Influence of mix proportions on rheology of cement grouts containing limestone powder

Cement&ConcreteComposites25(2003)

737–749

/locate/cemconcomp

In uenceofmixproportionsonrheologyofcementgrouts

containinglimestonepowder

LucieSvermova,MohammedSonebi*,PeterJ.M.Bartos

DepartmentofCivilEngineering,AdvancedConcreteandMasonryCentre,UniversityofPaisley,PaisleyPA12BE,UK

Received27May2002;accepted30September2002

Abstract

Inthispapertheparametersofcementgrouta ectingrheologicalbehaviourandcompressivestrengthareinvestigated.Factorialexperimentaldesignwasadoptedinthisinvestigationtoassessthecombinede ectsofthefollowingfactorson uidity,rheologicalproperties,inducedbleedingandcompressivestrength:water/binderratio(W/B),dosageofsuperplasticiser(SP),dosageofviscosityagent(VA),andproportionoflimestonepowderasreplacementofcement(LSP).Mini-slumptest,Marshcone,Lombardiplatecohesionmeter,inducedbleedingtest,coaxialrotatingcylinderviscometerwereusedtoevaluatetherheologyofthecementgroutandthecompressivestrengthsat7and28daysweremeasured.Atwo-levelfractionalfactorialstatisticalmodelwasusedtomodelthein uenceofkeyparametersonpropertiesa ectingthe uidity,therheologyandcompressivestrength.Themodelsarevalidformixeswith0.35–0.42W/B,0.3–1.2%SP,0.02–0.7%VA(percentageofbinder)and12–45%LSPasreplacementofcement.Thein uencesofW/B,SP,VAandLSPwerecharacterisedandanalysedusingpolynomialregressionwhichcanidentifytheprimaryfactorsandtheirinteractionsonthemeasuredproperties.Mathematicalpolynomialsweredevelopedformini-slump,platecohesionmeter,inducingbleeding,yieldvalue,plasticviscosityandcompressivestrengthasfunctionofW/B,SP,VAandproportionofLSP.Thestatisticalapproachusedhighlightedthelimestonepowdere ectandthedosageofSPandVAonthevariousrheologicalcharacteristicsofcementgrout.

Ó2002ElsevierLtd.Allrightsreserved.

Keywords:Compressivestrength;Grout;Inducedbleeding;Limestonepowder;Mini-slump;Rheology;Superplasticizer;Viscosityagent;Water/binder

1.Introduction

Cement-basedgroutsarewidelyusedininjectiongroutingofcracksinmassivestructuressincetheirphysicalandmechanicalpropertiescanbeeasilycon-trolled.Therheologicalbehaviourofspecialcementgroutsintendedfortheunderwatersealingofcracksindams,o shorestructures,massivefoundations,or s-suresinrockcanbeenhancedbytheincorporationofviscosityagent(VA)[1,2].GroutscontainingVAarealsousedfor llingpost-tensioningducts,whereitisimportanttoensurehighresistancetosedimentationandbleeding,henceensuringcorrosionprotectionofstressedtendons[3].Admixturesmainlya ectthe owbehaviourofthecementpastewithoutalteringthecomposition.Therefore,itseemsreasonabletotryto

Correspondingauthor.Tel.:+44-141-849-4216;fax:+44-141-848-3275.

E-mailaddress:mohammed.sonebi@paisley.ac.uk(M.Sonebi).0958-9465/$-seefrontmatterÓ2002ElsevierLtd.Allrightsreserved.doi:10.1016/S0958-9465(02)00115-4

*

studythee ectofwater/binderandadmixtures,chemi-calandmineral,byonlytestingthecementpaste.Therheologicalmeasurementsoncementpastewereusedtoassessthefreshproperties.Viscosityagentsarealsousedingroutsfortherepairofdeterioratedstructuresbyinjection.

Viscosityagentsarerelativelynewadmixturesusedtoenhancethecohesionandstabilityofcement-basedsys-tems[4–10].SuchVAsarewater-solublepolysaccharidesthatenhancethewaterretentioncapacityofpaste[4,5,7–10].Theuseofviscosityagentincreasestheyieldvalueandplasticviscosityofcement-basedgrout,thusneces-sitatingincreaseinwater/binderorsuperplasticiserdos-agetoinsurealowyieldstressnecessaryforproperpenetrability,spreadingandcontrolsedimentation[4,10].Severalresearchershaverelatedtheimprovementinrheologicalpropertiesandtheperformanceofcement-basedgrouttotheadditionofsuperplasticiser(SP)andVA[4–10].Forexample,forunderwatercement-basedgrout,mini-slump,washoutresistanceandresidual

738L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

compressivestrengtharehighlyin uencedbywater/binderratio(W/B),VAandSPdosages[7–9].Thewash-outresistanceimprovesasVAcontentincreasesforagivenW/BdespitethegreaterdosageofSPnecessarytomaintain uidity[7–9].ForagivenW/Bof0.40,anin-creaseinmini-slumpduetoagreaterdosageofSPcanincreasethewashoutmasslossandreduceresidualstrength,regardlessofthedosageofVA[9].Anoptimi-sationisthereforenecessarytoestablishabalancebe-tweenthedosagesofVAandSP,theW/Bratioandtheproportionofmineraladmixture(limestonepowder, yash,silicafume,etc.)toensuresuitable owandpene-trationandreducethebleeding.ThemixoptimisationofgroutoftennecessitatesseveraltrialbatchestoachieveabalancebetweenthemineralandchemicaladmixturesandW/Btoensuresuitable uidity,stabilityandme-chanicalproperties,someofwhichparametershaveoppositee ects.Forthisreasonthestatisticalmodellingapproachwasusedinordertoreducethenumberoftrialbatches.Nehdiandco-workers[11,12]reportedthatthelimestonemicro llerreplacementofcementslightlyincreasedtheyieldvalueofcementpasteanddecreaseditsplasticviscosity,whichimpliesbetterstabilityand owabilityofthecementpaste.However,increasingthelimestonemicro llercontentsreducedtheinducedbleedingofcementpasteonlyathighW/Bratiosanddidnotseemtohaveasigni cante ectatlowW/B[11,12].Theaimofthisstudyistoevaluatethee ectoftheW/B,thedosagesofSPandVA,andtheproportionoflimestonepowderreplacementofcementontherhe-ologicalpropertiesandthecompressivestrength(fc0)at7and28daysofgroutsusingastatisticaldesignapproachandanalysisofexperiments[13].Themini-slumptest,Marshcone,Lombardiplatecohesionmeter,inducedbleedingtest,andcoaxialrotatingcylinderviscometer

