基于有限元和神经网络法的桩基础沉降分析  

The Research of Settlement of Composite Piled Foundations Based on Finite Element Method and Neural Network Method

作　　者：蒋青青[1] 唐振[1] 孟庆玲[2] 林峰[1] 

机构地区：[1]中南大学资源与安全工程学院,湖南长沙410083 [2]上海宝钢工程技术公司,上海201900

出　　处：《南华大学学报（自然科学版）》2009年第2期107-110,共4页Journal of University of South China：Science and Technology

基　　金：国家自然科学基金资助项目(50490270)

摘　　要：以沉降控制复合桩基础的沉降为研究对象,介绍了有限元和神经网络法的数值分析方法,并运用这两种方法对桩的沉降量进行了预测,对比了它们的结果,分析其产生差异的原因,结果表明:两种方法的沉降预测数据与实测数据误差均小于10%,说明预测精度较高,可以在节省资源的情况下获得较理想的预测结果.The settlement of piled foundation is the subject of research. The finite element method and neural network method are introduced and both of them are used for settlement prediction. The results are compared and they indicate that prediction data is close to the measured data, both methods are proved reasonable and feasible.

关 键 词：桩基础 有限单元法 神经网络 沉降预测 

分 类 号：TU473.1[建筑科学—结构工程]