Copyright 1999, Society of Petroleum Engineers Inc.

This paper was prepared for presentation at the 1999 SPE Annual Technical Conference andExhibition held in Houston, Texas, 36 October 1999.

This paper was selected for presentation by an SPE Program Committee following review ofinformation contained in an abstract submitted by the author(s). Contents of the paper, as presented,have not been reviewed by the Society of Petroleum Engineers and are subject to correction by theauthor(s). The material, as presented, does not necessarily reflect any position of the Society ofPetroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject topublication review by Editorial Committees of the Society of Petroleum Engineers. Electronicreproduction, distribution, or storage of any part of this paper for commercial purposes without thewritten consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in printis restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstractmust contain conspicuous acknowledgment of where and by whom the paper was presented. WriteLibrarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax 01-972-952-9435.

AbstractIFP water shutoff technology is based on the use of relativepermeability modifiers (RPMs). The technique consists ofbullhead injection of polymer solutions into existingcompletions, usually without zone isolation. The polymer canbe swelled or weakly crosslinked in situ to increasepermeability reduction to water. The chemistry of the differentprocesses is explained. Each process covers a specific domainof temperature and salinity. All systems are designed to affectoil or gas relative permeability only slightly.

By reviewing typical field cases, i.e. water shutoff in gasstorage wells, heavy-oil horizontal wells, offshore gravel-packed wells and multilayer-waterflooded wells in bothsandstone and limestone reservoirs, several guidelines arepresented, dealing with candidate well selection, processdesign, operational aspects and treatment evaluation. Crucialfor a successful treatment is the placement of the chemicals.Therefore in order to make RPM treatments more reliable,future focus of research should be oriented towards diversionaspects. Some solutions are suggested and discussed.

IntroductionAlmost all oil or gas reservoirs produce water. Since naturedoesnt like vacuum, water usually replaces oil as hydrocarbonreserves decline in the field. In mature or old fields, most ofproduced fluid is water, with oil or gas representing a fewpercent of total production. Moreover, many reservoirs aresubmitted to water injection, which provides pressuremaintenance and improves sweep efficiency. A continuousincrease in water production is thus a normal behavior in thelifetime of a field.

Often, water flow paths in the reservoir, especially close tothe wellbore, are irregular, by-passing large hydrocarbon-saturated zones and inducing undesirable high water-cutlevels. In such situations, we are dealing with "bad"undesirable water, as opposed to "good" water producedunder normal conditions. The causes of excessive waterproduction are multiple. Seright1 proposes the following list,whose order corresponds to increasing difficulty of treatmentby gels:1. Tubing/casing/packer leaks2. Flow behind pipe3. Layered reservoirs with vertical flow barriers4. Individual fractures between injectors and producers5. "2-D coning" through fractures6. Channeling through naturally fractured reservoirs7. 3-D coning or cusping8. Layered reservoirs without vertical flow barriers.

In this list, cases 1 and 2 correspond to completion failures,a workover problem. Gels have the advantage over cements ormechanical plugs to be able to penetrate the formation overseveral feet, and thus create a deeper barrier. Moreover, theycan easily be removed from the borehole by water recirculation.Since case 3 is frequently encountered in field situations and isalready difficult to solve, we will consider it as our base case(Fig. 1).

Lets consider a two-layer reservoir with a strongpermeability contrast (for example 1/10) and a horizontalcontinuous barrier which prevents cross-flow. Since the high-permeability layer is swept first, it has a tendency to overtakethe oil production from the low-permeability layer. Thissituation calls for a treatment which intends to decrease waterinflux from the high-permeability layer, thus favoring low-permeability layer production.

When the different layers are clearly separated andworkover costs are acceptable, a water shutoff treatment aimsat sealing off the watered-out layer with strong gels placed bymechanical tools (coil tubing, packers, etc.). Nevertheless, inpractice, bullheading is often the only option for the operatordue to several problems like poor identification of the differentzones surrounding the wellbore, multilayered production,unfavorable completion (gravel pack, slotted liners, etc.) or

SPE 56740

Water Shutoff by Relative Permeability Modifiers: Lessons from Several FieldApplicationsA. Zaitoun, SPE, N. Kohler, SPE, D. Bossie-Codreanu and K. Denys, SPE, Institut Franais du Ptrole

2 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

excessive workover costs (offshore wells, marginal wells). Therecent development of horizontal, multilateral, subsea wellscalls for more bullhead treatments in the future. One viabletechnique, confirmed by numerous field successes, is the useof "Relative Permeability Modifiers" (RPMs). High-molecular-weight water-soluble polymers or weak gels reduce selectivelythe relative permeability to water with respect to the relativepermeability to oil or to gas (Fig. 2).2 They are more suitablefor matricial than for fractured reservoirs, and are usuallybullheaded into the existing completion without zonal isolation.

Since the early 1980's, the Institut Franais du Ptrole (IFP)has had a continuous activity on water shutoff by RPMs, fromresearch to process design and field applications. More than100 well treatments have been performed worldwide with IFPprocesses. This paper tries to establish the state-of-the-art ofthe technique resulting from this experience.

RPM mechanismsAdsorption in reservoir rocks of water-soluble polymers orgels induces a selective reduction of the relative permeabilityto water with respect to the relative permeability to oil or togas. Fig. 3 plots end point permeability reduction for oil vs.end point permeability reduction for water of various gelsystems.3 All data except one verify this property.

