191Applied RheologyVolume 13 Issue 4Viscoelastic
Characterization of Polymer-Modified AsphaltBinders of Pavement
ApplicationsWaheed UddinThe University of Mississippi, Carrier 203,
University, MS 38677-1848, USAEmail: [email protected]:
001.662.915.5523Received: 20.2.2003, Final version:
20.6.2003Abstract:Rutting is a primary reason of premature
deterioration of asphalt highway pavements. Pavements
constructedwith polymer and other modifiers are showing improved
performance. The virgin asphalt and modified asphaltbinders and
mixes used on several test sections of the I-55 highway
rehabilitation project in northern Missis-sippi are compared. The
laboratory creep compliance data for these binders were measured at
low temperaturesusing a modified test procedure adapted for the
Bending Beam Rheometer device. Dynamic Shear Rheometerwas used at
high service temperatures. The creep compliance data of the binder
was used as an input to simu-late creep compliance behavior of the
mix using a micromechanical model. The field evaluation confirms
therelatively poor performance of the virgin asphalt section with
respect to rutting, compared to modified
bindersections.Zusammenfassung:Fahrrinnenbildung ist ein Hauptgrund
fr die vorzeitige Alterung von Asphaltbelgen auf Autobahnen.
Belgemit Polymeren und anderen Modifikatoren zeigen eine
verbesserte Gte. Der unbehandelte Asphalt und modi-fizierte
Zusatzstoffe und Mischungen, welche auf verschiedenen
Testabschnitten des Sanierungsprojektes aufder I-55 im Norden
Mississippis benutzt wurden, werden verglichen. Die im Labor
ermittelten Daten der Kriech-nachgiebigkeit fr diese Zusatzstoffe
wurden bei niedrigen Temperaturen gemessen, wobei eine
modifiziertesTestverfahrenbenutztwurde,welchesandasBiegebalkenRheometerangepasstist.DasdynamischeScher-rheometer
wurde bei hohen Temperaturen eingesetzt. Die
Kriechnachgiebigkeitsdaten der Zusatzstoffe
wur-denalsEingabedatenzurSimulationdesKriechnachgiebigkeitsverhaltensderMischungbenutzt,wobeieinmikromechanisches
Modell herangezogen wurde. Die Auswertung der Feldstudie belegt die
relativ geringe Gtedes Primrasphaltes in Bezug auf die Bildung von
Fahrrinnen, verglichen mit den Abschnitten, welche mit
modi-fiziertem Bindemittel asphaltiert wurden.Rsum:La premire cause
de dtrioration prmature des revtements asphaltiques des autoroutes
est la
formationdornires.Lesrevtementsobtenusavecdupolymreetdautresadjuvantsprsententdesperformancesaccrues.
Lasphalte vierge, lasphalte modifi avec des liants ainsi que des
mlanges utiliss sur plusieurs sec-tions tests de lautoroute I-55
dans le cadre dun projet de rhabilitation de cette autoroute du
nord du Missis-sipi, ont t compars. Des donnes de relaxation en
fluage ont t obtenues en laboratoire pour ces mlanges basses
tempratures, laide dun rhomtre de fluage dont le protocole
exprimental a du tre modifi. Unrhomtre de cisaillement dynamique a
t utilis pour lobtention de donnes hautes tempratures. Les don-nes
de compliance en fluage obtenues pour lasphalte modifi avec du
liant ont t utilises comme paramtresdans un modle micromcanique
pour simuler le comportement en fluage du mlange. Lvaluation faite
surleterrainconfirmelesperformancesrelativementpauvresdelasphalteviergecomparlasphaltemodifiavec
du liant, en ce qui concerne la formation dornires.Key Words:
asphalt, binder, polymer-asphalt, laboratory, pavement, Superpave
Appl. Rheol. 13 (2003) 191-199Applied Rheology
Vol.13/4.qxd09.09.200311:53 UhrSeite 191192Applied RheologyVolume
13 Issue 41 INTRODUCTIONAsphalt is predominantly used to construct
pave-mentsforroads,highways,andairports.Bothasphaltbinderandasphalt-aggregatemixtureshow
temperature- and time-dependent
behav-ior.Ruttingorpermanentdeformationistheleading cause of
pavement deterioration in
tem-perateandwarmclimaticregionsoftheworldandlow-temperaturecrackingisacommonproblemincoldregions[1-4].Asthehighwaypavementinfrastructurehasagedanddeterio-rated,morepavementsareinneedofmainte-nance,
