首页 | 本学科首页   官方微博 | 高级检索  
     检索      

软土盾构施工地表变形的小样本进化神经网络预测
引用本文:安红刚,胡向东,赵永辉.软土盾构施工地表变形的小样本进化神经网络预测[J].岩土力学,2003(Z2).
作者姓名:安红刚  胡向东  赵永辉
作者单位:同济大学地下建筑与工程系 上海200092 (安红刚,胡向东),同济大学地下建筑与工程系 上海200092(赵永辉)
基金项目:上海市教委重点学科基金资助项目。
摘    要:上海地区软土具有高压缩性和易塑流等特性,在盾构机的挤压和不当施工扰动下将会引起较大的土层移动和地表隆陷。以盾构施工实测位移资料为学习样本,通过遗传算法搜索具有最优预测效果的神经网络结构及学习参数。利用获得的进化神经网络在小样本训练情况下建立模型,对下一步施工的地表变形进行合理的预测。对上海市某盾构隧道的施工地表变形预测表明该模型可获得较高预测精度。

关 键 词:盾构施工  进化神经网络  小样本  地表变形预测

A study of prediction of ground deformation with small samples in shield construction in soft clays using ENN method
AN Hong-gang,HU Xiang-dong,ZHAO Yong-hui.A study of prediction of ground deformation with small samples in shield construction in soft clays using ENN method[J].Rock and Soil Mechanics,2003(Z2).
Authors:AN Hong-gang  HU Xiang-dong  ZHAO Yong-hui
Abstract:Concerning the property of high compression and easy rheology of soft clay in Shanghai, a great movement and hogging or hagging of ground may be produced during shield construction due to extrusion and unreasonable disturbance. A new method is proposed to predict the deformation by evolutionary neural network(ENN). The Genetic Algorithms is used to study the optimal structure and parameters of ENN. Learning from small samples, the deformation of next step is obtained by ENN. The case study shows the method makes a good performance in the prediction of ground deformation.
Keywords:shield construction  ENN  small samples  prediction of ground deformation
本文献已被 CNKI 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号