FIGURE 1: Dover project area.

SAGP is a thermal oil recovery process that is similar to Steam-Assisted Gravity Drainage (SAGD) but which involves the addition of a small concentration of a non-condensable gas to the steam. This paper is a continuation of parts 1 and 2 presented at the 48 th and 49 th Annual Technical Meetings of the Petroleum Society.Theoretical developments and laboratory experiments continue to show significant improvements for the process as compared to SAGD. Experimental results have now been obtained with Athabasca crude oil as well as Cold Lake and Lloydminster type oils.In SAGP much of the oil displacement is caused by the flow of fingers of gas/steam rising counter-currently to the draining oil, rather than by the simple advance of a continuous steam chamber. The rising gas fingers raise the pressure in the reservoir above and this increase in pressure towards the top of the reservoir tends to push the oil down. Gas accumulates in the upper part of the reservoir and oil drains to the production well near to the bottom. The mechanism is discussed in the paper together with results from recent scaled, physical model experiments.The work demonstrates that SAGP may be expected to produce oil at rates nearly equivalent to SAGD but with much lower steam consumption. FIGURE 1: SAGP-Transfer regions. Journal of Canadian Petroleum Technology FIGURE 3: 1D grid blocks.FIGURE 2: Transport of water latent heat by saturated CH 4 /Steam mixtures. Journal of Canadian Petroleum TechnologyFIGURE 17: Vapour chamber at end of SAGP test (4.0 hrs, Lloydminister Crude; B3 = 5.2). FIGURE 18: Vapour chamber at end of SAGP Test (8.0 hrs, Cold Lake Crude; B 3 = 3.6). FIGURE 19: Vapour chamber at end of SAGP Test (8.0 hrs, Athabasca Crude; B 3 = 2.4).

Most of Canada's trillion barrels of petroleum consists of bitumen, and to a lesser extent, heavy oil. This total may be compared with the total Canadian production of light-medium crude oil to date, which is only 12 billion barrels. Mining is effective for the production of bitumen, but is limited to the minor fraction of the resource that is shallow; also mining involves significant environmental difficulties. The challenges of efficient in situ production are like those for other petroleum production activities:to find and define suitable reservoirs,to create conditions for oil to flow at economic rates, andto drain the reservoir systematically to obtain high recoveries. This paper discusses the following in terms of production rate, recovery, energy requirements and economic factors. Each topic is a step, sometimes a sideways step, in the search for a means to achieve high rates and recovery within the bounds of economic constraints.Cold production using vertical wells and horizontal wells.Stimulation by wellbore heating.Cyclic steaming using conventional wells.Steam Assisted Gravity Drainage (SAGD).Steam and Gas Push (SAGP).Cyclic Steaming with horizontal wells.Vapour Extraction (Vapex). Processes resulting in the displacement of oil by gas to a lower horizontal well show the most promise. The viscosity in the region around the horizontal well should be reduced to allow economic rates without gas coning. In SAGD, injected steam heats the oil and fills the reservoir as it drains, and rates of 79 – 159 m3/d (500 – 1,000 B/d) can be achieved with bitumen recoveries greater than 50%. Heat savings can be achieved by building a substantial gas concentration within the depleted region (SAGP). In Vapex, viscosity reduction is obtained by dilution with a olatile solvent; this is a promising approach for lower viscosity heavy oils. Another promising approach is cyclic steaming with horizontal wells, combined with gas addition to the steam to maintain drive. Introduction As conventional oil reserves become depleted, interest continues to grow in the improved recovery and utilization of Canadian tar sands and heavy oil. The resources are enormous in magnitude, and there have been great strides in technology. One approach, the mining and upgrading of shallow Athabasca deposits has, with the success of the Suncor and Syncrude projects, already become a major source of Canadian oil. Major expansions to both of these projects, as well as other new tar sand mining projects are underway. While mining overcomes the problem of moving the oil to the surface and of obtaining high recoveries, it requires the handling and disposal of vast amounts of solids and sludge, and it is only economic for the shallowest of deposits. The major part of the Canadian oil sand resource is too deeply buried for mining to be practical. This paper is concerned with the recovery of bitumen and heavy oils by in situ methods, i.e., by means of wells drilled from the surface.