rehabilitation, and reconstruction. This
istraditionallydonebyplacinghot-mixasphaltoverlay.TheMississippiDepartmentofTrans-portation(MDOT)hasadoptedtheSuperpavePerformance
Grade (PG) binder specification andmix design system [1], developed
by the
Strate-gicHighwayResearchProgram(SHRP),thatrelatestheasphaltmixdesigntoperformance.Polymer-modifiedasphalthasbeenclaimedtoresist
rutting and has been used in a side-by-sideexperimental study in
Mississippi. This comprehensive study included lab-oratory testing
for creep compliance characteri-zation of asphalt layers,
deflection testing, in situpavementperformanceeval uati on,
andadvancedfiniteelementcomputersimulationsof an in-service asphalt
highway pavement on I-55 highway near Grenada in northern
Mississip-pi[1-2].TheanalysisandresultsoflaboratorybindertestswereconductedusingSuperpaveequipment
for different asphalt binders used inthis study. The objectives of
this paper are to: syn-thesize the laboratory test results of
binders andmixes, compare the laboratory test results
withthefieldperformanceintermsofresistancetorutting, and evaluate
the effect of measured vis-coelastic properties of the asphalt
layer on finiteelementdynamicresponseanalysisofthecon-structed
asphalt pavement.The pavement test sections were pavedin 1996 with
virgin asphalt and eight other mod-ified binders including six
polymers, two rubbers,and one chemical. Based upon the MDOT
pave-ment design 76 mm thick (38 mm binder and
38mmsurface)asphaltlayerwasremovedandreplaced by the same thickness
of new modifiedmix in each test section. The control and
transi-tionsectionswereconstructedusingAC-30grade,thestandardunmodifiedvirginasphaltused
by the MDOT in Mississippi [1]. The optimumbinder content of these
mixes, determined usingthe Superpave level I mix design method, was
5.2% by weight of total mix. Other details of the
con-structedsectionsareprovidedbyUddinandYamini [2].2 SUPERPAVE
BINDER TESTSThe Superpave binder PG specifications and
mixdesignarerecommendedbytheFederalHigh-wayAdministration(FHWA)forimprovedandlonglifehighwaypavements[5,6].LaboratorySuperpavebindertestswereconductedattheUniversityofMississippi[7]usingasphaltandmodifiedasphaltbindersamplesfromI-55testsections.
The PG binder for the test was
PG58-10(fasttraffic),where58-10representsthehigh(58C)andlow(-10C)pavementtemperaturesfor
northern Mississippi.2.1 DYNAMIC SHEAR RHEOMETER TESTSDynamic Shear
Rheometer (DSR) tests were con-ducted on all nine kinds of binders.
In the DSR testoperation,theasphaltsampleiscompressedbetween two
parallel plates, one of which is
fixedandtheotheroneoscillates,asshowninFig.1.The DSR test measures
the complex shear
mod-ulus,G*,andphaseangle,d,ofthebinder.Thecomplexshearmodulusisameasureoftotalresistance
of a material to deform when
exposedtorepeatedpulsesofshearstress.Thephaseangle is an indicator
of the relative amounts
oftherecoverableandnon-recoverabledeforma-tion. The DSR tests were
conducted at 46, 58,70, 82C for Novophalt and control asphalt
sam-ples.Thetestswerealsoconductedatseveraltemperatures 46, 52, 58,
64, 70, 76, 82C on Cry-opolymer. For the rest of the asphalt binder
sam-ples the SHRP Spec Tests was conducted at 46CAAACB1
cyclePosition of Oscillating PlateTimeAC BApplied Stressor
StrainOscillatingPlateAsphaltFixed PlateFigure 1:Dynamic
ShearRheometer (DSR) principle.Applied Rheology
Vol.13/4.qxd09.09.200311:53 UhrSeite 192193Applied RheologyVolume
13 Issue 4only because of time constraints. The DSR ruttingfactor,
G*/sind, represents a measure of the
hightemperaturestiffnessorruttingresistanceofasphalt binders. The
SHRP Superpave PG binderspecifies minimum values for G*/sind of
1000 Pafororiginalasphaltbinderand2200Paafteraging the binder using
the Rolling Thin Film Oven(RTFO) procedure [7, 8]. The G*/sind
results for all nine types ofbinders at 46C are shown in Fig. 2.
