Characterization of asphalt mixtures and bitumen to minimize shear-related distresses in asphalt pavement : state of the art

(1)

Characterization of asphalt

mixtures and bitumen to

minimize shear-related

distresses in asphalt pavement

State of the art

ograf Satu AB, V

TI VTI rapport 1056A_{Published 2020}

vti.se/publications Jiqing Zhu

Ehsan Ghafoori Yared Dinegdae

(2)

(3)

VTI rapport 1056A

Characterization of asphalt mixtures and

bitumen to minimize shear-related

distresses in asphalt pavement

State of the art

Jiqing Zhu

Ehsan Ghafoori

Yared Dinegdae

(4)

Authors: Jiqing Zhu (VTI), Ehsan Ghafoori (VTI), Yared Dinegdae (VTI) Reg. No., VTI: 2018/0519-9.2

Publication No.: VTI rapport 1056A Published by VTI, 2020

(5)

Publication Information – Publikationsuppgifter

Title/Titel

Characterization of asphalt mixtures and bitumen to minimize shear-related distresses in asphalt pavement – State of the art

Karaktärisering av asfaltmassor och bitumen för att minimera skjuvrelaterade skador i asfaltbeläggning – En litteraturstudie

Author/Författare

Jiqing Zhu (VTI, http://orcid.org/0000-0003-1779-1710) Ehsan Ghafoori (VTI, http://orcid.org/0000-0002-5526-5896) Yared Dinegdae (VTI, http://orcid.org/0000-0001-7174-7214) Publisher/Utgivare

Swedish National Road and Transport Research Institute (VTI) VTI, Statens väg- och transportforskningsinstitut

www.vti.se/

Publication No./Serie och nr VTI rapport 1056A

Published/Utgivningsår 2020

Reg. No., VTI/VTI:s diarienr 2018/0519-9.2

ISSN 0347–6030

Project/Projektnamn

BVFF Bitumen selection protocol to minimize shear-related distresses of asphalt pavement BVFF Framtagning av urvalsmodell för bitumen kopplat till skjuvrelaterad nedbrytning Commissioned by/Uppdragsgivare

Swedish Transport Administration (Trafikverket) Language/Språk

English

No. of pages/Antal sidor 79 incl. appendices

(6)

Abstract

The shear stress and strain induced by traffic loading play a critical role in the development of major distresses in asphalt pavements. Aiming to facilitate further experimental studies on bitumen selection for asphalt pavements to minimize the shear-related distresses, this report presents a state-of-the-art literature study on the shear characterization of asphalt mixtures and bitumen. The significance of shear-related distresses in asphalt pavement is analysed. The advance in shear tests of asphalt mixtures and rheological characterization methods of bituminous binders is reviewed. The potential correlation existing between bitumen properties and the shear resistance of asphalt mixtures is discussed.

Considering the shear-related pavement distresses, many previous studies have indicated the importance of asphalt mixture properties such as shear strength, shear modulus, shear phase angle, shear viscosity at maximum phase angle, shear creep compliance and recovery capacity after the loading. The related bitumen properties to these mixture properties are identified and reviewed in this report by looking into models and past experimental studies, including the adhesive and cohesive strength, viscosity, complex shear modulus, phase angle, shear creep compliance and recovery. With this review, a ground is established for future developments on related research topics towards a performance-related bitumen selection protocol.

The recommended future studies include the enhanced understanding of the effects of shear loading on pavement materials and the roles of bitumen and mineral aggregates for shear resistance of asphalt mixtures. Another necessary development is towards new test methods and simple approaches to incorporate the shear-related aspects in practice for asphalt mixture and bitumen characterization through extensive linear viscoelastic test and evaluation of the non-linear failure behaviours, especially for polymer-modified bitumen (PMB) and asphalt mixtures with PMB.

