Loading…

Verification benchmark for a single-phase flow hydro - mechanical model comparison between COMSOL Multiphysics and DuMuX

Numerical modelling of hydromechanical processes in geological environments has become an invaluable tool in understanding and predicting system behaviour. However, due to the different algorithms and numerical schemes implemented in the different models, model reliability may vary considerably. Mod...

Full description

Saved in:
Bibliographic Details
Main Authors: Zhou Dejian, Tatomir Alexandru, Tomac Ingrid, Sauter Martin
Format: Conference Proceeding
Language:English
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Numerical modelling of hydromechanical processes in geological environments has become an invaluable tool in understanding and predicting system behaviour. However, due to the different algorithms and numerical schemes implemented in the different models, model reliability may vary considerably. Modelling of single- and multi-phase flow in porous media has been widely employed in various engineering applications such as geological disposal of nuclear waste, geological storage of carbon dioxide, high-temperature geothermal systems, or hydraulic fracturing for shale gas exploitation. Coupled hydro-mechanical (H-M) processes play a key role in the prediction of the behaviour of geological reservoirs during their development and testing operations. In this paper, we present a benchmark test on a single-phase flow problem in a hydrogeological reservoir with 5 horizontal layers of different properties. The aim is to compare two hydromechanical (H-M) models that use a vertex-centred finite-volume discretization and a finite element discretization. The first model is constructed with the free-open source simulator DuMuX, and the second with the commercial software COMSOL Multiphysics. The verification study suggests general confidence in the model reliability, but also highlights and discusses several areas of discrepancies between two models.
ISSN:2267-1242
DOI:10.1051/e3sconf/202020502002