Coupling geological and numerical models to simulate groundwater flow and contaminant transport in fractured media

  • Authors:
  • Daniela Blessent;René Therrien;Kerry MacQuarrie

  • Affiliations:
  • Département de Géologie et Génie Géologique, Université Laval, Ste-Foy, Québec, Canada G1K 7P4;Département de Géologie et Génie Géologique, Université Laval, Ste-Foy, Québec, Canada G1K 7P4;Department of Civil Engineering, University of New Brunswick, P.O. Box 4400, Fredericton, New Brunswick, Canada E3B 5A3

  • Venue:
  • Computers & Geosciences
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

A new modeling approach is presented to improve numerical simulations of groundwater flow and contaminant transport in fractured geological media. The approach couples geological and numerical models through an intermediate mesh generation phase. As a first step, a platform for 3D geological modeling is used to represent fractures as 2D surfaces with arbitrary shape and orientation in 3D space. The advantage of the geological modeling platform is that 2D triangulated fracture surfaces are modeled and visualized before building a 3D mesh. The triangulated fractures are then transferred to the mesh generation software that discretizes the 3D simulation domain with tetrahedral elements. The 2D triangular fracture elements do not cut through the 3D tetrahedral elements, but they rather form interfaces with them. The tetrahedral mesh is then used for 3D groundwater flow and contaminant transport simulations in discretely fractured porous media. The resulting mesh for the 2D fractures and 3D rock matrix is checked to ensure that there are no negative transmissibilities in the discretized flow and transport equation, to avoid unrealistic results. To test the validity of the approach, flow and transport simulations for a tetrahedral mesh are compared to simulations using a block-based mesh and with results of an analytical solution. The fluid conductance matrix for the tetrahedral mesh is also analyzed and compared with known matrix values.