项目名称: 密肋复合墙结构基于贝叶斯理论的恢复力模型参数识别及动力可靠度分析
项目编号: No.51208030
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 建筑环境与结构工程学科
项目作者: 刘佩
作者单位: 北京交通大学
项目金额: 25万元
中文摘要: 模型参数识别问题需要考虑模型本身及模型参数的不确定性。根据贝叶斯理论进行模型参数识别,不仅可以得到模型参数的最有可能值,还能得到定量描述的模型参数不确定性。本项目首先拟探讨贝叶斯理论在采用实测数据时在非线性模型参数识别中的应用,其中拟建立密肋复合墙的恢复力模型,推导贝叶斯理论用于恢复力模型参数识别的计算框架,依据密肋复合墙低周反复加载试验的滞回曲线数据,识别得到模型参数的最有可能值和协方差矩阵;其次本项目拟提出考虑模型不确定性的动力可靠度计算新方法,使模型误差对可靠度计算结果的影响定量化,并且结合恢复力模型参数的识别结果及相似理论,根据可靠度计算结果确定密肋复合墙结构的弹塑性层间位移角限值;另外本项目拟对一幢已建六层密肋复合墙结构进行环境振动测试,应用贝叶斯理论识别其自振频率、振型、阻尼等参数,根据识别结果分析实际密肋复合墙结构的动力特性,并对其可靠度进行评估。
中文关键词: 密肋复合墙;恢复力模型;参数识别;模型选择;环境振动
英文摘要: Uncertainty of the model and model parameters should be considered for the model parameter identification problems. According to Bayesian theorem, not only the most probable values, but also uncertainty of the model parameters can be obtained from the model parameter identification. In this project, firstly application of Bayesian theory to the identification of nonlinear model parameters with observed data will be made. Restoring force model for multi-grid composite walls will be proposed. Computation frame for identification of restoring force model parameters based on Bayesian theorem will be derived. According to the hysteretic curve data obtained from the cyclic loading tests of multi-grid composite walls, the most probable values and covariance matrix of the model parameters will be obtained. Secondly dynamic reliability calculation method will be proposed considering model uncertainty. Effect of model prediction errors on the reliability results will be determined. With the identification results of the restoring force model parameters and theory of similarity, limit of elastic-plastic drift ratios of multi-grid composite wall structures will be determined according to the reliability results. In addition, ambient vibration test will be done in an existing six-story multi-grid composite wall structure. Na
英文关键词: multi-grid composite wall;restoring force model;parameter identification;model selection;ambient vibration