The results ofG*/sind
valueswerecomparedforcontrolasphalt,NovophaltandCryopolymerat46,58,70,
and 82C. At 46 and 58C Novophalt
performsbetterthaneithercontrolasphaltorCryopoly-mer. At higher
temperatures (70 and 82C), all
ofthethreebindersperforminthesamewayasshown in Fig. 3. The results
in Fig. 2 and Fig. 3 show thatNovophalt may offer better resistance
to ruttingat high service temperature. It is observed
that,withtheexceptionofUltrapave,themodifiedasphalt binders and the
control asphalt meet
theSHRPPGspecificationrequirements.Thisdoesnotsupportthepoorfieldperformanceofcon-trol
asphalt, compared to the modified binders.Therefore, it can be
concluded that the SHRP
SpecTest,originallydesignedonlyforvirginasphaltbindersmaynotbeapplicabletomodifiedasphalt
binders. The DSR test can be adopted toconduct creep compliance
tests of these
bindersathightemperaturesrepresentingin-serviceconditions in warm
climatic regions.2.2 BENDING BEAM RHEOMETER
TESTTheBendingBeamRheometer(BBR)isusedtoaccuratelyevaluatebinderpropertiesatlowtemperaturesatwhichasphaltbindersaretoostifftoreliablymeasurerheologicalpropertiesusingtheparallelplategeometryoftheDSRequipment.
Used together, the DSR and the
BBRtestsproviderheologicalbehaviorofasphaltbinders over a wide
range of in-service
tempera-turesincoldtowarmclimaticregions.Fig.4shows a close up of
the BBR equipment. The BBRtest measures creep stiffness value at
the lowestin-servicetemperature,S(t)asdesignatedbySHRP, which is
indicative of the susceptibility tolow-temperature cracking. To
prevent this
crack-ing,S(t)hasamaximumlimitof300MPa,asrequiredbytheSuperpavePGbinderspecifica-tions.Sincelow-temperaturecrackingoccursonly
after the pavement has been in-service
forsometime,thisspecificationaddressestheseproperties using binder
aged in both the RollingThin Film Oven (RTFO) and Pressure Aging
Vessel(PAV) devices.Therateofchangeofbinderstiffnesswith time is
represented by the m-value,
whichistheslopeofthelogstiffnessversuslogtimecurve from the BBR
test results, as shown in Fig.5a. A high m-value is desired because
as the
tem-peraturedecreasesandpavementcontractionbeginstooccur,thebinderwillrespondasamaterialthatislessstiff.Thisdecreaseinstiff-5741,7001,8901,5901,2402,3404,2001,5801,42001,0002,0003,0004,0005,0006,0001
2 3 4 5 6 7 8 9Test Temperature:
460CUltrapaveStyrelfNovophaltKratonCryopolymerSealoflexMultigradeControl
AsphaltRouse RubberG'/sind [Pa]Binder
Type1,0002,0003,0004,0005,00040 50 60 70 80 90NovophaltContr
AsphaltCryopolymerTest Temperatures: 46, 58, 70, 82 deg
CTemperature [C]G'/sind [Pa]ol1101001000Log Creep Stiffness S(t)8
30 60 120 240Slope = m-valueLog Loading Time t[s]Figure 2 (left
above):Comparison of G*/sindvalues for asphalt andmodified asphalt
bindersamples at 46C.Figure 3 (left below):Comparison of
G*/sindvalues for control asphalt,Novophalt, and Cryopoly-mer
binders at varioustemperatures.Figure 4 (right above):Bending Beam
Rheometer(BBR) equipment.Figure 5 a) (right below):Rate of change
of binderstiffness with time,represented by the m-value.Applied
Rheology Vol.13/4.qxd09.09.200311:53 UhrSeite 193194Applied
RheologyVolume 13 Issue 4ness leads to smaller tensile stresses in
the binderand less chance for low-temperature cracking.
