基于广义Morse同步挤压小波和子结构法的结构时变参数识别

Engineering structures under long-term opration are essenially time-varying and nonlinear systems, which indicates thta their responses are non-stationary and their characteristic parameters are time-varying. The project aims to process the responses of time-varying and nonlinear structures under ambient vibration by employing continuous generalized Morse synchrosqueezing wavelet and discrete wavelet multi-resolution analysis substructuring approach so that the instantaneous featured parameters (i.e., instantaneous frequencies, stiffness and damping etc.) can be estimated effectively. This project is expected to propose new theories, algorithms and realize with the aid of computers through theoretical, numerical and experimental studies. In detail, this project will be focused on developing the improved synchrosqueezing transform algorithm, signal-noise separation technique using fast iterative shrinkage-thresholding algorithm (ISTA) in time-frequency domain, the system identification algorithm based on wavelet multi-resolution analysis and substructuring approach as well as determination of the layers for wavelet decomposition based on Akaiks's information criterion (AIC). The achievements of this project are believed to be useful during the research of disaster mechanism, health monitoring and safety evaluation of important civil structure, whose theoretical significance and engineering practical value are sound.

处于长期工作状态下的土木工程结构，本质上属于时变和非线性结构系统，结构响应表现出非平稳特性，其典型特征就是结构参数是时变的。项目以时变和非线性结构的环境振动响应处理为出发点，采用连续广义Morse同步挤压小波变换和离散小波多分辨率分析子结构法，以识别结构的瞬时特征参数（瞬时频率、瞬时刚度和瞬时阻尼系数等）为突破口，从理论、数值模拟和试验三个方面，对结构时变参数识别理论、算法和计算机实现进行深入研究，重点研究改进同步挤压变换技术、小波时频域信噪分离、基于小波多分辨率分析和子结构法的识别算法、基于AIC信息准则的小波变换分解层数确定方法等关键技术问题；项目成果可广泛应用于重大土木结构灾变机理、健康监测和安全评估的研究，具有较大的理论意义和工程实用价值。

处于长期工作状态下的土木工程结构，本质上属于时变和非线性结构系统，结构响应表现出非平稳特性，其典型特征就是结构参数是时变的。项目以时变和非线性结构的环境振动响应处理为出发点，采用连续广义Morse同步挤压小波变换和离散小波多分辨率分析子结构法，以识别结构的瞬时特征参数（瞬时频率、瞬时刚度和瞬时阻尼系数等）为突破口，从理论、数值模拟和试验三个方面，对结构时变参数识别理论、算法和计算机实现进行深入研究。研究提出了基于同步挤压小波变换的结构时变模态参数识别技术；基于小波多分辨率分析和子结构法的时变结构物理参数识别技术；小波函数的选取方法；基于AIC准则的小波分解层数的确定技术；基于GCV和改进阈值函数的小波去噪技术。另外也初步研究了基于经验小波变换和广义Morse小波的结构瞬时频率识别技术。项目成果可应用于重大土木结构灾变机理、健康监测和安全评估的研究，具有较大的理论意义和工程实用价值。