Study of constitutive relationships of multi-phase flow for geological carbon sequestration, The

Abstract

To develop effective strategies for safely storing CO2 in deep geological formations, it is necessary to understand the fundamental processes that contribute to stable entrapment in naturally heterogeneous subsurface formations. However, controlled experiments to generate data at all scales of interest are not feasible, and it is challenging to use supercritical CO2 (scCO2) in laboratory settings. Therefore, numerical modeling will be an essential tool for properly characterizing the capillary trapping of scCO2 in saline aquifers as well as improving the understanding of complex processes in multiphase flow systems. To produce an effective numerical model to simulate scCO2 behavior, it is essential to have reliable constitutive relationships such as the capillary pressure (Pc) -- saturation (Sw) and relative permeability (kr) -- saturation (Sw) relationships. Although previous studies have used constitutive models to derive constitutive relationships, the accuracy of the constitutive models for various fluids and porous media remains uncertain. Therefore, this research project developed column experiments to derive constitutive relationships by using surrogate fluids instead of scCO2 and brine and to investigate the applicability of the constitutive models. Having the surrogate fluids makes it more feasible and cost effective to run the experiments because they can be conducted at ambient conditions. Next, the results of an intermediate tank experiment conducted to explore trapping mechanisms of scCO2 and a numerical simulation using TOUGH2 T2VOC, using the derived constitutive relationships, were compared to verify the applicability of the derived constitutive relationships. Establishing the applicability of surrogate fluids to both measure Pc--Sw and kr--Sw relationships and producing accurate constitutive relationships will allow researchers to develop a greater understanding of the multiphase flow system and the trapping mechanisms of scCO2 in saline aquifers. Leverett scaling, the entry pressure scaling method, and the dielectric sensor method were used to obtain the Pc--Sw relationships. Both scaling methods use Pc--Sw relationships from an air/water system in which data was collected in previous studies. All three methods yielded similar curves for the Pc--Sw relationship. However, the entry pressure scaling method shows a slightly higher entry pressure value compared to other two methods. Moreover, the irreducible saturation and residual saturation for the Soltrol 220/ Glycerol -- water mixture differed from the air/water systems. Therefore, the study required measuring these two saturations independently from the scaling methods. The long column method and the hydrostatic method were used to measure the kr--Sw relationships. The experimentally obtained kr--Sw relationships differed greatly from the constitutive model kr--Sw relationships such as the van Genuchten -- Mualem model. It can be concluded that the van Genuchten -- Mualem model cannot be used for this test fluid and sand. Also, the result demonstrated that it is important to measure the kr--Sw relationships independently from the constitutive models. Moreover, the results did not show any hysteresis behavior. Lastly, the experimentally obtained constitutive relationships were applied to the numerical model to investigate the applicability of measured constitutive relationships and to examine the impact of constitutive relationships on the two-phase flow system. The simulation which used the experimentally obtained constitutive relationships better replicated the tank experiment compared to the one using the constitutive models without investigating the accuracy of its model. However, the numerical model could not reproduce the movement of the front of the plume because of the inappropriate entry pressure value. Overall, the study demonstrated the impact of the constitutive relationships on the two-phase flow system and the importance of having the applicable constitutive relationships to obtain a better numerical model.