Development of a Morphology-based AnalysisFramework for Asphalt Pavements

IBRAHIM ONIFADE

Licentiate ThesisStockholm, Sweden 2015

TRITA-BYMA 2015:1ISSN 0349-5752

KTH Division of Building MaterialsSE-100 44 Stockholm

SWEDEN

Akademisk uppsats som med tillstånd av Kungl Tekniska högskolan framlägges tilloffentlig granskning för avläggande av teknisk licentiatexamen fredag den 8 maj2015 klockan 10.00 i sal B26, KTH, Brinellvägen 23, Stockholm.

© Ibrahim Onifade, May 2015

Tryck: Universitetsservice US AB

Abstract

The morphology of asphalt mixtures plays a vital role in their proper-ties and behaviour. The work in this thesis is aimed at developing a funda-mental understanding of the effect of the asphalt morphology on the strengthproperties and deformation mechanisms for development of morphology-basedanalysis framework for long-term response prediction. Experimental and com-putational methods are used to establish the relationship between the mixturemorphology and response. Micromechanical modeling is employed to under-stand the complex interplay between the asphalt mixture constituents resultingin strain localization and stress concentrations which are precursors to damageinitiation and accumulation. Based on data from actual asphalt field cores,morphology-based material models which considers the influence of the mor-phology on the long-term material properties with respect to damage resistance,healing and ageing are developed. The morphology-based material models areimplemented in a hot-mix asphalt (HMA) fracture mechanics framework forpavement performance prediction. The framework is able to predict top-downcracking initiation to a reasonable extent considering the variability of the inputparameters. A thermodynamic based model for damage and fracture is pro-posed. The results from the study show that the morphology is an importantfactor which should be taken into consideration for determining the short- andlong-term response of asphalt mixtures. Further understanding of the influ-ence of the morphology will lead to the development of fundamental analyticaltechniques in design to establish the material properties and response to loads.This will reduce the empiricism associated with pavement design, reduce needfor extensive calibration and validation, increase the prediction capability ofpavement design tools, and advance pavement design to a new level science andengineering.

Keywords: Morphology, damage, X-ray computed tomography, top-down crack-ing, fracture

i

ii

iii

Sammanfattning

Asfaltblandningars morfologi har en avgörande betydelse för deras egen-skaper och beteenden. Arbetet i denna avhandling syftar till att utveckla engrundläggande förståelse för effekten av asfaltsmorfologin för deras hållfast-hetsegenskaper och deformationsmekanismer och utveckling av ramverksana-lysmorfologi baserat på långsiktig förutsägelse. Experimentella beräkningsme-toder används för att fastställa sambandet mellan blandningens morfologi ochrespons. Mikromekanisk modellering används för att förstå det komplexa sam-spelet mellan asfaltmassans beståndsdelar som resulterar i spänningslokalise-ring och spänningskoncentrationer som är föregångare till initiering av ska-dor och ackumulation. Morfologibaserade materialmodeller beaktar påverkanav morfologin på de långsiktiga materialegenskaperna med avseende på ska-demotstånd, helande samt åldrande, och är utvecklade från data hos verkli-ga asfaltsfältskärnor. Morfologinbaserade materialmodeller är implementeradei en varmblandad asfalt-(HMA)-brottmekanik-ramverk för förutsägelse av be-läggningsprestanda. Ramverket kan i rimlig utsträckning förutspå variationen iingångsparametrarna ’top-down’ sprickbildningsinitiering. En termodynamisk-baserat ramverk föreslås för skador och brott. Resultaten från studien visaratt morfologin är en viktig faktor som bör beaktas för att bestämma responsav asfaltblandningar på kort och lång sikt. Ytterligare förståelse av inverkanav morfologin kommer att leda till utvecklingen av grundläggande analytis-ka tekniker i design för fastställning av materialegenskaper och belastningarsrespons. Detta kommer att minska empirism som förknippas med beläggnings-konstruktionen, minska behovet av omfattande kalibrering och validering, ökaförutsägelseförmågan av designverktyg för beläggningen, samt avancera belägg-ningsdesign till en ny vetenskaplig nivå och ingenjörskonst.

Nyckelord : Morfologi, skador, röntgendatortomografi, ’top-down’ sprickbild-ning, fraktur

Preface

The work presented in this thesis has been carried out partly at the Division ofHighway and Railway Engineering and the Division of Building Materials at KTHRoyal Institute of Technology, Stockholm. I would like to express my gratitude to mysupervisors Prof. Björn Birgisson, Associate Professor Nicole Kringos and AssistantProfessor Denis Jelagin for their assistance during the course of the work. I will alsolike to appreciate the financial support of the Swedish Transport Administration(Trafikverket). My appreciation also goes to Gerald Huber and Bill Pine of theHeritage group, Indianapolis for their insights and discussions with respect to thework in this thesis. I also appreciate the support of my colleagues at the departmentfor providing an enabling and conducive environment.

