Key Aspects of Project Design for PolymerFlooding at the Daqing Oil FieldDongmei Wang, Exploration and Development Research Institute of Daqing Oil Field Company; R.S. Seright,New Mexico Petroleum Recovery Research Center; Zhenbo Shao and Jinmei Wang, Exploration andDevelopment Research Institute of Daqing Oil Field CompanySummaryThispaperdescribesthedesignproceduresthatledtofavorableincremental oil production and reduced water production during 12years of successful polymer flooding in the Daqing oil field. Spe-cial emphasis is placedonsomenewdesignfactors that werefoundtobeimportantonthebasisofextensiveexperiencewithpolymer flooding. These factors include (1) recognizing when pro-filemodificationis neededbeforepolymer injectionandwhenzone isolation is of value during polymer injection, (2) establishingtheoptimumpolymer formulations andinjectionrates, and(3)time-dependent variation of the molecular weight of the polymerused in the injected slugs.For some Daqing wells, oil recovery can be enhanced by 2 to4% of original oil in place (OOIP) with profile modification beforepolymer injection. For some Daqing wells with significant perme-ability differential between layers and no crossflow, injecting poly-mer solutions separately into different layers improved flow pro-files, reservoir sweep efficiency, and injection rates, and it reducedthe water cut in production wells. Experience over time revealedthat larger polymer-bank sizes are preferred. Bank sizes grew from240380 mg/LPV during the initial pilots to 640 to 700 mg/LPVin the most recent large-scale industrial sites [pore volume (PV)].Economics and injectivity behavior can favor changing the poly-mer molecular weight and polymer concentration during thecourseof injectingthepolymer slug. Polymers withmolecularweights from 12 to 35 million Daltons were designed and suppliedto meet the requirements for different reservoir geological condi-tions. Theoptimumpolymer-injectionvolumevariedaround0.7PV, depending on the water cut in the different flooding units. Theaverage polymer concentration was designed approximately 1000mg/L, but for an individual injection station, it could be 2000 mg/Lor more. At Daqing, the injection rates should be less than 0.140.20 PV/year, depending on well spacing.IntroductionMany elements have long been recognized as important during thedesignofapolymerflood(LiandNiu2002;JewettandSchurz1970; Sorbie1991; Velaet al. 1976; Taber et al. 1997; Maitin1992; Koning et al. 1988; Demin et al. 1995, 2002; Wang and Qian2002; Wanget al. 2008). This paper spells out someof thoseelements,usingexamplesfromtheDaqingoilfield.TheDaqingoil fieldislocatedinnortheast Chinaandisalargeriver-delta/lacustrine-facies, multilayer, heterogeneous sandstone in an inlandbasin. The reservoir is buried at a depth of approximately 1000 m,withatemperatureof45C. Themainformationunderpolymerflood (i.e., the Saertu formation) has a net thickness ranging fromfrom2.3to11.6mwithanaverageof 6.1m. Theaverageairpermeabilityis 1.1m2, andtheDykstra-Parsons permeabilitycoefficient averages 0.7. Oil viscosity at reservoir temperature av-erages approximately 9 mPas, and the total salinity of the forma-tion water varies from 3000 to 7000 mg/L. The field was discov-ered in 1959, and a waterflood was initiated in 1960. The worldslargest polymerfloodwasimplementedat Daqing, beginninginDecember 1995. By2007, 22.3%of total productionfromtheDaqingoil fieldwas attributedtopolymer flooding. Polymerflooding should boost the ultimate recovery for the field to morethan 50% OOIP10 to 12% OOIP more than from waterflooding.Attheendof2007,oilproductionfrompolymerfloodingattheDaqingoilfieldwasmorethan11.6millionm3(73millionbbl)per year (sustained for 6 years). The polymers used at Daqing arehigh-molecular-weight partially hydrolyzed polyacrylamides(HPAMs).During design of a polymer flood, critical reservoir factors thattraditionally receive consideration are the reservoir lithology, stra-tigraphy, important heterogeneities (such as fractures), distributionof remaining oil, well pattern, and well distance. Critical polymerproperties include cost-effectiveness (e.g., cost per unit of viscos-ity), resistance todegradation(mechanical or shear, oxidative,thermal, microbial), toleranceofreservoirsalinityandhardness,retentionbyrock,inaccessibleporevolume,permeabilitydepen-dence of performance, rheology, andcompatibilitywithotherchemicals that might be used. Issues long recognized as importantfor polymer-bank design include bank size (volume), polymer con-centrationandsalinity(affectingbankviscosityandmobility),andwhether (andhow) togradepolymer concentrationsinthechase water.This paper describes the design procedures that led to favorableincremental oil production and reduced water production during 12years of successful polymer flooding in the Daqing oil field.Zone Management Before/During PolymerFloodingProfile