Workshop: New approaches to address pavement failure more realistically in asphaltic pavement design methods

… to promote a discussion about different approaches to address pavement failure, primarily cracking, in design methods.

Jorge Soares & Michael Wistuba

June 8th, 2016

Pavementdesign:past,present,future,whereisthecrack?byDr.LaurentPorot

(TheNetherlands)

Germandesignandmanagementapproachesaddressingasphaltpavementcracking,

byDr.MichaelWistuba(Germany)

FaGguecrackingwithinthenewmechanisGc-empiricalpavementdesignmethod

inBrazil,byDr.JorgeSoares(Brazil)

RecentDevelopmentsinAcceleratedPavementTesGng(APT)asaPavementDesign

ToolinCostaRica,byDr.LuisLoria-Salazar(CostaRica)

Considera*onsofuse

Que

stfo

rfund

amen

talund

erstan

ding

Edison

PasteurBohr

Pasteur’s Quadrant Good Science, Well Appplied

Fatigue cracking prediction within the new M-E pavement design guide

in Brazil

Workshop: New approaches to address pavement failure more realistically in asphaltic pavement design methods

Jorge Soares Luis Nascimento

June 8th, 2016

Nooffi

cial

metho

d

PhDThesis:Nascimento(2015),NCSUExperimentalSec*ons/Na*onwideDataBaseM-EPavementDesignGuide

1966 AASHORoadTest+USACE;EngineerMurilloLopesdeSouza;EmpiricalPavementDesignMethod(Official)

2016

1950 EngineerHenryMills-CBRandIGDesign

1928 Rio-PetrópolisHighway

1861 BrazilEmpire-UniãoeIndústriaRoad

1981 �FlexiblePavementDesignMethod�Review(Official)

1983PhDThesis:Preussler�ResilientMethod�inpavementrehabilita\ondesign(towardsmechanis\cal)

BrazilianPavem

entD

esignEvolu\

on

1991 PhDThesis:LauraMo]a�ResilientMethod�inNewflexiblepavementDesign(towardsmechanis\cal)

2007 PhDThesis:FelipeFrancoSISPAV(computa\onalframework;newflexiblepavements)

Empiricalofficialm

etho

d

Pavement Analysis Background in Brazil

-Empiricalapproacheswithoutthoroughvalida\on

-M-EpavementdesignguideunderdevelopmentbytheAsphaltResearchNetwork(Petrobrassince2005)è16universi*es+DNIT(BrazilianFHWA)

-VECDmodelforasphaltmixturecharacteriza\onandpavementdamageanalysis

OBJECTIVES:

-Tobuildandtoevaluatethefieldperformanceofasphaltpavementsec\onsthroughoutthecountry,subjectedtorealtrafficloading

-Todevelopand/ortovalidatepavementmaterialsperformancemodels;Todevelopananalysisframeworktool

Brazilian Database

TS Location Traffic,

number of ESALs

(first year)

Cracked area at last survey, %

Construction technique

Designed asphalt

thickness, mm Asphalt mixture

UFSM 1 Santa Maria, RS 4.49E+05 32.9 Overlay, no milling 50 UFSM 50/70 2 UFSM 3 Santa Maria, RS 8.72E+05 0.0 New pavement 70 UFSM 50/70 1

UFRGS 2 Porto Alegre, RS 7.60E+06 0.0 New pavement 50 UFRGS 60/85 USP 4 Ribeirão Preto, SP 1.40E+05 0.0 New pavement 40 USP 50/70 USP 5 Ribeirão Preto, SP 1.40E+05 0.0 New pavement 40 USP 30/45 ND 1 Aparecida do Norte, SP 2.56E+06 9.0 Overlay, partial milling 140 ND 15/25 ND 2 Aparecida do Norte, SP 2.56E+06 3.0 Overlay, partial milling 200 ND 15/25 ND 3 Aparecida do Norte, SP 2.56E+06 4.8 Overlay, partial milling 140 ND 15/25 ND 4 Aparecida do Norte, SP 2.56E+06 12.5 Overlay, partial milling 200 ND 15/25 ND 5 Aparecida do Norte, SP 2.56E+06 19.0 Overlay, partial milling 200 ND HM 1 ND 6 Aparecida do Norte, SP 2.56E+06 5.9 Overlay, partial milling 200 ND HM 1 ND 7 Aparecida do Norte, SP 2.56E+06 10.0 Overlay, partial milling 140 ND HM 1 ND 8 Aparecida do Norte, SP 2.56E+06 0.0 Overlay, partial milling 140 ND HM 2 ND 9 Aparecida do Norte, SP 2.56E+06 0.0 Overlay, partial milling 200 ND HM 2

