作　　者： ; ; ;

机构地区： 中国地震局工程力学研究所

出　　处： 《振动与冲击》 2010年第4期59-63,共5页

摘　　要： 从结构动力特征方程出发,以结构主模态参数为观测量,推得结构物理参数线性回归模型。对该模型应用贝叶斯估计理论得到物理参数后验联合分布,再结合马尔可夫蒙特卡罗抽样方法给出各个物理参数的边缘概率分布和最优估计值,而提出了基于结构主模态参数的结构物理参数识别贝叶斯估计马尔可夫蒙特卡罗方法。对五层剪切型结构的数值研究表明,此方法能够利用少数主模态参数给出结构质量和刚度参数的概率分布和最优识别值,而且在主模态参数较准确时识别误差很小。 The linear regression models of physical parameters of a structure were deduced from its dynamic characteristic equations with the modal parameters of its dominant modes as observation values.Based on these models,Bayesian statistics theory was used to obtain the posterior joint probability density function(PDF)of structural physical parameters,then the marginal PDF and optimal estimation of the physical parameters were obtained with the Markov Chain Monte Carlo(MCMC)method.The theoretical analysis and numerical simulation of a 5DOF shear type building showed that the approach presented can not only give the optimal estimation and the associated probability of the physical parameters using modal data of few primary modes but also get higher identification accuracy with relatively exact modal parameters observed.

关 键 词： 物理参数识别 贝叶斯估计 马尔可夫蒙特卡罗方法 模态参数

领　　域： [建筑科学]