A De-Noising Algorithm for EMAT on Wheels

Presently electro-magnetic acoustic technology (EMAT) has been widely adopted by the China Railway High-speed (CRH). In this paper, the principle of wheels tread detection by using EMAT is introduced. To solve the noise problem existing in the detecting procedure, an integrated algorithm combining the wavelet algorithm and the phase difference algorithm is presented. Under the premise of saving the necessary signal wave, the Fast Fourier Transform based on the Blackman window is adopted to calculate the signal phase, and then the integrated algorithm is chosen to process the signals. Compared with the wavelet algorithm or the phase difference algorithm, the rate of the noised data after being processed by the integrated one is reduced by more than 45%. This algorithm will improve the waveform quality and has an enlightening function to raise the detection accuracy in the EMAT system.