Field-scale simulation of non-condensable gas injection in a hybrid SAGD process for the heavy oil and bitumen reservoirs requires a comprehensive numerical model that considers different mechanisms involved in the processes. The complexity of modelling such processes is due to the interrelation of oil-gas phase behavior and fluid transport mechanisms such as advection, dispersion, and dissolution. The non-condensable gas in the vapour phase is transported into the reservoir by the advection process which is dominantly controlled by the rock-fluid interaction parameters.
The steam and non-condensable fluid in the gaseous phase may have different relative permeability and end-point saturations associated with their interaction with oil/liquid. The diffusion and dispersion of non-condensable gas control how it spreads within a phase and its dissolution into oil and water determines its transfer across phases. Although, it is still debated that non-condensable gas accumulates at the steam front reducing heat transfer into the oil phase and lowering oil production rate, some field production data show contrary results. That is, the co-injection of non-condensable gas with steam does not affect the oil production rate while it reduces the steam-oil ratio. The non-condensable gas may accumulate at the top of the reservoir providing an insulation effect, reducing heat losses to the overburden and increasing the thermal efficiency of steam.
This study attempts to provide insights about field scale simulation of non-condensable gas injection in the hybrid SAGD process with a comprehensive numerical simulation model. Analysis will be presented regarding important mechanisms that should be considered, their impact on the process, and techniques to model them. Some of the important mechanisms that will be discussed are diffusion and dispersion, and solubility of gas in bitumen and water. The intent is to identify factors important to the modeling of these processes at field scale and how robust simulation models can be developed to replicate observed field behaviour.