人工神经网络方法辨识岩体力学参数的可辨识性及其稳定性探讨  

DISCUSSION ON IDENTIFIABILITY AND STABILITY OF IDENTIFYING MECHANICS PARAMETERS OF A ROCK MASS WITH ARTIFICIAL NEURAL NETWORK

作　　者：刁心宏[1] 冯夏庭[1] 张士林[1] 赵炳祁[1] 

机构地区：[1]东北大学,辽宁沈阳110004

出　　处：《矿冶》2001年第3期11-14,19,共5页Mining And Metallurgy

基　　金：霍英东教育基金会资助 (710 48) ;高等学校博士点基金 (990 145 2 3)

摘　　要：探讨了以巷道围岩变形观测为依据 ,用神经网络反推岩体物理力学性质和初始地应力环境参数时各参数的可辨识性 ,以及参数辨识的稳定性。找出了待辨识参数个数与所需最少巷道变形观测项目数之间的关系 。Based on the survey to displacement of wall rocks of an opening,the identifiability and stability of identifying the mechanics parameters of a rock mass and the parameters of original stress condition with artificial neural network were investigated in this paper The relationship between parameter number to be identified and item of displacement survey was found out Suggestions for selecting survey items of opening displacement was put forward

关 键 词：人工神经网络 岩体力学参数 可辨识性 稳定性 

分 类 号：TD315[矿业工程—矿井建设]