GRAVITY DRAINAGE VITAL FACTOR FOR UNDERSTANDING STEAM FLOODS

Jack Y. Vogel
Consultant Houston

Gravity drainage of hot viscosity-reduced oil is a dominant recovery mechanism that should be considered in steam flood design, heat calculations, and field operations. This fact has significant and sometimes surprising implications.

Important considerations include the benefits of early steam breakthrough, need to control steam production, steam stimulation of production wells, and declining rate steam injection.

VISCOUS FORCE

Traditional steam flood calculations (Fig. 1) are based on oil displacement by steam pressure. These methods can compute the growth of steam zones and although useful are essentially pure heat balances that fail to account for viscous forces and pressure drops in the cold oil zone ahead of the steam front.

One serious difficulty'-' then arises with the frequently made assumption that the rate calculated for steam zone expansion provides the rate of oil displacement ahead of the steam.

In a Midway Sunset project, for example, oil displacement rates calculated by Marx and Langenheim methods 4 for an expanding 500 psia steam zone would require a totally unrealistic pressure drop of 9,500 psi across the cold oil zone.

Similar results can be obtained for almost any commercial steam flood with published data.

These calculations demonstrate that heavy cold viscous oil is practically immobile. Therefore, pressure drive of cold oil does not adequately explain the recovery mechanism.

For this reason, among others, operators now commonly base designs on computer simulations and/or field experience in similar reservoirs. New approaches to predict producing rates for the case of complete steam overlay have also been published .5-1

However, there are still frequently used assumptions that were derived from experience with drive or displacement processes such as waterfloods. Some of these assumptions should not be applied to drainage processes such as steam floods.

EARLY BREAKTHROUGH

One assumption is that early steam breakthrough is detrimental to a steam flood.

Instead of forming a vertical or near vertical front, injected steam generally rises to the top and spreads as a thin layer across the reservoir. The result frequently is early steam breakthrough at the producing wells.

This has led to a concern that much oil will be bypassed. However, field case studies. 9-17 show that sustained oil response occurs only after the injected heat is at or near the producing well. Furthermore, contrary to assumptions with frontal displacement, most of the oil recovery reported in field case histories is after heat breakthrough rather than before.

Reported steam flood recovery efficiencies are high, usually better than 50%.18 Thus, oil bypassing is not a problem except in thin separated oil layers where one layer preferentially takes the injected steam.

The conclusion is that early steam breakthrough generally should be considered beneficial, rather than detrimental to the steam flood process because:

Frontal displacement of viscous oil would require enormous steam pressures.
Sustained oil response requires that producing wells be in heat communication with the injectors.
Excellent recovery efficiencies are obtained after steam reaches the producers.

In fact, deliberate promotion of early breakthrough, as recommended by Farouq Ali,18 can reduce the time required to bring a steam flood up to peak production.

Two methods to encourage early heat communication are: large initial steam stimulations in producing wells and initial high-rate steam injection in the injectors. 16

GRAVITY DRAINAGE

Instead of considering a steam flood as a displacement process, the steam flood should be visualized as a melting process. Heat is supplied overhead and carried downward by conduction and convection (Fig. 2).

As heat penetrates the cold, viscous oil, a thin layer of hot, low-viscosity oil is created. This oil drains by gravity down to producing wells. The process is aided somewhat by steam drag.

A pertinent quotation from Myhill and Stegemeier is: "Often, a more-or-less continuous steam layer then spreads across the reservoir, and heated oil flows to the wells as a result of gravity drainage and steam drag. In high-permeability reservoirs, little dip or sand thickness is required to make gravity drainage the dominant mechanism." 17

A simulation for a 1.25 acre zero dip pattern in Canada's Athabasca oil sands illustrates this condition.

Tables 1 and 2 list the simulation input. After 3 years of injection:

Oil production rate was 43 b/d and declining.
Steam injection was 130 b/d (cold water equivalent).
A 25-ft thick steam layer had formed at the top.

As shown in Fig. 3, the computer calculated only a 0.5 psi pressure drop (steam drag force) across the entire top of the reservoir from injector to producer. In comparison, the gravity head available from the 45-ft oil column was equivalent to 17.3 psi, or 35 times more.

The gravity forces in a dipping reservoir would be even more predominant.

Although the gravity forces in a steam flood are larger than the drag forces, the gravity forces are still very small compared to the pressure drawdowns available in most conventionally produced oil fields. As a result, good permeability and thick or steeply dipping reservoirs are generally required for attaining commercial rates.

Where this is the case, high recovery efficiencies can occur even in nonthermal production. Some gravity-drained zones have efficiencies better than 50% and 19-23 as high as 85%.

