State Assessment Of Oil-Paper Insulation For Power Transformer Based On Relaxation Response Characteristic Quantity

The subject derives from the sub-project of National Natural Science Foundation of China project’the diagnostic method of oil-paper insulation aging based on circuit analysis and return voltage response(Grant No.61174117)’.Large power transformer is the key equipment of power transmission and transformation.The research of oil-paper insulation condition evaluation is of great engineering significance for the stable operation of power system.Taking the oil paper insulation state of power transformer as the research object,based on the study of relaxation response characteristics,a new method to identify the relaxation mechanism parameters of oil-paper insulation considering free relaxation is proposed.At the same time,multivariate mathematical function model by regression analysis to build furfural content in oil is established,which is by in-depth analyzing the recovery voltage line,trap density and the relaxation response characteristic parameters of the equivalent mechanism parameters.On this basis,a comprehensive diagnosis system of oil paper insulation state is established by using the grey target classification method of oil-paper insulation state section based on sample set.The main work of this paper is as following:(1)Based on information of return voltage line relaxation parameters,that approach introduces free relaxation factor beta from microscopic dynamics,containing different dielectric relaxation response information of branch line,and determines the topological structure of transformer insulation.That method can not only identify the topological structure of oil-paper insulation more accurately,but also reflect the physical meaning of relaxation process,which can improve the accuracy of the transformer insulation topology modeling,and lay the foundation for subsequent extraction of new features to evaluate the status of oil and paper.(2)Based on spectrum analysis of topology of transformer oil-paper insulation,this paper calculation of equivalent circuit parameters of oil-paper insulation using pseudo physics algorithm.On this basis,the characteristics of oil-paper insulation state reflected by equivalent circuit parameters are deeply studied.Firstly,find the relationship between the number of relaxation mechanism and aging of oil-paper insulation through the measured data.Secondly,a new characteristic quantity,named average time constant,for diagnosing oil-paper insulation state is proposed.and its internal relation with furfural content in oil is analyzed.Finally,this research puts forward a new feature,named free relaxation factor,which provides a new way to determine insulation state of oil and paper by actual test and calculation.(3)Starting from time domain dielectric spectroscopy diagnostic technique,the characteristic spectrum of recovery voltage spectrum is extracted and the correlation between it and furfural content in the oil is studied.In order to mine more feature information on spectral lines,starting from the depolarizing current test technology and trap density theory,this research proposes depolarization trap density spectrum,and then extracts the new features of the two new features,which are peak value and peak time constant,to evaluate the state of transformer insulation.Finally,it is proved by examples that different numerical ranges of characteristic quantities for different oil paper insulation states can be used as reference for diagnosis of transformer oil paper insulation in the future,which makes up for the gap of characteristic quantity extracted from depolarization current spectrum lines.(4)This thesis integrates the time domain dielectric spectrum characteristics,introduces a comprehensive evaluation system of insulation state.comprehensive relaxation response characteristics.Firstly,optimal comprehensive weight weighting which based on sum of squares of deviations,can improve fixed value problems of the traditional grey target evaluation scores approaching degree and the fixed value of linear superposition by distributing the subjective and objective weights of calculation results.Secondly,the grey target classification method based on the sample set is used to calculate the grey target classification table,which solves the problem that the critical value of the intermediate grade is difficult to be determined in the present state assessment method.Finally,a comprehensive evaluation method of transformer oil-paper insulation based on the combination of comprehensive weighting method and grey target theory is proposed,which solves the pattern recognition problem of transformer oil-paper insulation state in the absence of standard failure mode.