Availableonlineat
ScienceDirect
www.sciencedirect.comwww.e-ache.com HormigónyAcero2018;69(S1):15–19 www.elsevierciencia.com/hya
Changes
in
mechanical
properties
of
concrete
due
to
ASR
Cambios
en
las
propiedades
mecánicas
del
hormigón
debido
a
la
ASR
Hans
W.
Reinhardt
a,∗,
Hasan
Özkan
b,
Oliver
Mielich
caProf.Dr.-Ing.,DepartmentofConstructionMaterials,UniversityofStuttgart,Stuttgart,Germany bDipl.-Ing.,MaterialsTestingInstitute,UniversityofStuttgart,Stuttgart,Germany
cDr.-Ing.,MaterialsTestingInstitute,UniversityofStuttgart,Stuttgart,Germany
Received24August2017;accepted9February2018 Availableonline25April2018
Abstract
Alkali-SilicaReaction(ASR) isoneofthe mostseveredamagereactionsof concrete.Thereareatleasttwodifferentmechanisms ofASR: Themost commonisthe rapiddisruptionof siliceousaggregates atthe surfaceof the grain,whichoccurs in flintand opalsandstone, for example:Theothersarecalledslowreactingaggregates,andrefer,forexample,tocrackingofthegrainofgreywackeandquartzporphyry.Both reactionsareduetothedissolvingofSiO2byalkalis.Thepresentpaperreportsontestsonslowreactingaggregates.
Whilemoststudiesintheliteratureconcentrateontheexpansionbehavioroftheaggregates,inthisarticlefocusiscenteredonthemechanical properties,suchasstrengthandcreep.Theaggregatesusedaregreywacke,quartzporphyry,andcrushedsedimentsfromtheUpperRhineValley. ©2018PublishedbyElsevierEspa˜na,S.L.U.onbehalfofAsociaci´onEspa´ooladeIngenier´oaEstructural,ACHE.
Keywords:Concrete;Alkali-silicareaction(ASR);Expansion;Mechanicalproperties;Creepmodulus
Resumen
Lareacciónálcali-sílice(ASR)esunodelosdeteriorosmásimportantesqueseproduceenelhormigón.Existenprincipalmentedosmecanismos deASR:Elmáscomúneslairrupciónrápidadelosáridossilíceosenlasuperficiedelgranoqueporejemplosucedeconelpedernalylaarenisca opal,mientrasquelosotrosmecanismossonlosdeáridosdereacciónlentay,porejemploserefierenalagrietamientodelgranodegrauvacay pórfidodecuarzo.AmbasreaccionessedebenaladisolucióndelSiO2porlosálcalis.Elpresentetrabajoversasobreensayosenáridosdereacción
lenta.
Mientrasquelamayoríadelasinvestigacionesenlaliteraturaseconcentranenelcomportamientodeexpansióndelosáridos,nosotrosnos centramosenlaspropiedadesmecánicastalescomoresistenciamecánicayfluencia.Losáridosutilizadossongrauvaca,pórfidodecuarzoy sedimentostrituradosdelValleAltodelRin.
©2018PublicadoporElsevierEspaña,S.L.U.ennombredeAsociaciónEspañoladeIngenieríaEstructural(ACHE).
Palabrasclave: Hormigón;Reacciónálcali-sílice(ASR);Expansión;Propiedadesmecánicas;Módulodefluencia
1. Introduction
ASRmanifestsitselfmainlybyexpansionandcracking.The surfaceofconcretestructuresshowsrandomcrackingor, depen-dentoftheloadingdirectionlongitudinalortransversecracking.
∗Correspondingauthor.
E-mailaddress:reinhardt@iwb.uni-stuttgart.de(H.W.Reinhardt).
Ithappensvery frequentlythatsuch structuresareabandoned althoughtheloadingcapacityisstillsufficient.Thiswasthe start-ingpointoftheinvestigationwithseveralquestions:Whatare theremainingcompressivestrengthandtensilestrength,what are thedeformations duetoexpansion of the aggregates,due to shrinkage andcreep when ASR happens.It may be more economictoanalyzethestateofastructuretakingaccountof inherentmechanicalpropertiesandmonitorthefurther
behav-https://doi.org/10.1016/j.hya.2018.02.001
iorinstead of demolishing the structure.That means that the mechanicalpropertieshavetobeknown.
