联合CEEMD和改进阈值去噪的初至波识别算法  

Combined CEEMD and Improved Threshold Denoising Algorithm for First Break Recognition

作　　者：艾文强 范锦彪[1] 王燕[1] AI Wenqiang;FAN Jinbiao;WANG Yan(Instrument Science and Dynamic Test Laboratory, North University of China, Taiyuan 030051, China)

机构地区：[1]仪器科学与动态测试教育部重点实验室,太原030051

出　　处：《兵器装备工程学报》2019年第3期137-142,共6页Journal of Ordnance Equipment Engineering

摘　　要：提出了一种新的阈值函数并联合互补集合经验模态分解算法,用改进能量算法对初至波进行提取。将信号分解为几个本征模态函数和一个残余信号,计算每个本征模态函数分量的自相关函数及频域的主频,并依据自相关函数特点和主频分布情况判断噪声信号主导的本征模态函数,对这些分量用新阈值函数去噪,将信号重构使用能量比值法提取初至点。仿真信号计算结果及实际实验结果表明,该方法能有效提取出低信噪比信号的初至波,高精度拾取初至时间。A new threshold function combined with complementary set empirical mode decomposition algorithm was proposed. The first arrival wave was extracted by improved energy algorithm. In this algorithm, the signal was decomposed into several eigenmode functions (Intrinsic Mode Function, IMF) and a residual signal by CEEMD, and the autocorrelation function of each IMF component and the main frequency of the frequency domain were calculated. According to the characteristics of the autocorrelation function and the distribution of the main frequency, the IMF of the dominant modal of the noise signal was judged, and these components were analyzed. The improved threshold was denoise, and finally the signal reconstruction was used to extract the initial to the point by using the energy ratio method. The simulation results and the actual experimental results show that the method can effectively extract the first break of low signal- to-noise ratio signal and achieve high-precision pickup of the first break time.

关 键 词：阈值函数 联合互补集合经验模态分解 阈值去噪 能量比值法 初至时间 

分 类 号：TP30[自动化与计算机技术—计算机系统结构]