Content of PetroWiki is intended for personal use only and to supplement, not replace, engineering judgment. SPE disclaims any and all liability for your use of such content. More information
Thermal recovery by steam injection
The most common method used to enhance oil production over primary rates is water injection, commonly referred to as secondary oil recovery. Common practice in the industry is to refer to all other methods as tertiary enhanced oil recovery. According to Prats, thermal enhanced oil recovery (TEOR) is a family of tertiary processes defined as "any process in which heat is introduced intentionally into a subsurface accumulation of organic compounds for the purpose of recovering fuels through wells." This article provides an introduction to the mechanisms by which steam can enhance oil recovery.
The most common vehicle used to inject heat is saturated steam. Hot water and heated gasses have been tried, but none are as effective as quality steam. According to a 2000 Oil & Gas Journal survey, steam enhanced oil recovery (EOR) projects accounted for 417,675 barrels of oil per day (BOPD), or 56% of the total for all tertiary enhanced recovery methods. That production rate has been essentially flat for more than 15 years. Hydrocarbon gas injection and CO2 gas injection are the only other significant contributors and amount to only 17 and 24%, respectively.
Properties of saturated steam
Like other substances, water can exist in the form of a solid (ice), as a liquid (water), or as a gas (loosely called steam). Steamflood processes are concerned with the liquid and gas phases, and the change from one phase to the other. The phase change region, in which water coexists as liquid and gas, is where our interest lies when considering steam for use in the oil field. The term "steam" is an imprecise designation because it refers to a water liquid/gas system that can exist from 32°F to any higher temperature; from 0.1 psia to any higher pressure; and from nearly all liquid to 100% gas. Steam quality refers to the phase change region of liquid to gas and is defined as
Heat capacity is expressed in units of Btu/(lbm-°F). A "Btu" is defined as the amount of heat required to raise 1 lbm of water from 60 to 61°F. All liquids and solids are compared to pure water, which has the highest heat capacity of any substance at 1 Btu/(lbm-°F). By calculating a ratio of the heat capacity of water divided by that of another substance, a convenient fraction called "specific heat" is obtained. Notice that petroleum has a specific heat of 0.5, or half that of water, and sandstone is only 20% of water on a per pound basis. No other liquid or gas carries as much heat per pound as water. Also, the temperature range at which this high heat carrying performance is achieved, 34 to 700°F, is ideal for many processes, including steamflooding.
Enthalpy is a useful property defined by an arbitrary combination of other properties and is not a true form of energy. The absolute value of enthalpy is of no practical value. Changes in enthalpy are extremely useful, however, and are the basis for steamflood energy calculations. The total enthalpy held by each pound of liquid water at any temperature is called sensible heat, hf. The heat input, which produces a change of state from liquid to gas without a change of temperature, is called the "latent heat of evaporation" and is shown by hfv. The total heat, hv, in each pound of 100% quality or saturated steam is the sum of these two, hv = hf + hfv.
In the phase change or "saturation" region, steam can only exist at one temperature for a given pressure regardless of quality or latent heat content, as shown in Fig. 1 . Steam increases in volume as latent heat increases, as in Fig. 2. This is a useful property in displacing oil in a steamflood process. The volume occupied by 1 lbm of steam at any pressure is its specific volume in ft3/lbm and is represented by υs. Values for these thermal properties of water are published widely. Fig. 3 is a chart of the phase change region. Equations have been derived that approximate the values to acceptable accuracy for most steamflood calculations.
Simple versions that are accurate to within a few percent in the normal pressure ranges encountered in steamflooding projects are
These are recommended for hand calculations or simple analytical equations. There are more precise versions in the literature, but for most purposes, Eqs. 2 through 8 are more accurate than any other available information that goes into steamflood calculations and are more than adequate for most calculations. The more accurate equations are cumbersome and lend themselves to computer applications. However, in a computer application, a lookup table is easy to create and is much more accurate than even the advanced analytical equations.
The basis for reservoir heat transfer calculations is traced back to the published solution to a mechanical engineering problem. Marx and Langenheim were the first to publish an adaptation of this solution to a steamflood process. They assumed that the equations for temperature response in a thin plate, backed in perfect contact to a semi-infinite solid after sudden exposure to constant-heat input, were analogous to steam injection into an oil-bearing reservoir.
Fig. 4 shows the temperature distribution anticipated in this model. The entire flood zone is at steam temperature upstream of the steam front, and the reservoir downstream of the steam front is at initial reservoir temperature. The true temperature profile is much less defined.
Fig. 4 - Temperature profile resulting from convection and conduction caused by piston displacement of steam (after Marx and Langenheim).
