Numerical Simulation Study Of CO₂ And Brine Core Flooding In Porous Media

Global warming caused by greenhouse effect has been a hot topic.Carbon dioxide?CO₂?is the major component of greenhouse gases.Thus,reducing CO₂ emission has become a necessary method in sustainable development of society.CO₂ sequestration technology in saline aquifers is considered as one of the most promising technologies to reduce CO₂ emissions.The broad application prospect has attracted the attention of many countries and scientific research institutions.The saline aquifer is a complex hydrogeology process affected by some factors such as pore structure and petrophysical parameters,etc.Core-scale research on CO₂sequestration in saline aquifers can help to provide the essential understanding of the rule of CO₂/brine multiphase flow in porous media and the mechanism of local trapping.It also provides a theoretical basis for improving the prediction accuracy of ultimate fate and reliably evaluating the storage capacity and security in saline aquifers.In this study,we construct five 3D heterogeneous rock cores whose local petrophysical properties are determined from measurement values and accordingly correlated well with pore structure,and evaluate the effects of petrophysical heterogeneity on numerical analysis of CO₂core flooding.Results indicate:1)The model error from the simplification of 3D heterogeneity into 1D axial-direction one could make the predicted local CO₂ saturation(S_CO2)largely deviate from the actual magnitude and variation trend.In a 2D cross-section from a 3D heterogeneous core,porosity inhomogeneity setting is also demonstrated to have no specific correlation with the predicted SCO₂ distribution.2)The setting of the local petrophysical properties for a low-porosity heterogeneous structure has a significant impact on the predicted S_CO2 distribution in front of it.The uncertainty in its local porosity may be related to permeability magnitude of the whole core,influencing the modeling of fluid pressure transfer and the computation of S_CO2.The error of the local capillary entry pressure can not only lead to underestimation of SCO₂ but also change variation rule of the local S_CO2 near the structure.3)Both magnitude and variation trend of the local SCO₂ could be numerically sensitive to 3D spatial distribution of porosity,notably increasing the probability of mismatching between model prediction and experimental observation.Besides,the effects of different parameter settings and heterogeneity distribution of relative permeability?k_r?curve on CO₂ local sequestration are also numerically simulated.Our results indicate:1)As two important parameters of kr functions,the error in residual gas saturation(S_gr)can cause almost the same degree of the deviation in predicted S_CO2 distribution as that in residual liquid saturation(S_lr);2)Different k_r functions significantly influence numerical analysis of CO₂ local trapping.Comparably,in the case of Corey's curve,the predicted CO₂ local accumulation is most obvious,whereas it is converse if using van-Genuchten's curve.Due to much larger effects on the model prediction,specifying a proper k_r function is more important for improving predictive precision of S_CO2 than eliminating errors in S_lr and S_gr;3)k_r curve can be various for relatively dense and loose zones inside sandstone cores.In this case,the effect of the heterogeneity in k_r curve on predicted S_CO2 distribution diminishes with the increased contrast of capillary entry pressure between these two zones.When such contrast is big enough,the effect of k_r curve heterogeneity can be negligible;4)Although the errors and uncertainties in k_r curves can affect S_CO2 prediction,they cannot change the locations where CO₂ local capillary trapping occurs.