Preformed-Particle-Gel Extrusion Through Open Conduits During Conformance-Control Treatments.pdf

11
Preformed-Particle-Gel Extrusion Through Open Conduits During Conformance-Control Treatments Abdulmohsin Imqam, Baojun Bai, Mustafa Al Ramadan, and Mingzhen Wei, Missouri University of Science and Technology; and Mojdeh Delshad and Kamy Sepehrnoori, University of Texas at Austin Summary Millimeter-sized (10 lm–mm) preformed particle gels (PPGs) have been used successfully as conformance-control agents in more than 5,000 wells. They help to control both water and CO 2 production through high-permeability streaks or conduits (large pore openings), which naturally exist or are aggravated either by mineral solution or by a high injection pressure during the flood- ing process. This paper explores several factors that can have an important impact on the injectivity and plugging efficiency of PPGs in these conduits. Extensive experiments were conducted to examine the effect of the conduit inner diameter and the PPG strength on the ratio of the particle size to the opening diameter, injectivity index, resistance factor, and plugging efficiency. Five- foot tubes with four internal diameters were designed to emulate the opening conduits. Three pressure taps were mounted along the tubes to monitor PPG transport and plugging performance. The results show that weak gel has less injection pressure at a large particle/opening ratio compared to strong gel. PPG strength affected injectivity more significantly than did particle/opening ratio. Resistance factor increased as the brine concentration and conduit inner diameter increased. PPGs can significantly reduce the permeability of a conduit, and their plugging efficiency depends highly on the particle strength and the conduit inner di- ameter. The particle size of PPGs was reduced during their trans- port through conduits. Experimental results confirm that the size reduction was caused by both dehydration and breakdown. On the basis of the laboratory data, two correlations were developed to quantitatively calculate the resistance factor and the stable injec- tion pressure as a function of the particle strength, particle/open- ing ratio, and shear rate. This research provides significant insight into designing better millimeter-sized particle-gel treatments intended for use in large openings, including open fractures, caves, worm holes, and conduits. Introduction Excess water production in oil fields is becoming a challenging economic and environmental problem as more reservoirs mature. An estimated average of 3 bbl of water is produced for each barrel of oil produced worldwide (Bailey et al. 2000). It is estimated that the total cost to separate, treat, and dispose of this water is approx- imately USD 50 billion per year (Hill et al. 2012). Water can flow into the wellbore as a result of either near-wellbore problems or reservoir-related problems (Seright et al. 2001). The mechanisms that contribute to this undesired water production must be fully understood before the appropriate treatment can be chosen. Unde- sirable water channeling, one of the primary reservoir conform- ance problems, is caused by reservoir heterogeneities that lead to the development of high-permeability streaks. These streaks include open fractures and fracture-like features, such as caves, worm holes, and conduits (Smith et al. 2006). These high-conduc- tivity features inside the reservoir occupy only a small fraction of the reservoir but will capture a significant portion of injected water. As a result, large amounts of oil remain unswept because a large waterflood will bypass oil-rich unswept zones/areas. Gel treatments proved to be a cost-effective chemical con- formance-control technology to reduce the fluid flow in these large-opening features. The application of these technologies can not only control water production but also significantly increase the oil production and extend the economic life of a reservoir. Traditionally, in-situ bulk gels were used for this purpose. How- ever, preformed particle gels (PPGs) recently have attracted much attention because they can solve some of the problems associated with in-situ gel systems, such as the dilution and dispersion of the gelant and chromatographic separation of the gelant solution (Coste et al. 2000; Chauveteau et al. 2001, 2003; Bai et al. 2007a,b). Commercial PPGs currently available include milli- meter-sized particle gel (PPG), microgels, and submicron gel. Their differences are mainly in particle size, swelling ratio, and swelling time. The millimeter-sized particles are not only more distinguishable but also more reliable than other types of particle gels for plugging large-pore-opening features. It is estimated that PPGs have been used to treat more than 5,000 wells (Bai et al. 2013). The success of gel treatments depends heavily on the gel’s ability to extrude through fractures and channels during the place- ment process (Seright 1999). Thus, understanding both the mech- anism and the behavior of gel extrusion is the key to a successful conformance-control treatment. Numerous studies were con- ducted to evaluate in-situ gel propagation through fractures. Seright (1995, 1997, 1998, 1999, and 2001) studied both bulk gel placement and the mechanism behind gel propagation through fracture systems. Liu and Seright (2000) identified a correlation between gel rheology and the extrusion properties of gels in frac- tures. Ganguly et al. (2001) conducted a series of experiments to determine the effects of fluid leakoff on gel strength when placed in fractures. Sydansk et al. (2005) characterized the transport of partially formed gels in fractures. Wang and Seright (2006) exam- ined whether the use of rheology measurements to evaluate gel properties in fractures is an acceptable substitute for extrusion experiments as a way to reduce costs. Wilton and Asghari (2007) worked to determine how to improve bulk gel placement and per- formance through fractures. They applied Cr (III) acetate preflush to investigate whether a stable gel can be formed within a fracture without gelant leakoff. McCool et al. (2009) investigated the effect of shear on flow properties during the placement of gelants in fractures. No one, however, has studied either the performance or the mechanism of PPG extrusion and placement in conduit sys- tems. The only study of PPG extrusion was conducted by Zhang and Bai (2011), who investigated the effect of PPG extrusion through fractures on the injectivity and plugging efficiency when the fracture width was less than the gel-particle size. The objective of this work was to conduct an in-depth exami- nation of several factors that can have an important impact on the PPG-extrusion mechanism and -placement performance in open- ing conduits. This paper intensively examines the effect of the conduit inner diameter and brine concentration (PPG strength) on the injectivity index, resistance factor, gel dehydration, particle/ Copyright V C 2014 Society of Petroleum Engineers This paper (SPE 169107) was accepted for presentation at the SPE Improved Oil Recovery Symposium, Tulsa, 12–16 April 2014, and revised for publication. Original manuscript received for review 1 April 2014. Revised manuscript received for review 17 July 2014. Paper peer approved 28 July 2014. 2014 SPE Journal 1

Transcript of Preformed-Particle-Gel Extrusion Through Open Conduits During Conformance-Control Treatments.pdf

  • Preformed-Particle-Gel ExtrusionThrough Open Conduits During

    Conformance-Control TreatmentsAbdulmohsin Imqam, Baojun Bai, Mustafa Al Ramadan, andMingzhen Wei,Missouri University of Science

    and Technology; andMojdeh Delshad and Kamy Sepehrnoori, University of Texas at Austin

    Summary

    Millimeter-sized (10 lmmm) preformed particle gels (PPGs)have been used successfully as conformance-control agents inmore than 5,000 wells. They help to control both water and CO2production through high-permeability streaks or conduits (largepore openings), which naturally exist or are aggravated either bymineral solution or by a high injection pressure during the flood-ing process. This paper explores several factors that can have animportant impact on the injectivity and plugging efficiency ofPPGs in these conduits. Extensive experiments were conducted toexamine the effect of the conduit inner diameter and the PPGstrength on the ratio of the particle size to the opening diameter,injectivity index, resistance factor, and plugging efficiency. Five-foot tubes with four internal diameters were designed to emulatethe opening conduits. Three pressure taps were mounted along thetubes to monitor PPG transport and plugging performance. Theresults show that weak gel has less injection pressure at a largeparticle/opening ratio compared to strong gel. PPG strengthaffected injectivity more significantly than did particle/openingratio. Resistance factor increased as the brine concentration andconduit inner diameter increased. PPGs can significantly reducethe permeability of a conduit, and their plugging efficiencydepends highly on the particle strength and the conduit inner di-ameter. The particle size of PPGs was reduced during their trans-port through conduits. Experimental results confirm that the sizereduction was caused by both dehydration and breakdown. On thebasis of the laboratory data, two correlations were developed toquantitatively calculate the resistance factor and the stable injec-tion pressure as a function of the particle strength, particle/open-ing ratio, and shear rate. This research provides significant insightinto designing better millimeter-sized particle-gel treatmentsintended for use in large openings, including open fractures,caves, worm holes, and conduits.

