STEAM ASSISTED GRAVITY DRAINAGE WITH ADDED OXYGEN ("SAGDOX") IN DEEP RESERVOIRS

Abstract

A process to recover hydrocarbons, from a hydrocarbon reservoir having a bottom, using a substantially horizontal production well, the substantially horizontal production well having a toe and a heel, the process including: (a) injecting oxygen into the hydrocarbon reservoir, the horizontal production well having at least one perforation zone for contact with the reservoir; (b) injecting steam into the hydrocarbon reservoir; the oxygen producing in situ heat and in situ carbon dioxide by combustion and the steam producing in situ heat by conduction and condensation; the in situ carbon dioxide dissolving into the liquid hydrocarbon, lowering its viscosity; (c) recovering the reservoir liquid hydrocarbons of lowered viscosity using the substantially horizontal production well; and (d) optionally conveying the recovered liquid hydrocarbons to the surface; where the process is absent a removal step of any non-condensable gas from the reservoir.

Description

BACKGROUND OF THE INVENTION

Steam Assisted Gravity Drainage (SAGD) is a commercial, thermal enhanced oil recovery (“EOR”) process. The SAGD process uses saturated steam injected into a horizontal well, where latent heat is used to heat bitumen in the reservoir. The heating of the bitumen lowers its viscosity, so it drains by gravity to an underlying parallel, twin, horizontal well completed near the reservoir bottom.

Since the process inception in the early 1980's, SAGD has become the dominant, in situ process to recover bitumen from Alberta's bitumen deposits (Butler, R., “Thermal Recovery of Oil & Bitumen”, Prentice-Hall, 1991). Today's SAGD bitumen production in Alberta is about 300 Kbbl/d with installed capacity at about 475 Kbbl/d (Oilsands Review, 2010). SAGD is now the world's leading thermal EOR process.

FIG. 1 (PRIOR ART) shows the “traditional” SAGD geometry, using twin, parallel horizontal wells 2,4 drilled in the same vertical plane. There is a 5-metre spacing between the horizontal wells 2,4, which are about 800 metres long with the lower well 1 to 2 metres above the (horizontal) reservoir floor. Circulating steam 6 in both wells starts the SAGD process. After communication is established, the upper well 2 is used to inject steam 6, and the lower well 4 produces hot water and hot bitumen 8. Fluid production is accomplished by natural lift, gas lift, or submersible pump.

After conversion to “normal” SAGD operations, a steam chamber 10 forms around the injection 2 and production wells 4 where the void space is occupied by steam 6. Steam 6 condenses at the boundaries of the chamber 10, releases latent heat (heat of condensation), and heats bitumen, connate water and the reservoir matrix. Heated bitumen and water 8 drain by gravity to the lower production well 4. The steam chamber 10 grows upward and outward as bitumen is drained.

FIG. 2 (PRIOR ART) shows how SAGD matures. A “young” steam chamber 10 has bitumen drainage from steep chamber sides and from the chamber ceiling. When the chamber growth hits the top of the reservoir, ceiling drainage stops, bitumen productivity peaks, and the slope of the side walls decreases as lateral growth continues. Heat loss increases (and steam-to-oil ratio (“SOR”) increases) as ceiling contact and the “surface area” of the steam chamber increases. Drainage rates slow down as the side wall angle decreases. Eventually, the economic limit is reached, and the end-of-life drainage angle is small (10-20°).

Produced fluids are near saturated-steam temperature, so it is only the latent heat of steam that contributes to the process in the reservoir. But, some of the sensible heat can be captured from surface heat exchangers (a greater fraction at higher temperatures), so a useful rule-of-thumb for net heat contribution of steam is 1000 BTU/lb. for the P, T range of most SAGD projects (FIG. 3 PRIOR ART).

The operational performance of SAGD can be characterized by measurement of the following parameters: 1) saturated steam P, T in the steam chamber (FIG. 4 PRIOR ART); 2) bitumen productivity; 3) SOR, usually at the well head; 4) sub-cool target, the T difference between saturated steam and produced fluids; and 5) Water Recycle Ratio (“WRR”), the ratio of produced water to steam injected.

During the SAGD process, the SAGD operator has two choices to make: 1) the sub-cool target T difference and 2) the operating pressure in the reservoir. A typical sub-cool of about 10 to 30° C. is meant to ensure no live steam breaks through to the production well. Process pressure and temperature are linked (FIG. 4 PRIOR ART) and relate mostly to bitumen productivity and process efficiency.

Bitumen viscosity is a strong function of temperature (FIG. 5). SAGD productivity is proportional to the square root of the inverse viscosity (FIG. 6 PRIOR ART) (Butler (1991)). Conversely if pressure (and T) is increased, the latent heat content of steam drops rapidly (FIG. 3). More energy is used to heat the rock matrix and is lost to the overburden or other non-productive areas. So, increased pressure increases bitumen productivity but harms process efficiency (increases SOR). Because economic returns can be dominated by bitumen productivity, the SAGD operator usually opts to target operating pressures higher than native or hydrostatic reservoir pressures.

Despite becoming the dominant thermal EOR process, SAGD has some limitations and detractions. The requirements for a good SAGD project are:

• • a horizontal well completed near the bottom of the pay zone to effectively collect and produce hot draining fluids.
  • the injected steam, at the sand face, has a high quality (latent heat drives the process).
  • the process start up is effective and expedient.
  • the steam chamber grows smoothly and is contained.
  • the reservoir matrix is good quality (porosity (φ)>0.2);
  • Initial Oil Saturation (S_io)>0.6; Vertical permeability (k_v)>2D).
  • net pay is sufficient (>15 metres).
  • proper design and control must achieved to simultaneously; 1) prevent steam breakthrough to the production well and injector flooding; 2) stimulate steam chamber growth to productive zones; and 3) inhibit water inflows to the steam chamber.
  • there must be absence of significant reservoir baffles or barriers.

