基于遗传神经网络的CFG桩复合地基承载力预测  

The bearing capacity prediction of CFG pile composite foundations based on genetic-neural network

作　　者：姜伟[1,2] 马令勇[3] 刘功良[3] 

机构地区：[1]中国地震局工程力学研究所,黑龙江哈尔滨150080 [2]黑龙江八一农垦大学工程学院,黑龙江大庆163319 [3]东北石油大学土木建筑工程学院,黑龙江大庆163318

出　　处：《世界地震工程》2010年第S1期263-266,共4页World Earthquake Engineering

基　　金：黑龙江省教育厅科学技术研究项目(11513015)

摘　　要：由于影响复合地基承载力的因素众多,各因素又互相作用,很难建立准确的承载力计算公式,现在设计中广泛使用的公式及其他理论公式仍是借助于试验基础上的半经验半理论公式,在实际工程中设计往往偏于保守造成不必要的浪费。文中提出利用MATLAB工具箱建立的基于遗传神经网络的CFG桩复合地基承载力预测模型来对CFG桩复合地基承载力进行预测,从预测的结果可知该方法运行简单,操作方便,准确度高。Due to the multi-influence interacted factors on bearing capacity of composite foundation,it is hard to set up accurate calculation formula of the foundation capacity.Now the formula and other theory formula widely used in design are the partly experimental,partly theoretical formula based on tests,which are conservative and wasteful in engineering design.Based on MATLAB genetic toolbox,the neural network based prediction model is established and applied to CFG pile composite foundation bearing capacity prediction.It is shown that this method is easy to work and convenient to operate,and has high accuracy.

关 键 词：遗传神经网络 CFG桩复合地基 地基承载力 MATLAB 

分 类 号：TU473.11[建筑科学—结构工程]