Study On The Leakage Current Characteristics And State Detection Of Contamination For Polluted Insuloator Of Catenary

In recent years with the development of science and technology of railway and in order to satisfy people's growing demand for rail travel,railway construction scale is also expanded year by year in China,and Electrified railway occupies an important position in all kinds of railway.Catenary is an important part of the electrified railway,catenary insulator as the important equipment which connects electrically charged components and other uncharged parts of catenary.Since railway spans fairly long,catenary insulator has a wide distribution along the railway,most of them work at the scene of the outdoors and complex environment for a long time,their surface are easily contaminated.Especially in recent years,many cities are suffering from the fog haze weather.Practice has proved that if catenary insulator has been exposed to bad weather conditions,such as rain,fog,dew,which will no doubt increase degree of the contamination on insulator's surface.And will cause its creeping discharge easily,even develop to the insulator pollution flashover accident,so it will endanger the safety of the railway operation.In order to improve the early warning ability of catenary insulator pollution flashover accident.The effective method is to use leakage current online detection combined with equivalent salt deposit density(ESDD)testing method.So studying the characteristics of the insulator surface leakage current and ESDD,and analysis leakage current the waveform and spectrum characteristics within the scope of time domain and frequency domain respectively,and extract the feature of ESDD and leakage current characteristics.It has a certain reference value and practical significance in insulator contamination degree detection.This paper is based on artificial contamination test cannot fully reflect contamination accumulation rule and external surface insulation.First conducted a year-long.Built leakage current characteristics,environmental relative humidity and ESDD database in insulator natural contamination depositing test.Catenary insulator's contamination accumulation overall relatively quickly in relatively dry,cold climate and more industrial and mining enterprises and fog haze all the year round these relatively poor conditions especially in spring and winter,the accumulation is faster.However because of the climate is relatively wet with rain sludge accumulation rate slows and even reduced in summer.Next,for leakage current signal is collected with the train-pass interference and other noise disturbance,use the catenary insulator leakage current on-line detection device to develop train-pass experiment.We have been used the way which is based on EMD(Empirical Mode Decomposition)combined with power to Identify interference data when train cross the insulator column.And removed the train-pass data segment directly from the original leakage current data segment in order to achieve the purpose of removal of vehicle interference.Then analyzed the advantages and disadvantages of the EMD de-noising method,and combined with the advantages of wavelet threshold denoising method,used the wavelet threshold denoising based on EMD method for denoising leakage current.Practice shows that the wavelet threshold denoising method based on EMD relative to the EMD method in denoising to maintain the effective information of leakage current signals at the same time,it's suitable for insulator leakage current signal denoising processing.Then based on the leakage current signal data after pretreatment,Analysis the time domain and frequency domain characteristics of it.extracted the leakage current of effective value,maximum value,three times and the harmonic amplitude ratio and the total harmonic distortion rate of four characteristics under different degree of contamination.Analyzed the respective changes in the relationship between ESDD and leakage current characteristics in curve figure and data fitting method.Results show that there have a close relationship between ESDD changes on surface of the insulator and leakage current characteristics.And result of fitting are all over 90%,which meet the demand of engineering,can be used as a leakage current characteristics for further analysis.Finally using the leakage current characteristics,environmental relative humidity and the ESDD database under insulator natural contamination depositing test,which have been established before.Made leakage current characteristics and environment relative humidity as an input parameter,the ESDD as an output parameter,built the support vector machine(SVM)regression model to predict and assess the contamination accumulated degree of catenary insulator surface.Used 20 groups of training samples for training support vector machine(SVM)model to get the best model.And Used 8 groups of test samples for support vector machine(SVM)model to forecast evaluation,the results showed that the model output predictive value relative to the actual value have a high correlation coefficient,a smaller mean square error,and the error wasn't exceed permitted extent in practical engineering.Therefore,this model can effectively predict the degree of catenary insulator surface contamination.