The RPM pore scale mechanism is still controversial. Thereare two main "schools", one of which relying on "fluidpartitioning", whereas the other relies on "wall effects." Thefluid partitioning theory claims that there are segregated flowpaths for oil and for water inside the porous medium, and thegel tends to invade water flow paths, thus reducing watermobility preferentially.4-6 In the wall effect theory, the basicassumption is that after gel injection, a film covers the porewalls and changes dramatically two-phase flow properties bywettability-, steric- and lubrication effects. Fig. 4 gives apicture of end point pore situations. For the wall effects, thereare two hypotheses, one assuming that the polymer/gel film isalmost rigid,2, 7-8 the other that the film can be squeezed by oilflow through pore channels.9

Due to hydration water, polymer adsorption increases theirreducible water saturation. Furthermore, for a formationproducing both oil and water, a reduction of permeability towater induces automatically an increase in water saturation inthe zone invaded by the RPM. The combination of theseeffects, both inducing an increase in water saturation,decreases oil permeability.10 Thus in practice it is veryimportant to evaluate these unfavorable water saturationeffects and to minimize them whenever possible. As aconsequence, RPM treatments are more suitable in wellshaving zones with high oil saturation surrounding the wellborethan in wells where all zones produce at the same water cuts.

Due to the reduction of both water and oil permeabilities,RPM treatments always induce a loss in the well productivityindex. If this productivity loss is not counterbalanced by anincrease in the drawdown on the well (by activation or by a

lightening of the well fluid column), there is an obvious risk oflosing oil production, even with the water cut stronglyreduced. For RPM bullhead treatments in matricial reservoirs,it is thus risky to drop the relative permeability to water by afactor greater than 10. For fractured formations, due to thesuperficial invasion of the matrix blocks, RPM treatments haveless impact on the well productivity index.

Candidate selectionSeveral factors have to be taken into account for RPMcandidate well selection, i.e.,1) Heterogeneity For both permeability and saturation issues, strong verticalheterogeneity is a positive factor for the choice of a candidatewell. As explained earlier, the presence of both highly oil-saturated and highly water-saturated layers producingtogether is preferable than having all the layers producing atthe same water cut. Also, a strong permeability contrastbetween the layers is advantageous because the placement ofthe gel will be favored. In bullhead treatments, the gelant willinvade more deeply the high-permeability watered out layers(to be plugged) and less far the low-permeability oil-saturatedlayers (to be protected). From a more general point of view,since vertical heterogeneity is a factor enhancing waterbreakthrough, it makes at the same time the well a goodcandidate for a water shutoff treatment.2) Crossflow When there is crossflow between the layers, water can rapidlybypass the gel in place and therefore will return to the samerate as before treatment. Crossflow is thus a negative factor forcandidate choice. As a consequence, wells with a water coningare, in principle, bad candidates for RPM treatments. On theother hand, multilayered wells with no communication betweenthe layers are good candidates.3) Production mode Since a gel treatment reduces the well productivity index,maintaining well production requires a higher drawdown on thewell either through more activation (pumping, gas lift) orthrough reduction of the water cut (for eruptive wells) bylightening of the fluid column. Also, a good pressuremaintenance in the reservoir (active aquifer, gas cap) is anadvantage for maintaining well productivity.4) Technical constraints The gel should withstand reservoir conditions for long periodsof time. Thermal stability is often a major factor for treatmentselection. Also local environmental regulations, wellaccessibility etc. may play an important role in candidateselection.5) Economical constraints Water shutoff treatments are usually considered as workoveroperations. Treatment decision is based on comparison ofcosts vs. expected returns. It is very important to evaluateboth at an early stage. A candidate well should have apotential of incremental oil production sufficient to cover

SPE 56740 WATER SHUTOFF BY RELATIVE PERMEABILITY MODIFIERS: LESSONS FROM SEVERAL FIELD APPLICATIONS 3

treatment cost and make significant profit. An expensivetreatment can be perfectly suitable for a big offshore well, butcompletely inadequate for a small onshore well. Although inmost cases producing more oil is the target, sometimes theoperator can tolerate some loss in oil production providedwater production is strongly reduced. This is frequently thecase under offshore conditions when water handling capacitiesare limited. As a rule of thumb, treatment costs should be paidout by three months of post-treatment production.6) Origin of water production Some methods have been proposed in the literature to identifythe origin of water production in a given well.11 Although noneof them has reached commercial practice, these methods mayhelp in the selection of candidate wells. For example, theprofiles of WOR plots are markedly different for coning thanfor multilayer production.12 Recently, a method based on watercut analysis at the level of a field pattern has been proposed toidentify the contribution of surrounding wells (injectors andproducers) to the productivity of the candidate well.13,14

7) LogsLog analysis is a good indicator of the configuration of thepart of the reservoir surrounding the wellbore. Resistivity logsgive the saturation of the different layers. Gamma-ray logspoint to the presence of shale barriers and help to evaluateclay vertical distribution. Whenever possible production logsare run before and after the gel treatment in order to identifythe contribution of each individual layer in terms of total fluidflow and water production.

