Potential of Asphalt Pavement Analyzer to Predict Rutting of Hot Mix Asphalt

8/9/2019 Potential of Asphalt Pavement Analyzer to Predict Rutting of Hot Mix Asphalt

http://slidepdf.com/reader/full/potential-of-asphalt-pavement-analyzer-to-predict-rutting-of-hot-mix-asphalt 1/27

POTENTIALOFASPHALTPAVEMENTANALYZERTOPREDICTRUTTINGOFHOTMIXASPHALT

By

PrithviS.Kandhal

RajibB.Mallick

PaperpublishedintheProceedingsoftheInternationalConferenceonAcceleratedPavementTesting,Reno,Nevada,October1999

277TechnologyParkway •Auburn,AL36830



POTENTIALOFASPHALTPAVEMENTANALYZERTOPREDICTRUTTINGOFHOTMIXASPHALT

By

PrithviS.Kandhal

AssociateDirectorNationalCenterforAsphaltTechnology

AuburnUniversity,Alabama

RajibB.Mallick

AssistantProfessorWorcesterPolytechnicUniversity

Worcester,Massachusetts

PaperpublishedintheProceedingsoftheInternationalConferenceonAcceleratedPavementTesting,Reno,Nevada,October1999



DISCLAIMER

Thecontentsofthisreportreflecttheviewsoftheauthorswhoaresolelyresponsibleforthefactsandtheaccuracyofthedatapresentedherein.ThecontentsdonotnecessarilyreflecttheofficialviewsandpoliciesoftheNationalCenterforAsphaltTechnologyofAuburnUniversity.Thisreportdoesnotconstituteastandard,specification,orregulation.

i



ABSTRACT

Ruttingisacommonprobleminhotmixasphaltpavements,particularlyinhotclimatesandat

intersections.TheAsphaltPavementAnalyzer(APA)isalaboratoryacceleratedloadingequipmentthatcanbeusedtoevaluateruttingpotentialofHMA.ThisstudywascarriedouttoevaluatethepotentialofAPAtopredictrutting.Specifically,theobjectivesweretofindthesensitivityoftheequipmenttochangesinaggregatetypeandgradation,performancegrade(PG)ofasphaltbinder,andevaluatetheequipmentbycomparingthetestresultswiththetestresultsfromSuperpavesheartester(SST).Mixesfrompoor,fairandgoodperformingpavementswerealsotestedwiththeAPAtodeveloparutdepthcriteriaforevaluationofmixes.

Binderandsurfacecoursemixesweremadewithgranite,limestoneandgravelaggregates,with

gradationsabovethemaximumdensityline,gradationsthroughtheSuperpaverestrictedzoneincloseproximityofthemaximumdensityline,andgradationsbelowthemaximumdensityline.

Resultsfromtestswithdifferentaggregates,gradations,andbindertypesshowthattheAPAis

sensitivetothesefactorsand,therefore,hasapotentialtopredictrelativeruttingofhotmixasphaltmixtures.TheAPAhadafaircorrelationwiththerepeatedshearconstantheighttestconductedwiththeSuperpavesheartester.

KEYWORDS:asphaltpavementanalyzer,APA,loadedwheeltester,rutting,hotmixasphalt,

restrictedzone,Superpavesheartester

ii





Kandhal&Mallick

Table1.GradationofAggregates

PercentPassingCourse

SieveSize(mm) ARZ TRZ BRZWearing 19.0 100 100 100

12.5 95 95 95

9.5 86 86 86

4.75 61 61 61

2.36 45 39 33

1.18 35 29 23

0.6 26 21 16

0.3 19 16 13

0.15 11 10 90.075 4.0 4.0 4.0

25.0 100 100 100

Binder 19.0 95 95 95

12.5 80 80 80

9.5 68 68 68

4.75 45 45 45

2.36 41 35 29

1.18 31 25 19

0.6 24 19 140.3 17 14 11

0.15 11 10 9

0.075 4.0 4.0 4.0

ManystudieshaveshownthatthereisaninteractionoftheeffectofgradationandaggregateshapeandtextureonruttingpotentialofHMA.Mixescontainingdifferentaggregates,butwithsamegradationcanshowsignificantlydifferentruttingpotential.Inordertotesttheeffectofaggregatetype,itwasdecidedtotestmixeswiththreetypesofaggregates:granite,limestone,andgravel.ThepropertiesoftheaggregatesareshowninTable2.Allthreeaggregatesare

crushedaggregate.However,thepercentageofcrushedfacesingravelislowerthanthepercentageofcrushedfacesingraniteandlimestone(thelattertwobeing100percent).

Apartfromgradationandtypeofaggregatethetopsizeofaggregateisalsobelievedtohave

significanteffectonruttingpotential.Experienceshowsthatstiffbindercoursewithbiggeraggregateshavelessruttingpotentialcomparedtorelativelymoreflexiblewearingcourseswithfineraggregatesandhigherbindercontent.Hence,anypass/failcriteriafortestingmixeswiththeAPAmustbedevelopedseparatelyforwearingandbindercourses.Itwasplannedtotestbothwearingandbindercourses,withmaximumnominalsizeof12.5mmand19.5mm,respectively,withtheAPA.BothbinderandwearingcoursegradationsareshowninTable1.Similartothegradationofthewearingcourse,thegradationofthebindercoursedifferonlyneartherestrictedzone.

2



Kandhal&Mallick

Table2.PropertiesofAggregates

Property

BulkSpecificGravityofCoarse

AggregateBulkSpecificGravityofFineAggregate

FracturedFace(%)

2Face1

Face

NAAVoids(%)

Granite

2.6882.712

100

100

49.3

Limestone

2.7272.639

100

100

45.8

Gravel

2.6112.623

90.3

95.7

46.0

Allofthetestsampleswerepreparedat4percentairvoidswiththeSuperpavegyratorycompactor(SGC).AllofthemixdesignswereconductedbycompactingHMAsamplesto

Ndesign.TheNdesignvaluewasselectedas76,correspondingtoadesigntrafficlevelof0.3-1millionESALS.Thiswasdonetoavoiddiscrepanciesinoptimumasphaltcontentduetovariationincorrectionfactors.MixesweresubsequentlycompactedtoNmax,usingoptimumasphaltcontent,tocheckdensityatNmax.SincetherateofdensificationofHMAduringsamplepreparation,asindicatedbyslopeofgyrationversusdensityplot,maypossiblyindicatetheruttingpotentialofHMA,itwasdecidedtocorrelateslopeofgyrationplotwithrutdepthsfromAPAtests.