    Introduction

    Excess water production in oil fields is becoming a challengingeconomic and environmental problem as more reservoirs mature.An estimated average of 3 bbl of water is produced for each barrelof oil produced worldwide (Bailey et al. 2000). It is estimated thatthe total cost to separate, treat, and dispose of this water is approx-imately USD 50 billion per year (Hill et al. 2012). Water can flowinto the wellbore as a result of either near-wellbore problems orreservoir-related problems (Seright et al. 2001). The mechanismsthat contribute to this undesired water production must be fullyunderstood before the appropriate treatment can be chosen. Unde-sirable water channeling, one of the primary reservoir conform-ance problems, is caused by reservoir heterogeneities that lead tothe development of high-permeability streaks. These streaksinclude open fractures and fracture-like features, such as caves,worm holes, and conduits (Smith et al. 2006). These high-conduc-

    tivity features inside the reservoir occupy only a small fraction ofthe reservoir but will capture a significant portion of injectedwater. As a result, large amounts of oil remain unswept because alarge waterflood will bypass oil-rich unswept zones/areas.

    Gel treatments proved to be a cost-effective chemical con-formance-control technology to reduce the fluid flow in theselarge-opening features. The application of these technologies cannot only control water production but also significantly increasethe oil production and extend the economic life of a reservoir.Traditionally, in-situ bulk gels were used for this purpose. How-ever, preformed particle gels (PPGs) recently have attracted muchattention because they can solve some of the problems associatedwith in-situ gel systems, such as the dilution and dispersion of thegelant and chromatographic separation of the gelant solution(Coste et al. 2000; Chauveteau et al. 2001, 2003; Bai et al.2007a,b). Commercial PPGs currently available include milli-meter-sized particle gel (PPG), microgels, and submicron gel.Their differences are mainly in particle size, swelling ratio, andswelling time. The millimeter-sized particles are not only moredistinguishable but also more reliable than other types of particlegels for plugging large-pore-opening features. It is estimated thatPPGs have been used to treat more than 5,000 wells (Bai et al.2013).

    The success of gel treatments depends heavily on the gelsability to extrude through fractures and channels during the place-ment process (Seright 1999). Thus, understanding both the mech-anism and the behavior of gel extrusion is the key to a successfulconformance-control treatment. Numerous studies were con-ducted to evaluate in-situ gel propagation through fractures.Seright (1995, 1997, 1998, 1999, and 2001) studied both bulk gelplacement and the mechanism behind gel propagation throughfracture systems. Liu and Seright (2000) identified a correlationbetween gel rheology and the extrusion properties of gels in frac-tures. Ganguly et al. (2001) conducted a series of experiments todetermine the effects of fluid leakoff on gel strength when placedin fractures. Sydansk et al. (2005) characterized the transport ofpartially formed gels in fractures. Wang and Seright (2006) exam-ined whether the use of rheology measurements to evaluate gelproperties in fractures is an acceptable substitute for extrusionexperiments as a way to reduce costs. Wilton and Asghari (2007)worked to determine how to improve bulk gel placement and per-formance through fractures. They applied Cr (III) acetate preflushto investigate whether a stable gel can be formed within a fracturewithout gelant leakoff. McCool et al. (2009) investigated theeffect of shear on flow properties during the placement of gelantsin fractures. No one, however, has studied either the performanceor the mechanism of PPG extrusion and placement in conduit sys-tems. The only study of PPG extrusion was conducted by Zhangand Bai (2011), who investigated the effect of PPG extrusionthrough fractures on the injectivity and plugging efficiency whenthe fracture width was less than the gel-particle size.

    The objective of this work was to conduct an in-depth exami-nation of several factors that can have an important impact on thePPG-extrusion mechanism and -placement performance in open-ing conduits. This paper intensively examines the effect of theconduit inner diameter and brine concentration (PPG strength) onthe injectivity index, resistance factor, gel dehydration, particle/

    CopyrightVC 2014 Society of Petroleum Engineers

    This paper (SPE 169107) was accepted for presentation at the SPE Improved Oil RecoverySymposium, Tulsa, 1216 April 2014, and revised for publication. Original manuscriptreceived for review 1 April 2014. Revised manuscript received for review 17 July 2014. Paperpeer approved 28 July 2014.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 1 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2014 SPE Journal 1

  • opening ratio, gel washout, and plugging efficiency. It also dis-cusses new correlations that were developed to predict the PPGresistance factor and the initial stable injection pressure during gelextrusion through conduit systems.

    Experimental Description

    Materials. Preformed Particle Gel (PPG). A superabsorbentpolymer was used as a PPG to conduct the experiments. The parti-cle was synthesized with acrylamide, acrylic acid, and N, N0-methylenebisacrylamide by a free-radical process. Dry particleswith a mesh size of 30 were swollen in different concentrations ofNaCl brine (0.05, 0.25, 1, and 10%). The brine concentration wascarefully selected on the basis of the swelling ratio and the gelstrength after swelling, as shown in Table 1. PPG swollen inlower-NaCl-concentration brine will have higher swelling ratioand lower gel strength. PPG concentration was determined withthe initial weight of dry gel divided by the final weight of com-pletely swollen gel. PPG concentration is changed as a result ofthe brine-concentration effect.

    Tubes. Tubes 5 ft (1.5 m) in length with varying internaldiameters (10.922, 3.048, 1.752, and 0.774 mm) were used toemulate different conduit sizes. Three pressure taps were mountedalong the tube to monitor PPG-propagation performances. The in-ternal diameters were carefully selected to be larger than, equalto, and smaller than the swollen particles.

    Microscope. A microscope was used to determine the particlesize before and after particle extrusion through the conduit mod-els. An image-analysis technique was used to obtain the particle-gel size distribution.

    Rheoscope Device. Storage moduli (G0) for PPG prepared indifferent brine concentrations were measured at room temperature(approximately 23C) with a rheoscope. The PPG strength wasmeasured before and after gel propagation into the conduit todetermine the effect of the extrusion process on strength. The sen-sor used for measurements is PP335 TiPoLO2 016 with a gap of0.2 mm between the sensor and the plate. G0 were measured at afrequency of 1 Hz for each sample.

    Experimental Setup

    Fig. 1 provides a schematic of the conduit model used to conductthe experiments. This model contained a syringe pump that wasused to inject brine and gel through the accumulator into a 5-ft tube.The tube was divided into three sections: The first two were 2 ftlong and the last section was 1 ft long. Effluent gel and brine wereboth collected to evaluate the gels properties after the extrusion.

    Experimental Procedure

    Dry preformed particle gels (PPGs) of 30-mesh size were placedin different-concentration brines and left overnight to swell fully.