IFP processesIFP has designed several RPM processes which have beenapplied in the field. All processes use non-toxic materials andcan be bullheaded into the well. The processes are based onthe use of high-molecular-weight water-soluble-polymerswhich can be either swollen or weakly crosslinked in situ. Ashort description of the processes is given hereafter. Fig. 5shows a temperature/salinity diagram indicating the applicationdomains of the processes.Process AApplicable in low-salinity, low-temperature matricial reservoirs.Hydrolyzed polyacrylamides are injected in a high-salinitybrine. After production release, low-salinity formation waterreplaces progressively injection brine and swells the polymeradsorbed on pore walls. Advantages are a low viscosity duringinjection, a large adsorption and a high permeability reductionto water without the risk of well impairment by gels. Moredetails are given in Ref. 15.

Process BDepending on produced brine salinity, nonionicpolyacrylamides are injected with either a caustic swellingagent (that hydrolyzes the polymer in situ) or an organiccrosslinker (glyoxal). For higher temperatures acrylamide

copolymers can be crosslinked by zirconium lactate. A processdescription with different options is given in Ref. 15-17.Process CApplicable in high-temperature matricial reservoirs. It usesscleroglucan, a polysaccharide with strong shear-thinningrheology and excellent thermal stability. Polymer swelling canbe simply obtained by the release of shear forces between highinjection rates and low production rates. The polymer can beweakly crosslinked by zirconium lactate. Process description isgiven in Ref. 18, 19.

MethodologyThe preparation of a RPM water shutoff treatment requireslaboratory experiments, numerical simulations and on-fieldadjustments.

The laboratory study aims at (1) verifying polymer/additives injectivity and compatibility, (2) optimizing chemicalformulations, (3) running two-phase flow corefloods underreservoir conditions to measure end point relativepermeabilities before and after polymer treatment.

Numerical simulations are run in three phases, i.e. (1)establishment of a history match of fluid production from thecandidate well with a simplified near-wellbore reservoirdescription, (2) simulation of polymer injection, (3) post-treatment production forecasts. Numerical simulations aim atsizing treatment slug volume and evaluating expectedperformances.

Onfield adjustments are made with a light lab equipmentenabling treatment survey during field operations. Thisequipment includes BrookfieldTM type viscometer, pH meterand in-flow wellhead pressure recorder. Quick compatibilitychecks with actual fluids are usually done before startingoperations. During polymer injection, solution viscosity has tobe adapted to the actual wellhead pressure, which of coursehas to remain below the fracturation pressure (a 20% safetyfactor is usual).

Field casesIn the following, different field cases are discussed. Appliedguidelines dealing with candidate well selection, processdesign and operational aspects are presented and correlatedwith treatment evaluation.

1) Gas storage wells: Treatments of sandstone and limestonereservoirs (France)Several water shut-off treatments on gas storage wells insandstone reservoirs have been performed. Well treatmentsbased on RPM technology have proven to be effective in mostcases.15,20

For example the treatment of well VA 48 of the Cerville-Velaine gas storage reservoir by Process A reduced thewater/gas ratio for at least 3 years. The main characteristics ofthe well treatment are shown in Table 1. The treatment isdocumented in Ref. 15.

4 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

A candidate well from a gas storage limestone reservoirwas proposed for a polymer treatment. Usually carbonatereservoirs are at least slightly fissured and require a geltreatment to reduce water encroachment. In the case of well VN21 from the Saint-Clair-sur-Epte gas storage the formationconsisted of a superposition of alternating layers of highpermeability grainstone deposits (k = 0.7 mm2) and layers of lowpermeability packstones (k = 0.01 mm2) (Table 1). Due to thehigher formation brine salinity, Process B was preferred. Labstudies performed on model carbonate cores demonstrated thefeasibility of a polyacrylamide based adsorption processfollowed by in situ swelling of the adsorbed polymer by KOH.It was concluded from numerical simulations and laboratorytests that the optimal RPM slug volume was 250-300 m3 andthat injection had to be done in the zone below a packerlocated in the middle of the pay-zone.

Consequently a slug of 248 m3 of polymer solution (averageconcentration 2 kg/m3) was injected. Due to the use of riverwater, a bactericide was added. Unexpected compatibilityproblems between the bactericide and the swelling agentinduced a premature crosslinking of a part of the polymer inboth surface installations and in the wellbore. Polymerinjection was nevertheless continued without swelling agentbut with a reduced injection rate and followed by a river waterpostflush.

Gas injection (started one week after treatment) showedthat injection rate in the zone below the packer was lower (10000 std m3/h) than before the treatment (15 000 std m3/h).Besides, a productivity test, performed at the end of thefollowing winter producing campaign, showed that waterproduction rate from the well was almost unchanged (Table1). It was concluded that due to the premature gelling of a partof the injected polymer solution, some damage was done to theformation, thus explaining treatment failure. Consequently thepolymer adsorption process did not proceed as intended.

2) Horizontal wells: Treatments in the Pelican Lake andSouth Winter fields (Western Canada) Four heavy-oil horizontal wells from the Pelican Lake field,namely wells 11-15A, 11-15B, 14-10A and 14-10B, were treatedby Process B (Table 2).21 Although the same injectionprocedure was applied, i.e. bullheading the chemicals into theslotted liner drainhole, only well 11-15A gave a good response.After treatment, the water cut dropped immediately from 85 to50% and remained low afterwards (Fig. 6). For this well both anincrease in oil production and a decrease in water productionwere observed for two years following the treatment (Fig. 7).

