Study On Two-dimensional Adaptive Grid Generation Method Of Pore Ground-water Flow Finite Element Simulation

So far, Finite Element Method (FEM) is main method to numerically simulate the pore groundwater and the accuracy of space discrete items of regional triangular grid cells simulated determines the effectiveness of pore groundwater simulation. One important principle of the discrete space is "Where there is large hydraulic gradient of groundwater, there are dense grid cells". The FEM of pore groundwater simulation mainly develops on the inflexible triangular spatial discrete grids.This method would not be effective to deal with the engineering and scientific problems with increasing complexity since its inflexible triangular spatial discrete grids generated in the simulation is unable to express the variation of hydraulic gradient. However, hydraulic gradient of pore groundwater system can be expressed as a dynamic function of the intensity of groundwater exploitation, recharge time, intensity and mining. Different lag size of time would lead to spatial variation of hydraulic gradient. Therefore, if the large size of time in simulation models still stay static, the models could not meet well with the principles of discrete space in pore groundwater simulation. Therefore, it is necessary to propose an adaptive method to generate the discrete grid cells based on the features of simulation models of the finite element of pore groundwater. The main work in this paper includes:(1) Generation of triangular space discrete grid cells in numeric simulation of pore groundwater. Considering the characteristics of complexity of the boundary of simulating region in numeric simulation of pore groundwater and the demanding of adaptive grid cells, an advancing front technique proposed in this paper takes advantage of the adaptability of boundary and strong control capabilities of grid nodes and thus can be used to generate the finite element of grid cells. Through analyzing the advance front technique and the advancing process of the forefront of side, the change of morphology and its influence on construction of grid cells could be summarized as different situation, and on the basis of geometric principles, the solutions are proposed for each situation. Based on automation, quality of grid cells and applicability of algorithms, a method is proposed to improve the existing AFT algorithms and to effectively adapt to dynamic and automatic generation of spatial discrete grid cells with hydraulic gradient of pore groundwater changes.(2) Automatic extraction of geologic parameters. According to the spatial distribution of hydrogeological parameter, hydrogeological parameters in the numeric simulation of pore groundwater, e.g., ground water levels, water-conductive system, storativity, leakage, development strength of ground water, could be divided into " point", "line", and " polygon". Combining the mechanism and spatial structure of dynamic generation of triangular discrete grid cells, the parameters of pore groundwater simulation could be automatically extracted based on theories and methods of spatial analysis in GIS.(3) Adaptively adjustment of triangular spatial discrete grid cells. Considering the requirements of accuracy in pore groundwater simulation, the error of simulation is extracted and then is used as heuristic parameter to redo the triangulation on those regions with low accuracy. It is in this way that the discrete grid cells could be adaptively generated.