节点文献
用Hopfield神经网络辨识岩土流变本构模型
IDENTIFICATION OF RHEOLOGICAL CONSTITUTION MODEL OF ROCK AND SOIL USING HOPFIELD NEURAL NETWORK
【Author】 Chen Yuanjiang;Pan Changliang;Wang Wenxing;Cao Ping Institute of Mining and rock soil mechanics Engineering,Collgeg of Resources,Environment and Civil Engineering,Central South University, Changsha 410083 China
【机构】 中南大学资源环境与建筑工程学院采矿与岩土工程研究所；
【摘要】 该文基于Hopfield神经网络的优化计算原理,探讨了将其用于SISO线性定常系统模型辨识的可能性及算法,然后把该算法引入岩土流变本构模型的辨识,并在matlab软件包中开发了相应的计算程序,通过有关考题验证表明,该算法是成功可信的,能有效用于加载条件可控、位移输出可测的可近似描述为线性时不变系统的大量岩土工程流变变形破坏实际中。
【Abstract】 Based on the optimization computing theory of Hopfield Neural Networks(HNN),the paper stuided the probability of using HNN toidentify the model of the single-input-single-output liear stationary system and its algorithmic principle,and then used the method to identifythe rheological constitutive model of rock and soil and programmed for the identification algorithm on the base of MATLAB software.The test ofusing an examination question showed that the identification method was successful and believable and could be effectively used in the geotech-nical engineering practice of a lot of creep and failure which could be approximately linear stationary system whose loading condition control-lable and displacement output is measurable.
【Key words】 hopfield neural network； rock and soil material； rheology； constitutive model model identification；
- 【会议录名称】 岩石力学新进展与西部开发中的岩土工程问题——中国岩石力学与工程学会第七次学术大会论文集
- 【会议名称】中国岩石力学与工程学会第七次学术大会
- 【会议时间】2002-09
- 【会议地点】中国西安
- 【分类号】TU459
- 【主办单位】中国岩石力学与工程学会