Keywords

(7)

Referat

Skjuvspänningen och skjuvtöjningen som framkallas av trafikbelastning spelar en avgörande roll i utvecklingen av stora skador i asfaltbeläggningar. Denna rapport presenterar en litteraturstudie om skjuvkaraktärisering av asfaltmassor och bitumen. Syftet är att underlätta för experimentella studier om val av bitumen till asfaltbeläggningar för att minimera skjuvrelaterade skador. I rapporten analyseras betydelsen av skjuvrelaterad nedbrytning i asfaltbeläggning. Skjuvtester av asfaltmassor och reologiska karakteriseringsmetoder för bituminöst bindemedel presenteras och granskas. De möjliga korrelationerna som finns mellan bitumenegenskaper och asfaltmassans skjuvmotstånd diskuteras.

Med tanke på de skjuvningsrelaterade beläggningsskadorna har många tidigare studier visat på vikten av egenskaper hos asfalten, såsom skjuvhållfasthet, skjuvmodul, fasvinkel, skjuvviskositet vid maximal fasvinkel, skjuvkrypning och återgång efter belastningen. De motsvarande

bitumenegenskaperna till dessa egenskaper hos asfalten har identifierats genom att undersöka modeller och tidigare experimentella studier, inklusive vidhäftning mot ballasten, kohesion hos bitumen,

viskositet, komplex skjuvmodul, fasvinkel, skjuvkrypning och återgång. Med denna granskning skapas en grund för utvecklingen mot en ny urvalsmodell för bitumen.

De rekommenderade framtida studierna behöver syfta till att förbättra förståelsen för påverkan av skjuvbelastning på vägmaterial och hur bitumen respektive ballast påverkar skjuvmotståndet hos asfaltbeläggningar. En annan nödvändig utveckling är mot nya och enklare metoder för att

karaktärisera asfaltbeläggningars skjuvmotstånd för att kunna inkorporera skjuvrelaterade aspekter i praktiken. För bitumen, särskilt för polymermodifierad bitumen rekommenderas viskoelastisk provning och utvärdering av icke-linjära egenskaper och motstånd mot brott.

Nyckelord

(8)

(9)

Foreword

The study presented in this report is a part of the research project “Bitumen selection protocol to minimize shear-related distresses of asphalt pavement”. The project is financed by the BVFF (Bana väg för framtiden) programme with funding from the Swedish Transport Administration

(Trafikverket) and co-funding from Nynas AB, Peab Asfalt AB and VTI. All support is gratefully acknowledged.

As the project leader, I planned this report, carried out and coordinated the writing work with my colleagues Ehsan Ghafoori and Yared Dinegdae at VTI. Ehsan Ghafoori drafted Chapter 3 of this report while Yared Dinegdae drafted Chapter 2. Their hard work is appreciated. During the period of this study, Abubeker Ahmed, Andreas Waldemarson, Henrik Bjurström and Shafiqur Rahman at VTI provided valuable help to me. I would like to express my gratitude to each of them.

An external expert group was formed for steering the project, consisting of the following members: • Henrik Arnerdal, Trafikverket

• Xiaohu Lu, Nynas AB

•_{Anders Gudmarsson, Peab Asfalt AB} • Michael Langfjell, Peab Asfalt AB

The experts gave me constructive advice and suggestions for planning and organizing this report. I would like to thank all the expert members for their input.

Linköping, April 2020 Jiqing Zhu

(10)

Quality review

An internal peer review was conducted on 14 May 2020 by Safwat Said. Jiqing Zhu has made

adjustments to the final report. Research Director Björn Kalman has thereafter reviewed and approved the report for publication on 11 August 2020. The conclusions and recommendations in the report are those of the authors and do not necessarily reflect the views of VTI as a government agency.

Kvalitetsgranskning

Intern peer review har genomförts 14 maj 2020 av Safwat Said. Jiqing Zhu har genomfört justeringar av slutligt rapportmanus. Forskningschef Björn Kalman har därefter granskat och godkänt

publikationen för publicering 11 augusti 2020. De slutsatser och rekommendationer som uttrycks är författarnas egna och speglar inte nödvändigtvis myndigheten VTI:s uppfattning.