In the current investigation, fourtypesof aggregateshave beenused:acrushedgreywackefromtheHarz(GW),crushed gravelfromtheUpperRhineValley(OR),quartzporphyryfrom Halle(QP),andanotherquartzporphyryfromtheBlackForrest (QPSW).Accordingtothe40◦Cfog-roomtest[1]thefirstthree aggregateswerealkalisensitivewhiletheforthaggregatewas insensitive. The first three aggregates belong tothe category of slowreactingaggregatesi.e.anASRdamage wouldoccur inpracticeonlyafterabout15years.Howevertherearemany structuresmainlyintheEastpartofGermany,whichsufferfrom suchsignsofdeterioration.
2. Concretecomposition
ACEMI32.5RaccordingtoEN197-1[2]wasusedwith additionofK2SO4tothemixingwatersuchthattheNa2Oequi. amountedto1.30M-%.Theconcretewasmadewith400kg/m3 cementandawatertocementratioof0.45.30M-%ofthe aggre-gateswereinertquartzsandupto4mm,70M-%werethecoarse aggregatesupto16mmgrainsize.Thegradingcurvefollowed almostaFullerparabola.
3. Testingprogram
For strength testing, 150mm cubes and cylinders with 150mmdiameterand300mmlengthwereproduced.Modulus ofelasticity,creepandshrinkageweredeterminedoncylinders, compressive strength on cubes, tensile strength on dog-bone specimens,deformationduetoASRinthe40◦Cfogroomon cylinders.Testingtookplacebeforeanddirectlyafter installa-tionofthe28daysmoistcuredspecimensinthefogroomand after140,280and560daysstorageinthefogroom.
Forcreeptesting,cylindersweremountedinthetestrig.The compressiveforcewasappliedbyahydraulicramandheld con-stantbyBellevillesprings.Thedeformationwasmeasuredwith dialgauges.Thestresswasdecidedtobeonethirdofthe charac-teristicstrengthofconcreteafter28daysstandardcuring.The specimenswere takenout ofthe fogroomafterthe predeter-minedtime,transportedtothecreeproomwithconstantclimate of23◦Cand80%RH,mountedinthetestrig,andloaded.80% RHwasusedbecauseASRhadtocontinue,fromexperienceit isknownthatASRwouldstopatlowerrelativehumidity[3].
4. Testingresults
4.1. Expansioninthefogroom
ConcretewhichisaffectedbyASRtypicallyexhibits expan-sioninmoistenvironment.Thespecimenswerestoredat40◦C andalmost100%RHforacertaintime.Thelengthchangeis plottedinFig.1.Thedotsgivetheexperimentalresults,thelines aregeneratedwiththeapproximationfunctionofLarive[4].
Obviously, the crushed gravel from the Upper Rhine Val-leyexpandedmost,the greywackeissecond,quartzporphyry fromHalleisthree,andquartzporphyryfromtheBlackForest
Figure1.Expansionofspecimensinthefogroom,functionofLarive[4]:ε(t)=
ε∞ 1−e−τCaract
1+e−
(t−τLatence)
τCarac
.
Figure2.Rateofexpansion.
doesnotexpand.TheexpansionfollowsalwaysanS-curveas known fromLarive[4].Alllinesstartatalowlevel,increase thenstronglyandfinallystaywithanasymptoticmaximum.The timeofstrongincreasedependsonthereactivityofthe aggre-gate.TherateofexpansionisplottedinFig.2,itrepresentsthe firstderivativeoftheLarivefunction.
Threeaggregatesweresuitableforthedeterminationofthe expansion rate, crushed gravel from the Upper Rhine Valley (OR),greywacke(GW)andquartzporphyry(QP).ORexhibits themaximumat130dwhileGWshowsthemaximumat190d. QP islesspronouncedwithamaximumat392d.Thetypical shapeofthecurveisveryimportantandcanlaterbeusedfor thepredictionofexpansioninrealstructures.
4.2. Modulusofelasticity
Table1
Staticmodulusofelasticityofconcretemeasuredafteracertainstoragetime inthefogroominMPa(inbrackets:therelativevaluein%),meanofthree specimens.