TheSuperpavesystemhasintroducedtheuseofPerformanceGraded(PG)asphaltbinders.The

gradeofasphaltbindershouldcorrespondtotheexpectedhighandlowtemperaturesofthelocationofthepavement.Forexample,aPG64-22asphaltbindershouldbeusedwheretheexpectedmaximumhighandlowpavementtemperaturesare64/Cand-22/C,respectively.Theasphaltbindersarerequiredtoexhibitspecificminimumandmaximumvalueswhentestedfordifferentpropertiesataparticulartemperature,tobepermittedforuseatthatparticular

temperature.Forexample,tobeusedwithsufficientreliabilityatalocationwherethemaximumhighpavementtemperatureis58/C,theasphaltbinder,whentestedat58/C,mustexhibitadynamicshearrheometerstiffnessofatleast1.0kPa.Becauseoftheinfluenceofbinderstiffness,mixeswithsameaggregategradationbutdifferentasphaltbindersshouldexhibitdifferentruttingpotentialatthesametemperature.However,mixeswithsameaggregatebuttwodifferentbinders(ofgradePGx-zandPGy-z)shouldexhibitsimilarruttingpotentialwhentestedatx/Candy/C(PGx-ztestedatx/C,PGy-ztestedaty/C),respectively.Toevaluatetheeffectofbinderonruttingpotentialofmixes,itwasdecidedtotestmixeswithPG64-22andPG58-22asphaltbinderat64/Cand58/CwiththeAPA.ResultsfromlowandhightemperaturebindercharacterizationtestsforthetwoasphaltbindersareshowninTable3.

Table3.AsphaltBinderProperties

Test

Test

PG58-22

Value

Test

PG64-22

Value

Temperature Temperature

G*/sin*(original) 58°C 1.24kPa 64°C 1.76kPa

G*/sin*(RTFO) 58°C 2.91kPa 64°C 3.24kPa

G*sin*(RTFO-PAV) 19°C 2195kPa 22°C 4567kPa

Stiffness,S(RTFO- -12°C 118MPa -12°C 255MPaPAV)

Slope,m(RTFO-PAV) -12°C 0.43 -12°C 0.32

3



Kandhal&MallickTheSuperpavemixdesignandanalysissystemrecommendstheuseofSuperpaveShearTester(SST)todeterminetheruttingpotentialofHMA.TheSSTisbelievedtobeaverysensitive,sophisticatedmaterialcharacterizationequipmentwiththecapabilityofidentifyingthefundamentalpropertiesofHMA.TocomparetheresultsofAPAwiththeresultsfromtheSST,itwasdecidedtotestsomeselectedmixeswiththeSSTaswell.TwoSSTtestswereselectedfortheirusefulnessandsimplicity:therepeatedshearatconstantheight(RSCH)andrepeatedshearatconstantstressratio(RSCSR).TheRSCHcangiveanestimateofrutdepth,whereastheRSCSRiscapableofidentifyingmixessusceptibletoruttingatlowairvoids.

Anylaboratoryruttester,howeversensitiveitmightbe,isboundtohavescaleeffectsontest

results.Becauseofthedifferenceinlayerthickness,underlyingsupport,confiningpressure,andstressdistribution,amongotherthings,theresultsofruttestsinalaboratoryruttesterwillbedifferentfromactualrutdepthsinpavement.However,torecommendaspecificrutdepthforacceptance/rejectionofHMA,thereisaneedtocorrelatetheresultsfromtheAPAtestandactualrutdepthsinpavements.MixeswereobtainedbytheAlabamaDepartmentofTransportation(ALDOT)frompavementswithmajor,intermediateandminorrutting.Itwas

decidedtotestthesemixeswiththeAPA,andcorrelatetheresultswithactualrutdepths.Inthisway,laboratoryrutdepthscorrespondingtomajor,intermediateandminorruttingcanbeusedasbasisforspecificationofacceptance/rejectioncriteria.

TESTPLAN

Totestmixeswithdifferentaggregates,gradation,nominalmaximumsizeaggregatesandbinder,mixeswerepreparedwithgranite,limestone,gravel,withgradationabove,throughandbelowtherestrictedzone,fortypicalALDOTwearingandbindercourses,andwithPG64-22andPG58-22asphaltbinders.Table4showsthemixtestmatrix.Inthefirststep,dryruttestswereconductedondifferentmixes.MixeswithPG64-22andPG58-22asphaltbindersweretestedat64/Cand58/C,respectively.Mixeswithhighandlowrutdepths,asobtainedfromdryruttests,weretestedwiththeSSTunderrepeatedshearasconstantheightandrepeatedshearat

constantstressratioconditions.Themixesexhibitinghighruttingpotentialsindryruttestsweretestedunderwater,andalsotestedwiththeAASHTOT283(ModifiedLottman)procedure.Testswerealsoconductedunderdryconditionswithmixesobtainedfromhigh,intermediateandlowruttingpavements.AllAPAtestswereconductedwithawheelloadof445Nandahosepressureof690kPa.

Thedatawasanalyzedtoanswerthefollowingspecificquestions:

1.a.IstheAPAsensitivetoaggregategradation?b.Ifyes,howarethegradationscharacterizedaccordingtotheirruttingpotential?

2.Howdotherutdepthsfromwearingandbindercoursescompare?DoestheAPAshowlessrutdepthsforbindercourses,asexpected?

3.Whatistheeffectofasphaltbinderonruttingpotential?DoestheAPAshowsimilarrutdepthsformixeswithdifferentbinderstestedattheircorresponding(PGgrade)

hightemperatures?4.IsthereanycorrelationbetweenAPArutdepthandgyratorycompactionslopesof

differentmixes?AremixesmeetingN initialandNmaxcriterialikelytoshowlessruttingpotentialcomparedtomixeswhichdonotmeetthesecriteria?

5.Isthereanycorrelationbetweenrutdepthandbinderfilmthickness?6.HowdotheresultsfromtestswithAPAcomparewiththeresultsfromtestswith

SST?

