Effect of trapped gas saturation on oilrecovery during the application ofsecondary recovery methods in exploitationof petroleum reservoirsA. Feigl

REVIEW

The paper presents the effect of trapped gas during the application of secondary processes, or moreprecisely, in the period of application of conventional waterflooding of partially depleted oil reservoirs.These processes can be simulated by analogous tests on reservoir rock samples, on synthetic models underlaboratory conditions or by numerical models. An explanation of physical mechanisms that take place duringsuch processes is required to understand better the benefits of gas phase saturation, and additionally tothat, favourable and unfavourable factors present during such processes are explained. Besides, here are aswell explained procedures for finding optimal trapped gas saturation, by which, for the selected partiallydepleted reservoir and the selected combination of fluids, and by waterflooding with maintenance of anearly constant reservoir pressure, maximal ultimate recovery is achieved. At the same time, it is a guidelinethat waterflooding should start when, during the natural process of recovery by solution gas drive,mentioned optimal gas saturation is accomplished. Finally, the paper gives a short review of the results ofdifferent simulation studies of such processes in which, inter alia, three-phase relative permeability curveswith expressed effects of hysteresis are used.

Key words: maximal recovery, optimal saturation, trapped gas, waterflooding

1. INTRODUCTION

In petroleum reservoirs with dissolved gas drive thepressure decreases due to depletion and the resultingconsequence is liberation of progressively higher quanti-ties of gas from oil, i.e. production is carried out with in-creasingly higher gas-oil ratios. If the method of reservoirpressure maintenance by flooding is applied to such res-ervoirs, the gas present in the pore space has a certain in-fluence on the quantity of oil remaining in the reservoirafter the completion of that process, as well as on the fi-nal oil recovery.

In such processes for macroscopic study of displace-ment efficiency, the most often is used linear flow model.In simplified calculations, it implies the application ofBuckley-Leverett's model of frontal displacement, whilstin more complex calculations one or multidimensionalnumerical models can be used.

Deviations from the linear flow model occur during theinitial period of flooding, when the fluid flow is radial,and remains such until the moment when the banks ofneighbouring wells connect into a joined bank. Naturally,it refers to the line injection-production well pattern,when assembling of individual waterflooding banks ofneighbouring injection wells hereafter implies linearflow.

Reservoirs with large gas caps are generally poor candi-dates for waterflooding due to the possibility of waterbank to bypass the oil through the gas cap area, and be-cause of the risk that oil might be, by waterflooding, in-

jected into the gas cap area, where it would remaintrapped.

2. DESCRIPTION OFWATERFLOODING PROCESSESWITH THE PRESENCE OF TRAPPEDGAS

Many professional papers and books in considerable de-tail describe the physical characteristics of thewaterflooding processes with the presence of trappedgas.1,3,7,11,12,13 In many cases it has been confirmed thatthe presence of free gas phase during waterflooding leadsto lower residual oil saturation than duringwaterflooding with no-gas presence. It also applies to thenatural water drive process, but due to the characteris-tics of such drive to maintain the pressure to a lesser orhigher degree, it is simultaneously accompanied by liber-ation of only small volumes of gas from oil, which mayalso have a favourable impact. Additionally, if the averagereservoir pressure is higher than the saturation pres-sure, gas is liberated only locally and in even smallerquantities.

2.1. Macroscopic aspects of waterfloodingprocesses

The favourable effect of free gas during the waterfloodingprocesses arises from the forming of immobile i.e.trapped gas saturation in the waterflooding bank area. Ifcontinuous gas phase is present in the reservoir prior to

NAFTA 62 (5-6) 153-159 (2011) 153

the advance of oil bank, duringwaterflooding a part of gas is forcedforward by the oil bank and a part re-mains trapped within it. Fig. 1 sche-matically presents distribution ofsaturation in such a process, whileFig. 2 illustrates its simplified linearmodel.

