Structural health diagnosis method based on a system identification technique of self-learning type for bi-axial nonlinear vibration system

本课题提出了一种基于系统辨识技术的结构健康诊断方法。该结构健康诊断方法根据使用所识别的结构模型估计的建筑物的地震响应来评估建筑物的损坏程度。利用结构健康诊断法对部分建筑物的震害进行了评估。建筑物安装了地震反应观测系统，并对振动反应进行了监测。对利用结构对弱风反应进行结构健康监测的可行性进行了研究，提出了一种基于地震反应记录的结构非线性层间剪切变形模型参数估计方法，并在地震后自动更新参数。非线性楼层-变形关系考虑了双向变形。基于随机理论的结构健康诊断方法根据模型参数的估计误差来评估损伤程度估计的可靠性，利用该方法进行结构健康诊断的结果表明了损伤检测的有效性。结构在弱风作用下的响应数据对结构健康诊断具有重要意义。

This project proposed a structural health diagnosis method using a system identification technique. The structural health diagnosis method evaluates the damage level of a building from the earthquake responses of a building estimated using the identified structural model. The earthquake damages of some buildings were estimated by the structural health diagnosis method. The buildings are installed the observation system of earthquake response and monitored the response of vibration. The feasibility study of the structural health monitoring using the structural response to weak wind was performed.The proposed identification technique estimates the parameters of nonlinear story shear-deformation models for a building from earthquake response records and updates automatically the parameters after earthquakes. The nonlinear story-deformation relationship takes bi-axial deformation into account. The structural health diagnosis method evaluates the reliability of the estimated damage level from the estimate error of the model parameters based on stochastic theory.The results of the structural health diagnosis using the proposed method illustrate the affectivity in damage detection. The data of the response to weak wind is useful in the structural health diagnosis.