Bitumen from thermal recovery processes is produced mainly in the form of water-in-oil emulsions. Generally, it has been believed that these emulsions form al restrictions in and around the wellbore, and/or in the surface sampling ports, but not in the reservoir. Thus, it has been assumed to be appropriate to model the flow of bitumen and water using standard two phase relative permeability concept.
In this paper, we present both laboratory and field evidence for the in situ formation and flow of emulsions in porous media. Laboratory experiments demonstrate that emulsions form readily when steam or water mist directly contacts the Cold Lake bitumen. Furthermore, careful sampling shows that emulsions are formed when dehydrated bitumen and water flow through a sandpack. The injection or dewatered bitumen into a Cold Lake core removes water from the core by emulsification. Water-In-oil emulsions are also generated during cyclic steam stimulation in a 3D scaled physical model.
Field samples from Esso Resources Canada Ltd.s cyclic steam stimulation operations at Cold Lake indicate that produced bitumen is commonly in the form of water-in-bitumen emulsions. Four fluid flow regimes are identified which are in agreement with those identified from field oil and water production rates. Production from wells before steaming, completed with or without casing and perforations, also shows the presence of water-in -bitumen emulsions.
The implication of emulsion flow in porous media is that the existing two phase relative permeability concept commonly used to describe immiscible displacement should be modified to allow for the mixing of the two phases to flow as an emulsion.
It is common practice to add demulsitiers to remove water from the produced water-in -bitumen emulsions before the bitumen is transported. While studies exist of emulsions flow in porous media1–5, it has been widely accepted that the emulsions form at either the perforations, the pumps, and/or at the valves of sampling instruments.
Recently, Vitloratos6 studiedfield production records, and discussed the difficulties encountered in trying to understand production behaviour based on immiscible flow. He concluded that accepting the possibility or insitu emulsion flow provides us with a simpler interpretation of the production data. His hypothesis was based on field oil and waler production rate data and our data on cold Lake wellhead emulsion samples in general, we see four flow regimes during production in cyclic steam stimulation(CSS) from our wellhead sampling. Initially only free water is produced (regime 1). This is clearly seen both in wellhead samples and In the fluid production data. In regime 2, we see a mixture of water-in-oil emulsions and free water. In the production data this regime is rellected as as a period with ilucluating oil and water rales. In regime 3 we see only water-in-oil emulsions and no free water. Finally in regime 4, we see a mixture of free water and water-in oil emulsions. Since both the emulsion samples and the production data were taken at the surface they do not necessarily indicate emulsion formation and emulsion flow in the reservoir.