基于神经网络理论的砂土液化的判别
详细信息 本馆镜像全文    |  推荐本文 | | 获取馆网全文
摘要
本文探讨了用神经网络方法来研究土的抗液化性能与土参数之间的非常复杂的非线性关系 ,并对实例进行网络测试。测试结果与其它液化判别方法所得的结果相比要更为可靠
The complicated nonlinear realations between parameters of soils and anti-liquefaction behaviors of soils are discussed by means of Artificial Neural Network(ANN) After training of network,based on the weights obtained the engineering examples are tested.The result of testing indicatestlat the methods of ANN to estimate sand liquefactions is better and more reliable than other methods such as the seed's method.
引文
[1 ]  Seed,H.B,Tokimatsu,H,Harder,L.F,and Chung,R.M."Influence of SPT procedure in seismic liquefactionresistance evaluations.”J.Geotech.Engrg.ASCE,1 985,1 1 1 ( 2 ) ,1 42 5-1 3 3 5.
    [2 ]  Stein,R.“Selecting data for neural networks.”AIExpert,1 993 ,8( 2 ) ,3 2 -4 7.
    [3 ]  Garson,G.D,“Interpreting neural network connec-tion weights.”AI Expert,1 991 ,6( 7) ,3 7-51 .
    [4 ] 周锡元、王广军、苏经军 .场地、地基、设计地震 [M].北京 :中国地震出版社 ,1 990
    [5] 楼顺天、施阳 .基于 MATLAB的系统分析与设计——神经网络 [M].西安 :西安电子科技大学出版社 ,1 998
    [6] 谢定义 .土动力学 [M].西安 :西安交通大学出版社 ,1 987

版权所有:© 2023 中国地质图书馆 中国地质调查局地学文献中心