混凝土坝结构动力响应信息挖掘与安全性态评价

At present, the safety behavior of in-service concrete dam is mainly monitored by observation data under static load condition, less more by dynamical response information. In nature, the dynamical response information has higher sensibility to structure damage evolution and its overall performance. Moreover the dynamical response information is usually captured under ambient excitation, such as earthquake, water flow and operation of hydropower unit due to the large volume of concrete dam generally. The structural health monitoring methods based on “only-output” dynamical response data for concrete dam is needed due to the input signal is difficult to obtain under ambient excitation. In the project, the change of the dynamical response information in damage evolution of concrete dam will be discussed based on dynamical response record decomposed and reconstructed through the dynamical response information mining combined with model test and numerical simulation analysis. Then the relationship between the characters of dynamical response information and the structural safety status of the concrete dam will be established. And the effects of the environment loads such as water level and temperature on the dynamical response information of concrete dam will be given. The characterization methods based on dynamical response modal and non-modal parameters for concrete dam structural safety status will be obtained. Finally, the structural safety monitoring model and diagnosis theory for concrete dam will be developed using its dynamical response information.

在役混凝土坝运行状态，目前主要依据静力状态监测信息进行监控，应用动力响应监测信息监控混凝土坝运行状态方面研究较少。动力响应信息对结构的损伤演化和整体性能的变化更具敏感性。鉴于混凝土坝一般体积较大，应用地震、泄流和机组运行等环境激励产生的动力响应信息是自然选择。但环境激励的输入信息难以获取，需要研究仅依据输出响应信息诊断混凝土坝安全性态的理论和方法。项目基于实测动力响应信息，通过信息分解、重构和挖掘，结合模型试验和数值仿真分析，研究混凝土结构损伤演变过程的动力响应信息变化，建立动力响应信息特征与混凝土坝结构安全性态的关系，揭示水位和温度等环境荷载对混凝土坝动力响应信息的影响规律，提出混凝土坝安全性态的动力响应非模态参数表征方法，建立混凝土坝安全性态动力响应信息监控模型，提出基于动力响应信息的混凝土坝安全性态评价方法。

在役混凝土坝运行状态，目前主要依据静力状态监测信息进行监控，应用动力响应监测信息监控混凝土坝运行状态方面研究较少。动力响应信息对结构的损伤演化和整体性能的变化更 具敏感性。鉴于混凝土坝一般体积较大，应用地震、泄流和机组运行等环境激励产生的动力响应信息是自然选择。但环境激励的输入信息难以获取，需要研究仅依据输出响应信息诊断混凝土坝安全性态的理论和方法。项目提出采用迭代Hankel矩阵特征值分解法，分解动力响应信息的环境激励优势频率影响，压制环境激励优势频率对结构固有振动响应的影响，结合非线性去噪技术，重构了环境激励下混凝土坝的动力响应信息；通过模型试验和数值仿真分析，研究混凝土坝损伤演变过程的动力响应信息特征变化，提出了仅输出型模态参数盲源辨识方法和基于Hankel矩阵联合近似对角化的模态识别方法，采用非线性映射和流形学习方法，建立了动力响应信息特征参数与混凝土结构安全状态的关系，揭示水位和温度等环境荷载对混凝土坝结构动力响应信息的影响规律，提出了混凝土坝结构安全状态的动力响应时间序列模型及其参数表征方法，基于状态空间理论，建立混凝土坝结构状态动力响应信息监控模型；研究了动力响应信息的波动理论分析方法，分析了波走时和波数对混凝土坝损伤的敏感性，提出混凝土结构健康诊断的动力响应波动参数指标；形成了基于动力响应信息的混凝土坝安全状态评价理论方法体系。这对充分挖掘混凝土坝整体安全状态信息，准确评估混凝土坝的安全度和承载能力，具有重要的理论意义和工程实际应用价值。