Aminimumm-valueof0.3after60secondsisrequiredbytheSuperpavePGbinderspecifica-tion
[7]. As shown in Fig. 5b, all binders show high-er than the
required minimum m-value. The S(t)test results of all these binders
satisfy the SHRPSuperpave PG binder specifications [7]. Fig. 6
compares the S(t) test results at -12C. Excessively high stiffness
values, shown bysome polymer-asphalt binders (Novophalt,
Kra-ton,andMultigrade),indicatethattheasphaltmixproducedusingthesebindersmaybesus-ceptible
to low-temperature cracking, especiallyin the areas of cold
climate. On the other
hand,controlasphaltandothermodifiedasphaltbindersincludingRouseRubber,Sealoflex,andStyrelf
may perform better. 3 BINDERCREEPCOMPLIANCETESTSUSING
BBRAdditionallaboratorytestswereconductedusingBBRtoevaluatecreepcomplianceoftheselected
binders for longer duration of 3600 sec-onds. The creep compliance,
as a function of time,is defined as:(1)where e(t) is the
time-dependent strain under
aconstantstress,s.Thecreepcomplianceisthereciprocal of the
relaxation modulus,
E.AmodifiedtestprocedurefortheBBRequipmentwasadaptedforasphaltandmodi-C
tt( )( )= fied binders at different temperatures. The mod-ified
test was conducted by increasing the totaltest duration for a
longer time of 3600 seconds.The BBRAN computer program was
developed toanalyze and summarize the output generated bythe BBR
machine. Creep compliance and relaxation modu-lus values were
calculated for several binders at atemperature of -12C. Tests at
temperatures of -18,-12 and 0C were also conducted during this
study.It is apparent from this study that the binder
creepcompliance data at 0C may not be reliable usingthe BBR
equipment. Therefore, future binder
creepcompliancetestswillbeconductedat0Cusingthe DSR equipment. Fig.
7 shows the binder relax-ation modulus calculated from BBR tests at
-12Cand at 1 second. The relaxation modulus value
forSealoflexmodifiedbinderwassimilartothatofthe control asphalt
binder, an indication of betterperformance at low temperatures.
Fig.8plotforcontrolasphaltbindershowsanincreaseintherelaxationmoduluswhenthetesttemperatureisdecreased.Fig.9showsacomparisonoftherelaxationmodulusfor
control asphalt and modified binders at -12C Figure 5 b) (left
above):The m-value results at -12C.Figure 6 (left below):Comparison
of measuredS(t) values for asphaltbinder samples at -12C.Figure 7
(right above):Binder relaxation modulusfor control asphalt
andmodified binders at 1 secondfrom BBR tests at -12C.Figure 8
(right middle):Relaxation modulus forcontrol asphalt binder.Figure
9 (right below):Relaxation modulus forcontrol asphalt andmodified
binders at -12C.Applied Rheology Vol.13/4.qxd09.09.200311:53
UhrSeite 194195Applied RheologyVolume 13 Issue 4test temperature.
Fig. 9 shows that the
asphaltmodulusattenuationisrelativelymorebothinthe short-term and
long-term ranges at -12C testtemperature. This test temperature is
well belowthe low-temperature regime in northern
Missis-sippi.Thebindercreepcompliancetestresultsand elastic
properties of aggregate were used topredict creep compliance of the
compacted mixusing the micromechanical analysis approach [9,10], as
discussed in the following section.4 MICROMECHANICAL MODEL
TOPREDICT MIX CREEP COMPLIANCEThe micromechanics approach for
material mod-eling examines the interaction of the
constituentmaterials directly on a microscopic scale [9 - 11].The
micromechanics approach used in this
studyisbasedupontheapplicationofAboudismethodofcellsmicromechanicalanalysistopredict
viscoelastic response of resin matrix com-posites [9, 12]. The
coupling of the method of
cellswiththetime-steppingalgorithmprovidesanaccuratemodelfortheanalysisofcompositematerials
displaying viscoelastic behavior. The most attractive advantage of