If these conditions are not attained or other limitations are experienced, SAGD can be impaired, as follows:

(1) The preferred dominant production mechanism is gravity drainage, and the lower production well is horizontal. If the reservoir is slanted, a horizontal production well will strand significant resources.
(2) The SAGD steam-swept zone has significant residual bitumen content that is not recovered, particularly for heavier bitumens and low pressure steam (FIG. 7). For example with a 20% residual bitumen (pore saturation) and a 70% initial saturation, the recovery factor is only 71%, not including stranded bitumen below the production well or in the wedge zone between recovery patterns.
(3) To contain a SAGD steam chamber, the oil in the reservoir must be relatively immobile. SAGD cannot work on heavy (or light) oils with some mobility at reservoir conditions. Bitumen is the preferred target.
(4) Saturated steam cannot vaporize connate water. By definition, the heat energy in saturated steam is not high enough quality (temperature) to vaporize water. Field experience also shows that heated connate water is not usually mobilized sufficiently to be produced in SAGD. Produced Water-to-Oil Ratio (“PWOR”) is similar to SOR. This makes it difficult for SAGD to breach or utilize lean zone resources.
(5) The existence of an active water zone—either top water, bottom water or an interspersed lean zone within the pay zone—can cause operational difficulties or project failures for SAGD (Nexen Inc., “Second Quarter Results”, Aug. 4, 2011) (Vanderklippe, N., “Long Lake Project Hits Sticky Patch”, CTV News, 2011). Simulation studies concluded that increasing production well standoff distances can optimize SAGD performance with active bottom waters, including good pressure control to minimize water influx (Akram, F., “Reservoir Simulation Optimizes SAGD, American Oil and Gas Reporter, September 2010).
(6) Pressure targets cannot (always) be increased to improve SAGD productivity and SAGD economics. If the reservoir is “leaky”, as pressure is increased beyond native or hydrostatic pressures, the SAGD process can lose water or steam to zones outside the SAGD steam chamber. If fluids are lost, the Water Recycle Ratio (WRR) decreases, and the process requires significant water make-up volumes. If steam is also lost, process efficiency drops and SOR increases. Ultimately, if pressures are too high, if the reservoir is shallow, and if the high pressure is retained for too long, a surface breakthrough of steam, sand, and water can occur (Roche, P., “Beyond Steam”, New Tech. Mag., September 2011).

(7) Steam costs are considerable. If steam “costs” are over-the-fence for a utility including capital charges and some profits, the costs for high-quality steam at the sand face is about $10 to 15/MMBTU. High steam costs can reflect on resource quality limits and on ultimate recovery factors.

(8) Water use is significant. Assuming SOR=3, WRR=1, and a 90% yield of produced water treatment (i.e. recycle), a typical SAGD water use is 0.3 barrels (bbls) of make-up water per barrel (bbl) of bitumen produced.
(9) SAGD process efficiency is poor, and CO₂ emissions are significant. If SAGD efficiency is defined as [(bitumen energy)−(surface energy used)]/(bitumen energy), where 1) bitumen energy=6 MMBTU/bbl; 2) energy used at sand face=1 MMBTU/bbl bitumen (SOR ˜3); 3) steam is produced in a gas-fired boiler at 85% efficiency; 4) there are heat losses of 10% each in distribution to the well head and delivery from the well head to the sand face; 5) usable steam energy is 1000 BTU/lb (FIG. 3 PRIOR ART); and 6) boiler fuel is methane at 1000 BTU/SCF, then the SAGD process efficiency=75.5% and CO₂ emissions=0.077 tonnes/bbl bitumen.
(10) Practical steam distribution distance is limited to about 10 to 15 km (6 to 9 miles), due to heat losses, pressure losses, and the cost of insulated distribution steam pipes (Finan, A., “Integration of Nuclear Power . . . ”, MIT thesis, June 2007), (Energy Alberta Corp., “Nuclear Energy . . . ”, Canada Heavy Oil Association, pres., Nov. 2, 2006).
(11) Lastly, there is a natural hydraulic limit that restricts well lengths or well diameters and can override pressure targets for SAGD operations. FIG. 8 shows what can and has happened. In SAGD, a steam/liquid interface 12 is formed. For a good SAGD operation with sub-cool control, the interface is between the injector 2 and producer wells 4. The interface is tilted because of the pressure drop in the production well 4 due to fluid flow. There is little/no pressure differential in the steam/gas chamber. If the fluid production rates are too high (or if the production well is too small), the interface can be tilted so that the toe 14 of the steam injector is flooded and/or the heel 16 of the producer is exposed to steam 6 breakthrough (FIG. 8). This limitation can occur when the pressure drop in the production well 4 exceeds the hydrostatic head between steam injector 2 and fluid producer 4 (about 8 psi (50 kPa) for a 5 metre spacing).

In some cases, for deeper bitumen reservoirs, SAGD has the following issues:

(1) Hydrostatic and native reservoir pressures increase. The critical pressure for water/steam is 218 atm (3208 psia, 22 MPa (Table 3)). This corresponds (at 0.5 psi/ft hydrostatic gradient) to a hydrostatic depth of about 6416 feet or 1955 metres. Beyond this depth, a steam EOR process, at hydrostatic pressure, would need to use supercritical steam (FIGS. 3, 9).
(2) Because SAGD produced fluids, including water, are near saturated steam temperature, the SAGD process operates by delivering (net) latent heat to the reservoir. FIG. 9 shows steam latent and sensible heat content as a function of reservoir depth, assuming that SAGD operates with saturated steam pressures equal to hydrostatic reservoir pressures. FIG. 9 shows how SAGD becomes inefficient as depth increases. For example, if 1 MMBTU of latent heat per barrel of bitumen produced needs to be delivered, at 300 metres reservoir depth 1325 lbs. of steam is needed, while at 1500 metres in depth, 2778 lbs. of steam (a factor or more than two) is needed.
(3) Not only is more steam required as depth increases, but as pressure increases, the cost of steam generation and water treatment increases significantly (Smith (2005)).
(4) Heat losses in the vertical well bore section of the wells also increase significantly for two reasons—1) the pipe length and residence time of steam is increased and 2) steam temperature is increased. Vertical well bore heat losses can be a strong function of steam temperature (Radiation losses are proportional to T⁴).
(5) Capital expense (“Capex”) increases because the wells are longer, unit steam demand is increased, and steam/water capex increases with pressure.
(6) Operating expense (“Opex”) increases because SAGD efficiency drops, heat losses increase, and steam/water opex increases with pressure.

Other steam EOR processes (e.g. ISC SF, CSS . . . ) that don't rely totally on latent heat transfer can still work in reservoirs of increasing depth. But heat transfer via conduction and bitumen flow by flooding is slower and/or less efficient than steam condensation and/or gravity drainage.

