基于BP神经网络的李家峡拱坝材料参数反演  

The convolute analysis of Lijiaxia arch dam material parameter based on BP neural network

作　　者：练继建[1] 王春涛[1] 赵寿昌[2] 

机构地区：[1]天津大学建筑工程学院,天津300072 [2]黄河上游水电开发有限责任公司,西宁810000

出　　处：《水力发电学报》2004年第2期44-48,共5页Journal of Hydroelectric Engineering

摘　　要：利用神经网络模型对大坝材料参数进行反演 ,可避免建模因子的选择不当而造成的误差 ,精度较高。本文利用有限元模型建立神经网络的训练样本 ,采用BP神经网络模型对李家峡拱坝进行了坝体和坝基材料参数反演 ,最后以反演后的材料进行正向计算得到大坝的变形 ,并与实测变形对比 ,结果令人满意。The error caused by the unsuitable choice of model factor can be avioded by the convolute analysis of dam material parameter based on the model of neural network.The precision of this model is high.By use of FEM the sample of neural network model can be set up, and by use of BP neural network model the convolute analysis of Lijiaxia arch dam material parameter is presented in this paper. Finally, Compare the dam deformation calculated by convolute analysis of dam material parameter with the actual deformation, the result is satisfactory.

关 键 词：水工结构 材料参数反演 神经网络 确定性模型 

分 类 号：TV4[水利工程—水工结构工程]