(11)

Summary

Characterization of asphalt mixtures and bitumen to minimize shear-related distresses in asphalt pavement – State of the art

by Jiqing Zhu (VTI), Ehsan Ghafoori (VTI) and Yared Dinegdae (VTI)

Asphalt pavement is a common type of flexible pavement that can provide a smooth road surface for comfortable ride. However, as the pavement ages and experiences repeated loading from the traffic, the road condition may degrade, and distresses begin to accumulate. Pavement distresses may be in different forms and can be caused by different mechanisms. Among others, the traffic loading is one of the main causes to many pavement distresses. Despite numerous past studies on compression, tension and bending types of loading, the impacts of shear loading on pavement structure and materials have long been overlooked. Yet the shear stress and strain induced by traffic loading play a critical role in the development of major distresses in asphalt pavements.

For minimizing the shear-related pavement distresses, the selection of bitumen is very important to ensure a high shear resistance of the asphalt mixture. Aiming to facilitate further experimental studies on bitumen selection for asphalt pavement, this report presents a state-of-the-art literature study on the shear characterization of asphalt mixture and bitumen. The significance of shear-related distresses in asphalt pavement is analysed. The advance in shear tests of asphalt mixtures and rheological

characterization methods of bituminous binder is reviewed and the potential correlation existing between bitumen properties and the shear resistance of asphalt mixtures is discussed.

In previously reported studies, the vertical shear stress in asphalt pavement, mostly due to the traffic loading, has been linked to not only the rutting failure mode but also the top-down cracking, bottom-up cracking, and reflective cracking. The loading-induced horizontal shear stress has been connected to distresses such as corrugation and shoving. The shear properties of asphalt mixtures like shear strength and shear modulus are particularly crucial in the development of permanent deformation. Various test methods have been developed to characterize and evaluate the shear behaviour of asphalt mixtures. The most common ones include the axial compression tests, direct shear test and simple shear test.

The shear behaviour of asphalt mixtures is dependent on its internal structure as well as properties of the raw materials, i.e. both the bituminous binder and mineral aggregates. For a given asphalt mixture, the interaction between the relatively fixed aggregate properties and temperature- and time-dependent bitumen properties determines how resistant the asphalt mixture is to shear loadings. Due to the viscoelastic nature of bitumen, the selection of proper binder is particularly important to ensure that the asphalt mixture has a high shear resistance, especially at high temperatures or low frequencies. The currently common test methods for bitumen evaluation and selection are mostly empirical methods. They may work well for unmodified bituminous binders but can be problematic when it comes to polymer-modified bitumen. New test methods that work for both unmodified and modified binders still need to be developed, evaluated, and validated.

The correlation between bitumen properties and asphalt mixture performance is the basis for bitumen selection. Considering the shear-related pavement distresses, many previous studies have indicated the importance of asphalt mixture properties such as shear strength, shear modulus, shear phase angle, shear viscosity at maximum phase angle, shear creep compliance and recovery capacity after the loading. The related bitumen properties to these mixture properties are identified and reviewed in this report by looking into models and past experimental studies, including the adhesive and cohesive strength, viscosity, complex shear modulus, phase angle, shear creep compliance and recovery. With this review, a ground is established for future developments on related research topics towards a

(14)

performance-related bitumen selection protocol. The recommended future studies include the

enhanced understanding of the effects of shear loading on pavement materials and the roles of bitumen and mineral aggregates for shear resistance of asphalt mixtures. Another necessary development is towards new test methods and simple approaches to incorporate the shear-related aspects in practice for asphalt mixture and bitumen characterization through extensive linear viscoelastic test and evaluation of the non-linear and failure behaviours, especially for polymer-modified bitumen (PMB) and asphalt mixtures with PMB.