Table1

Mixproportionforgroutsusedinthetwo-levelfractionalfactorialdesign

Codedvalues

W/B

Levelsoffactors

1234567891011121314151617

)11)11)11)1100000.140.430.710.43)0.14

SP)1)111)1)11100000.000.330.440.33)0.56

LSP)1)1)1)1111100000.09)0.820.390.24)0.52

wereusedfortestingthebehaviouroffreshcementgrouts.Thecompressivestrengthsofgroutsat7and28daysofagewerealsomeasured.Theestablishedmodelscanidentifyparametersandthetwo-wayinteractionsthathavesigni cante ectontherheologicalpropertiesandcompressivestrengthofgrouts.Themodelscanbeusedtoevaluatethepotentialin uenceofadjustingmixvariablesongroutpropertiesrequiredtoensuresuc-cessfuldevelopmentofgrout.Suchsimulationcanfa-cilitatethetestprotocolneededtooptimizegroutwithagivensetofperformancecriteriathatcanbetriedinthelaboratory.

2.Statisticaldesignapproach

Thetechniqueofanalysisusedwasastatisticalanalysisoftheresultsobtainedfromasetofexperiments[13].Thistechniqueappliedtocementmaterialgroutcangivealotofinformationfromafewexperiments.A24À1fractionalstatisticalexperimentaldesign(2kÀ1¼8)wasusedtoevaluatethein uenceoftwodi erentlevelsforeachindependentvariableontherelevantgroutproperties.Fourkeyparameters(k¼4)thatcanhavesigni cantin uenceonmixcharacteristicsofcementgroutwereselectedtoderivemathematicalmodelsforevaluatingrelevantproperties.ThefourvariableswereW/B,dosagesofSPandVA,andproportionofLSPasreplacementofcement.Themodelledexperimentalre-gionconsistedofmixesrangingbetweencodedvariablesof)1toþ1.ThederivedstatisticalmodelsarevalidformixesmadewithrangesofW/Bof0.35–0.42,dosagesofVAof0.02–0.07%,bymassofbinder(or0.057–0.166%ofwater),SPof0.3–1.2%,bymassofbinder,andtheproportionofLSPfrom12to45%(Table1).Themodel

Actualvalues

VA

W/B

SP(%)

LSP(%)

VA(%)

)10.350.312.00.0210.420.312.00.0710.351.212.00.07)10.421.212.00.0210.350.345.00.07)10.420.345.00.02)10.351.245.00.021.042.12.45.00070.0385.075.28.500450.0385.075.28.500450.0385.075.28.500450.0385.075.28.50045)0.60)0.201.00)0.20)0.60

0.390.400.410.400.38

0.750.900.950.900.50

30.015.035.032.520.0

0.030.040.070.040.03

Centrepoints

Pointsofveri cation

L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749739

consistedofeightfactorialspointswhereeachvariablewas xedattwodi erentlevels.

Fourreplicatecentralpointswerepreparedtoesti-matethedegreeofexperimentalerrorforthemodelledresponses.Thecentralpointsconsistedofmixeswithvariablescorrespondingto0.385,0.75%,0.045%,and28.5%forW/B,dosageofSP,VAandLSP,respectively.Finally, verandommixeswereproducedtoestablishtheaccuracyofthederivedmodels.Themixeswereproducedandtestedinrandomorder,whichisoneoftherequirementsoffactorialexperimentaldesign.

Thefreshcementgroutsweretestedwithmini-slumptest,Marshcone,Lombardiplatecohesionmeter,in-ducedbleedingtest,andcoaxialrotatingcylindervis-cometer.Thecompressivestrengthsweremeasuredat7and28days.

The17mixcombinations,expressedincodedandactualvalues,consideredintheexperimentaldesignofgroutsarelistedinTable1.Thecodedfactorsofvari-ablesarecalculatedasfollows:

CodedFactor¼ðActualvalueÀFactormeanÞ=

ðRangeofthefactorialsvalues=2Þ

CodedW=B¼ðActualW=BÀ0:385Þ=0:035CodedVA¼ðActualVAÀ0:045Þ=0:025CodedSP¼ðActualSPÀ0:75Þ=0:45CodedLSP¼ðActualLSPÀ28:5Þ=16:5

3.Materialproportionsandtestingprocedures

ThegroutsinvestigatedinthisstudywerepreparedusinganordinaryPortlandcementandlimestonepowder.Thechemicalandphysicalpropertiesofce-mentandlimestonepowderarepresentedinTable2.Thelimestonepowderwasproducedfromcarbonifer-ouslimestoneofaveryhighpurityandwas nerthancement.Thelimestonehadgradingof98%<45and25%<5lm.

Anewgenerationofsuperplasticiseronthebasisofmodi edpolycarboxylateswasusedwithasolidcontentof30%andspeci cgravityof1.11.TheviscosityagentwastheKelco–Cretewelangumthatisahighmolecularweight,microbialpolysaccharide.Thewelangumwassuppliedinapowdergum.

