Modeling brine-rock interactions in an enhanced geothermal systemdeep fractured reservoir at Soultz-Sous-Forets (France): a joint approachusing two geochemical codes: frachem and toughreact

Abstract

The modeling of coupled thermal, hydrological, and chemical (THC) processes in geothermal systems is complicated by reservoir conditions such as high temperatures, elevated pressures and sometimes the high salinity of the formation fluid. Coupled THC models have been developed and applied to the study of enhanced geothermal systems (EGS) to forecast the long-term evolution of reservoir properties and to determine how fluid circulation within a fractured reservoir can modify its rock properties. In this study, two simulators, FRACHEM and TOUGHREACT, specifically developed to investigate EGS, were applied to model the same geothermal reservoir and to forecast reservoir evolution using their respective thermodynamic and kinetic input data. First, we report the specifics of each of these two codes regarding the calculation of activity coefficients, equilibrium constants and mineral reaction rates. Comparisons of simulation results are then made for a Soultz-type geothermal fluid (ionic strength {approx}1.8 molal), with a recent (unreleased) version of TOUGHREACT using either an extended Debye-Hueckel or Pitzer model for calculating activity coefficients, and FRACHEM using the Pitzer model as well. Despite somewhat different calculation approaches and methodologies, we observe a reasonably good agreement for most of the investigated factors. Differences in the calculation schemes typically produce less difference in model outputs than differences in input thermodynamic and kinetic data, with model results being particularly sensitive to differences in ion-interaction parameters for activity coefficient models. Differences in input thermodynamic equilibrium constants, activity coefficients, and kinetics data yield differences in calculated pH and in predicted mineral precipitation behavior and reservoir-porosity evolution. When numerically cooling a Soultz-type geothermal fluid from 200 C (initially equilibrated with calcite at pH 4.9) to 20 C and suppressing mineral precipitation, pH values calculated with FRACHEM and TOUGHREACT/Debye-Hueckel decrease by up to half a pH unit, whereas pH values calculated with TOUGHREACT/Pitzer increase by a similar amount. As a result of these differences, calcite solubilities computed using the Pitzer formalism (the more accurate approach) are up to about 1.5 orders of magnitude lower. Because of differences in Pitzer ion-interaction parameters, the calcite solubility computed with TOUGHREACT/Pitzer is also typically about 0.5 orders of magnitude lower than that computed with FRACHEM, with the latter expected to be most accurate. In a second part of this investigation, both models were applied to model the evolution of a Soultz-type geothermal reservoir under high pressure and temperature conditions. By specifying initial conditions reflecting a reservoir fluid saturated with respect to calcite (a reasonable assumption based on field data), we found that THC reservoir simulations with the three models yield similar results, including similar trends and amounts of reservoir porosity decrease over time, thus pointing to the importance of model conceptualization. This study also highlights the critical effect of input thermodynamic data on the results of reactive transport simulations, most particularly for systems involving brines.