Modeling of Laboratory Gas Flooding in Tight Chalk with Different Non-equilibrium Treatments
This study focuses on proper modeling of bypassed oil in tight chalk during gas injection, caused partly by the small-scale heterogeneity and the non-equilibrium contact especially in low permeable chalk.
Conventional compositional simulators using the local equilibrium assumption tend to predict excessive vaporization of the residual oil. We present the laboratory gas flooding results in tight chalk and discuss how different non-equilibrium treatments can provide more realistic simulation results.
Composite core flooding experiments with low-permeable tight chalk and natural gas were conducted at different pressures below the minimum miscibility pressure of the live oil used. The ECLIPSE compositional simulator E300, using an EoS model tuned with the swelling data, was used to history match the results. It was found that the simulation without considering non-equilibrium effects over-predicted the oil production in the late stage. Two methods were tested to avoid the excessive vaporization of oil: the Sorm method (excluding the residual oil from flash calculations) and the transport coefficients (alpha factors) method together with pseudo-relative permeability curves.
Our results show that the sub-grid non-equilibrium effect is significant in tight chalk. Compositional simulation without considering this effect leads to unrestricted vaporization and over-prediction of the oil recovery in gas injection into tight chalk even for laboratory experiments. Both methods tested here are suitable for reproducing the flooding results, in particular, the residual oil in the late stage. For the experiments studied here, the Sormmethod seems to show a better performance in maintaining no further mass transfer between the residual oil and gas after the ultimate recovery is reached, since it excludes the bypassed oil fraction from flash calculations and models the immobile saturation explicitly.
For the alpha factors method, oil production keeps a slow increase at the late stage as long as gas is being injected. In addition, the use of pseudo-relative permeability method can lead to obtaining irrational trends in some cases. We therefore propose an alternative method by adjusting the alpha factors of the mobile components, which avoids the difficulties of modifying the relative permeability curves.