Allgroutmixeswerepreparedina5lplanar-actionhigh-shearmixer.Themixingtapwaterhadatemper-atureof16Æ1°C,whichwasmeasuredbeforemixingstarted.Theviscosityagentwasmixedwithcement.Thesuperplasticiserwasaddedtothewaterandmixedtogether.Mixingtimewasmeasuredfromwhenthelimestonepowder(the rstsolidcomponent)wasaddedintothemixofwaterandsuperplasticiser.Finally,themixofcementandviscosityagentwasaddedandallcomponentsweremixedfor7minfromthestartof

Table2

Chemicalandphysicalpropertiesofcementandlimestonepowder

Cement

LimestonepowderSiO220.8–Al2O35.0

–Fe2O3

3.2–CaO63.7–MgO2.6

0.2Na2Oeq.0.39–FreeCaO1.6–LOI0.65–CaCO3–99Relativedensity3.142.65Speci csurfacearea385

–

(m2/kg)

Standardcompressivestrength(MPa)Age(d)Cement741.528

57.8

measuringtime.Thegrouttemperaturefollowingtheendofmixingwasmaintainedat20Æ2°C.

Followingtheendofmixing,thepropertiesofthefreshcementgroutweremeasured.Thefollowingtestsofthefreshcementgroutwerecarriedout(the guresinbracketsshowtherangeoftimeswhentheindividualtestsstartafter nishingofmixing):mini-slumptest(1–2min),Marshcone(4–5min),Lombardiplatecohesionmeter(10–15min),coaxialcylinderrotationviscometer(10–20min),inducedbleedingtest(20–30min)andunitweight.Threecylinderswith50Â55-mmdiameterandheightwerecasttodetermine7and28dayscompressivestrength.

Themini-slumptestisbasedonthemeasurementofthespreadofgroutplacedintoacone-shapedmould.Themini-slumpconehasanupperdiameterof19mm,alowerdiameterof38.1mm,andaheightof52.7mm.Theconeisplacedinthecentreofasmoothplateandthespreaddiameterofthegroutafterliftingoftheconeismeasured.

TheMarshconetestisbasedonmeasuringthetimenecessaryforthe owofaparticularvolumeofgroutthrougha ow-cone.Nowadays,di erenttypesof ow-coneareused.Aplasticfunnelwithacapacityof1200mlwasusedinthiscase.Onehalfoftheupperpartofthefunnelwascoveredwithasieve.Thegroutwasplacedinthefunnelthroughthissieve,whichpreventedlargeparticlesblockingtheoutlet.Thefunnelwaswet-tedbeforeeachtest.Avolumeof1100mlofthegroutwasplacedintheconewiththeoutletsealedandthenthetimeforthe owofeach100mlofgroutwasre-corded.The owtimeofMarshconeat700mlwasevaluated.

740L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

ThecohesivenessofthegroutwasmeasuredwithaLombardiplatecohesionmeter[15].Theapparatusconsistsofathinsteelplate(100mmÂ100mmÂ1mm),onwhichthegroutcanstick,andanelectronicscale.Thecleandryplatewasweighedandthensub-mergedonceintothegrout.Theplatewasthenwith-drawnandweighedagainafteranydroppingofgroutstopped(Fig.1).

Thespeci cweightofthegroutwasmeasuredbyamudbalance.Thismudbalanceconsistsofaconstant-volumesamplecupwithlidconnectedtoabalancearm.Areaderismovedalongthebalancearmtoindicatethescalereading.Thereisaknifeedgeattachedtothearmnearthebalancecupandabubblelevelbuiltintothisknifeedgeforlevellingthearm.Itwaspossibletocal-culatethethicknessofgroutoneachsideoftheplatefromtheunitweightandtheamountofgroutstickingtothesteelplate.

Theresistanceofthefreshgrouttoinducedbleedingwasevaluatedusingapressure lter.Theequipmentconsistsofapressurevessel, lterpaper,whichisplacedonasieve,andagraduatedcylinder.A200mlgroutsampleisplacedinthepressurevessel.Afterclosingthecell,thegraduatedcylinderisplacedundertheoutletofthecell.Thecellispressuredbycompressedairto0.55MPa.Thevolumeofwatergoingoutthroughtheoutletonthebottomofthecellisrecordedat15and30s,thenateveryminuteupto10min,andthenatevery5minupto30min[16].Theresultsofthistestarepresentedasareaundercurveresponsetimevs.volumeofwater(Fig.2).

Theviscosityofcementgroutisdeterminedusingacoaxialrotatingcylinderviscometer(smoothcylinders,noserration)thatenabledthedeterminationofapparentviscosityatdi erentshearrates[17].Thetestiscon-tainedintheannularspacebetweenanoutercylinder(rotor)withradiusof18.415mmandabobwithradiusof17.245mmandheightof3.80cm.Therotorandthebobareplungedintoacupwhichcontains350mlofsample(Fig.3)[17].Viscositymeasurementsaremade

Fig.1.Platecohesionmeterapparatus[15].W=B¼0:35,SP¼1:2%,VA¼0:02%,LSP¼

45%.

Fig.3.Coaxialrotatingcylinderviscometer[17].

whentheoutercylinder,rotatingataknownspeed,causesaviscousdragtobeexertedbythe uid.Thisdragcreatesatorqueonthebob,whichistransmittedtoaprecisionspringwhereitsde ectionismeasuredandthencomparedwiththetestconditionsandtheinstru-mentÕsconstants.Themeasurementismadefor12speedsofrotorfrom0.9to600rpmwhenthevaluesofviscometerreading¼harerecorded.Thevalueofshearstress¼s(Pa)iscalculatedbyincludingk1¼torsionconstantofspringperunitde ection(Ncm/deg.),k2¼shearstressconstantforthee ectivebobsurface(cmÀ3)andk3¼shearrateconstant(sÀ1/rpm)[17].