My sincere gratitude also goes to my wife Busola Odubonojo for her understand-ing, love, care and support.

Ibrahim OnifadeStockholm, Sweden.

v

List of appended papers

Paper I

Dinegdae, Y., Onifade, I., Jelagin, D., Birgisson, B. (2015). Mechanics-based Top-down Fatigue Cracking Initiation Prediction Framework for Asphaltic Pavements.Submitted to Road Materials and Pavement Design.

Paper II

Onifade, I., Jelagin, D., Birgisson, B., Kringos, N. (2015). Towards Asphalt MixtureMorphology Evaluation with the Virtual Specimen Approach. Submitted to EATA2015 conference for publication in Special Edition of Road Materials and PavementDesign Journal.

Paper III

Onifade, I., Balieu, R., Birgisson, B. (2015). Energy-Based Damage and FractureFramework for Viscoelastic Asphalt Concrete. Submitted to the Journal of Engi-neering Fracture Mechanics.

In addition to the appended papers, the work has resulted in thefollowing conference publications and presentations:

Onifade, I., Jelagin, D., Guarin, A., Birgisson, B., Kringos, N. (2013). Asphalt In-ternal Structure Characterization with X-Ray Computed Tomography and DigitalImage Processing, in: Kringos, N., Birgisson, B., Frost, D., Wang, L. (Eds.), Multi-Scale Modeling and Characterization of Infrastructure Materials, RILEM Book-series. Springer Netherlands, pp. 139-158.

Onifade, I., Jelagin, D., Guarin, A., Birgisson, B., Kringos, N. (2014). Effect ofMicro-scale Morphological Parameters on Meso-scale Response of Asphalt Concrete,in: Asphalt Pavements. CRC Press, pp. 1775-1784.

vii

viii

Contents

Abstract i

Sammanfattning iii

Preface v

List of appended papers vii

Contents ix

1 Introduction 1

2 Methods 52.1 Asphalt Morphology Framework . . . . . . . . . . . . . . . . . . . . 52.2 Multi-scale modeling approach . . . . . . . . . . . . . . . . . . . . . 7

3 Results and Discussion 93.1 Paper I - Mechanics-based Top-down Fatigue Cracking Initiation Pre-

diction Framework for Asphaltic Pavements . . . . . . . . . . . . . . 93.2 Paper II - Towards Asphalt Mixture Morphology Evaluation with the

Virtual Specimen Approach . . . . . . . . . . . . . . . . . . . . . . . 123.3 Paper III - Energy-Based Damage and Fracture Framework for Vis-

coelastic Asphalt Concrete . . . . . . . . . . . . . . . . . . . . . . . . 16

4 Summary and Conclusions 19

References 21

Appended Papers 23

ix

Chapter 1

Introduction

Asphalt concrete is a composite material made up of aggregates, bitumen binderand air voids. The morphology of the asphalt mixtures may be defined as a set of pa-rameters describing the geometrical characteristics of the constituent phases, theirrelative proportions as well as spatial arrangement in the mixture. In particular,at the meso-scale, the morphology of asphalt mixtures is of considerable practicalimportance, as it has been shown in several studies that the deficient internal struc-ture of the material results in compromised performance in the field, e.g. Eppset al. (2002). In order to ensure the adequate internal structure of the material, theasphalt mixture design methods put requirements on aggregate size distribution,their angularity and texture as well as binder, voids in mineral aggregates (VMA)and target air void content, e.g. E-C124 Transportation Research Board (2007).These requirements are however formulated primarily based empirical observationsand also considers the contribution of the individual morphology component, and asa result, they cannot be used to fully optimize the internal structure of the mixture.

The effect of the asphalt mixture internal structure on its performance receivedconsiderable attention in the literature. Recent attempts in the field are focusedon combining experimental investigations with numerical modeling. For instance,Souza et al. (2012) studied the effects of aggregate angularity and binder contenton bituminous mixture and related these effects to materials fracture resistance. Itwas concluded that the fracture energy is increased as the aggregate angularity isdecreased and the binder content increased. In Chen et al. (2005), the effect ofcoarse aggregate shape on rutting performance of asphalt concrete mixtures hasbeen evaluated. The authors proposed a measure of the combined effect of theparticle shape, angularity, and surface texture referred to as Particle Index (PI)was used in the study to define the stability of an aggregate in the mix. The airvoids content, and bitumen content and voids in mineral aggregates (VMA) havealso been related to mixture performance, e.g. Epps et al. (2002); Kandhal andChakraborty (1996b). The VMA has been related to the durability of mixtures andits ability to resist changes in the hot mix asphalt (HMA) properties. InadequateVMA can result to rapid oxidization of the asphalt which could make the pavement

1

2 CHAPTER 1. INTRODUCTION

too brittle, e.g. Chadbourn et al. (2000). Inadequate air voids content can havean adverse effect on the mixture performance and is mainly manifested in asphaltpavements as bleeding and rutting of the asphalt pavement, e.g. Epps et al. (2002).The effect of asphalt binder film thickness on mixture performance was investigatedin Kandhal and Chakraborty (1996a). It was found that there is a correlationbetween the asphalt binder film thickness and the mixture tensile strength, tensilestrain, resilient modulus and ageing susceptibility. Mixtures with low asphalt binderfilm thickness have higher strength and stiffness properties.

