NCATReport17-08
PAVEMENTMEDESIGN–IMPACTOFLOCAL
CALIBRATION,FOUNDATIONSUPPORT,ANDDESIGNAND
RELIABILITYTHRESHOLDS
Dr.NamTran,P.E.Dr.MaryM.RobbinsDr.CarolinaRodezno
Dr.DavidH.Timm,P.E.
September2017
Tran,Robbins,Rodezno,andTimm
PAVEMENTMEDESIGN–IMPACTOFLOCALCALIBRATION,FOUNDATIONSUPPORT,ANDDESIGNANDRELIABILITYTHRESHOLDS
Dr.NamTran,P.E.AssociateResearchProfessor
NationalCenterforAsphaltTechnology
Dr.MaryM.Robbins*ResearchEngineer
OhioResearchInstituteforTransportationandtheEnvironment(*WorkcompletedwhileatNationalCenterforAsphaltTechnology)
Dr.CarolinaRodezno
AssistantResearchProfessorNationalCenterforAsphaltTechnology
Dr.DavidH.Timm,P.E.
BrasfieldandGorrieProfessorofCivilEngineeringPrincipalInvestigator
SponsoredbyNationalAsphaltPavementAssociation
September2017
Tran,Robbins,Rodezno,andTimm
iii
ACKNOWLEDGEMENTSThe authors wish to thank the National Asphalt Pavement Association for sponsoring thisresearchaspartof theOptimizing FlexiblePavementDesignandMaterial Selection researchprojectandforprovidingtechnicalreviewofthisdocument.
DISCLAIMERThecontentsof this reportreflect theviewsof theauthorswhoareresponsible for the factsandaccuracyofthedatapresentedherein.Thecontentsdonotnecessarilyreflecttheofficialviews or policies of the sponsoring agency, the National Center for Asphalt Technology orAuburn University. This report does not constitute a standard, specification, or regulation.Commentscontained inthispaperrelatedtospecifictestingequipmentandmaterialsshouldnotbeconsideredanendorsementofanycommercialproductorservice;nosuchendorsementisintendedorimplied.
Tran,Robbins,Rodezno,andTimm
iv
TABLEOFCONTENTS
1 Introduction.............................................................................................................................5
2 Mechanistic-EmpiricalPavementDesignGuide.......................................................................5
3 LocalCalibrationofMEPDG.....................................................................................................6
4 CaseStudies.............................................................................................................................84.1 LocalCalibrationResults...................................................................................................94.2 SelectionofPerformanceandReliabilityLimits..............................................................124.3 ComparisonofDesignResultsUsingGlobalandLocalCalibrationCoefficients.............134.3.1 DesignProjects..........................................................................................................134.3.2 DesignInputs.............................................................................................................134.3.3 DesignSimulationandEvaluation.............................................................................164.3.4 DesignResults...........................................................................................................17
4.4 Summary.........................................................................................................................24
5 EffectofFoundationSupport.................................................................................................255.1 FoundationMaterialsInputsRequiredinPavementMEDesign.....................................265.1.1 InputsforUnboundandSubgradeLayers.................................................................265.1.2 InputsforStabilizedLayers.......................................................................................285.1.3 EffectofFoundationSupportonPavementMEDesignResults...............................30
6 PerformanceCriteriaandReliability......................................................................................336.1 PerformanceCriteriaandReliabilityLevels.....................................................................336.2 EffectofPerformanceCriteriaandReliabilityonPavementDesign...............................346.2.1 SensitivityofPermanentDeformationinUnboundLayerstoPavementDesign
Thickness...................................................................................................................366.2.2 SensitivityofBottom-UpFatigueCrackingtoPavementDesignThickness..............386.2.3 ProposedPerformanceCriteriaandReliabilityLevelsforPavementMEDesign......40
7 ConclusionsandRecommendations......................................................................................41
References....................................................................................................................................44
AppendixAPerformanceModelsforFlexiblePavementDesign.................................................46
Tran,Robbins,Rodezno,andTimm
5
1 INTRODUCTION
The Mechanistic-Empirical Pavement Design Guide (MEPDG) and the accompanyingAASHTOWarePavementMEDesignsoftware(hereafterreferredtoastheMEDesignsoftware)havebeendevelopedtoreplacetheempiricalAASHTOPavementDesignGuides.TheMEPDGrepresentsaquantumleapforwardfromtheempiricalpavementdesignprocedures(1,2).Asindicated in a survey of state agencies conducted in 2013, 43 agencies were evaluating theMEPDG, and 15 agencies planned to implement the new design procedure in the next twoyears (3). The implementation plans of these agencies include, among other elements,importantstepsfor(1)conductinglocalcalibrationtoaccountfordifferencesinstatepractices,policies, and local conditions, and (2) selecting design thresholds and reliability levels foracceptable pavement designs (4, 5). Without properly conducting these importantimplementationsteps,theadoptionoftheMEPDGwillnotmakethepavementdesignprocess“better.”Infact,ithasbeensuggestedthatuseofthegloballycalibratedMEDesignsoftwaremaypotentiallyyieldinaccurateasphaltpavementdesigns(6,7).
Recognizing the importance of local calibration and selection of design thresholds andreliability levels, this study provides information and evidence to support the need for localcalibrationoftheMEPDGandcarefulconsiderationofdesignthresholdsandreliabilitylevelsintheimplementationprocess.Theresultsofthisresearcheffortarepresentedintworeports.Apreviousreportwaspreparedtodiscussthegeneralapproachto localcalibrationundertakenby state agencies and to summarize results of their local calibration efforts andrecommendations for implementing the locally calibrated MEPDG (8). This (second) reportpresents resultsof a case study that comparespavementdesigns conductedwith global andlocalcalibrationcoefficientsto illustratethe importanceofconducting localcalibrationoftheMEPDG in the implementation process. In addition, it provides results of sensitivity analysesthat show the effect of performance criteria, reliability levels, and foundation support onpavementdesign.
In the following sections, the MEPDG and local calibration methodologies are brieflydiscussed, followedby case studiesand resultsof sensitivity analyses. Finally, key findingsofthisresearcheffortaresynthesized,andrecommendationsareoffered.
2 MECHANISTIC-EMPIRICALPAVEMENTDESIGNGUIDE
TheMEPDG was developed to design new and rehabilitated pavement structures based onmechanistic-empirical principles. Figure 1 illustrates basic steps for conducting a pavementdesign using the ME Design software. Based on the inputs and trial pavement designinformation, the ME Design software “mechanistically” calculates pavement responses(stresses and strains) and uses those responses to compute incremental damage overtime. The program then utilizes the cumulative damage to “empirically” predict pavementdistresses for each trial pavement structure. A trial pavement structure is accepted as thefinaldesignwhenitspredictedpavementdistressesmeetthedesigncriteriaattheselectedreliabilitylevels.
Tran,Robbins,Rodezno,andTimm
6
Figure1.BasicStepsofPavementMEDesign(7)
The mechanistic analysis utilizes the Enhanced Integrated Climatic Model, structural
responsemodels, and time-dependentmaterial propertymodels. The empirical analysis usesthedistressprediction (regression)models, sometimescalled transfer functions, representingrelationships between the cumulative damage and observed pavement distresses.While themechanistic models are assumed to be accurate and to correctly simulate field conditions,inaccuracies still exist andaffect the resultsofdistressprediction functioncomputationsandfinal distress predictions (5). The local calibration process, which is briefly discussed in thefollowing section, is often related to the distress prediction functions, but it essentiallyaddressestheerrorsofboththemechanisticandempiricalanalyses.
3 LOCALCALIBRATIONOFMEPDG
Under the NCHRP 1-37A and 1-40 projects, the MEPDG was “globally” calibrated using arepresentativedatabaseofpavementtestsitesacrossNorthAmerica.Mostofthesetestsiteshave been monitored through the Long-Term Pavement Performance (LTPP) program. TheywereusedbecauseoftheconsistencyinthemonitoreddataovertimeandthediversityoftestsectionsspreadthroughoutNorthAmerica.However,constructionandmaterialspecifications,pavementpreservationandmaintenancepractices,andmaterialsandclimaticconditionsvaryacrossNorthAmerica.ThesedifferencesarenotcurrentlyconsidereddirectlyintheMEDesignsoftware but are indirectly considered through local calibration in which the calibrationcoefficientsofdistresspredictionfunctionsintheMEDesignsoftwarecanbeadjusted(5).
Duetothedifferencesinthelocalpracticesandconditions,thedistressespredictedbytheglobally calibrated distress prediction models may have higher bias and/or lower precisionwhen compared with the locally measured pavement distresses. As illustrated in Figure 2,
Traffic
Foundation
Climate
Materials
TrialDesignStructuralResponse
Models
DistressPredictionModels
MeetDesignCriteria?
FinalDesign
No
No
YesInputs
Analysis
Results
Tran,Robbins,Rodezno,andTimm
7
throughlocalcalibration,thecoefficientsofthesedistresspredictionmodelsmaybeadjustedtoimprovethebiasandprecisionofthemodelsintheMEDesignsoftware.Inpracticalterms,biasisthedifferencebetweenthe50%reliabilitypredictionandthemeasuredmean.Precisiondictateshowfarthepredictedvaluesataspecifieddesign reliability levelwouldbe from thecorresponding predicted values at the 50% reliability. The locally calibratedmodels are thenvalidatedusinganindependentsetofdata.Themodelsareconsideredsuccessfullyvalidatedtothe local conditions if the bias and precision statistics of the models are similar to thoseobtainedfrommodelcalibrationwhenappliedtothevalidationdataset.
Figure2.ImprovementofBiasandPrecisionthroughLocalCalibration
Basicstepsforcalibratingadistresspredictionmodeltoimproveitsbiasandprecisionare
showninFigure3.TheMEDesignsoftwarewithglobalcalibrationfactorsisfirstconductedtodesign pavements at 50% reliability using the inputs available from the pavement segmentsthathavebeenselectedforlocalcalibration.Thepredicteddistressesarethencomparedwiththemeasureddistressesoftheselectedpavementsegments,anddiagnosticstatistics,includingR-square, bias, and the standarderrorof theestimate (Se), aredetermined. If thediagnosticstatisticsarenotacceptable,themodelcalibrationcoefficientsareadjusted,andtheanalysisisrepeatedusingtheadjustedcoefficientsuntilthediagnosticstatisticsaredeemedacceptable.
A step-by-step procedure for local calibration is described in the Guide for the LocalCalibration of the MEPDG (5). The procedure includes detailed steps for (1) selecting andcollecting inputs for local calibration, (2)determining local calibrationcoefficients to improvethe bias and precision of each distress prediction function, and (3) reviewing the calibrationresults tomake sure theexpectedpavementdesign life is “reasonable” for theperformancecriteriaandreliabilitylevelsselectedforfutureusebytheagency.
Higherbias
Lowerprecision
Lowerbias
Higherprecision
Pred
icted
Pred
icted
Measured Measured
Local
Calibration
Tran,Robbins,Rodezno,andTimm
8
Figure3.LocalCalibrationofPavementMEDesign
Since the first releaseof theMEPDG, several stateshave sponsored studies to verifyand
calibratetheMEPDGtolocalmaterialsandconditions.Detailedresultsofthelocalcalibrationeffortscompletedthrough2015werereviewedinthisstudyandpresentedinapreviousreport(8). Despite the availability of the Guide for the Local Calibration of the MEPDG (5), manycalibration efforts did not follow the step-by-step procedure or the terminology used in theguide, in part due to the timing of the publication (i.e., in 2010) relative to the timing ofcalibrationeffortsineachstateandthetimededicatedtosuchefforts.Inthefollowingsection,resultsofthelocalcalibrationeffortsintwostates,MissouriandColorado,aresummarizedintwocasestudies thatcomparepavementdesignsconductedwithglobaland local calibrationcoefficientstoillustratetheimpactoflocalcalibrationonMEPDGpavementdesigns.
4 CASESTUDIES
Aftercompleting the localcalibrationof theMEPDG,somestateshaveconsidered taking thenext step of the implementation process, which is adopting the locally calibrated designprocedureforsomeroutinepavementdesigns.Amongthestateagenciesthathavecompletedthe local calibration, the Missouri Department of Transportation (MoDOT) and ColoradoDepartment of Transportation (CDOT) have implemented the MEPDG for routine pavementdesigns.BothMoDOTandCDOThaveusedtheMEPDGtodesignnewasphaltpavementsandjointed plain concrete pavements (JPCPs) as well as asphalt and concrete overlays. In thefollowing sections, results of the local calibration sponsored by these agencies are firstdiscussed, followed by performance and reliability limits selected by the states for futuredesigns.Finally,acomparisonofpavementdesignsconductedwithglobalandlocalcalibrationcoefficientsispresentedtoillustratehowthelocalcalibrationcoefficientsaffectthePavementMEDesignresults.
