This paper presents a method to estimate the maximum inter-story drift and relative displacement time history of multi-degree-of-freedom (MDOF) shear structures underground-motion excitation by using only one accelerometer installed on a floor when the structures are damaged during the event and hence modal parameters change and become unavailable. First, the absolute acceleration and relative displacement are formulated in modal coordinates and a state-space expression is derived. Then, a scheme to determine the unknown natural frequencies of the damaged structure is devised with the help of the genetic algorithm (GA) and a reasonably chosen fitness function, and the mode shapes are updated by solving a set of nonlinear equations with respect to each group of frequency variables in the GA. The applicability of this method was studied in numerical simulations that examined the effect of the extent of damage to the structure and the selection of the lower bound of the GA variables. Further simulations were conducted to investigate the modal truncation error and robustness against measurement noise of the proposed approach. Finally, the method was validated in a shake table experiment. The results indicate that the time history of the relative displacement and maximum inter-story drift can be accurately estimated in the case of a significant reduction in structural stiffness and a large search range of GA variables. The results also show the method is robust against noise and performs well even when only the lower modes are included in the model.

本文提出了一种方法来估计多自由度 (MDOF) 剪切结构地下运动激发的最大层间漂移和相对位移时程，当结构损坏时，仅使用一个安装在地板上的加速度计。事件和模态参数发生变化并变得不可用。首先，绝对加速度和相对位移在模态坐标中被公式化，并导出状态空间表达式。然后，在遗传算法的帮助下，设计了一种确定受损结构未知固有频率的方案。(GA) 和合理选择的适应度函数，通过求解一组关于 GA 中每组频率变量的非线性方程来更新模式形状。在数值模拟中研究了该方法的适用性，该数值模拟检查了结构损坏程度和 GA 变量下限选择的影响。进行了进一步的模拟以研究所提出的方法的模态截断误差和对测量噪声的鲁棒性。最后，该方法在振动台实验中得到了验证。结果表明，在显着减少的情况下，可以准确估计相对位移和最大层间漂移的时间历程。结构刚度和大搜索范围的 GA 变量。结果还表明，即使模型中只包含较低的模式，该方法对噪声也具有鲁棒性，并且表现良好。