ND 10 Aparecida do Norte, SP 2.56E+06 0.0 Overlay, partial milling 140 ND HM 2 UFSC 2 Araranguá, SC 1.33E+07 20.0 New pavement 150 UFSC 50/70

UFSC 3 Araranguá, SC 1.33E+07 0.0 New pavement 114 (top) 56 (bottom)

UFSC 50/70 (top) UFSC RA (bottom)

Pavement Analysis Background in Brazil

Basic Level Advanced Level

Flow Number (traffic)

Flow Number (traffic)

(confinement?) Permanent Deformation

Fatigue

Microcracking described by state

variable S

Typical evolution of a

damage test

ViscoelasGcConGnuum

DamageTheory

VECDModel

Mode-of-loading independent !!!

Widerangeofasphaltmixtures

E=E’ < E0

Controlled-strain Fatigue Testing Simulation

S-VECD+GR-basedfailurecriterion

NfèPredictedvs.Experimental

Inputs

Outputs

Pavementresponses

Damagecontours

Materialsproper6es

Traffic/climate

ComputacionalFramework

Layered ViscoElastic Pavement Analysis for Critical Distresses (1.1Beta program)

GR(Nf)

E*MastercurveShiiFactors

Analysis Framework (FaGgueCracking)

Transfer Function

New Pavement Rehabilitation

Climate (EICM)

Traffic (Load and Volume Distribution)

Materials Properties (E* Mastercurve

Shift Factors)

Damage Contour

Fatigue Analysis S-VECD Model

(asphalt mixture characterization)

LVECD Program

INPU

T St

ruct

ural

A

naly

sis

Rel

iabi

lity-

Bas

ed

Perf

orm

ance

A

naly

sis

% Fatigue Cracking prediction

Damage Index Stresses & Strains Output

Final Output

LVECD Validation and Calibration in Brazil-44pavementsec*onsèRiodeJaneiro,SãoPaulo,SantaCatarina,RioGrandedoSul,andCeará-Widerangeofpavementstructuresandtrafficlevels-Differentdamagelevelsèfa\guecrackedareas-Asphaltlayersthicknesses:40mmto200mm-Widerangeofasphaltmixtures-Materialsanddataavailability

TS ESALs(firstyear)

Crackedarea,%

Overlaytechnique

Designedoverlaythickness,mm Asphaltmixture

19 7.28E+05 7.0 Fullmilling 5060

30/4519.1mm(top)65/9019.1mm(bo]om)

34 1.11E+06 42 Par\almilling 70 30/4519.1mm37 1.18E+06 87 Par\almilling 50 30/4512.5mm

40 1.39E+06 0 SAM/Nomilling 50 30/4512.5mm

Damage Index Proposed

-N/Nfra\ovaluesunderneaththeloadedareaareaveragedusingMiner’slaw

Gridof110points(10x11) Averageddamagevs.TimeCrackPredic*on:

pavementstructure;traffic;materials;climate

Averaged Damage vs. Time

(a)

(b)

Differentaverageddamagegrowthpa]erns:-Asphaltmixtureproper\esandthicknesses-Pavementstructure-Trafficlevel

Averageddamagevs.\me

Crackedareavs.Averageddamage

Thehighertheini\aldamagegrowthrate(fastercracking),thehigherthedamagelevelwhenthecrackingbegins

Crackedareavs.numberofESALs

èFa\guecrack“start”

Averaged Damage vs. Time

Averaged Damage vs. Cracked Area

Averageddamagevs.Crackedarea:Canwecollapseofallthecurvesatonesinglecorrela\on?