    A sieve was used to allow the swollen gel to separate from theexcess brine solution. The gel then was packed into a stainless-steel accumulator so that it could be injected into a conduit model.The gel-injection process at ambient temperature is summarizedas follows:

    The PPGs were injected into different internal tubes at thesame designed velocity. Table 2 summarizes the velocitiesused for the different inner diameters.

    The gel initially was injected at a high velocity, which thenwas reduced gradually for all experiments. The pressureneeded to be stable for each gel-injection velocity.

    Following pressure stabilization, gel samples were taken foreach gel-injection velocity to measure gel strength and parti-cle size.

    Finally, when the gel-injection process was complete, thesame-concentration brine was injected into the tube filledwith particles from a low to a high velocity to determine gelresistance to water flow.

    Results and Analysis

    Preformed particle gels (PPGs) swollen in four different-concen-tration brines were injected into three sizes of conduits at variousinjection velocities to investigate the effect of brine concentration(related to gel strength) on injectivity, the resistance factor, andthe threshold pressure. The resistance-factor and gel-injection-pressure data were used to develop new correlation models forPPG to predict the resistance factor and the initial stable injectionpressure during gel extrusion in conduits.

    Injectivity-Index-Calculation Results. An injectivity index wasobtained as a function of the brine concentration, injection veloc-ity, and conduit inner diameter to observe the behavior of pre-formed particle gels (PPGs) that had been extruded through theconduit systems. Fig. 2 shows the effect of the brine concentra-tions and gel-injection velocity on the gel-injection pressurethrough three different sizes of conduits. At the same injection ve-locity, the gel-injection pressure increased as the brine concentra-tion increased. This occurred because PPG swollen in low brineconcentration swelled more and became weaker than the PPGswollen in high brine concentration. The gel-injection volumerequired to achieve stable pressure is varied, depending on thebrine concentration and the conduit inner diameter. A large vol-ume of gel was injected as the gel become stronger and the con-duit inner diameter become smaller. For the conduit with theinner diameter of 1.752 mm, the injected PPG pore volume (PV)required to get stable pressure increased from 11.5 PV to 33.9 PVwhen brine concentration increased from 0.05% NaCl to 10%NaCl. The results also show that the gel-injection pressureincreased as the injection velocity increased. This increase in thegel-injection pressure became insignificant when the gel-injectionvelocity exceeded 500 ft/D. This suggests that the gel-injectionpressure did not increase linearly through all of the gel-injectionvelocities, but rather tended to reach a plateau after a certaininjection velocity. This insignificant increase most likely occurredbecause of the gel slip that can occur when extruding through con-duits at a high velocity (Seright 1997). Our results were consistentwith Seright (1997, 1998) for gel extrusion through tubes in whichhe observed that gel-injection pressure became independent ofinjection velocity after a specific velocity value.

    The data from Fig. 2 were used to obtain the gel-injectivityindex through the conduit systems. PPGs with a high injectivityindex required a lower injection pressure to be propagated through

    No. Brine Concentration(%NaCl)

    PPG Concentration (wt%)

    Swelling Ratio

    Gel Strength (Pa)

    515 561 06.0 50.0 1 756 08 52.1 52.0 2 078 05 0.2 1 3 0031 52 0.4 01 4

    Table 1Swelling and gel-strength measurements of 30-mesh PPG.

    DI water

    Syringe Pump

    PistonBrine PPG Pressure Gauges

    Tubing (5 ft)

    2 ft

    P1 P2 P3

    2 ft 1 ft

    Effluent

    Fig. 1Schematic of the conduit model used in this experiment.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 2 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2 2014 SPE Journal

  • the conduit. In this study, the injectivity index increased as thebrine concentrations decreased, as shown in Fig. 3. This likelyoccurred as a result of the swelling-ratio effect. PPGs swollen inlow brine concentrations contain a high percentage of aqueous

    phase and a low percentage of solid phase. This compositionallows PPGs swollen in low brine concentrations to be moreinjectable than PPGs swollen in high brine concentrations. Theseresults also indicate that the injectivity index increased as both theconduit inner diameter and the velocity increased. For the conduitsize, this behavior is easy to understand, but for the velocity, thisrelationship occurred because the gel followed the shear-thinningor pseudoplastic behavior in the conduit systems.

    Resistance-Factor (Fr) -Calculation Results. In analogy to aporous-media experiment, the Fr was estimated from the injectiv-ity index and the geometry of the conduit: QL/ADP for the brineand the gel injection. One can define it as the ratio of the particle-gel-injection pressure drop to the brine-injection pressure drop atthe same flow rate, and one can calculate it from the followingequation:

    Fr DpPPG=Dpbrine; 1

    where Q is the injection flow rate, L is the length of the conduit,DpPPG is the PPG-injection-pressure drop, and Dpbrine is the brine-injection-pressure drop before PPG placement.

    Effect of Brine Concentration and Conduit Inner Diameteron Fr. PPGs swollen in four different concentrations of brineswere injected into three conduits at various injection velocities todetermine the effect of brine concentration and conduit inner di-ameter on the Fr. The injection began with the highest injectionvelocity until the injection pressure became stable. Then, theinjection continued at reduced velocities. A stable pressure wasrecorded at each injection velocity. Fig. 4 indicates that for allgel-velocity injections, Fr increased as the brine concentrationand conduit inner diameter increased. The Fr measured across allthree conduits became an independent factor on velocity when itexceeded 500 ft/D. The Fr value for the gel swollen in 10% NaClextruded in 10.922 mm was 99,133; it then decreased

    . . . . . . . . . . . . . . . . . . . . . . . . .

    Conduit InnerDiameter

    (mm)

    Injection Velocity (ft/D)

    979,12.93674,12.92

    0792.915948.97429.48919.3

    999.1941

    10.922

    012.0829,13644,13.2

    4695.128457.014273.039103.0

    6951.08470.0

    3.048

    0110.0139,11844,157.0

    6695.038452.0142521.03911.0

    6950.084520.0

    1.752

    01500.0

    Injection FlowRate (cm3/min)

    Table 2Gel velocities designed for each conduit inner diameter.

    Conduit inner diameter 1.752 mm Conduit inner diameter 3.048 mm Conduit inner diameter 10.922 mm

    1

    10

    100

    1000

    0 1000 2000 3000

    Gel

    -Injec

    tion

    Pres

    sure

    (psi

    )

    Gel-Injection Velocity (ft/D)0 1000 2000 3000

    Gel-Injection Velocity (ft/D)

    10% NaCl 1% NaCl0.25% NaCl 0.05% NaCl

    0 1000 2000 3000Gel-Injection Velocity (ft/D)

    Fig. 2PPG-injection pressure as a function of brine concentration and conduit diameter.

    0.0001

    0.001

    0.01

    0.1

    1

    10

    1 100 10000Gel-Injection Velocity (ft/D)

    Conduit inner diameter 1.752 mm Conduit inner diameter 3.048 mm Conduit inner diameter 10.922 mm

    1 100 10000

    10% NaCl 1% NaCl0.25% NaCl 0.05% NaCl

    1 100 10000Gel-Injection Velocity (ft/D) Gel-Injection Velocity (ft/D)

    Injec

    tivity

    In

    dex

    (cm3 /m

    in/p

    si)

    Fig. 3Injectivity-index results.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 3 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2014 SPE Journal 3

  • substantially to 3,364 as the velocity increased from 10 to 500 ft/D. However, as the velocity increased more than 500 ft/D, the Frvalues decreased only slightly during the gel-injection process.