Thespeedofrotorwasincreasedstepbystepfrom0to600rpmandreadingonviscometerwithincreasingrotatingspeedwasrecorded.Thereadingofhwastakenwhentheneedleintheviscometerwasstabilised,or30safterthechangeofspeedincaseswhentheneedlehasnotstabilisedwhichwascausedbythethixotropyofthecementgrout.Themeasurementofthereadingbyde-creasingrotatingspeedstepbystepallowedtheassess-mentofthethixotropyofgroutbetweentheascendinganddescendinglegsoftheshearstress–shearraterheo-grams.Theupcurvewaschosenfor nalevaluation

for

L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749741

betterdescriptionofrheologicalbehaviourofthegroutsincludingastructuralbreakdownphenomenonofinnerforcesamongparticles[18,19].

Thevaluesofyieldstressandplasticviscosityareobtainedfrommodi edBinghammodel[20],whichisdescribedbytheequation:_þcc_2s¼s0þlpc

wheres0¼yieldstress(Pa),lp¼plasticviscosity(Pas),

_¼shearrate(sÀ1),c¼constant.c

Thevalueoftheconstantcismostlyabout10À3andless,whichismuchsmallerthanthevalueofplasticviscositylpandyieldstresss0,andforthisreasonthisconstantisconsideredtoequalzero.

Thecompressivestrengthwasdeterminedon55Â50-mmcylinders(diameterÂheight).Thespecimensweredemouldedonedayaftercastingandwerestoredinwateruntiltestingat7and28days.4.Testresultsanddiscussion4.1.Derivedstatisticalmodels

ThetestresultsformixesinvestigatedinthisstudyaregiveninTable3.Thevariousresponseswhichresultedfromthedesignedexperimentalprogrammewereanal-ysedandplottedusingastatisticalsoftwarepackage[14].Thederivedstatisticalmodelsforallthesetestsresultswithcorrelationcoe cient,Prob:>jtjvaluesandprospectivetransformationwereshowninTable4.Theestimatesforeachparameterrefertothecoe cientsofthemodelfoundbyaleast-squaresapproach.The

Prob:>jtjistheprobabilityofgettinganeventgreatertstatistic,inabsolutevalue,thattestswhetherthetrueparameteriszero.Probabilitieslessthan0.05areoftenconsideredassigni cantevidencethattheparameterisnotzero,i.e.thatthecontributionoftheproposedparameterhasahighlysigni cantin uenceonthemeasuredresponse.

ThepresentationinTable4enablesthecomparisonofvariousparametersaswellastheinteractionsofthemodelledresponses.Forthemajorityofparameters,theprobabilitiesthatthederivedcoe cientsofthevariousparametersin uenceeachresponsearelimitedto5%.Thissigni esthatthereislessthan5%chance,or95%con dencelimit,thatthecontributionofagivenpa-rametertothetestedresponseexceedsthevalueofthespeci edcoe cient.Anegativeestimatesigni esthatanincreaseofthegivenparameterresultsinareductionofthemeasuredresponse.

Transformationwasusedforstabilisingofthemodelintwocases.Forexample,thetransformationofnaturallogarithmwasusedforplatecohesionmeterandplasticviscosity.Inordertoillustratethemethod,assumethatresponsesY1andY2arefunctionsofW/B,dosagesofSPandVA,andtheproportionofLSP,thenLinearmodel:Y1¼a0þa1W=Bþa2SPþa3LSPþa4VAþa5W=BSPþa6W=BLSPþa7W=BVAþeNaturallogarithm:

lnY2¼a0þa1W=Bþa2SPþa3LSPþa4VA

þa5W=BSPþa6W=BLSPþa7W=BVAþe

Table3

ResultsoftestingmethodsusedforindividualmixesMix

Mini-slump(mm)77.061.577.0179.564.5118.0142.0172.5113.0126.0116.0111.5150.5104.5134.0142.0117.0

Flowtime(s)

Cohesionmeter(mm)1.3501.4091.0680.0821.7420.5700.0790.1640.4450.3620.3180.2710.1440.4300.2210.1530.296

ViscometerYieldvalue(Pa)11.188.4010.771.814.4311.262.555.4610.9810.189.819.096.1510.259.776.479.00

Plasticvis-cosity(Pas)0.520.401.170.110.790.170.170.130.350.330.270.290.190.240.200.160.31

Inducedbleeding(Lmin)1.091.220.930.510.921.291.291.401.241.301.251.251.281.521.011.321.35

fc07days(MPa)40.326.738.329.723.017.025.517.735.029.132.327.829.934.824.230.034.4

fc028days(MPa)47.534.739.233.531.423.726.820.037.939.234.133.540.045.328.933.842.5

1234567891011121314151617NMNMNM81.4NM184.3174.3197.5321.7295.9474.2566.961.0NM>20min242.4702.4

NM:notmeasurable.

742L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

Table4

Parameterestimatesofsevenstatisticalmodels

Mini-slump(mm)R2¼0:99EstimateProb:>jtj

a1a2a3a4a5a6a7a8

TransformInterceptW/BSPLSPVAW/BSPW/BLSPW/BVA

none111.50

21.370.0031.250.0012.750.00)17.630.0011.880.00––––

Platecohe-sionmeter

(mm)R2¼0:97EstimateProb:>jtjnaturallog)0.77)0.370.01)0.930.00)0.320.020.560.00––0.270.04––

Yieldvalue

(Pa)R2¼0:97EstimateProb:>jtjnone6.98

)0.250.37)1.840.00)1.060.010.280.31)1.260.012.680.00––

Plasticvis-cosity(Pas)R2¼0:97Estimate

Prob:>jtjnaturallog)1.18)0.560.00)0.280.01)0.280.010.420.00––––)0.160.06

Induced

bleeding(Lmin)R2¼0:99EstimateProb:>jtjnone1.08

0.0240.08)0.0480.010.140.000.0380.02)0.10.000.10.000.170.00

fc07days

(MPa)R2¼0:91EstimateProb:>jtjnone27.28)4.500.00––)6.480.00––––––––

fc028days

(MPa)R2¼0:93EstimateProb:>jtjnone32.10)4.130.00)2.230.03)6.630.00––––––––

wherea0denotestheoverallmean;coe cientsanrep-resentmodelconstants(contributionofindependent

variablesontheresponse),andeistherandomerrortermrepresentingthee ectsofuncontrolledvariables.Thethirdorderinteractionisusuallyneglected.