LocalTraffic
Foundation
LocalClimate
LocalMaterials
M-EAnalysis
AdjustCalibrationCoefficients
MeetAllowableTolerance?
FinalCalibrationCoefficients
No
Yes
PredictedStresses
MeasuredDistresses
Tran,Robbins,Rodezno,andTimm
9
4.1 LocalCalibrationResults
BothMoDOTandCDOThiredaconsultanttocalibratetheMEPDGdistresspredictionmodelstolocal conditions. The two agencies conducted local calibration using Version 1.0 of thePavementMEDesignsoftware.Thepavementtypesconsideredforlocalcalibrationinclude(1)newasphaltpavementandasphaltoverlayoverexistingasphaltandconcretepavements;and(2)new JPCP, JPCPoverlayoverexistingasphaltpavement, andunbonded JPCPoverlayoverexistingJPCP.Asummaryofthe localcalibrationeffortsforbothnewasphaltpavementsandasphalt overlays over existing asphalt and concrete pavements is presented in this section.Moredetailed informationabout the local calibrationefforts in these states canbe found inotherreports(8-10).
The local calibration efforts sponsored by the agencies included two main activities—preparing required information for local calibration and calibrating performance predictionmodels.BothstatesdeterminedfromprevioussensitivitystudiestheinputsthathaveamajorimpactondistressandInternationalRoughness Index(IRI)predictions;therefore,theyshouldbe characterized accurately using the highest possible hierarchical input level. Using moreaccurateinputswouldhelpreducepredicteddistress/IRIstandarderrorordeviation,whichisakeycomponentofthevariabilitytermsusedincalculatingdesignreliability.Ahigherstandarderrorwouldresultinhigherpredicteddistress/IRIataspecifiedreliabilitylevelgreaterthan50percent,whichinturnwouldrequireathickerpavementstructureorbettermaterials,bothofwhichcanaffecttheeconomyofthepavement.
To prepare inputs essential to the local calibration of performance prediction models,several researcheffortswereconductedorsponsoredbythestates to (1)characterize trafficinputs, (2) determine laboratory material properties of typical asphalt mixtures, aggregatebases,andsubgradesoils, (3)conduct in-situtestingofpavements,and(4)analyzepavementperformancedata.Table1summarizesthehierarchicalinputlevelsusedinlocalcalibrationineachstate.InTable1,Level1inputsrequirethehighestlevelofaccuracyandarelaboratory-and/or field-measureddata. Level2 inputs requirean intermediate levelofaccuracyandaredeterminedusingprocedures (or correlations) similar to thoseused in theempiricalAASHTOPavementDesignGuides. Level 3 inputs require the lowest level of accuracy andaredefaultinputspreviouslyusedbystateorprovidedintheMEDesignsoftware(9,10).
In addition to the inputs listed in Table1, automatedandmanualdistress surveysof theselectedpavementsectionswereconductedbyeachstateforlocalcalibration.Thedatawereused to characterize pavement condition (measured fatigue cracking, transverse cracking,rutting,andIRI).Table2providesthenumberof500-ftand1,000-ftpavementsectionsselectedbyMoDOTandCDOT,respectively,forlocalcalibration.
Tran,Robbins,Rodezno,andTimm
10
Table1.InputLevelsforLocalCalibration(9,10)InputGroup MoDOT CDOT
AllTrafficInputs(ExceptasNoted) Level3 Level3AADTT Level1 Level1AxleLoadDistribution Level1 Level1VehicleClassDistribution Level1 Level1TruckWheelBasePercentages Level3 Level1ClimaticInputs Level2 Level2AllAsphaltLayerInputs(ExceptasNoted) Level3 Level3MixtureVolumetrics Level1 Level3(CDOTDefaults)MechanicalProperties Level2 Level2AllUnboundAggregateLayerInputs(ExceptasNoted) Level3 Level3Classification Level3 Level1*ResilientModulus Level2 Level2Moisture-DensityRelationships Level2 Level2AllSubgradeLayerInputs(ExceptasNoted) Level3 Level3Classification Level3 Level1*ResilientModulus Level2 Level2*Laboratory-measureddatautilizedTable2.PavementSitesSelectedforLocalCalibration(9,10)
PavementMoDOT(500-ft) CDOT(1,000-ft)
LTPP Agency Total LTPP Agency TotalNewAsphalt 14 6 20 30 16 46AsphaltOverlayoverAsphalt 11 0 11 20 21 41AsphaltOverlayoverJPCP 9 0 9 2 5 7Total 34 6 40 52 42 94
The results of laboratory and field evaluation, layer thickness measurement, and fieldperformancesurveyoftheselectedpavementsectionswereusedtoevaluateandcalibratetheperformancepredictionmodels intheMEDesignsoftwareto localconditions inMissouriandColorado.A summaryof theglobal and local calibration coefficients forMoDOTandCDOT isshown inTable3 throughTable6.ForMoDOT, fourmodelswere locallycalibrated, includingasphalt rutting, total rutting, transverse cracking, and IRImodels. The fatigue crackingmodelwasfoundtobeappropriateforuseinMissouri;thus,itscoefficientswerenotadjustedduringlocal calibration. Five models were calibrated by CDOT, including fatigue cracking, asphaltrutting,totalrutting,transversecracking,andIRI.TheinformationpresentedinTable3throughTable6wasusedlaterincasestudiestocomparepavementdesignsconductedwithglobalandlocal calibration coefficients. More information about the performance models and theircoefficientsispresentedinAppendixA.
Tran,Robbins,Rodezno,andTimm
11
Table3.LocalCalibrationResultsforFatigueCrackingModel(8,9,10)PerformanceIndicator Coefficient Global* MoDOT CDOT
FatigueCracking
kf1 0.007566
NotCalibrated
0.007566kf2 -3.9492 -3.9492kf3 -1.281 -1.281βf1 1 130.3674βf2 1 1βf3 1 1.218C1 1 0.07C2 1 2.35C4 6000 6000
GoodnessofFitR2,% 27.5 62.7Se(%) 5.01 9.4N 405 56
Biasp-value(pairedt-test) NotReported 0.7566
p-value(slope) NotReported 0.3529*GloballycalibratedcoefficientsshowninMEDesignsoftwareTable4.LocalCalibrationResultsforRuttingModels(8,9,10)PerformanceIndicator Coefficient Global* MoDOT CDOT
AsphaltRutting
kr1 -3.35412NotReported
-3.3541kr2 1.5606 1.5606kr3 0.4791 0.4791βr1 1 1.07 1.34βr2 1 1 1βr3 1 1 1
FineGradedSubmodelks1 1.35 NotReported 0.84βs1 1 0.01 NotReported
GranularSubmodelks1 2.03 NotReported 0.4βb1 1 0.4375 NotReported
GoodnessofFitR2,% 57.7 52 41.7Se(in) 0.107 0.051 0.147N 334 183 137
Biasp-value(pairedt-test)
NotReported0.943 0.4306
p-value(intercept) 0.05 0.0898p-value(slope) 0.322 NotReported
*GloballycalibratedcoefficientsshowninMEDesignsoftware
Tran,Robbins,Rodezno,andTimm
12
Table5.LocalCalibrationResultsfor(Thermal)CrackingModel(8,9,10)PerformanceIndicator Coefficient Global* MoDOT CDOT
TransverseCrackingK(Level1) 1.5 0.625 7.5K(Level2) 0.5 K(Level3) 1.5
GoodnessofFitR2,%
NotReported
91(Level1) 43.1Se(ft/mi) 51.4 194
N 49 12
Biasp-value(pairedt-test) 0.0041 0.529p-value(intercept) 0.907 NotReportedp-value(slope) <0.0001 0.339
*GloballycalibratedcoefficientsshowninMEDesignsoftwareTable6.LocalCalibrationResultsforInternationalRoughnessIndexModel(8,9,10)PerformanceIndicator Coefficient Global* MoDOT CDOT
Top-Down(Longitudinal)Cracking
C1 40 17.7 35C2 0.4 0.975 0.3C3 0.008 0.008 0.02C4 0.015 0.01 0.019
GoodnessofFitR2,% 56 53 64.4
Se(in/mi) 18.9 13.2 17.2N 1,926 125 343
Biasp-value(pairedt-test)
NotReported0.6265 0.1076
p-value(intercept) 0.0092 0.3571p-value(slope) 0.225 NotReported
*GloballycalibratedcoefficientsshowninMEDesignsoftware
4.2 SelectionofPerformanceandReliabilityLimits
After the performance prediction models were calibrated, the states selected performancecriteria and reliability levels thatmet their needs for designing new asphalt pavements andasphaltoverlays.Usingtheselectedperformancecriteriaandreliabilitylevels,theyconductedtrial designs using the local calibration coefficients to ensure that the expected pavementdesignlifewasreasonableforfutureuse.
Table7liststheperformancecriteriaandreliabilitylimitsselectedbyMoDOTfordesigningnew asphalt pavements and asphalt overlays. Even though four performance models werelocally calibrated, includingasphalt rutting, total rutting, transversecracking,and IRI,MoDOTcurrentlydesignspavementsbasedonlyonfatiguecrackingandrutting inasphalt layers.Theperformancecriterionforfatiguecrackingwassettominimize/eliminatebottomupcrackinginasphaltlayers,andthecriterionforasphaltruttingwasdeterminedbasedontheapproximatedepthtoreducethepotentialforhydroplaning.MoDOThasnotadoptedtheIRIcriteriainthepavementdesignprocess.
Tran,Robbins,Rodezno,andTimm
13
Table7.MoDOTPerformanceCriteriaandReliabilityLimits(11)PerformanceIndicator Years PerformanceCriteria Reliability(%)FatigueCracking 30 2%LaneAreaMaximum 50AsphaltRutDepth 20 0.5in.Maximum 50
Table 8 shows the reliability levels and performance criteria chosen by CDOT for theimplementation of PavementME Design in Colorado. The selected thresholds are similar tothoserecommendedintheManualofPractice(4).Thereliabilityandperformancelimitsvarybased on the functional classification with the thresholds for higher traffic roadways beingmorestringent.Thecriteriaarealsodifferentfornewpavementandoverlaydesigns.Fornewpavement designs, the thresholds for terminal IRI, total rutting, AC rutting, and top-downfatiguecrackingareonlyrequiredfortheyearstothefirstrehabilitation(withaminimuminitialperformanceperiodof12years)whilethecriteriaforbottom-upfatiguecrackingandthermalcrackingarerequiredfortheentiredesignlifeofthenewpavement.Foroverlaydesigns,allofthecriteriaarerequiredfortheyearstotheendoftheoverlaydesign life,andtheminimumagefortheoverlaysunderrehabilitationconsiderationis10years.
4.3 ComparisonofDesignResultsUsingGlobalandLocalCalibrationCoefficients
4.3.1 DesignProjects
Toillustratetheimpactoflocalcalibrationonpavementdesigns,theresearchteamcontactedCDOT andMoDOT to obtain information for case studies. CDOT suggested a reconstructionproject on I-25 at Cimarron Boulevard in Colorado Springs, Colorado, and MoDOTrecommendedanewrealignmentprojectonUS-50 inOsageCounty,Missouri.Bothagenciesconducted initial flexible pavement and jointed plain concrete pavement (JPCP) designs forthesesectionsusingthePavementMEDesignsoftwarewiththeirlocalcalibrationcoefficients.CDOTandMoDOTsharedthedesignfilesforconcreteandasphaltpavementswiththeresearchteamtoconductadditionaldesignsusingtheglobalcalibrationcoefficientsforcomparison.
4.3.2 DesignInputs
Table9showsthedesign lifeandbasic traffic inputs fornewasphaltandconcretepavementdesignsconductedbyMoDOTandCDOT.MoDOTdesignedboththenewflexiblepavementandJPCPfor45years.TheyearsshowninTable9(30yearsforflexiblepavementand25yearsforJPCP) represented the last time within the design period when MoDOT would performrehabilitation.CDOTdesignedthenewasphaltpavementfor20yearsandthenewJPCPfor30years. Inadditiontothetraffic inputssummarized inTable9,MoDOTuseddefaultvaluesfortheother traffic inputs in thePavementMEDesignsoftware,whereasCDOTused its specifictraffic inputs for vehicle class distribution, monthly adjustment, and axle load distributionfactorsdeterminedduringlocalcalibration.MoreinformationabouttheseinputsisdescribedinCDOT’sM-EPavementDesignManual(12).