For deep (>500 metres) heavy oil or bitumen resources where thermal EOR is the preferred recovery process and steam EOR is the perceived preferred process choice, heat losses from steam injection tubing have been a serious, long-time issue. The issue is complex with the following highlights:

• • (1) The traditional steam injection geometry is the use of centralized tubing to inject steam into the reservoir. Injection down the annulus has significantly higher heat losses. Heat losses from the tubing to the annular fluid to the casing and through the casing to the overburden can be significant. Heat losses can amount to over 20% for steady-state design conditions and much worse for non-design operations such as start-up (Herrera, J. O. et al, “Wellbore Heat Losses in Deep Steam Injection Wells,” The Society of Petroleum Engineers Regional Mtg., Apr. 12, 1978),
  • (2) Prior to start-up, for a simple completion without a packer, the annulus contains water. If saturated steam is injected, some steam condenses to provide for heat losses. But, as long as some steam survives at bottom hole, the saturated steam temperature remains constant. Currently there is no down hole instrument that can measure steam quality, so the operator cannot determine heat loss by simple down hole measurements (Satter, A. “Heat Losses during Flow of Steam Down a Wellbore,” The Journal of Petroleum Technology, July 1965). Water in the annulus is vaporized (contributing to slow start-up), so that most of the annulus will be filled with steam.
  • (3) Heat transfer for heat loss in this simple system is due to conduction, convection, and radiation from the outside of the steam tubular to the casing wall. Considering these mechanisms, early analysis showed that heat losses were dominated by radiation (Huygen, H. A. et al. “Wellbore Heat Losses and Leasing Temperatures during Steam Injection,” APRI meeting April 1966). Since radiative losses are proportional to surface area, larger sized steam tubing increases heat losses. Also at lower steam flows, fractional heat losses increase because of increased residence times. Radiation is a strong function of temperature (˜T⁴), so heat losses also increase with increased steam temperature and pressure.
  • (4) A simple solution to reduce heat losses was to paint the outside of the steam tubular with a low emissivity paint, such as aluminum paint (Huygen (1966)) (Pacheco, E. F et al. “Wellbore Heat Losses and Pressure Drop in Steam Injection,” The Journal of Petroleum Technology Feb. 12, 1972).
  • (5) This can reduce radiative heat losses by about a factor of two, but, unless the paint is applied in situ, it can easily be scraped off during installation and reduce the effectiveness considerably (Huygen (1966)) (Pacheco (1972)).
  • (6) Another way to reduce heat losses is to place a thermal packer downhole to isolate the annulus and steam tubulars. The packer can prevent steam from entering the annulus. Packers are best for cyclic steam processes (CSS) where the packer can reduce casing temperature increases and reduce the chance of casing failure. But, thermal packers are expensive and are known to leak (Satter, A. “Heat Losses During Flow of Steam Down a Wellbore,” Journal of Petroleum Technology July 1965) (Willhite, G. P. et al. “Wellbore Refluxing in Steam Injection Wells,” The Journal of Petroleum Technology March 1987).
  • (7) The simplest and most direct way to reduce heat losses is to insulate steam tubulars on the outside, reducing conduction losses and radiation heat losses. It has been shown that insulation can reduce heat losses by about ⅔ considering simple conduction, convection, and radiation mechanisms (Huygen (1966)). However, insulation is expensive. It may be difficult to install. Also, it can be ineffective if it is wet.
  • (8) So far, heat losses due to steam reflux in the annulus have been ignored, although this mechanism has been shown to be important (Willhite (1987)) (Satter (1964)). If steam is in the annulus, it can condense on casing walls, thus releasing latent heat as well as running downward by gravity until it finds a hot spot where it is re-vaporized. Hot spots can be uninsulated (or poorly insulated) gaps, couplings, and collars (etc.) in contact with the steam tubular. This reflux mechanism can cause heat losses three to six times higher than anticipated if only conduction, convection, and radiation are considered (Willhite (1987)).
  • (9) A potential solution to reduce reflux heat losses is to install a packer to isolate steam tubulars from the annulus and to fill the annulus with a (pressurized) gas (e.g. nitrogen) that has low heat conductivity. The purpose of the gas is to have a lower heat transfer than steam and to keep steam out of the annulus. But, packers have some leakage, and field experience shows that the annulus did not dry out for such systems. Water continued to reflux in the annulus (Willhite (1987)).
  • (10) A potential further solution is to continuously inject inert gas (e.g. nitrogen) into the annulus, with or without a packer or insulated tubing, to dry out the annulus. This can minimize annular water reflux heat losses and dry out insulation (if used).

Carbon dioxide (“CO₂”) is the primary non-condensable gas product of in situ combustion (excluding inert N₂ in air). CO₂ is partially soluble in reservoir fluids (oil and water). Deep heavy oil reservoirs have higher native pressures than shallow resources. For example, a 2000-metre deep reservoir has a hydrostatic pressure of about 3280 psi (22.5 MPa), while a 200-metre deep reservoir has a hydrostatic pressure of only 328 psi (2.3 MPa).

CO₂ solubility in reservoir fluids is not an important issue for shallow resources. But, increased pressure can significantly increase the impact of dissolved gases on thermal EOR processes. Solubility behaviour is not necessarily intuitive for dissolution into water. The normal expectation is that gas solubility in fluids (e.g. water) drops as temperature increase (FIGS. 10, 11, 12), based on the “soda-pop” visualization of gas dissolution in water. However, even at fixed pressures, as temperature is increased beyond about 200° F. (i.e. near the boiling point of water), gas solubility can increase substantially (FIG. 13). If pressures are also increased to follow the saturated pressure curve of water, gas solubility can increase even more with temperature.

FIGS. 10, 11, and 12 show that CO₂ is much more soluble in water than other gases by a factor of about 25 compared to methane. FIG. 14 shows that for elevated pressures and temperatures (above about 3000 psia or 20.5 MPa), CO₂ solubility in water is expected to be more than 160 SCF/bbl (Lake, L. W. “Enhanced Oil Recovery,” Prenctice Hall 1989). For deep thermal EOR processes involving CO₂ production, dissolution of CO₂ in water can be a significant sink.

CO₂ is also soluble in oil and dissolved CO₂ can significantly reduce oil viscosity to improve oil mobility. FIGS. 15, 16, 17, 18, 19 shows that CO₂ solubility in oils (including heavy oils) is expected to be about double the solubility in water (Issever, K. et al. “Use of CO₂ to Enhance Heavy Oil Recovery,” No. 1998. 14.1, 1998), (Bennion, D. B. et al “The Use of CO₂ as EOR Agent for Heavy Oil,” Joint Canadian Romania Heavy Oil Symposium, March 1993), (Lake (1989)). Again, for deep thermal EOR processes involving CO₂ production, dissolution of CO₂ in oil can be a significant sink for CO₂ (FIG. 20).

