| 注册
首页|期刊导航|铁道科学与工程学报|基于支持向量机沉降代理模型复合地基优化设计方法

基于支持向量机沉降代理模型复合地基优化设计方法

陈昌富 李欣 龚晓南 俞建霖

铁道科学与工程学报2018,Vol.15Issue(6):1424-1429,6.
铁道科学与工程学报2018,Vol.15Issue(6):1424-1429,6.

基于支持向量机沉降代理模型复合地基优化设计方法

Composite foundation's optimization design of settlement's surrogate models based on support vector machines

陈昌富 1李欣 1龚晓南 2俞建霖2

作者信息

  • 1. 湖南大学 岩土工程研究所,湖南 长沙 410082
  • 2. 浙江大学 岩土工程研究所,浙江 杭州 310000
  • 折叠

摘要

Abstract

During the optimization design of the complex composite foundations, it was difficult to establish a mathematical model with explicit expressions as a result of lacking explicit settlement formulas. On the other hand, a large amount of computation was required if only numerical simulation method was used to establish the optimization models. Therefore, a composite foundation's optimization design method of settlement's surrogate models was introduced. To establish the three dimensional finite element models of the composite foundations, many numerical experimentations were designed based on uniform design, and the settlements and differential settlements of the composite foundations were calculated. The surrogate models of the composite foundations were established based on the least squares support vector machine. The mathematical models of the composite foundations were established by using the costs of the composite foundations as the object functions and the settlements as the constraints. Based on the genetic algorithms, the most economic strengthening scheme was searched and exported, which satisfies the allowable settlements. The reliability of this method has been proved by using in the optimization design of the composite foundation in shore.

关键词

地基处理/均匀设计/有限元/支持向量机/遗传算法

Key words

foundation treatment/uniform design/finite-element/support vector machines/genetic algorithms

分类

建筑与水利

引用本文复制引用

陈昌富,李欣,龚晓南,俞建霖..基于支持向量机沉降代理模型复合地基优化设计方法[J].铁道科学与工程学报,2018,15(6):1424-1429,6.

基金项目

国家重点研发计划资助项目(2016YFC0800200) (2016YFC0800200)

国家自然科学基金资助项目(51278186) (51278186)

铁道科学与工程学报

OA北大核心CSCDCSTPCD

1672-7029

访问量0
|
下载量0
段落导航相关论文