Modification Before Polymer Injection. Under some cir-cumstances, use of gel treatments or other types of profile modi-ficationmethods maybeof valuebeforeimplementation-of apolymer or chemical flood (Seright et al. 2003). If fractures causeseverechanneling, gel treatmentscangreatlyenhancereservoirsweepifappliedbeforeinjectionoflargevolumesofexpensivepolymer or surfactant formulations (Trantham et al. 1980). Also, ifone or more high-permeability strata are watered out, there may beconsiderablevalueinapplyingprofile-modificationmethodsbe-fore starting the enhanced-oil-recovery (EOR) project.For some Daqing wells with layers with no crossflow, numeri-cal simulation demonstrated that oil recovery can be enhanced by24% of OOIP with profile modification before polymer injection(Chen et al. 2004). (1012% of OOIP was the typical EOR becauseof polymer flooding alone.) As expected, the benefits from profilemodificationdecreaseifitisimplementedtowardthemiddleorend of polymer injection.OnthebasisoffieldexperiencewithprofilemodificationatDaqing, the following basic principles were observed.Wells That Are Candidates for Profile Modification.1. The downhole pressure in injection wells during intial poly-mer injection is lower than the average level for injectors in the site(Yuan 2005).2. Thepressureinjectionindex, PI, (apressure-transient-testresult)islessthantheaveragevalueinthepilotsite(Qiaoetal.2000). PI, is defined byPI = 1t 0tptdt pt, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (1)Copyright 2008 Society of Petroleum EngineersThis paper (SPE 109682) was accepted for presentation at the 2007 SPE Annual TechnicalConferenceandExhibition,Anaheim,California,1114November,andrevisedforpubli-cation. Original manuscript received for review 30 July 2007. Revised manuscript receivedfor review 27 March 2008. Paper peer approved 14 April 2008.1117 December 2008 SPE Reservoir Evaluation & Engineeringandisthewell pressureaftertheinjectorisshut infortime, t,duringapressure-transient test (at Daqing, thesetransient teststypically last 90 minutes).3. Injection pressures are less than the average level andthewatercutsat theoffset productionwellsarelargerthantheaverage level.Layers That Are Candidates for Profile Modification.1. Chooselayersthatshowgoodlateralconnectivitybetweenwells, that have high permeability differential with respect to ad-jacentlayers,thathavehighpermeabilityandthicknetpay,andthat exhibit effective-permeability barriers between adjacentzones. (Here, permability differential is defined as permeability ofthe high-permeability zone divided by the permeability of the low-permeability zone.)2. Choose layers with a high water cut, a high water saturation,or that appear watered-out.3. Chooselayerswithalargedifferenceinwaterintakefromother layers. (Water-intake index is given by injection rate dividedbyp times net pay of the injectors.) (Yuan 2005).Separate-Layer Injection. If crossflow can occur between adja-cent strata, sweepintheless-permeablezonescanbealmost asgreatasthatinthehigh-permeabilityzonesifthemobilityratiotimesthepermeabilitydifferential islessthanunity(ZhangandSeright 2007). However, if nocrossflowoccursbetweenstrata,sweep in the less-permeable zone will be no better than approxi-mately the square root of the reciprocal of the permeability differ-ential (Zhang and Seright 2007). At Daqing, a means was devisedto improve this sweep problem when crossflow does not occur. OnthebasisoftheoreticalstudiesandpracticalresultsfromDaqingpilot tests (Wu et al. 2005; Wang and Qian 2002), separate-layerinjection was found to improve flow profiles, reservoir-sweep ef-ficiency, andinjectionrates, andit canreducethewater cut inproduction wells. Numerical-simulation studies reveal that the ef-ficiencyofpolymerfloodingdependsimportantlyontheperme-abilitydifferentialbetweenlayersandatthetimeatwhichsepa-rate-layer injection occurs.Anexample basedonnumerical simulationis providedinTable 1, in which the permeability differential was 2.5 and flood-ingoccurreduntil 98%watercut wasreached. Inthiscase, theincremental recovery using layer separation was 2.04% more thanthe case with no layer separation. The simulations were performedwith an in-house simulator called Polymer. This simulator was animproved version of the UTCHEM (University of Texas, Austin,Texas, Version 6.0, 1997), and is a 3D multiphase multicomponentcompositional variable-temperature finite-difference numericalsimulator. For the simulations described in Table 1, we used a 3Dmodel that had 17172 gridblocks.Based on numerical simulation, Fig. 1 shows ultimate-recoveryresults for various conditions whentheinjectionratewas heldconstant in different layers (i.e., the same injection rate per unit ofnet pay occurred in all layers). In this figure, the final polymer slughad a polymer-concentration/volume product (i.e., a polymermass) of 600 mg/LPV (Wu et al. 2005). The x-axis plots