As discussed above, CO₂ dissolved in oil has the added benefit of reducing oil viscosity (FIG. 21). However, in order to get reasonable bitumen productivity, in situ viscosity has to be dropped to about 20 cp or about 5 orders of magnitude for the bitumen shown in FIG. 5 (Nexen (2011)). By itself, CO₂ dissolution can reduce bitumen or heavy oil viscosity by about one order of magnitude (FIG. 17).

FIG. 16 shows that the saturate (paraffinic) fractions of oil are more effective dissolving CO₂, so one would expect that paraffinic crudes would have the highest CO₂ solubility (Marufuzzanan, M. “Solubility and Diffusivity of Carbon Dioxide, Ethane, and Propane in Heavy Oil” University of Regina, M.A.Sc. Thesis November 2010).

FIG. 21 shows the combined effects of heat and CO₂ dissolution to reduce bitumen viscosity (Bennion (1993)). As temperature is increased, CO₂ solubility is decreased, so that viscosity reductions due to CO₂ decrease.

The combination of CO₂+steam for thermal EOR has also been evaluated using simulation models for CSS EOR (Balog (1982)). A mixture of 7% (v/v) CO₂ in steam was injected in a CSS process using a simulator and a Cold Lake Alberta reservoir. Compared to steam alone, the CO₂ increment increased bitumen productivity by 35%, and after 3 cumulative cycles, bitumen production increased by over 50%. Further, reservoir CO₂ inventory was established equal to about 2 MSCF/bbl bitumen produced. The inventory could either be blown down at project end or sequestered. (FIGS. 20 and 22). Reservoir pressure was about 1200 psi.

SUMMARY OF THE INVENTION

The current invention involves application and simplification of the SAGDOX process applied to deep, high-pressure bitumen reservoirs. Shallow (<500 metres) average depth reservoirs employing SAGDOX processes require separate removal of non-condensable combustion gases (mostly CO₂) using vent gas wells or segregated vent gas sites. But, for deep reservoirs, preferably having a depth average greater than about 500 metres in depth from surface level, the non-condensable vent gas generated by the SAGDOX process may be left to dissolve in the reservoir or production fluids, so that separate (non-condensable) gas removal is not necessary. In addition, CO₂ dissolution in bitumen can reduce viscosity and increase bitumen productivity.

According to one aspect, there is provided a process to recover hydrocarbons, from a hydrocarbon reservoir having a bottom, using a substantially horizontal production well, preferably said substantially horizontal production well has a toe and a heel, said process comprising:

• (1) injecting oxygen into said hydrocarbon reservoir, said horizontal production well having at least one perforation zone, for contact with said reservoir;
• (2) injecting steam into said hydrocarbon reservoir;
  said oxygen producing in situ heat and in situ carbon dioxide by combustion and said steam producing in situ heat by conduction and condensation; said in situ carbon dioxide dissolving into the liquid hydrocarbon, lowering its viscosity;
• (3) recovering said reservoir liquid hydrocarbon using said substantially horizontal production well; and
• (4) optionally conveying said recovered liquid hydrocarbon to the surface; wherein said process is absent a removal step of any non-condensable gas from said reservoir.

In one embodiment, said hydrocarbon reservoir comprises at least one characteristic selected from the group consisting of:

i) an average depth greater than about 500 metres;
ii) an average pressure greater than about 800 psia, and combinations thereof.

According to another aspect, there is provided a process to recover liquid hydrocarbons, from a hydrocarbon reservoir, using a horizontal production well, wherein:

• • (1) the horizontal production well is used to produce water and liquid hydrocarbons and is completed within 2 metres of the reservoir bottom;
  • (2) oxygen gas is injected into the hydrocarbon reservoir within 50 metres from the horizontal production well and with a perforation (reservoir contact) zone less than 50 metres in length;
  • (3) steam is injected within 20 metres from the horizontal production well;
  • (4) the oxygen gas produces in situ heat and in situ carbon dioxide by combustion and steam produces in situ heat by conduction and by condensation;
  • (5) carbon dioxide dissolves into the reservoir hydrocarbon, lowering its viscosity;
  • (6) in situ heat causes the liquid hydrocarbon to be heated, also lowering its viscosity;
  • (7) the lower-viscosity reservoir liquid hydrocarbon drains by gravity to the horizontal production well where it is conveyed (or pumped) to the surface; and
  • (8) the process pressure (in the reservoir) is greater than 800 psia.

In one embodiment, the oxygen to steam injected is controlled so that produced water to oil (v/v liquid) has a ratio greater than 0.5, preferably the ratio of produced water to oil (v/v liquid) is between 0.5 and 2.0.

In another embodiment, the hydrocarbon reservoir is positioned at least 500 metres below the ground surface.

In another embodiment, said steam is injected within 10 metres from the horizontal well, preferably said steam is injected using a parallel horizontal well in the same vertical plane as the horizontal production well and located about 3 metres to 8 metres above the well, more preferably said steam is injected into the reservoir using at least one substantially vertical well selected from the group consisting of a single well or a plurality of substantially vertical wells.

In one embodiment, said oxygen is injected into the reservoir using at least one well selected from the group consisting of a single substantially vertical well or a plurality of substantially vertical wells.

In one embodiment, said steam and oxygen are comingled on the surface and injected into the reservoir.

In another embodiment, said steam and oxygen are segregated, preferably using packers, and injected separately into the reservoir, preferably said steam and oxygen are segregated using concentric tubing and packers with steam in a central tubing of said concentric tubing surrounded by oxygen in an adjacent annulus and said oxygen is injected at a higher elevation of said steam injected into the reservoir.

In yet another embodiment, said steam and oxygen are injected into said reservoir using a single substantially vertical well, wherein said single substantially vertical well is completed within 50 metres from the toe of the horizontal production well.

According to one of the embodiments of the invention the pressures of the process are sufficient so that substantially no free CO₂ is produced in the liquid production well. More preferably, the operating in situ pressure and the ratio of oxygen/steam (v/v) are adjusted so there is substantially no free CO₂ gas found in the horizontal section of the horizontal production well.

These and other benefits of the invention will be apparent from the review of the illustrations, descriptions and the claims of the invention.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 illustrates the “traditional” SAGD geometry.

FIG. 2 schematically illustrates the SAGD life cycle.