At the 48th Annual Technical Meeting of the Petroleum Society, one of us presented a paper that showed that there was a possibility of making the SAGD process more efficient by adding a small concentration of a non-condensible gas such as methane to steam(1). For this to be effective the steam injection well should be located slightly above the production well. With this configuration, and with a small continuous production of gas with the produced oil and condensate, the non-condensible gas becomes concentrated in the upper part of the chamber and the heat loss to the overburden, and for the heating of the chamber, is greatly reduced; the steam oil ratio is much lower. Another configuration involves the continuous injection of a small stream of non-condensible gas from a well or wells near the top of the chamber with steam injection from a lower well or even into the production well. The heat is confined to the near wellbore region and again there is a considerable economy. The present paper discusses further analysis of these configurations and also results from physical model tests that are being carried out at the University of Calgary. The results of these experiments have been very positive and it appears that the concept may be even more effective than was predicted originally. The reason for this appears to be that the introduction of gas with steam invokes a new mechanism as the gas flows counter currently to the falling liquids; this mechanism involves the creation of a large surface area for mass transfer. As a result, the steam chamber is not only much lower in temperature, particularly at the top, but it also rises more slowly and spreads laterally more quickly. A larger volume is draining at a much lower temperature. Measurements made in our model show large improvements in the steam/oil ratio. The observation of the new mechanism suggests that this approach may have economic applications in fields having top water such as Surmont as well as in more normal type reservoirs. In general, the improved performance should broaden the range of reservoirs that can be produced economically. Introduction It is estimated that there are 273 billion m3 heavy oil and bitumen in place in Canada. They are deposited mainly in the Athabasca, Cold Lake, Lloydminster, and Peace River areas. Conventional heavy oil (10 to 20 ° API), which is partially recoverable by conventional in situ methods, is less than 2﹪ of the total resources. Most of the bitumen has a viscosity ranging from 100,000 to over 1.0 million mPa ⋅s at reservoir temperature. It is present in a solid or semi-solid state in the porous media and there is almost no mobility at the initial reservoir conditions. The effective recovery of bitumen by in situ methods is difficult. The current production of bitumen is over 400,000 barrels a day, which constitutes about 70﹪ of the total heavy oil and bitumen production(2) in Canada.

Injection into unstimulated heavy oil reservoirs generally results in disturbance to the soil matrix. To correctly model the fluid distribution, the fluid flow behaviour must be coupled to the mechanical behaviour of the sands. A model has been developed representing the conditions present during isothermal leak-off from fracture face, accounting for the physics of this coupling. The model assumes that the minimum effective stress controls the soil behaviour and that the fluids are linearly compressible. The model satisfactorily represents fracture growth inferred from pressure observation wells and measured mean times for communication between cyclic steam stimulation wells. Several new features generated by the model can be correlated with field behaviour. Nonlinear compressibility of the oil sand at low effective confining stresses causes increases in porosity which explain observed injectivity with zero initial mobility. Shear failure occurs around a fracture primarily due :0 decreasing effective stresses as the local pore pressure increases: Dilatant failure behaviour increases porosity and permeability, producing plateaus of increased water saturation. The coupling of the fluid and solid mechanics in the new model represents a significant advance in the realism of modelling of fluid flow and fracturing in oil sands. Introduction Injection of hot water or steam in oil sands results in complex interactions of the soil mechanics of the uncemented matrix and fluid flow and heat transfer in porous media. In the past, the recovery mechanisms were studied mostly using multiphase flow models developed for consolidated porous media(1–3). Such models are not adequate for modelling oil sands with low initialobility and one has to resort to artificial means of providing injectivity. In 1980s the importance of the geotechnical aspects was recognized, but the applications were limited mostly to surface mining problems. In most in situ projects, the formation is fractured during the first, or several of the injection cycles. Early work on fracture modelling(4, 5) showed that the proper representation of the fracturing and soil mechanics will be the key to any realistic modelling effort in oil sands. To date, all models including the fracture have been severely compromised by poor treatment of fracture mechanics or assumptions of linear elasticity(6, 7). This work describes part of the results of an ongoing research effort for the development of a practical, but realistic model for thermal processes in oil sands. In this paper we will restrict the discussion to isothermal behaviour for the following reasons:It is necessary to understand isothermal behaviour before studying more complex thermal process.Fracture is established on the first cycle of steam injection during which the temperatures will be lowest. Due to heat transfer in the fracture, the injected fluid at the tip will have a temperature close to that of the reservoir. Even for the isothermal case, one has to deal with several facts observed in the field, which contradict classical reservoir and fracture mechanics:Observed fracture dimensions are relatively small and fracture widths are large(8).Injectivity is larger than what would correspond to in situ mobility of fluids(5).