usingthemicromechanicsapproach,incorporatedinthe ASPHALT program
[10], is that overall
prop-ertiesofthemixmaybedeterminedfromtheknown properties of the
constituents. The
time-dependentbehaviorofasphaltormodifiedbinder,asanonlinearviscoelasticcomponent(bindercreepcompliancedataandbulkmodu-lus)ischaracterizedbyatime-steppingalgo-rithm
through the use of the hereditary
integralformoftheconstitutiveequations.Theaggre-gate is
characterized as a linearly elastic materi-al by its Youngs modulus
and Poissons ratio. Fig.10showsmixrelaxationmodulusresults
predicted by the micromechanical analy-sis for the control asphalt
and several modifiedbinders at -12C.5 TEMPERATURE
CONTROLLEDSTRESS-STRAIN TESTS ON COMPACTEDASPHALT MIXThe
stress-strain behavior of asphalt and modi-fied asphalt highway
mixes is dependent on timeof loading and test temperature. For
accuratelymodeling asphalt mixes it is important to recog-nize the
temperature- and time-dependent
vis-coelasticmaterialbehaviorthatcanbecharac-terizedinthelaboratorybysimplestaticcreeptests
[10]. The environmental controlled
AsphaltTestingMachine(ATM)wasusedtoconductthesestress-straintestsatdifferenttempera-tures
on compacted asphalt mix specimens
pro-ducedusingtheU.S.ArmyCorpsofEngineersGyratory Testing Machine
(GTM). The GTM wasdeveloped to produce laboratory test
specimensbyakneadingcompactionprocessusinganappropriate gyration
angle to simulate the fieldwheel rolling condition [13]. The
principle of gyra-tory compaction has been adopted in the
SHRPSuperpaveasphaltmixdesignsystem.SeveralGTMcompactedspecimenswerefabricatedusing
the binders from the I-55 test site
[1].TheATMtestswereconductedontheGTMspecimensattemperatureof10Cforthe01,0002,0003,0004,0005,0006,0007,0008,0000
300 600 900 1,200 1,500 1,800Time, SecondsE,
MPaControlAsphaltSealoflexKratonNovophaltRouse RubberTest
Temperature = -12.0 oC
4,535MPa4,703MPa5,004MPa5,277MPa5,064MPa01,0002,0003,0004,0005,0006,0007,0008,0001
2 3 4 5 6 7 8 9Relaxation Modulus Values,
MPaUltrapaveStyrelfNovophaltKratonCryopolymerSealoflexMultigradeControl
AsphaltRouse Results at 1.0 second Test Temperature = -12.0 oC
Binder Figure 10 (left):Mix relaxation moduluspredictions from
theASPHALT program.Figure 11 (right): Comparisonof lab relaxation
modulusresults for virgin asphaltand modified asphaltmixes at
10C.Applied Rheology Vol.13/4.qxd09.09.200311:53 UhrSeite
195196Applied RheologyVolume 13 Issue 4control asphalt and several
modified binders andat 25C for the control asphalt. In this paper,
onlythe results of the tests conducted at 10C are pre-sented for
comparing the behavior of the
controlasphaltwiththebehaviorofthemodifiedbinders. The test
procedure consisted of an
appli-cationofaconstantloadonthesurfaceoftheasphalt compacted mix
specimen. This load
wasappliedfor3600secondsandthenreleasedtomeasurethepermanentdeformationoftheasphaltspecimen.Thedataacquisitionequip-ment
reads the displacement measured by twovertical Linear Variable
Displacement Transduc-ers (LVDTs). Fig. 11 shows a comparison of
the
relax-ationmoduluscalculatedfromthetestdatameasuredforthefirst1800seconds.Therelax-ation
modulus values are normalized by the ini-tial relaxation modulus
calculated at 0.5 secondsfor the virgin asphalt mix. These plots
show
thetime-dependentpropertiesandviscoelasticbehaviorofthemixspecimens.Themodifiedasphalt
mixes perform better than the
unmodi-fiedcontrolasphaltmix.Moreruttingoftheasphalt section after
four years, measured in thefield, confirms this finding.6
COMPARISON OF FIELD EVALUATIONWITH SUPERPAVE BINDER TEST
RESULTSField rutting measurements were made on peri-odic basis by
the MDOT staff for all nine types ofasphalt binder sections on
highway I-55. Fig. 12 isthe box plot of the mean rut depth measured