Gravity drainage probably accounts for the excellent recovery efficiencies reported for most steam floods. The acceptance of gravity as the drive mechanism avoids the need for the conjecture generally offered with the frontal displacement concept. This conjecture says that good steam flood recovery efficiencies result from:

Gas drive by many reservoir volumes of steam *Efficient miscible drive by light ends distilled by the steam and condensing at the heat front.

UNCONTROLLED STEM

To meet environmental requirements and reduce the back pressure against the formation, many steam floods include a casing vapor-recovery system that gathers steam and noncondensible gases from the production well casings.

These systems (Fig. 4) should be operated cautiously. A minimum back pressure is not always desirable. While early heat communication is recommended, uncontrolled steam production is detrimental.

During the flood's early life while the steam layer is thin, a little steam flow (about 2-20 b/d/well) may help production.

But after substantial steam encroachment has reached the production wells, the steam flood is essentially in a gravity drainage mode. Attempts to increase oil production by decreasing the casing vapor pressure have no effect on the liquid gravity head that provides the dominant force contributing to oil production.

Instead, the pressure affects only the small steam drag forces that contribute little to oil production. Significant increases in the drag force would require that uneconomic amounts of steam be wasted.

Edmonds performed simulations to investigate recovery by steam drag alone (gravity forces set to zero) and concluded that steam drag could provide very good recovery but only by using astronomical amounts of stream. 24

Steam blow capacities as high as 100-200 b/d were measured in some Kern River wells with only 25 psig steam pressure in the reservoir.' Other reports indicate more typical steam production potentials of 50-75 b/d at Kern River and San Ardo. 26 27

The heat content of steam gathered from the casing is generally transferred to the atmosphere. Although the steam passed through the reservoir, the steam did not heat the reservoir.

Also to note is that steam produced up the casing annulus is almost 100% in quality 27 and thus the heat value of 100 bbl produced steam is equivalent to about 130 or more bbl of the lower quality steam commonly injected.

Because California steam flood wells now average less than 25 bo/d, 28 it is obvious that uncontrolled steam production, additive to the steam required to heat the reservoir, can lead to uneconomic steam/oil ratios.

With fuel costs around $2/Mcf, unnecessary steam production can increase production costs by as much as $2-3/bbl of oil. The higher costs cause wells to reach an early economic limit with attendant loss of reserves.

Condensible hydrocarbon vapors, about 3-5% by liquid volume, accompany the produced steam." These may be condensed and sold. This may appear as offering an additional source of income to offset the cost of the wasted steam, but this hydrocarbon production would not be lost if steam production is controlled at lower rates.

If not produced through the casing, the light hydrocarbons will stay in solution and be produced with the heavy oil.

A simple and low-cost method to control steam rates from a producing well is to install a small critical flow orifice on the casing. The orifice is sized to limit steam production to any desired limit.

These orifices will increase back pressure against the well, but this will not affect the gravity component of flow.

Reduction of steam production can also help decrease downhole problems with sanding and cutting of tubulars as well as surface environmental costs for disposal of associated contaminant gases.

STEAM STIMULATION

Because gravity provides the dominant force for oil recovery, it is important to maximize the gravity head.

Downhole pump intakes should be set at or below the producing interval. The pumps should have sufficient capacity to keep the wells pumped off. Otherwise, the liquid level inside the well bore will reduce the effective gravity head.

These measures are not always sufficient to provide maximum producing rates, As shown in Fig. 5, unless the reservoir around the producing well is hot from top to bottom, a barrier of cold viscous oil may reduce the effective gravity head on the hot oil at the top.

To heat the whole reservoir interval, frequently an effective remedy is to inject a moderate amount of steam into the producer.

Sometimes the top steam zone acts as a thief that takes most of the steam. This theft prevents adequately heating the lower portion of the well. Some success has been reported using packers, chemical foams, and gels to divert steam to the lower part of the well bore.

Where good vertical permeability permits, another effective measure is to complete the well in only the lower art of the producing p 1619272930 Even here interval .

steam stimulation may be required 31 if, as shown in Fig. 6, cold viscous oil at or above the open interval prevents hot oil from draining from the top.

Steam stimulation can also be needed if the injected steam spreads in an irregular pattern because of reservoir heterogeneities, directional permeabilities, or dip effects. Cold areas can then prevent one or more wells from responding to the steam flood.

In such cases, 121631 the injection of large volumes of steam in the producing well has often succeeded in linking the cold producers with the existing steam chests (Fig. 7). Sometimes more than one of these stimulations is required before sustained steam flood response is obtained.