Recently, a new morphology framework was developed by Lira et al. (2012);Yideti et al. (2013) which can be used to characterize the internal structure of un-bound granular and asphalt concrete materials. The framework considers the sizegradation and distribution of the stones, the distribution of the bitumen and thedistribution of the air voids in the asphalt mixture matrix. It allows for the de-termination of a morphological parameter called "Primary Structure" (PS) coatingthickness (tps) which is the characteristic average value of the thickness of the masticcoating around the main load carrying structure. Asphalt mixtures with differentgradation and volumetric properties have distinct tps values. Das et al. (2013) havestudied the influence of the changing morphological parameter on the performanceof asphalt mixtures. It was found that there exist a relationship between the mor-phology parameter and the resilient modulus, creep compliance, and fracture energyof the mixtures.

The characterization and influence of the morphology on the long-term mate-rial response is not an integral part of existing mechanistic-empirical (ME) designtools. These ME tools are used to predict the damage in an asphalt pavement asa function of age or accumulated traffic loads. The measure of damage usuallyconsidered in these tools are fatigue cracking (bottom-up), permanent deformationand thermally induced cracking. However, advances in the use of non-destructiveX-ray Computed Tomography (CT) techniques provide the means to digitally cap-ture, quantify and interactively modify the internal structure of asphalt. X-Ray CTand micromechanical modeling techniques has been used to investigate the meso-scale response of asphalt mixtures to different loading conditions, e.g. Bažant et al.(1990); Onifade et al. (2013); Tashman et al. (2002); You et al. (2012, 2008).

The objective of the work in this thesis is to investigate the influence of themorphology on the short-term and long-term behaviour of asphalt concrete usingexperimental and computational modeling tools. The morphology framework devel-oped by Lira et al. (2012) and Yideti et al. (2013) is used for the characterization ofthe internal structure. Relationships between the morphology and the key mixtureproperties for the determination of the long-term material response are established.These relationships are implemented in an analysis framework to predict the top-down crack initiation time in asphalt pavements. Micromechanical modeling is usedto develop more fundamental understanding of the influence of morphology on thestrength and deformation mechanism in the mixture. A first step for the devel-opment of a morphology-based homogenized macro-scale model for damage andfracture characterization in asphalt pavements is developed and proposed.

3

The thesis is divided into three parts. The first part is focused on the im-plementation of morphology-based material models in an analysis framework forpredicting top-down cracking initiation. Relationships between morphological pa-rameter and key mixture properties are developed using data from asphalt fieldcores. The second part is focused on the investigation of the influence of the as-phalt morphology on the strength and degradation mechanisms in asphalt mixturesusing micro-mechanical modeling techniques. The third part is focused on the de-velopment of a thermodynamic framework for damage and fracture characterizationbased on the understanding from the previous parts.

Chapter 2

Methods

A multi-scale modeling approach and the asphalt morphology framework are em-ployed in this thesis to develop an understanding of the influence of morphologyon mixture performance. The asphalt morphology framework is used to character-ize the internal structure of the asphalt mixture and micro-mechanical modelingtechnique is used to investigate the response to load at the meso-scale. Continuumdamage mechanics and the thermodynamics of irreversible processes using internalstate variables are used for the development of a macro-scale damage and fracturemodel. The relationships between the multi-scale model characterization and themixture morphology is relied on for the development of morphology-based macro-scale model for asphalt pavements.

2.1 Asphalt Morphology Framework

In this thesis, the asphalt concrete morphology is characterized using a morphologyframework presented in Das et al. (2013); Lira et al. (2012); Yideti et al. (2013).The morphology framework can be used to categorize the aggregates in the asphaltmixture into four different groups namely, the primary structure, the secondarystructure, the oversized particles, and the filler particles. The primary structureis the main load carrying structure in the mix and it is made up of interactingaggregate particle sizes. The main idea is that there is a range of particle sizes thatinteracts to form the main load carrying structure in the mixture. The secondarystructure consists of particle sizes smaller than the primary load carrying structuresand fills the voids inside the primary structure. The secondary structure can helpin stabilizing the primary structure but too much of the secondary structure canresult in disruption of the load carrying structures. The oversized particles arethose particles with sizes greater than the primary load carrying structure anddo not contribute to the load carrying capacity. An interaction check betweenconsecutive sieve sizes beginning with the largest sieve sizes is used to check ifthe particle sizes retained on the consecutive sieve sizes are interacting with eachother to transfer load in the mixture. Once the four different structures have been

