| 饱和粘土本构关系的神经网络模型 |
| |
| 引用本文： | 曾静,王靖涛. 饱和粘土本构关系的神经网络模型[J]. 华中科技大学学报(自然科学版), 2002, 30(3): 68-70 |
| |
| 作者姓名： | 曾静 王靖涛 |
| |
| 作者单位： | 华中科技大学土木工程与力学学院,430074;华中科技大学土木工程与力学学院,430074 |
| |
| 摘 要： | 基于不同应力中径下饱和粘土的三轴试验和反问题理论,提出了用改进BP神经网络和径向基函数（RBF）神经网络这两种方法来建立粘土的本构模型。实例分析表明,两种模型对具体本构关系都能够很好逼近和预测,比较起来,径向基函数神经网络模型的稳定性好,且逼近速度快,而改进BP神经网络稳定性好,但逼近速度过慢。
|
| 关 键 词： | 应力路径 本构模型 BP神经网络 径向基函数神经网络 |
| 文章编号： | 1671-4512(2002)03-0068-03 |
| 修稿时间： | 2001-10-24 |
Two neural network models for saturated clay constitutive relations under different stress paths |
| |
| Zeng Jing Wang Jingtao Doctor Candidate, College of Civil Eng. & Mechanics,Huazhong Univ. of Sci. & Tech.,Wuhan ,China.. Two neural network models for saturated clay constitutive relations under different stress paths[J]. JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE, 2002, 30(3): 68-70 |
| |
| Authors: | Zeng Jing Wang Jingtao Doctor Candidate College of Civil Eng. & Mechanics Huazhong Univ. of Sci. & Tech. Wuhan China. |
| |
| Affiliation: | Zeng Jing Wang Jingtao Doctor Candidate, College of Civil Eng. & Mechanics,Huazhong Univ. of Sci. & Tech.,Wuhan 430074,China. |
| |
| Abstract: | Based on the triaxial experiments of saturated clay under different stress paths and inverse problem theory, the reformed back propagation (BP) neural network and the radial basis function (RBF) neural network are used to establish the constitutive model of clay. It is shown that these two methods can also approach the stress strain curve correctly. Compared with these two methods, the RBF has good stability and rapid speed, although the BP has good stability but the speed is too slow. These methods are valuable for developing the numerical soil mechanics. |
| |
| Keywords: | stress paths constitutive model BP neural network RBF neural network |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