4



Kandhal&Mallick

Table4.MixTestMatrixAGGREGATE

RoundedGravel Granite Limestone

Wearing Binder Wearing Binder Wearing BinderCourse Course Course Course Course Course

A T B A T B A T B A T B A T B A T BR R R R R R R R R R R R R R R R R RZ Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z Z

CylinderforAPA@4%VTM

TotalsCylinderforRS@CSR3%VTM

Totals

CylinderforRS@CH6%VTM

Totals

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

x

x

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

x

x

x

x

x

x

(6)xx

(2)

xx

(2)

AdditionalworkincludedtestingthreesectionsonI-85(southofGeorgia/Alabamaborder)whichwereshowinggood,fairandpoorperformanceintermsofrutting,wereidentified.Cores

wereobtainedfromeachofthesesectionsfromthetravellane,about300mmawayfromthepavementedge.MixesA,B,andCarecharacterizedasgood(norutting),fair(6mmrutting),andpoor(12.5mmruttingormore),respectively.Inthelaboratory,thewearingcoursesweresawedofffromthecores,andthebulkspecificgravitiesweredetermined.Thecoreswerethenheatedandpartofthemixeswereusedfordeterminingthetheoreticalmaximumdensity,andasphaltcontent.Tengyratorysampleswerethencompactedwitheachtypeofmix,at4%airvoids.ThesampleswerethentestedwiththeAPAfordeterminingtheruttingpotential.

TESTRESULTSANDANALYSIS

DatafromtestingwiththeAPAwereanalyzedasdiscussedinthefollowingsections.

DifferencesBetweenRutDepthsofMixeswithGradationsPassingAbove,Throughand

BelowRestrictedZoneStatisticalanalyseswereconductedtodetermineifdifferencebetweenrutdepthsofmixeswithgradationspassingabove(ARZ),through(TRZ),andbelow(BRZ)therestrictedzonearesignificant.Specifically,ananalysisofvariance(ANOVA)( "=0.05)andDuncan'smultiplerangetest(meanseparationtechnique)wereconductedwiththedata.Table5showsthemeanandstandarddeviationofrutdepthsfromdifferentmixes.Analysisofwholedatasetindicatessignificanteffectofaggregatetype,asphaltbindertype,gradation,coursetype,andaninteractionofaggregateandgradation(Table6).

5



Kandhal&Mallick

Table5.RutDepthsforMixeswithDifferentGradationsAsphalt Course Aggregate Gradation MeanRut Standard Ranking(A

Depth(mm) Deviation, hasmoreRutDepth ruttingthan(mm) B);

Significancelevel=5%

PG64-22 Wearing Granite ARZ 4.48 0.737 AB

TRZ 4.30 0.825

BBRZ 5.35

0.561 A

Limestone ARZ 3.77 0.608 B

TRZ 3.90 0.452

BBRZ 6.23

1.036 A

Gravel ARZ 6.46 0.656 ATRZ 5.77 0.342 AB

BRZ 5.64 0.776 B

Binder Granite ARZ 3.48 1.205 A

TRZ 1.62 0.348

BBRZ 3.43

0.567 A


TRZ 3.98 0.287

BBRZ 5.62

1.531 AGravel ARZ 5.19 1.034 A

TRZ 4.35 0.678

ABRZ 4.53

0.492 A

PG58-22 Wearing Granite ARZ 6.59 1.191 A

TRZ 3.81 0.442

BBRZ 6.01

0.622 A


TRZ 5.47 1.148

BBRZ 7.16

0.949 A

Gravel ARZ 7.95 0.539 A

TRZ 6.036 0.477

BBRZ 5.24

0.708 C

Binder Granite ARZ 3.4 0.446 A

TRZ 2.8 0.283

ABRZ 2.85

0.707 ALimestone ARZ 4 0.186 B

TRZ 5.04 0.581

BBRZ 9.49



2.021 A

Gravel ARZ 6.41 1.005 A

TRZ 5.23 0.621

BBRZ 4.65

0.375 B6



Kandhal&MallickTable6.AnalysisofVarianceforRutDepthsofMixeswithDifferentGradations,Binder,

andCourses

Source DF MeanSquare FValue Pr>F

Aggregate 2 51.59 45.64 0.0001Asphalt 1 34.96 30.94 0.0001

Gradation 2 24.35 21.54 0.0001

Course 1 57.56 50.92 0.0001

Aggregate*Gradation 4 33.79 29.90 0.0001

Consideringalldata,mixeswithgravelandlimestoneaggregatesgenerallyshowhigherruttingthangraniteandmixeswithasphaltPG58-22showedmoreruttingcomparedtoasphaltPG64-22.Also,forgraniteandlimestone,mixeswithgradationbelowrestrictedzonegenerallyshowedhighestamountofrutting,whereasthroughrestrictedzonegenerallyshowedlowestrutdepth,andaboverestrictedzonegenerallyshowedintermediaterutting.Forgravel,inmostcasesthemixeswithbelowrestrictedzonegradationshowtheleastamountofrutting,whereasmixeswithaboverestrictedzonegradationshowhighestamountofrutting;mixeswithgradationsthroughtherestrictedzoneshoweitherhigherorsimilarruttingasmixeswithgradationbelowtherestrictedzone.

Analysisofindividualgroupsofdatashowedthat:

1.TheeffectofgradationongraniteandlimestonewearingandbindercourseswithPG64-22asphaltissignificant,withbelowrestrictedzonegradationshowinghigherruttingcomparedtoaboveandthroughrestrictedzone.TheeffectissimilarandsignificantforgranitePG58-22wearingcoursesbutnotsignificantforgranitebindercourse.

2.TheeffectofgradationisnotsignificantforruttingofgravelwearingandbindercoursemixeswithPG64-22.Theaboveandthroughrestrictedzonemixesshowedslightlyhigherruttingcomparedtobelowzonemixes.

However,thedataforPG58-22wearingandbindercoursemixesshowedsignificanteffectof

gradation,andtheARZ,TRZandBRZgradationshowedlowest,intermediate,andhighestamountofrutting,respectively.Thetestdataandstatisticalanalysis,therefore,showthattheAPAissensitivetomixgradation.

ComparisonofRutDepthsofMixeswithPG64-22andPG58-22Binder

PairedttestswereconductedtocomparerutdepthsofmixeswithPG64-22(testedat64/C)and

PG58-22(testedat58/C)asphaltbinder.Table7showsatableofaveragerutdepthsforeachmix;mixwithPG64-22binderpairedagainstsamemixwithPG58-22binder.Sincetherewerethreeaggregates,threegradations,andtwocourses,thereare18pairsofdata.

