Research On High-speed Railway Inclined Cracks Feature Recognition Using 3d Mfl Technology

The inclined cracks that have a certain angle to the rail trends are the major damage mode of the high-speed railway rail surface, the rapid and accurate identification of which is of great significance. 3D magnetic flux leakage detection technology has the characteristic of fast detection, and without the crack main plane perpendicular to the magnetization direction. What's more,it can carry on the rapid scanning a large area with the sensor array structure and imaging technique, which can reduce the non-detection zone. Therefore 3D array sensor with magnetic flux leakage technology is used in this paper to achieve accurate identification of rail inclined crack.This paper firstly investigated the major damage modes and features of the high-speed railway rail surface and probed into the current detection means and their advantages and disadvantages, which also proposed to accurately identify the inclined cracks with the 3D magnetic flux leakage inspection technique. Then the principle of magnetic flux leakage testing, the signal feature and causes of the magnetic flux leakage detection are discussed theoretically. The leakage magnetic field distribution of magnetic field lines of the inclined crack is analyzed and put forward a method to identify the angle between the inclined cracks and rail trends. Ansoft finite simulation software is used to carry on the 3D static simulation in the inclined cracks of rail surface and verified the feasibility of the identification method.Based on the above, three-dimensional array of sensors leakage magnetic system experimental platform is designed ,including the signal excitation, three-dimensional array of Hall sensors, signal amplification and conditioning, encoder position sensor, data acquisition part and computer processing part. Using the experimental platform, the experiments on the railway rail with artificial cracks have been undertaken. The detecting signals are processed by displacement reconstruction and wavelet de-noising method. Then the signal eigenvalues are extracted from the crack leakage field Bx and By in the time domain respectively to calculate the angle of the crack, then the sensor array data imaging analysis to verify the accuracy of the method.Finally, the work of this paper is summarized and the suggestion for further research is provided.