Semi-logscale

Arithme\cscale

LVECDsimula6on

Fieldobserva6on

Fatigue Cracking Starting Point Averaged Damage Growth Rate Dependency

T0.35èsimulated\meforaverageddamage(N/Nf)of0.35

N/Nf@10%ofobservedcrackedarea(field)vs.T0.35(simula6on)

-T0.35isstronglycorrelatedtotheaverageddamagewhencrackingstarts-ForT0.35>80months(slowdamagegrowing)èN/Nf@10%tendstostabilize

Fatigue Cracking Starting Point Averaged Damage Shifting

-Mul\plica\veshiifactorè

-Referenceaverageddamageforshiiingè0.5(allcurvesareshiiedinsuchawaythattheir10%ofcrackedareahappensforN/Nfsimulated=0.5)-Thefactors(S)forshiiingtheaverageddamageat10%ofcrackedareatothereferencecondi\onweredetermined:

NN f

red = N N f× S

BTAS +×= 35.0

Defining the Damage-to-Cracked Area Transfer Function

Based on T0.35

Crackedareavs.AveragedreduceddamageCrackedareavs.Averageddamage

BeforeshiOing AOershiOing

-CrackedareaandAveragedreduceddamagepresentedauniqueandstrongrela\onship(consideringfieldvariability)

R2≅0.75

2

1

C

fredN

NCCA ⎟⎠

⎞⎜⎝

⎛×=

(44 National Sections) (curvesareshiiedsothat10%ofcrackedareahappensforN/Nfsimulated=0.5)

Predicted vs. Observed cracked areas

Predic\oncapability:

TS Crackedareaonlastsurvey,% Predictedcrackedarea,%

UFSM1 32.9 23.2UFSM3 0.0 0.0UFRGS2 0.0 0.0USP4 0.0 0.0USP5 0.0 0.0ND1 9.0 10.0ND2 3.0 0.4ND3 4.8 2.9ND4 12.5 0.5ND5 19.0 7.1ND6 5.9 7.1ND7 10.0 44.3ND8 0.0 0.4ND9 0.0 0.0ND10 0.0 0.2UFSC2 20.0 14.8UFSC3 0.0 0.5

Samemixture

Bindercontent=4.7%

Bindercontent=5.2%Rubberasphaltatthebo]omlayer

Fieldproblem

CrackedareaPredictedvs.ObservednumberofESALs(onlycrackedsec\ons)

Predicted vs. Observed cracked areas

Largedifferencesbetweentheobservedandpredictedcrackedareasarenotnecessarilyassociatedwithlargedifferencesregardingtrafficlevel

⎟⎟

⎠

⎞

⎜⎜

⎝

⎛+

+=⎟⎟⎠

⎞⎜⎜⎝

⎛××−⎟⎟

⎠

⎞⎜⎜⎝

⎛redfN

NdcredfN

N

e

baSe100log

1p

redfN

Nedictedliab ZSeCACA ×+=⎟⎟⎠

⎞⎜⎜⎝

⎛Pr.Re

Fatigue cracking prediction standard error Final Performance Cracking Function

LVECD in Brazil

-NCHRPReport457:predic\oncapabilitywithR2>0.65èverygood

-Thepredic\onerrorisaboutfive\messmallerthantheerrorobtainedthroughtheAASHTOMEPDG

AASTHO MEPDG

R2≅0.75R2≅0.27

LVECD - Brazil

Next Steps

-Morena*onalpavementsec*onswillbeincludedinthecalibra\on

-Introducingasphaltmixtureself-healingproper\esintheanalysis

-Permanentdeforma*onpredic\onmodelsvalida\onforasphaltmixtureandstructurallayers

-AnalysisofPortlandcementtreatedmaterials.

ACKNOWLEDGEMENTS

UFAM

UFPA UEMA

UFC

UFPB, UFCG

UFS

UFRJ, IME, IPR

USP, EESC

UFSC

UFPR, UEM

UFG

UFMT

UFBA

UFJF, UFOP, CEFET-MG, UFV

UFSM

UFRGS

JoinusinFortalezafortheISAPConferencein2018...

ISAP Workshop in 2009 (≈ 50 people) ISAP Workshop in 2012 (≈ 50 people) ISAP Symposium in 2012 (≈ 150 people) RMPD Workshop sponsored by ISAP in 2016 (≈ 50 people) ISAP Conference in 2018 …