Resultsofpairedttests(Table8)showthatatasignificantlevelof5%,thereisasignificant

differencebetweenrutdepthsofmixeswithPG64-22andPG58-22asphaltbinder.RutdepthsofmixeswithPG58-22asphaltbinder(testedat58oC)arehigherthanmixeswithPG64-22asphaltbinder(testedat64oC).Pairedttestswerealsodonewithmixesofwearingandbindercoursesseparatelyandmixescontainingdifferentaggregates.OneofthepossiblereasonsforgreaterrutdepthsformixeswithPG58-22asphaltisrelativelylowerG*/sin*valueofPG58-22asphaltbindercomparedtotheG*/sin*valueforPG64-22asphalt.Thedynamicshearrheometer(DSR)stiffness(RTFOTcondition)forthePG58-22binderat58oCis2.9kPa,

7



Kandhal&MallickwhereastheDSRstiffnessforthePG64-22binderat64oCis3.2kPa(Table3).Thetestdataandthestatisticalanalysis,therefore,indicatesthattheAPAissensitivetobindertype.

Table7.RutDataforMixeswithPG64-22andPG58-22AsphaltBinder

Mix*

WARZGRN

WTRZGRN

WBRZGRN

WARZLMS

WTRZLMS

WBRZLMS

WARZGRV

WTRZGRVWBRZGRV

BARZGRN

BTRZGRN

BBRZGRN

BARZLMS

BTRZLMS

BBRZLMS

BARZGRV

BTRZGRVBBRZGRV

PG64-22

4.48

4.31

5.35

3.77

3.91

6.24

6.46

5.775.64

3.48

1.62

3.43

4.07

3.98

5.62

5.19

4.354.53

PG58-22

6.59

3.81

6.02

4.53

5.47

7.16

7.95

6.035.24

3.40

2.80

2.85

4.00

5.04

9.49

6.41

5.234.65

Note: *Firstletterindicatescourse:Nextthreelettersindicategradation:

Lastthreelettersindicateaggregate:

W-Wearing,B-BinderARZ-AboveRestrictedZoneTRZ-ThroughRestrictedZoneBRZ-BelowRestrictedZoneGRN=Granite,LMS-Limestone,GRV-Gravel

Table8.ResultsoftTestsforComparingMixeswithPG64-22andPG58-22Binders

Comparison Mean* StandardError T Probability>T

Consideringallmixes -0.804 0.255 -3.149 0.0059Considering All -0.763 0.294 -2.599 0.0317

wearingcoursesGranite -0.760 0.755 -1.007 0.420

Limestone -1.080 0.244 -4.419 0.048

Gravel -0.450 0.554 -0.812 0.502

Consideringbindercourses -0.844 0.436 -1.936 0.089onlyNote:*=(rutdepthofmixeswithPG64-22asphaltbinder-rutdepthofmixeswithPG58-22)

8



Kandhal&MallickCorrelationofRutDepthswithDensityatN initialandNmaxNinitialandNmaxcriteriahavebeenspecifiedbySuperpaveinordertoavoidtendermixesandmixespronetorutting,respectively.Thedatawasanalyzedtodetermineifrutdepthsarelower(orlowest)whenthemixmetdensity#89%ofTMD(theoreticalmaximumdensity)criteriaatNinitialanddensity#98%ofTMDcriteriaatNmax.AnanalysisofvariancewasconductedtoobserveanysignificanteffectofdifferencebetweendensityatNintialand89(x=89-densityatNintial),anddifferencebetweendensityatNmaxand98(y=98-densityatNmax),onrutting.ThecalculatedxandyvaluesareshowninTable9.Theanalysisindicatednosignificanteffectofxandyonrutdepths(Table10).

Alloftheyvalueswerepositivenumbers,whichindicatesthatnoneofthemixeshaddensity

higherthan98%GmmatNmax.Thewearingcoursewithgranitehastwomixeswithnegativexvalues(densityatNinitialhigherthan89%).Therutdepthversusxandyvaluesshownoapparentcorrelationbetweenx,y,andrutdepth.However,observationofwearingcourseofgraveldoessuggestsomeeffectofxonrutdepth.Thisdatawaspooledwiththebindercoursegraveldatato

runaregressionbetweenrutdepthandxandy(Table11).However,nosignificantmodelwasobserved.

Inmostcases,exceptforbinderlimestoneitdoesseemthatcomparedtotherutdepthfora

densitylessthan89%ofTMDatNinitial,therutdepthstendtobehigherforthosecasesinwhichthedensityishigherthan89%ofTMD(Table12).However,thedatadoesnotsuggestthatamixwillhavethelowestrutdepthwhenitmeetstheNinitialcriteria,comparedtomixeswhichdonotmeetNinitialcriteria.OneobservationisthatinthosecasesinwhichthemixeswhichmeettheNinitialcriteriabuthavemaximumrutdepth(foraparticularaggregate),thedifferencebetweenthedensityatNmaxand98%(y)isobservedtobyverysmall.TheexceptionsareWearing-Gravel-BRZ,Binder-Granite-BRZandBinder-Gravel-BRZ(Table13).However,inthecaseoftheexceptions,thedifference

betweenthedensityatNmaxand98%ofTMDarehigher.Thedataindicatesthatifthedensityiswithin0.1-0.2%of98%ofGmmatNmax,onemightexpectrelativelyhigheramountofrutting.

9



Kandhal&Mallick

Table9.CalculatedxandyValues

Asphalt Mix* Density Density Rut (x=89-density y(y=98-densityatNinitial atNmax Depthx atNinitial) atNmaximum)

