Long-Term Performance of Heat Mining in Hot Aquifers: Water-Rock Interaction and Permeability Changes

Abstract

We simulate the effect of chemical induced changes in the pore space structure of porous aquifers on flow and transport using a new numerical code developed in an interdisciplinary project of 3-year duration. The three key elements of this new simulation approach are: (1) a relationship between changes in porosity and permeability which is based on petrophysical principles; (2) a chemical model for precipitation and dissolution reactions at high temperature and salinity; (3) a numerical simulation tool for coupled flow, heat and multi-component transport, and chemical water-rock interaction. The derivation of the new relationship between porosity and permeability comprises the application of a fractal model of the pore space structure and its change due to precipitation and dissolution of minerals. Permeability is expressed as a power series of porosity. Based on the Kozeny-Carman equation (Carman, 1956) the exponents of this series are calculated from the fractal dimension of the fluid-rock interface which is a fundamental structural parameter in our approach. This expression was calibrated and validated with large petrophysical data sets from clean and shaly sandstones. This way, a set of curves was obtained in a permeabilityporosity plot in which each curve is characteristic for one type of sandstone. The geochemical reaction model is based on the speciation code PHRQPITZ (Plummer et al., 1988). Following our code extension and update of its thermodynamic data base it is now valid for brines of low to high (5 mol/l) ionic strength and temperatures of 0ñ150 ?C. Reaction kinetics was added to account for the fact that under- or over-saturation does not always trigger dissolution or precipitation, which was observed in the laboratory as well as in geothermal installations under operation. With this new numerical tool coupled, threedimensional simulations can be run for total operation times of 30-100 years and different production/injection regimes, types of fluid mineralization, and fractal dimensions of the precipitating mineral surface. It turns out that most changes in the system occur at the propagating temperature front and in the vicinity of the injection well due to the complex interaction of all coupled processes. Finally, we identify critical scenarios for the operation of a hydrothermal heat mining installation over a period of 30ñ50 years.