in the use of arti ficia1 intelligence analysis ofvibralion signa1，high voltage circuit breaker closing coi1 by using neura1 network and the waveform ofthc corresponding time for training，as well as the rcso1ution based on rough sel matrix and so 011，bul a1so the existence of signal exlraction more difficu1t，the diagnosis resu1ts can be afIected by the the influence of the surrounding environment and fauJt detection is some charactcristic quantity difficult 10 extract and sometimes local optimal solution of the probJem ，at the same time in the presence of without special training data will be lower when the problem of Jow accuracy in diagnosis. Jn this paper，the combination of genetic algorithm (GA) of the rough set and Bayesian network are combined together，the first use of GA rough sets reduction of attributes，and according to the prior probability of signs and causes of the failure，calculate in the Bayesian network，without the need for training to the network，do not extract precision vibration signa1 sore vulnerable to environme