Study On Chaotic Characteristics Of Partial Discharge Of Electrical Treeing In Oil-paper Insulation

Oil-paper insulation system is used in most power transformers,and partial discharge of oil-impregnated pressboard due to defects and other factors is a very common phenomenon.Electrical treeing is an important insulation defect that will eventually lead to the insulation breakdown as the partial discharge develops.Some features of partial discharge are similar to chaotic systems.Chaotic identification and characteristic analysis of partial discharge of electrical treeing have significant necessity for the study of discharge and the diagnosis of material degradation.In this paper,the data samples of 36 groups of partial discharges of electrical treeing under the six voltage levels of 24kV to 29kV are taken as research objects,which are obtained in the laboratory.Firstly,the development laws of characteristics with deterioration are statistically analyzed for 5 kinds of discharge,such as pulse frequency and average discharge amount.These parameters cannot directly provide effective information for the end stage and pre-breakdown stage,and the average discharge does not even have a uniformity.And this paper constructs time series,which are time series of pulse discharge and time series of reverse polarity discharge.It is proved that the partial discharge under positive and negative feedback of discharge residual charge in oil-paper insulation has deterministic chaotic characteristics,by the power spectrum analysis,phase space attractor reconstruction and Poincare cross-section construction of two kinds of time series.Moreover,with the development of degradation,both the chaotic attractor and the Poincare cross section show obvious changes,and the significant change of the morphogenesis of the attractor can be used as a strong criterion for the diagnosis of deterioration.The calculation and analysis of the chaotic feature quantities of the two kinds of time series are further carried out.The correlation dimension of the two time series shows a significant growth trend with the deepening of the deterioration.The thickening and merging of the“branches" at the end of the deterioration lead to the decrease of the correlation dimension,which can be used as effective information for the end of the deterioration.The Kolmogorov entropy of the time series of pulse shows an overall increase with the development of degradation,and has a decreasing trend at the end of the degradation.It can also be used as the characteristic information of the end-stage and pre-breakdown stages of the insulation pressboard.The Kolmogorov entropy of the inverted polarity discharge sequence and the Lyapunov exponent of the two sequences are highly volatile,and there is no consistency law.On the whole,the pulse discharge sequence has better effect on the diagnosis of pressboard deterioration and the end of deterioration.There is no corresponding relationship between the development law of chaotic features such as Lyapunov exponent and correlation dimension and the development of traditional statistical discharge parameters such as pulse frequency,average discharge,1 second discharge,cumulative discharge and maximum discharge.