    Gel Threshold Pressure vs. Particle/Opening Ratio. The parti-cle/opening ratio is defined as the ratio of the gel-particle diame-ter (Dg) before the extrusion to the pore-opening conduit diameter(Dp). The 30-mesh preformed particle gels swollen in different-concentration brines had different sizes and strengths, as shown inTable 1. The gel threshold pressure (Pt) is the minimum pressurerequired to initiate gel flow through the conduit. Fig. 5 illustratesthe relationship between the threshold pressure and the particle/opening ratio. Strong gel requires a higher threshold pressure thanweak gel in order for it to pass through an opening. The resultobtained agrees with Seright (1997, 1998), who observed thatsome threshold pressure was required before the gel wouldextrude through a given opening size. The data also suggest thatwhen the particle/opening ratio exceeded 2:1, the threshold pres-sure for both strong and weak gel increased much less comparedto when the ratio was below 2:1. This may have occurred for tworeasons. First, the swollen-particle dehydration during the extru-sion process may have reduced the size of the particles as the ratioincreased. Second, the gel particles broke into small pieces, whichmay lead to smaller increases in the threshold pressure withincreasing particle/opening ratio.

    Stabilized-Gel Injection Pressure vs. Particle/Opening Ratio.After the preformed particle gels passed through the conduit, gelwas injected continuously until the injection pressure stabilized.The injection pressure of the stable gel was measured as a func-tion of the gel strength and particle/opening ratio, as shown in

    Fig. 6. The results show that the stable injection pressureincreased with the gel strength and particle/opening ratio. The gelstrength had a significant effect on the stability of the injectionpressure, more than did the particle/opening ratio. The gel-injec-tion pressure increased by approximately 10 times (100 to 1,320psi) when the gel strength approximately doubled from 515 to1300 Pa. The injection pressure only tripled (191.7 to 590 psi)when the particle/opening ratio approximately doubled from 0.72to 1.26 at the gel strength of 1300 Pa.

    Correlation Models. To have correlation models that can predictthe resistance factor (apparent viscosity) and stabilized injectionpressure for preformed particle gels (PPGs) during gel treatmentsis important to quantify the gel-transport process. Such modelsnot only can be inserted into a simulator to yield better predictionsof PPG performance, but also can provide results more quickly,because conducting all of these experiments in the laboratorywould be time consuming and would require a great amount ofeffort to achieve reliable results.

    PPG Resistance-Factor (Fr) Model. Polymer or polymer-gelviscosity is often expressed as a function of shear rate; therefore,we tried to correlate the Fr with shear rate. Here, we use the maxi-mum shear rate at the pore wall to obtain shear-rate values, andthe equation is given as follows (Zaitoun et al. 2012):

    c 8v=D; 2

    where c is the shear rate, v is the superficial velocity, and D is theconduit inner diameter.

    The data in Fig. 4 were reorganized to Fig. 7 after convertingvelocity to shear rate. One can see that all data are in the same linefor the particle prepared by the same-concentration brine eventhough their particle/opening ratios are different, indicating that Frwas independent of conduit inner diameter. This phenomenon isnot very surprising because we know that the polymer-gel viscosityis a function of shear rate but does not depend on the gap betweencylinder and spindle when we measure the bulk-gel viscosity.

    A good fit was noticed with the power-law equation for Frresults plotted against shear rate. The fit is even better with a par-ticle/opening ratio greater than unity or equal to unity. Therefore,the developed model will include the data for particle/openingratios greater than and equal to unity. One can express the Frobtained as a function of the shear rate as

    Fr McE; 3

    where M and E are constant coefficients related to brine concen-tration and particle/opening ratio; both were obtained from gelextrusion through conduits. Table 3 summarizes the resultsobtained for both M and E for each brine concentration.

    To develop a general correlation that one can use to predict theFr for all brine concentrations, both constant coefficients M and E

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    Conduit inner diameter 10.922 mm

    100

    1000

    10000

    100000

    1000000

    1 100 10000

    Resi

    stan

    ce Fa

    ctor

    Gel-Injection Velocity (ft/D)1 100 10000

    10% NaCl 1% NaCl0.25% NaCl 0.05% NaCl

    1 100 10000Gel-Injection Velocity (ft/D) Gel-Injection Velocity (ft/D)

    Conduit inner diameter 3.048 mmConduit inner diameter 1.752 mm

    Fig. 4Resistance factor as a function of brine concentration and conduit inner diameter.

    0.01

    0.1

    1

    10

    100

    0 2 4 6 8

    Thre

    shol

    d Pr

    essu

    re (p

    si)

    Particle-Opening Ratio (Dg/Dp)

    515 Pa Gel strength657 Pa Gel strength870 Pa Gel strength1300 Pa Gel strength

    Fig. 5Effect of particle/opening ratio on threshold pressure.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 4 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    4 2014 SPE Journal

  • need to be determined. Table 3 indicates that E was not affectedvery much by brine concentration, but M was strongly affected.To obtain these coefficients, both constants were plotted as a func-tion of the brine concentration (C), as shown in Fig. 8. Thepower-law equation was used again to obtain the proper fittingcorrelation for the coefficients.

    The constant coefficient (M) was fitted with correlation coeffi-cient squared (R2) of 99%:

    M 3831:3 C0:2709: 4

    The constant coefficient (E) was fitted with the R2 of 96%:

    E 0:6001 C0:01: 5

    Then, one can write the general form of the correlation (whichone can use to predict the Fr in conduit systems) as

    Fr 3831:3 C0:2709c0:6001 C 0:01: 6

    One can also express the obtained correlation as a function of thegel strength. Table 1 clearly indicates that the gel strengthdepends heavily on the brine concentrations; the following corre-lation was fitted with R2 of 99.7% to express the relationshipbetween brine concentrations and gel strength:

    C 3 1017G0 5:6391: 7

    Then, one can modify the correlation to be a function of gelstrength (G0):

    Fr 3831:3 3 1017 G05:63910:2709

    c0:6001 31017 G05:63910:01 : 8PPG Stabilized-Injection-Pressure Model. The data in Fig. 6

    were drawn in log-log scale, as shown in Fig. 9, and were fittedwell by use of the following power-law equation:

    Pinj aDg=Dpb; 9

    where Pinj is the initial stable injection pressure in psi, and a andb are coefficient factors obtained for PPGs extruded through. . . . . . . . . . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . . . . . . . . . .

    . . . . . . . . . . . . . . . . . . . . . . . . . .

    1

    10

    100

    1000

    10000

    0 2 4 6 8

    Stab

    le In

    jectio

    n Pr

    essu

    re (p

    si)

    Particle-Opening Ratio (Dg/Dp)

    515 Pa Gel strength657 Pa Gel strength870 Pa Gel strength1300 Pa Gel strength

    Fig. 6Stabilized injection pressure as a function of particle/opening ratio and gel strength.