the delay(expressed in mg/LPV) between the start of initial polymer injec-tion(withnoseparationof injection) andwhenseparate-layerinjection was initiated. The figure shows that separate-layer injec-tion did not affect ultimate oil recovery if the polymer mass at thetimeoflayerseparationwaslessthan200mg/LPV. However,above this value, the total effectiveness decreased with delay of thestart of separate injection. Increased vertical heterogeneity accen-tuated this effect.Resultsfromtheoreticalstudiesandfrompracticalpilottestsindicatethat separate-layerinjectionshouldbeimplementedbe-forethepolymer-concentration/volumeproduct (mass) becomes200mg/LPVifthistechnologycanbeimplementedinthesites(Zhanget al. 2004). For thosewellswhereinjectioncannot beseparatedbecause of technical or other reasons, flowprofilesshouldbecontrolledas well as is practical, andseparate-layerinjection should be implemented when it becomes feasible.Our theoretical studies and pilot tests revealed that the condi-tions that favor separate-layer injectionat Daqinginclude(Wuet al. 2005)1. The permeability differential between oil zones2.5.2. The net pay for the lower-permeability oil zones should ac-count for at least 30% of the total net pay.3. Layers shouldbe separatedbya barrier that is at least1 m thick.Optimization of the Polymer-Injection FormulaImportant factors to optimize when formulating the polymer bankinclude polymer-solutionviscosity, polymer molecular weight,polymer concentration, polymer volume, and injection rate.Polymer-Solution Viscosity. The polymer-solution viscosity is akey parameter to improve the mobility ratio between oil and waterandtoadjust thewater-intakeprofile. Asinjectionviscosityin-creases, the effectiveness of polymer flooding increases. The vis-cosity can be affected by a number of factors. First, for a given setof conditions, solution viscosity increases with increased polymerFig.1Effectonultimaterecoveryoftimeatwhichseparate-layer injection is initiated. Constant total rate.1118 December 2008 SPE Reservoir Evaluation & Engineeringmolecular weight. Second, increased polymer concentration leadstohigherviscosityandincreasedsweepefficiency.Third,asthedegreeof HPAM-polymer hydrolysis increases uptoacertainvalue, viscosity increases. Fourth, as temperature increases, solu-tionviscositydecreases. Polymer degradationcanalsodecreaseviscosity. Fifth, increasedsalinityandhardnessinthereservoirwater also decreases solution viscosity for anionic polymers.The effectiveness of a polymer flood is determined directly bythe magnitude of the polymer viscosity. The viscosity depends onthe qualityof the water usedfor dilution. Achange inwaterquality directly affects the polymer solution viscosity. At Daqing,the water quantity changes with the rainfall, ground temperature,and humidity during the seasons. The concentrations of Ca2+andMg2+inthe water source are lower insummer andhigher inwinter. Consequently, the polymer viscosityis alsorelativelyhigher in summer and lower in winter.Usingamedium-MwHPAMpolymer, theinjectionpolymerconcentration and solution viscosity can be adjusted according toFig. 2. These curves were used during project design for the pilotsite in the center of Saertu at Daqing. The curves were invaluablein adjusting polymer concentrations to respond to changes in waterquality(salinity). Inthisapplication, foramedium-Mwpolymer(12to16millionDaltons), 40mPas was recommended. Thisviscositylevel was sufficient toovercome (1) the unfavorablemobility ratio (i.e., 9.4) and (2) permeability differential up to 4:1.For a high-Mw polymer (17 to 25 million Daltons) or extrahigh-Mw polymer (25 to 38 million Daltons), 50-mPas viscosity couldbe provided cost-effectively. For new polymers that provide spe-cial fluid properties, additional laboratory investigations areneeded before implementation in a polymer flood.Polymer Molecular Weight. The effectiveness of a polymer floodis affectedsignificantlybythe polymer Mw. As illustratedinFig. 3, polymerswithhigher Mwprovidegreater viscosity. Formany circumstances, larger polymer Mw also leads to improved oilrecovery. Results fromnumerical simulation of corefloods(Table 2)verifythisexpectationforcasesofconstant polymer-slug volume and concentration.Ourlaboratorytestswithafixedvolumeofpolymersolutioninjected confirmed that oil recovery increases with increased poly-merMw(Wuetal. 