基于信息更新的在役桥梁结构模糊随机可靠度研究

As researches about reliability assessment of inservice bridges are impractical, the bayes method based on uncertain model is proposed to identify and update the structure parameters timely Combining with the Markov Monte Carlo sampling.The concept of Fuzzy-Random variable is introduced and defined. Based on the kriging model, the mean maximum improved algorithm will be used to ensure the efficiency of approximation, and the adaptive Monte Carlo interval optimization algorithm will be also uesd to improve the calculation efficiency. By means of adaptive iterative strategy, the fuzzy anlysis will nested in the random anlysis prosess, thereafter, the unified anlysis method mixed by fuzzy and random could be obtained, which based on Saddle Point approximation algorithm. Taking Jiujiang cable-stayed bridge as the engineering background, the bridge intelligent monitoring system based on the advanced Imote2-based wireless sensor networks will be used to collect the observation information for the fuzzy random reliability analysis of the cable-stayed bridge in Jiujiang. Based this, an effective methodology used for reliability assessment of inservice bridge could be established.

针对当前在役桥梁可靠度计算往往不能符合实际的问题，本项目采用基于不确定性模型的贝叶斯（Bayes）方法，并结合马尔可夫蒙特卡罗抽样，对桥梁的结构参数特性进行实时识别和信息更新；引入模糊随机变量的概念，采用Kriging模型，提出平均最大改进算法以保证响应模拟效率，同时采用自适应蒙特卡罗区间优化算法，提高区间计算效率；通过自适应迭代策略，将模糊分析嵌套于随机分析，提出基于鞍点模拟的模糊性和随机性统一分析法。以九江斜拉桥为工程背景，利用已成功应用于九江斜拉桥状态监测的基于Imote2无线传感器技术的桥梁智能监测系统得到观测信息，完成对九江斜拉桥的模糊随机可靠度分析，继而形成一套适用于在役桥梁安全可靠度计算评估的有效理论方法。

本项目进一步完善了基于Imote2高性能无线传感器技术的桥梁状态监测系统。研究了基于贝叶斯估计-马尔可夫蒙特卡罗的参数识别方法。利用贝叶斯更新思想对基于Kriging响应面的结构可靠度分析方法进行改进，同时采用自适应蒙特卡罗区间优化算法，提高区间计算效率；通过自适应迭代策略，将模糊分析嵌套于随机分析，提出基于鞍点模拟的模糊性和随机性统一分析法。以九江斜拉桥为工程背景，完成对九江斜拉桥的模糊随机可靠度分析，继而形成一套适用于在役桥梁安全可靠度计算评估的有效理论方法。