PG64-22 WARZGRN 89.72 97.14 4.48 -0.72 0.86

WTRZGRN 89.05 97.32 4.31 -0.05 0.68WBRZGRN 87.34 97.59 5.35 1.66 0.41

WARZLMS 88.58 97.34 3.77 0.42 0.66

WTRZLMS 87.13 97.71 3.91 1.87 0.29

WBRZLMS 85.95 97.86 6.24 3.05 0.14

WARZGRV 89.98 97.22 6.46 -0.98 0.78

WTRZGRV 89.37 97.36 5.77 -0.37 0.64

WBRZGRV 88.83 97.45 5.64 0.17 0.55

BARZGRN 89.95 97.17 3.48 -0.95 0.83

BTRZGRN 89.00 97.19 4.62 0 0.81BBRZGRN 87.46 97.45 3.43 1.54 0.55

BARZLMS 88.42 97.42 4.07 0.58 0.58

BTRZLMS 90.60 97.08 3.98 -1.60 0.92

BBRZLMS 85.81 97.83 5.62 3.19 0.17

BARZGRV 90.16 96.91 5.19 -1.16 1.09

BTRZGRV 89.46 97.21 4.35 -0.46 0.79

BBRZGRV 87.72 97.46 4.53 1.28 0.54

PG58-22 WARZGRN 89.72 97.14 6.59 -0.72 0.86

WTRZGRN 89.05 97.32 3.81 -0.05 0.68

WBRZGRN 87.34 97.59 6.02 1.66 0.41

WARZLMS 88.58 97.34 4.53 0.42 0.66

WTRZLMS 87.13 97.71 5.47 1.87 0.29

WBRZLMS 85.95 97.86 7.16 3.05 0.14

WARZGRV 89.98 97.22 7.95 -0.98 0.78

WTRZGRV 89.37 97.36 6.03 -0.37 0.64

WBRZGRV 88.83 97.45 5.24 0.17 0.55

BARZGRN 89.95 97.17 3.40 -0.95 0.83

BTRZGRN 89.00 97.19 2.80 0.00 0.81

BBRZGRN 87.46 97.45 2.85 1.54 0.55BARZLMS 88.42 97.42 4.00 0.58 0.58

BTRZLMS 90.60 97.08 5.04 -1.60 0.92

BBRZLMS 85.81 97.83 9.49 3.19 0.17

BARZGRV 90.16 96.91 6.41 -1.16 1.09

BTRZGRV 89.46 97.21 5.23 -0.46 0.79

BBRZGRV 87.72 97.46 4.65 1.28 0.54

Note: *Firstletterindicatescourse:Nextthreelettersindicategradation:


W-Wearing,B-BinderARZ-AboveRestrictedZoneTRZ-ThroughRestrictedZoneBRZ-BelowRestrictedZone

GRN=Granite,LMS-Limestone,GRV-Gravel

10



Kandhal&Mallick

Table10.AnalysisofVarianceforRutDepthsVersusxandy

PG64-22

AsphaltPG58-22Asphalt

Source

Model

Error

CTotal

Source

Model

Error

DF

2

15

17

2

15

17

MeanSquare

2.58

1.257

3.227

30.88

FValue

2.055

1.045

Probability>F

0.163

0.376

Note: x=(89-DensityatNinitial)y=(98-DensityatNmaximum)Table11.AnalysisofVarianceforRutDepthsandxandyforGravelMixes

PG64-22Asphalt

SourceModel

Error

CTotal

DF2

3

5

MeanSquare0.935

0.439

FValue2.132

Probability>F0.265

Note: x=(89-DensityatNinitial)y=(98-DensityatNmaximum)Model:Response=TrueMean+Effectofx+Effectofy+EffectofExperimentalUnit

Table12.RutDepthandxValues

Mix*

WTRZGRNWARZGRN

WARZGRV

WTRZGRV

WBRZGRV

BTRZGRN

BARZGRN

BTRZGRV

BARZGRV

x

-0.05-0.72

0.17

-0.37

-0.98

0.00

-0.95

-0.46

-1.16

RutDepth(mm)

0.680.86

5.64

5.77

6.46

1.62

3.48

4.35

5.19Note::

x=(89-DensityatNinitial)*Firstletterindicatescourse:Nextthreelettersindicategradation:


W-Wearing,B-BinderARZ-AboveRestrictedZoneTRZ-ThroughRestrictedZoneBRZ-BelowRestrictedZoneGRN=Granite,LMS-Limestone,GRV-Gravel

11



Kandhal&Mallick

Table13.RutDepthsandyValues

Mix

WBRZGRNWBRZLMS

exception:WBRZGRV

exception:BBRZGRN

BBRZLMS

exception:BBRZGRV

MeetsNintialCriteria?

yes(onlyone)yes(allmeet)

yes(onlyone)

yes(2meet)

yes(2meet)

yes(onlyone)

RutDepth

6.02(2ndhighest)7.16(highest)

5.24(lowest)

2.85(2ndhighest)

9.49(highest)

4.65(lowest)

y

0.14(lowestofallthree)0.14(lowest)

0.55(lowest)

0.55(lowest)

0.17(lowest)

0.54(lowest)

Note:y=(98-DensityatNmaximum)

EffectofAsphaltBinderFilmThicknessonRuttingRegressionanalysesweredonetoobserveanypossiblerelationbetweenfilmthicknessandrutting.Inthefirststep,onlywearingcoursesofgraniteandlimestone(forPG64-22andPG58-22)wereconsidered.Thegravelmixeswerenotincludedsinceobservationofthedata(Table14)showedthatwhilegraniteandlimestonemixestendtohavemoreruttingwithanincreaseinfilmthickness,forgraveltheruttingdecreasedwithanincreaseinfilmthickness.

Table14.FilmThicknessandRutDepthsforDifferentMixes(withPG64-22

AsphaltBinder)

Course Aggregate Gradation FilmThicknessRutDepth(mm)

(micron)Wearing Granite ARZ 8.70 4.48

TRZ 9.36 4.31

BRZ 10.58 5.35

Limestone ARZ 6.96 3.77

TRZ 8.09 3.91

BRZ 11.01 6.24

Gravel ARZ 7.83 6.46

TRZ 8.47 5.77

BRZ 10.14 5.64

Binder Granite ARZ 9.74 3.48

TRZ 10.41 1.62

BRZ 12.92 3.43

Limestone ARZ 8.80 5.47

TRZ 10.41 3.98

BRZ 15.18 5.62

Gravel ARZ 9.27 8.04

TRZ 10.14 4.35

BRZ 12.6 4.53

12



Kandhal&Mallick

Thebestrelationwasobtainedbetweensquareof 2filmthicknessandrutdepthis:

Rutdepth=2.53+0.035(filmthickness).

Hence,forarutdepthof7mm,onewouldexpectafilmthicknessof11.9 .12:m.Prob>Fofmodel=0.0084Prob>*t* forintercept=0.0125

fo2r(filmthickness)2=0.0084

R=0.52

Forbindercourseswithgraniteandlimestone,thebestrelationwasfoundtobe:

Rutdepth=37.05-6.137(filmthickness)+0.2754(filmthickness)2Prob>Fofmodel=0.0108Prob>*t* ofintercept=0.0321

offilmthickness=20.0363

of 2(filmthickness)=0.0256R=0.63

Whilethevalidityofthissomewhatcomplexregressionequationisdebatable,itdoesindicate

thatforbindercoursesruttingmayactuallydecreasewithanincreaseinfilmthickness.However,theapplicabilityoffilmthicknessconcepttocoursesotherthanthewearingcourseisquestionable.