For the three other wells, the response to the treatment wasmuch weaker. A possible explanation for this difference inbehavior, could the less favorable placement of the polymer.Indeed, horizontal well profiles (Fig. 8) show that the lowestpoints of the drainhole (corresponding to a higher watersaturation) are at the heal for well 11-15A and further away forthe other wells. Probably, for these wells, an appreciable

amount of polymer invaded oil productive zones beforereaching water zones. Moreover, since the aquifer is not active,both oil and water production decrease strongly due to thepoor pressure maintenance. The long term trend for all wells,even without treatment, is a progressive reduction of the watercut.

Another horizontal well, B 4-10 from the South Winter field,was treated by Process B. This well was producing heavy oilfrom a high-permeability sand reservoir laying above a veryactive bottom aquifer (Table 2). Bottom water coning wasresponsible for the high water cut, rendering the productionuneconomic. It was decided to perform the treatment of well B4-10 in two steps, bullheading the first half of the injectedpolymer volume through a tubing placed at the heal of thehorizontal slotted liner and the second half at the tail. Due tothe high oil/water mobility ratio it was estimated that thepolymer would invade the water zones preferentially.

After treatment, the water cut decreased from 95 to 80% for2 months and then slowly increased again to reach theeconomic productivity limit. It was concluded that polymerdiversion to water zones was well achieved. However, thetreatment effect did not last due to the presence of a strongbottom aquifer and a too weak gel formulation.

Table 2 summarizes the treatments of the horizontal wells inboth fields.

3) Gravel-packed wells: Treatment of two offshore wells in theGulf CoastTwo gravel-packed oil wells in the Gulf Coast with bottomholetemperatures of 88C and 93C respectively, were submitted tobullhead treatments by Process C. These types of wells wereknown to water out quickly after breakthrough . A RPMtreatment was expected to stabilize or reduce the water cut.Moreover their gravel-packed completions called for RPMtechnology instead of cement squeeze or zonal isolation.

For the first well, the gamma ray log showed a permeabilitydecrease from bottom to top with an average permeability ofabout 1.25 mm2 enhancing the water coning effect. The netresult of the injection of 500 m3 of a 0.5 kg/m3 polysaccharidesolution was to reduce the water production substantiallyduring a 2.5 year period following the treatment (Fig. 9).22 Thebasic action of the polymer was to redistribute the fluid flowand allow reconstitution of some permeability barriers near thewellbore. As a result the water production was reduced and oilproduction from low permeability layers was stimulated.

Following this successful treatment, the operatingcompany proposed a second candidate well from the samefield. Although the average permeability of the 14 m thickreservoir was only 0.07 mm2, the water production was as highas 2000 BWPD. It was estimated that in this case waterencroachment proceeded through high permeability streaks.As a result of the injection of 305 m3 of polymer solution boththe water/oil ratio and the total fluid rate were dramaticallyreduced. Contrary to the first well, polymer treatment induced a

SPE 56740 WATER SHUTOFF BY RELATIVE PERMEABILITY MODIFIERS: LESSONS FROM SEVERAL FIELD APPLICATIONS 5

20% reduction in oil rate. For economical reasons this well wasthen recompleted in another sand.

4) Multilayer waterflooded wells: Treatments in theChagirtsk field (Russia)A number of wells from the Chagirtsk field were treated byProcess B. Candidate well selection was performed accordingto the earlier mentioned criteria, especially those concerningthe existence of stratifications and permeability anisotropy. Inthis extensively waterflooded field, Bobrick 2 (Bb 2) is the mainoil producing interval. All wells are usually perforated over thetotal height of the sandstone reservoir. Nevertheless some ofthe wells are also producing from the upper Tula 2b (Tl 2b) andBobrick 1a (Bb 1a) intervals (Table 3). Prior to polymertreatment a water injectivity logging test was performed oneach candidate well. For all wells except two, the injected waterentered the Bb 2 reservoir. The exceptions were wells C 325and C 1160, where brine injection affected respectively the topof reservoir Bb 1a and the total height of the reservoir Tl 2b(Table 3).

Process B is usually implemented in two sequences: itstarts with a single polymer treatment aiming at diverting thegelant, injected after, towards the more permeable waterbearing layers.

The size of each sequence is deduced from such data asreservoir thickness, injection rate, wellhead pressure duringwater injectivity test and production prior to the treatment.

Table 3 shows that for both wells C 325 and C 1160 thevolume of the second sequence was much larger than for theother wells. This was done in order to reduce waterproductivity from reservoirs Bb 1a or Tl 2b and favorproduction from reservoir Bb 2. Table 4 shows that this choiceseemed to be erroneous for well C 325 and right for well C 1160.For well C 325 the water cut was found to increase after thetreatment leading to an estimated loss of about 3,400 tons of oilover a 13 months period of time. For well C 1160, the water cutdecreased drastically during the first 4 months after treatmentand increased again to values close to 100%. Presumably forboth wells it would have been preferable to force the treatmentto enter the lower Bb 2 reservoir.

As can be seen in Table 4, the water cut in all other wellswas reduced, leading to appreciable amounts of incremental oil.

