Research On The Transformer Fault Diagnosis Based On Synergetic Theory

The transformer fault diagnosis is very important to the safety and economicoperationofpowersystem.DissolvedGasAnalysis(DGA)isoneofthemosteffectivemethods to predict and diagnose transformer faults. The transformer fault diagnosishasmadegreatprogressattheprinciplesandmethodsthroughthehardworkofallthedomesticandforeignscholars,buttherearestillmanyproblemswewant tosolve.Allthe methods that we use usually have their strongpoint; also they have weaknessesand deficiencies. Some methods have the weakness of coding incomplete ordiagnosing fault not comprehensive, other's structure is complex and it is difficult tomaintain, also there are some methods which is hard to converge. So it is veryimportanttofindanewmethodtodiagnosetransformerfaultefficiently.In this thesis, on the foundation of analyzing the transformer fault diagnosisprecisely, a new method is raised which can diagnose the transformer fault efficientlybased on synergetic theory. This method is like the pattern recognition in artificialneural network, but its algorithm is more flexible compared to the BP network. Alsothe training and the selecting neurons in hidden layer is very easy. First, somedynamic equations about the transformer fault diagnosis are built based on synergetictheory. Then, we can get some order parameter equations which can reflect all thefeaturesexistintheoriginalequations,bythedeductionandadiabaticelimination.Wecan judge the type of the transformer fault only by comparing the initial value of allthe order parameters. Summarize the process of analyzing the problem; the wholestepstojudgethetypeofthetransformerfaultbasedonsynergictheoryareobtained.At last, it is proved in some real examples that it is feasible to judge the faulttype with the synergetic method. When it is compared with the Three-Ratio Methodand BP network, we find that coding incomplete can be avoid in synergetic method,also there is higher accuracyin fault diagnosis. The synergetic method is simple to berealized. There is no pseudo-state in the process of the transformer fault diagnosis.Anyway,itisaveryidealmethodtobeusedinthetransformerfaultdiagnosis.