Their equation for the heated area, At, over time, t, is
G(tD) from the mechanical engineering problem is a function of dimensionless time, tD.
where tD is time of injection, t, multiplied by a few reservoir properties.
and the complementary error function, erfc(x), is common in heat conduction calculations. Several publications have tables of values for G(tD) vs. time, but van Lookeren offers a simple equation with sufficient accuracy for most calculations, which is written as
Other useful relationships derived from the Marx and Langenheim equations are heat loss to the adjacent formations,
rate of heated zone growth,
heat remaining in the reservoir,
cumulative heat loss to the adjacent formations,
and reservoir efficiency or fraction of injected heat remaining in the reservoir,
Two constants appear in the equations that need definitions. MR is the volumetric heat capacity of the composite formation including rock and fluids.
Note that there are two gas components: inert gas represented by the volumetric heat capacity, Mg, and steam represented by two terms, one owing to the latent heat of vaporization and one to the sensible heat.
Thermal diffusivity, α, is the ratio of the thermal conductivity to the volumetric heat capacity,
Farouq Ali showed it is appropriate to use average values for M and α when the thermal properties of various layers of formation and over/underburden differ significantly.
Ramey and Prats showed that Eq. 17, reservoir heating efficiency, Eh, is independent of reservoir and steam zone geometry. Further the relationships apply to any form of heat transport (convection, conduction, or even radiation) in the plane of the reservoir and when the heat transport to adjacent formations is three-dimensional (3D). Assuming the reservoir properties are constant and vary little with changing temperature, Eh is solely a function of dimensionless time.
Steam zone growth
Reservoir heating by steam injection translates into the formation of a swept zone of some shape and an oil bank consisting of a migrating zone containing the displaced oil. Two general concepts have been proposed to describe this process.
This is an extension of early water flooding concepts in which the displacement front is considered to be an advancing vertical front, the plane of which is normal to the bedding plane and extending vertically from the top to the bottom of a homogeneous flow section. Displacement of oil in this model is piston-like and is directly proportional to injection rate. The equations in the Reservoir Heating section, have been used to calculate steamfloods as frontal advance floods.
Mandl and Volek, followed by a slight modification by Myhill and Stegemeier, contributed the concept of a critical time beyond which the vapor front stagnates and heat is carried only by hot water traveling through the condensation front. Dimensionless critical time, tcD, is defined by
where fhv is the fraction of heat injected as latent heat.
and fs and hfv are at reservoir conditions.
The Marx and Langenheim equation (Eq. 10) is an upper bound to Eq. 22, as shown in Fig. 5.
Fig. 5 – Fraction of injected steam remaining in steam zone.
These models address the severe buoyancy forces that pertain when steam vapor, a gas, is injected into a liquid filled reservoir. In bypass models, the fronts are not vertical; the steam migrates quickly to the top of the reservoir. Van Lookeren describes the severity of this override with a buoyancy factor.
Average steam zone thickness is
The shape of the steam zone is described by
Note that steam zone thickness is not dependent on net zone thickness in Eqs. 24 and 25.
Fig. 6 shows the calculated injector liquid level as a function of time and five values of ARD. ARD is normally smaller than 0.6 in field projects; thus, the steam zone does not cover the vertical zone except for very thin reservoirs.
Fig. 6 – Effect of the ratio of viscous to gravity forces on the location of steam displacement fronts.
Neuman addresses the bypass or steam override concept by basically assuming that injected steam quickly rises to the top of the zone and spreads out evenly. Thickness of the steam zone, hs, is related to
Field projects usually include features of both frontal advance and bypass processes. Owens and Ziegler propose an analytical model that calculates the contribution of each process. They calculate total oil rate, qo, as the sum of gravity drainage oil rate, qog, after Clossman, and viscous oil rate, qov, as
This series of equations can be solved in a computer spreadsheet. The authors report that the viscous component of oil rate in the Kern River Ten-Pattern is significant and varied from 66% early in the project to 53% later. The method is also used to show the effect of operating with (supporting qov) and without (inhibiting qov) a casing vapor recovery system. Fig. 7 shows the impact of shutting in the casing vapor recovery on a selected project. This is not to say that shutting in casing vapor recovery will have this profound negative impact on every steamflood. Notice that this method does not account for steam injection rate. A project that is significantly over-injected (high-pressure drawdown) would be expected to respond as in Fig. 7, while a project with more controlled steam rate (modest-pressure drawdown) may show little, if any, reduction in production rate.
Fig. 7 – Effect of shutting in casing vapor recovery system on a selected steamflood project.
Kimber et al. found with a physical model that there is a viscous component of steamflooding that has potential to add significant oil recovery. Fig. 8 shows that there is an optimal steam quality for a steamflood and that some amount of overinjection may be beneficial also. Ward and Sharpe studied the subject at the Kern River field using a numerical simulator and found that shutting in the casing vents resulted in a permanent 18% loss in oil production rate as shown in Fig. 9. Oil production could be restored if either the vents were reopened or a depleted zone in the wellbore were perforated, allowing the excess steam to vent by crossflow.