(15)

Sammanfattning

Karaktärisering av asfaltmassor och bitumen för att minimera skjuvrelaterade skador i asfaltbeläggning – En litteraturstudie

av Jiqing Zhu (VTI), Ehsan Ghafoori (VTI) och Yared Dinegdae (VTI)

Asfaltbeläggning är en flexibel beläggning som kan packas vid läggningen till en slät vägyta. När beläggningen åldras och utsätts för upprepade belastningar från trafiken börjar skador ackumuleras och vägen försämras. Beläggningsskador kan vara i olika former och orsakas av olika mekanismer men trafikbelastningen är en av de främsta orsakerna till många av de olika beläggningsskadorna. Många tidigare studier av beläggningsskador har fokuserat på effekten av tryck, drag och böjning orsakat av trafikbelastningen. Effekter från skjuvning på vägkonstruktionen och beläggningsmaterial har emellertid ofta förbisetts. Ändå spelar skjuvspänningen och skjuvtöjningen som framkallas av trafikbelastning en avgörande roll för utvecklingen av flera typer av nedbrytningsmekanismer hos asfaltbeläggningar.

För att minimera de skjuvrelaterade beläggningsskadorna är valet av bitumen mycket viktigt för att säkerställa ett högt skjuvmotstånd hos asfaltmassan. Denna rapport presenterar en litteraturstudie om skjuvkaraktärisering av asfaltmassor och bitumen. Syftet är att ytterligare underlätta för framtida experimentella studier kring val av bitumen till asfaltbeläggningar. I rapporten analyseras betydelsen av skjuvrelaterade skador hos asfaltbeläggning. Den utveckling som gjorts av skjuvtester för asfalt-beläggningar och reologiska karakteriseringsmetoder för bituminöst bindemedel granskas. De potentiella korrelationerna som finns mellan bitumenegenskaper och asfaltbeläggningens skjuv-motstånd diskuteras.

I tidigare rapporterade studier har den vertikala skjuvspänningen i asfaltbeläggningen, främst på grund av trafikbelastningen, inte bara kopplats till spårbildning men också till sprickbildning (inklusive top-down, bottom-up och reflektionssprickor). Den horisontella skjuvspänningen som orsakas av trafik-belastning har kopplats till skador såsom korrugering och valkbildning. Skjuvningsegenskaperna, såsom skjuvhållfasthet och skjuvmodul, hos asfaltmassan är särskilt avgörande för utvecklingen av permanent deformationer. Olika provningsmetoder har utvecklats för att karakterisera och utvärdera skjuvbeteendet hos asfaltmassor. De vanligaste är axiala kompressionstester, direkt skjuvtest och enkelt skjuvtest.

Asfaltbeläggningens skjuvningsbeteende är beroende av dess inre struktur såväl som råmaterialens egenskaper, dvs. både det bituminösa bindemedlet och ballasten. För en asfaltmassa bestäms

resistensen mot skjuvbelastningar av interaktionen mellan de relativt fasta ballastegenskaperna och de temperatur- och tidsberoende bitumenegenskaperna. På grund av bitumens viskoelastiska egenskaper är valet av lämpligt bindemedel mycket viktigt för att säkerställa att asfaltmassan får ett högt skjuv-motstånd, särskilt vid höga temperaturer eller låga trafikhastigheter. De för närvarande vanligaste provningsmetoderna för utvärdering och val av bitumen är mestadels empiriska. De kan fungera bra för omodifierade bituminösa bindemedel men kan vara problematiska när det gäller

polymer-modifierade bitumen. Nya metoder som fungerar för både opolymer-modifierade och polymer-modifierade bindemedel behöver fortfarande utvecklas, utvärderas och valideras.

Korrelationen mellan egenskaperna hos bitumen och asfaltmassor är grunden för val av bitumen. Med tanke på skjuvningsrelaterade beläggningsskador har många tidigare studier visat på vikten av asfalt-beläggningens egenskaper, såsom skjuvhållfasthet, skjuvmodul, fasvinkel, skjuvviskositet vid maximal fasvinkel, skjuvkrypning och återgång efter belastning. I rapporten granskas bitumen-egenskaperna relaterade till dessa egenskaper hos asfaltmassan genom att undersöka modeller och tidigare experimentella studier, inklusive vidhäftning mot ballasten, kohesion hos bitumen, viskositet,