Forexample,mini-slump,platecohesionmeter,yieldvalue,plasticviscosity,inducedbleedingandfc0at28daysaregiveninEqs.(1)–(6),respectively.Mini-slumpðmmÞ¼111:5þ31:3SPþ21:4W=B

À17:6VAþ12:75LSPþ11:9W=BSP

lnPlatecohesionmeterðmmÞ

¼À0:77À0:93SPþ0:56VAÀ0:37W=BÀ0:32LSPþ0:27W=BLSP

YieldvalueðPaÞ¼7þ2:7W=BLSPÀ1:8SP

À1:3W=BSPÀ1:1LSPþ0:3VAÀ0:3W=B

lnPlasticviscosityðPasÞ¼À1:2À0:56W=B

þ0:42VAÀ0:28SPÀ0:28LSPÀ0:16W=BVA

InducedbleedingðLminÞ

¼1:1þ0:17W=BVAþ0:14LSPÀ0:1W=BSPþ0:1W=BLSPþ0:05SPÀ0:04VAþ0:03W=B

0fC

ð1Þ

ð2Þ

ð3Þ

ð4Þ

plasticviscosity,inducedbleedingandcompressivestrengthat7and28daysare0.99,0.97,0.97,0.97,0.99,0.91and0.93,respectively.Thehighcorrelationcoe -cientofmostresponsesdemonstratesexcellentcorrela-tionwhereitcanbeconsideredthatatleast95%ofthemeasuredvaluescanbeaccountedforwiththeproposedmodels.

Table5indicatestheaveragemeasuredresponseofthefourreplicategrouts,coe cientofvariation,estimatederrorwith95%con dencelimit,aswellasrelativeerrorforeachofthemeasuredproperties.Theestimatederrorofcementgroutformini-slump,platecohesionmeter,yieldvalue,plasticviscosity,inducedbleeding,andcompressivestrengthsat7and28dayswereÆ6.6mm,Æ0.08mm,Æ0.80Pa,Æ0.04Pas,Æ0.026Lmin,Æ3.3,andÆ2.9MPa,respectively.

Therelativeexperimentalerrorsformini-slump,yieldvalue,plasticviscosityandcompressivestrengthareshowntobelimitedto2–12%.Ontheotherhand,therelativeerrorfortheplatecohesionmeterwas22%.Thisvalueisexpectedtodecreasewiththeincreaseinrelativecohesionplatevalue,sincethemeanofcohesionplatevalueofthegroutcorrespondingtothecentralpointswasslightlylow.

The owtimemodelisnotgivenastheresultsoftheMarshconecouldnotbeusedfor nalevalua-tionbecausethegroutswithhighlevelofviscosityagent(0.07%)hadsuchhighviscositythattheywereunableto owthroughthe5-mm-outletofthefunnel.Inthiscase,anoutletbiggerthan5mmisrecom-mended.

4.2.Accuracyoftheproposedmodels

Theaccuracyofthestatisticalmodelswasdeterminedbycomparingaveragevaluesofpredicted/measuredratioformixesofgroutwhichwereproducedforveri -cationoftwo-levelfractionalfactorialdesign.Theaver-

ð5Þ

at28daysðMPaÞ¼32:1À6:6LSPÀ4:1W=B

À2:2SP

ð6Þ

Thecorrelationcoe cientsoftheproposedmodelsformini-slumptest,platecohesionmeter,yieldvalue,

L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

Table5

Repeatabilityoftestparameters

Mini-slump

Mean(n¼4)

Coe cientofvariation(%)Estimatederror(95%con dencelimit)Relativeerror(%)

116.6mm5.66.6mm5.7

Platecohesionmeter0.35mm21.30.08mm21.7

ViscometerYieldvalue10.0Pa7.90.80Pa8.0

Plasticviscosity0.31Pas12.30.04Pas12.5

Inducedbleed-ingin30min1.26Lmin2.0

0.026Lmin2.1

fc07days31.1MPa10.43.3MPa10.6

743

fc028days36.2MPa7.8

2.9MPa7.9

agepredicted/measuredratiosformini-slumptest,platecohesionmeter,yieldvalue,plasticviscosity,inducedbleedingtest,andcompressivestrengthsat7and28daysaresummarisedinTable6.

Theratiosbetweenpredictedandmeasuredproper-tiesofcementgroutrangedbetween0.83and1.05,thusindicatinggoodaccuracyforestablishedmodelstopredictthemini-slumptest,yieldvalue,plasticviscosity,inducedbleeding,andcompressivestrengthat7and28days.Ingeneral,theproposedmodelstopredictmini-slump,yieldvalue,plasticviscosity,inducedbleedingandcompressivestrengthappeartobesatistifactoryinpredictingthe uidity,rheology,inducedbleedingandcompressivestrength.Ontheotherhand,theaveragevalueofpredicted/measuredratioofplatecohesionmeterwasslightlyhigher(1.45).