5

6 CHAPTER 2. METHODS

identified, another term called the "Primary Structure" coating thickness (tps) canbe defined. The Primary Structure coating thickness is average value of the coatingof the secondary structure, the mineral filler and the bitumen all mixed togetheraround the primary structure. The "Primary Structure" coating thickness is referredto as the "mastic coating thickness" in this thesis. Figure 2.1 shows the conditionfor the identification of the different groups in the aggregate structure.

The relationship for the mastic coating thickness and the porosity of the primarystructure is shown in Equation 2.1.

tps = 0.95(ηps)1.28 × dp/2 (2.1)

Where:

tps: is the mastic coating thickness

ηps: is the porosity of the primary structure

dp: is the weighted average diameter of the primary structure particles

Figure 2.1: Identification of different groups based on aggregate gradation Das et al.(2013)

2.2. MULTI-SCALE MODELING APPROACH 7

2.2 Multi-scale modeling approachAsphalt mixtures are heterogeneous materials at several length scales. At meso-scale, the asphalt mixture is considered as a heterogeneous material with the aggre-gates, mastic and air voids modeled separately and assigned appropriate materialproperties. Meso-mechanical analysis provides information about the stress andstrain fields due to the interaction of the constituents in the mix. It takes intoaccount the influence of the interaction at the interface between the aggregate andthe mastic on the overall mixture performance. Adhesive damage due to breakingof the bond between the mastic and the stones as well as cohesive damage which isdue to the lose of integrity of the mastic phase can be captured as well.

At the macro-scale, the asphalt mixture is considered a homogeneous materialand modeled with effective material properties. The macro-scale models do notaccount for the distribution of the constituents, stress concentration and strainlocalization as well as the behaviour at the stone-mastic interface. A great deal ofdetails are not accounted for in the macro-scale models.

The characterization of the morphology as well as understanding of its influenceon material response using micromechanical modeling techniques provide a wayto systematically take into account meso-scale morphological parameters in macro-scale models. This provides the possibility to develop morphology-based macro-scalemodels to scale up from the meso-scale to the macro-scale. The morphology-basedmacro-scale models will then take into account the influence of the morphology onthe short-term and long-term material response. Figure 2.2 shows a schematic ofthe scaling up approach from meso-scale to macro-scale based on the morphology.

8 CHAPTER 2. METHODS

Figure 2.2: Scaling up from the meso-scale to the macro-scale using morphology-based models

Chapter 3

Results and Discussion

3.1 Paper I - Mechanics-based Top-down FatigueCracking Initiation Prediction Framework forAsphaltic Pavements

The paper presents a morphology-based top-down fatigue cracking analysis frame-work for asphalt pavements. Top-down cracking has been the predominant failuremode observed in core samples evaluation in many parts of the world includingJapan, UK, Europe and the United States, e.g. Gerritsen et al. (1987); Uhlmeyeret al. (2000). It was also reported that over 90% of the cracking in asphalt pave-ments in the state of Florida was in the form of top-down fatigue cracking, e.g.Uhlmeyer et al. (2000). Existing mechanistic-empirical ME design tools have notbeen optimized for the prediction of top-down cracking.

The morphology-based top-down fatigue cracking analysis framework is basedon the hot-mix asphalt (HMA) fracture mechanics. The HMA fracture mechanicsidentifies the existence of a fundamental mixture property i.e. Dissipated CreepStrain Energy (DCSE), below which any damage induced in the material is heal-able. Relationships between the morphology and the key mixture properties suchas the DCSE and healing potential are established to predict the changes in keymixture properties over the pavement service life. These established relationships(morphology-based sub-models) are used in the framework to predict the varia-tions in the material properties with time. Figure 3.1 shows the description of thetop-down cracking initiation prediction framework.

Twenty-eight different pavement sections which include state roads, turnpikesand interstates were selected for the calibration and validation of the framework.The relevant information needed as input in the analysis framework include asphaltmixture gradation and volumetrics, binder type, cross-sectional properties and di-mensions, traffic, and temperature profile. The observed crack initiation time of thepavement sections were obtained from the FDOT database, Florida department oftransportation (FDOT) (2013). Sixteen pavement sections were selected for modelcalibration and categorized into 2 different groups. Pavement sections with an an-

9

10 CHAPTER 3. RESULTS AND DISCUSSION

 

 

                                                                                                                                                                                                               

   

                                                                                                                                                                                           

 

 

 

 

 

 

 

Inputs Module

Mixture, environmental and cross- sectional properties

Traffic

Material-property sub-models

Pavement response sub-model Damage accumulation and recovery sub-model

Crack initiation prediction sub-model

Figure 3.1: Top-down cracking initiation prediction framework

nual traffic volume of 100,000 ESALS or less are referred to as "low traffic volumegroup" while those with annual traffic volume of more than 100,000 ESALS arereferred to as "high volume group".