In such processes reservoir pres-sure mostly remains constant, inso-much effects of gas compressibilityand gas solubility in oil may bedisregarded.

The results of core tests performedunder laboratory conditions can betaken as confirmation of the abovestated.7 Fig. 3 shows oil and gas satu-ration profiles depending on cumula-tive injected quantity of water. Thetest was run on the core with initialmobile gas saturation. Only free gaswas produced at the beginning offlooding. Afterwards production ofgas ceased completely and oil with-out gas flowed at the outlet from thecore (marked by O-O' line). It was tobe expected due to the favourableoil-gas system mobility ratios (M =0,1),13 and because displaced gas al-ways plays a role of non-wettablefluid.7 After the water breakthrough(marked by W-W' line) oil continuesto be produced, but with significantlyhigher water ratio in total fluid, untilthe completion of waterflooding test.

Fig. 3, and description of the testreferring to that figure, indicate that

154 NAFTA 62 (5-6) 153-159 (2011)

A. FEIGL EFFECT OF TRAPPED GAS SATURATION ON OIL RECOVERY...

Fig. 1. Schematic presentation of areas saturated with different fluids betweenproduction and injection wells under conditions of displacement withimmiscible fluid

Sl. 1. Shematski prikaz podruèja zasiæenih razlièitim fluidima izmeðu proizvodnih iutisnih bušotina u uvjetima istiskivanja s nemješljivim fluidom

Fig. 2. Saturation profile that shows displacement of oil during initial saturation with mobile gasSl. 2. Profil zasiæenja koji prikazuje istiskivanje nafte u prisutnosti poèetnog zasiæenja pokretnim plinom

two different banks are formed dur-ing such process, firstly an oil bank,followed by water bank. These twobanks jointly advance through thecore sample during water injection atthe inlet core end, and it is in theform of saturation profiles illustratedin the same figure.

Fig. 4 shows the result of a test,where at the beginning ofwaterflooding the core sample is, ad-ditionally to oil, saturated only withtrapped gas. It is evident from the fig-ure that gas saturation of the core re-mains constant during the whole testand that only oil is produced at thebeginning, followed by oil and waterwith gradual increasing of water-oilratio in the total liquid. Thereby,waterflooding in the presence oftrapped gas is actually analogous tothe two-phase water-oil displace-ment that occurs in the part of thepore space unsaturated by gasphase.7

The curves of referent waterflood-ing in Figs. 3 and 4 (thinner lines)pertain to waterflooding tests with-out gas phase saturation of the core.The difference in oil saturationswhen gas is not, or is present (thickerlines), points to the favourable effectof gas during water injection in asense of higher decrease of residualoil saturation at all times during thetest duration.

2.2. Microscopic aspects ofwaterflooding processes

The trapped gas as non-wettablephase occurs in the reservoir as dis-continued, in the form of mutuallyseparate globules or discontinuousfilaments.3 Within the oil bank andahead of the water bank, oil fills thepore space around trapped gas, ex-cept in the part of the reservoir whichis saturated with connate water. Inthe water wettable system which con-tains oil, water and gas, it is to be ex-pected that gas will be trapped in oil.It is a minimum total free surface en-ergy position, as interfacial tensionbetween gas and oil is lower than in-terfacial tension between gas and wa-ter.3 Fig. 5 schematically showspositions of oil and gas in intergranu-lar space at residual oil saturation.3

In the previous subchapter, the ex-plained phenomena refer to wa-ter-wettable rocks, i.e. to tests onwater-wettable reservoir rock sam-

EFFECT OF TRAPPED GAS SATURATION ON OIL RECOVERY... A. FEIGL

NAFTA 62 (5-6) 153-159 (2011) 155

Fig. 3. Waterflooding in the presence of initial mobile gas saturation7

Sl. 3. Zavodnjavanje u prisutnosti poèetnog zasiæenja pokretnim plinom7

Fig. 4. Waterflooding when only trapped gas is present 7

Sl. 4. Zavodnjavanje u prisutnosti zarobljenog plina 7

ples. Tests run on oil-wettable samples show that pres-ence of trapped gas has no effect on final residual oilsaturation. Explanation why trapped gas reduces resid-ual oil saturation in water-wettable samples, but does notreduce it in oil-wettable ones, is as follows:

• In water-wettable rock the final residual oil saturationis trapped by water and residual gas is trapped withinoil, thus occupying a part of pore space which wouldotherwise be saturated with additional residual oil.

