Increasing demand of oil and gas worldwide is promoting a new, fast growth in the oil industry where the presence of experienced engineers is limited. Exploration for hydrocarbons is reaching limiting frontiers and the near future and long term challenge will be to maximize recovery from the existing fields. Enhanced oil recovery offers an alternative to improve recovery by means of introducing an external agent which enhances oil sweeping at a pore level scale.
While the EOR concept is not new; field implementation has been scarce. As a consequence the physics governing the displacement processes have not been completely understood, posing a challenge for the design and modeling of the process. This is enhanced when dealing with numerical models, which, typically are designed for primary and/or secondary depletion processes, with grid orientations and dimensions suitable for these field conditions. Very often though, these same models are used to design and evaluate the potential for field EOR. This paper addresses the main challenges of modeling the fine scale displacement mechanism with a full field model, highlighting the typical errors in recovery efficiency that can occur and suggesting scales at which screening models can be built.
Displacement processes in the reservoir are dominated by the combination of the viscous and capillary forces, the efficiency and ultimately the amount of displaced oil is controlled by the balance of these forces. During a core scale displacement process, viscous forces are dominant and most of the oil is contacted by the injected agent. This displacing mechanism is different from the one experienced at reservoir conditions where gravity forces play an important role, influencing the amount of oil which is contacted by the EOR agent, where under and overrunning may occur. Modeling of these displacements requires a greater resolution than the one used in for the full field model. The impact of model size and force balance during an EOR displacement process is presented is this paper.