摘要
结合桥梁深水群桩基础在运营期间的受力特征,研究了深水群桩基础基桩轴力混合模型的一般原理.提出了基于多种环境因素影响下群桩基础监测数据的运营期混合模型,利用有限元模拟与PSO-SVM统计方法在苏通大桥群桩基础中实现了混合模型的构建.为了方便比较,建立了径向基函数(RBF)人工神经网络模型,对比了混合模型与RBF模型的预测结果.研究表明,混合模型预测精度较高,在受力情况不同的3根基桩上都有较为稳健的预测能力,具有较强的泛化能力,混合模型可适用于深水群桩基础运营期轴力的预测.
Combined with the stress characteristics of the bridge deep-water pile foundation during the operation period, this paper studied the general principles of the mixed model for pile axial force in deepwater pile foundation. The mixed model was firstly presented during the operation period based on the monitoring data of pile group foundation considering the influence of multiple environment factors. The mixed model of Sutong bridge pile group foundation was built by using the finite element simulation and PSO- SVM statistical method. In order to facilitate comparison, the Radial Basis Function （RBF） artificial neural network model was built, whose results were compared with the prediction results of the mixed model. The results showed that the mixed model had higher prediction accuracy and more robust predictive ability for the three piles under different loading conditions, and it exhibited better generalization ability. The mixed model could be applied to the prediction of the axial force in the deep-water pile group foundation during the operation period.
出处
《湖南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2016年第3期155-160,共6页
Journal of Hunan University:Natural Sciences
基金
'十一五'国家科技支撑计划资助项目(2006BAG04B05)
国家重点基础研究发展计划(973计划)资助项目(2002CB412707)
关键词
有限元方法
深水群桩基础
运营期
混合模型
多因素
轴力预测
finite element method
deep-water pile group foundation
operation period
mixed model
multiple factors
axial force prediction
