Partial Discharge Signal Analysis Of Transformer Bushings And Identification Of Air Gap Size Of Oil-impregnated Paper

The partial discharge that occurs in electrical equipment refers to an electrical discharge,which occurs locally and limitedly in their insulating parts when the electrical insulation is under strong enough electric effect.This phenomenon of discharge can only cause short circuit partially between conductors without forming a complete circuit.In the meanwhile,different levels of PD(partial discharge)bring about different degrees of damage to the insulating medium.Therefore,the equipment insulation strength will drop rapidly while the PD proves serious enough.The most momentous device in the power system must be the transformer which has a wide range of uses.At present,oil-immersed transformers are used widely.The insulating oil-impregnated paper and insulating sleeve in it play an important role in the insulation of the transformer.Partial discharge is one of the most important causes of insulation degradation,and it even results in destruction of power transformers.Thus,it is of great engineering value to study further how to detect partial discharge in transformer accurately and sensitively.So far,the commonly used methods for PD pattern feature extraction include statistical feature parameter method,wavelet feature parameter method,fractal feature parameter method,digital image moment feature parameter method,and waveform feature parameter method.The difficulty of UHF PD detection is to determine the intrinsic relationship between the characteristics of all kinds of signals and their discharge parameters.Now,there are few studies on the related aspects at home and abroad.Using the finite-difference time-domain method to calculate Maxwell's equation,the electromagnetic wave propagation process can be quantitatively analyzed.In this paper,XFDTD is used to analyze the bushing partial discharge signal of the transformer,and the propagation law of the high-frequency electromagnetic wave is compared with the setting position of the sensor.Through the modeling and analysis of the casing,it was found that with the increase of the propagation distance of the electromagnetic wave in the casing,there is a significant attenuation phenomenon.Moreover,there is a certain degree of attenuation not only in the axial direction but also in the longitudinal direction.In addition,from the simulation results,as the inner diameter of the electromagnetic wall becomes smaller and smaller,the reflection phenomenon of electromagnetic waves becomes more and more obvious,and there is a significant increase in the electromagnetic wave at the smallest inner diameter.Therefore,the algorithm for detecting the position of a partial discharge defect by using the attenuation speed of the electromagnetic wave should take into account the phenomenon of the electromagnetic reflection in the waveguide system.Since transformer oil-impregnated paper does not have a good waveguide system,analysis of transformer oil-impregnated paper cannot use high-frequency electromagnetic waves that analyze the discharge of the casing layout.Therefore,the analysis of the partial discharge of oil-impregnated paper uses the characteristic parameter method of discharge and the three-dimensional map to jointly analyze,and the size of the air gap in the oil-impregnated paper is identified using the K-means clustering method.In this paper,the experimental design of the high-pressure discharge test,the artificial manufacturing department of the three different air gap size of the oil-impregnated paper discharge test,through a step-up method to insulation defects in the oil-impregnated paper breakdown.During the interval,a data collection of discharge information on oilimpregnated paper was made every ten minutes.The collected data information includes the phase ? of the voltage at the time of partial discharge,the amplitude of discharge q,and the number of times of discharge n.Use the collected data to calculate the PD parameters at this time.Finally,K-means clustering algorithm was used to train and classify oil-impregnated papers with different air gap sizes.The experimental results show that the K-means clustering algorithm has very high precision in the identification of the air gap size.By extracting the statistical parameters of partial discharge of oil-impregnated paper with two different air gaps in the experimental data,it is input to the trained system and its accuracy is >89%.