0

5

10

15

I75-1A

I75-1B I75-3

I75-2

SR80-1

SR80-2

I-75SB

I-75SB2

I-10EB

US-301SB

CI t

ime

(yea

r)

Pavement section

Observed Predicted

Figure 3.2: Predicted vs observed results for medium to high volume roads

3.1. PAPER I - MECHANICS-BASED TOP-DOWN FATIGUE CRACKINGINITIATION PREDICTION FRAMEWORK FOR ASPHALTIC PAVEMENTS11

0  

5  

10  

15  

SR222 SR 16-6 US 19-2 SR16-4 SR89 SR18

CI t

ime

(yea

rs)

Pavement sections

Observed Predicted

Figure 3.3: Predicted vs observed results for low volume roads

Due to the variations in the input factors that influence the long term perfor-mance of asphalt pavement sections, a crack initiation time of ±3 years from theobserved crack initiation time in the field is considered a "good" prediction. Figures3.2 and 3.3 show the calibration results for the medium to high volume roads andthe low volume roads respectively. For the "medium to high volume roads", it canbe seen that the predicted crack initiation time is consistent with the observed crackinitiation time except for SR80-1 that deviates from the observed value by 3.7 years.For the "low volume roads", the predicted crack initiation time is also in the rangeof acceptable prediction except for US19-2 and SR89 that deviated by more than 3years. Twelve pavement sections are used for the validation of the model. Figure3.4 shows the result of the model validation. The result of the validation showsthat the framework is capable of predicting the top-down crack initiation time to areasonable extent.

0  

5  

10  

15  

TPK2

NW39-1

NW39-2

SR483

SR563 US41

SR823 US27

A1A SR60

US1 SR46

CI t

ime

(yea

r)

Pavement section

Observed Predicted

Figure 3.4: Model validation result using 12 different pavement sections

12 CHAPTER 3. RESULTS AND DISCUSSION

With the consideration of the influence of the morphology on the key mixtureproperties, the framework is able to predict the top-down crack initiation time withreasonable accuracy. The model has accounted for the fundamental mechanism ofpavement degradation on which further study will be based. More accurate trafficcharacterization will further improve the prediction capability of the framework.

3.2 Paper II - Towards Asphalt Mixture MorphologyEvaluation with the Virtual Specimen Approach

The paper presents the study of the effect of different internal structures on thestrength properties and deformation mechanisms of asphalt mixtures. Using X-ray computed tomography (CT), the internal structure with the distribution of thedifferent constituents in an asphalt concrete sample is captured. Image processingtechnique is used to identify, segment and quantify the constituent into air voidphase, mastic phase and aggregate phase. The morphology framework is used tocharacterize the internal structure of the mixture. The framework is also used tocharacterize the three dimensional (3D) distribution of the average mastic thicknessaround the aggregate main load carrying structure. This mastic thickness is referredto as the "mastic coating thickness". The mastic coating thickness is used as amorphological parameter to characterize the morphology of the asphalt sample.Figures 3.5a and 3.5b show the scanned asphalt concrete sample and a typicalslice from the X-ray CT scan. The image processed asphalt concrete sample andthe segmentation of the different phases is shown in Figure 3.6.

(a) (b)

Figure 3.5: a)Asphalt concrete sample, b) X-Ray CT slice from the scanned asphaltconcrete sample.

3.2. PAPER II - TOWARDS ASPHALT MIXTURE MORPHOLOGYEVALUATION WITH THE VIRTUAL SPECIMEN APPROACH 13

(a) Asphalt concrete filtered image (b) Air-void phase

(c) Mastic phase (d) Aggregate phase

Figure 3.6: Image processed asphalt concrete sample and segmentation results

The morphology of the scanned asphalt concrete sample is then virtually modi-fied using erosion morphological operation tool to make the 12.5mm and 9.5mm ag-gregate sizes finer. The scanned asphalt concrete sample is referred to as "Structure1". The Erosion morphological operation is used to modify the aggregate gradationinside the mixture. The 12.5mm particle size in "Structure 1" is eroded by one pixelto obtain "Structure 2" while the 9.5mm particle size in "Structure 2" is eroded byone pixel to obtain "Structure 3". The percentage air voids is dilated to compensatefor the increase in percentage binder when the 12.5mm and 9.5mm aggregate sizesare made finer. In this way, the three different resulting structures have relativelythe same amount of binder but different morphological structures. The effect ofthe varying morphological structure to mechanical loading at 0oC is studied usingthe finite element method by subjecting the three structures to the same boundaryconditions. The stones are modeled as isotropic linear elastic materials and the

14 CHAPTER 3. RESULTS AND DISCUSSION

mastic modeled as a viscoelastic material using the generalized Maxwell’s model.The mastic coating thickness and the primary structure range for the structures areshown in Table 3.1. Figure 3.7 show the three different morphological structuresobtained after the modification.

