Research on a LMS adaptive filtering algorithm for acoustic emission signal processing

Acoustic emission technology, one of non-destruction testing methods, occupies a significant position in failure diagnosis field. And one key issue to signal processing in acoustic emission is how to denoise the signals collected in acoustic emission system. However, with traditional wavelet methods applied, the acoustic emission signals overlapping spectrum with original ones raise serious problems to optimum filtering. According to this problem, a Least Mean Square (LMS) adaptive filter model based on wavelet analysis is established, combining LMS adaptive filtering thesis and wavelet analysis together. This module realizes noise reduction or elimination to real-time monitoring signal, increases Signal-to-Noise Ratio (SNR) obviously, and achieves optimum result of signal-noise separation. Series of application experiments of acoustic emission signal processing on rock-like materials indicate the effectiveness of this denoising method.