• In oil-wettable rock the final residual oil saturation is incontact with the rock surface, and residual saturationwith gas occupies the space which would be otherwiseoccupied with water.

With regard to water-wettable reservoir system, de-crease of residual oil saturation due to saturation withtrapped gas is the most beneficial effect during thewaterflooding process for increase of ultimate recovery.Because of that, this subchapter is mostly dedicated tothe description and explanation of such effect. Other,sometimes unfavourable effects present during theseprocesses, are briefly described in further text.

3. ADDITIONAL FACTORS AFFECTINGRESERVOIR WATERFLOODINGPROCESSES WITH THE PRESENCEOF TRAPPED GAS

In addition to the main effect, previously described de-crease of residual oil saturation, there are a number ofother factors which in a lesser or greater degree, favour-ably or unfavourably, affect the waterflooding processesin the presence of trapped gas. Some of the more impor-tant ones are listed below:

1. Oil viscosity: the greater the oil viscosity, the lesstrapped gas it contains.7,12 However, efficiency oftrapped gas saturation is the highest in waterfloodingof viscous oils.7

2. Increase of reservoir pressure: it is an unfavourableeffect if we take into consideration the loss of trappedgas. If the oil bank is of greater thickness, it requires

a higher displacement pressure,which causes gas compression anddissolution of gas in oil. Decrease oftrapped gas in places where oil bankchanges into water bank isaccompanied by increase of residualoil saturation and simultaneousdecrease of ultimate recovery. Inother words, the process is in alesser or greater extent changed intotwo-phase waterflooding - withoutgas presence. Besides, advance of oilbank to production wells is delayedfor a quantity of oil which, due to thepressure increase in the reservoirduring flooding, occupies the porespace previously saturated withgas.13 Increase of reservoir pressureat the same time represents afavourable effect, since dissolving ofgas trapped in oil causes its swellingand reduction of viscosity.7 Oil

swelling reduces its residual saturation, whilereduction of viscosity leads to more favourable oiland water mobility ratio. Therefore, if we plan toutilize the effect of trapped gas during waterflooding,then the process should take place at pressuregradient that is very low, compared to absolutepressure in the whole system.

3. Decrease of relative permeability to oil: it is anunfavourable effect caused by the presence oftrapped gas, which consequently leads to earlierwater breakthrough. However, this effect is smallerthan the favourable effect of residual oil saturationdecrease, explained in the previous chapter. Fig. 6shows an example of relative permeability to oildecrease due to gas presence.

4. Rock consolidation: in well-consolidated water-wettable rocks the effect of trapped gas saturation onresidual oil saturation is significantly higher incomparison with unconsolidated rocks.13

5. Speed of running waterflooding process: at higherspeeds, due to higher pressure gradient, gasdissolves in oil and the process turns into atwo-phase waterflooding, and gas saturation effect ismissing.2 Because of that, at all cumulative injectedwater quantities, average oil saturation remaining inthe reservoir increases with the increase ofdisplacement rates. In other words, at all cumulativeinjected water quantities, oil recovery is the highestat the lowest displacement rates.2

6. Gravitational effects during waterflooding have anunfavourable effect: introduction of gravitationaleffect in calculation of waterflooding with thepresence of gas causes markedly differentperformance from that of the one-dimensionalhorizontal systems. It is due to non-uniform verticaldistribution of initial gas saturation and gravityeffects on the injected water distribution.2