2001). Thereasonissimplythatforagivenpolymer concentration, solutionviscosityandsweepefficiencyincreasewithincreasedpolymerMw. Statedanotherway, tore-cover a given volume of oil, less polymer is needed using a high-Mw polymer than with a low-Mw polymer.Theaboveargumentmustbetemperedbecausethelevelsofmobility and permeability reduction (i.e., the resistance factor andresidual resistancefactor)forpolymerwithagivenMwcanin-crease with decreasing permeability (Vela et al. 1976). (Resistancefactor, Fr, is defined as brine mobility divided by polymer-solutionmobility. Residualresistancefactor, Frr, isdefinedasbrinemo-bilitybefore polymer floodingdividedbybrine mobilityafterpolymer flooding.) This effect is accentuatedas Mwincreases.Mechanical entrapment canretardpolymer propagationsignifi-cantly if the pore-throat size and permeability are too small. Thus,dependingonMwandpermeabilitydifferential, this effect canreducesweepefficiency(ZhangandSeright 2007). Atrade-offmustbemadeinchoosingthehighest-Mwpolymerthatwillnotexhibit pore plugging or significant mechanical entrapment in theless-permeable zones.Two factors should be considered when choosing the polymerMw. First, choosethepolymer withthehighest Mwpractical tominimizethepolymer volume. Second, theMwmust besmallenough so that the polymer can enter the reservoir rock and propa-gate effectivelythroughit. For a givenrockpermeabilityandpore-throat size, athresholdMwexists, abovewhichpolymersexhibit difficulty in propagation.OnthebasisoflaboratoryresultsandpracticalexperienceatDaqing, themediumpolymer Mw(12to16millionDaltons) isapplicable for oil zones with average permeability greater than 0.1m2and net pay greater than 1 m. The high polymer Mw (17 to 25million Daltons) is appropriate for oil zones with average perme-abilitygreaterthan0.4m2. Table 3showsrecoveryresultsforvarious combinations of polymer Mwand core permeability.Table 4 lists Fr and Frr for different combinations of polymer Mwand core permeability.Economics andinjectivitybehavior canfavor changingthepolymer Mw during the course of injecting the polymer slug. Thispoint can be appreciated from Table 5, which considers six caseswhere a bank of high-Mw (17.9 million Daltons) HPAM polymerwas injected before switching to a bank of medium-Mw (12 millionDaltons) HPAMpolymer. Inthis simulationexample, thetotalpolymer concentration (1000 mg/L) and bank size (570 mg/LPV)wereheldconstant. Theleft-handcolumnof Table5lists thepercentage of the total bank that involved injection of the high-Mwpolymer, whilethemiddlecolumnliststhetotal EORfromthepolymer flood. The right-hand column lists the incremental EOR,compared with using only the medium-Mw polymer.Details of the numerical-simulationmodel canbe foundinShao et al. 2001. In this simulator, polymer retention is modeledboth by adsorption and by mechanical entrapmentvarying withFig.2Viscosityvs.concentrationfordifferentsalinitieswithmedium-Mw (15 million Daltons) polymer (Wang et al. 2001).Fig. 3Viscosityvs. concentrationandvs. Mwfor polymersused in the central part of Xing4-5 (Gao and Su 2004).1119 December 2008 SPE Reservoir Evaluation & Engineeringpermeability, salinity, andpolymerconcentration. Polymersolu-tionsareallowedtoreduceboththemobilityofdisplacingfluid(i.e., resistance factor) and the effective permeability of the porousmedium (i.e., residual resistance factor). Polymer-solution viscos-ity depends on the salinity and polymer concentration (Sun et al.1989). Both shear thinning and viscoelastic polymer rheology arebuilt into the models polymer properties. Furthermore, our modelincludesaviscoelasticpolymer propertythat actstoreducere-sidual-oil saturation below the value expected from waterfloodingor conventional polymer flooding.InTable5, theEORfrominjectionof onlythehigher-Mwpolymer was 3.0% of OOIP greater than that from injection of onlythe medium-Mw polymer (12 million Daltons). Most of the benefit(i.e., 2%) was achieved if only 20% of the bank had the high-Mwpolymer. Increasing the high-Mw-polymer fraction in the bank pro-videdlittle additional increase inoil recovery. If injectivityislower andcost ishigher for thehigh-Mwpolymer thanfor thelower-Mw polymer, a significant benefit can be achieved by chang-ing (i.e., decreasing) the polymer Mw during injection of the poly-merslug. Fortheparticularexamplehere, thehigh-Mwpolymercosts 1.1 to 1.3 times more than the medium-Mw polymer. For thesame polymer concentration, injectivity for the high-Mw polymerwas 70%of that for themedium-Mwpolymer. Givenapproxi-mately the same incremental oil recovery, greater injectivity, andlowerpolymercostleadstoacceleratedoilrecoveryandgreaternet present value for the project.Polymer-SolutionConcentration. Polymer-solutionconcentra-tion determines the polymer-solution viscosity and the size of therequiredpolymer-solutionslug. Thepolymer-solutionconcentra-tion dominates every index that changes during the course of poly-mer flooding.1. Higherinjectionconcentrationscausegreaterreductionsinwater cut and can shorten the time required for polymer flooding.For a certain range, they can also lead to an earlier response timeintheproductionwells, afasterdecreaseinwatercut, agreaterdecreaseinwatercut, lessrequiredPVofpolymer, andlessre-quired volume of water injected during the overall period of poly-mer flooding. On the basis of numerical simulation, Table 6 showstheeffectivenessof polymer floodingasafunctionof