FIG. 3 illustrates the properties of saturated steam.

FIG. 4 illustrates saturated steam conditions.

FIG. 5 illustrates bitumen viscosity vs. temperature.

FIG. 6 shows the equation for SAGD bitumen productivity.

FIG. 7 depicts residual bitumen in the steam swept zone.

FIG. 8 schematically illustrates the hydraulic limitations of the SAGD Process.

FIG. 9 illustrates the properties of saturated steam.

FIG. 10 illustrates Carbon monoxide solubility in water at different temperatures.

FIG. 11 illustrates Methane solubility in water at different temperatures.

FIG. 12 illustrates Carbon dioxide solubility in water at different temperatures.

FIG. 13 illustrates Nitrogen solubility in water at different temperatures.

FIG. 14 illustrates Carbon dioxide solubility in water at different pressures.

FIG. 15 illustrates Carbon dioxide solubility in bitumen at different pressures.

FIG. 16 illustrates Carbon dioxide solubility in heavy oil at different pressures.

FIG. 17 illustrates Carbon dioxide solubility and viscosity reduction.

FIG. 18 illustrates Carbon dioxide solubility in bitumen.

FIG. 19 illustrates Carbon dioxide solubility in oil.

FIG. 20 illustrates CSS retention of Carbon dioxide.

FIG. 21 illustrates bitumen viscosity dependent on Carbon dioxide.

FIG. 22 illustrates CSS using steam and Carbon dioxide.

FIG. 23 illustrates In Situ combustion minimum air flux rates.

FIG. 24 depicts SAGDOX mechanisms.

FIG. 25A,B,C depicts SWSAGDOX piping schemes using centralized packers.

FIG. 26 illustrates the connection of Oxygen requirements on peak temperature.

FIG. 27 illustrates correlation between Oxygen Pressure and Carbon gas content.

FIG. 28 illustrates a preferred SAGDOX geometry.

FIG. 29 illustrates a preferred SAGDOX geometry.

FIG. 30 illustrates the basic geometry of a SAGDOX process.

FIG. 31 illustrates the basic geometry of a SAGDOX process with packers.

FIG. 32 illustrates a preferred geometry of a SAGDOX process with packers on the injection and production wells.

FIG. 33 illustrates a first preferred embodiment of the Deep well SAGDOX geometry with packers on the injector well.

FIG. 34 illustrates a second preferred embodiment of the Deep well SAGDOX geometry.

FIG. 35 illustrates a combustion heat release chart.

DETAILED DESCRIPTION OF THE INVENTION

SAGDOX is an improved thermal enhanced oil recovery (EOR) process for bitumen recovery. The process can use geometry similar to SAGD (FIG. 31), but it also has versions with separate vertical wells or segregated sites for oxygen injection and/or non-condensable vent gas removal (FIGS. 28, 29, 30, and 32). The process can be considered as a hybrid SAGD+ISC process.

One objective of SAGDOX is to reduce reservoir energy injection costs, while maintaining good efficiency and productivity. Oxygen combustion produces in situ heat at a rate of about 480 BTU/SCF oxygen, independent of fuel combusted (FIG. 35 Butler (1991)). Combustion temperatures are independent of pressure and they are higher than saturated steam temperatures (FIGS. 3, 26). The higher temperature from combustion vaporizes connate water and refluxes some steam. Steam delivers EOR energy from latent heat released by condensation with a net value, including surface heat recovery of about 1000 BTU/lb. (FIG. 3).

Table 1 presents thermal properties of steam+oxygen mixtures. Per unit heat delivered to the reservoir, oxygen volumes are ten times less than steam, and oxygen costs including capital charges are one half to one third the cost of steam.

The recovery mechanisms are more complex for SAGDOX than for SAGD. The combustion zone is contained within the steam-swept zone 170. Residual bitumen, in the steam-swept zone 170, is heated, fractionated and pyrolyzed by hot combustion gases to produce coke that is the actual fuel for combustion. A gas chamber is formed containing steam combustion gases, vaporized connate water, and other gases (FIG. 24). The large gas chamber can be subdivided into a combustion-swept zone 100, a combustion-zone, a pyrolysis zone 120, a hot bitumen bank 130, a superheated steam zone 140 and a saturated steam zone 50 (FIG. 24). Condensed steam drains from the saturated steam zone 150 and from the ceiling and walls of the gas chamber. Hot bitumen drains from the ceiling and walls of the chamber and from the hot bitumen zone 130 at the edge of the combustion front 110 (FIG. 24). Condensed water and hot bitumen 8 are collected by the lower horizontal well 4 and conveyed (or pumped) to the surface (FIG. 30).

Combustion non-condensable gases are collected and removed by vent gas 22 wells or at segregated vent gas sites (FIGS. 29, 30, 31, 32). Process pressures can be controlled (partially) by vent gas 22 production, independent of fluid production rates. Vent gas 22 production can also be used to influence direction and rate of gas chamber growth.

Because SAGDOX delivers both steam and oxygen energy and oxygen gas has 10 times the energy density as steam (Table 1), pipe/tubing sizes for SAGDOX can be smaller (and less costly) than SAGD or other steam EOR processes. This can also reflect on production well sizes because reduced steam injection (with SAGDOX) results in less water production compared to SAGD.

Table 5 shows calculated pipe diameters for various SAGD and SAGDOX streams. Design criteria are presented in the table. When fluids use concentric pipe systems and annular flow, the total size of the combined pipe is indicated by brackets.

Often pipe costs are proportional to the diameter of the pipe. The total of pipe diameters can also be proportional to total costs. Table 5 shows total pipe diameters can be reduced by using SAGDOX and related processes.