Table 3.1: Mastic coating thickness and primary load carrying structure range forthe three different morphology structures

Structure 1 Structure 2 Structure 3Mastic coating thickness (mm) 2.13 2.47 2.48PS Range (mm) 12.5 - 4.75 9.5 - 4.75 9.5 - 4.75Aggregate proportion inPrimary Structure (%) 75.5 56.0 54.8

(a) Structure 1 (b) Structure 2 (c) Structure 3

Figure 3.7: The three different morphological structures with cropped region shownin Figure 3.6a

The results show that the structure with the lowest mastic coating thicknesshas better stress distribution patterns with less stress concentrations in the masticregions. High strain localizations can be observed in the structure with the highestmastic coating thickness. It can be observed that the structure with the lowest mas-tic coating thickness has the highest effective modulus and the modulus decreasesas the mastic coating thickness increases. Figure 3.8 shows the effective relaxationmodulus and the stress-strain response for the three different structures. Figure3.9 show the finite element mesh and the first principal stress streamlines at theaggregate boundaries.

3.2. PAPER II - TOWARDS ASPHALT MIXTURE MORPHOLOGYEVALUATION WITH THE VIRTUAL SPECIMEN APPROACH 15

0 10 20 30 40 50 60 70 80 90 1000

1

2

3

4

5

6

7

time [s]

Effe

ctiv

e re

laxa

ion

mod

ulus

[GPa

]

Structure 1Structure 2Structure 3

(a)

0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.160

0.5

1

1.5

2

2.5

3

3.5

strain [%]

stre

ss [M

Pa]

Structure 1Structure 2Structure 3

(b)

Figure 3.8: a)Effective relaxation modulus, b) stress-strain response.

(a) Finite element mesh (b) Structure 1

(c) Structure 2 (d) Structure 3

Figure 3.9: FE mesh and First principal stress streamlines at aggregate boundariesafter application of a uniaxial tensile stress of 2MPa

16 CHAPTER 3. RESULTS AND DISCUSSION

Micromechanical modeling provides fundamental understanding of the influenceof the internal structure on the material response. The work is a step towards theimplementation of a digital specimen approach for optimization of mix designs tomeet desired functional requirements. The work also highlights the importance ofthe morphology of the asphalt mixture in the characterization of mixture perfor-mance. The morphology parameter can be used in homogenized asphalt concretemodels to take into account the effect of the morphology on short-term and long-term material response.

3.3 Paper III - Energy-Based Damage and FractureFramework for Viscoelastic Asphalt Concrete

The paper is focused on the development of a unified damage and fracture frame-work for asphalt concrete mixtures with particular attention paid to micro-crackinitiation and accumulation. The paper provides a step towards a more funda-mental pavement response prediction. The work suggests the development of twodifferent potential-based models to accurately characterize the material responseat low and high temperatures respectively. One of the models will focus on thecharacterization of cracking at low temperatures, while the other will be used tocharacterize the plastic deformation at high temperatures. Both models can then becoupled to characterize the material behaviour at intermediate temperature. Theintegrated model will provide improved prediction of material response at extremetemperatures i.e. low and high temperatures, while minimizing material predictionerrors at intermediate temperatures.

The proposed damage model is based on continuum damage mechanics andthe thermodynamics of irreversible processes using internal state variables. Theinternal state variable is used to characterize the distributed damage in viscoelasticasphalt materials in the form of micro-crack initiation and accumulation. At lowtemperatures and high deformation rates, micro-cracking is considered as the sourceof non-linearity in the material response and thus the cause of deviation from linearviscoelastic response. Using a non-associated evolution law, a damage initiationcriterion is used to identify the instance of micro-crack initiation while anothercriterion is used to derive the micro-crack evolution by means of restrictions imposedby the second law of thermodynamics.

The Superpave IDT test is used to characterize the performance of six differentasphalt concrete mixtures used in this study. The Superpave IDT test procedureconsists of three different tests (resilient modulus test, creep test and strength test)which can be used to characterize the samples nondestructively and destructively.In this study, the Superpave IDT tests are carried out at three different tempera-ture (−20oC, −10oC and 0oC) for each test setup. The low temperature range isconsidered to minimize the effect of plasticity. The linear viscoelastic response andthe strength characteristics of the mixtures are obtained from the Superpave IDTtest results.

