基于马氏链样本模拟的渡槽结构系统可靠度分析  

System reliability analysis of aqueduct structure based on Markov chain sample simulation

作　　者：吴剑国[1] 金伟良[2] 张爱晖[2] 王盛[2] 

机构地区：[1]浙江工业大学建筑工程学院,浙江杭州310014 [2]浙江大学建筑工程学院,浙江杭州310027

出　　处：《水利学报》2006年第8期985-990,共6页Journal of Hydraulic Engineering

基　　金：水利部重点创新项目资助(SCX2002-09)

摘　　要：提出了一种适合于渡槽结构系统可靠度计算的重要样本法。该方法根据Metropolis准则构造马尔可夫链模拟样本,通过在失效域中进行预抽样,获得对失效概率的计算贡献较大的重要区域的分布信息,然后利用此信息构造重要样本函数,采用重要样本技术计算结构系统的失效概率。对一典型梁式渡槽进行各种失效模式的分析,建立了相应的功能函数,并就本文算法与已有算法计算得的可靠度进行了比较。结果表明,该方法由于不涉及设计点的概念,解决了应用重要样本法计算系统可靠性时重要样本函数构作的难题,具有较好的效率和精度。An adaptive important sampling methodology is proposed to calculate the structural system reliability of aqueducts. To populate the region of higher probability for a large variety of shapes of the failure region, the preparative sampling technique is processed in which the Markov simulation samples are simulated according to Metropolis criterion. An important sampling function is then constructed from the information of the higher probability region. The typical failure modes and reliability of a typical beam type aqueduct are analyzed. The system reliability of the aqueduct is discussed by using the proposed method and other methods. The comparison indicates that this method is precise, efficient, having good robustness and has the advantage of making an important sampling function because no related concept of design viewpoint is involved. This method can be used to calculate the system reliability of other types of structure.

关 键 词：结构系统可靠度 重要样本法 METROPOLIS准则 马尔可夫链模拟 梁式渡槽 失效模式 

分 类 号：TV672[水利工程—水利水电工程] TV31