Study On The Damage Monitoring Based On The Denoising Of Acoustic Emission Signal

Acoustic emission (AE) detection is a new member of nondestructive examination. Due to various reasons, AE detection cannot be applied well in practical engineering at first. But with the rapid development of hardware and software and some of its unique advantages, the AE detection has been widely used in many kinds of fields.The health monitoring based on the AE technique is the research objective in this study, including denoising algorithms and monitoring algorithm. The AE detection and characteristic parameters, including some kinds of parametric analysis and waveform analysis are sum up. We also elaborate the fast Fourier transform, namely Mallat algorithm. These serve as the theoretical underpinning for the following researches. By comparing the energy distribution and the simulation time of different wavelet coefficients, we discuss the impact of the vanishing moments on Mallat algorithm. The vanishing moments need to be chosen suitably. And Mallat algorithm uses periodic extension to analysis finite length signal. At the same time, the merit and demerit of Donoho, K-means clustering, and spatially selective noise filtration (SSNF) are all studied in details.We propose an improved SSNF base on the obvious frequency division property of wavelet transform, the thought of image compression, and the normalization of the correlation coefficient. And we also propose a damage monitoring algorithm by using the b-value method. Compared with the original SSNF, the improved SSNF can avoid "glitches" and realize real-time denoising of AE signal. And then we analyze the affection of b-value curves by using different denoising algorithms. We determine to use the improved SSNF in the damage monitoring algorithm. The monitoring algorithm meets requirements of AE health monitoring. It can both eliminate the interference of noise and realize real-time monitoring.