Supercritical CO2 flow through fractured low permeability geological media: experimental investigation under varying mechanical and thermal conditions

Abstract

To ensure secure geological storage of carbon dioxide it is necessary to establish the integrity of the overlying sealing rock. Seal rock fractures are key potential leakage pathways for storage systems; understanding their behaviour in the presence of CO2 under reservoir conditions is therefore of great importance. This thesis presents experimental investigations into the hydraulic behaviour of discrete fractures within low permeability seal rocks during single phase supercritical CO2 flow, under varying mechanical and thermal conditions representative of in-situ conditions. An experimental rig was designed and built to enable the controlled study of supercritical CO2 flow through 38 mm diameter samples under high pressures and temperatures. Samples are placed within a Hassler-type uniaxial pressure cell and CO2 flow is controlled via high precision syringe pumps. Flow experiments with supercritical CO2 within the pressure range 10-50 MPa were undertaken at temperatures of 38°C and 58°C with confining pressures of 35-55 MPa. The effects of stress loading and temperature change on the hydraulic properties of the fractured sample were studied; continuous differential pressure measurement enabled analysis of hydraulic response. Experiments were undertaken on a pre-existing Wissey field Zechstein Dolomite fracture and three artificial fractures (two East Brae field Kimmeridge Clay samples and one Cambrian shale quarry sample). Fracture permeabilities ranged from 8 X 10-14 m2 to 6 X 10-11 m2 with higher permeabilities observed within the harder rock samples. A broadly linear flow regime, consistent with Darcy's law, was observed in the lowest permeability sample (East Brae). A Forchheimer-type non-linear flow regime was observed in the other samples. Transmissivity variations during experiments were used to infer the mechanical impact of stress and temperature changes. An increase in effective stress resulted in transmissivity reduction, suggesting fracture aperture closure. During initial stress loading cycles, and subsequent higher temperature stress loading, a component of this transmissivity reduction was found to be inelastic, suggesting permanent modification of fracture geometry during closure. Pre- and post-experiment fracture surface characterisation provides further evidence for the occurrence of plastic deformation. Transmissivity-stress relationships were elastic during subsequent external stress-loading cycles, suggesting elastic closure and opening of fractures without additional permanent fracture geometry changes. The impact of fluid property variations on fracture hydraulic conductivity, Kfrac, was also analysed. Under constant effective stress Kfrac was found to be higher within high temperature and low fluid pressure scenarios, due to higher density/viscosity ratios. However, under constant confining pressure, fluid pressure changes are coupled both to mechanical effects (from effective stress alteration) and hydraulic effects (from viscosity variation), with opposing impacts on fracture hydraulic conductivity. At lower effective stresses mechanical effects were found to be dominant, with fluid pressure increase resulting in a notable increase to Kfrac due to aperture opening. At higher effective stresses, mechanical changes are much smaller due to increased contact area between fracture surfaces, and thus increased stiffness of fractures. Under such conditions hydraulic effects may be dominant and result in a small Kfrac reduction as fluid pressure increases, due to a reduction in the density/viscosity ratio. These results highlight that CO2 fluid property variation can have a notable influence on hydraulic conductivity under certain in-situ conditions. The single phase CO2 fracture flow experiments undertaken during this study were designed to enable a study of hydraulic and mechanical processes in isolation, without the influence of chemical processes. In-situ, the additional presence of brine and thus multiphase fluid behaviour and associated chemical processes makes the hydraulic behaviour of fractures considerably more complex. Coupled process modelling enables the relative influence of these processes to be simulated, but relies on experiments for validation. These unique experimental findings are of great value for enabling validation of such models as well as for informing analyses of geological and field studies.