Tran,Robbins,Rodezno,andTimm
14
Table8.CDOTPerformanceCriteriaandReliabilityLimits(12)
Classification Reliability(%)
ToDeterminetheYearstoFirstRehabilitation*
MaximumValueattheEndoftheDesignLife
TerminalIRI
(in/mi)
TotalRutting(in)
RuttinginAClayers
(in)
Top-DownCracking(ft/mi)
Bottom-UpCracking(%lane)
ThermalCracking(ft/mi)
ReflectiveCracking(%lane)
Interstate 80-95 160 0.40 0.25 2,000 10 1,500 5PrincipalArterials(Freeways/Expressways) 75-95 200 0.50 0.35 2,500 25 1,500 10PrincipalArterials(Others) 75-95 200 0.50 0.35 2,500 25 1,500 10MinorArterial 70-95 200 0.65 0.50 3,000 35 1,500 15MajorCollectors 70-90 200 0.65 0.50 3,000 35 1,500 15*Maximumvalueusedtodeterminetheyearstothefirstrehabilitationfornewpavementdesignsormaximumvalueattheendofthedesignlifeforoverlaydesigns.Theminimumagetothefirstrehabilitationforflexiblepavementsshallbe12years.
Table9.GeneralandTrafficInputs
DesignInputsMoDOT’sInputs CDOT’sInputs
NewFlexible NewJPCP NewFlexible NewJPCPDesignLife 30Years* 25Years* 20Years 30YearsTrafficInputsTwo-wayAADTT 1056 11,975No.ofLanes 2 3TrucksinDesignDirection 50% 50%TrucksinDesignLane 95% 60%OperationalSpeed 60mph 60mphGrowthRate 2.2%Compound 1.43%Linear*TheseyearsrepresentedthelasttimewithinthedesignperiodwhenMoDOTwouldperformrehabilitation.Bothpavementsweredesignedfor45years.
Tran,Robbins,Rodezno,andTimm
15
TheclimatefilesforJeffersonCity,MissouriandColoradoSprings,ColoradowereselectedforutilizationintheMoDOTandCDOTdesigns,respectively,topredictpavementtemperatureandmoisture.Thesoftwareusedthesepredictionstomodifytheasphaltconcretemodulusasafunction of temperature and the granular materials as a function of moisture content. Theannualaveragedepthofwatertablewas3ft. fortheMoDOTdesignand10ft. fortheCDOTdesign.
Table10throughTable13showsthepavementstructuresselectedforthenewflexibleandrigidpavementdesignsbyMoDOTandCDOT.MoDOTusedLevel3materialinputswhileCDOTusedLevel1site-specificmaterialinputsintheirdesigns.Table10.PavementStructureforMoDOT’sNewFlexiblePavementDesign
PavementStructure MoDOT’sInputs
Layer1:AsphaltConcreteThickness:1.8in.
12.5-mmSuperpave,PG70-22Level3Inputs
Layer2:AsphaltConcreteThickness:3in.
25-mmSuperpave,PG70-22,Level3Inputs
Layer3:AsphaltConcreteThickness:VariedtoMeetDesignCriteria
25-mmSuperpave,PG64-22,Level3Inputs
Layer4:CrushedStoneThickness:18in.
Classification:A-1-aMr=30,000psiGradation&OtherPropertiesforA-1-a
Layer5:SubgradeSubgradeA-7-6Mr=8,000psiGradation&OtherPropertiesforA-7-6
Table11.PavementStructureforMoDOT’sNewJPCPDesign
PavementStructure MoDOT’sInputs
Layer1:PortlandCementConcreteThickness:VariedtoMeetDesignCriteria
LimestonewithTypeICementCoefficientofThermalExpansion=5.5x10-6in/in/oFLevel3Inputs
Layer2:CrushedStoneThickness:18in.
Classification:A-1-aMr=30,000psiGradation&OtherPropertiesforA-1-a
Layer3:SubgradeClassification:A-7-6Mr=8,000psiGradation&OtherPropertiesforA-7-6
Tran,Robbins,Rodezno,andTimm
16
Table12.PavementStructureforCDOT’sNewFlexiblePavementDesign
PavementStructure CDOT’sInputs
Layer1:AsphaltConcreteThickness:2in.
StoneMatrixAsphalt,PG76-28,Mix#FS1919-2Level1Inputs
Layer2:AsphaltConcreteThickness:VariedtoMeetDesignCriteria
Superpave,PG64-22,Mix#FS1938-1Level1Inputs
Layer3:CrushedGravelThickness:6in.
Classification:A-1-aMr=41,424psiGradation&OtherPropertiesforA-1-a
Layer4:SubgradeThickness:120in.
Classification:A-2-4Mr=13,808psiGradation&OtherPropertiesforA-2-4
Layer5:BedrockHighlyFracturedandWeatheredE=500,000psi
Table13.PavementStructureforCDOT’sNewJPCPDesign
PavementStructure CDOT’sInputs
Layer1:PortlandCementConcreteThickness:VariedtoMeetDesignCriteria
GranitewithTypeICement,Mix#2009105CoefficientofThermalExpansion=4.86x10-6in/in/oFLevel1Inputs
Layer2:CrushedGravelThickness:6in.
Classification:A-1-aMr=44,445psiGradation&OtherPropertiesforA-1-a
Layer3:SubgradeThickness:120in.
Classification:A-2-4Mr=28,905psiGradation&OtherPropertiesforA-2-4
Layer4:BedrockHighlyFracturedandWeatheredE=500,000psi
4.3.3 DesignSimulationandEvaluation
MoDOT’sandCDOT’snewflexibleandrigidpavementdesignswereconductedusingglobalandlocal calibration coefficients in this study, resulting in eight unique designs. For each design,layerthicknesseswerevariedwhiletheotherdesigninputswerekeptconstant.Foraparticularsetoflayerthicknesses,thesoftwareperformedsimulationstopredictpavementperformanceover the selected design period. The simulation output consisted of incremental pavementdamageanddistressesover timeat the50-percent reliability level andat the reliability levelselectedforeachdesign.Apavementcross-sectionwasacceptableifalldistressespredictedattheselectedreliabilitylevelwerebelowthedesigncriteria.ThecriteriaandreliabilitylevelsforthedesignsinthecasestudiesarepresentedinTable14.
Tran,Robbins,Rodezno,andTimm
17
Table14.DesignCriteriaandReliabilityLevels
PerformanceCriteriaMoDOT CDOT
Limit Reliability Limit Reliability
NewFlexibleInitialIRI(in./mi) 63 NA 50 NATerminalIRI(in./mi) 172 50 160 90ACTop-downFatigueCracking(ft/mile) NA NA 2,000 90ACBottom-upFatigueCracking(percent) 2 50 10 90ACThermalCracking(ft/mile) NA NA 1,500 90PermanentDeformation-TotalPavement(in.) 0.75 50 0.4 90PermanentDeformation-ACOnly(in.) NA NA 0.25 90NewJPCPInitialIRI(in./mi) 63 NA 75 NATerminalIRI(in./mi) 172 50 160 90JPCPTransverseCracking(PercentSlabs) 1.5 50 7 90MeanJointFaulting(in.) 0.15 50 0.12 90*NA=NotAvailable4.3.4 DesignResults
Table15 shows the finalpavementdesignsobtained from theMEDesign softwareusing theglobal and local calibration coefficients. For MoDOT’s designs, the final thicknesses are thesamefornewflexiblepavementsusingglobalandlocalcalibrationcoefficients,butthedesignthicknessusingtheglobalcalibrationcoefficientsisslightlygreaterthantheoneusingthelocalcalibration coefficients for the new JPCP designs. For CDOT’s designs, the local calibrationcoefficients yielded a thinner asphalt concrete layer but the samePortland cement concretestructure as the global calibration coefficients. A detailed analysis of the performancepredictionresultsfollows.Table15.ComparisonofDesignThicknessesusingGlobalandLocalCalibrationCoefficients
PavementLayerMoDOT CDOT
Global Local* Global Local*
NewFlexibleAsphaltConcrete(in.) 8 8 11.5 10.5CrushedAggregateBase(in.) 18 18 6 6NewRigid(JPCP)PortlandCementConcrete(in.) 8.5 8** 7.5 7.5CrushedAggregateBase(in.) 18 18 6 6*BasedonMoDOT’sandCDOT’sdesigns**DuetochangesinlocallycalibratedIRImodel
Figure4andFigure5comparethepredicteddistressesandIRIresultsforflexibleandrigidpavementdesignsfortherealignmentprojectonUS-50inOsageCounty,Missouriusingglobaland local calibration coefficients. Based on these results, the following observations can bedrawn.
Tran,Robbins,Rodezno,andTimm
18
• The flexible pavement designs using the global and local calibration coefficients havethe same thickness designs (Table 15) because the designs are governed by thepredictedbottom-upcrackingdistresses,asshowninFigure4.Thepredictedbottom-upcrackingresultsarethesameforthetwodesignssincethecoefficientsforthebottom-upcrackingmodelwerenotadjustedduringlocalcalibration.
• Also, shown in Figure 4, the predicted total rutting and IRI results basedon the localcalibrationcoefficientsarelowerthanthosebasedontheglobalcalibrationcoefficientsastheperformancepredictionmodelsforasphalt layerrutting,baselayerrutting,andIRIwerelocallycalibrated.NotshowninFigure4arethepredictedtransverse(thermal)crackingresultsfortheflexiblepavementdesignssincetransversecrackingisnotusedasadesigncriterioninMissouri.
• ThelocalcalibrationJPCPdesign(Figure5)is0.5in.thinnerthantheglobalcalibrationJPCPdesign,asshowninTable15.Withathinnerthicknessdesign,thelocalcalibrationcoefficients yield similar predicted PCC cracking and faulting results but higher IRIpredictionsthantheglobalcalibrationcoefficients.
SimilartoFigure4andFigure5,Figure6andFigure7comparethepredicteddistressesandIRIresultsforflexibleandrigidpavementdesignsusingglobalandlocalcalibrationcoefficientsforthereconstructionprojectonI-25atCimarronBoulevardinColoradoSprings,Colorado.Thefollowingobservationscanbedrawnfromtheseresults.
• With a 1-in. thinner asphalt layer (Table 15), the local-calibration design yields lowerruttingandcrackingpredictionsandslightlyhigher IRI results (Figure6).More flexiblepavementdesignresultsareshowninTable16andTable17.Asshowninthesetables,all of the performance predictions pass the corresponding design criteria except thepredicted total rutting and AC rutting results. The asphalt pavement design was stillacceptedbyCDOTasruttinghadnotbeenfoundtobeaperformance issue insimilarpavements in thearea.This suggests that theCDOTdesign is largelygovernedby thepredictedbottom-upcrackingresults.
• FortheJPCPdesigns(Figure7),theglobalandlocalcalibrationcoefficientsyieldsimilarperformancepredictions,suggestingthatthelocalcalibrationhasaminimumeffectonJPCPdesignforthisproject.
