水布垭面板堆石坝变形反馈分析  

Back analysis on deformation of Shuibuya CFRD

作　　者：王彭煦[1] 宋文晶[1] 

机构地区：[1]清华大学水沙科学与水利水电工程国家重点实验室,北京100084

出　　处：《水力发电学报》2009年第3期75-79,共5页Journal of Hydroelectric Engineering

摘　　要：根据水布垭面板坝实测变形,在试验参数基础上采用神经网络和遗传算法反馈得到堆石料的清华K-G模型参数,对大坝蓄水后的变形和应力进行分析。预测认为,正常高蓄水位下坝体最大沉降约为坝高的1%,面板最大法向位移约为457mm,应力状态表现为河谷部位受压,周边坝肩部位受拉,接缝体系的变形都在止水承受能力以内。The direct back-analysis method combined with the neural networks and genetic algorithm is adopted to compute Tsinghua K-G model parameters of Shuibuya concrete face roekfill dam（CFRD） according to the measured settlement in Dec. 2006. Tsinghua K-G model is used for 3-D FEM analysis to predict deformation and stresses when the water level reaches the normal high water level of 40Ore. The maximum computed settlement of roekfill dam is approximately 2.39m, which is only 1% of the dam height. The max normal displacement of face slab is computed about 457mm, and the result of stress analysis indicates that the face slab is compressed in central and otherwise around abutment. The displacements of vertical and peripheral joints are less than 5 cm within the capability of waterstop strips.

关 键 词：水工结构 反演分析 神经网络 遗传算法 清华K-G模型 堆石坝 

分 类 号：TV31[水利工程—水工结构工程]