polymerconcentrationwhentheinjected-polymermassis640mg/LPV.As polymer concentration increases, EOR increases and the mini-mum in water cut during polymer flooding decreases.2. Aboveacertainvalue, theinjected-polymer concentrationhas little effect on the efficiency of polymer flooding. For a pilotproject, the economics of injecting higher polymer concentrationsshouldbeconsidered. Thepolymer-solutionconcentrationhasalargeeffect onthechangeinwatercut. However, considerationshouldalsobegiventothefact that higher concentrationswillcause higher injection pressures and lower injectivity. Consideringthe technical feasibilityandconditions at Daqing, the averageinjectionpolymerconcentrationrangesfrom1000to1400mg/Lfor our projects. For individual wells, theconcentrationcanbeadjusted to meet particular conditions.3. Additional steps can increase effectiveness when using slugswithhigher polymer concentrations. First, effectiveness canbeimprovedbyinjectingpolymersolutionswithhigherconcentra-tions during the initial period of polymer flooding. The increase ineffectivenesscomesfromthewellsortheunitsthatexperiencedin-depthvertical-sweepimprovement duringtheearlystagesofpolymer flooding. Second, theincreaseinwater cut duringthethird stage of polymer flooding (i.e., after the minimum in watercut) can be controlled effectively using injection of higher polymerconcentrations. Onthebasis of pilot tests intwoareas of theDaqing field (the Western of Central area and the 1-4# Stationin the Beixi of Lamadian area), the water-intake profile becamemuchmoreuniformafterinjecting2200-to2500-mg/Lpolymersolution in 2004 (Fulin et al. 2004; Yang et al. 2004).Polymer retention also plays an important role in determining theappropriate polymer concentration. Sufficient polymer must beinjected to allow the polymer to propagate most of the way throughthe reservoir. Laboratory measurements using Daqing core mate-rial revealed retention values of 126 g/g for a 15-million-DaltonHPAM and 155 g/g for a 25-million-Dalton HPAM. Calculationssuggest that injectionof 1PVof 1000-mg/Lpolymer solutionwould experience 65% depletion by retention for the 15-million-Dalton polymer and 80% depletion for the 25-million-Dalton poly-mer. Polymer-banksizes at Daqingare designedfor effectivepropagation using these polymer-retention levels. The actual vis-cosities of the polymer solutions that flow in the formation and thatareproducedfromtheformationhavebeenmeasured. Flowingpolymer solutions within the formation typically have a viscosityof approximately 20 cp. Mechanical degradation of the polymer isbelieved to be the primary cause of viscosity loss from the surface(where the originally injected polymer typically has a viscosity ofapproximately 40 cp) to locations deep within the reservoir (Wanget al. 2008; Seright 1983).Polymer Volume. Oil-recovery efficiency decreases with in-creased mobility of the injectant (Jang 1994; Craig 1991). A con-1120 December 2008 SPE Reservoir Evaluation & Engineeringtinuous polymer flood could be used instead of a waterflood. How-ever, becausepolymersolutionsaremoreexpensivethanwater,economics limit the volume of polymer that should be injected.For the first polymer pilot tests at Daqing, the polymer-volume/concentration product (mass) was designed to be 240 to 380 mg/LPV. Later, the polymer mass was increased to 570 mg/LPV. Atpresent, the polymer mass is 640 to 700 mg/LPV in the large-scaleindustrial sites.Onthebasisofourtheoreticalresearchandpracticalexperi-ences, thepolymer volumeshouldbedeterminedbythegrosswater cut of the flooding unit. Generally, when the gross water cutreaches 9294%, the polymer injection should be stopped.On the basis of statistics from the 14# Station of the easternLamadian area of the Daqing oil field, the rate of increase in thewater cut isapproximatelythesameduringthelast part of thepolymer flood as it is during the follow-up waterdrive. GenerallyatDaqing,polymerinjectionisswitchedtowaterdrivewhenthewater cut reaches 92 to 94%. On the basis of our observations ofthe response to polymer flooding, we characterize the entire poly-mer-floodingprocessinfivestages,asillustratedin Fig. 4(Guoet al. 2002; Liao et al. 2004) and described in the following:1. The initial stage of the polymer flood in which water cut hasnot yet started to decrease. This stage ranges from the very begin-ning of polymer injection typically to 0.05 PV. During this time,the polymer solution has not begun to work.2. The response stage in which a decrease in water cut can beseen. At Daqing, this stage typically occurs from 0.05 to 0.20 PVof polymer injection. During this time, the polymer solution pen-etrates deep into the formation and forms the oil bank. Typically,approximately 15% of the EOR is produced during this stage.3. The period in which the water cut change is relatively stable.Theminimumwater cut wasobservedduringthisperiod. Thisstagetypicallylastsfrom0.20to0.40PVofpolymerinjection.The