    0.05% brine concentration 0.25% brine concentration

    1% brine concentration 10% brine concentration

    10

    100

    1000

    10000

    100000

    0.01 0.1 1 10 100

    Res

    ista

    nce

    Fact

    or

    Dg/Dp = 0.44Dg/Dp = 1.6Dg/Dp = 2.78

    10

    100

    1000

    10000

    100000

    0.01 0.1 1 10 100

    Res

    ista

    nce

    Fact

    or

    Dg/Dp = 0.32Dg/Dp = 1.16Dg/Dp = 2.03

    10

    100

    1000

    10000

    100000

    0.01 0.1 1 10 100

    Res

    ista

    nce

    Fact

    or

    Dg/Dp = 0.29Dg/Dp = 1.04Dg/Dp = 1.82

    10

    100

    1000

    10000

    100000

    0.01 0.1 1 10 100

    Res

    ista

    nce

    Fact

    or

    Dg/Dp = 0.2Dg/Dp = 0.72Dg/Dp = 1.26

    Shear Rate (1/sec) Shear Rate (1/sec)

    Shear Rate (1/sec) Shear Rate (1/sec)

    Fig. 7Resistance factor for gel swollen in brine concentrations as a function of both shear rate and particle/opening ratio.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 5 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2014 SPE Journal 5

  • different particle/opening ratios. Table 4 shows the resultsobtained for these two factors for the different gel strengths.

    With fitting equations obtained for the various particle/openingratios and gel strengths, one can evaluate PPG stable injectionpressure quantitatively to obtain a better prediction of the PPGinjection pressure in conduit systems.

    To develop a general correlation that can predict the PPG sta-ble injection pressure for all gel strengths, another regressionanalysis was performed to correlate these two coefficients withthe gel strengths, as shown in Fig. 10. Then, a and b were substi-tuted into the new general fitting equations.

    Finally, the general form of the new correlation that can beused to predict the initial stable injection pressure in conduit sys-tems can be written as

    Pinj 2 106 G02:5988 Dg=Dp2:3561 G00:053

    : 10

    Resistance to Water Flow After Gel Placement inConduits. After gel placement within the conduit system, brinewas injected with different velocities, from low to high, to extrudethe gel inside a conduit. In this way, four parameters were system-atically obtained to characterize particle-blocking behavior to

    water. These four sequence parameters include the pressure-gradi-ent peak (PGP), critical water breakthrough pressure (Pcw), resid-ual resistance factor (Frrw), and plugging efficiency (E).

    PGP. PGP is defined as the pressure gradient at which the gelbegan to move and wash out from the conduit as a result of brineinjection. Fig. 11 provides an example of the brine-injection-pres-sure gradient at each section through the gel swollen in a 0.05%-concentration brine within a conduit inner diameter of 10.922 mm.Brine was injected through the gel at a velocity of 9.89 ft/D. Gelwashout and water movement were measured by observing thepressure changes in all three sections, and monitoring both theeffluent produced gel and brine. In all experiments, we noticed thatthe injection-pressure gradient in all sections increased sharplyuntil reaching a certain peak, at which point it began to decline.This peak indicates the point at which gel failure and washoutbegan to occur in each section (Seright 2003). After each peak, thepressure gradient declined significantly before becoming stable inall sections. In the first section, the peak occurred at 1.85 psi/ft af-ter a 0.03-pore-volume (PV) injection of brine. In the second sec-tion, the PGP occurred at 1.05 psi/ft after injecting 0.04 PV ofbrine. In the last section, the peak occurred at 0.53 psi/ft afterinjecting 0.05 PV of brine. Then, after injecting 0.15 PV of brine,the water-pressure gradient in all sections became stable. The pres-sure-gradient variations in all three sections exhibited a differencein gel movement and washout along the conduit systems.

    The conduit inner diameter was checked visually after thebrine-injection process was complete. For the PPG swollen in0.05% NaCl, approximately 20% of gel was found remaininginside the conduit whereas for the PPG swollen in 10% NaCl,approximately 70% of gel was found remaining inside the sameconduit-inner-diameter size. This remaining volume suggests thatthe conduit was filled with a concentrated immobile gel.

    Table 5 provides a summary and comparison of the resultsobtained from the first section for brines of all concentrations.These results include the brine volume injected, as associatedwith its pressure gradients, for gels placed inside a conduit withinan opening size of 3.0488 mm. The results suggest that gel swol-len in high brine concentrations exhibit more stability inside theconduit than gel swollen in low brine concentrations when sub-jected to the same injection velocity.

    Table 6 summarizes the results obtained from the injection ofbrine through gel swollen in 10%-concentration brine for threeconduit inner diameters. Differing from the results obtained in alarge conduit opening, these results indicate that gel washout

    . . . . . . .

    Brine Concentration (%NaCl) M E R 2

    89.026.04.816,150.079.016.09.719,252.089.0695.07.346,3169.0885.05.821,701

    Table 3Fitting equations for Fr for each brine concentration.

    M = 3831.3C 0.2709R = 0.986

    10

    100

    1000

    10000

    0.01 0.1 1 10

    M

    Brine Concentrations (%)

    E = 0.6001C 0.01R = 0.9608

    0.56

    0.58

    0.6

    0.62

    0.64

    0.01 0.1 1 10

    E

    Brine concentrations (%)

    Fig. 8Shows constant coefficients (M) and (E) as a function of brine concentration.

    1

    10

    100

    1000

    10000

    100000

    0.1 1 10

    Stab

    le In

    jectio

    n Pr

    essu

    re (p

    si)

    Particle-Opening Ratio (Dg/Dp)

    515 Pa Gel strength657 Pa Gel strength870 Pa Gel strength1300 Pa Gel strength

    Fig. 9Stable injection pressure as a function of particle/open-ing ratio and gel strength in log-log scale.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 6 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    6 2014 SPE Journal

  • began to occur in a small conduit inner diameter when both ahigh-injection-pressure gradient and large volume of water wereapplied. Gel washout began to occur through an opening of 1.752mm when 1.17 PV of brine was injected, and the pressure gradientreached 245.3 psi/ft. In contrast, the gel injected through a largeropening size (10.922 mm) began to move when 0.16 PV of brinewas injected and the pressure gradient was only 4.9 psi/ft. Thesefindings indicate that less gel movement occurred in smaller con-duit diameters than in larger conduit diameters.

    Critical PCW. PCW is defined as the pressure at which the firstdrop of water can be seen from the outlet. Fig. 12 provides infor-mation about this variable as a function of both the brine concen-tration and conduit inner diameter. The small water-breakthroughpressure indicates that water could start propagating easilythrough the gel. This result suggests that as the gel became stron-ger (swollen in high brine concentration), the water-breakthroughpressure increased. Differences in water breakthrough are clearwhen comparing weak gel (swollen in low brine concentrations)to strong gel. Fig. 12a shows the water-breakthrough measure-ment for gel swollen in different-concentration brines when gelwas placed within a 3.0488-mm opening. When gel was swollenin 0.05% brine, water was able to pass through it at 8.8 psi. Watercould not pass through gel swollen in 10% brine until the pressurereached 46 psi. Fig. 12b shows the results obtained for waterbreakthrough through gel swollen in a 0.05% brine concentration

    as a function of different conduit sizes. Water was less likely topass through a smaller pore opening than a larger opening. Waterpassed through a 10.922-mm opening at a pressure of 2.1 psi, andthrough a 1.752-mm opening at a pressure of 60.2 psi.

    Frrw and E. Frrw is defined as the ratio of water-phase perme-ability before and after particle-gel treatment, and E refers to thepercentage of permeability reduction, which one can calculatefrom E (%) [1 (1 / Frrw)] 100. The stabilized water-injectionpressures at different velocities were used to calculate Frrw and E.Fig. 13 shows the residual resistance factor as a function of brineconcentration and brine velocity. Frrw increased as the gelstrength and conduit inner diameter increased. Fig. 14 shows thegel-plugging efficiency as a function of the brine concentrationand brine velocity. The preformed-particle-gel (PPG) plugging ef-ficiency increased when a strong gel was selected as a pluggingagent for large conduit sizes. The results suggest that gel swollenin 10% brine can provide a 97% plugging, as compared to 76%plugging for gel swollen in 0.05% brine for a conduit with aninner diameter of 1.752 mm. This percentage increased to 98%for the former and 93% for the latter when gel was placed into alarge opening (10.922 mm). These findings indicate that the plug-ging efficiency of the PPG did not decrease significantly in spiteof the gel washout occurring after gel placement.