Fig. 8 – The effect of latent heat injection rate on strategy economics.
Fig. 9 – Results of a numerical simulation showing 18% loss in production resulting from shutting in casing vapor recovery.
- Prats, M. 1982. Thermal Recovery, No. 7. Richardson, Texas: Monograph Series, SPE.
- Moritis, G. 2000. EOR Weathers Low Oil Prices. Oil & Gas J. 98 (12): 39.
- Keenan, J.H. et al. 1969. Steam Tables—Thermodynamic Properties of Water Including Vapor, Liquid, and Solid Phases, 2. New York City: John Wiley & Sons Inc.
- Chien, S.-F. 1992. Empirical Correlations of Saturated Steam Properties. SPE Res Eng 7 (2): 295-303. SPE-20319-PA. http://dx.doi.org/10.2118/20319-PA.
- Carslaw, H.S. and Jaeger, J.C. 1950. Conduction of Heat in Solids, 373. Amen House, London: Oxford U. Press.
- Marx, J.W. and Langenheim, R.H. 1959. Reservoir Heating by Hot Fluid Injection. In Trans., AIME, 216, 312–314.
- van Lookeren, J. 1983. Calculation Methods for Linear and Radial Steam Flow in Oil Reservoirs. SPE J. 23 (3): 427–439. SPE-6788-PA. http://dx.doi.org/10.2118/6788-PA.
- Farouq Ali, S.M. 1970. Oil Recovery by Steam Injection, 51. Bradford, Pennsylvania: Producers Publishing Co.
- Ramey, H.J. Jr. 1959. Discussion of Reservoir Heating by Hot Fluid Injection. In Trans., AIME, 216, 364.
- Prats, M. 1969. The Heat Efficiency of Thermal Recovery Processes. J Pet Technol 21 (3): 323-332. SPE-2211-PA. http://dx.doi.org/10.2118/2211-PA.
- Mandl, G. and Volek, C.W. 1969. Heat and Mass Transport in Steam-Drive Processes. SPE J. 9 (1): 59-79. SPE-2049-PA.
- Myhill, N.A. and Stegemeier, G.L. 1978. Steam-Drive Correlation and Prediction. J Pet Technol 30 (2): 173-182. SPE-5572-PA. http://dx.doi.org/10.2118/5572-PA
- Neuman, C.H. 1985. A Gravity Override Model of Steamdrive. J Pet Technol 37 (1): 163-169. SPE-13348-PA. http://dx.doi.org/10.2118/13348-PA
- Owens, B.K. and Ziegler, V.M. 1995. An Oil Production Model for a Well Producing by Both Gravity Drainage and Viscous Flow From a Mature Steamflood. Presented at the SPE Western Regional Meeting, Bakersfield, California, 8-10 March 1995. SPE-29656-MS. http://dx.doi.org/10.2118/29656-M
- Closmann, P.J. 1995. A Simplified Gravity-Drainage Oil-Production Model for Mature Steamfloods. SPE Res Eng 10 (2): 143-148. SPE-25790-PA. http://dx.doi.org/10.2118/25790-PA
- Kimber, K.D., Emerson, G.G., Luce, T.H. et al. 1995. The Role of Latent Heat in Heat Management of Mature Steamfloods. Presented at the SPE Western Regional Meeting, Bakersfield, California, 8-10 March 1995. SPE-29659-MS. http://dx.doi.org/10.2118/29659-MS
- Ward, R.C. and Sharpe, H.N. 1997. Mechanistic Modeling of Casing Blow Shut-In Effects and Strategies for the Recovery of Lost Oil Production. Presented at the International Thermal Operations and Heavy Oil Symposium, Bakersfield, California, 10-12 February 1997. SPE-37561-MS. http://dx.doi.org/10.2118/37561-MS
Noteworthy papers in OnePetro
Morrow, A. W., Mukhametshina, A., Aleksandrov, D., & Hascakir, B. 2014. Environmental Impact of Bitumen Extraction with Thermal Recovery. Society of Petroleum Engineers. http://dx.doi.org/10.2118/170066-MS
Mukhametshina, A., Morrow, A. W., Aleksandrov, D., & Hascakir, B. 2014. Evaluation of Four Thermal Recovery Methods for Bitumen Extraction. Society of Petroleum Engineers. http://dx.doi.org/10.2118/169543-MS
Mukhametshina, Albina, & Hascakir, B. 2014. Bitumen Extraction by Expanding Solvent-Steam Assisted Gravity Drainage (ES-SAGD) with Asphaltene Solvents and Non-Solvents. Society of Petroleum Engineers. http://dx.doi.org/10.2118/170013-MS
Use this section to provide links to relevant material on websites other than PetroWiki and OnePetro