Gravitational effects are related to the rate of process

156 NAFTA 62 (5-6) 153-159 (2011)

A. FEIGL EFFECT OF TRAPPED GAS SATURATION ON OIL RECOVERY...

Fig. 5. The concept of oil and gas position in the pore space at residual oilsaturation (water wettable systems)3

Sl. 5. Koncept polo�aja nafte i plina u pornom prostoru kod preostalog zasiæenjanaftom (vodomoèivi sistemi)3

running, reservoir thickness, physical characteris-tics of the reservoir's and injected fluids (solubility,viscosity and compressibility), relative permea-bilities and areal waterflooding pattern (allocation ofproduction and injection wells).2 When we considergravitational effects, lower waterflooding rates over acertain range result in lower oil recovery rates. Anyhorizontal reservoir has an optimal waterfloodingrate due to opposing effects of gravitational influenceand initial free gas saturation.2

7. Spreading coefficient: if positive, it has a favourableeffect. Although in the paper which describes thiscoefficient10 tests are run in reverse order from theprocesses that have to be dealt with in this paper(double drainage mechanism: injection of gas instrictly water-wettable media, saturated with waterand residual oil), the effect of that coefficient can alsobe reviewed here. Spreading coefficient isrepresented by the following equation:

Sow wg og ow� � �� � � (1)

where:Sow = oil/water spreading coefficient, dyne/cm

�wg = interfacial tension of water/gas system, dyne/cm

�og = interfacial tension of oil/gas system, dyne/cm

�ow = interfacial tension of oil/water system, dyne/cm

If that coefficient is positive, oil tends to form a continu-ous film between water and gas. If the film is thickenough, oil flows through that film, reducing its residualsaturation. If the coefficient is negative, the oil filmbreaks, and so does the flow through it. Tests haveshown that in cases of positive spreading coefficients,markedly higher recoveries were accomplished than incase of negative ones.

4. OPTIMAL GAS SATURATION FORACHIEVEMENT OF MAXIMAL OILRECOVERY BY WATERFLOODING

Investigations of influence of a larger number of favour-able and unfavourable factors, shown in the two previouschapters, indicated undisputable usefulness of free and,related to that, trapped gas in the reservoir. Several au-thors therefore tried to establish the optimal trapped gassaturation which, after waterflooding that follows the pri-mary depletion phase, yields maximal additional oilrecovery.

One of the authors, using the program for calculation ofregressions with multiple variables, tried to establish acorrelation for calculation of optimal gas saturation.5 Heobtained a regression equation with 4 coefficients, whichsatisfies all sets of data with average absolute variation of4.6%.

In a more recent paper another author used variousmodels of relative permeability curves.8 He applied the

ones from earlier literature to de-scribe the primary depletion phasewith dissolved gas drive and flow inthe drainage direction. The others,calculated by him, were used for de-scription of waterflooding and forprocesses in the imbibition direc-tion. He checked his calculations byapplying the Buckley-Leverett frontaladvance equation. Based on the cal-culations, he developed differentnomograms to obtain optimal gassaturations depending on differencesin oil densities, saturation pressures,pore size distributions and degreesof waterflooding (at water break-through, after injection of one or twopore volumes of water). Fig. 7 pres-ents one of the typical results of cal-culation, which he used, among theothers, for development of above de-scribed nomograms. The figureshows results of calculation obtainedby using oil with density of 876 kg/m3

(30 °API) and average relative perme-ability curves. It is evident from thefigure that for such fluid and poresystem, all curves indicate maximaloil recovery at initial gas saturation,or more precisely, at gas saturationpresent in the moment when waterinjection started, of about 20%. Useof other combinations of above pa-

EFFECT OF TRAPPED GAS SATURATION ON OIL RECOVERY... A. FEIGL

NAFTA 62 (5-6) 153-159 (2011) 157

Fig. 6. Relative permeability to oil and water for the two systems in the Nellie Blysandstone3

