Research On The Rail Irregularities Detection And Surface Defect Identification Based On In-service Vehicles

To achieve high-speed, high-load, high-density goals and protect the rail transportation safety, the rail health state is managed and monitored strangely. But the traditional detection methods can’t reach the requirements of dynamic and real-time. The track irregularity detection system based on in-service vehicles has many advantages, such as high efficiency, short detection period, little impact on the in-service vehicles, and it has caused great attention in the world. However, the track irregularity detection method based on in-service vehicles has some technical problems to solve, such as, without a common gauge management standard of detection, lack of researching on surface defect, etc. Therefore, this paper puts forward a method of how to draw up the standard of the track irregularity detection system based on in-service vehicles, there are also many researches on various methods of monitoring the rail health state and identifying the rail surface defect, which provide the new thought on the rail irregularity detection and the rail surface defect identification.The main research results and techniques as follows:(1) The gauge management standard of the track irregularity detection system based on in-service vehicles is drawn up. When the vehicle operating on the irregularity or surface defects rail, there are more or less impacts on the train. So, the abnormal rail at the same position can be identified by the track inspection car and the track irregularity detection system, but the standards of the two methods are different, which is resulted from the different detection devices in the two methods. Based on this principle, lots of rail detected signals between the Shanghainanzhan station and the Jinjiangleyuan station are processed and compared with the track irregularity measured by the track inspection car, which provide the standard of the track irregularity detection system based on in-service vehicles. And the standard usability and the repeatability of the track irregularity detection system detecting rail are validated.(2) The rail health state monitored based on the bogie acceleration. In this paper, the rail irregularity signal is processed by the fast Fourier transform to acquire the amplitude spectrum. The rail irregularity signal described by the amplitude spectrum, reflecting the amplitude distribution characteristics of the rail irregularity signal in the frequency domain, and describing the rail irregularity consists of wavelength that influence the rail health state strangely. While, the spatial distribution characteristics of the rail health state cannot be reflected by the amplitude spectrum, then it becomes necessary to calculate the root mean square and the peak to peak becomes necessary. The root mean square analysis can identify the abnormal extreme amplitude of the rail detection signal, and the peak to peak analysis can be describe the vibration intensity in this segment much more directly. Through the root mean square and the peak to peak analysis reflect the distribution characteristics of the rail health state with the change of the position are reflected.(3) The rail surface defect identification methods are researched by the bogie acceleration. As the rail surface defect signal frequency is complicated, the different wavelength defect corresponds correspond the different frequency signal. The rail detected signal is decomposed into variant frequency bands signal by wavelet packet decomposition, part of the nodes represent the typical rail surface defect signal, and all kinds of the rail defect signals are processed through the amplitude spectrum analysis to identify the type of the rail defect in the segment. Finally, continuous wavelet transform analysis of the rail surface defect signal of each defect node is done to confirm the defect position.The main contribution of this paper:The gauge management standard of the track irregularity detection system based on in-service vehicles is drawn up; the identification method of the rail health state is provided by the bogie acceleration detection based on the in-service vehicle; the identification method of the rail surface defect is provided by the continuous wavelet transform analysis, etc.