Modeling Single Well Injection-Withdrawal (SWIW) Tests for Characterization of Complex Fracture-Matrix Systems

An essential condition for performance evaluation of enhanced geothermal systems (EGS) resides in the ability to reliably predict fluid flow and heat transport in fractured porous rocks, where fast convection-dispersive transport through the fracture network can be strongly affected by heat conduction into the adjacent rock matrix. SWIW tests are single-well tracer tests that involve an initial period of fluid and tracer injection followed by a period of fluid withdrawal. As a result of the flow field reversal, the measured breakthrough curves tend to be less sensitive to advective heterogeneities and more sensitive to matrix diffusion and sorption, making this method very valuable in characterizing fracture-matrix interaction and evaluating matrix properties. In particular, we propose using SWIW tests before and after hydrofracking operations, to help assess the means by which hydrofracking increases permeability and enhances fracture-matrix interaction. In the present study, we have modeled single-well injection-withdrawal (SWIW) tests for non-sorbing and sorbing tracers, using the mixed Eulerian-Lagrangian transport simulator TRIPOLY, which solves tracer advection and dispersion in fracture networks together with solute exchange processes between the fractures and the porous matrix. Our simulations were conducted for hypothetical but workable SWIW test designs considering a variety of statistically generated 2D fracture-matrix systems. Parameter sensitivity studies were completed on three physical parameters of the rock matrix, namely porosity, diffusion coefficient and retardation coefficient, in order to investigate their impact on the fracture-matrix solute exchange process. Hydraulic fracturing, or hydrofracking, was modeled in two different ways, one by increasing the fracture aperture for flow and the other one by adding a new set of fractures to the fracture network. The results of all these different tests were analyzed by studying the population of matrix blocks, the tracer spatial distribution, and the breakthrough curves (BTCs), while performing mass balance checks to ensure numerical accuracy. The possibility of inferring from SWIW-test BTCs relevant information on the physical parameters of the fracture-matrix system was investigated. Our results clearly demonstrate the importance of matrix effects in the solute transport process. The sensitivity studies illustrate the increased importance of the matrix as providing a retardation mechanism as matrix porosity, diffusion coefficient, or retardation coefficient increase. Heat convection and conduction can be shown to be mathematically equivalent to advection, diffusion, and sorption of tracer, making these tracer studies directly useful for analysis of EGS. Interestingly, preliminary results before and after hydrofracking are insensitive to adding more fractures while somewhat sensitive to aperture increase, making SWIW tests a possible means of discriminating between these two potential hydrofracking effects. However, our base case fracture network is highly connected, potentially minimizing the effect of hydrofracking. Further study is needed using a sparser network to study hydrofracking under more realistic conditions.