ParametersSensitivity Analysis and Computation of the Embankment Dam and its Application to Inversion and Locations Optimization of Displacement Monitoring Points
This paper aimed at the problems of huge computation burdens and poor accuracy because of lots of parameters in the embankment dams inversion.Authors designed a mixed-level test for several parameters of the Duncan-Chang Model through an orthogonal test method and numerical analysis with FEM.Using the different directional displacement of two nodes and principal stress of element as the test index and through variance analysis, we can find out the main sensitive parameters which control the strain and stress of dam, and which should be accurately obtained by back-analysis.In a dam system, the explicit functional relationship between material parameters and dam displacement is too complex to be derived and the analytical solution of the parameters sensitivity is unable to be computed.The common-used simplified method, the finite difference algorithm, also has the disadvantage of low accuracy due to being a rough model.In this paper, an analytical solution based on neural network was introduced to compute the parameters sensitivity of the embankment dam and numerical results on one simple beam structure were presented to show the efficiency and reliability of the introduced method.Then, this analytical method was employed in the sensitivity computation of main parameters which control the strain and stress of the embankment dam, the sensitivity s distribution of every parameter in dam body are obtained.Authors propose that the optimal location of the monitoring points should be in the place that has both the larger displacement and the higher sensitivity.If this is done, we can not only obtain the displacement information which has the higher signal-to-noise ratio, but also gain the larger parameters s ensitivity information.These monitoring points can provide the most valuable information for inversiug dam parameters and evaluating dam safety.Finally, an inversion example of dam parameters validates its correctness of the optimal location of the monitoring points.
embankment dam orthogonal test neural network sensitivity computation parameters inversion optimization of monitoring points
Song Zhiyu Dong Lili
Yellow River Engineering Consulting Co., Ltd., Zhengzhou,450003, China
国际会议
昆明
英文
493-511
2013-11-01（万方平台首次上网日期，不代表论文的发表时间）