TypeofaggregateAfter28dcuring Storagetimein40◦Cfogroomindays
140 280 560
OR 37,702(100) 15,805(42) 20,750(55) 28,934(77) GW 40,302(100) 43,391(108) 16,440(41) 25,931(64) QP 36,031(100) 41,078(114) 41,551(115) 22,342(62) QPSW 31,729(100) 38,682(122) 39,543(125) 40,191(127)
Table2
CompressivestrengthofconcreteinMPameasuredafteracertainstoragetime inthefogroom(inbrackets:therelativevaluein%),meanofthreespecimens.
Typeofaggregate After28dcuring Storagetimein40◦Cfogroomindays
140 280 560
OR 52.1(100) 57.9(111) 61.5(118) 65.4(125) GW 50.6(100) 69.2(137) 71.4(141) 72.7(143) QP 53.4(100) 69.6(130) 74.5(140) 81.5(153) QPSW 54.5(100) 72.5(133) 74.2(136) 83.8(154)
Theelasticmodulus of crushedgravelof theUpper Rhine valleydecreasesstronglyafter140dandexhibitsonly42%of theoriginalvalue.Thereafteritincreasesagain.Theother aggre-gatesshowanoppositebehavior.Themodulusincreasesdueto goodhydration conditions (40◦C andwater saturationof the ambientair)firstlyanddecayonlyafter280or560days.The insensitivequartzporphyryfromtheBlackForestincreases con-tinuously.ComparingwithFig.2showsthattheminimumvalue of the elastic modulus coincides roughly withthe maximum reactionrate.
4.3. Compressivestrength
Compressivestrengthhasbeendeterminedon150mmcubes.
Table2showstheresultsasmeanofthreespecimens.
Allconcretesstartatabout50MPaandincreasecontinuously to65–84MPadependingontheaggregate.Thestrengthincrease is also dueto the favorablehydration conditions.Obviously, ASRdidnotaffectthecompressivestrengthwhichmeansthat thecompressivestrengthcannotbeusedasindicatorofASR.
4.4. Tensilestrength
Theinvestigation of thesplitting tensilestrength wasvery limited. The results of aggregate OR and QP are shown in
Table3.
ThestrengthofORdecaysafter140 daysinthefogroom to1.15MPawhichisonly25%oftheoriginalvalue,it recov-ersto31%after280daysandto44%after560days.Fig.1with theexpansionofconcretehasshownastrongexpansionatabout 140dayswhichexplainsthe reductionofthetensilestrength. Slow/lateaggregategrainsexhibitfractureinthegrain[8]and, since the grain is responsible for the load transfer onemust concludethatASRhascausedthestrengthloss.Withongoing
Table3
TensilestrengthofconcreteinMPameasuredafteracertainstoragetimeinthe fogroom(inbrackets:therelativevaluein%),meanoftwospecimens.
TypeofaggregateAfter28dcuring Storagetimein40◦Cfogroomindays
140 280 560
OR 4.68(100) 1.15(25) 1.43(31) 2.04(44) QP 3.99(100) 3.40(85) – 3.19(80)
Figure3.CreepstrainofaggregateOR.
hydration strengthrecovers partlybut cannotreach the origi-nalvalueagain.QPshowsrespectivevaluesof117%after280 daysand80%after560days.AlooktoFig.1revealsthatASR hasstartedonlyafterabout320dayswhichmeansthatongoing hydration hascausedthestrength increaseat280 days.Later strengthdecreasesto80%duetoASR.Comparingthe devel-opmentofcompressivestrengthwithtensilestrength,the two propertiesdonotfollowthesamemanner.Thishasbeenreported alsobyotherresearchers[9].Thus,thetensilestrengthisabetter indicatorforASRthancompressivestrength.
4.5. Creepundercompressiveload
Figs.3–6showthedevelopmentofstrainwithtimeasmeanof twospecimens.Thestressisconstantduringthetimeunderload, howeveritdependsontheaggregate.Assaidbeforethestress levelhasbeenchosenasonethirdofthecharacteristicstrength after28daysstandardcuring.Thediagramsgivethe informa-tion.Thelegendinformsaboutthepre-storageinthe40◦Cfog room. Ofcourse,thissituation isartificial,however sincethe climateinthetestingroomistakenconstantwith80%RHone canassumethat ASRisstill progressing. (Remark:Although thecreepstrainsarenegative(shortening)theyareplottedwith aplussign,i.e.asabsolutevalues.)