3.3. PAPER III - ENERGY-BASED DAMAGE AND FRACTUREFRAMEWORK FOR VISCOELASTIC ASPHALT CONCRETE 17

A micro-crack initiation threshold is identified below which the material responseis purely linear viscoelastic. Temperature coupling is introduced to predict the dam-age parameters and critical micro-crack initiation threshold at other temperaturesnot tested for. It was observed that the critical micro-crack damage threshold in-creases as the temperature increases. This results in an increase in resistance tomicro-crack formation at higher temperatures. The proposed damage model showsthe capability to characterize the damage in both conventional and unconventionalasphalt mixtures. Figure 3.10 shows the evolution of the micro-cracking damageinitiation threshold for AG1-mixtures from −40oC to 40oC

Figure 3.10: Evolution of micro-cracking damage initiation threshold for AG1-mixtures from −40oC to 40oC

Internal scalar damage variable is used to estimate the distributed damage inthe material. The damage variable D is estimated from the Superpave IDT strengthtest as a function of the experimental observed stress and the theoretical predictedlinear viscoelastic stress response. Figures 3.11 and 3.12 show the experimentaldamage evolution and the model predicted damage evolution for the conventionalmixtures, AG1-0 and AG2-0 mixtures.

18 CHAPTER 3. RESULTS AND DISCUSSION

0 200 400 600 800 1000 1200 1400 1600 1800 20000

0.05

0.1

0.15

0.2

0.25Damage evolution vs strain

strain (microstrain)

dam

age

(D)

experimentalmodel prediction

Figure 3.11: Micro-crack damage evolution: model and experiment for AG1-O

0 200 400 600 800 1000 1200 1400 1600 18000

0.05

0.1

0.15

0.2

0.25Damage evolution vs strain

strain (microstrain)

dam

age

(D)

experimentalmodel prediction

Figure 3.12: Micro-crack damage evolution: model and experiment for AG2-O

The proposed damage evolution law is consistent with the estimated damageevolution in the Superpave indirect tensile (IDT) strength test with a good accuracyat the temperatures considered. The energy-based formulation enables the extensionof the model to a wide range of temperature and easy incorporation of the healingand ageing phenomena.

Chapter 4

Summary and Conclusions

The thesis is focused on developing fundamental understanding of the influence ofasphalt morphology on its response. Experimental investigation of field cores isused to determine the effect of the morphology on resistance to damage and otherkey mixture properties. Micromechanical modeling is used to study the complexinteraction between the asphalt mixture constituents at the meso-scale. It wasfound that the morphology doesn’t only have effect on the instantaneous effectiveproperties but also, it determines the load distribution pattern and deformationmechanisms within the mixture which influences the long-term performance.

The understanding of the deformation mechanism and interaction between con-stituent composition over a wide range of temperature provided the motivation forthe development of a thermodynamic based damage and fracture model. The modelpresented focuses more on the characterization of micro-crack initiation and accu-mulation. The model identifies the existence of a critical micro-crack damage initi-ation threshold below which the response is purely linear viscoelastic. Temperaturecoupling is integrated into the model to characterize the material at temperaturesnot tested for.

The results highlight the importance of the morphology for improved materialcharacterization. Fundamental understanding of the influence of the morphologyon the short- and long term bulk material response will enable the developmentof new morphology-based models for improved pavement performance analysis andprediction. The models will provide the basis for more robust analytical techniquesin the design and analysis of asphalt pavements with which mixture morphology canbe optimized to meet certain functional requirements. This development will presentthe opportunity for the realization of performance-based specifications for asphaltmix design and as well reduce or totally eliminate the black boxes in pavementdesign. This will as well reduce the empiricism associated with pavement design,reduce need for extensive calibration and validation, and advance pavement designto a new level science and engineering.

19

References

Bažant, Z. P., Tabbara, M. R., Kazemi, M. T., and Pijaudier-Cabot, G. (1990).Random particle model for fracture of aggregate or fiber composites. Journal ofEngineering Mechanics, 116(8):1686–1705.

Chadbourn, B. A., Skok Jr, E., Newcomb, D.E. Crow, B., and Spindler, S. (2000).The Effect of Voids in Mineral Aggregate (VMA) on Hot-mix Asphalt Pavements.Minnesota Department of Transportation, Office of Research & Strategic Services.

Chen, J.-S., Chang, M. K., and Lin, K. Y. (2005). Influence of coarse aggregateshape on the strength of asphalt concrete mixtures. Journal of the Eastern AsiaSociety for Transportation Studies, 6:1062–1075.

Das, P. K., Birgisson, B., Jelagin, D., and Kringos, N. (2013). Investigation of theasphalt mixture morphology influence on its ageing susceptibility. Materials andStructures, 48(4):987–1000.