Tran,Robbins,Rodezno,andTimm
19
Figure4.MoDOT’sFlexiblePavementMEDesignResultsusingGlobal(Left)andLocal(Right)CalibrationCoefficients
Tran,Robbins,Rodezno,andTimm
20
Figure5.MoDOT’sRigidPavementMEDesignResultsusingGlobal(Left)andLocal(Right)CalibrationCoefficients
Tran,Robbins,Rodezno,andTimm
21
Figure6.CDOT’sFlexiblePavementMEDesignResultsusingGlobal(Left)andLocal(Right)CalibrationCoefficients
Tran,Robbins,Rodezno,andTimm
22
Figure7.CDOT’sRigidPavementMEDesignResultsusingGlobal(Left)andLocal(Right)CalibrationCoefficients
Tran,Robbins,Rodezno,andTimm
23
Table16.DistressPredictionResultsforCDOTFlexiblePavementMEDesignusingGlobalCalibrationCoefficients
DistressType(ACThickness=11.5in.)DistressatSpecifiedReliability Reliability(%)
CriterionSatisfied?Target Predicted Target Achieved
TerminalIRI(in./mile) 160.00 133.12 90.00 98.81 PassPermanentDeformation-TotalPavement(in.) 0.40 0.66 90.00 13.86 FailACBottom-upFatigueCracking(percent) 10.00 9.35 90.00 91.48 PassACThermalCracking(ft/mile) 1500.00 84.34 90.00 100.00 PassACTop-downFatigueCracking(ft/mile) 2000.00 309.69 90.00 100.00 PassPermanentDeformation-ACOnly(in.) 0.25 0.42 90.00 37.67 FailTable17.DistressPredictionResultsforCDOTFlexiblePavementMEDesignusingLocalCalibrationCoefficients
DistressType(ACThickness=10in.)DistressatSpecifiedReliability Reliability(%)
CriterionSatisfied?Target Predicted Target Achieved
TerminalIRI(in./mile) 160.00 149.52 90.00 94.78 PassPermanentDeformation-TotalPavement(in.) 0.40 0.53 90.00 53.04 FailACBottom-upFatigueCracking(percent) 10.00 9.09 90.00 92.19 PassACThermalCracking(ft/mile) 1500.00 537.38 90.00 100.00 PassACTop-downFatigueCracking(ft/mile) 2000.00 330.61 90.00 100.00 PassPermanentDeformation-ACOnly(in.) 0.25 0.43 90.00 32.52 Fail
Tran,Robbins,Rodezno,andTimm
24
4.4 Summary
Thecase studieswereconducted for twopavementdesignprojects,one inMissouriand theotherinColorado.Thetwostatescompletedtheirlocalcalibrationprocessesandhaveutilizedthe Pavement ME Design software for routine pavement designs. For the local calibrationprocessescompletedthusfar,CDOTspentmoretimeandeffortondeterminingmoreaccuratedesigninputsandselectedmorepavementsites.Asmoreinformationisavailablesincethelastlocal calibration, both CDOT and MoDOT have planned for recalibrating their MEPDGprocedures.
MoDOTcalibratedfourmodelsforasphaltrutting,totalrutting,transversecracking,andIRI;the fatigue cracking was found to be appropriate for use in Missouri. CDOT calibrated fivemodelsforfatiguecracking,asphaltrutting,totalrutting,transversecracking,andIRI.
Even though four performancemodels were locally calibrated,MoDOT currently designspavements based only on fatigue cracking and rutting in asphalt layers. The performancecriteriaforthesedistressesweresettominimize/eliminatebottomupcrackinginasphaltlayersandtoreducethepotentialforhydroplaning.MoDOTconductsallflexiblepavementdesignsata50%reliabilitylevel.
CDOTselecteditsdesigncriteriaandreliabilitylevelssimilartothoserecommendedintheManualofPractice(4);theyvarybasedonthefunctionalclassificationwiththethresholdsforhighertrafficroadwaysbeingmorestringent.Thecriteriaarealsodifferentfornewpavementandoverlaydesigns.Fornewpavementdesigns,thethresholdsforIRI,totalrutting,ACrutting,andtop-downfatiguecrackingarerequiredfortheyearstothefirstrehabilitationwhereasthecriteriaforbottom-upfatiguecrackingandthermalcrackingarerequiredfortheentiredesignlifeofthenewpavement.Foroverlaydesigns,allofthecriteriaarerequiredfortheyearstotheendoftheoverlaydesignlife.
Thelocalcalibrationresultswereusedinacomparativeanalysistoillustratetheimpactoflocal calibration on pavement designs for the twopavement design projects inMissouri andColorado.Forboththeprojects,flexiblepavementdesignswereconductedusingtheglobalandlocalcalibrationcoefficients. Inaddition, JPCPdesignswerealsoconductedfortheseprojectsforcomparison.
For the project in Missouri, the new flexible pavement designs were governed by thepredicted bottom-up cracking distresses. The bottom-up cracking model was not adjustedduring local verification as its predictions agreed with the early age performance of deep-strength flexible pavements. Thus, the final thicknesses were the same for new flexiblepavementsusingglobalandlocalcalibrationcoefficients.ForthenewJPCPdesigns,thedesignusingtheglobalcalibrationcoefficientswas0.5in.thickerthanthatusingthelocalcalibrationcoefficients.
FortheprojectinColorado,thelocalcalibrationcoefficientsyieldeda1-in.thinnerasphaltstructure but the same Portland cement concrete structure as the global calibrationcoefficients.ThenewflexiblepavementdesignsfortheColoradoprojectweredeterminedbythe predicted bottom-up cracking results. The design failed the rutting performance criteria,
Tran,Robbins,Rodezno,andTimm
25
butitwasacceptedbyCDOTasruttingwasnotaperformanceissueinsimilarpavementsinthearea.
The above comparative analysis was conducted for one type of aggregate base andsubgrade;furtheranalysiswasconductedtodeterminetheeffectofbaseandsubgradesupportonasphaltpavementdesign,anditispresentedinthenextsection.
5 EFFECTOFFOUNDATIONSUPPORT
Aflexiblepavementstructuretypicallyconsistsofsurface,base,andsubbaselayersplacedonsubgrade to carry traffic loads. The performance of a flexible pavement relies upon theperformance of individual layers. Since the stresses induced by traffic loads in a pavementstructure arehighest in the top layers anddecreasewithdepth, higherqualitymaterials aregenerallyusedintheupperlayers,andlowerqualitymaterialsareusedinthelowerlayers.
At the bottom, the subgrade provides a platform and supports the pavement structure.When subgrade is of lowquality andproperbasematerial is not locally available, a subbaselayer is often required. Constructed on top of the subgrade, the aggregate subbase can beunboundor treatedwith cement, lime,or fly ash to improve its strength characteristics. Thesubbaselayercontributestothestructuralcapacityofthepavement,preventsintrusionoffinegrainedsubgradesoils intothebase layer,minimizesthedamagingeffectof frostaction,andprovidesdrainageforfreewaterthatmayenterthepavementstructure.
Thebaselayerisplacedonthesubbaseordirectlyonthesubgradetoprovidesupportforthe surfacecourseandother functions similar to thoseof the subbase.Highqualitymaterialsuch as gravel and crushed stone is normally used for this layer. The material may also bestabilizedwithcement,lime,orotheradmixtures.
Builtontopofthebaselayer,thesurfacelayerconsistsofoneormoreasphaltlayersthatprovide structural capacity to support traffic loads, distribute the loads to the lower layers,minimizewaterinfiltration,andprovideasmoothandskidresistantsurface.
The performance of an asphalt pavement is usually based upon observations of surfacedistresses, which can be due to deficiencies in the surface layers, but very often, they arecaused by deficiencies in the underlying layers. Table 18 shows how a foundation-relatedproblemcancauseothersurfacedistressesforanasphaltpavement(17).Table18.SurfaceDistressesPotentiallyCausedbyFoundation-RelatedProblems(17)Foundation-RelatedProblems
SurfaceDistressesResultedfromFoundationProblemsCracking Rutting Corrugation Bumps Depressions Potholes Roughness
InsufficientStrength x x x x xMoisture/Drainage x x x x xFreeze/Thaw x x x x x xSwelling x xContamination x x x x
Tran,Robbins,Rodezno,andTimm
26
ThePavementMEDesignsoftwareallowsanasphaltpavement foundationsupport tobedesigned with unbound aggregate or chemically stabilized materials for base and subbaselayersandallowsfortheselectionofdifferentsoiltypesforsubgrade.Inthefollowingsections,a brief summary of foundationmaterials inputs required in the PavementMEDesign is firstpresented followed by a sensitivity analysis to illustrate the effect of foundation support onasphaltpavementdesign.
5.1 FoundationMaterialsInputsRequiredinPavementMEDesign
5.1.1 InputsforUnboundandSubgradeLayers
The hierarchical design approach in the ME Design software also applies to foundationmaterials inputs. The required inputs for unbound layers and subgrade include resilientmodulus,physical/engineeringproperties(suchassoilclassification,moisturecontent,anddrydensity),andhydraulicproperties.Eventhoughthree input levelswereplanned,onlyLevel2and3inputsareallowedfortheselayersinthecurrentPavementMEDesignsoftware(Version2.0). Table 19 summarizes the properties required for Level 2 and 3 inputs for unboundaggregateandsubgradesoils.
Table19InputsforUnboundAggregateandSubgradeSoils(4)Property Description Level2 Level3ResilientModulusofUnboundLayersandSubgradeMr MeasuredorEstimatedResilientModulus ü ü CBR CaliforniaBearingRatio ü R R-value ü ai LayerCoefficient ü DCP DynamicConePenetrationIndex ü PI PlasticityIndex ü P200 PercentagePassingNo.200Sieve ü Soilclass AASHTOorUSCSSoilClass ü ν Poisson’sRatio ü ü PhysicalPropertiesGs SpecificGravity Defaultγdmax MaximumDryDensity Defaultwopt OptimumMoistureContent DefaultPI PlasticIndex ü ü HydraulicPropertiesaf,bf,cf,hr SoilWaterCharacteristicCurveParameters DefaultKsat SaturatedHydraulicConductivity(Permeability) DefaultThe resilient modulus (Mr) input is required in the ME Design software to calculate thestructural response of the pavement. It can bemeasured directly from laboratory testing orobtainedthroughtheuseofcorrelationswithothermaterialstrengthpropertiessuchasCBRandR-value (Level2)orbyusing typicalvalues (Level3).Table20showsmodels thatcanbeusedtoestimateMrbasedonothermaterialandstrengthpropertiesforLevel2inputs.Listed
Tran,Robbins,Rodezno,andTimm
27
inTable21aretypicalresilientmoduli forunboundgranularandsubgradeforLevel3 inputs.OtherphysicalpropertiesforLevel2and3inputsareshowninTable22.Table20ModelsforEstimatingLevel2MrInputs(4)
Property Model TestStandardCBR Mr(psi)=2555(CBR)0.64 AASHTOT193-CaliforniaBearingRatioR-Value Mr(psi)=1155+555R AASHTOT190-ResistanceR-value
AASHTOLayerCoefficient
⎟⎠
⎞⎜⎝
⎛=0.14a
30000(psi)M iR
ai=LayerCoefficient
1993AASHTOGuidefortheDesignofPavementStructures
PIandGradation1 )(728.0175
wPICBR
+=
wPI=P200*PI
AASHTOT27-SieveAnalysisofAggregatesAASHTOT90-PlasticLimitandPlasticityIndexofSoils
DCP1 12.1292
DCPCBR = ASTMD6951-DynamicConePenetrometer
1EstimatesofCBRtoestimateMr.Table21Level3MrForUnboundGranularandSubgradeatOptimumMoistureContent(4)AASHTOSoilsClassification Base/Subbase EmbankmentandSubgrade
A-1-a 40,000 29,500A-1-b 38,000 26,500A-2-4 32,000 24,500A-2-5 28,000 21,500A-2-6 26,000 21,000A-2-7 24,000 20,500A-3 29,000 16,500A-4 24,000 16,500A-5 20,000 15,500A-6 17,000 14,500A-7-5 12,000 13,000A-7-6 8,000 11,500
Table22OtherLevel2and3InputsforUnboundLayersandSubgrade(4)
Property RecommendedInputSpecificGravity
EstimatedUsingGradation,PlasticityIndex,andLiquidLimit
MaximumDryDensityOptimumMoistureContentSaturatedHydraulicConductivitySoilWaterCharacteristicCurveParameters SelectedBasedonAggregate/SubgradeClass
Inaddition,theManualofPractice(4)recommendsthatwhengranularbase/subbaselayersare used, the resilientmodulus of these layers be a function of the resilientmodulus of theunderlying layers, includingsubgradelayers.Theinitialresilientmodulusofthegranular layershould not exceed three times the resilient modulus of the supporting layers to avoid
Tran,Robbins,Rodezno,andTimm
28
decompactionandtensilestressesintheunboundlayers.Figure8showsthemaximumresilientmodulusofanunboundmateriallayerasafunctionofitsthicknessandtheresilientmodulusofthesupportinglayers.
Figure8LimitingModulusCriteriaofUnboundAggregateBaseandSubbaseLayers(4)
5.1.2 InputsforStabilizedLayers
Someunboundmaterialsandsubgradesoilssusceptibletofluctuationsinstrengthandstiffnesspropertiesmay require stabilization due to fluctuations inmoisture content. TheMEDesignsoftware allows inputs for chemically stabilized layers, including elastic/resilient modulus,flexural strength, and physical and thermal properties. Table 23 summarizes the propertiesrequired for each stabilized material at Levels 1, 2, and 3. Table 24 summarizes therecommendedtestprotocolsandrelationshipsforLevel1and2inputs.
