Influence of capillary pressure boundary conditions and hysteresis on CO2-water relative permeability

Abstract

This study examines the effect of capillary pressure boundary conditions and hysteresis on CO2-water relative permeability. We perform drainage and imbibition injections into a Berea sandstone core and a sintered glass core. The cores receive three injection stages: water, then water-saturated CO2 representing drainage, and finally CO2-saturated water representing imbibition. Pressure difference and fluid production are recorded. To generate relative permeability using these data, we propose that the conventional Johnson–Bossler–Naumann (JBN) method be modified to incorporate capillary pressure boundary conditions. We then compare the results yielded by the JBN and modified-JBN methods, which generate relative permeability for post-breakthrough water saturation. To generate relative permeability for all states of saturation, history matching is used, in which the relative permeability of a numerical model is tuned until the simulated data match the experimental data.For drainage, history-matched relative permeability curves are found to be significantly closer to modified-JBN estimates than to JBN estimates. This indicates that the proposed modified-JBN method can provide a more accurate initial guess for history matching. In addition, using the initial guess generated by modified-JBN is found to reduce the chance of non-unique history-matched relative permeability curves.For imbibition, the CO2-water relative permeability curve is generated directly from history matching because the experimental data are too sparse to apply the JBN or modified-JBN method. Both CO2 and water imbibition relative permeability are found to differ from the drainage relative permeabilities. This so-called relative permeability hysteresis is attributed to contact angle hysteresis. In addition, imbibition CO2-water relative permeability is found to be sensitive to imbibition capillary pressure boundary conditions.