    Discussion

    When investigating particle injection, many researchers are inter-ested in the passing ratio, which is the ratio of the particle size tothe pore-throat size at which the particle can pass through a con-striction. For stiff, hard particles, this question is easy to answer.Extensive experimental results have shown that stiff particles canpass through pore throats only if their particle sizes are less than 1/9 of the pore size. However, swollen gel particles are deformableand breakable, so they can pass through porous media much easierthan stiff particles. Swollen gel-particle transport mechanismthrough porous media exhibits different patterns of behavior (Baiet al. 2007b). Table 7 provides the ratio of particle size to openingsize (Dg/Dp), as well as the particle size before and after extrusion.Weak particles still were able to transport through the openingwhen the Dg/Dp was as high as 6.3 but required a relatively highinjection pressure-gradient of 12.5 psi/ft to do so. These results areconsistent with Seright (1997) in which he observed that pressuregradient increased significantly with decreased tube diameter.

    The results shown in Table 7 indicate that gel-particle size wasreduced after extrusion when Dg/Dp is larger than, equal to, and

    Gel Strength (Pa) ba R2

    99.07896.1966.4251599.06866.1550.1675699.04846.1317.0907889.06516.194.5030031

    Table 4Fitting equations for stable injection pressure.

    a = 2E06G`2.5988R = 0.9803

    10

    100

    1000

    100 1000 10000

    a

    Gel Strength (Pa)

    b = 2.3561G`0.053R = 0.9897

    1.59

    1.62

    1.65

    1.68

    1.71

    100 1000 10000

    b

    Gel Strength (Pa)

    Fig. 10Shows correlation coefficients (a) and (b) as a function of gel strength.

    0

    0.5

    1

    1.5

    2

    0 0.05 0.1 0.15 0.2

    Pres

    sure

    Gra

    dien

    t (ps

    i/ft)

    Brine-Injection Volume (PV)

    section onesection twosection three

    Fig. 11Brine-injection gradient through gel in three sectionsfor 10.922-mm conduit.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 7 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2014 SPE Journal 7

  • Brine Concentration (%NaCl)

    Brine-InjectedVolume for Peak (PV)

    PGP Through BrineInjection (psi/ft)

    52.421.050.058.713.052.050.2125.01

    6.5276.001

    Table 5Effect of brine concentrations on gel movement in conduit size of 3.0488 mm at 9.62 ft/D.

    9.461.0229.016.5276.0840.33.54271.1257.1

    Inner Diameter (mm)Brine-Injected

    Volume for Peak (PV) PGP Through

    Water Injection (psi/ft)

    Table 6Effect of conduit inner diameter on gel extrusion for gel swollen in 10% NaCl.

    (a) Conduit inner diameter 3.0488 mm (b) Gel swollen in 0.05% brine concentration

    05

    101520253035404550

    0.05 0.25 1 10Wat

    er-B

    reak

    thro

    ugh

    Pres

    sure

    (psi

    )

    Brine Concentration (% NaCl)

    0

    10

    20

    30

    40

    50

    60

    70

    10.922 3.0488 1.752

    Wat

    er-B

    reak

    thro

    ugh

    Pres

    sure

    (psi

    )

    Conduit Inner Diameter (mm)

    Fig. 12Critical water-breakthrough pressure as a function of brine concentration and conduit inner diameter.

    Conduit inner diameter 1.752 mm Conduit inner diameter 3.048 mm Conduit inner diameter 10.922 mm

    1

    10

    100

    1000

    10000

    100000

    1000000

    1 100 10000

    Res

    idua

    l-Res

    ista

    nce

    Fact

    or

    Brine-Injection Velocity (ft/D)1 100 10000

    0.05% NaCl 10% NaCl

    1 100 10000Brine-Injection Velocity (ft/D) Brine-Injection Velocity (ft/D)

    Fig. 13Residual resistance factor as a function of brine concentration and conduit inner diameter.

    Conduit inner diameter 1.752 mm Conduit inner diameter 10.922 mmConduit inner diameter 3.048 mm

    0

    20

    40

    60

    80

    100

    120

    0 1000 2000 3000

    Plug

    gin

    g Ef

    ficie

    ncy

    (%)

    Brine-Injection Velocity (ft/D)0 1000 2000 3000

    Brine-Injection Velocity (ft/D)

    0.05% NaCl10% NaCl

    0 1000 2000 3000

    Brine-Injection Velocity (ft/D)

    Fig. 14Plugging efficiency as a function of brine concentration and conduit inner diameter.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 8 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    8 2014 SPE Journal

  • even smaller than unity. Fig. 15 shows particle-size measurementbefore and after extrusion for the sample with gel strength of 515Pa. The weak gel particles experienced a significant decrease inparticle size, up to 81.5%, when they moved through conduit witha 0.774-mm opening for Dg/Dp equal to 6.3. However, a strong geldecreased only by 21.7% when moving through the same conduitsize but with Dg/Dp equal to 2.85. On the basis of previous knowl-edge (Bai et al. 2007b), this particle-size reduction could beexplained by three reasons: breakdown, dehydration, or both. Todetermine if the particle-size reduction was caused by gel dehydra-tion, we collected effluent particle-gel samples from a 3.048-mmconduit, in which Dg/Dp are smaller than and equal to unity, meas-

    ured their strength, and also placed them in the same-concentrationbrine to observe their reswelling. Fig. 16 shows how much the gelvolume increased at different injection rates for four different-strength gels. The results show that the weakest particles canregain 50% of water, whereas the strong gels can regain onlyapproximately 20% of water, indicating that weak gel can be dehy-drated more than strong gel during conformance-control treat-ments. In other words, the weakest particles shrank 50% of theiroriginal volume whereas the strongest ones did shrank 20% whenthey passed through the conduit. The strength measurement takenafter extrusion, shown in Fig. 17, also indicates that the gelbecame more concentrated because of water loss from its cluster.

    Gel Strength (Pa)

    Gel-Particle Size Before Extrusion

    (mm)

    Conduit-Opening

    Size (mm)

    Dg/Dp Gel-Particle Size After Extrusion

    (mm)

    Gel-Particle-Size

    Decrease (%)

    Gel-Threshold Pressure Gradient (psi/ft)

    Gel-Injection Stable Pressure

    (psi)

    10.922 0.44 3.57 26.8 0.1 6.8 3.048 1.60 2.58 47.13 0.65 51 1.752 2.78 2.391 51 1.1 112

    515 4.88

    0.774 6.3 0.902 81.51 12.5 680 10.922 0.32 2.90 18.5 0.2 9.2 3.048 1.16 2.17 39.04 0.85 76

    657 3.56

    1.752 2.03 2.19 38.4 2.1 203 10.922 0.29 2.30 28.1 0.21 11.5 3.048 1.04 1.99 37.81 1.2 105

    870 3.2

    1.752 1.82 1.87 41.5 2.8 230 10.922 0.20 1.945 11.99 0.25 20.1 3.048 0.72 1.808 18.19 2.3 191.7 1.752 1.26 1.923 12.98 3.5 590

    1300 2.21

    0.774 2.85 1.73 21.7 26.2 1320

    Table 7Particle/opening-ratio measurements results.