5) Carbonate wells: Treatments in the Kudryachevo field(Russia)Process B was also implemented on three wells in theKudryachevo field producing from the Tournaisien formation(limestone reservoir). This formation is characterized by asuperposition of three to seven oil producing layers more orless clearly differentiated. According to the operator theexistence of large fractures in this reservoir is not proven. Theformation has thus to be considered as essentially matricialwith production characteristics quite similar for the threecandidates.

A water injectivity logging test on these wells showed areasonable injectivity, equally distributed over the entireheight of the perforated intervals. Wellhead pressure duringthis test was nevertheless much lower for wells K 2 and K 9,indicating the presence of microfissures or high permeabilitystreaks (Table 5). Consequently the slug size of thepolymer/crosslinker sequence for these wells was chosen to belarger than for well K 3. Actual wellhead pressure duringtreatments remained quite similar for all the wells (5000 - 10000 kPa).

Table 6 gives main treatment results. Both wells K 2 and K9 maintained their overall productivity after treatment and hada strong reduction in water cut. On the contrary well K 3showed a loss in average production rate over 9 monthsresulting from an initial mechanical failure of the pumpingequipment and an irreversible damage of formationproductivity. Although the treatment of wells K 2 and K 9produced the same amount of incremental oil, their behaviorwas quite different during the 9 months following the release ofproduction (Figs. 10, 11). For both wells production rate tookabout 3 months to reach the level before treatment. Well K 9showed an improvement in oil productivity during this periodwhile the water cut of well K 2 remained high. After 3 monthsthe opposite was observed: the water cut of well K 2 droppedsharply and remained low, while for well K 9 it increased to pre-treatment level. The reason for this difference in behavior isnot clear.

Treatment evaluationFor most operators, a successful treatment should induce botha significant reduction in water production and an increase inoil production. From the examples given, it follows that theeffect of a RPM treatment can last for several months up toseveral years (for example Pelican Lake, well 11-15 A).21

Generally, the primary criterion for a technical success isthe water cut reduction. However, some wells do react quickly,while others show a smoother behavior. This differentbehavior can even occur within the same field (for exampleKudryachevo). Thus a definite water cut evaluation mayrequire several months of monitoring. In some cases due to thereduction of the well productivity index, a strong decrease inthe water cut may coincide with a significant decrease in oilproduction. In such a case, a post-flush of a breaking agent(peroxyde, acid) can be attempted to restore well productivity.

A convenient tool to evaluate incremental oil production isgiven by plots of cumulative oil vs. cumulative water (Fig. 7).A successful water shutoff treatment induces a break in theplot, with a strong increase in the slope. Incremental oil can beestimated then by the difference between the actual cumulativeoil curve and the extrapolated values from pre-treatmentslope.13,14 However, this plot does not indicate the productionrate evolution, which has to be checked also to evaluatetreatment success. As a rule of thumb, a treatment can beconsidered as successful when the pay out is obtained in a

6 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

three month period and the returns largely exceed treatmentcosts. In our best cases (for example Well 11-15A in PelicanLake) the pay-out was obtained in a couple of weeks and thereturn was more than ten times treatment cost.

Discussion, guidelines and future research focusFrom the field cases presented in this paper we can concludethat RPM water shutoff treatments have a wide range ofapplication and could lead to remarkable successes. Howeverfor the users the results are often disappointing as thetechnology still suffers from a lack of confidence.

Here are some guidelines to possibly improve the chancesof success of a RPM treatment:(1) Candidate selectionThe mechanism of RPM implies that in the zone invaded by thegel water saturation increases, thus reducing oil permeability.The magnitude of this reduction increases with the fraction ofwater produced from the zone invaded. Therefore, wells havingnear-wellbore "virgin" oil layers are good candidates (forexample young wells having a sudden water encroachment).On the contrary, old wells, having produced for long periods oftime at high water cut, are often bad candidates.

Heterogeneous reservoirs without crossflow between thelayers and with high permeability contrast are also goodcandidates, because (i) water and oil flow through differentpathways, so both permeability and saturation contrastsbetween the layers are high, and (ii) once the RPM is placed,the absence of communication between the layers preventswater flow in oil productive zones.

Each candidate well has to be considered as a case per se.This implies that (i) all available data (which could be collectedin the form of a check list) have to be analyzed before taking adecision, (ii) a successful treatment obtained in a first well doesnot guarantee success in a neighboring one (although itincreases chances of success), and (iii) some wells which looka priori as poor candidates may be acceptable or even goodcandidates after careful analysis. This last situation is attestedby two remarkable successes obtained from our ownexperience, i.e. the first horizontal well (11-15A) treated inPelican Lake field (where a water path was located close to theheal) and the first well treated in the Gulf Coast where shalelaminations existed close to the roof of the reservoir in aconing situation.(2) Operational issuesObviously the treatment has to be adapted to well/reservoirconditions. Temperature vs. salinity diagrams (Fig. 5) arehelpful for process screening. Once the process is chosen, theconcentrations of the different chemicals (polymer, crosslinkeror swelling agent, biocide) are defined by laboratory studies. Itis worthwhile to perform quick lab tests on site with actualfluids. During treatment the chemical composition has to beadjusted according to well response during injection. Someunexpected behavior may occur with actual fluids andchemicals. This happened in Saint-Clair-sur-Epte gas storage,

where a premature gelation occurred due to interactionbetween the swelling agent and the biocide (Table 1).