Cumulative SAGDOX pipe diameters are 82% of SAGD for the case studied (35% oxygen in gas mix). THSAGDOX cumulative pipe diameters are 59% of SAGD, and SWSAGDOX cumulative diameter is only 42% of SAGD

Preferred parameters in SAGDOX geometries include:

(1) Use Oxygen (rather than air) as the oxidant injected

• • If the cost of treating vent gas to remove sulphur components and to recover volatile hydrocarbons is factored in, even at low pressures the all-in cost of oxygen is less than the cost of compressed air, per unit energy delivered to the reservoir.
  • Oxygen occupies about one fifth the volume compared to air for the same energy delivery. Well pipes/tubing is smaller and oxygen can be transported further distances from a central plant site.
  • In situ combustion (ISC) using oxygen produces mostly non-condensable CO₂, undiluted with nitrogen. CO₂ can dissolve in bitumen to improve productivity. Dissolution is maximized using oxygen.
  • Vent gas, using oxygen, is mostly CO₂ and may be used for sequestration.
  • There is a minimum oxygen flux to sustain HTO combustion (FIG. 23)
  • It is easier to attain/sustain this flux using oxygen
    (2) Keep oxygen injection at a concentrated site
  • Because of the minimum O₂ flux constraint from in situ combustion (FIG. 23), the oxygen injection well (or a segregated section) should have no more than 50 metres of contact with the reservoir
    (3) Segregate Oxygen and steam injectants, as much as possible
  • Condensed steam (hot water) and oxygen are very corrosive to carbon steel.
  • To minimize corrosion, either 1) oxygen 26 and steam 6 are injected separately (FIGS. 30, 31, and 32) 25); 2) comingled steam 6 and oxygen 26 have limited exposure to a section of pipe that can be a corrosion resistant alloy; 3) the section integrity is not critical to the process (FIG. 25(a)); or 4) the entire injection string is a corrosion resistant alloy (FIG. 25(a)).
    (4) The vent gas well (or site) is near the top of the reservoir, far from the oxygen injection site
  • Because of steam movement and condensation, non-condensable gas concentrates near the top of the gas chamber.
  • The vent gas well should be far from the oxygen injector to allow time/space for combustion.
    (5) Vent gas should not be produced with significant oxygen content
  • To mitigate explosions and to foster good oxygen utilization, any vent gas production with oxygen content greater than 5% (v/v) should be shut in.
    (6) Attain/retain a minimum amount of steam in the reservoir
  • Steam is added/injected with oxygen in SAGDOX because steam helps combustion. It preheats the reservoir so ignition, for HTO, can be spontaneous. It adds OH⁻ and H⁺ radicals the combustion zone to improve and stabilize combustion (FIGS. 26 and 27, Moore (1994)). This is also confirmed by the operation of smokeless flares, where steam is added to improve combustion and reduce smoke (Stone (2012), EPA (2012), Shore (1996)). The process to gasify fuel also adds steam to the partial combustor to minimize soot production (Berkowitz (1997)).
  • Steam also condenses and produces water that “covers” the horizontal production well and isolates it from gas or steam intrusion.
  • Steam condensate adds water to the production well to improve flow performance—water/bitumen emulsions—compared to bitumen alone.
  • Steam is also a superior heat transfer agent in the reservoir. When one compares hot combustion gases (mostly CO₂) to steam, the heat transfer advantages of steam are evident. For example, if one has a hot gas chamber at about 200° C. at the edges, the heat available from cooling combustion gases from 500° C. to 200° C. is about 16 BTU/SCF. The same volume of saturated steam contains 39 BTU/SCF of latent heat—more than twice the energy content of combustion gases. In addition, when hot combustion gases cool, they become effective insulators impeding further heat transfer. When steam condenses to deliver latent heat, it creates a transient low-pressure that draws in more steam—a heat pump, without the plumbing. The kinetics also favour steam/water. The heat conductivity of combustion gas is about 0.31 (mW/cmK) compared to the heat conductivity of water of about 6.8 (mW/cmK)—a factor of 20 higher. As a result of these factors, combustion (without steam) has issues of slow heat transfer and poor lateral growth. These issues may be mitigated by steam injection.
  • Since one can't measure the amount of steam in the reservoir, SAGDOX sets a steam minimum by a maximum oxygen/steam (v/v) ratio of 1.0 or alternately 50% (v/v) oxygen in the steam and oxygen mix.
    (7) Attain (or exceed) a minimum oxygen injection
  • Below about 5% (v/v) oxygen in the steam and oxygen mix, the combustion swept zone is small and the cost advantages of oxygen are minimal. At this level, only about a third of the energy injected is due to combustion.
    (8) Maximum oxygen injection
  • Within the constraints of (6) and (7) above, because per unit energy oxygen is less costly than steam, the lowest-cost option to produce bitumen is to maximize oxygen/steam ratios.
    (9) Use preferred SAGDOX geometries
  • Depending on the individual application, reservoir matrix properties, reservoir fluid properties, depth, net pay, pressure and location factors, there are three preferred geometries for SAGDOX (FIG. 28 a-c).
  • FIGS. 28b (THSAGDOX) and 28c (SWSAGDOX) are most preferred for thinner pay resources, with only one horizontal well required. Compared to SAGD, THSAGDOX and SWSAGDOX have a reduced well count and lower drilling costs. Also, internal tubulars and packers should be usable for multiple applications.

(10) Control/operate SAGDOX by:

• • Sub-cool control on fluid production rates where produced fluid temperature is compared to saturated steam temperature at reservoir pressure. This assumes that gases, immediately above the liquid/gas interface, are predominantly steam.
  • Adjust oxygen/steam ratios (v/v) to meet a target ratio, subject to a range limit of 0.05 to 1.00.
  • Adjust vent gas removal rates so that the gases are predominantly non-condensable gases, oxygen content is less than 5.0% (v/v), and to attain/maintain pressure targets.
  • Adjust steam and oxygen injection rates (subject to (ii) above), along with (iii) above, to attain/maintain pressure targets.

Preferred parameters in SAGDOX for deep reservoirs include the following:

1. Increased Pressures

For shallow reservoirs, because of the risk of fluid losses and the risk of surface blowouts, thermal EOR processes operate close to native reservoir pressures (Roche (2011). As reservoirs become deeper, there is less risk of surface blowouts, but fluid losses can still be an issue.

At a 0.5 psi/ft hydrostatic gradient, shallow reservoirs (200-300 m depth) have hydrostatic pressures of 330 to 490 psia (2.5 to 3.4 MPa). For deep reservoirs (500-2000 metres), hydrostatic pressures are much higher (820 to 3280 psia, 5.6 to 22.5 MPa).

If saturated steam is used (or is a component) and latent heat delivery is important (i.e. SAGD), there is an efficiency loss as pressure is increased (FIG. 9). For a pure steam process (i.e. SAGD), if 1 MMBTU of latent heat is needed, at 300 metres depth 1325 lbs. of steam is needed, while at 1500 metres depth 2778 lbs. steam is needed—a factor of more than two.

