Comparison of lab experiments and numerical experiments of single fracture flow

Mona Ribbers1, Sascha Frank1, Thomas Heinze1, Stefan Wohnlich1
1 Lehrstuhl Hydrogeologie, Ruhr-Universität Bochum

14.1 in Young Hydrogeologists Forum

To enable sustainable long-term use of a geothermal system, one aim of the ReSalt project is the investigation of scaling effects on permeability in jointed reservoirs. Within the scope of the project, hydrothermal water circulates in a high-pressure-high-temperature column containing a low-permeable sandstone core with a single fracture as hydraulic pathway. To identify the primary flow path along the fracture, we conduct lab experiments (including Darcy and tracer tests) and numerical simulations. We determine hydraulic parameters in the laboratory and compare those results to parameters required by numerical models reproducing the hydraulic behavior observed in the lab.

The hydraulic parameters are determined by Darcy tests using a common Darcy constant head laboratory setup. For each sample, we adjust the hydraulic head to obtain sufficiently small Reynolds number, securing laminar flow conditions. Water passes through the rock cores surrounded by a shrinking tube from bottom to top. The effective hydraulic aperture of the fracture is calculated based on the observed flow rate.
We determine transport parameters such as estimated velocity, dispersivity and dispersion coefficient using moment analysis of recorded breakthrough curves. A 2 molar NaCl solution as tracer is injected as a Dirac pulse and tracer concentration is estimated from electric conductivity measurements directly behind the core.

For the numerical model, fracture geometry is obtained through a laser scan of the surfaces, which are then digitally reassembled, and through a Micro-CT scan of the sandstone core. Both methodologies are compared for consistency, resolution and suitability for the numerical simulations. To simulate the water flow inside the single fracture under the stationary conditions of the lab experiments, we solve the Navier-Stokes equation assuming laminar flow conditions with no flow boundary conditions along the fracture surfaces. In- and outflow conditions, temperature, tracer concentrations etc. are chosen accordingly to the laboratory setup. Tracer transport is simulated by solving the advection-dispersion equation using a stationary flow field neglecting possible density driven flow.

In this work, we present the different sets of hydraulic parameters obtained from the lab experiments and the numerical simulation, identify bottlenecks, uncertainties, and discuss consequences for future single-fracture flow experiments and simulations. We further discuss the benefits and drawbacks of both scan methods with respect to the applicability for numerical simulation also with regard towards considered precipitation experiments.