(16)

komplex skjuvmodul, fasvinkel, skjuvkrypning och återgång. Med denna granskning skapas en grund för den framtida utvecklingen av en ny urvalsmodell för bitumen. De rekommenderade framtida studierna inkluderar att förbättra förståelsen för hur skjuvbelastning påverkar vägmaterial och hur rollerna fördelar sig mellan bitumen och ballast för att skapa skjuvmotståndet hos asfaltmassan. En annan nödvändig utveckling är mot nya och enklare metoder för att karaktärisera asfaltbeläggningars skjuvmotstånd för att kunna inkorporera skjuvrelaterade aspekter i praktiken. För bitumen, särskilt för polymermodifierad bitumen rekommenderas viskoelastisk provning och utvärdering av icke-linjära egenskaper och motstånd mot brott.

(17)

1. Introduction

1.1. Background

The road network is an essential type of transport infrastructure in the modern society. It facilitates the movement of people and goods from one geographic location to another and, thus, supports the economic and social developments. Good pavement conditions are important for maintaining the road network reliable. As roads age and experience repeated loading from the traffic, however, the

pavement condition may degrade, and distresses begin to accumulate. This has become a global challenge, as the renewal of road infrastructure usually costs a lot of resources for the society.

Pavement distresses may occur in different forms, e.g. rutting, cracks, corrugation, shoving, etc. Their presence is due to various reasons and can be attributed to, for example, construction-, environment- and loading-related factors. One of which, traffic loading is the main cause to many pavement

distresses, as it generates complex stress states dynamically and continuously in the layered pavement structure. Numerous past studies have investigated the mechanical responses of pavement structures and materials under traffic loading. The focus was mostly placed on the compression, tension and bending responses of the structure and materials. However, the impact of shear loading on pavement performance evaluation has long been overlooked, leading to a lack of consideration on shear behaviour in most pavement analysis and design tools.

The traffic-induced shear stress is closely related to major pavement distresses, which can impact the long-term performance of roads. Many previous studies have observed that the magnitude, direction, and distribution of shear stresses play a significant role in the development of various pavement distress types. Vertical shear stress has been linked to rutting failures, while load-induced horizontal shear stress has been connected to distresses such as corrugation and shoving. Thus, more attention to the influence of shear loading is currently needed and further research efforts are still necessary for a better understanding of the shear behaviour of pavement structures and materials.

Asphalt mixture is the predominant material used for road surfaces all over the world. An in-depth and up-to-date understanding of asphalt mixture shear behaviour, from the perspective of material

selection, can assist in minimizing the potential of shear-related distresses in asphalt pavement and thus contribute to maintaining good pavement conditions. Asphalt mixtures usually consist of the bituminous binder, mineral filler, and granular aggregates. Its shear behaviour is dependent on its internal structure as well as properties of the raw materials, i.e. both the bituminous binder and mineral aggregates.

As a binder-bound material, however, asphalt mixture has a very different shear behaviour from unbound materials. The binder in an asphalt mixture significantly influences its response under shear loading. In addition, the shear resistance of an asphalt mixture is highly affected by its rheological properties, yet its viscoelasticity mainly comes from the viscoelastic binder, namely bitumen. Thus, it is of great importance to understand the shear response of asphalt mixtures, especially the response under dynamic shear loading, and find the link between the asphalt mixture response and bitumen properties. This may assist in the bitumen selection process to minimize the potential of shear-related distresses in asphalt pavement. As a ground for the related future developments, this report presents a state-of-the-art literature study on the shear characterization of asphalt mixtures and bitumen.

1.2. Objectives and scope

During the past decades, various characterization methods have been developed to study the shear resistance of asphalt mixtures. At the same time, many previous studies experimentally investigated the bituminous binder and its effects on asphalt mixture properties, particularly with the rheological test methods. In order to facilitate further experimental studies on bitumen selection for asphalt pavement, as well as to establish a state-of-the-art ground for future developments on related research

(18)

topics, this literature study analyses the significance of shear-related distresses in asphalt pavement and summarises the advances in shear tests of asphalt mixtures and rheological characterization methods of bituminous binders.

