Research On Fault Prediction Method Of Contact Wire Based On Data Drive

Overhead Contact System(OCS)is an important link to transmit electric energy to the electric locomotive or EMU.Its quality and working state will directly affect transport capacity of electrified railway.The contact wire is in direct friction contact with pantograph,which needs to bear impact,wear and working tension.Once the operation status is abnormal or damaged,it will directly affect the safe operation of high-speed train.Due to the influence of complex working environment,the fault of contact wire occurs from time to time.If there is an intuitive prediction of the fault before it occurs,it will be helpful to prevent the trouble and deal with the fault quickly and accurately.Therefore,based on the historical fault data of contactwire,it is of great significance to study the fault characteristics of contact wire based on data-driven,and to predict the future operation status of contact wire.Firstly,this paper summarizes the research status of fault prediction of contact wire.The technology of fault prognostic and health management(PHM)is to obtain the running state and information of the studied object through advanced state monitoring means,and to detect,analyze and predict with the help of intelligent algorithm and model,so as to manage the working state of the system or equipment.The fault prediction of OCS is a research hotspot in the field of railway at present;A large amount of data has been accumulated in the detection,monitoring and maintenance of railway catenary,which lays a foundation for the fault prediction of contact wire by using data-driven method.With the rapid development of data mining technology,it is no longer a problem to use complex algorithm and intelligent model to deeply mine and analyze historical data of contact wire,and then predict the fault.Secondly,recognize contact wire fault and analyze the data characteristics of contact wire fault.In this paper,the fault of contact wire is defined as the defect of characteristic parameters of OCS directly related to contact wire and the defect of technical state of contact wire,so as to carry out a more comprehensive research on the operation state of contact wire.The statistical analysis of catenary fault data of Chengdu Railway Bureau from January 2016 to March 2018 shows that: contact wire fault is the most frequent occurrence of catenary fault.In the time sequence analysis of contact wire fault data,the concept of contact wire fault intensity is defined,and the statistical analysis is carried out according to the monthly fault amount of contact wire;The periodic detection and state maintenance are carried out in the railway catenary system.The statistical analysis shows that the periodicity of the contact wire fault intensity is related to the period length of the periodic detection of the contact wire related equipment,and the contact wire fault intensity has a dual trend of growth and fluctuation.At last,using the combination prediction method with the idea of integrated algorithm,a prediction model of contact wire fault intensity based on time combination is proposed.The time series of contact wire fault intensity is divided into monthly development series and multiple quarterly development series.Holtwinters model is used to solve the monthly development series with seasonal variation and the quarterly development series with small range fluctuation.GM(1,1)model solves the quarterly development sequence with uncertain factors,and then combines the prediction results of two series to get the final prediction results;Through the example data verification,it is concluded that the combination prediction model has good fitting effect and high prediction accuracy for the development trend of contact wire fault.The work done in this paper has a certain reference value for understanding the rule of contact wire fault occurrence changing with time,predicting the future fault situation of contact wire,and guiding the detection and maintenance of contact wire.