    0

    0.2

    0.4

    0.6

    0.8

    1

    1.2

    1.4

    0 1 2 3 4 5 6 7 8 9 10

    NORM

    DIST

    PPG Radius (mm)

    Before Extrusion0.774 mm Conduit inner diameter1.752 mm Conduit inner diameter3.048 mm Conduit inner diameter10.922 mm Conduit inner diameter

    10.922 mm 1.752 mm

    3.048 mm 0.774 mm

    Before Extrusion

    Fig. 15Preformed-particle-gel size distributions and images for gel with strength of 515 Pa before and after extrusion.

    0

    10

    20

    30

    40

    50

    60

    70

    0 1 2 3 4

    Gel

    -Vo

    lum

    e In

    crea

    se (%

    )

    Gel-Injection Rate (mL/min)

    515 Pa Gel strength657 Pa Gel strength870 Pa Gel strength1300 Pa Gel strength

    Fig. 16Particle volume increased after soaking in same brine.

    0

    500

    1000

    1500

    2000

    0 1 2 3 4

    Stor

    age

    Mod

    ulus

    (G`)

    (Pa)

    Gel-Injection Rate (mL/min)

    515 Pa 657 Pa870 Pa 1300 Pa

    Fig. 17Particle-storage moduli (G0) after extrusion.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 9 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2014 SPE Journal 9

  • When we compared the significant reduction in particle size to thegel-particle volume shrinkage, we observed that gel-particle sizereduction was caused by both particle breakdown and dehydration.

    Conclusion

    Numerous factors that affect preformed-particle-gel (PPG) extru-sion and blocking efficiency through conduit systems were exten-sively examined in this study. The mechanisms associated withgel propagation and placement, such as dehydration and gel wash-out, were investigated during the experiments. The following con-clusions can be drawn from the research: PPG-injection pressure increased as the brine concentration and

    injection velocity increased. This increase, however, after a cer-tain velocity level was reached, became unsubstantial.

    The resistance factor (Fr) increased when the gel strengthincreased and/or when the conduit inner diameter becamewider, but it decreased if the velocity increased.

    Both the gel-threshold pressure and the stable injection pressureincreased as the particle/opening ratio increased. Both pres-sures, however, would not increase significantly after a specificratio value was reached. In addition, the gel strength affectedthe gel-injection pressure more than did the particle/openingratio.

    Two new empirical correlation models were successfully devel-oped to predict both PPG Fr and stable injection pressure.

    The Fr measurements are not dependent of the particle/openingratio when it is measured against shear rate.

    PPG blocking performance increased as the gel strength andconduit inner diameter increased. This finding reveals that theconduit-size conductivity can significantly decrease if a stronggel is selected for the conformance treatment.

    Weak gels can be injected into large particle/opening ratio withrelative small increase in injection pressure compared to stronggels. Weak gels break into small sizes, so they are able to passthrough.

    Weak gels tend to dehydrate more than strong gels. The gelbecomes stronger after the extrusion process, as a result of thedehydration mechanism.

    PPG size can be reduced during transportation through con-duits, by dehydration and by breakdown.

    Nomenclature

    A cross-sectional area, cm2a, b coefficient factors, dimensionlessC brine concentration, %NaClD inner-conduit diameter, mm

    Dg/Dp particle/opening ratio, dimensionlessE plugging efficiency, %Fr resistance factor, dimensionless

    Frrw residual-resistance factor to water, dimensionlessG0 storage moduli, PaL length of conduit, cm

    M, E constant coefficients, dimensionlessPinj initial stable injection pressure, psiQ injection flow rate, cm3/s

    DPbrine brine-injection-pressure drop, psiDpPPG preformed-particle-gel-injection-pressure drop, psiDP differential pressure, atmc shear rate, 1/secv superficial velocity, ft/D

    Acknowledgments

    Funding for this project is provided by RPSEA authorized by theUS Energy Policy Act of 2005. RPSEA (www.rpsea.org) is a non-profit corporation with the mission of providing a stewardshiprole in ensuring the focused research, development, and deploy-ment of safe and environmentally responsible technology that caneffectively deliver hydrocarbons from domestic resources to thecitizens of the United States. The first author wishes to thank theMinistry of Higher Education and Scientific Research in Libya for

    its support. The corresponding author for the paper is Baojun Bai,Lester Birbeck Chair Professor at Missouri University of Scienceand Technology.

    References

    Bai, B., Li, L., Liu, Y. et al. 2007a. Preformed Particle Gel for Conform-

    ance Control: Factors Affecting Its Properties and Applications. SPERes Eval & Eng 10 (4): 415421. SPE-89389-PA. http://dx.doi.org/10.2118/89389-PA.

    Bai, B., Liu, Y., Coste, J.-P. et al. 2007b. Preformed Particle Gel for Con-

    formance Control: Transport Mechanism Through Porous Media. SPE

    Res Eval & Eng 10 (2): 176184. SPE-89468-PA. http://dx.doi.org/10.2118/89468-PA.

    Bai, B., Wei, M. and Liu, Y. 2013. Field and Lab Experience With a Suc-

    cessful Preformed Particle Gel Conformance Control Technology. Pre-

    sented at the SPE Production and Operation Symposium, Oklahoma

    City, Oklahoma, 2328 March. SPE-164511-MS. http://dx.doi.org/

    10.2118/164511-MS.

    Bailey, B., Crabtree, M., Tyrie, J. et al. 2000. Water Control. Oilfield

    Review 12 (1): 30.Chauveteau, G., Omari, A., Tabary, R. et al. 2001. New Size-Controlled

    Microgels for Oil Production. Presented at the SPE International Sym-

    posium on Oilfield Chemistry, Houston, Texas, 1316 February. SPE-

    64988-MS. http://dx.doi.org/10.2118/64988-MS.

    Chauveteau, G., Tabary, R., le Bon, C. et al. 2003. In-Depth Permeability

    Control by Adsorption of Soft Size-Controlled Microgels. Presented at

    the SPE European Formation Damage Conference, The Hague, The

    Netherlands, 1314 May. SPE-82228-MS. http://dx.doi.org/10.2118/

    82228-MS.

    Coste, J.-P., Liu, Y., Bai, B. et al. 2000. In-Depth Fluid Diversion by Pre-

    Gelled Particles. Laboratory Study and Pilot Testing. Presented at the

    SPE/DOE Improved Oil Recovery Symposium, Tulsa, Oklahoma, 35

    April. SPE-59362-MS. http://dx.doi.org/10.2118/59362-MS.

    Ganguly, S., Willhite, G.P., Green, D.W. et al. 2001. The Effect of Fluid

    Leak-off on Gel Placement and Gel Stability in Fractures. Presented at

    the SPE International Symposium on Oilfield Chemistry, Houston,

    Texas, 1316 February. SPE-64987-MS. http://dx.doi.org/10.2118/

    64987-MS.

    Hill, F., Monroe, S., and Mohanan, R. 2012. Water ManagementAn

    Increasing Trend in the Oil and Gas Industry. Presented at the SPE/

    EAGE European Unconventional Resources Conference and Exhibi-

    tion, Vienna, Austria, 2022 March. SPE-154720-MS. http://

    dx.doi.org/10.2118/154720-MS.