4.3.Isoresponsesoftheproposedmodelsofthekeyvariables

4.3.1.Mini-slump

Theproposedstatisticalmodelscanthereforebeusedtoevaluatethee ectofagroupofvariablesonthepropertiesa ectingthequalityofcementgrout.Thispermittedthecalculationoftheisoresponsecurvesfromtheparametersunderstudyovertheexperimentaldo-mainandtheoptimisationoftheire ects.AsshowninTable4,themini-slumpisin uenced,inorderofsig-ni cance,bythedosageofSP,W/B,thedosageofVA,theproportionofLSPandtheinteractione ectofW/BSP.ThedosageofSPisshowntoexhibitthegreateste ectonthemini-slump.TheincreaseinSPhasap-proximately1.8and1.5timesgreaterin uenceonin-creasingmini-slumpthanthedecreaseinthedosageofVAandtheincreaseinW/B,respectively(31.3vs.)17.6and21.4).Forexample,thee ectofincreasingW/Bratioonmini-slumpvs.theproportionofLSP,whendosagesofSPandVAwere xedat0.75%and0.05%,respectively,orvs.dosageofSPof0.75%andtheproportionofLSPof28.5%,respectively,orvs.thedosageofVAof0.05%andtheproportionofLSPof28.5%,isshowninFig.4.ForanygivenW/BratioanddosagesofSPandVA xedat0.75%and0.05%,re-spectively,themini-slumpincreasedsigni cantlywhentheproportionofLSPincreased(Fig.4(a)).Sim-ilarly,themini-slumpincreasedwhenthedosageofSPincreasedwhilethedosageofVAandtheproportionofLSPwere xed(Fig.4(c)).Ontheotherhand,the

Table6

Predicted/measuredratiosformixesofestablishedmodelsTestingmethodVeri cationpoints

Mini-slump0.97

Cohesionmeter1.45

ViscometerYieldvalue0.85

Plasticviscosity1.05

0.87

0.87

0.83

Inducedbleeding

fc07days

fc028days

744L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

mini-slumpreducedwhenthedosageofVAincreasedwhilethedosageofSPandtheproportionofLSPwerekeptconstant(Fig.4(b)).

4.3.2.Platecohesionmeter

AsshowninTable4,platecohesionmeterisin u-enced,intheorderofsigni cance,bythedosagesofSPandVA,theW/BandtheproportionofLSP.Thein-creaseinW/Bhasafairlysimilarin uenceontheplatecohesionmeterastheincreaseintheproportionofLSP()0.37vs.)0.32).Bycomparingthee ectofSPandVAdosagesontheplatecohesionmeter,theincreaseofdosageofSPcanthenbeinterpretedtohaveapproxi-mately1.7timesgreaterin uenceonthereductionoftheplatecohesionvaluethantheincreaseinVA,giventhattheW/BandtheproportionofLSPareheldcon-stant.Fig.5showsanexampleofisopresponsecurvesoftheplatecohesionmetervs.W/Bfor xedvaluesofSPandVAorSPandLSPorLSPandVA.For xedvaluesofSPandVAat0.75%and0.045%,respectively,theplatecohesionmeterdecreasedwhentheW/BincreasedortheproportionofLSPincreased(Fig.5(a)).Thein-creaseofthedosageofSP,foranygivenW/Band xedLSPproportionandVAdosage,ledtoareductionintheplatecohesionmeter(Fig.5(c)).

4.3.3.Yieldvalue

AsshowninTable4,thequadratice ectofW/BLSPisshowntoexhibitthegreateste ectontheyieldvaluefollowingbythedosageofSP(2.7and)1.8).Thein-

creaseofW/BandtheproportionofLSPreducedtheyieldvalue.Themodel(Eq.(3))showsthattheincreaseofthedosageofSPismoree cientinreducingyieldvaluethananincreaseintheproportionofLSP()1.8vs.)1.1).Forexample,thee ectofincreasingW/Bratioonyieldvaluevs.theproportionofLSP,whendosagesofSPandVAwere xedat0.75%and0.05%,respectively,orvs.dosageofSPof0.75%andtheproportionofLSPof28.5%,respectively,orvs.thedosageofVAof0.05%andtheproportionofLSPof28.5%,isshowninFig.6.TheyieldvalueseemedtodecreasewithincreasedW/Bup0.40,thentendedtoincreasebeyondthisthresholdvalue(Fig.6(a)).For xeddosagesofSPandVA,theyieldvalueincreasedsigni cantlyuptoabout0.40ofW/BandhigherproportionofLSP(upto29%).AtlowerproportionsofLSP,however,adecreaseofyieldvaluewasobservedwithanincreaseinW/B(Fig.6(a)).4.3.4.Plasticviscosity

Plasticviscosityisin uenced,inorderofsigni cance,bytheW/B,thedosageofVA,thedosageofSPandtheproportionofLSP.TheW/Bisshowntohavethegreateste ectontheplasticviscosity(Eq.(4)).TheincreaseofW/Bhasapproximately1.3timesgreaterin uenceonreducingtheplasticviscositythanthede-creaseinthedosageofVA()0.56vs.0.42).Themodel(Eq.(4))showsthatthee ectofchangingthedosageofSPontheplasticviscosityissimilartothatofthepro-portionofLSP()0.28vs.0.28).Bycomparingthee ectsofSPandVAontheplasticviscosity,itcanbe

observed

L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749745

thatthee ectofVAishigherthanthatofSP(0.42vs.)0.28).Forexample,thee ectofincreasingW/Bratioonplasticviscosityvs.theproportionofLSP,whendosagesofSPandVAwere xedat0.75%and0.05%respectively,orvs.dosageofSPof0.75%andthepro-portionofLSPof28.5%,respectively,orvs.thedosageofVAof0.05%andtheproportionofLSPof28.5%,isshowninFig.7.TheincreaseinW/Band/orthepro-portionofLSPledtoareductioninplasticviscosity(Fig.7(a)).

