基于小波包变换的时间序列模型模态参数识别  

Modal parameter identification of the time series model based on wavelet packet transform

作　　者：徐晓霞[1] 任伟新[2] 韩建刚[3] 

机构地区：[1]福州大学土木工程学院,福建福州350002 [2]中南大学土木建筑学院,湖南长沙410075 [3]海南大学土木工程系,海南海口570228

出　　处：《福州大学学报（自然科学版）》2008年第2期266-271,共6页Journal of Fuzhou University(Natural Science Edition)

基　　金：国家自然科学基金资助项目(50678173);教育部新世纪优秀人才支持计划资助项目(NCET-04-0612)

摘　　要：提出了基于小波包变换的时间序列模型结构模态参数识别方法.该方法以线性的离散时间序列方程为基础,对结构的振动响应数据进行小波包变换分解,利用小波包函数的正交特性,建立量测点间的离散化运动方程,最后利用该离散化运动方程的系数矩阵,估算结构的模态参数(自振频率、阻尼比与振型).用数值模拟算例对此方法进行了验证,并与随机子空间识别方法结果进行了比较.结果表明,该方法可以正确地识别出结构的模态参数.A method of modal parameter identification of the time series model by using wavelet packet transform is suggested which is based on linear discrete time series equation. The measured responses of a structure are first decomposed by wavelet packet transform. Appropriate discrete equations of motion for measured degrees of freedom are established from the transformed responses by using the orthonormality of the wavelet package function. Finally, the modal parameter of the structure can be extracted from the co-efficient matrix of the established discrete equations of motion. A numerical study is performed to demonstrate the proposed technique and verify its accuracy. The results obtained are comparable with those previously obtained from the stochastic subspace identification. It is show that the proposed technique is reliable and efficient, which can be used to identify the modal parameter of structure quite effectively.

关 键 词：小波包变换 时间序列 ARMA模型 模态参数识别 

分 类 号：U441.3[建筑科学—桥梁与隧道工程]