基于BP神经网络的面板堆石坝参数反演分析  

Parameters Inversion Analysis of Facing Rock Fill Dam Based on BP Neural Network

作　　者：杨金强[1] 程壮[1] 肖蕾[1] 马莉[1] 

机构地区：[1]三峡大学三峡库区地质灾害教育部重点实验室,湖北宜昌443002

出　　处：《水电能源科学》2012年第6期89-91,101,共4页Water Resources and Power

基　　金：国家科学基金资助项目(90610029);湖北省自然科学基金资助项目(2008CD366);"十一五"国家科技支撑计划基金资助重点项目(2008BAB29B03-4)

摘　　要：高面板堆石坝的有限元计算分析中本构模型与模型参数的选取至关重要,尤其模型参数的选取与实际参数存在一定的差异性。根据面板堆石坝的特点,结合实测资料,基于BP神经网络选用合理方法构造参数训练样本,并给出了参数反演步骤,对坝体主、次堆石区的邓肯E-B模型参数反演与三维有限元进行了比较分析,计算结果与实测资料较为吻合。In the finite element calculation of high facing rock fill dam,it is essential to select the appropriate construction model,particularly the model parameters,which are difference from the actual ones.According to the characteristics of facing rock fill dam and the observed data,this paper designs reasonable training samples on the basis of BP neural network and gives specific steps for Duncan E-B model parameters inversion of the major and minor rock fill zones.Then the author makes a three-dimensional finite element calculation under the inversed parameters.By comparison,calculation result is consistent with monitoring data.

关 键 词：堆石坝 模型参数 反演分析 BP神经网络 

分 类 号：TV641.43[水利工程—水利水电工程]