The Method Research Of Fault Diagnosis For The Transmission System Of CRH5 Basing On Second Generation Wavelet Transform

At the beginning of the 21st century, high-speed train technology is developing rapidly in China, the safety operation of high-speed trains becomes more and more important.Due to the special structure of transmission system(equipped with a overhung motor and a universal shaft which is 2 meters long)and its high speed service state(the design speed is 200Km/h and actual operation speed is 250Km/h),in order to ensure its safety operation, it needs online fault diagnose.A vibration sensor is equipped on technology hole in gear box of CRH5, in order to collect vibration signal of transmission during running process. According to the structure of CRH5, failure mode and the frequency of vibration signal caused by failure can be concluded. It is used to completely reconstruct original signal and has an ability to extract different frequency components of signal of the second generation wavelet difference subdivision algorithm, the features of signal in different frequency are extracted out by using this algorithm. Because of the existence of spectrum aliasing of the equivalent filter from this algorithm, singular value decomposition algorithm is used to reconstruct eigenvalues which are located in specific frequency band. In this way, noises and spectrum aliasing are reduced. A bench test was designed whose mechanical structure is same with the transmission system of CRH5. Cardan shaft of different unbalance value were used to excite on gear box in bench test. Signals were analyzed to verify whether the algorithm is right or not, then the way to compute parameter is out forward.In combination with SVM (support vector machine) algorithm and running train data samples, classify model is figured out through training. It can be judged whether the transmission system is in safe state or not by the classifying of data using classify model.