Tran,Robbins,Rodezno,andTimm
29
Table 25 summarizes the recommended input values for elastic/resilient modulus andflexuralstrengthwhenaLevel3analysis isused.Finally,Table26presentstherecommendedunit weight and thermal properties values for Level 2 and 3 inputs for chemically stabilizedlayers.Table23ChemicallyStabilizedMaterialsInputRequirements
MaterialType PropertyLevel
1 2 3LeanConcreteandCementTreatedAggregate
ElasticModulus ü ü ü FlexuralStrength ü ü ü
LimeCement-FlyAshElasticModulus ü ü FlexuralStrength ü ü ü
SoilCementElasticModulus ü ü FlexuralStrength ü ü ü
LimeStabilizedSoilElasticModulus ü ü ü FlexuralStrength ü ü
All
UnitWeight ü ThermalConductivity ü Poisson’sRatio ü HeatCapacity ü
Table 24 Recommended Test Protocols and Relationships for ElasticModulus and FlexuralStrengthvaluesforChemicallyStabilizedLayers(Level1and2Inputs)
DesignType Material Level1–
Modulus
Level1–FlexuralStrength
Level2–Relationshipfor
Modulus
Level2–RelationshipforFlexuralStrength
New
LeanConcreteandcementtreatedaggregate
ASTMC469 AASHTOT97 E=57,000(f’c)0.5
Forf’c:AASHTOT22
Use20%oftheCompressiveStrength(LabSamplesorCores)
Lime-Cement-FlyAsh N/A AASHTOT97 E=500+qu
Forqu:ASTMC593
SoilCement N/A ASTMD1635 E=1200(qu)Forqu:ASTMD1633
LimeStabilizedSoil AASHTOT307 N/A Mr=0.124(qu)+9.98
Forqu:ASTMD5102
Existing All
ModulusfromFWDAASHTOT256andASTMD5858
N/A SameasNewDesign SameasNewDesign
Tran,Robbins,Rodezno,andTimm
30
Table 25 Recommended Elastic/Resilient Modulus and Flexural Strength Values forChemicallyStabilizedLayersforLevel3Analysis
Material Modulus(psi) Material FlexuralStrength(psi)LeanConcrete,E 2,000,000 ChemicallyStabilizedMaterial
UsedasBase 750CementStabilizedBase,E 1,000,000SoilCement,E 750,000 ChemicallyStabilizedMaterial
UsedasSubbaseorSubgrade 250LimeStabilizedSoil,Mr 45,000Table26RecommendedUnitWeightandThermalPropertiesValuesforLevel2and3InputsforChemicallyStabilizedLayers
RequiredInput RecommendedValueUnitWeight DefaultValue150pcfThermalConductivity DefaultValue1.25BTU/h-ft°FHeatCapacity DefaultValue0.28BTU/lb-°F5.1.3 EffectofFoundationSupportonPavementMEDesignResults
Inthissection,asensitivityanalysisispresentedtoevaluatetheeffectoffoundationsupportonthePavementMEDesignresults.TheCDOTnewflexiblepavementdesignwithlocalcalibrationcoefficientswasutilizedforthisanalysis.Table27showsthepavementstructureandmaterialinputsthatwerevariedinthesensitivityanalysis.Theinputsforthebaseandsubgradelayerswereselectedtocoverthetypicalrangesofthesematerialsdiscussedintheprevioussections.Otherinputs,includingtrafficinputs,climaticinputs,reliabilitylevels,andperformancecriteria,werekeptthesameasthoseintheCDOTpavementdesignshownearlier.Table27.InputsforSensitivityAnalysistoEvaluateEffectonFoundationSupport
PavementStructure Material InputsforSensitivityAnalysis
Layer1:AsphaltConcrete StoneMatrixAsphalt,PG76-28 Thickness:2in.
Level1InputsLayer2:AsphaltConcrete Superpave,PG64-22 ThicknessOptimized
Level1Inputs
Layer3:Base
UnboundandCementStabilized(CTB)ClassificationOtherPropertiesforUnboundandCTB
Thickness:6&12in.Mr(Unbound):30,000;40,000&50,000psi;Mr(CementTreatedBase):100,000psi
Layer4:Subgrade
Classification:A-7-4,A-4andA-2-4Gradation&OtherPropertiesforA-7-6,A-4andA-2-4
Thickness:120in.(toBedrock)Mr:10,000;15,000&20,000psi
Layer5:Bedrock HighlyFracturedandWeathered Semi-infinite
E=500,000psiTable 28 shows results of the sensitivity analysis to evaluate the effect of foundation
support on the thickness design, which in this case is the thickness of Layer 2 (Superpaveasphaltbase).Basedontheanalysisresults,thefollowingobservationscanbedrawn:
Tran,Robbins,Rodezno,andTimm
31
• Asthestiffnessofsubgradeincreasedfrom10,000psito20,000psiandthestiffnessoftheunboundaggregatebasewas selectedasa functionof the subgrade stiffnessandthe thickness of the aggregate base layer (Figure 8), the thickness of the asphaltstructurewasreducedby1inch.
• Whenincreasingthethicknessoftheaggregatebasefrom6inchesto12inches(allelseequal), the thickness of the asphalt structure changed by 0.5 inches for only onescenario with subgrade resilient modulus of 15,000 psi and unbound base resilientmodulusof40,000psi.
• Comparedtoa6-inchunboundaggregatebase,the6-inchstabilizedbasecouldyielda1.5-inch,1.0-inch,and0.5-inchthinnerasphaltstructurefor10,000psi,15,000psi,and20,000psisubgrade,respectively.
• Thethicknessoftheasphaltstructuredecreasedby3incheswhenthethicknessofthecementstabilizedbaseincreasedfrom6inchesto12inches.
• Allof thedesignswith theunboundaggregatebaseweregovernedby thebottom-upfatiguecracking,andallofthedesignswiththestabilizedbaseweregovernedbytheIRIcriteria.Table29andTable30showthefinalpredictionresultsforScenarios1and12.Eventhoughtheruttingpredictions failedthedesigncriteria, the locallycalibratedACmodelswerefoundbyCDOTtoover-predicttheAClayerruttinginthefield.
Table28.EffectonFoundationSupportonLayer2Thickness
ScenarioLayer5:Bedrock
Layer4:Subgrade
Layer3:Base
Layer2:Superpave
Layer1:SMA Design
GovernedByE(psi) H(in) Mr(psi) H(in) Mr(psi) H(in) H(in)
UnboundAggregateBase1 500,000 120 10,000 6 30,000 8.5 2 Bottom-up2 500,000 120 15,000 6 40,000 8.0 2 Bottom-up3 500,000 120 20,000 6 50,000 7.5 2 Bottom-up4 500,000 120 10,000 12 30,000 8.5 2 Bottom-up5 500,000 120 15,000 12 40,000 7.5 2 Bottom-up6 500,000 120 20,000 12 50,000 7.5 2 Bottom-up CementStabilizedBase7 500,000 120 10,000 6 100,000 7.0 2 IRI8 500,000 120 15,000 6 100,000 7.0 2 IRI9 500,000 120 20,000 6 100,000 7.0 2 IRI10 500,000 120 10,000 12 100,000 4.0 2 IRI11 500,000 120 15,000 12 100,000 4.0 2 IRI12 500,000 120 20,000 12 100,000 4.0 2 IRI
Tran,Robbins,Rodezno,andTimm
32
Table29.DistressPredictionResultsforScenario1SensitivityAnalysis
DistressTypeDistressatSpecified
Reliability Reliability(%) CriterionSatisfied?
Target Predicted Target AchievedTerminalIRI(in./mile) 160.00 149.06 90.00 94.94 PassPermanentDeformation-TotalPavement(in.) 0.40 0.53 90.00 53.54 FailACBottom-upFatigueCracking(percent) 10.00 9.96 90.00 90.10 PassACThermalCracking(ft/mile) 1500.00 494.69 90.00 100.00 PassACTop-downFatigueCracking(ft/mile) 2000.00 297.51 90.00 100.00 PassPermanentDeformation-ACOnly(in.) 0.25 0.41 90.00 40.63 FailTable30.DistressPredictionResultsforScenario12SensitivityAnalysis
DistressTypeDistressatSpecified
Reliability Reliability(%) CriterionSatisfied?
Target Predicted Target AchievedTerminalIRI(in./mile) 160.00 151.61 90.00 93.99 PassPermanentDeformation-TotalPavement(in.) 0.40 0.65 90.00 19.51 FailACBottom-upFatigueCracking(percent) 10.00 1.28 90.00 100.00 PassTotalCracking(Reflective+Fatigue)(percent) 15 4.48 - - PassACThermalCracking(ft/mile) 1500.00 189.22 90.00 100.00 PassACTop-downFatigueCracking(ft/mile) 2000.00 645.01 90.00 100.00 PassPermanentDeformation-ACOnly(in.) 0.25 0.61 90.00 3.70 FailChemicallyStabilizedLayer-FatigueFracture(percent) 25.00 0.84 - - -
Tran,Robbins,Rodezno,andTimm
33
6 PERFORMANCECRITERIAANDRELIABILITY
6.1 PerformanceCriteriaandReliabilityLevels
AsfortheempiricalAASHTOGuides,theMEPDGallowstheusertosettheperformancecriteriaandreliabilitylevelstoevaluatetheadequacyofeachdesign.Table31andTable32showtheperformancecriteriaandreliability levelsrecommendedintheManualofPractice(4).AbriefdescriptionofthereliabilityconceptemployedintheMEPDGispresentedbelow.Table31.PerformanceCriteriaRecommendedforFlexiblePavementDesign(4)
PerformanceIndicator
MaximumValueatEndofDesignLife Performance
IndicatorMaximumValueatEndofDesignLife
TerminalIRI(Smoothness)
Interstate:160in/miPrimary:200in/miSecondary:200in/mi
FatigueCracking
Interstate:10%laneareaPrimary:20%laneareaSecondary:35%lanearea
TotalRutting(InWheelPaths)
Interstate:0.40in.Primary:0.50in.Others(<45mph):0.65in.
Transverse(Thermal)Cracking
Interstate:500ft/miPrimary:700ft/miSecondary:700ft/mi
Table32.LevelsofReliabilityforDifferentFunctionalClassificationsoftheRoadway(4)
FunctionalClassification
LevelofReliabilityUrban Rural
Interstate/FreewaysPrincipalArterialsCollectorsLocal
95908075
95857570
In the PavementMEDesign software, reliability is applied to the individual performance
indicators, which include total rutting, fatigue cracking, thermal cracking, IRI, and otherperformance predictions including longitudinal cracking, reflective cracking, and AC rutting.Reliability(R)isdefinedastheprobability(P)thateachofthekeydistresstypesandIRIwillbelessthanaselectedcriticalleveloverthedesignperiod,asshowninEquation1.
R=P[DistressoverDesignPeriod<CriticalDistressLevel] (1)
For each performance indicator, the PavementME Design software first determines themean prediction using the corresponding performance (transfer) model. The software thenincreasesthepredictionbythepredictionreliability,whichisdependentonthestandarderroroftheperformancemodeldeterminedwhenitwascalibratedandthereliabilitylevelselected.Thisprocedureisutilizedforalltheperformanceindicators.
Tran,Robbins,Rodezno,andTimm
34
6.2 EffectofPerformanceCriteriaandReliabilityonPavementDesign
Since the performance criteria and design reliability greatly affect the final design, initialconstruction costs, long-term performance, and life cycle costs, they should be carefullyselectedinbalancewitheachother.Forthesamesetofperformancecriteria,alowerreliabilitylevelwouldyield lowerpredicteddistresses, resulting ina thinnerpavement section. For thesame design reliability level, a set of criteria with lower allowable cracking, rutting, and IRIwould result in a thicker pavement section. For example, sinceMoDOT set lower allowablecrackingandrutting,itsreliabilitylevelwassetlowerthanthatofCDOT.