Liberal use of producer steam stimulations for the cases discussed, and also for occasional small volume warm ups, can increase steam flood producing rates. This recommendation would seem to require close analysis to ensure the cost effectiveness of extra steam. But in a gravity-dominated system, the steam for stimulation should not be considered extra or additive to the continuous injection.

In a successful steam flood, substantially all of the reservoir volume must ultimately be heated. In particular this includes the region around the producing well, from top to bottom of the vertical thickness.

As previously noted, it can be advantageous to heat this region with steam injection into the producing wells rather than to rely solely on heat from the continuous injectors.

If more steam than needed is injected into the producer, the surplus steam just enters the main steam chest and does the same job of heating as if it had been injected into a continuous injection well in the first place.

Accordingly, volumes of steam used for stimulating producing wells may be deducted from the amount that would otherwise be used for continuous injection.

STEAM INJECTION RATE

Heat requirement calculations for the frontal displacement case are not applicable to the overlay case. For the overlay case, rates of steam injection that diminish with time are preferable 32 to constant rate injection.

Assuming due allowance for surface losses, well bore losses, and heat produced back to the surface, the following equation (see nomenclature box) can be used for calculating diminishing injection rates that can satisfy formation heating and provide for conductive losses above and below the steam zone.

[SEE FORMULA]

As further support for the equation, Fig. 8 shows close agreement between steam requirements calculated with this equation and those from a computer simulation using input data from Tables I and 2.

In using this equation for initial steam flood design, one must estimate future oil producing rates as indicated by the dashed line in Fig. S. This is not necessary when the equation is used to optimize and adjust steam injection rates in a mature steam flood. The actual oil producing rates can then be used.

The equation does not require past injection or production history other than the length of time the steam flood has been in operation.

Because of the high steam transmissibility between patterns, it is more practical to calculate injection rates on a project-wide basis rather than attempting to fine tune individual injection rates for each pattern.

Again, it should be remembered that steam volumes produced from the production well casings must be added to the calculated injection rates.

CONTRASTING CONCEPTS

Many operating and design practices accepted as conventional wisdom for a displacement or drive process are not applicable to a gravity drainage process. Measures suitable for one process may, in fact, be directly opposite to those prudent for the other process. Several of these contradictory concepts are compared in the list of Table 3.

This comparison of two different processes should not suggest that a steam flood operator is free to choose which process to put into effect underground. Instead, field experience has shown that gravity effects have almost invariably caused the steam to go overhead, forcing the steam flood into a drainage mode, whether this was desired or not.

The choices remaining to the operator concern only which operating and design principles will give the best results.

Some of the drive concepts and drain concepts in Table 3 can be considered as conjectural, i.e., not firmly supported by specific field testing. But on the whole most steam flood behavior reported in field case histories seems much more readily explained by the drainage concepts.

