基于CEEMD和改进小波阈值的机械密封声发射信号降噪方法

A Noise Reduction Method for Mechanical Seal AE Signal Based on CEMD and Improving Wavelet Threshold

摘要:

声发射作为一种无损检测结束被广泛应用于多个领域，针对声发射信号难以从背景噪声种分离的问题，提出了一种基于互补集合经验模分解（CEEMD）与改进小波阈值相结合的降噪方法。首先对声发射信号进行CEEMD分解，通过峰宽占比确定信噪分量分界点，对噪声分量进行改进的小波阈值降噪，将降噪后的分量与其余分量进行信号重构得到最终降噪结果。通过仿真信号和机械密封声发射实验信号论证了本文方法相较于传统小波降噪和CEEMD强制降噪更有效。

Abstract:

As a nondestructive testing end, acoustic emission is widely used in many fields. Aiming at the problem that acoustic emission signals are difficult to separate from background noise, a noise reduction method based on Complementary Ensemble Empirical Mode Decomposition (CEEMD) and improved wavelet threshold is proposed. Firstly, the acoustic emission signal is decomposed by CEEMD. The demarcation point of signal-noise component is determined by the ratio of peak width. The noise component is denoised by improved wavelet threshold. The denoised component is reconstructed with the rest of the component to get the final denoising result. The simulation signal and the mechanical seal AE experimental signal demonstrate that the proposed method is more effective than the traditional wavelet denoising and CEMD forced denoising.