    Liu, Jin, and Seright, R.S. 2000. Rheology of Gels Used For Conformance

    Control in Fractures. Presented at the Improved Oil Recovery Sympo-

    sium, Tulsa, Oklahoma, 35 April. SPE-59318-MS. http://dx.doi.org/

    10.2118/59318-MS.

    McCool, C.S., Li, X., and Willhite, G.P. 2009. Flow of a Polyacrylamide/

    Chromium Acetate System in a Long Conduit. SPE J. 14 (1): 5466.SPE 106059-PA. http://dx.doi.org/10.2118/106059-PA.

    Seright, R.S. 1995. Reduction of Gas and Water Permeabilities Using

    Gels. SPE Prod & Fac 10 (2): 103108; Trans., AIME, 299. SPE-

    25855-PA. http://dx.doi.org/10.2118/25855-PA.

    Seright, R.S. 1997. Use of Preformed Gels for Conformance Control in

    Fractured Systems. SPE Prod & Fac 12 (1): 5965. SPE-35351-PA.

    http://dx.doi.org/10.2118/35351-PA.

    Seright, R.S. 1998. Improved Methods for Water Shut-off Annual

    Report (US DOE Report DOE/PC/91008-4), US DOE Contract DE-

    AC22-94PC91008 (September).

    Seright, R.S. 1999. Mechanism for Gel Propagation Through Fractures. Pre-

    sented at the SPE Rocky Mountain Regional Meeting, Gillette, Wyom-

    ing, 1518 May. SPE-55628-MS. http://dx.doi.org/10.2118/55628-MS.

    Seright, R.S. 2001. Gel Propagation Through Fractures. SPE Prod & Fac16 (4): 225231. SPE-74602-PA. http://dx.doi.org/10.2118/74602-PA.

    Seright, R.S., Lane, R.H., and Sydansk, R.D. 2001. A Strategy for Attack-

    ing Excess Water Production. Presented at the SPE Permian Basin Oil

    and Gas Recovery Conference, Midland, Texas, 1516 May. SPE-

    70067-MS. http://dx.doi.org/10.2118/70067-MS.

    Seright, R.S. 2003. Washout of Cr(III)-Acetate-HPAM Gels From Frac-

    tures. Presented at the SPE International Symposium on Oilfield

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 10 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    10 2014 SPE Journal

  • Chemistry, Houston, Texas, 57 February. SPE-80200-MS. http://

    dx.doi.org/10.2118/80200-MS.

    Smith, D., Giraud, M., Kemp, C. et al. 2006. The Successful Evolution of

    Anton Irish Conformance Effort. Presented at the SPE Annual Techni-

    cal Conference and Exhibition, San Antonio, Texas, 2427 September.

    SPE-103044-MS. http://dx.doi.org/10.2118/103044-MS.

    Sydansk , R.D., Xiong, Y., Al-Dhafeeri, A. et al. 2005. Characterization of

    Partially Formed Polymer Gels for Application to Fractured Produc-

    tion Wells for WaterShutoff Purposes. SPE Prod & Fac 20 (3):240249. SPE-89401-PA. http://dx.doi.org/10.2118/89401-PA.

    Wang, Y. and Seright, R.S. 2006. Correlating Gel Rheology With Behav-

    ior During Extrusion Through Fractures. Presented at the SPE/DOE

    Symposium on Improved Oil Recovery, Tulsa, Oklahoma, 2226

    April. SPE-99462-MS. http://dx.doi.org/10.2118/99462-MS.

    Wilton, R. and Asghari, K. 2007. Improving Gel Performance in Frac-

    tures: Chromium Pre-Flush and Overload. J Can Pet Technol 46 (2):SPE-07-02-04-PA. http://dx.doi.org/10.2118/07-02-04-PA.

    Zaitoun, A., Makakou, P., Blin, N. et al. 2012. Shear Stability of EOR

    Polymers. SPE J. 17 (2): 335339. SPE-141113-PA. http://dx.doi.org/10.2118/141113-PA.

    Zhang, H. and Bai, B. 2011. Preformed Particle Gel Transport Through

    Open Fracture and Its Effect on Water Flow. SPE J. 16 (2): 388400.SPE-129908-PA. http://dx.doi.org/10.2118/129908-PA.

    Abdulmohsin Imqam, SPE, is a PhD degree candidate at Mis-souri University of Science and Technology majoring in petro-leum engineering. He holds graduate research and teachingassistant positions. Previously, Imqam was a reservoir engineerat the Suncor Energy Company. He has 4 years of experienceworking in oil-industry companies. Imqam research interestsinclude reservoir engineering, enhanced oil recovery (EOR),and conformance-control treatments. He holds MS degrees inpetroleum engineering from Tripoli University and in engineer-ing management from Missouri University of Science andTechnology.

    Baojun Bai, SPE, is an associate professor of petroleum engi-neering at Missouri University of Science and Technology. Heholds the Lester R. Birbeck endowed Chair position. Previously,Bai was a reservoir engineer and head of the conformance-control team at the Research Institute of Petroleum Explora-tion and Development, Petro China. He also was a post-doc-

    toral scholar at the California Institute of Technology andgraduate research assistant at the New Mexico Petroleum Re-covery Research Center for EOR projects. Bai has nearly 20years of experience in the area of EOR, especially in gel treat-ments. He holds PhD degrees in petroleum engineering fromNew Mexico Institute of Mining and Technology and in petro-leum geology from China University of GeoscienceBeijing.Bai has published more than 80 papers in the area of EORmethods and application. He serves on the JPT Editorial Com-mittee and is a technical editor for SPE Journal and SPE Reser-voir Evaluation & Engineering.

    Mustafa Al Ramadan, SPE, is an MS degree student at MissouriUniversity of Science and Technology majoring in petroleumengineering. His research interests include EOR and conform-ance-control treatments. Ramadan holds a BS degree in pe-troleum engineering from King Fahd University of Petroleumand Minerals.

    Mingzhen Wei, SPE, is an assistant professor of petroleum engi-neering at Missouri University of Science and Technology. Shewas an assistant professor in petroleum engineering at ChinaUniversity of Geoscience before she started studies at NewMexico Institute of Mining and Technology in pursuit of a PhDdegree in computer science in 2000. Wei holds BS and MSdegrees in petroleum engineering from China University of Pe-troleum. She has expertise in data management and datamining and in numerical simulation.

    Mojdeh Delshad, SPE, is a research professor in the Depart-ment of Petroleum and Geosystems Engineering at the Univer-sity of Texas at Austin and vice president of Ultimate EOR LLC.She holds MS and PhD degrees in petroleum engineering fromthe University of Texas at Austin. Delshads research interestsinclude reservoir engineering, modeling petrophysical proper-ties, simulation of chemical and CO2 EOR processes, and nu-merical simulation of CO2 storage in saline aquifers.

    Kamy Sepehrnoori is the Bank of America Professor at theDepartment of Petroleum and Geosystems Engineering at theUniversity of Texas at Austin. His research interests and teachinginclude computational methods, reservoir simulation, parallelprocessing, EOR modeling, and inverse modeling. Sepehrnooriholds a BS degree in mechanical engineering, an MS degreein engineeringmechanics, and a PhD degree in petroleum en-gineering, all from the University of Texas at Austin.

    J169107 DOI: 10.2118/169107-PA Date: 30-October-14 Stage: Page: 11 Total Pages: 11

    ID: jaganm Time: 14:30 I Path: S:/3B2/J###/Vol00000/140071/APPFile/SA-J###140071

    2014 SPE Journal 11