Incaseofwearinggravelcourses,thebestrelationwasobtainedas:

Rutdepth=19.39-14.017log 10filmthicknessProb>F=0.0832(notsignificantat"=5%)Prob>*t*=0.0281

ofintercept

of 2log10filmthickness=0.0832

R=0.63

Thisindicatesthatrutdepthdecreaseswithanincreaseinfilmthickness.

Forbindercourseforgravel,nosignificantmodelwasfoundbetweenrutdepthandfilm

thickness.

Thedifferenceintheeffectoffilmthicknessonrutdepthforgraniteandlimestone,andgravel

indicatesadifferenceinthewaytheaggregatesandasphaltbinderarepackedtogetherinamix.Oneexplanationisthatinthecaseofrelativelyroundedandsmoothtexturedgravelparticles,increasedfilmthicknesshelpsinlubricationofparticlesduringcompaction,bringsthemcloser(lowVMA)andthushelpsinmakingatightlyinterlockedstructure.Ontheotherhand,inthecaseofrelativelyangularandroughtexturedgraniteandlimestonepresenceoftoomuchasphaltfilmtendtomovetheparticlesapartandbreakthetightlyinterlockedaggregatestructure.

Sincethefinematerials,particularlymaterialpassing0.15mmand0.075mmsievemayactually

beembeddedinasphaltmatrixandnotprovidesurfaceareaforcoating,filmthicknesswasalsocalculatedbyneglectingthesurfaceareaofmaterialpassing0.15mmand0.075mmsieves.However,noimprovementinthemodelbetweenfilmthicknessandrutdepthwasobtained.Inthenextstepfilmthicknesswascalculatedbyneglectingmaterialpassing0.075mmsieveonly,butconsideringmaterialpassing0.15mmsieve.Again,nosignificantimprovementwasobtained.

13



Kandhal&MallickComparisonofRutDevelopmentSlopewithGyratorySlopeEachplotofpassesversusruttingresultingfromtestswithAPAconsistsofthreelineswithdifferentslopes,between0-1000,1000-4000,4000-8000passes.Thenatureoftherutdevelopmentcurveisverysimilartothenatureofthegyratorycompactioncurve.Therutdevelopmentplotgraduallyappearstoleveloff(andhavealowerslope)justlikethegyratorycompactionplot,inwhichthedensityappearstoleveloffbeyondNdesign.Hence,itwasdecidedtoexamineslopeofeachpartoftheplotandtrytocorrelatewithslopeofgyratoryplot.

Sincetheslopeofgyratoryplotisfromlogofgyrationversusdensity,itwasdecidedtousethe

logofpassversusruttingplotfordeterminingtheslopeofrutdevelopmentplot.Also,toconsiderinitialzerorutting,theinitialpassnumberwaschangedfromzeroto10.

Hence,threeslopesweredeterminedforeachrutdevelopmentplot;slopesbetween0-1000,

1000-4000,and4000-8000.Eachoftheseslopeswerecorrelatedwithgyratorycompactionslope(Table15).

Bothlogandnormalslopesdidnotshowanycorrelationwithgyratorycompactionslope

(betweenNdesignandNinitial)(Table16).

Table15.GyratoryCompactionSlopeandRutDepthforDifferentMixes(WithPG64-22

AsphaltBinder)Course Aggregate Gradation Gyratory

CompactionSlope

(between

Log

RutDevelopmentSlope

Normal

Ninitialand 0-1000 1000- 4000- 0- 1000- 4000-Ndesign) 4000 8000 1000 4000 8000

Wearing Granite ARZ 6.066 1.014 3.074 2.015 2.028 0.617 0.152

TRZ 6.761 1.382 1.845 1.428 2.764 0.370 0.108BRZ 8.389 1.725 2.400 1.503 3.451 0.482 0.113

Limestone ARZ 7.16 0.964 1.954 2.198 1.928 0.392 0.165

TRZ 8.653 0.936 1.412 3.931 1.872 0.283 0.296

BRZ 9.737 1.860 2.519 3.316 3.720 0.506 0.250

Gravel ARZ 5.918 2.160 2.526 2.068 4.319 0.507 0.156

TRZ 6.531 1.779 2.317 2.710 3.558 0.465 0.204

BRZ 7.055 1.705 1.924 3.491 3.431 0.386 0.263

Binder Granite ARZ 5.907 0.825 1.89 2.290 1.650 0.379 0.173

TRZ 6.701 0.485 0.615 0.914 0.970 0.123 0.069

BRZ 8.175 0.923 1.362 2.552 1.845 0.273 0.192Limestone ARZ 7.355 0.973 2.383 6.943 1.945 0.478 0.173

TRZ 5.299 0.793 2.356 3.233 1.585 0.473 0.243

BRZ 9.83 1.738 2.223 2.669 3.475 0.446 0.201

Gravel ARZ 5.518 1.634 2.234 1.905 3.268 0.448 0.143

TRZ 6.332 1.183 2.029 2.525 2.367 0.407 0.190

BRZ 7.963 1.205 2.323 2.392 2.410 0.466 0.180

14



Kandhal&MallickTable16.RegressionEquationsforRutDevelopmentSlopeVersusGyratoryCompaction

SlopeforDifferentMixes(WithPG64-22AsphaltBinder)

RutDevelopmentSlope Course Modela R2

Log 0-1000 Wearing y=-0.0063x+1.55 0.00031000-4000 y=-0.1057x+2.99 0.08

4000-8000 y=0.3185x+0.1725 0.21

Normal 0-1000 y=-0.0125x+3.1001 0.0003

1000-4000 y=-0.0212x+0.6016 0.08

4000-8000 y=0.024x+0.013 0.21

Log 0-1000 Binder y=0.105x+0.3481 0.15

1000-4000 y=-0.0061x+1.9775 0.0002

4000-8000 y=-0.6841x+8.6716 0.09

Normal 0-1000 y=0.21x+0.6962 0.15

1000-4000 y=-0.0012x+0.3969 0.0002

4000-8000 y=0.003x+0.1524 0.009a"x"isslopeofgyratorycompactioncurve

VoidsinMineralAggregates(VMA)versusRutDepthRutdepthdataandVMAdataofthedifferentmixesareshowninFigure1.Ingeneral,forgraniteandlimestone,thereisanincreaseinrutdepthwithanincreaseinVMA.Incaseofgravel,thetrendisreverse-thereisadecreaseinrutdepthwithanincreaseinVMA.Atthistimethedifferenceinbehaviorcannotbeexplained.