An analysis was conducted in this study to evaluate the sensitivity of pavement designthickness to the performance criteria and reliability levels recommended in the Manual ofPractice (4). The sensitivity analysis was conducted for the CDOT’s new flexible pavementstructure shown in Table 12 using CDOT’s local calibration coefficients. Itwas conducted forfour roadway classifications including interstate, principal arterial, minor arterial, and majorcollector.ThedesignlifeandtrafficinputsselectedforeachroadwayclassificationareshowninTable33.Table33.DesignLifeandTrafficInputsSelectedforSensitivityAnalysisDesignInputs Interstate PrincipalArterial MinorArterial MajorCollectorsDesignLife,years 20 20 20 20TrafficInputs
Two-wayAADTT 11500 8640 4320 1920No.ofLanes 3 3 2 1TrucksinDesignDirection 50% 50% 50% 60%TrucksinDesignLane 60% 60% 90% 100%OperationalSpeed,mph 70 65 55 45GrowthRate 1.43% 1.43% 1.43% 1.43%
Table 34 shows the performance criteria and reliability levels used in the analysis. TheperformancecriteriaweresetbasedontherecommendationsintheManualofPractice(4),asshowninTable31,andthoseutilizedbyCDOT,asshowninTable8.Also,showninTable34aretheperformancecriteriafortheunboundlayersthatwereobtainedbysubtractingpermanentdeformation in the AC layer from the total permanent deformation in the pavement. Thesecriteriaarethesame(0.15in.)foralltheroadwayclassifications.
AsshowninTable34,eachdesignwasconductedatseveralreliabilitylevelsstartingfrom50%tothereliabilitylevelrecommendedforeachroadwayclassificationshowninTable32.Inaddition, the AC thicknesswas varied for each design, which also includes a 6-in. aggregatebase,toevaluatethesensitivityofACthicknesstotheperformancecriteriaandreliabilitylevelsasfollows:
• Interstatedesign:9–12in.AC,including2in.SMA• Principalarterialdesign:7.5–10.5in.AC,including2in.SMA• Minorarterialdesign:7–9.5in.AC,including2in.SMA• Majorcollectordesign:6.5–9.5in.AC,including2in.SMA
Tran,Robbins,Rodezno,andTimm
35
Table34.PerformanceCriteriaandReliabilityLevelsSelectedforSensitivityAnalysisPerformanceCriteria Interstate PrincipalArterial MinorArterial MajorCollectors
Limit Reliability Limit Reliability Limit Reliability Limit ReliabilityNewFlexible InitialIRI(in./mi) 50 50 50 50 TerminalIRI(in./mi) 160 50—95 200 50—90 200 50—85 200 50—80
ACTop-downFatigueCracking(ft/mile) 2,000 50—95 2,500 50—90 3,000 50—85 3,000 50—80
ACBottom-upFatigueCracking(percent) 10 50—95 25 50—90 35 50—85 35 50—80
ACThermalCracking(ft/mile) 1,500 50—95 1,500 50—90 1,500 50—85 1,500 50—80
PermanentDeformation-TotalPavement(in.) 0.4 50—95 0.5 50—90 0.65 50—85 0.65 50—80
PermanentDeformation-ACOnly(in.) 0.25 50—95 0.35 50—90 0.5 50—85 0.5 50—80
PermanentDeformation-Unbound(in.) 0.15 50—95 0.15 50—90 0.15 50—85 0.15 50—80
Tran,Robbins,Rodezno,andTimm
36
Based on the analysis, two types of pavement distresses, including bottom-up fatiguecrackingandpermanentdeformationintheunboundlayers,werefoundtobemoresensitivetothechangesinpavementdesignthickness.Thus,thesensitivityofthesepredicteddistressestothechangesinpavementdesignthicknessisfurtherdiscussedinthefollowingsubsections.
6.2.1 SensitivityofPermanentDeformationinUnboundLayerstoPavementDesignThickness
The PavementMEDesign software predicted total permanent deformation in the pavementand permanent deformation in the AC layer. Based on this information, the predictedpermanent deformation in the unbound layers, including aggregate base and subgrade, wasobtained by subtracting the permanent deformation in the AC from the total permanentdeformationinthepavement.Figure9throughFigure12showstheeffectofACthicknessonpredicted rutting in theunbound layers for the four roadway classifications. Thedashed lineshown in these figures is the design limit formaximum rutting in the unbound layers. Sinceeachdesignwasalsoconductedatseveralreliabilitylevels,thesefiguresalsoshowtheeffectofreliabilitylevelonpredictedruttingintheunboundlayers.Basedontheseresults,thefollowingobservationscanbeoffered:
• TheeffectofACthicknesswasslightlyhigherforthepavementswiththinnerAClayers.However,thedifferenceinruttingintheunboundlayersbecamepracticallyinsignificantwhentheACthicknessincreasedabove9inches,approachingthedesignthicknessofaperpetualasphaltpavement.
• Theimpactofreliabilitylevelonpredictedruttingintheunboundlayerswasminimalforeachdesign.
• The predicted permanent deformation in the unbound layers passed the design limit(0.15in.),suggestingthatthedesignAClayercouldbethinner,thatthedesignlimitforrutting in theunbound layerscouldbe lower,or that thedesignwasgovernedby theotherdistresstype.
Figure9EffectofACThicknessonPredictedRuttinginUnboundLayers(Interstate)
0.08
0.10
0.12
0.14
0.16
9 9.5 10 10.5 11 11.5 12
Unb
ound
Ru@
ng(Inche
s)
ACThickness(Inches)
UnboundRuUng(95%Rel)UnboundRuUng(90%Rel)UnboundRuUng(80%Rel)UnboundRuUng(70%Rel)UnboundRuUng(60%Rel)UnboundRuUng(50%Rel)UnboundRuUngLimit
Tran,Robbins,Rodezno,andTimm
37
Figure10EffectofACThicknessonPredictedRuttinginUnboundLayers(PrincipalArterial)
Figure11EffectofACThicknessonPredictedRuttinginUnboundLayers(MinorArterial)
Figure12EffectofACThicknessonPredictedRuttinginUnboundLayers(MajorCollector)
0.08
0.10
0.12
0.14
0.16
7.5 8 8.5 9 9.5 10 10.5
Unb
ound
Ru@
ng(Inche
s)
ACThickness(Inches)
UnboundRuUng(90%Rel)UnboundRuUng(85%Rel)UnboundRuUng(80%Rel)UnboundRuUng(70%Rel)UnboundRuUng(60%Rel)UnboundRuUng(50%Rel)UnboundRuUngLimit
0.08
0.10
0.12
0.14
0.16
7 7.5 8 8.5 9 9.5
Unb
ound
Ru@
ng(Inche
s)
ACThickness(Inches)
UnboundRuUng(90%Rel)UnboundRuUng(80%Rel)UnboundRuUng(70%Rel)UnboundRuUng(60%Rel)UnboundRuUng(50%Rel)UnboundRuUngLimit
0.08
0.10
0.12
0.14
0.16
6.5 7 7.5 8 8.5 9 9.5
Unb
ound
Ru@
ng(Inche
s)
ACThickness(Inches)
UnboundRuUng(80%Rel)UnboundRuUng(70%Rel)UnboundRuUng(60%Rel)UnboundRuUng(50%Rel)UnboundRuUngLimit
Tran,Robbins,Rodezno,andTimm
38
6.2.2 SensitivityofBottom-UpFatigueCrackingtoPavementDesignThickness
Figure13throughFigure16showstheeffectofACthicknessandreliabilitylevelonthebottom-upfatiguecrackingusingCDOT’slocalcalibrationcoefficients.Thedashedlinesinthesefiguresrepresent the performance criteria listed in Table 34. Based on these results, the followingobservationscanbedrawn:
• TheeffectofACthicknessonpredictedbottom-upfatiguecrackingwasmoresignificantforthepavementswiththinnerAClayers.Also,theeffectofACthicknessonpredictedbottom-up fatigue crackingwasmore significant than thatonpredicted rutting in theunboundlayers.
• The impact of reliability level on predicted bottom-up fatigue cracking was moreprofoundthanthatonpredictedruttingintheunboundlayers.
Figure13EffectofACThicknessonPredictedBottom-UpCracking(Interstate)
Figure14EffectofACThicknessonPredictedBottom-UpCracking(PrincipalArterial)
0
10
20
30
40
50
60
9 9.5 10 10.5 11 11.5 12
FaMg
ueCracking(%
)
ACThickness(Inches)
Bo[om-UpCracking(95%Rel)Bo[om-UpCracking(90%Rel)Bo[om-UpCracking(80%Rel)Bo[om-UpCracking(70%Rel)Bo[om-UpCracking(60%Rel)Bo[om-UpCracking(50%Rel)Bo[om-UPCrackingLimit
0
10
20
30
40
50
60
7.5 8 8.5 9 9.5 10 10.5
FaMg
ueCracking(%
)
ACThickness(Inches)
Bo[om-UpCracking(90%Rel)Bo[om-UpCracking(85%Rel)Bo[om-UpCracking(80%Rel)Bo[om-UpCracking(70%Rel)Bo[om-UpCracking(60%Rel)Bo[om-UpCracking(50%Rel)Bo[om-UPCrackingLimit
Tran,Robbins,Rodezno,andTimm
39
Figure15EffectofACThicknessonPredictedBottom-UpCracking(MinorArterial)
Figure16EffectofACThicknessonPredictedBottom-UpCracking(MajorCollector)
Figure17showsthelevelsofreliabilitythatcouldbeachievedforvariousACthicknessesin
this analysis if the performance limit for bottom-up fatigue cracking was set at 10% forinterstates, 25% for principal arterials, and 35% for minor arterials andmajor collectors, asshowninTable34.Basedontheseresults,thefollowingobservationscanbedrawn:
• For the interstatedesign (Figure17),when the reliability level achievedwasbetween80%and90%,a0.5-in. increase inAC thicknesswould improve the reliability levelbyapproximately 10%. When the reliability level achieved was around 90%, a 0.5-in.increase inAC thicknesswouldonly improve the reliability levelbyapproximately5%.When the reliability level achievedwas above 95%, a 0.5-in. increase in AC thicknesswould have a minimum effect on the design reliability level. This suggests that thereliabilitylevelforinterstatescouldbesetbetween80%and95%.
• For the principal arterial design (Figure 17), the design AC thickness would not beaffectedwhenthereliabilitylevelwasselectedbetween80%and90%.
0
10
20
30
40
50
60
7 7.5 8 8.5 9 9.5
FaMg
ueCracking(%
)
ACThickness(Inches)
Bo[om-UpCracking(90%Rel)Bo[om-UpCracking(80%Rel)Bo[om-UpCracking(70%Rel)Bo[om-UpCracking(60%Rel)Bo[om-UpCracking(50%Rel)Bo[om-UPCrackingLimit
0
10
20
30
40
50
60
6.5 7 7.5 8 8.5 9 9.5
FaMg
ueCracking(%
)
ACThickness(Inches)
Bo[om-UpCracking(80%Rel)Bo[om-UpCracking(70%Rel)Bo[om-UpCracking(60%Rel)Bo[om-UpCracking(50%Rel)Bo[om-UPCrackingLimit
Tran,Robbins,Rodezno,andTimm
40
• Fortheminorarterialandmajorcollectordesign(Figure17),theACthicknessforeachdesign would be the same when the reliability level was selected between 70% and90%.
Figure17ReliabilityLevelversusACThicknessatSelectedPerformanceCriteria
6.2.3 ProposedPerformanceCriteriaandReliabilityLevelsforPavementMEDesign
Table35 lists theproposedperformancecriteriaand reliability levels thatcanbeused in thefuture forPavementMEDesign.Thesevalueswereselectedbasedon thesensitivityanalysisresults and considering the performance criteria and reliability levels recommended in theManual of Practice, shown in Table 31 and Table 32, and those adoptedby CDOT, shown inTable8.Furtherjustificationfortheproposeddesigncriteriaandreliabilitylevelsfollow:
• TheproposedreliabilitylevelsaresimilartothoseintheManualofPractice(Table32),exceptforruralinterstatewhenareliabilitylevelof90%wasfoundtobeappropriateinthesensitivityanalysis.
• Theproposedperformancelimitsforbottom-upfatiguecrackingarethesameasthoseadoptedbyCDOT.
• The proposed performance criteria for rutting in the lower layers were determinedbased on the total rutting andAC layer rutting criteria adopted by CDOT, except forinterstatewhere a rut depth of 0.1 in.was found to be achievable in the sensitivityanalysis.However, it shouldbenoted that thesecriteriawouldonlybeutilizedwhentheruttingmodelsfortheunboundlayershavebeenproperlycalibrated.Thenationallycalibratedmodelshavebeenfoundtoover-predictruttingintheunboundlayers.
