Research On Detection And Diagnosis System For The Lubrication State Of Axle Box Of High Speed Train

With the continuous extension of the high-speed railway line,people choose the way of rail travel is also moving in the direction of high speed,high comfort and long distance,this poses a more stringent requirement for the performance of high-speed electric multiple unit(EMU)trains,especially the safety and reliability of the transmission device.In the case of high speed driving,the wear particles generated by the relative movement of the rolling pair and the friction pair in the axle box carry a lot of characteristic information of wear fault.By analyzing the characteristics of wear particles in lubricating grease,such as morphology,concentration,texture and material composition,it can be used to judge the wear condition of the contact surface.Therefore,it is an important way to ensure the safe operation of the train,reduce the frequency of shutdown and reduce the maintenance cost.This paper based on the thorough understanding of the axle box wear mechanism and lubrication conditions of high speed EMU.First,the application of electrostatic monitoring technology in the wear particle monitoring of the axle box is studied in order to grasp the lubrication condition of each axle box of the EMU and provide the early warning information for the maintenance personnel in time;Then,the grease sample of the worn axle box is extracted,and the types of abrasive grains are analyzed by image processing technology,in order to determine the degree and the causes of the wear of the key parts of the axle box,to provide theoretical basis to achieve the purpose of judging the lubrication condition of the axle box.The main contents of this paper can be summarized as follows:Firstly,the filtering method of background noise and random pulse in electrostatic induction signal of wear particles is studied.The singular value decomposition of the original signal matrix,and the singular value of the effective signal and the interfering signal is obtained respectively,then the denoising of electrostatic induction signal can be realized by reconstructing the signal component corresponding to the larger singular values,finally,set reasonable threshold line on the output signal after denoising,when the amplitude exceeds the preset threshold line,the warning information of the wear fault will be issued,to remind the staff to enter the wear particle image processing module to analyze the wear type;Secondly,the digital processing algorithm of wear particle image is studied.The geometric,texture and color parameters of eight kinds of typical wear particles are extracted to improve the accuracy of wear particle recognition.A particle classifier is designed by using support vector machine,and the extracted characteristic parameters of wear particle are normalized to form a complete sample set is used as the input of the classifier,and then the encoding of the wear state is taken as the target output,and the fault diagnosis of train axle box is realized according to the mapping relationship between wear and lubrication;Finally,a set of comprehensive detection system of oil and grease based on the modular design and humanized design idea is designed,which includes particle electrostatic signal warning module,image processing module,wear fault diagnosis module of the axle box and historical data analysis module,provide technical support for determining the axle box oil change cycle and for comprehensive assessment of the wear condition of the key parts.The results show that the proposed method of the diagnosis of axle box wear state can effectively avoid the interference of human factors,improve the early warning capability of equipment wear fault,and it has high generalization ability of particle type recognition,it can be used in the wear state diagnosis system of train axle box.