Time-spatial Correlation Based Adaptive Wavelet Packet Method of Acoustic Emission (AE) Signal De-noising

Existing wavelet transform method in the strong noise interference AE signal de-noising, due to ignore the correlation between the wavelet coefficients, easily lead to signal distortion. To solve the problem, a time-spatial correlation based adaptive wavelet packet method was proposed to Acoustic Emission (AE) signal de-noising. Adaptive redundant wavelet transform was used to maximize the time-spatial features in the algorithm. The correlation of the wavelet coefficients in time-spatial domain were used to improve the discrimination ability of AE signal and noise, It were improves the effect of noise reduction. Meanwhile, the correlation of wavelet coefficients in spatial domain was used to find the best decomposition level to further improve the adaptability of the algorithm. In order to validate the effectiveness of the proposed algorithm, different levels of noise are added to the simulated AE signal and the disconnected lead signal respectively, and the noise reduction experiment is carried out. Experimental results show that the SNR of the signal processed by this algorithm is greatly improved compared with that of the other algorithms, and the RMSE is also reduced to a certain extent.