On-line Monitoring Technical Development Of Rail Deformation

With the countrywide increasing in train speed and running heavily loaded trains, the security issue of the railway is becoming prominence. It is especially important to ensure the security and celerity of the railway transport.An on-line rail deformation detection system has been established in this subject. The system consists of an on-train module and a ground station. The on-train module was aimed to measure vibrations of the train in different directions, to obtain the speed, time, milestone, section and train number information of the running train, and to send the gained information to the ground station by CDMA. The ground station further processes the acceleration and sends the information to the maintenance unit of the railway for confirmation and repair.A rail deformation detection algorithm based on multi-threshold feature extraction was presented in this paper. The waveforms'amplitude character will be more distinct through accelerations'feature extraction, and this is propitious to rail deformation detection.The vibration source identification was realized in this paper by analyzing and comparing the train's and the bogie's lateral acceleration. The proposed method can divide into four steps: first adopting the acceleration's peak-peak value entropy comparison method, second peak-peak value cross correlation comparison method, third peak-peak value weight center position comparison method, and the last adopting peak-peak value max-value position comparison method, if still unidentified refuse it.The decision tree was applied in the rail deformation detection in this paper. The decision tree (or rules) used for rail deformation detection was generated by learning the train data.The driver's operation level assessment method based on the instrument for impact detection was discussed, and the method was simply tested in this paper.