Based On Wavelet Analysis And Artificial Immune Power Transformer Fault Diagnosis Method

In the present age electric system,the electric power transformer is one of important,the most expensive equipment,it is related to the stability and reliability that electric system runs directly,therefore, occurrence our necessary high limit field being guarded against and cutting down the electric power transformer malfunction and the accident. That part discharges is important transformer inner insulation degradation cause,is also to reflect important transformer insulation situation portent,transformer part is discharged monitoring the method reaching principles of pattern recognition being to check the transformer inner insulation defect and appraising insulation situation importance on line,various being how to remove a scene is disturbed whose maximal difficult problem,is drawn and by the principles of pattern recognition measuring the part discharging signal characteristic and carrying out discharging.The author distinguishes the main body of a book on analysing home and abroad electric power transformer part discharging basis about technology disturbing restraint and principles of pattern recognition monitoring in gleam,to the disturbance shifting restraint cyclicity narrow belt and chirping vainly,statistics characteristic amounts drawing small wave in part discharging signal,with statistics characteristic amounts discharge pattern method and so on be in progress system study.1,The thesis have analysed transformer noise signal source and characteristic property,the signal diversity frequency characteristic property having studied noise signal and part discharge, studies the signal making use of the frequency region that the minor wave shifts to carve up characteristic property,a bureau being set free complies with method extracted out in collecting the signal arriving at by scene.And have been in progress simulating studying and have analysed to this method,whose result indicates wavelet analysis law restraint disturbing an ability strong,distribution not needing ascertaining a disturbance in advance,can various accurate detecting part discharging signal information.2,The thesis has been studied apply to transformer malfunction diagnose principles of pattern recognition technology method,immune manpower algorithms among them are that one kind of random appearing in recent years optimizes an algorithm,that it copies immune living things system principle's example,produces an antibody by a series of complicated mechanism such as clone,duplication,dissociation, achieves die out antigen's purpose,is one kind of fairly good principles of pattern recognition method.3,This article uses the shelf depreciation characteristic quantity which the wavelet transformation withdraws,unifies it with artificial immunization's method,carries on the pattern recognition,the discrimination different electric discharge type,its result indicated that this method,can distinguish the shelf depreciation signal effectively the pattern,provides intuitively,the effective theory basis for the transformer failure diagnosis.That the technology the main body of a book is distinguished mainly owing to algorithmic immune pattern,combines its wavelet analysis law each other,to realize electric power transformer part discharging type diagnose,makes that become a method both have theory value and have project practical value.