Keywords

Groundwater flow; Adaptive mesh refinement; Posteriori error estimation; Non-linear simulations; Unconfined aquifers

Location

Session F5: Advances in Environmental Software Systems

Start Date

18-6-2014 10:40 AM

End Date

18-6-2014 12:00 PM

Abstract

The simulation of groundwater flow based on numerical methods requires sufficient discretization in areas of the domain where large changes of the hydraulic head gradient are expected. Mesh generation is the first step of numerical modeling, whereby the domain is discretized into a mesh of elements. The density of the mesh is defined prior to the simulation based primarily on user experience. In addition, a priori mesh generation for non-linear unconfined flow problems becomes very inefficient as there is no information regarding the water table. To alleviate the subjective and experimental mesh generation procedure in non-linear problems we propose a coupled scheme which combines adaptive mesh generation with a non-linear solver. Starting with a very coarse initial mesh, the mesh nodes are modified after every iteration of the non-linear solver so that the top boundary is equal with the hydraulic head. In addition, a refinement/coarsening step is executed where certain cells with very large or very small errors are refined or coarsened, respectively. The method is applied to a two dimensional hypothetical aquifer using three different distributions of hydraulic conductivity. The results show that the method is capable of producing optimized meshes by adding necessary degrees of freedom. In addition the non-linear problem converges during the first iterations when the mesh is coarser thus a faster convergence is achieved.

Share

COinS
 
Jun 18th, 10:40 AM Jun 18th, 12:00 PM

Simulation of Groundwater Flow based on Adaptive Mesh Refinement

Session F5: Advances in Environmental Software Systems

The simulation of groundwater flow based on numerical methods requires sufficient discretization in areas of the domain where large changes of the hydraulic head gradient are expected. Mesh generation is the first step of numerical modeling, whereby the domain is discretized into a mesh of elements. The density of the mesh is defined prior to the simulation based primarily on user experience. In addition, a priori mesh generation for non-linear unconfined flow problems becomes very inefficient as there is no information regarding the water table. To alleviate the subjective and experimental mesh generation procedure in non-linear problems we propose a coupled scheme which combines adaptive mesh generation with a non-linear solver. Starting with a very coarse initial mesh, the mesh nodes are modified after every iteration of the non-linear solver so that the top boundary is equal with the hydraulic head. In addition, a refinement/coarsening step is executed where certain cells with very large or very small errors are refined or coarsened, respectively. The method is applied to a two dimensional hypothetical aquifer using three different distributions of hydraulic conductivity. The results show that the method is capable of producing optimized meshes by adding necessary degrees of freedom. In addition the non-linear problem converges during the first iterations when the mesh is coarser thus a faster convergence is achieved.