Fault Prediction Of Metro Door Transmission System Based On ARIMA Model

Metro door is one of the frequent startup components of Metro vehicles,and its reliability gets a lot of attention due to the safety of passengers and the order of metro network dispatch.As an important executive structure of Metro door,mechanical transmission system is the key to ensure Metro door can be opened and closed normally.Effective collection of degradation,modeling and prediction of degradation state and early failure of transmission system are particularly important to ensure the normal operation of Metro doors and orderly work of Metro network.New on-line vehicles refer to newly assembled vehicles which need a phase of operation debugging in order to adapt to the local operating environment.Therefore,for new on-line metro vehicles which are being debugged in the new environment in adaptation period,the probability of system failure is relatively high,and its degradation trend is different from that of vehicles in the mature operation period,because the mechanical transmission device is in the running-in period or part of the parameters are in the debugging period.Therefore,how to effectively collect,model,analyze and predict the degradation trend of new on-line vehicles can provide technical reference for effectively shortening the debugging period of new on-line vehicles.Based on this,the research contents of this thesis are as follows:(1)According to the structure and working principle o f the Metro door transmission system,the kinematics analysis of the motor-screw transmission is carried out,and the rationality of clamping force as the fault prediction data is determined.According to the characteristics of Metro doors,a clamping forc e measuring device with adjustable spacing and a clamping force data acquisition system is designed and developed.Then the functions of real-time acquisition,storage,display and query of the clamping force data of Metro doors are realized through the host computer analysis software.(2)Based on the modeling methods of different time series models,the transform relationships among the four time series models which include AR?MA?ARMA and ARIMA are analyzed.The four different time series models are compa red in terms of model identification,parameter estimation,model checking and model prediction,so as to select a prediction model which is suitable for the characteristics of the clamping force of subway doors.(3)For new metros,the clamping force data at different stages of operation are collected.Based on the collected clamping force data,ARIMA time series prediction model is applied to predict the trend of the mean and variance of the peak clamping force with cumulative running time,and the probability of early failure of the Metro door transmission system is determined by the failure threshold.By comparing the predicted data with the test data,the results show that ARIMA model can reasonably predict the changing trend of the peak of clamping force and the failure probability of the subway door,and shorten the debugging period of the subway door to a certain extent.The data acquisition system can support the establishment of the initial model of Metro door faults in the debugging stage,and provide theoretical support for the design of fault prediction scheme in the debugging period.