The normal sequence of a treatment starts with thepreparation of the well (pump release, wellbore clean up bybrine recirculation) followed by an injectivity test with brine.Wellhead pressure is monitored during brine injection atdifferent rates. The injectivity index is measured and comparedto expected values. If it is too low, a stimulation treatment maybe considered (acid, solvent squeezes). Polymer injection canonly proceed when the value of the injectivity index isacceptable. If the injectivity remains low, the polymerconcentration may be reduced. On the other hand, if it is high,polymer concentration has to be increased.

RPM treatments begin with injection of polymer alone (tocheck injectivity). A skin is then developed in the lowpermeability (oil-bearing) layers, thus preventing deeppenetration of the subsequent polymer/crosslinker mixture. Inour Russian treatments, we injected more or less equal volumesof single polymer and polymer/crosslinker mixture (except inthe special cases discussed). After polymer injection it isuseful to injected a small post-flush of brine, followed by dieseloil, to clean up and resaturate the near wellbore. Well shut-intime before production release is typically in the order of a fewdays. The release of well production has to be veryprogressive to avoid sudden pressure drawdown on thepolymer/gel bank.(3) Future focus of researchThe upper temperature limit of RPM is currently around 120C(Process C). For higher temperature applications there is aneed of new products, which could be low concentrationversions of high-temperature plugging gels (for exampleacrylamide copolymer/polyethyleneimine system,23 or lowmolecular weight polymers20). The consistency of RPMproducts can be adjusted by the concentrations of thedifferent compounds. The key issue to improve RPMtreatments concerns placement techniques. When thechemicals have to be bullheaded into the existing completion, itis very important to reduce gel penetration in oil-bearinglayers. Since most often these layers are also the lesspermeable ones, a promising way could be to inject a diverteras a preflush that can build-up a superficial skin on the low-permeability layers. The barrier created is expected to preventgelant penetration in these zones. The skin formed on the oil-productive zones must be destroyed before production release,for example by breaker postflush. Two diversion techniques forRPM treatments have been recently proposed in the literature.The first one is based on the use of flexible polymers with highadsorption energy, which can bridge pore throats.24-26 Thesecond is based on preflushing a highly viscous non-damaging fluid before the gelant.27,28 Both approaches arecurrently investigated in a European research program(WELGEL).

SPE 56740 WATER SHUTOFF BY RELATIVE PERMEABILITY MODIFIERS: LESSONS FROM SEVERAL FIELD APPLICATIONS 7

Conclusions

1) IFP has designed several RPM water shutoff processes inorder to treat various types of reservoir characteristics. Theyare based on adsorption of high-molecular-weight water-soluble polymers, which can be either swollen, or weakly gelledby organic crosslinkers. Their application domains in terms oftemperature and salinity are quite complementary. IFPprocesses can be applied at various salinities and temperaturesup to 120C.2) The processes have been used in different field situations,i.e., gas storage wells, heavy-oil horizontal wells, gravel-packedoffshore wells, multilayer waterflooded wells both in sandstoneand in limestone reservoirs. The paper reviews some of theseapplications, commenting successes and failures.3) Guidelines concerning candidate well screening andoperational issues are discussed. One of the key selectioncriteria is the presence of heterogeneities between the differentlayers surrounding the wellbore (both in terms of permeabilityand oil saturation).4) Due to the adverse effect on oil permeability of increasedwater saturation in the zones invaded by the RPM, it isimportant to prevent deep penetration in oil zones. Since thesezones are often the less permeable, the use of a diversionpreflush should be considered in the future.

Acknowledgments

The authors wish to acknowledge the managements ofChevron, Gaz-de-France, CS Resources and Permneft for theircooperation.

References1. Seright, R.S. in Minutes and Key points of SPE Applied

Technology Workshop on Water Conformance, Dunkeld,Scotland, 19-22 May 1997, prepared by Bob Eden, SPE.

2. Zaitoun, A., Bertin, H. and Lasseux, D.: "Two-Phase FlowProperty Modifications by Polymer Adsorption," paperSPE 39631 presented at the 1998 SPE/DOE IOR Symposium,Tulsa, OK, 19-22 April 1998.

3. Liang, J., Sun, H. and Seright, R.S.: "Reduction of Oil andWater Permeabilities Using Gels," paper SPE 24195presented at the 1992 SPE/DOE Symposium on EnhancedOil Recovery, Tulsa, 22-24 April.

4. Liang, J., Sun, H. and Seright, R.S.: "Why Do Gels ReduceWater Permeability More Than Oil Permeability?" SPERE(November 1995) 282-286.

5. Liang, J. and Seright, R.S.: "Further Investigations of WhyGels Reduce Water Permeability More Than OilPermeability," SPEPF (November 1997) 225-230.

6. Nilsson, S., Stavland, A. and Jonsbraten, H.C.:"Mechanistic Study of Disproportionate PermeabilityReduction," paper SPE 39635 presented at the 1998

SPE/DOE Improved Oil Recovery Symposium, Tulsa, 19-22April.

7. Barreau, P., Bertin, H., Lasseux, D., Glnat, P. and Zaitoun,A.: "Water Control in Producing Wells: Influence of anAdsorbed-Polymer Layer on Relative Permeabilities andCapillary Pressure," SPERE (November 1997) 234-239.

