Research On The Identification Of Typical Subway Line Condition Characteristics

It is self-evident importance of the subway in China. Along with the major urban construct subway, more and more people choose to travel by subway. But the metro bogie occurs cruck frequently before the service life, seriously affect the safe operation of the subway. This shows that, metro bogie designed and produced according to the international standard can not completely adapt to domestic subway line condition. The establishment of standard suitable for domestic subway line condition is imminent. Identification of subway line condition is very import of the study on the influence of different line conditions on the subway vehicle’s fatigue life. It also provides an important basis for the establishment of standard suitable for domestic subway line condition. In this paper, the research work and its conclusions are mainly the following three points:(1) With the finite element analysis, casting structure analysis, modal analysis, dynamic load transfer analysis and practical application, determine the position of measuring point on the metro bogie. Carry out the test on Vehicle strength of Beijing subway line six, and get the stress signal, acceleration signal, speed signal and gyroscope signal when the vehicle on the line. Through the pretreatment, remove the abnormal signal, retain the identified signal with features of the subway line condition.(2) Analyze and contrast of different method of signal feature extraction and signal feature identification. Signal feature extraction method is divided into time domain method and frequency domain method. Compared with the time domain method, the frequency domain method describes of the signal more detailed and deeply, including fourier transform, wavelet transform, Hilbert-Huang transform, and so on. The three methods have their own characteristics and advantages. Fourier transform can analyze the spectrum characteristics of the signal, wavelet transform can denoise and compress signal, Hilbert-Huang transform can adaptively decompose signal into several intrinsic mode functions (IMF). Signal feature identification methods mainly include traditional neural network and support vector machine (SVM), and so on. Through sample training, they can identify the signal characteristic intelligently to realize pattern identification. By contrast, SVM has better performance in identification of subway line condition.(3) Determine the method of subway line condition identification. According to the type of signal and the sensitivity of the signal to the subway line conditions, identify curve by gyroscope signal, and identify turnout by acceleration signal. The combination of them can identify curve and turnout in subway line. At the same time, analyze the rail gap signal’s characteristic of time and frequency domain. In the process of identification, it use spectrum analysis, wavelet packet de-noising, EMD and SVM. Multistage identification method is adopted to achieve the desired identification. By contrast, confirm several key identification parameters of the method. Apply the method to Beijing subway line six, and get a higher accuracy.