基于CEEMD和相关性函数特性的爆破震动信号滤波方法研究  

Research of filtering method for blasting vibration signals based on CEEMD and correlation function property

作　　者：张亮[1] 孙新建[1,2] 占其兵 

机构地区：[1]青海大学,青海西宁810016 [2]清华大学水沙科学与水利水电工程国家重点实验室,北京100084

出　　处：《水利水电技术》2017年第6期37-42,共6页Water Resources and Hydropower Engineering

基　　金：国家重点基础研究发展计划(973)项目(2015CB057904)

摘　　要：应用互补集合经验模态分解(Complementary Ensemble Empirical Mode Decomposition,CEEMD)方法对爆破震动信号进行分解,得到一组IMF分量,对各个分量进行相关性函数特性分析可知,高频噪声在0时刻有最大值,其他时刻取值约为0;低频噪声具有周期函数的自相关函数特性,是岩石爆破震动信号所固有的。据此去除噪声分量,使去噪后的信号更能真实地体现岩石爆破震动特性。以黄河上游某水电站引水发电隧洞实测爆破震动信号为例,将基于CEEMD算法与基于EEMD、EMD算法的信号去噪效果进行对比分析,结果表明:采用EMD、EEMD、CEEMD三种分解方法和相关性函数特性滤波后的信噪比分别为:39.48、40.63、49.07,采用CEEMD算法滤波的信噪比SNR与EMD、EEMD算法相比提高了24.31%和20.78%,从而验证了基于CEEMD和相关性函数特性的滤波方法的实用性与优越性。The principle and process of the signal fihering with the approach of complementary ensemble empirical mode decom- position （CEEMD） are studied herein, and then the blasting vibration signal is decomposed with the CEEMD algorithm, from which a set of IMF components is obtained and the correlation function properties of all the IMF components are analyzed as well. The high frequency noise has the maximum value at the zero time and the other time values are about 0, while the low frequency noise has the autocorrelation property of the periodic function which is inherent in rock blasting vibration signal. Removing the noise component can make the de-noised signal more truly embody the characteristics of rock blasting vibration. By taking the measured blasting vibration signal from the construction of the diversion tunnel of a hydropower station on the Upper Yellow River as the study case, the CEEMD algorithm-based de-noising effect and the effects from the de-noising based on EEMD and EMD al- gorithms are comparatively analyzed herein. The result shows that the signal-noise ratios （SNR） after filtering with decomposition methods of EMD, EEMD, CEEMD and correlation function property are 39.48, 4O. 63 and 49. 07 respectively, while the signal-noise ratio （SNR） from filtering with CEEMD algorithm is increased by 24. 31% and 20. 78%, if compared with those from EMD and EEMD algorithms, thus the practicability and superiority of the filtering method based on CEEMD and correlation function property are proved.

关 键 词：CEEMD 自相关函数 互相关函数 爆破震动 滤波 消噪 

分 类 号：TV312[水利工程—水工结构工程]