摘要
以龙滩水电站左岸进水口边坡为例,采用正交设计、神经网络和遗传算法建立了力学参数的联合位移反分析模型,反演了边坡岩体的综合力学参数。进一步针对高边坡施工期的工程特点,利用显式有限差分法程序FLAC3D对进水口边坡施工过程进行了三维弹塑性仿真模拟。基于该反演参数的正分析计算结果表明,计算位移与监测位移趋势一致,量值吻合,所得参数是合理的。最后分析了开挖完成后边坡岩体的应力分布和变形规律,并对边坡的整体稳定性进行了评价。
On the basis of the finite difference method considering the effect of excavation unloading, one back analysis model of mechanical parameters is established by using orthogonal design theory, neural network and genetical gorithm, and the relative mechanical parameters of the left-bank intake slope of Longtan Hydropower Project are back analyzed. Then, the excavation process of construction of the intake slope is simulated hy fast Lagrangian analysis of continuance in 3 dimensions （FLAC^3D）. The deformation pattern and stress state caused by excavation at section 1-1 of left bank slope are studied, and the slope stability is evaluated comprehensively.
出处
《长江科学院院报》
CSCD
北大核心
2008年第2期33-37,共5页
Journal of Changjiang River Scientific Research Institute
基金
国家重点基金研究发展规划(973)项目(2002CB412708)
国家自然科学基金面上项目(50579071)
关键词
岩石边坡
正交设计
神经网络
反分析
有限差分法
rock slope
orthogonal design
neural network
back analysis
finite difference method