Figure4.Creepstrainofaggregategreywacke.
Figure5.CreepstrainofaggregateQP.
Figure6.CreepstrainofaggregateQPSW.
largestforthespecimenswhichwerestoredinthefogroomfor 140days.LookingtoFig.2onecanseethatthisisthetimeof strongestreaction.Thespecimenswith280and560daysstorage showlesscreepwhichisatokenofgreaterhydration.
Fig.4belongstothegreywacke.
Table4
CreepmodulusofconcreteinMPameasuredafter365daysloadingandaftera certainpre-storagetimeinthefogroom,meanoftwospecimens.
Pre-storage Typeofaggregate
OR GW QP QPSW
28daysmoistcuring 10,520 12,518 10,967 13,214 140daysinfogroom 6117 21,393 14,159 25,056 280daysinfogroom 11,596 5956 23,195 28,021 560daysinfogroom 11,030 10,531 9480 25,154
ThecurvesmorespreadthanitwasinthecaseofOR.The largestcreep occursforconcretewhichwasstoredinthefog roomfor280days.Thistimecoincidesroughlywiththetime ofgreatestASRreaction.Thelowercurvefor140daysmustbe causedbymorehydrationcomparedtothecurvewithoutstorage inthefogroom.The560dayscurvemustbeattributedtothe ongoingASR.
Fig.5showstheresultsofthequartzporphyryfromHalle. The concretewhichhad nopre-treatmentinthe fog room showsthelargestcreep.Comparedtotheothersitistheyoungest concrete.Thelineforthe140dconcreteislowerandthelinefor the280dconcreteisevenlower.Thisbehaviorcanbeattributed tothe ongoinghydration withouteffectofASR.The line for the560dpre-treatmentapproachesalmostthelineofthezero concrete,i.e.ASRinfluencescreepconsiderably.
Fig.6referstothequartzporphyryfromtheBlackForrest, theaggregatewhichisinsensitivetoASR.
Thisfigureshowsbesttheinfluenceofprogressinghydration. Thezeropre-treatmentconcreteexhibitsthelargestcreepand theotherlineslieclosetogether.
Comparingthefourforegoingfigures,onecanseeacommon feature:TheaggregateswhichareASRsensitiveshowalarger creepthantheinsensitiveaggregateQPSW.TheaggregatesOR andGWwhichundergothelargestexpansionduetoASRshow largercreepstrainthanthelassexpandedQP.Ontheotherand, thesimultaneousinfluenceofabsoluteageofconcretebecame obvious.
5. Creepmodulus
Ifstructureshavetobeanalyzedwithrespect tolong-term deformation onecan have alook to polymertechnology. To thisend,theirengineersusetheso-calledcreepmodulus.The creepmodulusEcristhequotientofstressdividedbytheload induced strainEcr=σ/(εel+εcr)withεel the elasticstrainand
εcrthecreepstrain.Thecreepmoduluscouldalsobeusedfor concretestructures.Table4showsthenumberscalculatedwith thedeformationatoneyear(withoutshrinkage,ofcourse).
QPSWexhibitsthelargestvalues.Italsoreflectsthefeatureof
Figs.1and2.
6. Outlook
Theaimoftheinvestigationswastoincreasetheknowledge withrespecttoshort-timeandlong-timemechanicalproperties of ASR affectedconcrete. The knowledge issuitable for the analysisofthestatestructuresinthelightofloadingcapacityand deformation.However,thereareseveraldisciplinesquestioned: firstly,thetestingresultshavetobetransposedtoreallifeofthe structure,secondly,ASRdevelopmenthastobemodeledforthe structureintermsoftimeandspace,andthirdly,anASRaffected structureshouldbemonitoredcontinuously.Asthefirst point isconcerned thereareresults of ASRatreal exposurewhich showthatthedevelopmentofASRtakesabouttentimesmore timethaninthe40◦Cfogroom[10].Thesecondpointcanbe satisfiedwiththeaidofmodernfinite-elementcodeswhichtake accountofphysicalaspects(diffusion,permeation,temperature, humidity)andstress[11].Thenowadaysmonitoringtechniques areabletomonitorcontinuouslyandwirelessthecrucialspots ofstructures[12].Inanycasethereisastrongdemandforgood knowledgeofthematerial.
