Finite element free surface seepage analysis without mesh iteration
Finite element free surface seepage analysis without mesh iteration
复制标题
DOI:
10.1002/nag.1610030103
复制
发表时间:
1979
影响因子:
4
通讯作者:
K. Bathe;M. Khoshgoftaar
中科院分区:
文献类型:
--
作者:
K. Bathe;M. Khoshgoftaar
SUMMARY An effective solution procedure for the finite element analysis of free surface seepage problems is presented. The solution algorithm employs a non-linear permeability description of the material and avoids iteration with the finite element mesh. The results and experiences obtained in the analyses of some problems are presented to demonstrate the usefulness of the technique. The phenomena of fluid flow or seepage through porous media is observed in various disciplines of engineering.''2 It appears therefore natural that, as soon as the generality of the finite element method of analysis was recognized, emphasis was directed to develop the finite element method also for analysis of seepage problems in order to obtain a more genera1 analysis tooL3 Apart from being able to consider in an effective manner complex geometries and material properties, emphasis on the development of the finite element analysis pro- cedures is also important because of the potential of the technique for analysis of coupled stress and fluid flow problems.4s5 The current practice using the finite element method in the analysis of free surface fluid flow through porous media is to assume a free surface, discretize the domain below the free surface using finite elements, solve for the flow conditions in the finite element model, and check whether the free surface boundary conditions are satisfied with sufficient accuracy. If the flow conditions at the free surface are not satisfied to a specified tolerance, the free surface is adjusted and the problem is resolved until the free surface flow conditions are met. Depending on the problems considered, some 10 to 30 iterations may be necessary in steady-state analysis, and in transient analysis an iteration is carried out in the time steps of the time response calculation. In the iteration for the free surface, each iteration step represents a new problem, and a new finite element mesh could be established in each step. However, to keep the analysis effort to a minimum, usually the same basic finite element mesh is employed, but the geometric locations of the nodal points (possibly only near the free surface) are adjusted. The disadvantages of this scheme are that the elements can become very distorted, thus introducing severe errors in the analysis, and that a relatively large computational effort is required. These disadvantages are particularly pronounced in three-dimensional analysis. If non-linear stress and flow conditions * Associate Professor. t Research Assistant.