4.3.5.Inducedbleeding

Thein uencesoftheproportionofLSPandthedosagesofSPandVAarehighlysigni cantonthein-ducedbleedingaccordingtotheANOVA.Thepropor-tionofLSPisshowntoexhibitthegreateste ectasaprimaryvariableontheinducedbleedingcomparedtothedosagesofSPandVA(0.14vs.)0.05or0.04).However,theANOVAshowsthatthetwo-factorin-teractionofW/BVAishighlysigni cantandhasthegreateste ectoninducedbleeding.TheinteractionsbetweenW/BSPandW/BLSParesigni cantandhaveoppositee ect()0.1vs.0.1).TheincreaseinSPdosagehasagreaterin uenceonreducingtheinducedbleedingthantheincreaseinVAdosage()0.05vs.0.04).Forexample,thee ectofincreasingW/Bratiooninducedbleedingvs.theproportionofLSP,whendosagesofSPandVAwere xedat0.75%and0.05%,respectively,orvs.dosageofSPof0.75%andtheproportionofLSPof28.5%,respectively,orvs.thedosageofVAof0.05%andtheproportionofLSPof28.5%,isshowninFig.8.

For xeddosageofVAandtheproportionofLSP,theincreaseinSPledtoanincreaseininducedbleedingforlowerW/Bupto0.38.However,forhigherW/B(be-tween0.38and0.42),theincreaseofthedosageofSPresultedinareductionintheinducedbleeding(Fig.8(c)).Thisisduetotheimproveddispersionandpack-ingofcementgrainsassociatedwithgreaterSPdosage.Theresultingincreasein uidity(Fig.4)andparticlepacking(versus occulatedcementgrainsthathavealowerpackingdensity)cansubstantiallyreducethetendencyofwatertopercolateamongcementgrainsunderagivenhead,whichre ectsthepermeabilityofthefreshgrout.Theseresultsconcurwiththe ndingsofotherresearchers[3,4].ForlowerW/B(lowerthan0.38),theincreaseinVAdosageexhibitedareductionintheinducedbleeding,howeverforhigherW/Bbeyond0.39,theinducedbleedingseemedtoincreaseasthedosageofVAincreased(Fig.8(b)).

pressivestrength

TheANOVAsgiveninTable4showthatthee ectofW/Bishighlysigni cantoncompressivestrengthat7and28daysandhasthegreateste ectonstrength.Thecompressivestrengthdecreased,asW/Bincreased.TheproportionofLSPandthedosageofSParealsohighlysigni cantoncompressivestrength.TheincreaseoftheproportionofLSPorthedosageofSPledtoareductionincompressivestrengthat28days.Thee ectoftheproportionofLSPoncompressivestrengthwasgreaterthanthatofW/BandSP

dosage.

746L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

4.4.Trade-o betweenSPandVA

Contourresponsesshowingthein uenceofSPandVAdosagesonmini-slumpandplatecohesionmeter,andmini-slumpandyieldvalueforgroutsmadewith xedW/Bof0.35and12%ofLSParepresentedinFig.9.Asexpected,foragivenSPdosage,thecontourdia-gramsofFig.9(a)indicatethattheincreaseinVAdosagereducesthemini-slumpwhiletheplatecohesionmeterincreases.Forexample,formini-slumpof90mm,amixgroutwith0.6%ofSPand0.02%ofVAcanen-surethe uidityof90mm.TheincreaseinVAdosageto0.042%resultedinareductionofmini-slumpto75mm.However,byincreasingtheSPfrom0.6%to0.97%,themini-slumpcanbere-establishedto90mm.Forthesamemix,with0.6%ofSPand0.02%ofVA,thein-creaseinVAdosageto0.042%increasedtheplateco-hesionmeterfrom0.9to1.4mm.Theplatecohesionmetercanbere-establishedto0.9mmbyincreasingSPdosageto0.84%.

ForagivenVAdosage,theresultsfromFig.9(b)indicatethattheincreaseinSPledtoanincreaseinmini-slumpandareductionintheyieldvalue.Forex-ample,agroutmadewith0.05%ofVAand0.3%ofSPhadmini-slumpof60mmandyieldvalueof11.5Pa.TheincreaseofSPto1%resultedinanincreaseofmini-slumpto90mmandareductionofyieldvalueto10.6Pa.However,byreducingtheVAfrom0.05%to0.02%forthesameSPof0.3%,themini-slumpincreasedto75mmandtheyieldvaluedecreasedto11.25Pa.

Contourresponsesshowingthein uenceofSPandVAdosagesonplasticviscosityandinducedbleedingwithmixmadewith0.35ofW/BandLSPof12%and0.42ofW/BandLSP¼45%,arepresentedinFig.10(a)and(b),respectively.ForgivenVAandSPdosages,theincreaseofW/BandtheproportionofLSPresultedinanincreaseininducedbleedingandareductioninplasticviscosity.Forexample,forgroutmadewith0.42W/Band45%ofLSP,plasticviscosityandinducedbleedingof0.145Pasand1.40Lminareobtainedwith0.66%ofSPand0.052%ofVA.TheincreaseofSPto0.97%resultedinareductionofplasticviscosityandinducedbleedingto0.12Pasand1.30Lmin,respec-tively.With0.050%and0.3%ofVAandSP,agroutmadewith0.35ofW/Band12%ofLSPhadaplasticviscosityandinducedbleedingofapproximately0.9Pa

s

L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749747

and0.95Lmin,respectively.TheincreaseofW/BandLSPto0.42and45%,respectively,wouldresultinadropofplasticviscosityto0.17Pasandanincreaseininducedbleedingto1.50Lmin.4.5.Correlationoftestingmethods

The owofgroutisverysensitivetoitsshearhistory.Theabovetestswerecarriedoutwithextremecareinordertokeeptheshearhistory,theexperimentalpro-ceduresandtheirtimingasconstantaspossible.Hence,itisinterestingtoassessthevariouspossiblecorrelationsbetweenthedi erenttestscarriedout.