oil-production rate reaches its peak value, and approximately40% to the total EOR is produced during this stage. Oil productionbeginstodecrease, andtheproduced-polymerconcentrationbe-gins to increase.4. Thestageinwhichwatercutincreasesagainrapidly. Thisstagetypicallylastsfrom0.40to0.70PVofpolymerinjection.Areal sweepreachesitsmaximum; oil productiondeclines, thepolymer concentrationandtheinjectionpressurefollowsteadytrends.Approximately30%ofthetotalEORisproducedduringthis stage.5. The stage of the follow-up waterdrive. This stage lasts fromthe end of polymer injection to the point at which water cut reaches98%. Water cut increases continuously; the produced-polymerconcentration declines rapidly; and the fluid-production capabilityincreases a little. The EOR produced during this stage is approxi-mately 1012% of the total.Onthebasisof numerical simulation, ultimateoil recoverybe-comes less sensitive to bank size as the polymer mass increases tohigh values. Of course, increased polymer mass enhances the oilrecovery(middlecolumnof Table 7).However,theincrementalpolymer efficiency (defined in the right-hand column of Table 7)decreaseswithincreasedpolymermass.Consequentally,thereisan economic optimum in the polymer-bank size.For largepolymer banks, polymer wasproducedfromwellsafter the water cut increased back up to 92%. So, more extendedinjectionof polymer hurts income andeconomics because theproduced polymer is effectively wasted.Fig. 5 plots the incremental oil (expressed per ton of polymerinjected).Onthebasisofoureconomicevaluation,optimumef-fectiveness canbe obtainedif a suitable time toendpolymerinjection is chosen, followed by a water-injection stage. ForDaqing, the optimum polymer mass ranged from 640 to 700 mg/LPV. Projects at Daqing were profitable within this range, for theoil prices experienced over the past 12 years.Fig. 4Five stages of water cuts during a polymer flood.1121 December 2008 SPE Reservoir Evaluation & EngineeringNumerical simulationandour economicevaluationrevealedthatwhenincomefromthepolymerprojectmatchedtheinvest-ment (i.e., thebreak-evenpoint), theincremental oil was 55metric tons of oil per metric ton of polymer (64 m3of oil per m3of polymer) [when the oil price was 1280 Chinese Yuan per m3orapproximately USD 25.5/bbl, and the polymer mass was 750 mg/LPV (See Fig. 5)]. Of course, the optimum polymer mass dependson oil price. The break-even point is reduced to 36 metric tons ofoil per ton of polymer (42 m3of oil per m3of polymer) when theoil priceincreasestoUSD45/bbl (seeFig. 6). Withthecurrenthigh oil prices, greater polymer masses could be attractive.Injection Rate. The polymer-solution injection rate is another keyfactor in the project design. It determines the oil-production rates.Table 8 shows the effect of injection rate on the effectiveness ofpolymer flooding. It shows that the magnitude of the injection ratehas little effect on the final recovery. It also has a minor effect onthefractionoftheinjectedpolymermassthat isultimatelypro-duced (fourth column from left of Table 8). However, the injectionratehasasignificant effect onthecumulativeproductiontime.Lower injection rates lead to longer production times. So when weprogram the design, the injection rate should not be too small.Fig. 7 shows how reservoir pressure changes with the injectionrateafterthecompletionofpolymerinjection. Asexpected, theaveragereservoirpressureneartheinjectorsincreasesasthein-jectionrate increases while decreasingnear productionwells.Also, higher injection rates cause a larger disparity between injec-tion and production. Injection rates must be controlled (i.e., not toohigh)tominimizepolymerflowoutofthepatternoroutofthetarget zones.Changes inthebehavior of gross water cut withtimealsodepend on the injection rate (Fig. 8). Lower injection rates lead toslower increases in water cut and delay the time when lowest watercut beginstoincrease. Asaresult, thestableperiodwithlowerwater cut isextended. If Fig. 8werereplottedwithPVonthex-axis, the curves would coincide. In western nations, economicsmight favor accelerating injection and production to the maximumallowableextent.However,inChina,thistendencyismoderatedbyaneedforsustainabledevelopmentoftheoilfield, sothatamore protractedoil productionbecomes optimum. Inaddition,early polymer breakthrough becomes a greater risk for higher in-jection rates (which would significantly, and undesirably, alter theshape of a given curve in Fig. 8).The injection rate determines the time period when oil can berecoveredeconomically. Fig. 9showshowinjectionratesaffectthe oil-production rate. It also demonstrates how the term of eco-nomic production varies with injection rate. To maximize the termof oil production and maximize ultimate production, the injectionrate at Daqing should be maintained under 0.16 PV/yr with 250-mwell spacing.Fig. 