ontheoutsidewheelpath(OSWP)forthecontrolasphalt and modified
asphalt binders sections inNovember1999.Itshowsthatthelargestamount
of rutting (about 8 mm) occurred in thecontrol asphalt section and
the minimum
ruttingoccurredintheRouseRubbersectionandfol-lowedbytheSealoflex,Styrelf,KratonandNovophalt
polymer-modified asphalt binder
sec-tions.Themeanrutdepthmeasuredintheoutside wheelpath in November
1999 has
beencorrelatedwiththerheologicalpropertiesofthesebinders.TheG*/sind
valuesfortempera-turesof46,76Cfororiginalbinders,76CforRolling Thin
Film Oven (RTFO) aged binders, and22C for Pressure Aging Vessel
(PAV) have beenconsidered.Theseresultswereobtainedfromthe tests
conducted at the University of
Missis-sippi[7]andfromtheMDOTStateStudy123report [8]. From the
plots of these properties
andfieldrutdepthmeasurementsin1999,itisobservedthatthehighestcorrelationexistsbetween
the mean rut depth and the DSR ruttingfactor G*/sind at a
temperature of 76C for bothoriginal binders as well as the RTFO
aged binders.The BBR binder test result at -12C for Rouse
Rub-bershowsrelativelyhigherrelaxationmoduluscompared to the
control asphalt, Sealoflex, Kra-ton, and Novophalt binders.7
BACKCALCULATION OF IN SITUMODULUS VALUESIn this study
nondestructive evaluation (NDE)
ofthepavementsectionswasconductedbymea-suringsurfacedeflectiondatausingthefallingweightdeflectometer(FWD),operatedbytheMississippiDepartmentofTransportation.Thedeflection
data were measured in fall of 1997 and1998 at peak loads of around
4083 kgf (9000 lbf)and higher in the outer wheel path, about 1m
(3ft) from the pavement edge [1]. The FWD deflec-tion data measured
at the last drop height werenormalized to 4083 kgf (9000 lbf) peak
load foranalysis.A comparative study using several back-calculation
programs was done [1]. Based on theresults of this study, the PEDD
modulus backcal-culationprogramwasselectedbecauseitensures the
uniqueness of backcalculated
mod-ulusresultsbypredictingseedmodulusvaluesusing in-built
nonlinear deterministic equations.The I-55 pavement structure model
consists of a76 mm (3 in) asphalt overlay on top of a 190.5
mm(7.5in)asphaltlayer,a305mm(12in)cementtreatedbase(CTB)andlimetreatedsubgrade(LTS)
layer, and a semi-infinite subgrade layer. Insitu effective modulus
values were
backcalculat-0.14770.30110.15340.09090.05110.17610.06820.09660.05110.0455-0.050.000.050.100.150.200.250.300.350.40Ultrapave
Control AsphaltMultigradeMultigrade BSealoflexCryopolymer
KratonNovophaltStyrelfRouse RubberBinder TypeMean Rut Depth,
inInterstate 55 - Outside Wheel PathFigure 12:Mean rut depth
datameasured in 1999.Applied Rheology Vol.13/4.qxd09.09.200311:53
UhrSeite 196197Applied RheologyVolume 13 Issue 4ed by the PEDD
program using the FWD data col-lected on 9th September 1998 at in
situ test tem-peratures in the range of 21.7 to 26.1C (71 to
79F).The in situ modulus results for the overlay layerare
summarized in Fig.
13.TheinsituYoungsmodulusvaluesoftheoverlaybackcalculatedfromthedynamicdeflectiondataalsoshowsmallermodulusforthe
control asphalt compared to all other
binders.ThemodulusvaluesofRouseRubberandSealoflex pavement layers
are closer to the
con-trolasphaltmodulusindicatingadesirablebehavior at the low test
temperature for in-ser-vice pavements in northern Mississippi.
Howev-er,thelowermodulusvalueofthecontrolasphaltlayeratthetesttemperatureof26.1C(79F)
makes it more susceptible to rutting
dur-ingthewarmerin-servicetemperatures,whichmay exceed 50C (122F)
during hot summer days.8 3D-FE SIMULATION USING VIS-COELASTIC
MATERIAL
PROPERTIESKeylimitationsofmostbackcalculationpro-gramsaretheassumptionsoflinearelastic,homogenousandisotropicmaterialpropertiesfor
the pavement layers and static loading.
