【摘要】 影响桥梁地震安全的不仅是地震强度,还包括桥梁结构材料强度、结构尺寸,以及其离散分布特征等,这些因素也不是唯一确定的,而是随机的。传统的桥梁地震易损性分析方法已无法满足要求,需要充分考虑地震、材料等因素的随机性。为了更加全面而充分地确定在多种不确定因素综合影响下,桥梁结构的地震易损性分布及影响规律,采用全概率方法进行桥梁地震易损性分析。针对目前结构地震易损性分析方法的不足,引入人工神经网络和蒙特卡罗技术,采用动量增量分析方法和推倒分析方法,同时考虑了桥梁结构参数的随机性和地震动输入的随机性,提出了桥梁复合随机地震易损性分析方法。主要分析步骤为:首先,分别对随机桥梁结构的抗力和随机地震响应进行特征统计,在此基础上,再应用蒙特卡罗和神经网络技术对桥梁结构在不同损伤状态下的超越概率进行统计,绘制桥梁结构复合地震易损性曲线。复合随机地震易损性分析方法从全概率的角度比较全面地反映了桥梁结构的地震易损性,并具有以下特点:(1)采用神经网络仿真结构随机性的非线性分析,大大提高了计算效率;(2)采用充足地震波进行动量增量分析,从而考虑了地震强度的随机性;(3)采用蒙特卡罗随机抽样分析方法,避免了求解析解的困难。更多还原

【Abstract】 Factors influences the seismic safety of bridges are not only the earthquake intensity, but also material strength, size and discrete distribution characteristics of bridge structure, these factors are not uniquely deterministic but random. Traditional bridge seismic vulnerability analysis method is unable to meet the requirements, which need to give full consideration to the randomness of the factors such as earthquake and material. In order to more comprehensively and fully determine the vulnerability distribution and the influence rule of bridge structure under the comprehensive influence of uncertain factors, aiming at the shortcomings of the current structure seismic vulnerability analysis method, a method of compound stochastic seismic vulnerability analysis based on incremental dynamic analysis and pushover analysis,which adopts ANN and MC technology, is presented considering the randomness of structural parameters and seismic response. The analysis steps are:(1)first, the characteristic statistics of the resistance and the random seismic response of the random bridge structure are carried out;(2)on this basis, MC and ANN technology are applied to carry out statistics of the transcendence probability of the bridge structure under different damage states, and the composite seismic vulnerability curve of the bridge structure is drawn. The method of composite random seismic vulnerability analysis reflects the seismic vulnerability of bridge structure comprehensively from the perspective of full probability, and has the following characteristics:(1) the nonlinear analysis of the structure randomness by neural network simulation greatly improved the calculation efficiency;(2) the momentum increment analysis by sufficient seismic wave considered the randomness of earthquake intensity;(3) MC random sampling analysis method avoided the difficulty of finding analytic solution.更多还原