0
10
20
30
40
50
60
70
80
90
100
6.5 7 7.5 8 8.5 9 9.5 10 10.5 11 11.5 12
Reliability(%
)
ACThickness(Inches)
Interstate(10%Cracking)
PrincipalArterial(25%Cracking)
MinorArterial(35%Cracking)
MajorCollector(35%Cracking)
Tran,Robbins,Rodezno,andTimm
41
• For the other distresses in Table 35,when the distresses predicted by PavementMEDesign are greater than the performance criteria, the designer may need to verifywhetherthe levelsofdistresspredictedareobserved inthefieldor ifothermaterialsmaybeused to avoid these typesof distress. In general, increasing theAC thicknesswouldnotreducethesedistresseseffectively.Forexample,athickerAClayerwouldnotresultinlowerruttingintheAClayer.Rather,choosinganasphaltbinderwithalowerlowtemperaturegradewouldbemoreeffective inmitigatingthermalcrackingthanathickerAClayer.
Table35.ProposedPerformanceCriteriaandReliabilityLevelsforFutureDesign
Classification
Reliability(%) PerformanceCriteria
Urban RuralTerminal
IRI(in/mi)
RuttingTotal(in)
RuttingAC(in)
RuttingUnbound
(in.)
Top-Down(ft/mi)
Bottom-Up
(%lane)
ThermalCracking(ft/mi)
Interstate 95 90 160 0.5 0.4 0.1 2,000 10 1,500PrincipalArterials 90 85 200 0.5 0.35 0.15 2,500 25 1,500MinorArterial 80 75 200 0.65 0.5 0.15 3,000 35 1,500MajorCollectors 75 70 200 0.65 0.5 0.15 3,000 35 1,500
7 CONCLUSIONSANDRECOMMENDATIONS
State agencies have considered implementing PavementMEDesign to replace the empiricalPavementDesignGuides.Amongotherelements,theirimplementationplansoftenincludetwoimportantsteps—localcalibrationandselectionofperformancethresholdsandreliabilitylevelsforacceptingfuturedesigns.PaststudieshaveshownthatwithoutlocallycalibratingPavementMEDesign,usingthemoresophisticatedsoftwarewouldnotyieldbetterdesigns.Recognizingthe importanceof thesesteps, this reportpresents twocasestudies thatcomparepavementdesignsconductedwithglobalandlocalcalibrationcoefficientstoillustratetheimportanceofconductinglocalcalibrationofPavementMEDesignintheimplementationprocess.Inaddition,it discusses sensitivity analyses that show the effect of foundation support, performancecriteria,andreliabilitylevelsonpavementdesign.
MoDOTandCDOThavecompletedtheirlocalcalibrationprocessesandadoptedPavementME Design for routine pavement designs. The local calibration coefficients were used in acomparativeanalysisto illustratethe impactof localcalibrationonpavementdesignsfortwopavement design projects in Missouri and Colorado. For both projects, flexible pavementdesigns were conducted using the global and local calibration coefficients. In addition, JPCPdesignswerealsoconductedfortheseprojectsforcomparison.Thefollowingconclusionscanbeofferedfromthiscomparativeanalysis:
• For theproject inMissouri, thenew flexiblepavementdesignsweregovernedby thepredictedbottom-upcrackingdistresses.Thebottom-upcrackingmodelwas found tobeappropriateforpavementsinMissouri,anditscoefficientswerenotadjustedduringlocalverification.Thus,thefinalthicknesseswerethesamefornewflexiblepavementsusingglobalandlocalcalibrationcoefficients.ForthenewJPCPdesigns,thedesignusing
Tran,Robbins,Rodezno,andTimm
42
theglobalcalibrationcoefficientswas0.5in.thickerthanthatusingthelocalcalibrationcoefficients.
• For the project in Colorado, the local calibration coefficients yielded a 1-in. thinnerasphalt structure but the same Portland cement concrete structure as the globalcalibration coefficients. The final thickness designs were selected based on thepredictedbottom-upcrackingresults.Thesedesignsfailedtheruttingcriteria.However,ruttingwas not a performance issue in similar pavements in the area, so the designswerestillacceptedbyCDOT.
The above comparative analysis was conducted for one type of aggregate base andsubgrade;thus,asensitivityanalysiswaslaterperformedusingCDOT’sdesigntodeterminetheimpactofbaseandsubgradesupportondesignACthickness.Basedontheanalysisresults,thefollowingconclusionscanbeoffered:
• Asthestiffnessofsubgradeincreasedfrom10,000psito20,000psi,thethicknessoftheasphalt structurewould reduce by 1 inch. Also,when increasing the thickness of theaggregate base from 6 inches to 12 inches, the thickness of the asphalt structurechangedby0.5inches.
• Compared to a 6-inchunbound aggregate base, the 6-inch stabilizedbase could yield1.5-inch,1.0-inch,and0.5-inchthinnerasphaltstructurefor10,000psi,15,000psi,and20,000psisubgrade,respectively.Also,thethicknessoftheasphaltstructuredecreasedby3incheswhenthethicknessofthecementstabilizedbaseincreasedfrom6inchesto12inches.
• All the designs with the unbound aggregate base were governed by the bottom-upfatigue cracking, and the designs with the stabilized base were governed by the IRIcriteria.
Another sensitivity analysis was conducted using the CDOT’s new flexible pavementstructure and local calibration coefficients to evaluate the sensitivity of pavement designthickness to the performance criteria and reliability levels. Since these design parametersgreatlyaffectthefinaldesign,theyshouldbecarefullyselectedinbalancewitheachother.Theanalysiswasconducted for four roadwayclassifications including interstate,principalarterial,minorarterial,andmajorcollector.Basedontheanalysis,thereweretwotypesofpavementdistressesthatwerefoundtobemoresensitivetothechangesinpavementdesignthickness,includingbottom-up fatiguecrackingandpermanentdeformation in theunbound layers.Keyfindingsfromthisanalysisareasfollows:
• TheeffectofAC thicknessonpredictedbottom-up fatiguecrackingand rutting in theunbound layers was more significant for thinner pavements. Also, the effect of ACthickness was more significant on bottom-up fatigue cracking than on rutting in theunboundlayers.Theimpactofreliabilitylevelonpredictedbottom-upfatiguecrackingwasalsomoreprofoundthanthatonpredictedruttingintheunboundlayers.
• Fortheinterstatedesign,a0.5-in.increaseinACthicknesswouldimprovethereliabilitylevel from 80% to 90%. Or, it would improve the reliability level from 90% to 95%.However, itwould have aminimumeffect on the design reliability levelwhen itwasabove 95%. For the principal arterial design, the design AC thickness would not beaffectedwhen the reliability levelwas selectedbetween80%and90%.For theminor
Tran,Robbins,Rodezno,andTimm
43
arterialandmajorcollectordesign,theACthicknessforeachdesignwouldbethesamewhen the reliability level was selected between 70% and 90%. This analysis wasconductedwiththeperformancelimitforbottom-upfatiguecrackingbeingsetat10%for interstates, 25% for principal arterials, and 35% for minor arterials and majorcollectors.
• Basedon the sensitivity analysis results and considering theperformance criteria andreliability levels recommended in theManualofPracticeandthoseadoptedbyCDOT,the performance criteria and reliability levels that can be used in the future forPavementMEDesignareproposedinTable35.
Tran,Robbins,Rodezno,andTimm
44
REFERENCES
1. AASHTOGuideforDesignofPavementStructures.AmericanAssociationofStateandHighwayTransportationOfficials,Washington,D.C.,1993.
2. ARA,Inc.,ERESConsultantsDivision.GuideforMechanistic-EmpiricalDesignofNewandRehabilitatedPavementStructures.FinalReport,NCHRPProject1-37A.TransportationResearchBoardoftheNationalAcademies,Washington,D.C.,2004.http://www.trb.org/mepdg/guide.htm
3. Pierce,L.,andG.McGovern.NCHRPSynthesisReport457:ImplementationoftheAASHTOMechanistic-EmpiricalPavementDesignGuideandSoftware.TRB,NationalResearchCouncil,Washington,D.C.,2014.
4. Mechanistic–Empirical Pavement Design Guide, Interim Edition: A Manual of Practice.AASHTO,Washington,D.C.,2008.
5. GuidefortheLocalCalibrationoftheMechanistic-EmpiricalPavementDesignGuide.AASHTO,Washington,D.C.,2010.
6. Timm,D.,X.Guo,M.Robbins,andC.Wagner.M-ECalibrationStudiesattheNCATTestTrack.AsphaltPavementMagazine,Vol.17,No.5,NationalAsphaltPavementAssociation,2012,pp.45-51.
7. Carvalho,R.,andC.Schwartz.ComparisonsofFlexiblePavementDesignsAASHTOEmpiricalVersusNCHRPProject1-37AMechanistic-Empirical.InTransportationResearchRecord:JournaloftheTransportationResearchBoard,No.1947,TransportationResearchBoardoftheNationalAcademies,Washington,D.C.,2006,pp.167-174.
8. Robbins,M.,M.Rodezno,N.Tran,andD.Timm.PavementMEDesign–ASummaryofLocalCalibrationEffortsforFlexiblePavements.NCATReport17-07.NationalCenterforAsphaltTechnology,Auburn,Ala.,2017.
9. Mallela,J.,L.Titus-Glover,H.VonQuintus,M.Darter,M.Stanley,C.Rao,andS.Sadasivam.ImplementingtheAASHTOMechanisticEmpiricalPavementDesignGuideinMissouri,Vol.1StudyFindings,ConclusionsandRecommendations.MissouriDepartmentofTransportation,JeffersonCity,Mo.,2009.
10. Mallela,J.,L.Titus-Glover,S.Sadasivam,B.Bhattacharya,M.Darter,andH.VonQuintus.ImplementationoftheAASHTOMechanisticEmpiricalPavementDesignGuideforColorado.ReportCDOT-2013-4.ColoradoDepartmentofTransportation,Denver,Colo.,2013.
11. MEDesignManual.MissouriDepartmentofTransportation,JeffersonCity,Mo.,2005.http://sp.design.transportation.org/Documents/missouri_plan.pdf
12. CDOT2016M-EPavementDesignManual.ColoradoDepartmentofTransportation,Denver,Colo.,2015.
13. Darter,M.I.,L.Titus-Glover,H.VonQuintus,B.Bhattacharya,andJ.Mallela.CalibrationandImplementationoftheAASHTOMechanistic-EmpiricalPavementDesignGuideinArizona.ReportFHWA-AZ-14-606.ArizonaDepartmentofTransportation,Phoenix,Ariz.,2014.
14. Williams,C.andR.Shaidur.Mechanistic-EmpiricalPavementDesignGuideCalibrationForPavementRehabilitation.ReportFHWA-OR-RD-13-10.OregonDepartmentofTransportation,Salem,Ore.;FederalHighwayAdministration,Washington,D.C.,2013.
Tran,Robbins,Rodezno,andTimm
45
15. Li,J.,L.M.Pierce,andJ.Uhlmeyer.CalibrationofFlexiblePavementinMechanistic–EmpiricalPavementDesignGuideforWashingtonState.InTransportationResearchRecord:JournaloftheTransportationResearchBoard,No.2095,TransportationResearchBoardoftheNationalAcademies,Washington,D.C,2009,pp.73–83.
16. Darter,M.I.,L.T.Glover,andH.L.VonQuintus.DraftUser'sGuideforUDOTMechanistic-EmpiricalPavementDesignGuide.ReportUT-09.11a.UtahDepartmentofTransportation,SaltLakeCity,Utah,2009.
17. Christopher,B.,C.Schwartz,andR.Boudreau.GeotechnicalAspectsofPavements.ReportFHWANHI-05-037.NationalHighwayInstitute,FHWA,Washington,D.C.,2006.
Tran,Robbins,Rodezno,andTimm
46
APPENDIXAPERFORMANCEMODELSFORFLEXIBLEPAVEMENTDESIGN
A.1 Introduction
TheMEPDGincludesseveralperformance(transfer)modelstopredictthefollowingdistresses:• Rutdepth—total,asphaltandunboundlayers(in)• Transverse(thermal)cracking(non-loadrelated)(ft/mi)• Fatigue(bottom-upfatigue)cracking(percentlanearea)• Longitudinal(top-down)cracking(ft/mi)• Reflectivecrackingofasphaltoverlaysoverasphalt,semi-rigid,compositeandconcrete
pavements(percentlanearea)• Internationalroughnessindex(IRI)(in/mi)
These models are presented in this appendix to facilitate the discussion of the localcalibrationresults.TheinformationisadaptedfromtheManualofPracticefortheMEPDG(3)and the AASHTOWare Pavement ME Design software Version 2.0. When discrepancies arefoundbetweenthetworeferences,informationinthesoftwareispresented.