Sl. 6. Relativna propusnost za naftu i vodu za dva sistema u Nellie Bly pješèenjaku3

rameters gave curves resembling the ones shown in theFig. 7.8,11

5. THE APPLICATION OFTHREE-PHASE RELATIVEPERMEABILITIES IN SIMULATIONOF WATERFLOODING PROCESSESWITH PRESENT TRAPPED GAS

Several recent papers introduce the results of numericalsimulations based on three-phase relativepermeabilities.4,6,9 The first of them describes a simula-tion of waterflooding process with the presence of gas ina linear system.9 Reservoir model incorporates the ef-fects of hysteresis in permeability and the entrapment ofgas by the oil bank. As an illustration, Fig. 8 shows rela-tive permeability drainage curves for oil (kro) and gas(krg), with several relative permeabilities to gas in the im-bibition direction.

The curves in the imbibition direction allow simulationof gas trapping after the start of waterflooding. Which krgcurve will be activated depends on the degree of reservoirgas saturation achieved in the primary phase of depletionby dissolved gas drive, at which the waterflooding began.

The reservoir model incorporateseffects of gas compressibility andgas solubility in oil.

Another paper focused even moreon the effects of hysteresis, whichwere applied not only to relative per-meability curves, but to capillarycurves as well.6 To some extent, thenumerical model is based upon re-membering the saturation history ofthe reservoir, which implies that incontinuation of simulation themodel takes into consideration thepreviously achieved final saturationvalues where the simulated processchanges direction. In order to beable to do that, an algorithm that al-lows calculation of curves of smoothtransition (so-called scanningcurves) from drainage to imbibitionand vice versa, was incorporatedinto the model. Besides, the modelalso accounts for the influence oftrapped gas or oil saturation on rel-ative permeabilities and capillarypressures. By combining modelsthat include hysteresis effect andapplying the equation ofthree-phase relative permeability tooil, the simulator can more realisti-cally predict residual oil satura-tions.

The most recent paper describesadditional features incorporatedinto the numerical model.4 Apartfrom active hysteresis andthree-phase permeability options,effects of composition and interfa-

cial tension (capillarity) on relative permeabilities are

also taken into consideration. Correlation for calculation

of different relative permeability curves is based on inter-

pretation of the pore-level mechanisms that determine

fluid flow. The paper describes a mixed-wet gas reservoir,

i.e. a type of reservoir with transient wettability.

6. CONCLUSION

The practical meaning of trapped gas saturation effi-ciency is that it represents a direct measure of additionalrecovery in comparison with the recovery that wouldhave been achieved by waterflooding without the gaspresence during the entire process, or when gas phasedisappears during waterflooding due to compressionand dissolution in oil. In line with that, efficiency oftrapped gas saturation is expressed in the form of addi-tional oil recovery as a fraction or percentage of the gasquantity that remains in the pore space during thewaterflooding process. From the aspect of trapped gassaturation efficiency, review of data confirmed that inmost practical cases presence of free gas would certainlyhave a beneficial impact on oil recovery achieved bywaterflooding.

158 NAFTA 62 (5-6) 153-159 (2011)

A. FEIGL EFFECT OF TRAPPED GAS SATURATION ON OIL RECOVERY...

Fig. 7. Increase of total oil recovery depending on initial gas saturation for poresize distribution II, oil density 876 kg/m3 (30 °API), and at different oilsaturation pressures8,11

Sl. 7. Poveæanje ukupnog iscrpka nafte u ovisnosti o poèetnom zasiæenju plinom zaraspodjelu velièina pora II, gustoæu nafte 876 kg/m3 (30 °API),i pri razlièitim tlakovima zasiæenja nafte 8,11

REFERENCES:1. Craig, F. F. Jr. (1971): "Efficiency of Oil Displacement by Water" (pogl. 3);

''The Reservoir Engineering Aspects of Waterflooding'', Monograph Volume3, Henry L. Doherty Memorial Fund of AIME, Society of Petroleum Engi-neers of AIME, New York, Dallas, 141 str.