For SAGDOX, a mixture of steam 6 and oxygen 26 gas is injected (FIG. 29), or the mixture forms quickly in the reservoir (FIGS. 30, 31, 32). Post-combustion steam is diluted by a similar volume of CO₂. This has the immediate effect of reducing the partial pressure of steam in the reservoir and increasing the latent heat content of the steam fraction, so it can be a better heat transfer fluid using latent heat. This effect is partially mitigated because combustion will reflux some steam. Even in deep reservoirs, combustion temperature at 550° C. (FIG. 26) is significantly higher than steam (<370° C.).

2. CO₂ EOR

Carbon dioxide is produced as a result of in situ combustion. If oxygen gas is used, CO₂/O₂ ratios varying from about 0.85 to 0.96 are expected, depending on the fuel consumed and the reaction stoichiometry (Table 4). Some carbon monoxide may form, but it is likely to be converted to CO₂ in the reservoir (FIG. 10).

CO₂ will dissolve into bitumen to reduce its viscosity and increase bitumen productivity. By itself at high pressures (˜2000 psia), CO₂ can reduce bitumen (heavy oil) viscosity by about an order of magnitude (FIG. 17). If CO₂ dissolution is combined with heat, it can still contribute to bitumen viscosity reduction, particularly in the periphery of the reservoir where heat has not fully penetrated (FIG. 21).

3. Carbon Dioxide (Retention)

High operating pressures can drive gases (non-condensable gases) into solution in reservoir fluids (bitumen and water). Let's assume we use 1 MMBTU of combustion energy per bbl bitumen produced. Per MMBTU of combustion energy injected into the reservoir, 2083 SCF of oxygen is injected, and 1910 SCF CO₂ (in the worst case assuming a fuel consumed as CH_0.5 (Table 4)) is produced. If assume that produced fluids have a WOR=1.5 with some steam injection and some connate water production, CO₂ solubility in produced hot water is expected to be about 160 SCF/bbl hot water (FIG. 14, at ˜4000 psi.) (27 MPa), and CO₂ solubility in produced bitumen is expected to be about 200 SCF/bbl or more (FIGS. 15, 17). So, 1910 SCF CO₂ for 1 MMBTU of combustion energy is produced, and our produced fluids can remove about 440 SCF CO₂. This leaves 1470 SCF CO₂ that either resides in the gas chamber in the reservoir, or more likely dissolves in remaining reservoir fluids that are on the periphery of the gas chamber. FIG. 20 shows a CO₂ gas retention in the reservoir of about 1500 SCF/bbl bitumen produced (or more) for a lower pressure (˜2000 psia) process.

Based on the above example, no free CO₂ gas will be produced in the horizontal section of the production well. If free CO₂ gas is produced and if it is deemed harmful to the process, gas production can be reduced/eliminated by increasing SAGDOX reservoir pressures or by reducing O₂/steam ratios, eliminating the need for infrastructure venting non-condensable gases, in particular for venting CO₂ gas.

4. Heat Losses

As depth increases and saturated steam temperatures increase, heat losses from the vertical well sections to the overburden increase. As previously discussed, the optimum design to minimize heat losses, in this section, is to insulate the central steam injector with an annulus of continuously injected gas. For SAGDOX, this gas is oxygen, and the preferred designs are shown in FIGS. 29 and 32.

5. Carbon Monoxide

The combustion of in situ residual hydrocarbons and oxygen can produce a mixture of CO₂ and CO non-condensable gases (Table 4). Combustion tests (FIG. 27) show that CO can be produced in the combustion zone. Based on Le Chatelier's Principle, increased pressures should reduce CO formation, assuming some of the reaction steps are reversible. Also, in a reservoir (not a lab combustion tube), there is sufficient residence time and excess of steam (i.e. in SAGDOX) so that water-gas-shift reactions (CO+H₂O→CO₂+H₂) can occur, removing CO from the produced non-condensable gas mixture. Any hydrogen produced can then dissolve in bitumen and react with some components, so that for ISC projects CO is rarely seen in produced gases, and H₂ is even more rarely seen (Sarathi (1999)).

Nonetheless, the worst case CO production is about 140 SCF/MMBTU (Table 4). Assume CO solubility in water is similar to N₂ (FIG. 13) and Henry's law applies, at 600° F. (316° C.) and 2000 psia (13.7 MPa) the hot water solubility of CO in water is about 58 SCF/bbl water. As a worse case, assume cold-water solubility to be similar. So if 140 SCF of CO needs to be removed, 2.4 bbls of water needs to be contacted and saturated. This limit can be extended by changing Oxygen/steam ratios or by increasing system pressures.

In any case, undissolved CO should either be controllable or should not build up to levels that inhibit injectivity.

6. Deep SAGDOX Geometry

Because CO₂ need not be removed using separate vent wells or a segregated well section, the preferred geometry for deep SAGDOX processes can be simplified to three preferred cases:

• • the basic SAGDOX process with twin horizontal wells 2,4 can be simplified by removing the vent gas 22 annulus in FIG. 31.
  • similarly, the THSAGDOX version is also simplified as shown in FIG. 33
  • the SWSAGDOX version is also simplified as shown in FIG. 34.

Some of the differences between the prior art SAGDOX and the SAGDOX for deep reservoirs include:

• • SAGDOX has at least one vent gas well to remove non-condensable combustion gases; SAGDOX for deep reservoirs does not require same;
  • The target is deep, high pressure reservoirs (>500 m average depth or >800 psia average pressure); and
  • Non-condensable gas, preferably CO₂ is dissolved in bitumen; SAGDOX prefers venting of said gas.

Even further, SAGDOX for deep reservoirs allows for the following:

• • no vent wells required to remove non-condensable combustion gases;
  • pressure and oxygen/steam ratios can be adjusted, allowing for CO₂ being substantially dissolved in reservoir fluids
  • hydrocarbon recovery from deep reservoirs (>500 m average depth from surface)
  • hydrocarbon recovery from reservoirs with average pressure (>800 psia)
  • use of oxygen gas to insulate steam injector
  • two examples of preferred geometries as illustrated in FIGS. 33 and 34

TABLE 1
SAGDOX Injection Gases
	% (v/v) Oxygen in Steam + Oxygen Mixes
	0	5	9	31	50
% heat from O2	0	34.8	50.0	84.5	91.0
BTU/SCF mix	47.4	69.0	86.3	198.8	263.7
MSCF mix/MMBTU	21.1	14.5	11.6	5.0	3.8
MSCF O2/MMBTU	0.0	0.7	1.0	1.8	1.9
MSCF Steam/MMBTU	21.1	13.8	10.6	3.3	1.9
Where:
Steam @ 1000 BTU/lb.
Oxygen @ 480 BTU/SCF