VMAversusRutting

PG64-22,WearingCourse

7.00

6.005.004.00

3.00

2.001.000.00

4.49

4.31

5.35

3.77

3.91

6.24

6.46

5.77

5.64

15.4215.2415.99 12.3 12.4914.6814.4214.0314.88

VMA(%)

Figure1.PlotofVMAversusRuttingforPG64-22,WearingCourseMixes

15

R

u t t i n

m m



Kandhal&MallickComparisonofresultsfromtestswithSuperpaveShearTester(SST)andAPA Table17showstheresultsoftestswithRSCH.TheaveragepeakstrainvaluesshowthataccordingtotheSSTtest,forwearingcourse,theTRZmixesshowthelowestruttingpotential.

Table17.RSCHPeakShearStrainforDifferentMixes

Course AggregateGradation Strain AverageStrain

Sample1 Sample2

Wearing Granite ARZ 0.02676 0.01795 0.022355

TRZ 0.01892 0.0251 0.02201

BRZ 0.02294 0.02614 0.02454

Limestone ARZ 0.03824 0.03437 0.036305

TRZ 0.00954 0.0291 0.01932

BRZ 0.0511 -- 0.0511

Gravel ARZ 0.07194 -- 0.07194

TRZ 0.04932 0.05166 0.05049

BRZ 0.05049 0.08057 0.06553

Binder Granite ARZ 0.0064 0.02084 0.01362

TRZ 0.01269 0.02632 0.019505

BRZ 0.0144 0.02322 0.01881

Limestone ARZ 0.0405 0.02379 0.032145

TRZ 0.03399 0.0445 0.039245

BRZ 0.04854 0.07685 0.062695

Gravel ARZ 0.07154 0.06071 0.066125

TRZ 0.03779 -- 0.03779

BRZ 0.03634 0.07214 0.05424

Figure2showsacomparisonofresultsfromRSCHandAPAtest.Thedatashowsafaircorrelation(R2=0.62),whichindicatesthattheRSCHandtheAPAruttestshavecharacterized

the2mixesinthesameway.Thebindercoursedata(Figure3)showsaslightlybettercorrelation

(R=0.69).

Table18showstheresultsfromtestswithRSCSR.Thepeakshearstrainvaluesindicatethat

TRZmixesarenotalwaystheoneswiththeminimumruttingpotential—infact,inthecaseofgranitewearingcourseTRZmixshowsthehighestpeakstrain.Figures4and5show

comparisonofresultfromRSCSRamd2APAtest,forwearingandbindercourses,respectively.

Bothcorrelationsarerelativelypoor(R=0.55,0.44,respectively)comparedtoRSCH,indicatingthattheRSCSRtestdoesnotcomparewellwiththeAPAtest.

16



Kandhal&Mallick

7

6

5

4

3

WearingCourse,PG64-22Binder

y=39.761x+3.4937

2

1

0

0

0.02

0.04

R2=0.6235

0.06

0.08

PeakShearStraininRSCHTest

Figure2.PlotofPeakShearStraininRSCHTestversusRutDepthinAPAfor

WearingCoursewithPG64-22Binder

6

5

4

3

2

1

0

BinderCourse,PG64-22Binder

y=49.637x+2.1318

R2=0.6918

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

PeakShearStraininRSCHTest

Figure3.PlotofPeakShearStraininRSCHTestversusRutDepthinAPAforBinder

CourseWithPG64-22Binder

17

R u t D e

t h

m m

i n A P A

R u t D e

t h

m m

i n



Kandhal&Mallick

Table18.RSCSRPeakShearStrainforDifferentMixes

Course Aggregate Gradation Strain AverageStrain

Sample1 Sample2

Wearing Granite ARZ 0.0288 0.03814 0.03347

TRZ 0.03417 -- 0.03417

BRZ 0.02183 0.03188 0.026855

Limestone ARZ 0.02504 0.0429 0.03397

TRZ 0.0309 0.05407 0.042485

BRZ 0.04453 0.07859 0.06156

Gravel ARZ -- 0.08948 0.08948

TRZ 0.03893 0.08232 0.060625

BRZ -- 0.08457 0.08457

Binder Granite ARZ 0.01531 0.02651 0.02091

TRZ 0.01966 0.02218 0.02092

BRZ 0.0168 0.01761 0.017205

Limestone ARZ 0.02908 0.04326 0.03617

TRZ 0.02537 0.04491 0.03514

BRZ 0.04772 0.07686 0.06229

Gravel ARZ 0.03323 0.03655 0.03489

TRZ 0.02024 0.01649 0.018365

BRZ 0.03062 0.07929 0.054955

7

6

5

4

3

WearingCourse,PG64-22Binder

y=32.256x+3.42562

1

0

0

0.02

0.04

0.06

R2=0.5533

0.08

0.1

PeakShearStraininRSCSRTest

Figure4.PlotofPeakShearStraininRSCSRTestversusRutDepthinAPAfor

WearingCoursewithPG64-22Binder

18

R u t D e

t h

m m

i n A P A



Kandhal&Mallick

6

5

4

3

2

1

0

BinderCourse,PG64-22Binder

y=47.493x+2.4424

R2=0.4452

0 0.01 0.02 0.03 0.04 0.05 0.06 0.07

PeakShearStraininRSCSRTest

Figure5.PlotofPeakShearStraininRSCSRTestversusRutDepthinAPAfor

BinderCoursewithPG64-22Binder

ComparisonofIn-PlaceRuttingandResultsFromTestsWithAPA

Thepropertiesofthein-placemixesfromI-85areshowninTable19.Table20showstherutdepths,asobtainedfromthetestswiththeAPA,andthein-placerutdepthsforeachmix.Thegoodperformingmix(A)showstheleastamountofruttingfromtestswithAPA.However,thepoorperformingmix(C)showsslightlylessAPAruttingcomparedtothefair(B)performingmix.Thisdiscrepancymayhaveresultedduetothefollowingreasons:(a)althoughallthreeHMAsectionsareonthesameinterstateI-85,theywereplacedindifferentyearsand,therefore,haveagedtodifferentdegrees;(b)thesectionshavebeensubjectedtodifferentamountsofESALs;and(c)someruttingmayhavebeencontributedbytheunderlyingHMAcourseswhichwerenottestedbytheAPT.