8. Barreau, P., Lasseux, D., Bertin, H., Glnat, P. and Zaitoun,A.: "Polymer Adsorption Effect on Relative Permeabilityand Capillary Pressure: Investigation of a Pore ScaleScenario," paper SPE 37303 presented at the 1997 Int.Symposium on Oilfield Chemistry, Houston, 18-21February.

9. Mennella, A., Chiappa, L., Bryant, S.L. and Burrafato, G.:"Pore-Scale Mechanism for Selective PermeabilityReduction by Polymer Injection," paper SPE 39634presented at the 1998 SPE/DOE Improved Oil RecoverySymposium, Tulsa, 19-22 April.

10. Liang, J.T., Lee, R.L. and Seright, R.S.: "Gel Placement inProduction Wells," SPEPF (November 1993) 276-284.

11. Zaitoun, A., Kohler, N. and Bossie-Codreanu, D.: "WaterControl in Hydrocarbon Reservoirs and Storages: ALiterature Review - Chapter 4 : Reservoir Evaluationtechniques," Final Report, European CommissionDirectorate-General for Energy, Contract THERMIE - NDIS-1203-97-DE (1998).

12. Chan, K.S.: " Water Control Diagnostic Plots", SPE 30775,SPE Annual Technical Conference, Dallas, TX, 22-25October, 1995.

13. Renard, G., Dembele, D., Lessi, J. and Mari, J.L.: "SystemIdentification Approach Applied to Watercut Analysis inWaterflooded Layered Reservoirs, " paper SPE 39606presented at the 1998 SPE/DOE IOR Symposium, Tulsa,OK, 19-22 April 1998.

14. Kohler, N., Lessi, J. and Tabary, R. : "SuccessfulApplication Cases of Water Control Treatments in Russia," Revue de lInstitut Franais du Ptrole, Vol 50 (3), 381-390, May-June 1995.

15. Zaitoun, A., Kohler, N. and Guerrini, Y.: ImprovedPolyacrylamide Treatments for Water Control in ProducingWells, J.Pet.Techn., 862-867, July 1991.

16. Zaitoun, A., Rahbari, R. and Kohler, N.: ThinPolyacrylamide Gels for Water Control in High-PermeabilityProduction Wells, SPE 22785, 66th SPE Annual FallMeeting, Dallas TX, October 6-9, 1991.

17. Kohler, N., Zaitoun, A., Maitin , B.K. and Truchetet, R.:Selective Control of Water Production in Oil or GasProducing Wells," Oil and Gas in a Wider Europe, 4th ECSymposium, Berlin ,Germany, 1992.

18. Kohler, N. and Zaitoun, A.: Polymer Treatment for WaterControl in High-Temperature Production Wells, SPE21000, SPE International Symposium on Oilfield Chemistry,Anaheim, CA, February 20-22, 1991.

19. Kohler, N., Rahbari, R., Han, M. and Zaitoun, A.: Weak GelFormulations for Selective Control of Water Production in

8 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

High-Permeability and High-Temperature ProductionWells, SPE 25225, SPE International Symposium onOilfield Chemistry, New Orleans, LA, March 2-5, 1993.

20. Pusch, G., Kohler, N. and Kretzschmar, H.J.: PracticalExperience with Water Control in Gas Wells by PolymerTreatments, 8th European IOR Symposium, Proceedingsvol 2, 48-56, Vienna, Austria, May 15-17, 1995.

21. Zaitoun, A., Kohler, N. and Montemurro, M.A.: "Control ofWater Influx in Heavy-Oil Horizontal Wells by PolymerTreatment," paper SPE 24661 presented at the 1992 SPEAnnual Technical Conference and Exhibition, Washington,DC, October 4-7.

22. Gruenenfelder, M., Zaitoun, A., Kohler, N., Ali, S.A. andLinser, T.M.: Implementing New Permeability SelectiveWater Shutoff Polymer Technology in Offshore Gravel-Packed Wells, SPE/DOE 27770, SPE/DOE 9th Symposiumon Improved Oil Recovery, Tulsa, OK, April 17-20, 1994.

23. Hardy, M. and Botermans, C.W.: "New OrganicallyCrosslinked Polymer System Provides CompetentPropagation at High Temperatures in ConformanceTreatments," paper SPE 39690 presented at the 1998 SPESymposium on IOR, Tulsa, OK, April 19-22.

24. Zitha, P.L.J. , Chauveteau, G. and Zaitoun, A.:"Permeability-Dependent Propagation of PolyacrylamidesUnder Near-Wellbore Flow Conditions," paper SPE 28955presented at the 1995 SPE International Symposium onOilfield Chemistry, San Antonio, TX, February 14-17.

25. Zitha, P. L.J. and Botermans, C.W.: "Bridging Adsorptionof Flexible Polymers in Low-Permeability Porous Media,"paper SPE 36665 presented at the 1996 SPE AnnualTechnical Conference and Exhibition, Denver, CO, 6-9October.

26. Zaitoun, A. and Chauveteau, G.: "Effect of Pore Structureand Residual Oil on Polymer Bridging Adsorption," paperSPE 39674 presented at the 1998 SPE Symposium on IOR,Tulsa, OK, April 19-22.

27. Kvanvik, B.A., Litlehamar, T. and Stavland, A.: "GelantTransport and Placement in Heterogeneous Reservoirs," inRUTH Program Summary, Skjaeveland et al. (eds.),Norwegian Petroleum Directorate, Stavanger, 1996.