5Incremental oil vs. polymer mass (Shao et al. 2005).Fig. 6Incremental oil vs. investment/profitratio(Shaoetal.2005).1122 December 2008 SPE Reservoir Evaluation & EngineeringInsummary,theinjectionrateaffectsthewholedevelopmentand effectiveness of polymer flooding. Eq. 2 can be used to relatethe highest pressure at the injectionwellheadandthe averageindividual injectionratewiththepolymer-injectionrateandtheaverage apparent-water-intake index for different reservoir condi-tions. In general, the injection rate should not exceed the reservoir-fracture pressure.pmax = l2q180 Nmin, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . (2)wherepmaxishighest wellheadpressure, MPa; l isthedistancebetween injector and producer, m; is porosity, %; Nmin is lowestapparent-water-intake index, m3/(dmMPa); andqis injectionrate, PV/year.For the best results at Daqing, the polymer-injection rate shouldbe designed to be 0.14 to 0.16 PV/yr for 250-m well spacing, andto be 0.16 to 0.20 PV/yr for 150- to 175-m well spacing.ConclusionsOnthebasisofmorethan12yearsofexperienceattheDaqingoil field, we identified key aspects of project design forpolymer flooding.1. For some Daqing wells, oil recovery can be enhanced 2 to 4%of OOIP with profile modification before polymer injection.2. For some Daqing wells with significant permeability differentialbetweenlayersandnocrossflow, injectingpolymer solutionsseparatelyintodifferent layersimprovedflowprofiles, reser-voir-sweepefficiency, andinjectionrates, andit reducedthewater cut in production wells.3. Experienceovertimerevealedthat largerpolymer-banksizesare preferred. Bank sizes grew from 240 to 380 mg/LPV duringtheinitial pilots to640to700mg/LPVinthemost recentlarge-scale industrial sites.4. Economics andinjectivitybehavior canfavor changingthepolymer molecular weight and polymer concentration during thecourse of injecting the polymer slug. Polymers with molecularweights from 12 to 38 million Daltons were designed and sup-plied to meet the requirements for different reservoir geologicalconditions. The optimumpolymer-injection volume variedaround0.7PV, dependingonthewater cut inthedifferentflooding units. The average polymer concentration was designedto be approximately 1000 mg/L, but for an individual injectionstation, it could be 2000 mg/L or more.5. At Daqing, theinjectionratesshouldbelessthan0.140.20PV/yr, depending on well spacing.NomenclatureDznet net zone height, mFr resistance factorFrr residual resistance factor (permeability before/afterpolymer placement)kair permeability to air,m2keff effective permeability,m2kwater permeability to water,m2l distance between injector and producer, mMw molecular weight, DaltonsNmin lowest apparent-water-intake index, m3/(dmMPa)p pressure, MPapmax highest wellhead pressure, MPaPI pressure index for an injector, MPaPV pore volumesq injection rate, PV/yrt time, minutesVk Dykstra-Parsons coefficient of permeability variationp pressure difference from wellbore to formation, MPa porosityReferencesChen, F., Li, Y., andNiu, J. 2004. Summarizationonthetechnologyofmodification profile in-depth in Daqing. Petroleum Geology & OilfieldDevelopment in Daqing 23 (5): 9799.Craig, F.F. 1991. TheReservoirEngineeringAspectsof Waterflooding.Monograph Series, Richardson, Texas 3: 2977.Demin, W., Jingcun, Z., Fanru, M., Heng, L., Lin, L., and Bohui, H. 1995.CommercialTestofPolymerFloodinginDaqingOilField,DaquingPetroleumAdministrativeBureau. PaperSPE29902presentedattheInternationalMeetingonPetroleumEngineering,Beijing,1417No-vember. DOI: 10.2118/29902-MS.Demin, W., Jiecheng, C., Junzheng, W., andGang, W. 2002. ExperiencesLearnedAfterProductionofMoreThan300MillionBarrelsofOilbyPolymer Flooding in Daqing Oil Field. Paper SPE 77693 presented at theSPE Annual Technical Conference and Exhibition, San Antonio, Texas,29 September2 October. DOI: 10.2118/77693-MS.Fig. 8Water cut changes for different injection rates.Fig. 9Changes in oil-production rate vs. injection rate.Fig. 7Injection rate vs. reservoir pressure.1123 December 2008 SPE Reservoir Evaluation & EngineeringFulin, Y., Demin, W., Xizhi, Y., Xinguang, S., Qinghai, C., andIei, Z.2004. High Concentration Polymer Flooding is Successful. Paper SPE88454 presented at the SPE Asia Pacific Oil and Gas Conference andExhibition, Perth, Australia, 1820 October. DOI: 10.2118/88454-MS.Gao, S., andSu, Y. 2004. Thedevelopment project designof polymerfloodingfor the Central of Xing4-5inDaqing. Yearly Report 12:1617.Guo, W., Cheng, J., andLiao, G. 2002. Thecurrent situationonEORtechniqueinDaqing. PetroleumGeology&OilfieldDevelopment inDaqing 21 (3): 16.Jang, Y. 1994. Optimization Conditions for Polymer Flooding, Chap. 12,35. Beijing: Petroleum Industry Publishing Company of China.Jewett, R.L. andSchurz, G.F. 1970. PolymerFloodingACurrent Ap-praisal. JPT 22 (6): 675684. SPE-2545-PA. DOI: 10.2118/2545-PA.Koning, E.J.L., Mentzer, E., and Heemskerk, J. 