Threedimensional-finiteelement(3D-FE)analysisenablesfortheevaluationofthethree-dimen-sional
state of stress and strain in a
continuum.Dynamic,static,andnonlinearanalysiscanbeconducted using
3D-FE models [14]. In this study, the LS-DYNA finite elementcode
[15] was used to analyze and verify the back-calculated modulus
values from the PEDD back-calculation program shown in Fig. 13. The
3D-FEhalfmodeldevelopedfortheI-55testsectioncontained7313solidbrickelementsand8528nodes.Theboundaryconditionsforthis3D-FEmodel
were fixed at the bottom and roller sup-ports were used on the
sides. Material type 6 in the LS-DYNA finite ele-ment program was
used to simulate
viscoelasticbehavior.TheviscoelasticpropertiesfortheLS-DYNAinputfileare:thebulkmodulus(B),theshort-termandlong-termshearmodulus(G0andG1),andthedecayparameter,b.Theele-mentsshearmodulusvalueG(t)andthebulkmodulus,
B, are expressed by the following rela-tionships [15]:(2)(3)where E
is the Youngs modulus and is the Pois-sons
Ratio.ThesepropertieswerecalculatedfromtheATMtestsconductedat10ContheGTMcompacted
specimens. From the relaxation mod-ulus plots G0, G1, and b were
calculated. The G0value was calculated at the time of 0.5
secondsand G1at the time of 3000 seconds. A comparison between the
elastic analy-sis and the viscoelastic analysis was done for
thecontrolasphaltandSealoflexpavementssub-jectedtotheFWDimpulseloading.Detailedinputdataandresultsarenotshownhereforbrevity.
The Sealoflex modified binder plot
showslargerdeflectionsfromviscoelasticanalysis,compared to the
control asphalt plot, as
showninFig.14.ThisimpliessmallerinsituYoungsmodulus for the
modified binder pavement
layer.Theseresultsindicatethatcorrectproperties,appropri atemateri
al model s, and3D-FEresponse analysis are needed for correct
perfor-mance modeling of asphalt pavements.9 CONCLUSIONS AND
RECOMMENDA-TIONSSeveralmodifiedasphaltbinderandcontrolasphalt
samples from I-55 highway overlay
pro-jectsiteweretestedusingtheSuperpaveDSRand BBR equipment to
measure and analyze con-BE= 3 1 2 ( ) G t G G G et( ) = + ( )1 0
14,8292,9634,2883,1005,73411,3075,1688,0004,15102,5005,0007,50010,00012,50015,0001
2 3 4 5 6 7 8 9Backcalculated Modulus Values,
MPaUltrapaveStyrelfNovophaltKratonBinder
TypeCryopolymerSealoflexMultigradeControl AsphaltRouse
Rubber-8.00-6.00-4.00-2.000.002.000.00 0.02 0.04 0.06 0.08 0.10
0.12Linear Elastic PropertiesViscoelastic Properties-203.4 m0.0
m50.8 m-8.00-6.00-4.00-2.000.002.000.00 0.02 0.04 0.06 0.08 0.10
0.12Linear Elastic PropertiesViscoelastic
Properties-203.4m0.0m50.8mTime [s]Deflection [mils]Deflection
[mils]Time [s]Figure 13 (left): The PEDDbackcalculated in
situYoungs modulus values forthe overlay.Figure 14
(right):Comparison of FWDdeflection time historiesfrom 3D-FE
elastic andviscoelastic dynamicanalysis.Applied Rheology
Vol.13/4.qxd09.09.200311:53 UhrSeite 197trol asphalt and modified
asphalt binder proper-ties.Creepcomplianceandrelaxationmodulusdata
for different binders were measured usinga modified BBR test
procedure. The performanceof the modified asphalt binders was
better thanthevirginasphalt.Thiswasconfirmedbymorerutting measured
on the control asphalt sectionafter four years of traffic.The mix
creep compliance and relaxation
modulusresultspredictedusingthemicromechanicsapproachalsoindicatepoorperformanceofthecontrol
asphalt section compared to Rouse Rubber,Sealoflex, and other
modified binder sections. Thecorrect time- and
temperature-dependent
charac-terizationofasphaltmaterialsisimperativeforadvancedcomputermodelingandsimulation.Theseimprovedlaboratorymaterialcharacteriza-tion
and analysis efforts are needed for more accu-rate and meaningful
structural response analysis ofasphalt pavements and prediction of
