基于改进神经网络的边坡岩体弹性力学参数识别方法  

Identification algorithm of elastic parameters of rock body in slope based on improved BP neural networks

作　　者：李守巨[1] 刘迎曦[1] 刘玉晶[2] 

机构地区：[1]大连理工大学工业装备结构分析国家重点实验室,辽宁大连116024 [2]吉林工学院基础科学系,吉林长春130012

出　　处：《湘潭矿业学院学报》2002年第1期58-61,共4页Journal of Xiangtan Mining Institute

基　　金：国家自然科学基金资助项目 (编号 :1 0 0 72 0 1 4 ) ;工业装备结构分析国家重点实验室开放基金资助项目 (编号 :GZ990 8)

摘　　要：基于人工神经网络的 BP算法 ,建立了根据边坡开挖后岩体位移观测数据识别岩体弹性力学参数的数值方法 .在网络训练过程中采用改进的 BP算法 ,通过对学习算子的优化搜索 ,大大提高了网络的收敛速度 ,解决了 BP算法迭代过程中目标函数振荡问题 .通过算例表明 ,提出的改进的 BP算法有助于提高岩土材料参数识别收敛速度和识别精度 .图5 ,表 3,参Based on the artificial neural networks(ANN) and according to measured displacements after the rock slope excavation, the identification algorithm of rock slope elastic parameters was set up.The improved BP algorithm of neural networks was applied during the neural networks training process. By application of optimized learning operator, the convergence speed of the objective function has been fasted and the oscillation problem of objective function has also been solved during the iteration. The numerical computation results show that the convergence speed and the identification precision have been improved by making use of improved BP neural networks algorithm proposed in this paper.5figs.,3tabs.,15refs.

关 键 词：边坡 岩体 弹性力学参数 识别方法 神经网络 参数识别 学习算子 

分 类 号：TU457[建筑科学—岩土工程]