TABLE 2
MW OF OILS
Elemental Composition of Oils
	Composition (wt %)		Atomic ratios
Source	C	H	S	Molecular Weight	H/C	S/C
Italy	81.7	11.4	5.6	520	1.67	0.026
	81.1	11.3	5.9	506	1.67	0.027
	81.0	10.9	3.3	398	1.61	0.015
	84.1	12.7	3.1	550	1.81	0.014
	82.3	12.5	2.3	600	1.82	0.010
	74.3	9.6	3.9	335	1.55	0.020
	78.0	10.6	3.1	388	1.63	0.015
	83.7	12.1	1.4	428	1.73	0.006
	87.3	11.7	0.2	293	1.76	0.001
	86.4	12.8	—	270	1.78	—
	86.9	12.3	0.3	329	1.70	0.001
	86.7	13.1	—	263	1.81	—
	85.2	13.1	0.1	265	1.85	0.001
	85.7	13.1	—	305	1.83	—
	87.5	12.1	0.1	318	1.66	0.001
	86.1	13.2	0.3	336	1.84	0.001
	85.2	13.9	0.2	210	1.96	0.001
	87.6	11.4	0.3	300	1.56	0.001
	86.1	13.0	0.2	295	1.81	0.001
	84.4	11.0	4.2	380	1.56	0.019
	83.8	12.5	2.6	286	1.79	0.012
	84.8	12.3	1.4	295	1.74	0.006
Kuwait	83.7	11.6	3.9	—	1.66	0.017
Venezuela	82.8	11.7	2.3	—	1.70	0.010
Speight (1991)
Avg. M.W. = 358

TABLE 3
Critical Properties of Reservoir Gases
	Tc (° C.)	Pc (atm)	Gc (g/cm3)
	Air	−140.6	37.2	0.313
	Co2	31.04	72.85	0.468
	CO	−140.23	34.53	0.301
	CH4	−82.60	45.44	0.162
	O2	−118.38	50.14	0.419
	H2	−239.91	12.80	0.031
	Ar	−122.44	48.00	0.5307
	N2	−146.89	33.54	0.311
	H2S	100.4	88.9	0.310
	C2H6	32.28	48.16	0.203
	H2O	374.2	218.3	0.325
	Where: Lange, “Handbook of Chemistry”, McGraw Hill, 1973

As many changes therefore may be made to the embodiments of the invention without departing from the scope thereof. It is considered that all matter contained herein be considered illustrative of the invention and not in a limiting sense.

Claims

1. A process to recover hydrocarbons, from a hydrocarbon reservoir having a bottom, using a substantially horizontal production well, said substantially horizontal production well having a toe and a heel, said process comprising:

(a) injecting oxygen into said hydrocarbon reservoir, said horizontal production well having at least one perforation zone for contact with said reservoir;

(b) injecting steam into said hydrocarbon reservoir; said oxygen producing in situ heat and in situ carbon dioxide by combustion and said steam producing in situ heat by conduction and condensation; said in situ carbon dioxide dissolving into the liquid hydrocarbon, lowering its viscosity;

(c) recovering said reservoir liquid hydrocarbons of lowered viscosity using said substantially horizontal production well; and

(d) optionally conveying said recovered liquid hydrocarbons to the surface; wherein said process is absent a removal step of any non-condensable gas from said reservoir.

2. The process of claim 2 wherein said hydrocarbon reservoir comprises at least one characteristic selected from the group consisting of:

(a) an average depth greater than about 500 metres;

(b) an average pressure greater than about 800 psia, and combinations thereof.

3. A process to recover liquid hydrocarbons, from a hydrocarbon reservoir having a bottom, using a horizontal production well, comprising:

(a) Completing the horizontal production well within 2 metres of the reservoir bottom;

(b) injecting oxygen gas into the hydrocarbon reservoir within 50 metres from the horizontal production well and with a perforation (reservoir contact) zone less than 50 metres in length;

(c) injecting steam within 20 metres from the horizontal production well;

(d) the oxygen gas producing in situ heat and in situ carbon dioxide by combustion and the steam producing in situ heat by conduction and by condensation;

(e) said carbon dioxide dissolving into the reservoir hydrocarbon, lowering its viscosity;

(f) in situ heat causing the liquid hydrocarbon to be heated, also lowering its viscosity;

(g) the lower-viscosity reservoir liquid hydrocarbon draining by gravity to the horizontal production well where it is conveyed (or pumped) to the surface; and

(h) the process pressure (in the reservoir) is greater than 800 psia.

4. A process according to claim 1 or 3, wherein oxygen to steam injected is controlled so that produced water to oil (v/v liquid) has a ratio greater than 0.5.

5. A process according to claim 4, wherein the ratio of produced water to oil (v/v liquid) is between 0.5 and 2.0.

6. A process according to claim 3, wherein said hydrocarbon reservoir is positioned at least 500 metres below a ground surface.

7. A process according to claim 3, wherein steam is injected within 10 metres from the horizontal production well.

8. A process according to claim 1 or 7, wherein steam is injected using a substantially parallel horizontal well to the horizontal production well and located about 3 metres to 8 metres above the horizontal production well.

9. A process according to claim 1 or 7, wherein steam is injected into the reservoir using at least one well selected from the group consisting of a single substantially vertical well or a plurality of substantially vertical wells.

10. A process according to claim 1, wherein the oxygen is injected into the reservoir using at least one well selected from the group consisting of a single substantially vertical well or a plurality of substantially vertical wells.

11. A process according to claim 9, wherein steam and oxygen are comingled on the surface and injected into the reservoir.

12. A process according to claim 10, wherein steam and oxygen are segregated and injected separately into the reservoir.

13. A process according to claim 12 wherein said steam and oxygen are segregated using concentric tubing and packers with steam in a central tubing of said concentric tubing surrounded by oxygen in an adjacent annulus and said oxygen is injected at a higher elevation of said steam injected into the reservoir.

14. A process according to claim 13, wherein said steam and oxygen are injected into said reservoir using a single substantially vertical well, wherein said single substantially vertical well is completed within 50 metres from the toe of the horizontal production well.

15. A process according to claim 1 or 3 wherein said reservoir has a pressure such that substantially no free CO2 is produced in the production well.

16. A process according to claim 1 or 3 wherein said process has an operating in situ pressure and an oxygen/steam (v/v) ratio such that the substantially horizontal production well is substantially free of free CO2.