A.2 RutDepthforAsphaltandUnboundLayers
Two performancemodels are used to predict the total rut depth of flexible pavements andasphaltoverlays:onefortheasphalt layersandtheotheroneforallunboundaggregatebaselayers and subgrades. Equation A.1 shows the asphalt rutting model developed based onlaboratoryrepeatedloadplasticdeformationtests.∆!(!")= !!(!")ℎ(!") = !!!!!!!(!")10!!!!!!!!!!!!!!!!! (A.1)
where:
Dp(AC)= Accumulated permanent or plastic vertical deformation in the asphalt layer orsublayer,in;
εp(AC)= Accumulated permanent or plastic axial strain in the asphalt layer or sublayer,in/in;
εr(AC)= Resilientorelasticstraincalculatedbythestructuralresponsemodelatthemid-depthofeachasphaltlayerorsublayer,in/in;
h(AC) = Thicknessoftheasphaltlayerorsublayer,in;n = Numberofaxleloadrepetitions; T = Mixorpavementtemperature,°F; kz = DepthconfinementfactorshowninEquationA.2;
kr1,r2,r3= Global field calibrationparameters(fromtheNCHRP 1-40D recalibration;kr1 = -3.35412,kr2=1.5606,kr3=0.4791);and
βr1,r2,r3= Local or mixture field calibration constants; for the global calibration, theseconstantswereallsetto1.0.
!! = !! + !!! 0.328196! (A.2)
Tran,Robbins,Rodezno,andTimm
47
!! = −0.1039 !!"# ! + 2.4868!!"# − 17.342 (A.3)!! = 0.0172 !!"# ! − 1.7331!!"# + 27.428 (A.4)where:
D = Depthbelowthesurface,in;andH(AC)= Totalasphaltthickness,in.
Equation A.5 shows the field-calibrated transfer function for the unbound layers and
subgrade.
∆! !"#$ = !!!!!!!!ℎ!"#$ !!!!
!!!!! (A.5)
where:
Dp(Soil)= Permanentorplasticdeformationforthelayerorsublayer,in;n = Numberofaxleloadapplications;eo= Intercept determined from laboratory repeated load permanent deformation
tests,in/in; er = Resilientstrainimposedinlaboratorytesttoobtainmaterialproperties εo, β, and
r, in/in; ev = Averageverticalresilientorelasticstraininthelayerorsublayerandcalculatedby
thestructuralresponsemodel,in/in; hsoil = Thicknessoftheunboundlayerorsublayer,in;ksl = Globalcalibrationcoefficients;ks1=1.673forgranularmaterialsand1.35forfine-
grainedmaterials;andβs1 = Local calibration constant for the rutting in the unbound layers; the local
calibrationconstantwassetto1.0fortheglobalcalibrationeffort. !"#$ = −0.6119− 0.017638 !! (A.6)
! = 10! !!!! !"! !
!! (A.7)
!! = !" !!!!!!
!!!!!!
!! = 0.0075 (A.8)
where:
Wc = Watercontent,percent;Wr = Resilientmodulusoftheunboundlayerorsublayer,psi;
Tran,Robbins,Rodezno,andTimm
48
a1,9 = Regressionconstants;a1=0.15anda9=20.0;and b1,9 = Regressionconstants;b1=0.0andb9=0.0.
A.3 Transverse(Thermal)Cracking
The amount of thermal cracking is estimated using Equation A.9 based on the probabilitydistributionofthelogofthecrackdepthtoasphaltlayerthicknessratio.!" = !!"! !
!!!"# !!
!!"# (A.9)
where:
TC = Observedamountofthermalcracking,ft/mi;βt1 = Regressioncoefficientdeterminedthroughglobalcalibration(400);N[z] = Standardnormaldistributionevaluatedat[z];σd = Standarddeviationofthelogofthedepthofcracksinthepavement(0.769),in;Cd = Crackdepth,in;andHAC = Thicknessofasphaltconcretelayers,in.
Thecrackdepth(Cd)inducedbyagiventhermalcoolingcycleisestimatedusingtheParislawofcrackpropagation,asshowninEquationA.10.∆! = ! ∆! ! (A.10)where:
DC = Changeinthecrackdepthduetoacoolingcycle;DK = Changeinthestressintensityfactorduetoacoolingcycle;andA,n= Fractureparametersfor theHMAmixture,whichareobtainedfromthe indirect
tensilecreep-complianceandstrengthoftheasphaltmixtureusingEquationA.11.! = 10!!!! !.!"#!!.!"!"# !!"!!! (A.11)where:
m = 0.8 1+ !! ;
kt = Coefficientdeterminedthroughglobalcalibrationfor each input level (Level1 =5.0;Level2=3.0;andLevel3=1.5);
EAC= Asphaltconcreteindirecttensilemodulus,psi;sm = Mixturetensilestrength,psi;m = M-valuederivedfromtheindirecttensilecreepcompliancecurve;andβt = Localormixturecalibrationfactor(setto1.0).
Thestressintensityfactor,K,isdeterminedusingEquationA.12.
Tran,Robbins,Rodezno,andTimm
49
! = !!"# 0.45+ 1.99 !! !.!" (A.12)where:
s tip= Far-fieldstressfrompavementresponsemodelatdepthofcracktip,psi;andCo = Currentcracklength,ft.
The following equations for transverse (thermal) cracking are according to the
AASHTOWarePavementMEDesignsoftwareVersion2.1:
!! = 400×!!"# !
!!"! (A.13)
where:
Cf = Observedamountofthermalcracking,(ft/500ft);N[z]= Standardnormaldistributionevaluatedat[z];C = Crackdepth,in;hac = Thicknessofasphaltconcretelayers,in;andσ = Standarddeviationofthelogofthedepthofcracksinthepavements.
Thechangeinthecrackdepthduetoacoolingcycle,ΔC,iscalculatedasshowninEquationA.14.∆! = !×!! !!!×!×∆!! (A.14)where:
DC = Changeinthecrackdepthduetoacoolingcycle;k = Regressioncoefficientdeterminedthroughfieldcalibration(Level1=1.5;
Level2=0.5;andLevel3=1.5);βt = Calibrationparameter;DK = Changeinthestressintensityfactorduetoacoolingcycle;andA,n = Fractureparametersfortheasphaltmixture,AisdeterminedbyEquationA.15.
! = 10 !.!"#!!.!"×!"# !×!!×! (A.15)where:
E = Mixturestiffness;andsm = Undamagedmixturetensilestrength.
Tran,Robbins,Rodezno,andTimm
50
A.4 Fatigue(Bottom-Up)Cracking
Fatigue cracking is assumed to initiate at the bottom of the asphalt concrete layers andpropagate to the surface under truck traffic. The allowablenumberof axle load applicationsneededfortheincrementaldamageindexapproachtopredictbothtypesofloadrelatedcracks(fatigueandlongitudinal)isshowninEquationA.16asitisshownintheManualofPractice(4). !!!!" = !!! ! !! !!! !! !!!!!! !!" !!!!!! (A.16)where:
Nf-AC = Allowablenumberof axle loadapplicationsfor a flexiblepavement and asphaltoverlays;
εt = Tensilestrainatcriticallocationsandcalculatedbythestructuralresponsemodel,in/in;
EAC= DynamicmodulusoftheHMAmeasuredincompression,psi;kf1,f2,f3= Global field calibrationparameters(fromtheNCHRP1-40D re- calibration;kf1 =
0.007566,kf2=-3.9492,andkf3=-1.281);βf1,f2,f3 = Local or mixture specific field calibration constants; for the global calibration
effort,theseconstantsweresetto1.0;andCH = Thicknesscorrectionterm,dependentontypeofcracking.
! = 10! (A.17)
! = 4.84 !!"
!!!!!"− 0.69 (A.18)
where:
Vbe= Effectiveasphaltcontentbyvolume,%;andVa = PercentairvoidsintheHMAmixture.
TheallowablenumberofaxleloadapplicationsasitispresentedintheAASHTOWarePavementMEDesignsoftwareVersion2.1isshowninEquationA.19.EquationsA.17andA.18areappliedinthesamemannerasinEquationA.16.
!!!!" = 0.00432 ! !!! !! !!!
!!!!! !!
!!!!! (A.19)
where:
Nf-AC= Allowablenumberof axle loadapplicationsfor a flexiblepavement and asphaltoverlays;
Tran,Robbins,Rodezno,andTimm
51
ε1 = Tensilestrainatcriticallocationsandcalculatedbythestructuralresponsemodel,in/in;
E = DynamicmodulusoftheHMAmeasuredincompression,psi;k1,2,3= Globalfieldcalibrationparameters(k1 =0.007566,k2 =3.9492,andk3 =1.281);
andβf1,f2,f3 = Local or mixture specific field calibration constants; for the global calibration
effort,theseconstantsweresetto1.0.Theallowableaxleloadapplicationswerethenusedtodeterminethecumulativedamageindex(DI),whichisasumoftheincrementaldamageindicesovertimeasshowninEquationA.20.
!" = ∆!" !,!,!,!,! = !!!!!" !,!,!,!,!
(A.20)
where:
n = Actualnumberofaxleloadapplicationswithinaspecifictimeperiod;j = Axleloadinterval;m= Axleloadtype(single,tandem,tridem,quad,orspecialaxleconfiguration);l = TrucktypeusingthetruckclassificationgroupsincludedintheMEPDG;p = Month;andl = Median temperature for the five temperature intervals or quintiles used to
subdivideeachmonth,oF.TheareaoffatiguecrackingiscalculatedfromthecumulativedamageindexatthebottomoftheAClayerovertimeusingEquationA.21.
!"!"##"$ = !!!!! !!∗!!!!!!∗!!! ∗!"# !!∗!""
∗ ( !!") (A.21)
where:
FCbottom= AreaoffatiguecrackingthatinitiatesatthebottomoftheAClayers,percentoftotallanearea;
DIbottom = CumulativedamageindexatthebottomoftheAClayers,percent;andC1,2,4 = Transferfunctionregressionconstants;C4=6,000;C1=1;andC2=1.
!!! = −2.40874− 39.748(1+ ℎ!")!!.!"# (A.22)!!! = −2 ∗ !!! (A.23)where:
hAC = totalthicknessofasphaltlayer,in.
Tran,Robbins,Rodezno,andTimm
52
A.5 Longitudinal(Top-Down)Cracking
Longitudinal cracks are assumed to initiate at the surface and propagate downward. TheMEPDG uses Equations A.16 and A.20 to calculate the allowable number of axle loadapplications and cumulative damage index for fatigue and longitudinal cracks. The length oflongitudinalcrackingisthendeterminedusingEquationA.18.
!"!"# = !!
!!! !!!!!∗!"# !!"#∗ 10.56 (A.24)
where:
FCtop = Lengthoflongitudinalcrackingthatinitiatesatthesurface,in;DItop = Cumulativedamageindexatthesurface,percent;andC1,2,4 = Transferfunctionregressionconstants;C4=1,000;C1=7;andC2=3.5.
A.6 InternationalRoughnessIndex(IRI)
TheMEPDG uses Equation A.25 to predict IRI over time for AC pavements. This regressionequationwasdevelopedbasedondatafromtheLTPPprogram.!"! = !"!! + !! !" + !! !"!"#$% + !! !" + !!(!") (A.25)where:
IRI0 = InitialIRIafterconstruction,in/mi;RD = Averagerutdepth,in;
FCTotal = Total area of load-related cracking (combined fatigue, longitudinal, andreflectioncrackinginthewheelpath),percentofwheelpatharea;
TC = Length of transverse cracking (including the reflection of transverse cracks inexistingHMApavements),ft/mi;
C1,2,3,4 = Regressionconstants;C1=40;C2=0.4;C3=0.008;C4=0.015;andSF = Sitefactor(EquationA.26).
!" = !"#$%ℎ + !"#$$% ∗ !"#!.! (A.26)where:
IRI0= InitialIRIafterconstruction,in/mi;andAge= pavementage,year.
!"#$%ℎ = !"[ !"#$%& + 1 ∗ !"#$% ∗ (!" + 1)] (A.27)!"#$$% = !"[ !"#$%& + 1 ∗ !"#$ ∗ (!" + 1)] (A.28)
Tran,Robbins,Rodezno,andTimm
53
!"#$% = !!"#$ + !"#$ (A.29)where:
PI = subgradesoilplasticityindex,percentPrecip = averageannualprecipitationorrainfall,in