基于LMD-MFE-SVD的松动爆破降噪分析  

Noise Reduction Analysis of Loosening Blasting Vibration Signal based on LMD-MFE-SVD

作　　者：周红敏[1] 赵事成 王慧珍 余辉 李文豪 张宪堂[1,2] ZHOU Hong-min;ZHAO Shi-cheng;WANG Hui-zhen;YU Hui;LI Wen-hao;ZHANG Xian-tang(Shandong Key Laboratory of Disaster Prevention and Mitigation of Civil Engineering,Shandong University of Science and Technology,Qingdao 266590,China;Engineering Research Center of Underground Mine Construction,Ministry of Education,Anhui University of Science and Technology,Huainan 232001,China)

机构地区：[1]山东科技大学山东省土木工程防灾减灾重点实验室,青岛266590 [2]安徽理工大学矿山地下工程教育部工程研究中心,淮南232001

出　　处：《爆破》2023年第4期174-182,共9页Blasting

基　　金：国家自然科学基金项目(51874189);2021年度矿山地下工程教育部工程研究中心开放基金资助项目(JYBGCZX2021102)。

摘　　要：为提高松动爆破振动信号分析精度,在局部均值分解(LMD)的基础上,建立一种基于局部均值分解(LMD)-多尺度模糊熵(MFE)-奇异值滤波(SVD)的混合去噪方法。使用LMD方法对松动爆破振动信号进行分解,获得一系列乘积分量(PF);通过计算MFE和相关系数,对爆破振动信号进行初步降噪;针对主要PF分量的残留噪声,使用SVD滤波进行降噪处理,提取真实信号成分。通过上述处理,最终实现松动爆破信号降噪。结果表明:提出的LMD-MFE-SVD降噪方法具有可行性和应用价值,能够对含噪的PF分量进行有效处理;对于含多信号成分、多噪声的仿真信号,LMD类算法相较EMD类改进算法降噪效率更高,信噪比(SNR)、均根方误差(RMSE)和失真百分比(PRD)指标表现显著提升,而相较LMD算法,提出的LMD-MFE-SVD算法降噪效率进一步提高,依次提升11.73%、22.07%和9.25%,降噪效率显著;根据实测松动爆破振动信号去噪后的波形和频谱对比,提出的LMD-MFE-SVD降噪后的信号波形更为集中,能保留多数信号信息,信号频谱图更为清晰,有效显示信号频率波峰,更利于松动爆破振动信号的特征分析。In order to improve the analysis accuracy of loosening blasting vibration signals,a hybrid denoising method based on local mean decomposition(LMD),multiscale fuzzy entropy(MFE),and singular value filtering(SVD)was established.Firstly,the vibration signal was decomposed by LMD method to obtain a series of product components(PF).Then,the blasting vibration signal was preliminarily denoised by calculating MFE and correlation coefficients.Finally,the real signal components were denoised and extracted by SVD filtering on the residual noise of the main PF components.The results show that the proposed LMD-MFE-SVD denoising method can effectively deal with the noisy PF components.For the simulated signal with multiple components with noise,the LMD algorithm is more efficient than the EMD algorithm.Furthermore,the signal-to-noise ratio(SNR),root mean square error(RMSE)and percentage of distortion(PRD)of the proposed LMD-MFE-SVD method are significantly improved by 11.73%,22.07%and 9.25%,respectively,compared with the LMD algorithm,which indicates that the noise reduction efficiency is considerable.According to the waveform and spectrum comparison of the measured loosening blasting vibration signal after denoising by the proposed LMD-MFE-SVD method,the denoised signal waveform is more concentrated with most of the signal information retained.The frequency spectrum is clearer,and the signal frequency peaks are effectively displayed.

关 键 词：松动爆破 振动降噪 局部均值分解 多尺度模糊熵 奇异值滤波 

分 类 号：U455[建筑科学—桥梁与隧道工程]