Journal of Rock Mechanics and Geotechnical Engineering (Aug 2020)

Comparison of numerical codes for coupled thermo-hydro-mechanical simulations of fractured media

  • Ahmad Zareidarmiyan,
  • Hossein Salarirad,
  • Victor Vilarrasa,
  • Kwang-Il Kim,
  • Jaewon Lee,
  • Ki-Bok Min

Journal volume & issue
Vol. 12, no. 4
pp. 850 – 865

Abstract

Read online

Geo-energy and geo-engineering applications, such as improved oil recovery (IOR), geologic carbon storage, and enhanced geothermal systems (EGSs), involve coupled thermo-hydro-mechanical (THM) processes that result from fluid injection and production. In some cases, reservoirs are highly fractured and the geomechanical response is controlled by fractures. Therefore, fractures should explicitly be included into numerical models to realistically simulate the THM responses of the subsurface. In this study, we perform coupled THM numerical simulations of water injection into naturally fractured reservoirs (NFRs) using CODE_BRIGHT and TOUGH-UDEC codes. CODE_BRIGHT is a finite element method (FEM) code that performs fully coupled THM analysis in geological media and TOUGH-UDEC sequentially solves coupled THM processes by combining a finite volume method (FVM) code that solves non-isothermal multiphase flow (TOUGH2) with a distinct element method (DEM) code that solves the mechanical problem (UDEC). First, we validate the two codes against a semi-analytical solution for water injection into a single deformable fracture considering variable permeability based on the cubic law. Then, we compare simulation results of the two codes in an idealized conceptual model that includes one horizontal fracture and in a more realistic model with multiple fractures. Each code models fractures differently. UDEC calculates fracture deformation from the fracture normal and shear stiffnesses, while CODE_BRIGHT treats fractures as equivalent porous media and uses the equivalent Young's modulus and Poisson's ratio of the fracture. Finally, we obtain comparable results of pressure, temperature, stress and displacement distributions and evolutions for the single horizontal fracture model. Despite some similarities, the two codes provide increasingly different results as model complexity increases. These differences highlight the challenging task of accurately modeling coupled THM processes in fractured media given their high nonlinearity.

Keywords