Research on axle press-fit area defect detecting and denoising

With the rapid development of high-speed and heavy-haul railways, the security of railway is increasingly important. Under the combined actions of tension, compression, distortion, and shear stress, the railway axle material is prone to fatigue damage. Defects in the axle press-fit area can only be detected by using ultrasonic testing methods, with ultrasonic probes coupled to the axle journal, the axle end face or other detection schemes. No matter what schemes, compared with the bare axle, ultrasonic testing in the axle press-fit area is more complicated because certain factors, such as propagation of portion of ultrasonic energy into the wheel hub, resulting in less energy reflected back by the defect, and echoes from the axle press-fit interface, decrease the SNR. In order to improve the SNR and decrease the probability of false positives and false negatives, 2-D wavelet denoising is applied to axle journal and end face detection schemes. Denoising parameters are analyzed separately for the two schemes, and the performance of the denoising algorithm is assessed using axle press-fit area detection data.