REFERENCES

Jones, J., "Steam Drive Model for Hand-held Programmable Calculators," JPT, September 1981, pp. 1,583-98.
Chen, H.L., and Sylvester, N.D., "Appraisal of Analytical Steamflood Models," Paper No. SPE 20023, SPE California Regional Meeting, Ventura, Apr. 4-6, 1990.
Jensen, T.B., Sharma, M.P., and Harris, H.G. "A Critical Evaluation of Preliminary Design Techniques for Steam Drive Projects," AIChE Spring National Meeting, Houston, Apr. 2-6, 1989.
Marx, J.W., and Langenheim, R.H., "Reservoir Heating by Hot Fluid Injection," excerpted by Chu, C.M., in Thermal Recovery, SPE Petroleum Engineering Handbook, 1987 Edition, pp. 46-7 and 46-8.
Miller, M.A., and Leung, W.K., "A Simple Gravity Override Model of Steamdrive," Paper No. SPE 14241, SPE Annual Technical Conference, Las Vegas, Sept. 22-25, 1985.
Kumar, D., Patel, H.N., and Denbina, E.S., "Use of a Gravity Override Model of Steamflooding at Cold Lake," Paper No. CIM 86-37-62, CIM Annual Technical Meeting, Calgary, June 811, 1986.
Chhina, H.S., Arthur, J.E., and Pebdani, E.N., "Analytical Steamflood Model for Heavy Oil Reservoirs," Unitar Fifth International Conference on Heavy Crude and Tar Sands, Caracas, 1991.
Hsu, C.F., "A Pattern Steamdrive Model for Personal Computers," Paper No. SPE 24076, SPE Western Regional Meeting, Bakersfield, Mar. 30-Apr. 1, 1992.
Miller, K.A., Stevens L.G., and Watt, B.J., "Successful Conversion of the Pikes Peak Viscous-Oil Cyclic Steam Project to Steamdrive," JPT, December 1991, pp. 1,510-17.
Smith, R.Y., Bertuzzi, A.F., Templeton, E.E., and Clampitt, R.L., "Recovery of Oil by Steam Injection In the Smackover Field, Arkansas," JPT, August 1973, pp. 883-89.
Ault, J.W., Johnson, W.M., and Kamilos, G.N., "Conversion of Mature Steamfloods to Low-Quality Steam and/or Hot-Water Injection Projects," Paper No. 13604, SPE California Regional Meeting, Bakersfield, May. 27-29, 1985.
de Haan, H.J., and Schenk, L., "Performance Analysis of a Major Steam Drive Project in the Tia Juana Field, Western Venezuela," JPT, January 1969, pp. 111-19.
Prover, G., et al., "Results of a Steam Drive Pilot Project in a German Oil Field," Paper No. SPE 11956, SPE Annual Technical Conference and Exhibition, San Francisco, Oct. 5-8, 1983.
Blevins, T.R., Aseltine, R.J., and Kirk, R.S., "Analysis of a Steam Drive Project, Inglewood Field, California," JPT, September 1969, pp. 1141-50.
Duerkson, J.H., and Gomaa, E.E., "Status of the 26C Steamflood, Midway-Sunset Field, California," Paper No. SPE 6748, SPE Annual Technical Conference and Exhibition, Denver, Oct. 9-12, 1977.
Burger, D.H., and Kisman, K.E., "The Kearl Lake Pilot: Striving To Be Number One In Athabasca In Situ Recovery, Alberta Oil Sands Technology and Research Authority (Aostra) Conference Oil Sands 2000, Edmonton, Mar. 26-28, 1990.
Myhill, N.A., and Stegemeier, G.L., "Steam-Drive Correlation and Prediction," JPT, February 1978, pp. 173-82.
Farouq Ali, S.M., "Current Steamflood Technology," JPT, October 1979, pp. 1,332-42.
King, R.L., and Stiles, J.H. Jr., "A Reservoir Study of the Hawkins Woodbine Field," Paper No. SPE 2972, SPE Annual Meeting, Houston, Oct. 4-7, 1970.
Dykstra, H., "The Prediction of Oil Recovery by Gravity Drainage," JPT, May , 1978, pp. 818-30.
Garb, F.A., and Smith, G.L., "Estimation of Oil and Gas Reserves," SPE Petroleum Engineers Handbook, 1987 Edition, pp. 40-14 and 46-15.
Haldorsen, H.H., et al., "An Evaluation of the Prudhoe Bay Gravity Drainage Mechanism by Complementary Techniques," Paper No. SPE 13651, SPE California Regional Meeting, Bakersfield, Mar. 27-29, 1985.
Bachman, R.C., Papst, W., and Legere, R.F., "Investigation of Gas Cycling Schemes for the Westerose D-3 Pool," Paper No. CIM 87-38-53, Annual Technical Meeting of the Petroleum Society of CIM, Calgary, June 7-10, 1987.
Edmunds, N.R., "A Model of the Steam Drag Effect in Oil Sands," Journal of Canadian Petroleum Technology, September-October 1984, pp. 30-39.
Noran, D., "Enhanced recovery requires special equipment," OGJ, July 12, 1976, pp. 50-55.
Johnson, R.S., et al., "History Matching of High and Low-Quality Steamfloods in Kern River Field, California," SPE RE, Feb. 1992, pp. 75-82.
Traverse, E.F., Deibert A.D., and Sustek, A.J., "San Ardo-A Case History of a Successful Steamflood," Paper No. SPE 11737, SPE California Regional Meeting, Ventura, Mar. 23-25, 1983.
Moritis, G., "EOR increases 24% worldwide; claims 10% of U.S. Production," OGJ, Apr. 20, 1992, pp. 51-79.
Greaser, G.R., and Shore, R.A., "Steam Flood Performance in the Kern River Field," Paper No. SPE 8834, SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa, Apr. 20-23, 1980.
Lowe, K.W., and Donovan, T.E., "Thermal Zone Monitoring in an Athabasca Oil Sands Pilot," Paper No. SPE 19854, SPE Annual Technical Conference and Exhibition, San Antonio, Oct. 8-11, 1989.
Singh, K.S., "Development Strategies for a Heavy-Oil (3,500 cp) Reservoir at a Depth of 2,100 ft (ss)," Paper No. SPE/DOE 24199, SPE/DOE Symposium on Enhanced Oil Recovery, Tulsa, Apr. 22-24, 1992.
Vogel, J.Y., "Simplified Heat Calculations for Steamfloods," JPT, July 1984, pp. 1,127-36.