1988. Evaluation of a PilotPolymer Flood in the Marmul Field, Oman. Paper SPE 18092 presentedat the SPE Annual Technical Conference and Exhibition, Houston, 25October. DOI: 10.2118/18092-MS.Li, Y. and Niu, J. 2002. The study of adjustment measures during polymerflooding. Yearly Report 12.Liao, G., Niu, J., Shao, Z., andChen, P. 2004. Theeffectiveness andevaluation for industrialized sites by polymer flooding in Daqing. Pe-troleum Geology & Oilfield Development in Daqing 23 (1): 4851.Maitin, B.K. 1992. Performance Analysis of Several PolyacrylamideFloods in North German Oil Fields. Paper SPE 24118 presented at theSPE/DOEEnhancedOil RecoverySymposium, Tulsa, 2224April.DOI: 10.2118/24118-MS.Qiao, E., Li, Y., and Liu, P. 2000. Application of PI decision technique inPuCheng oilfield. Drilling & Production 23 (5): 2825.Seright, R.S. 1983. The Effects of Mechanical Degradation and Viscoelas-tic Behavior on Injectivity of Polyacrylamide Solutions. SPEJ 23 (3):475485. SPE-9297-PA. DOI: 10.2118/9297-PA.Seright, R.S., Lane, R.H., and Sydansk, R.D. 2003. A Strategy for Attack-ing Excess Water Production. SPEPF 18 (3): 158169. SPE-84966-PA.DOI: 10.2118/84966-PA.Shao, Z., Fu, T., and Wang, D. 2001. The determinate method of reason-ablepolymervolume.PetroleumGeology&OilfieldDevelopmentinDaqing 20 (2): 6062.Shao, Z., Chen, P., Wang, D., andWu, L. 2005. StudyoftheDynamicRules for Polymer Flooding in Industrial Sites in Daqing.TheThesisCollection for EOR Technology 12: 18.Sorbie, K.S. 1991. Polymer-ImprovedOil Recovery. Glasgow, Scotland:CRC Press.Sun, C., Wang, D., Zhou, Z., andShao, Z. 1989. TheDevelopment ofPolymer Flooding Simulator. Yearly Report 12: 26.Taber, J.J., Martin, F.D., and Seright, R.S. 1997. EOR Screening CriteriaRevisitedPart 1: IntroductiontoScreeningCriteriaandEnhancedRecoveryFieldProjects. SPERE12(3): 189198. SPE-35385-PA.DOI: 10.2118/35385-PA.Trantham, J.C., Threlkeld, C.B., andPatternson, H.L. 1980. ReservoirDescriptionfor a Surfactant/Polymer Pilot ina Fractured, Oil-WetReservoirNorthBurbankUnit Tract 97. JPT32(9): 16471656.SPE-8432-PA. DOI: 10.2118/8432-PA.Vela, S., Peaceman, D.W., and Sandvik, E.I. 1976. Evaluation of PolymerFlooding in a Layered Reservoir with Crossflow, Retention, and Deg-radation. SPEJ 16 (2): 8296. SPE-5102-PA. DOI: 10.2118/5102-PA.Wang, D. and Qian, J. 2002. The development project design of polymerflooding for Eastern in Sazhong in Daqing. Yearly Report 12: 3435.Wang, D., Han, P., Shao, Z., Hou, W., andSeright, R.S. 2008. SweepImprovement Options for the Daqing Oil Field. SPEREE 11 (1): 1826.SPE-99441-PA. DOI: 10.2118/99441-PA.Wang, Z., Han, C., Hou, J., andWu, W. 2001. Thepolymermolecularweightandthefactorsaffectingflowproperties. J. ofDaqingPetro-leum Transactions 25 (1): 1820.Wu, L., Chen, P., andLu, J. 2005. Studyof injectionparameters forseparate layers during the period of polymer flooding. Petroleum Ge-ology & Oilfield Development in Daqing 24 (4): 7577.Wu, W. et al. 2001. The polymer molecular weight and the factors affect-ing flow properties. J. of Daqing Petroleum Transactions 25 (1): 1820.Yuan, Q. 2005. Termsof Oil/GasReservoir Engineering. ChinaPatentNumber SY/T, 6174.Zhang, G. andSeright, R.S. 2007. ConformanceandMobilityControl:Foams vs. Polymers. Paper SPE 105907 presented at the InternationalSymposiumonOilfieldChemistry, Houston, 28February2March.DOI: 10.2118/105907-MS.Zhang, Y., Li, C., andWang, X. 2004. Discussionsonthetechnical ofstring with backflush for separate layers during the period of polymerflooding. Oil & Gas Ground Engineering 23 (9): 2627.SI Metric Conversion Factorsbbl 1.589 E01 m3cp 1.0* E03 Pasft 3.048* E01 min. 2.54* E+00 cmmd 9.869 233 E04 m2psi 6.894 757 E+00 kPa*Conversion factor is exact.Dongmei Wang is Department Chief Engineer and Senior Res-ervoir Engineer at the Exploration and Development Institute ofthe Daqing Oilfield Company in Daqing, China, where she hasworkedsince1987. E-mail: wangdongmei1@petrochina.com.cn. She holds a PhD degree in oil and gas developmentengineering from the Research Institute of Petroleum Explora-tion and Development at Beijing. Randy Seright is a senior en-gineer at the New Mexico Petroleum Recovery Research Cen-ter at New Mexico Tech in Socorro, New Mexico, where he hasworkedsince1987. E-mail: randy@prrc.nmt.edu. Heholds aPhD degree in chemical engineering from the University of Wis-consin at Madison. Zhenbo Shao is Department Director and asenior reservoir engineer at the Exploration and DevelopmentInstitute of the Daqing Oilfield Company, where he has workedsince 1983. E-mail: shaozhenbo@petrochina.com.cn. He holdsamasters degreeinoil andgas development engineeringfrom the China University of Petroleum at Beijing. Jinmei Wangis a senior engineer at the Exploration and Development Insti-tuteof theDaqingOilfieldCompany. E-mail: wangjinmei@petrochina.com.cn.1124 December 2008 SPE Reservoir Evaluation & Engineering