28. Thompson, K.E. and Kwon, O.: "Selective ConformanceControl in Heterogeneous Reservoirs Using Unstable,Reactive Displacements," paper SPE 39672 presented at the1998 SPE Symposium on IOR, Tulsa, OK, April 19-22.

Table 1 Treatments of gas storage wells

Characteristics Sandstone reservoirVA 48

Limestone reservoirVN 21

Reservoir parametersLithology

Thickness (m)Permeability (mm2)

Brine salinity (g/L TDS)Temperature (C)

Massive sandstone60

0.1-1 (top 55 m) 5 (bottom 5 m)

0.97230

Layered limestone28

grainstones: 0.7packstones: 0.01

1436

TreatmentRPM process

Polymer concentration (ppm)Brine salinity (g/L TDS)

Injected volume (m3)

Process A(HPAM + salinity gradient)

30008.2700

Process B(PAM + KOH)

2000river water

248

Results Water/gas ratio Water production strongly reducedGas injection/production unchanged

Gas injection ratebefore = 15 000 std m3/hafter = 10 000 std m3/h

Water production rate unchanged

SPE 56740 WATER SHUTOFF BY RELATIVE PERMEABILITY MODIFIERS: LESSONS FROM SEVERAL FIELD APPLICATIONS 9

Table 2 Horizontal well treatments in Pelican Lake and in South Winter

Characteristics Pelican Lake South Winter

Reservoir ParametersHorizontal length (m)

LithologyPermeability range (mm2)

AquiferOil viscosity (mPas)

Brine salinity (g/L TDS)

500Wabiskaw sand

> 1very weak

100010

800Dina sand

3 5very strong

300052

TreatmentRPM process

Injected volume (m3)Injected viscosity at 9.5 s -1

(mPas)

Process B(PAM + KOH)

60 110

50

Process B(PAM + Glyoxal)

400

15Results

Water cut (%) 85-90 50-70( 2 years)

95 80(2 months)

Table 3 Main characteristics of well treatments in Chagirtsk

Well Perforated interval(m)

Water injectivity test(m3)

Treatment(m3)

Layer

Tl 2b Bb 1a Bb 2 Total Rate(m3/day)

Pressure(kPa)

Intakeinterval

Polymeralone

Polymer+ X-linker

C 2131

C 336

C 1143

C 1177

C 325

C 1160

-

-

-

3

5.2

5

-

-

-

7

6.6

-

10

11

9

8

7.8

16

10

11

9

18

19.6

21

116

286

208

132

192

150

11 000

10 000

8000

11 000

9000

13 500

Bottom Bb 2

Total Bb 2

Bottom Bb2

Bottom Bb 2

Top Bb 1a

Total Tl 2b

50

50

44

48

36

25

42

33

42

36

66

67.5

10 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

Table 4 Main results of well treatments in Chagirtsk

Well Production data beforetreatment

Production data 4 months aftertreatment

Incremental oil(tons in

[ x] months)Rate

(m3/day)Water cut

(%)Rate

(m3/day)Water cut

(%)C 2131*

C 336

C 1143

C 1177

C 325

C 1160**

120

35

150

130

160

110

90

80

85

90

90

85

150

35

160

145

190

110

80

50

75

80

100

45

2149 [4]

3054 [13]

4988 [13]

5464 [13]

- 3399 [13]

3534 [4]

* Well C2131 was shut in after 4 months due to pump failure** The water cut of well C1160 increased strongly to near 100 % after 4 months

Table 5 Main characteristics of well treatments in Kudryachevo

Well Perforated interval

(m)

Water injectivity test Treatment (m3)

Rate(m3/day)

Pressure(kPa)

Polymer alone Polymer+ X-linker

K2

K3

K9

13

24.5

20

775

680

750

3000

8000

1000

25

30

20

59

41

67.9

Table 6 Main results of well treatments in Kudryachevo

Well Production data beforetreatment

Production data 6 months aftertreatment

Incremental oil(tons in

9 months)Rate

(m3/day)Water cut

(%)Rate

(m3/day)Water cut

(%)K2

K3

K9

110

115

120

90

95

90

110

60

120

50

95

60

6272

not evaluated

4772

SPE 56740 WATER SHUTOFF BY RELATIVE PERMEABILITY MODIFIERS: LESSONS FROM SEVERAL FIELD APPLICATIONS 11

Fig. 1 Principle of RPM treatment

Fig. 2 Modification of relative permeability after polymeradsorption in water-wet sandstone

Fig. 3 Disproportionate permeability reductionby polymers and gels

12 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

Fig. 4 Pore scale end point situations

Fig. 5 Field of application of water shutoff treatmentsversus formation temperature and salinity

Fig. 6 Pelican Lake 11-15A monthly production

SPE 56740 WATER SHUTOFF BY RELATIVE PERMEABILITY MODIFIERS: LESSONS FROM SEVERAL FIELD APPLICATIONS 13

Fig. 7 Pelican Lake 11-15A oil production versus water production

Fig. 8 Structural comparisonPelican Lake

14 A. ZAITOUN, N. KOHLER, D. BOSSIE-CODREANU SPE 56740

Fig. 9 WOR history for Gulf Coast well no.1

Fig. 10 Water cut history for well Kudryachevo 2

Fig. 11 Water cut history for well Kudryachevo 9