2. Dandona, A. K. & Morse, R. A. (1972): ''The Influence of Gas Saturation onWater Flood Performance - Variations Caused by Changes in Flooding Rate'',SPE 3881, 1-52.

3. Holmgren, C. R. & Morse, R. A. (1951): ''Effect of Free Gas Saturation on OilRecovery by Water Flooding'', T.P. 3055, Petroleum Transactions of AIME,Vol. 192, 135-140.

4. Jerauld, G. R. (1997): ''General Three-Phase Relative Permeability Model forPrudhoe Bay'', 36178 P, SPE Reservoir Engineering, 255-263.

5. Khelil, C. (1967): ''A Correlation of Optimum Free Gas Saturation with Rockand Fluid Properties'' (preprint), SPE 1983, Society of Petroleum Engineersof AIME, Dallas, Texas, 109-115.

6. Killough, J. E. (1976): ''Reservoir Simulation with History-Dependent satura-tion Functions'', SPE 5106, Society of Petroleum Engineers Journal, 37-48.

7. Kyte, J. R. et al. (1956): ''Mechanism of Water Flooding in the Presence ofFree Gas'', Petroleum Transactions of AIME, Vol. 207, 215-221.

8. Land, C. S. (1968): ''The Optimum Gas Saturation for Maximum Oil Recov-ery from Displacement by Water'' (preprint), SPE 2216, Society of PetroleumEngineers of AIME, Dallas, Texas, 1-12.

9. Land, C. S. & Carlson, F. M. (1972): ''Numerical Waterflood Calculations withCompressible Fluids'', SPE 4374, U.S. Bureau of Mines, Laramie, 1-23.

10. Oren, P. E. et al. (1992): ''Mobilization of Waterflood Residual Oil by Gas In-jection for Water-Wet Conditions'', SPE 20185, SPE Formation Evaluation,70-78.

11. Seèen, J. (2006): "Zavodnjavanje naftnih le�išta" (pogl. 2); "Metode poveæanjaiscrpka nafte", sveuèilišni ud�benik, INA-Industrija nafte d.d., SD Naftaplin,Zagreb, 608 str.

12. Stewart, F. M. (1958): ''Reduction of Residual Oil due to the Presence of FreeGas during Water Displacement'' (preprint), 1048-G, Society of PetroleumEngineers of AIME, Dallas, Texas, 1-4.

13. Willhite, G. P. (1986): "Background" (pogl. 1), "Microscopic Efficiency of Im-miscible Displacement" (pogl. 2), "Macroscopic Displacement Efficiency ofLinear Waterflood" (pogl. 3) & "Waterflood Design" (pogl. 6);''Waterflooding'', Society of Petroleum Engineers Textbook Series Vol. 3(ISBN 1-55563-005-7), Richardson, Texas, 326 str.

�

Author:

Alan Feigl, MSc., Reservoir Engineering Expert, INA Industrija nafte d.d., Oil& Gas Exploration and Production, Field Engineering & Operations Sector,Reservoir Monitoring & Surveillance Department, Reservoir Engineering BU,10 000 Zagreb, Šubiæeva 29, Croatia.Tel.: +385-1-4592282Fax: +385-1-4592224e-mail: alan.feigl@ina.hr

EFFECT OF TRAPPED GAS SATURATION ON OIL RECOVERY... A. FEIGL

NAFTA 62 (5-6) 153-159 (2011) 159

Fig. 8. Drainage relative permeabilities for oil and gas with relative permeabilities togas during imbibition9

Sl. 8. Drena�ne relativne propusnosti za naftu i plin s relativnim propusnostima za plinpri upijanju9