基于单通道盲源分离的结构模态参数识别  

Structural modal parameter identification based on single channel blind source separation

作　　者：甄龙信[1] 任良 董前程 ZHEN Longxin;REN Liang;DONG Qiancheng(Hebei Key Laboratory of Special Carrier Equipment,Yanshan University,Qinhuangdao 066004,China)

机构地区：[1]燕山大学河北省特种运载装备重点实验室,河北秦皇岛066004

出　　处：《振动与冲击》2023年第11期252-261,294,共11页Journal of Vibration and Shock

基　　金：国家自然科学基金资助项目(51675462)。

摘　　要：针对基于传统盲源分离算法的结构模态参数识别需要满足传感器数目不少于源信号数目的问题,提出一种基于单通道盲源分离的结构模态参数识别方法,该方法利用单个通道信号即可完成结构模态参数识别。利用同步提取变换(synchro extracting transform,SET)对单通道观测信号进行时频分析以确定变分模态分解(variational mode decomposition,VMD)参数K的取值;将观测信号利用VMD分解形成K个本征模态函数(intrinsic mode function,IMF);将K个IMF进行线性混合形成2维观测信号并与原单通道观测信号重构形成3维观测信号,利用基于信号稀疏性的源信号分离算法分离得到各单模态信号;利用单模态识别技术识别结构模态参数。仿真和实测信号数据表明所提方法的有效性。Aiming at the problem that the number of sensors is not less than that of source signals in structural modal parameter identification based on traditional blind source separation algorithm,a structural modal parameter identification method based on single channel blind source separation was proposed,which could complete the structural modal parameter identification by using a single channel signal.Firstly,the synchronous extracting transform(SET)was used to analyze the time-frequency of the single channel observation signal to determine the value of the variable mode decomposition(VMD)parameters;Secondly,the observed signal was decomposed by VMD to form an intrinsic mode function(IMF);Thirdly,the two IMF were linearly mixed to form a two-dimensional observation signal,and reconstructed with the original single-channel observation signal to form a three-dimensional observation signal.Each single-mode signal was separated by the source signal separation algorithm based on signal sparsity;Finally,the structural modal parameters were identified by single modal identification technology.Simulation and measured signal data verify the effectiveness of the proposed method.

关 键 词：单通道盲源分离 同步提取变换(SET) 变分模态分解(VMD) 信号稀疏性 模态参数识别 

分 类 号：TH17[机械工程—机械制造及自动化]