基于小波包分数幂阈值的声发射去噪研究  

Research of De-noising of Acoustic Emission Signal Based on the Fractional Power Model of Thresholding by Wavelet Packet

作　　者：康玉梅[1] 陈依珂 胡淑婷[1] 周家俊[1] 

机构地区：[1]东北大学资源与土木工程学院,辽宁沈阳110819

出　　处：《有色矿冶》2016年第5期16-20,共5页Non-Ferrous Mining and Metallurgy

基　　金：国家自然科学基金(No.51204029);辽宁省自然科学基金(2014020039);辽宁省科学事业公益研究基金项目(2015003001);教育部留学回国人员科研启动基金资助项目(50-2)

摘　　要：针对岩石声发射信号在传播和采集的过程中受到多种噪声干扰而具有复杂、瞬态等非平稳特性的特点,基于小波包硬阈值去噪法、小波包软阈值去噪法和小波包分数幂阈值去噪法,对岩石声发射信号进行去噪处理。研究表明,对于岩石声发射信号,小波包分数幂阈值去噪方法克服了传统的软阈值去噪使信号幅值失真和硬阈值去噪使去噪后信号振荡加剧的缺点。与软阈值去噪和硬阈值去噪相比,信号的信噪比和均方根误差都较小。小波包分数幂阈值去噪在保证原始信号特征的情况下,可以有效地滤除信号中的噪声,信号失真小,是一种较好的岩石声发射信号去噪方法。The rock acoustic emission signal is subjected to various kinds of noise interference in the process of propagation and acquisition,which has the characteristics of complicated transient non-stability.Aiming at the characteristics,the method of the fractional power model de-noising by wavelet packet decomposition is used to deal with the acoustic emission signal.The fractional power model threshold denoising method overcomes the shortcomings of the traditional soft threshold de-noising method,which makes the signal amplitude distortion and the signal oscillation increase by hard threshold de-noising.Compared with the soft threshold and hard threshold de-noising,the signal to noise ratio and the root mean square error are small.De-noising results show that the fractional power model threshold under the condition of ensuring the original signal characteristic can effectively remove signal noise,make small signal distortion and is a better rock acoustic emission signal de-noising method.

关 键 词：小波包分解 分数幂 阈值 声发射 

分 类 号：TU45[建筑科学—岩土工程]