摘要:采用改进的自动分组遗传算法,按现行规范实现了在概念设计阶段的桩基础优化设计。根据工程设计和施工的实际情况,提出了模块化方法,将整个基础划分为一定数量的模块,各模块以所含桩的数量、长度和直径为特征,兼顾桩距及布局等因素,能够实现桩基础拓扑、形状和尺寸的协同优化。在此基础上,建立了以桩基材料用量最少为目标,考虑承载力、沉降、偏心、软弱下卧层及分组等多个设计约束的桩基优化模型。求解上述优化问题,可获得具有最优设计变量取值和设计变量分组的可行设计。本文方法用于两个算例的优化,取得了良好效果。
This paper studies the code-based pile optimization with an improved automatic grouping genet- ic algorithm to obtain the optimal conceptual design. According to the engineering practice, a modular method is proposed to achieve the concurrent optimization of topology, geometry and size. This method divides the foundation into some modules and each module involves the pile length, pile diameter, pile number and pile layout. Based on the modular method, a minimum-volume pile optimization model sub- ject to the constraints of the bearing capacity, settlement, eccentricity, and variable grouping is built and solved to obtain the feasible design with the optimal design variables and optimal design variable group- ing. Two examples demonstrate the effectiveness of the proposed approach.