基于改进CEEMDAN方法确定微差爆破实际延期时间  

Identification of the real delay time of millisecond blasting based on an improved CEEMDAN method

作　　者：张亮[1,2] 凌同华 陈增辉[3] 张胜 余彬[1] ZHANG Liang;LING Tonghua;CHEN Zenghui;ZHANG Sheng;YU Bin(School of Civil Engineering,Changsha University of Science&Technology,Changsha 410114,China;College of Civil Engineering,Hunan City University,Yiyang 413000,China;Hangzhou Transportation Investment Construction Management Group Co.,Ltd.,Hangzhou 310012,China)

机构地区：[1]长沙理工大学土木工程学院,长沙410114 [2]湖南城市学院土木工程学院,湖南益阳413000 [3]杭州交通投资建设管理集团有限公司,杭州310012

出　　处：《振动与冲击》2020年第20期274-280,共7页Journal of Vibration and Shock

基　　金：国家自然科学基金(51678071,51608183,51878074);湖南省研究生科研创新(CX2018B530,CX2018B531)。

摘　　要：针对微差爆破实际延期间隔识别中容易出现的模态混叠和伪分量问题,提出了一种CEEMDAN算法结合排列熵和奇异值分解理论联合识别微差爆破延期时间的方法。以某微风化花岗岩场地实测爆破振动信号为例,分别采用EMD算法与CEEMDAN算法对仿真微差爆破信号和实测微差爆破信号实际延期间隔进行识别,在CEEMDAN算法识别的基础上,用排列熵定量检测保留爆破震动特征的有效IMF分量,对熵值大于0.5的IMF分量利用奇异值分解作处理,并采用Hilbert变换绘制相应IMF分量的包络谱。结果表明:CEEMDAN方法能识别微差爆破延期时间,且能有效克服模态混叠现象。经CEEMDAN法处理的数据依旧存在虚假分量的问题,通过排列熵检测和奇异值分解降噪处理有效地压制了噪声,使有效IMF分量的包络谱变得更光滑,从而进一步降低干扰,实现微差爆破延期时间的准确识别。In view of the problems of modal aliasing and pseudo-component which are easy to occur in the real delay interval identification of millisecond blasting,a method that combined the CEEMDAN algorithm with permutation entropy and singular value decomposition theory was proposed.Taking the measured blasting vibration signal of a slightly weathered granite site as an example,the EMD and the CEMDAN algorithms were applied in the identification of real delay interval in the simulated millisecond blasting signal and the measured millisecond blasting signal,respectively.Then on the basis of CEEMDAN algorithm recognition,the effective IMF component retaining the blasting feature was quantitatively detected by the permutation entropy,the IMF component with entropy greater than 0.5 was processed by singular value decomposition,and the envelope spectrum of the corresponding IMF component was drawn by the Hilbert transform.The results show that the CEEMDAN method can identify the real time of millisecond blasting and can effectively overcome the phenomenon of mode aliasing.The data processed by CEEMDAN still has the problem of false components,the entropy detection and singular value decomposition were used to suppress the noise and pseudo components,making the envelope spectrum of the effective IMF component smoother,thereby further reducing the interference and realizing accurate identification of the real delay time of millisecond blasting.

关 键 词：EMD CEEMDAN 排列熵 奇异值分解 微差延期间隔 

分 类 号：TD235[矿业工程—矿井建设]