Fig.11showstherelationshipbetweenthreerheo-logicalvalueswhicharecharacteristicofthegroutatlowshearrates:mini-slump,platecohesionmeterandyieldvalue.Thecoe cientofcorrelationR2betweenmini-slumpandplatecohesionmeter,andthemini-slumpandyieldvalueare0.95and0.56,respectively.Therelationshipseemedtofollowpolynomialsecond-ordermodelandshowedthatwhenthemini-slumpincreasestheplatecohesionmeterandtheyieldvaluedecrease.Therelationshipbetweenmini-slumpandtheplatecohesionmeterwasverygood.Fig.12(a)and(b)illustratethecorrelationsbetweenrheologicalcharac-teristicsofgroutsatlowandhighshearrates:theplatecohesionmeterandplasticviscosity,mini-slumpandplasticviscosity.Thecoe cientsofcorrelationbetweenplatecohesionmeterandplasticviscosity,andthemini-slumpandplasticviscosityare0.80and0.75,respec-tively(R2¼0:80wasobtainedwithouttheresultofmix3).Fig.12showsthattheincreaseinplasticviscosityledtoanincreaseinplatecohesionmeterandareductioninmini-slump.

InFig.13(a)and(b),therelationshipbetweenin-ducedbleedingandmini-slump,andinducedbleed-ingandplasticviscosityarepresented.Itseemsthatthereisnocorrelationbetweenmini-slumpandinducedbleeding.However,Fig.13(b)showsagoodrelation-shipbetweeninducedbleedingandplasticviscosity(R2¼0:85,withouttakingaccountoftheresultsfrommixes3,4and15).Theinducedbleedingwas

inversely

748L.Svermovaetal./Cement&ConcreteComposites25(2003)737–749

proportionaltotheplasticviscosity.This ndingcon- rmswhathasbeenreportedbyNedhietal.[12].

5.Conclusions

Thein uenceofdi erentW/B,dosageofSP,theproportionoflimestonepowderandthedosageofvis-cosityagentcombinationsonrheologybehaviourofcementgroutwereinvestigated.Basedontheresultspresentedinthispaper,thefollowingconclusionscanbedrawn:

(1)TheW/Bratioisshowntoexhibitagreate ect

onmini-slump,plasticviscosityandcompressivestrength.TheincreaseinW/Bratiohasanin uenceonincreasingmini-slumpandinducedbleeding,anddecreasingplasticviscosity,platecohesionmeterandcompressivestrength.

(2)Themini-slump,platecohesionmeter,andyield

valueofgroutsaredominatedprimarilybythedos-ageofSP.TheincreaseintheSPdosageledtoanincreaseinmini-slumpandareductioninplatecohesionmeter,yieldvalue,andplasticviscosity.However,theinducedbleedingseemedtoincreasewhentheSPdosageincreasedforlowW/B,andreducedwithhigherW/B.

(3)Theviscosityagentsigni cantlya ectedthemea-suredpropertiesofthisstudy,exceptcompressivestrength.TheincreaseinVAdosageisshowntoex-hibitareductioninmini-slumpandanincreaseinplatecohesionmeterandplasticviscosity.ForlowW/B,theincreaseinVAdosagereducedtheinducedbleeding,andincreaseditwhenW/Bishigher.

(4)ForagivenW/B,anddosagesofSPandVA,the

mini-slumpandinducedbleedingincreasedwhentheproportionofLSPincreased,whiletheplateco-hesionmeter,yieldvalue,plasticviscosityandcom-pressivestrengthreduced.TheLSPreplacementofcementhadagreatere ectoncompressivestrengththanthechangeofW/Bratio.

(5)Theproposedmethodcanbeusedwithothersetsof

materialssuchas yashorgroundgranulatedblast

slagasreplacementofcement,topredicttherheo-logicalpropertiesandcompressivestrengthofgroutbutthedi erencesbetweenthepredictedandmea-suredvalueswillthenindicatethee ectofthenewmaterialsontheaccuracyoftheproposedmodels.References

[1]KhayatKH,BallivyG.High-performancecementgroutforunderwatercrackinjection.In:MalhotraVM,editor.Proceed-ingsofThirdCANMET/ACIInternationalConferenceonPerformanceofConcreteinMarineEnvironment.1996.p.138–162.

[2]HoulsbyAC.Constructionanddesignofcementgrouting.In:Aguidetogroutinginrockfoundations.NewYork:JohnWileyandSons;1990.p.442.

[3]KhayatKH,YahiaA,Du yP.High-performancecementgroutforpost-tensioningapplications.ACIMaterJ1999;96(4):471–7.[4]KhayatKH,YahiaA.E ectofwelangum-high-rangewaterreducercombinationsonrheologyofcementgrout.ACIMaterJ1997;94(5):365–72.

[5]GhioVA,MonteiroPJM,DemsetzLA.Therheologyoffreshcementpastecontainingpolysaccharidegums.CemConcrRes1994;24(2):243–9.

[6]BuryJR,boratoryevaluationofauniqueanti-washoutadmixturesingrouts.In:MalhotraVM,editor.Pro-ceedingsofFifthCANMET/ACIInternationalConferenceonSuperplasticizersandotherChemicalAdmixturesinConcrete,ACISP-173,Rome,1997.p.445–473.

[7]SonebiM.Factorialdesignmodellingofmixproportionparam-etersofunderwatercompositecementgrouts.CemConcrRes2001;31(11):1553–60.

[8]SonebiM.Experimentaldesigntooptimizehigh-volumeof yashgroutinthepresenceofwelangumandsuperplasticizer.MaterStruct2002;35(250):373–80.

[9]YahiaA,KhayatKH.Experimentdesigntoevaluateinteractionofhigh-rangewater-reducerandantiwashoutadmixtureinhigh-performancecementgrout.CemConcrRes2001;31(5):749–57.[10]KhayatKH,YahiaA.Simple eldteststocharacterize uidity

andwashoutresistanceofstructuralcementgrout.CemConcrAggregates1998;20(1):145–56.[11]NehdiM,MindessS,A

本文来源：https://www.bwwdw.com/article/fz51.html