Research On Signal Processing Method Of Ultrasonic Detection In The Metal Plate And CRH Train Wheels

The development of modern ultrasonic nondestructive testing technology puts forward higher requirements for the signal collected by the instruments, so proper processing method is urgent in order to get better detection results. The research of this thesis is on two aspects of signal processing method. On the one hand, because of the frequency dispersion and multi-mode, Lamb wave signal of large metal plate detection requires exchanging frequency analysis on the ultrasonic signal to select more appropriate detection mode. On the other hand, with the more and more strict requirements of safety monitoring on railway operation, new CRH train wheel on-line ultrasonic testing system had higher standards on wheel position accuracy. Time sequence processing is required for the sampling data of wheel position, in order to achieve better imaging effect.The modal analysis of Lamb wave was made according to the dispersion properties of its propagation in the thin metal plate by comparing the two time scales of the signal collected in ultrasonic testing. Based on the dispersion curves of Lamb wave plotted by the semi analytical finite element method in the isotropic free aluminum, suitable ultrasonic transducers were selected to obtain modal signal propagation of Lamb wave in aluminum plate in different positions. In the overall time scale of Lamb wave, all the data were analyzed by two-dimensional Fourier transform and the modal and energy distribution would be confirmed. In the other time scale, the Hilbert-Huang transform was used to separate and distinguish the different modes accurately that were superimposed in a single signal under a reasonable filtering and noise reduction. In another part, the structure and mechanism of on-line ultrasonic inspection system for CRH train wheel was introduced. Based on the echo signal obtained from 768 channels of Angle probes and normal probes, a method was put forward to process the massive data using Hall sensors to locate the position of running wheels.The result showed that time-frequency analysis method based on Hilbert Huang transform can be used to identify Lamb wave mode and get higher time resolution, in order to select the appropriate mode for excitation. At the same time, the detection image in the practical application showed that the time sequence processed signal method can provide help for the effective detection of the internal and surface defects in the wheel. The work of this thesis provided a new idea for adaptive processing of stationary and non-stationary signals, and had the significance for the on-line nondestructive testing of large metal plate structural components and railway wheels.