their perfor-mance for resistance to
rutting.ACKNOWLEDGMENTSTheworkreportedhereinwasconductedasapart of
the research studies funded by the
Mis-sissippiDepartmentofTransportation,U.S.Department of
Transportation Federal
HighwayAdministration,andCenterforAdvancedInfra-structure
Technology at the University of
Missis-sippi.Theauthoracknowledgestheeffortsofseveralresearchassistantsandgraduatestu-dentsincludingBenSale,YaminiNangiri,andSergio
Garza.DISCLAIMERThe contents of this paper reflect the views of
theauthor who is responsible for the facts, findings,and data
presented herein.REFERENCES[1]
UddinW:ImprovedAsphaltThicknessDesignProcedures for New and
Reconstructed
HighwayPavements.ReportFHWA/MS-DOT-RD-99-122FinalReport-StateStudySS122.MississippiDepartment
of Transportation (1999). [2]
UddinWandNanagiriY:PerformanceofPoly-mer-Modified Asphalt Field
Trials in MississippiBased on Mechanistic Analysis and Field
Evalua-tion.IJP-InternationalJournalofPavements1(2002) 13-24.[3]
MonismithCLandTayebaliAA:PermanentDeformation(Rutting)ConsiderationsinAsphaltConcretePavementSections.Proceed-ings,AssociationofAsphaltPavingTechnolo-gists
57 (1988) 414-463. [4] Ayers Jr.M and Witczak MW: Resilient
ModulusPropertiesofAsphaltRubberMixesfromFieldDemonstrationProjectsinMaryland.InTrans-portationResearchRecord1492,TRB,NationalResearchCouncil,Washington,D.C.(1995)96-107.[5]
Harvey J, Lee T, Sousa J, Park J, and Monismith CL:Evaluation of
Fatigue and Permanent
Deforma-tionPropertiesofSeveralAsphalt-AggregateFieldMixesUsingStrategicHighwayResearchProgramA-300AEquipment.InTransportationResearchRecord1454,TRB,NationalResearchCouncil,
Washington, D.C. (1994) 123-133.[6] Tayebali AA, Deacon JA, and
Monismith CL: Com-parison of Axial and Diametral Resilient
StiffnessofAsphalt-AggregateMixes.InTransportationResearchRecord1492,TRB,NationalResearchCouncil,
Washington, D.C., (1995) 135-143.[7] Nanagiri Y: Pavement-Subgrade
Material Char-acterizationandAdvancedComputerSimula-tions. A Thesis
for the Master of Sciences at TheUniversity of Mississippi
(2001).[8] RichardSD:EvaluationofPolymerModifiedAsphalts for
Potential Application on
Mississip-piStateHighways.FinalReport-StateStudySS123.MississippiDepartmentofTransportation(1999).[9]
Plaxico CA, Hackett RM, and Uddin W: A
Micro-mechanicalAnalysisofTheViscoelasticResponse of Resin Matrix
Composites.
Proceed-ings,InternationalConferenceonFibreRein-forcedStructuralPlasticsinCivilEngineering,Madras,
India (1995) 99-108.[10] Plaxico CA, Uddin W, and Hackett RM: A
Micro-mechanical Model of Rubber-Modified Asphalt.Proceedings, Vol.
1, Material Congress,
AmericanSocietyofCivilEngineers,Washington,D.C.(1996) 761-770.[11]
Buttlar WG and Roque R: Evaluation of Empiri-cal and Theoretical
Models to Determine AsphaltMixture Stiffnesses at Low Temperatures.
PaperPreparedforPresentationattheAnnualMeet-ingofAssociationofAsphaltPavingTechnolo-gists,
Baltimore (1996).[12]
AboudiJ:MechanicsofCompositeMaterials.Elsevier Publishers
(1991).198Applied RheologyVolume 13 Issue 4Applied Rheology
Vol.13/4.qxd09.09.200311:53 UhrSeite 198199Applied RheologyVolume
13 Issue 4[13]
McRaeJL:UseofGyratoryTestingMachine(GTM)toEvaluateExistingAirfieldPavement,Base
and Subgrade for Use by Larger Military
Air-craft.Proceedings,FifthInternationalRILEMSymposium on
Mechanical Tests for BituminousMaterials, Lyon, France (1997)
575-581.[14] UddinW:SimulationofFallingWeightDeflec-tometer for In
Situ Material Characterization
ofHighwayandAirportPavements.Proceedings,SixthInternationalLS-DYNAUsersConference,Simulation
2000, Dearborn, Michigan (2000) 16-21 to 16-32.[15] LSTC: LS-DYNA
Keyword Users Manual
Version950.LivermoreSoftwareTechnologyCorpora-tion (1999).Applied
Rheology Vol.13/4.qxd09.09.200311:53 UhrSeite 199
