Research On The Scheme Of On-line Monitoring And Reliability Evaluation For The Stator Insulation Of Traction Motor Of Motor EMU

With the rapid development of high-speed railway in recent years,the safe operation of high-speed trains is particularly important.High-speed train refers to the train running at a speed of more than 200km/h.Traction motor,as the "Power Heart" of high-speed train,plays a very important role in the stable operation of the train.At present,high-speed trains supply power to traction motors mainly through the ac-dc-ac mode.In other words,the ac power transmitted from the power grid is first converted into dc through the rectifier,and then the dc is converted into ac power that can be used for frequency modulation and voltage regulation through the inverter.The inverter output has a pulse square wave with a steep rise/fall edge,which requires high insulation of traction motor.In addition,the traction motor has been in a bad working environment for a long time,which may cause the insulation of the traction motor to fail in advance due to the influence of electric aging,thermal aging,mechanical aging and other factors.Combined with the above background,this thesis mainly does the following work.Firstly,temperature field and electric field simulation of traction motor are carried out to explore the possibility of partial discharge of stator insulation.Motor-CAD software is used to simulate and analyze the temperature field.The results show that the high temperature of the stator insulation may cause thermal insulation aging.Then two-dimensional electric field simulation was carried out by Ansys software,and electric field distribution under four classical insulation defects was simulated.The results showed that partial discharge would be caused when insulation defects occurred in the stator insulation of traction motor,which would lead to the possibility of insulation failure.Secondly,the model partial discharge experiment and experimental data processing are carried out.According to four typical pd experimental models,a pd detection experimental platform is built.The principle of the experimental platform is that the discharge pulse data of the four typical PD models are collected by HFCT,transmitted by the sampler to the computer,and finally stored by the PD software.According to the pd data,PRPD spectra and grayscale images were generated and the features were extracted.The distance separability criterion and principal component analysis method were used to reduce the dimension.After dimension reduction,three classifiers were input,namely confidence interval,SVM vector machine and BP neural network,and the BP neural network with recognition success rate of 97.5% was selected by comparison.Finally,an experimental platform for partial discharge testing of traction motors is built,and an on-line monitoring and insulation evaluation scheme for traction motors is proposed.The partial discharge test scheme of traction motor was established based on the voltage withstand test of traction motor in TANGSHAN Plant of China Railway Rolling Stock Corporation.The discharge fingerprints of different discharge types are established as the premise of on-line monitoring of partial discharge of traction motor stator insulation.In this paper,the insulation reliability evaluation scheme of traction motor is proposed,which can reflect whether the insulation condition of traction motor stator under different pd types is beyond the current operating life by means of the weekly distortion coefficient,monthly distortion coefficient and annual distortion coefficient.