19

R u t D e

t h

m m

i n A P A



Kandhal&Mallick

Table19.PropertiesofMixesA,B,andC(In-Place)

Property Mix

VoidsinTotalMix(%)

AsphaltContent

TMD

Gradation

%Passing

25mm

19.5mm

12.5mm

9.5mm

4.75mm

2.36mm

1.18mm

0.600mm

0.300mm

0.150mm

0.075mm

A(Good)

5.61

5

2.493

100

85.9

75.2

60.5

44.1

33.5

23.3

13.9

7.9

4.6

B(Fair)

4.38

5.6

2.452

100

98.6

87.5

77.9

63.4

52.6

44.7

29.9

15.3

7.9

4.8

C(Poor)

3.08

6

2.454

100

86.3

75.8

61.3

49.7

41.9

28.4

15.3

8.4

5.2

Table20.In-PlaceRuttingandResultsfromTestswithAPAMix RuttingwithAPA(mm)

Samples

1 2 3 4 5 6 Average In-PlaceRutting(mm)

A 1.66 0.87 1.42 1.28 1.54 1.19 1.33 0.00

B 6.23 5.45 6.43 6.00 5.86 4.75 5.79 6.00

C 4.09 4.51 6.7 4.95 3.44 3.34 4.50 12.5

Thereisinsufficientdatainthisstudytoestablisharutdepthcriteria.However,basedontheborderlineperformanceofSectionsBandCandspecificationsusedbysomeDOTs,atentativecriteriaof4.5-5.0mmrutdepthafter8,000cyclesappearsreasonable.However,morefieldsectionsshouldbetestedtoconfirmthiscriteria.

20



Kandhal&MallickCONCLUSIONSThefollowingconclusionscanbedrawnfromthisstudy:

1.Theasphaltpavementanalyzer(APA)issensitivetoaggregategradationbasedonstatisticalsignificanceofdifferencesinrutdepths.Incaseofgraniteandlimestonemixesthegradationbelowtherestrictedzoneshowedhighestamountofruttingwhereasthegradationthroughtherestrictedzonegenerallyshowedlowestrutdepth.However,incaseofgravelmixes,thegradationbelowtherestrictedzoneshowedtheleastamountofruttingwhereasthegradationabovethezoneshowedhighestamountofrutting.

2.TheAPAwasalsofoundtobesensitivetotheasphaltbinderPGgradebasedonstatisticalsignificanceofdifferencesinrutdepths.TherutdepthsofmixeswithPG58-22asphaltbinder(testedat58°C)werehigherthanthoseofmixeswithPG64-22asphaltbinder(testedat64°C).ThisresultedfromrelativelylowerG*/sin*valueofPG58-22comparedtoG*/sin*ofPG64-22.

3.MixesmeetingNinitialandNmaxcriteriadidnotnecessarilyshowlessruttingpotential

thanmixeswhichdidnotmeetthesecriteria.4.NocorrelationcouldbeestablishedbetweenAPArutdepthsandthegyratorycompactionslopes(betweenNinitialandNdesign)ofallmixes.

5.Incaseofgraniteandlimestonewearingcoursemixes,theAPArutdepthincreasedwithanincreaseinasphaltfilmthickness.However,anoppositeeffectwasobservedincaseofgravelwearingcoursemixes,andbindercoursemixescontaininggraniteandlimestone.

6.TheAPAhadafaircorrelation(R2=0.62)withtherepeatedshearconstantheight(RSCH)testconductedwiththeSuperpavesheartester.Bothtestscharacterizedthemixesinthesameway.

7.ItappearsfromthisstudythattheAPAhasapotentialtopredicttherelativeruttingpotentialofhotmixasphaltmixes.

8.Basedonverylimiteddata,itappearsthattheAPArutdepthafter8000passesshould

belessthan4.5-5.0mmtominimizeruttinginthefield.However,morefieldtestsectionsneedtobeevaluatedtoestablishthiscriteria.

21



Kandhal&MallickREFERENCES1.

2.3.4.5.6.

7.8.

Lai,J.S.EvaluationofRuttingCharacteristicsofAsphaltmixesUsingLoaded-WheelTester.FinalReport,GaDOTProjectNo.8609,August1986.Lai,J.S.EvaluationoftheEffectofGradationofAggregateonRuttingCharacteristicsofAsphaltMixes.FinalReport,GaDOTProjectNo.8706,August1988.West,R.D.,G.C.Page,andK.H.Murphy.EvaluationoftheLoadedWheelTester. ResearchReportFL/DOT/SMO/91-391.StateofFloridaDept.ofTransportationDecember1991.Miller,T.,K.Ksaibati,andM.Farrar.UtilizingtheGeorgiaLoaded-WheelTestertoPredictRutting.Paperpresentedatthe73rdAnnualMeetingoftheTransportationResearchBoard,1995.Collins,R.,D.E.Watson,andB.Campbell.DevelopmentandUseoftheGeorgiaLoadedWheelTester.InTransportationResearchRecord 1492,TRB,NationalResearchCouncil,Washington,DC,1995,pp.202-207.Collins,R.,H.Shami,andJ.S.Lai.UseofGeorgiaLoadedWheelTestertoEvaluate

RuttingofAsphaltSamplesPreparedbySuperpaveGyratoryCompactor.InTransportationResearchRecord 1545,TRB,NationalResearchCouncil,Washington,DC,1996,pp.161-168.Shami,H.I.,J.S.Lai,J.A.D'Angelo,andT.P.Harmon.DevelopmentofTemperatureEffectModelforPredictingRuttingofAsphaltMixturesUsingGeorgiaLoadedWheelTester.InTransportationResearchRecord 1590,TRB,NationalResearchCouncil,Washington,DC,1997,pp.17-22.MissouriDepartmentofTransportation.RuttingSusceptibilityofBituminousMixturesbytheGeorgiaLoadedWheelTester .ReportNo.RDT98-01.May1,1998.

22

Documents

Potential of Asphalt Pavement Analyzer to Predict Rutting of Hot Mix Asphalt