论文部分内容阅读
考虑系统参数的随机性,将基于广义卡尔曼滤波的子结构法与贝叶斯更新方法相结合,提出了桥梁结构基于贝叶斯更新物理参数的剩余强度估计两步法:第一步,将子结构法与广义卡尔曼滤波算法相结合,成功识别出子结构及其相邻单元的物理参数;第二步,视识别出的结构物理参数为更新信息,对以蒙特卡罗仿真实验结果作为先验分布的参数进行贝叶斯更新并分别基于蒙特卡罗仿真参数和贝叶斯更新物理参数对结构进行了剩余强度估计。数值算例表明:基于贝叶斯更新物理参数估计得到的结构剩余强度明显低于基于蒙特卡罗仿真参数估计得到的结构剩余强度。该方法为测量响应信息不完备条件以及小样本抽样情况下桥梁结构剩余强度估计提供了一个较好的解决思路。
Considering the randomness of the system parameters, combining the substructure method based on generalized Kalman filter and the Bayesian updating method, a two-step method of estimating the residual strength of the bridge structure based on Bayesian update physical parameters is proposed. In the first step, The substructure method and the generalized Kalman filter algorithm are combined to successfully identify the physical parameters of the substructure and its adjacent units. In the second step, the structural physical parameters identified are the updated information. Based on Monte Carlo simulation results as The parameters of prior distribution are updated by Bayesian method and residual intensity is estimated based on Monte Carlo simulation parameters and Bayesian update physical parameters respectively. Numerical examples show that the structure residual strength estimated based on the Bayesian update physical parameters is obviously lower than that based on Monte Carlo simulation parameters. This method provides a better solution for measuring the incomplete information of response information and estimating the residual strength of bridge structure in the case of small sample size.