基坑土钉支护可靠性分析优化算法  

An optimization method of reliability analysis of foundation pit reinforced with soil nailing

作　　者：朱剑锋[1,2] 陈昌富[2] 徐日庆[1] 

机构地区：[1]浙江大学软弱土与环境土工教育部重点实验室,杭州310058 [2]湖南大学岩土工程研究所,长沙410082

出　　处：《岩土力学》2010年第7期2336-2341,共6页Rock and Soil Mechanics

基　　金：国家自然科学基金(No.50678158);教育部博士点专项基金(No.20050532021)资助项目

摘　　要：针对基坑工程中岩土参数存在随机性和变异性的特点,基于响应面重构法、遗传算法和禁忌搜索方法研究了土钉墙边坡可靠性分析方法。考虑土钉的加固作用,建立了适用于土钉墙边坡任意形状滑面安全系数计算的改进Morgenstern-Price法。基于响应面原理,将改进Morgenstern-Price法取代传统响应面法中的有限单元法来随机抽样构造响应面函数,建立了一种近似的土钉墙边坡可靠度计算方法。以土体的抗剪强度指标c、?为随机变量,提出了一种能同时确定土钉墙边坡最小可靠度指标βmin及相应最危险滑面的全局优化计算方法——土钉墙可靠性分析自适应禁忌搜索遗传算法(ATSGA)。结合算例,分别以土钉墙边坡的最小可靠度指标和最小中值安全系数为目标函数,采用ATSGA法搜索其相应的最危险滑动面,结果表明,二者相差较大。Aiming at the uncertainty and variation of geo-material parameters in foundation pit engineering, the reliability analysis method of soil-nailing wall is discussed based on the response surface method, genetic algorithm and taboo search techniques. A modified Morgenstem-Price limit equilibrium method for analysis of the interior stability of soil-nailing wall is presented by considering the reinforcement effect of soil nails. An approximate reliability calculation method of soil-nailing wall, in which the response surface function is constructed by these sample points that are obtained from the modified Morgenstern-Price method instead of traditional FEM numerical simulation method, is established based on the response surface principle. Taking soil shearing strength parameters c and φ as the random variables, a global optimal algorithm-adaptive taboo search genetic algorithm （ATSGA） is presented for soil-nailing wall reliability, which can effectively locate the critical slip surface associated with the minimum reliability index. An example is given; and the calculation results computed by proposed method show that if taking the minimum reliability index and the minimum safety factor as the target function respectively, the two critical slip surfaces are significantly different.

关 键 词：基坑 Morgenstern Price法 可靠度指标 响应面 遗传算法 禁忌搜索 

分 类 号：O213.2[理学—概率论与数理统计]