宋帅, 钱永久, 吴刚. 桥梁系统地震易损性分析的混合Copula函数方法[J]. 工程力学, 2017, 34(1): 219-227. DOI: 10.6052/j.issn.1000-4750.2015.12.1030
引用本文: 宋帅, 钱永久, 吴刚. 桥梁系统地震易损性分析的混合Copula函数方法[J]. 工程力学, 2017, 34(1): 219-227. DOI: 10.6052/j.issn.1000-4750.2015.12.1030
SONG Shuai, QIAN Yong-jiu, WU Gang. MIXED COPULA FUNCTION METHOD FOR SEISMIC FRAGILITY ANALYSIS OF BRIDGE SYSTEM[J]. Engineering Mechanics, 2017, 34(1): 219-227. DOI: 10.6052/j.issn.1000-4750.2015.12.1030
Citation: SONG Shuai, QIAN Yong-jiu, WU Gang. MIXED COPULA FUNCTION METHOD FOR SEISMIC FRAGILITY ANALYSIS OF BRIDGE SYSTEM[J]. Engineering Mechanics, 2017, 34(1): 219-227. DOI: 10.6052/j.issn.1000-4750.2015.12.1030

桥梁系统地震易损性分析的混合Copula函数方法

MIXED COPULA FUNCTION METHOD FOR SEISMIC FRAGILITY ANALYSIS OF BRIDGE SYSTEM

  • 摘要: 为了在桥梁系统易损性分析中考虑构件地震需求之间相关性的影响,采用贝叶斯加权平均方法构造混合Copula函数,将构件地震需求之间的相关结构和各构件的边缘分布函数进行分离;结合增量动力分析,建立了桥墩、支座等单个构件的易损性曲线及联合分布函数,提出了考虑构件地震需求相关性的桥梁系统易损性分析方法。结果表明:混合Copula函数能够准确描述构件地震需求间上、下尾相关并存的非对称相关结构,简化构件地震需求联合分布函数的建模过程;与Monte Carlo抽样方法相比,二者得到的桥梁系统易损性吻合良好,且混合Copula函数方法避免了大量的数值抽样,显著提高系统易损性分析的计算效率。

     

    Abstract: In order to consider the impacts of seismic demand dependence among structural components in the seismic fragility analysis of bridge system, a mixed Copula function is constructed by using Bayesian Model Average method, and then the dependence structure is separated from the marginal distribution functions of components. In combination with incremental dynamic analysis, the fragility curves and joint distribution functions of piers and bearings are developed. And a method for calculating the fragility of bridge system is proposed, which takes the seismic demand dependence among structural components into consideration. The results indicate that mixed Copula function can accurately depict the asymmetric dependence between the upper and lower tail distributions of the seismic demands of bridge components and simplify the modeling procedure of joint probability distribution function. A comparison between the system fragility curves obtained by the proposed method and those derived using Monte Carlo method shows good agreement. In addition, the computational efficiency is significantly improved by the mixed Copula function method because it significantly reduces the number of numerical samples.

     

/

返回文章
返回