基于概率充分因子的高效土质边坡可靠度优化设计  

Efficient reliability-based design optimization in soil slope using probability sufficiency factor

作　　者：辛立光 李典庆[1] 曹子君[1] 

机构地区：[1]武汉大学水资源与水电工程科学国家重点实验室,湖北武汉430072

出　　处：《武汉大学学报（工学版）》2016年第5期696-700,共5页Engineering Journal of Wuhan University

基　　金：国家杰出青年基金项目(编号:51225903);国家自然科学基金项目(编号:51409196;51579190;51528901);国家重点基础研究发展计划(973计划)项目(编号:2011CB013502)

摘　　要：提出了基于概率充分因子的高效土质边坡可靠度优化设计方法.首先,采用拉丁超立方抽样法抽取设计空间中的设计样本;其次,利用子集模拟确定设计样本的概率充分因子,建立概率充分因子与设计变量(如边坡倾角)的设计响应面;最后,通过优化设计确定既满足目标可靠度水平又满足经济性要求的设计方案.通过一个土质边坡算例说明了算法的有效性.结果表明,概率充分因子可以有效地表征设计是否满足可靠度要求、建立较为准确的设计响应面,设计响应面解耦了可靠度优化设计中双循环过程,有效地提高了可靠度优化设计效率.This paper proposes an efficient reliability-based design optimization（RBDO） approach using probability sufficiency factor Psf in soil slope design. First, Latin hypercube sampling technique is adopted to extract the design samplings from the design space. Second, probability sufficiency factors are obtained from the subset simulation for estabalishing design response surface （DRS） linking probability sufficiency factors to design variables, e.g. slope angle. Then, the final design is determined as the one that satisfies target reliability and economic requirements. The propsoed appraoch is illustrated using a soil slope exam- ple. The results indicate that the Psf can effectively represent whether the design meets the target reliability requirements and construct a relatively accurate DRS. Computaional efficiency is significantly improved using subset simulation and DRS that decouples the double-loop scheme of RBDO.

关 键 词：概率充分因子 设计响应面 可靠度优化设计 子集模拟 边坡稳定 

分 类 号：TU43[建筑科学—岩土工程]