一种结构参数识别的两阶段方法

针对测量信息不完备的剪切型结构 ,建立了一种两阶段系统识别的复合反演方法 ,这种方法包括两部分 :子结构地震动反演和结构参数识别。首先 ,选取可观测的子结构 ,利用一维地震动作用于结构的力学特性 ,将子结构动力方程的有限元列式进行变换 ,得到适合于最小二乘法的简单形式 ,解决了测量信息不完备及结构参数未知条件下的地震动反演问题。其次 ,根据子结构反演得到的地震动输入 ,采用结构参数时域识别技术中的加权整体迭代 -广义卡尔曼滤波器方法 ,成功地识别出了有限测量条件下单元水平结构参数

A two-stage method for identification of structural parameter

Most of the parameter identification techniques in time domain require information of ground motion and structure response at every degree of freedom, and this contradicts the usual practice of having a small number of sensors located on the structure. A method to obtain the time history of ground motion and to estimate the structural parameter down to the element level with incomplete measurements is proposed. The method consists of two stages: the sub|structural inversion of ground motion, and the parameter identification. By using of the mechanical characteristic of one dimension earthquake, the finite element formulation is transformed into the presentation suited to the least square estimation. An algorithm of the sub|structure solves the inversion problem of ground motion with the unknown structural parameters and limited measurements. The structural parameters are then identified in the time domain by extended Kalman filter|weighted global iteration (EKF|WGI) algorithm, which has good robustness and error tolerance. Several cases of a 6 story frame structure are numerically investigated. The results show that the proposed two stage approach can effectively identify the structural parameters from incomplete measurements.