7.Finalconclusions
The investigations have been performed on concrete with threetypesofslowreactingalkalisensitiveaggregatesandone insensitiveaggregate.Thefollowingconclusionscanbedrawn. TheexpansionduetoASRmeasuredafteracertainstorage timein the 40◦Cfog room follows an S-curve for the three concreteswithalkalisensitiveaggregate.Thetimeofmaximum reactionrateisstronglydependentonthetypeofaggregate.
ThestaticmodulusofelasticitydecreaseswithASRtoa mini-mumvalueandincreasesthereafteragain.Thetimeofminimum coincidesroughlywiththemaximumreactionrate.
ThecompressivestrengthseemsnottobeaffectedbyASR. Thetensilestrengthdecayswithtimeinthefogroom. Com-pressiveandtensilestrengthdonotfollowthesamebehavior.
Creep is enlarged by ASR. The largest creep rate goes togetherwiththelargestprogressofASR.
Asameanfortheanalysisoflong-termdeformationthecreep modulushasbeencalculatedwhichshowsasimilartendencyas creep.
Theoutlookprovidessomeideashowthegainedknowledge canbeusedintheanalysisofstructuresinrealenvironment.
Allconclusionsareonlyvalidforthematerialtested.Onehas tobeverycautiousindrawinggeneralconclusionsbecausethe aggregatesareofnatural originandvaryfromplacetoplace. It can happen that rocks taken from the same quarry differ strongly.
Acknowledgement
Thefirstauthorhasexperiencedalonglastingfriendshipwith CarmenAndrade.WemetatmanyeventsofRILEM,at commit-teemeetings,symposiaandworkshops.Carmenshowedalways muchinterestinvariousfieldswhichhavetriggeredresearchof commoninterest.Thankyou.Admultosannos!
TheGermanResearchCommunityhasgreatlysupportedthe projectunderRE691/39-1whichisthankfullyacknowledged.
References
[1]DeutscherAusschussfürStahlbeton(DAfStb),VorbeugendeMaßnahmen gegenschädigendeAlkalireaktionimBeton(AlkaliRichtlinie),Beuth Ver-lag,Berlin,DeutscherAusschussfürStahlbeton,Oktober,2013. [2]EN197-1:2011Cement.Composition,specificationsandconformity
cri-teriaforcommoncements.
[3]D.Stark,Themoistureconditionoffieldconcreteexhibitingalkalisilica reaction,in:DurabilityofConcrete,2ndInt.Conf.MontrealCanada,Vol. IIACI,SP126,Detroit,1991,pp.973–987.
[4]C.Larive,Apportscombinésdel‘expérimentationetdelamodélisationà lacompréhensiondel’alcali-réactionetdeseseffetsmécaniques,thèsede doctorat,L’ÉcoleNationaledePontsetChaussées,1997.
[5]H.W.Reinhardt,O.Mielich,Staticvs.dynamicYoung’smodulusas indi-catorforASRdegradation,in:C.Leung,K.T.Wan(Eds.),Advancesin ConstructionMaterialsthroughScienceandEngineering,2011,RILEM PRO79,S.104(Abstractbook)ISBN978-2-35158-116-2.
[6]P. Helene,M.Carvalho,J.Pachero,Engineering fieldtests for alkali-aggregatereaction,Struct.Concr.J.Fib.18(2017)349–355.
[7]DIN1048-5:1991Testingmethodsforconcrete;hardenedconcrete, spe-ciallypreparedspecimens.
[8]H.W.Reinhardt,O.Mielich,Fracturetoughnessofalkali-sensitiverocks inalkalinesolution,Int.J.RockMech.Min.Sci.70(2014)552–558. [9]“Alkali-aggregatereactionconcrete”(ICAAR),proceedingsvolumessince
1974.
[10]E.Siebel,M.Böhm,I.Borchers,C.Müller,J.Bokern,E.Schäfer. AKR-Prüfverfahren–VergleichbarkeitundPraxis-Relevanz,Betontechnische Berichte2004–2006,VDZDüsseldorf.
[11]J.Oˇzbolt,MASA–MAcroscopicSpaceAnalysis.Berichtzur Beschrei-bungdesFEProgrammesMASA,InstitutfürWerkstoffeimBauwesen, UniversitätStuttgart,Stuttgart,1998.