E-C124 Transportation Research Board (2007). Practical approaches to hot-mix as-phalt mixdesign and production quality control testing. Number E-C124 in Trans-portation Research Circular. Transportation Research Board, Washington, D.C.

Epps, J. A., National Cooperative Highway Research Program, National ResearchCouncil (U.S.), American Association of State Highway and Transportation Offi-cials, and United States, editors (2002). Recommended performance-related speci-fication for hot-mix asphalt construction: results of the WesTrack Project. Number455 in NCHRP report. National Academies Press, Washington, D.C.

Florida department of transportation (FDOT) (2013). Roadway designs / pavementmanagements / reports.

Gerritsen, A. H., Van Gurp, C. A. P. M., Van Der Heide, J. P. J., Molenaar, A.A. A., and Pronk, A. C. (1987). Prediction and prevention of surface cracking inasphaltic pavements.

Kandhal, P. S. and Chakraborty, S. (1996a). Effect of asphalt film thickness onshort and long-term aging of asphalt paving mixtures. Transportation ResearchRecord, 1535(1):83–90.

21

22 REFERENCES

Kandhal, P. S. and Chakraborty, S. (1996b). Evaluation of voids in the mineralaggregate for HMA paving mixtures. In Proceedings of the Annual Conference-Canadian Technical Asphalt Association, pages 78–101. Polyscience PublicationsInc.

Lira, B., Jelagin, D., and Birgisson, B. (2012). Gradation-based framework forasphalt mixture. Materials and Structures, 46(8):1401–1414.

Onifade, I., Jelagin, D., Guarin, A., Birgisson, B., and Kringos, N. (2013). Asphaltinternal structure characterization with x-ray computed tomography and digitalimage processing. In Kringos, N., Birgisson, B., Frost, D., and Wang, L., editors,Multi-Scale Modeling and Characterization of Infrastructure Materials, number 8in RILEM Bookseries, pages 139–158. Springer Netherlands.

Souza, L. T., Kim, Y.-R., Souza, F. V., and Castro, L. S. (2012). Experimentaltesting and finite-element modeling to evaluate the effects of aggregate angularityon bituminous mixture performance. Journal of Materials in Civil Engineering,24(3):249–258.

Tashman, L., Masad, E., D’Angelo, J., Bukowski, J., and Harman, T. (2002). X-raytomography to characterize air void distribution in superpave gyratory compactedspecimens. International Journal of Pavement Engineering, 3(1):19–28.

Uhlmeyer, J., Willoughby, K., Pierce, L., and Mahoney, J. (2000). Top-down crack-ing in washington state asphalt concrete wearing courses. Transportation ResearchRecord: Journal of the Transportation Research Board, 1730:110–116.

Yideti, T. F., Birgisson, B., Jelagin, D., and Guarin, A. (2013). Packing theory-based framework to evaluate permanent deformation of unbound granular mate-rials. International Journal of Pavement Engineering, 14(3):309–320.

You, T., Abu Al-Rub, R. K., Darabi, M. K., Masad, E. A., and Little, D. N.(2012). Three-dimensional microstructural modeling of asphalt concrete using aunified viscoelastic–viscoplastic–viscodamage model. Construction and BuildingMaterials, 28(1):531–548.

You, Z., Adhikari, S., and Dai, Q. (2008). Three-dimensional discrete elementmodels for asphalt mixtures. Journal of Engineering Mechanics, 134(12):1053–1063.

Appended Papers

Contribution of the work in the appended papers:

Paper I

Dinegdae, Y., Onifade, I., Jelagin, D., Birgisson, B. (2015), Mechanics-based Top-down Fatigue Cracking Initiation Prediction Framework for Asphaltic Pavements.Submitted to Road Materials and Pavement Design.

Onifade and Dinegdae were involved in the development of the top-down fatiguecracking initiation prediction framework and analysis of the data. Jelagin and Bir-gisson provided guidance during the work. Dinegdae wrote the paper.

Paper II

Onifade, I., Jelagin, D., Birgisson, B., Kringos, N., (2015). Towards Asphalt Mix-ture Morphology Evaluation with the Virtual Specimen Approach. Submitted toEATA 2015 conference for publication in Special Edition of Road Materials andPavement Design Journal.

Onifade was responsible for the data analysis and writing of the paper. Jelagintook part in writing the paper. Jelagin, Kringos and Birgisson provided guidanceduring the work.

Paper III

Onifade, I., Balieu, R., Birgisson, B. (2015), Energy-Based Damage and FractureFramework for Viscoelastic Asphalt Concrete. Submitted to the Journal of Engi-neering Fracture Mechanics.

Onifade was responsible for the development of the framework, writing the paperand the data analysis. Balieu and Birgisson provided guidance during the work.

23