The scope of this report is the characterization of asphalt mixtures and bitumen for asphalt pavement. Testing of asphalt mixtures in shear is a focus to characterize its response to the shear loading from traffic. Particularly, test methods with dynamic shear loading are prioritised in this report, as they can better reflect the complex stress-strain response within the viscoelastic asphalt mixture under traffic loading. The material properties of asphalt mixtures, for example shear strength, shear modulus, phase angle, etc, are within the scope of this report while the structural response of pavement layers and inter-layer bonding are out of the scope for this literature study.

As most past studies have correlated the shear resistance of asphalt mixtures to the potential of permanent deformation, the main discussion in this report regards the rutting issues of asphalt pavement at high temperatures, although other types of pavement distresses (e.g. cracks, corrugation, shoving) and the shear fatigue characterization of binder are analysed as well. It has previously been mentioned that both the bituminous binder and mineral aggregates have influences on the shear resistance of asphalt mixtures. In this report, however, the focus is placed on binder properties and tests while the effects of mineral aggregates are only discussed briefly. High-temperature properties and evaluation methods of bituminous binders are criticized extensively.

1.3. Report organization

This report consists of five chapters. After this introduction chapter, Chapter 2 analyses the significance of shear-related distresses in asphalt pavement by reviewing the basic knowledge of asphalt pavement structure, materials, traffic loading, stress distribution, distress mechanism as well as the latest developments of prediction models. Chapter 3 criticizes the advances in shear tests of asphalt mixtures, especially the ones with dynamic shear loading, and briefly discusses the role of mineral aggregates in resisting shear loading. Chapter 4 reviews the rheological characterization methods of bituminous binder and analyses the potential correlation existing between binder properties and asphalt mixture shear resistance. At the end, Chapter 5 summarises the main findings and gives recommendations for future work.

(19)

2. Shear-related distresses in asphalt pavement

2.1. Asphalt pavement distresses

2.1.1. Asphalt pavement structure and materials

Pavement structures are usually constructed to provide a smooth riding surface for vehicular transportation. The typical structure of pavement usually consists of several layers of engineered materials, which in addition to providing structural support, protect the native subgrade soil from excessive stress and strain that are induced by the traffic loading. Pavement structures are designed to fulfil several performance criteria, among others, riding quality, skid resistance, light reflective characteristic, and noise. They can be categorized into two broad types: flexible and rigid. Flexible pavement structures are the most common and can, as the name implies, deflect or flex under traffic loading. The flexibility of pavement structures facilitates the transmission of wheel loading from the surface to the subgrade without compromising the integrity of each layer. This is achieved by spreading out the wheel loading on a larger area, which reduces the magnitude, intensity and concentration of the induced stress and strain (Huang, 2004).

Asphalt pavement, which is a typical example of flexible pavement, is constructed by placing the binder-rich asphalt mixture layer(s) over the unbound granular materials. The typical asphalt pavement structure consists of four main “grand layers”. The asphalt mixture “grand layer” is at the top and usually made of several sublayers of asphalt mixtures with various bitumen types and granular aggregates. The base and subbase layers, which are commonly known as unbound granular layers, are usually made of fine and coarse aggregates. The subgrade or native soil, upon which the pavement super-structure is constructed, is considered as a foundation (Ameri et al., 2011). These layers are comprised of materials that exhibit complex behaviours under varying traffic and environmental loadings. The mechanical properties of these materials are highly dependent on factors such as temperature, moisture, rate of loading, stress level and stress history, thus making it difficult to accurately model and predict pavement response and performance. A typical Swedish pavement structure of medium- and high-traffic roads is shown below in Figure 1 (Trafikverket, 2014), with 3 bound sublayers of asphalt mixtures and 2 unbound layers of granular materials (base and subbase). It should be noted that the subgrade (“terrass” in Swedish) is not shown in Figure 1.

Figure 1.Typical Swedish pavement structure of medium- and high-traffic roads. From top to bottom: bound wearing course (“bundet slitlager” in Swedish); bound binder course (“bundet bindlager”); bound base course (“bundet bärlager”); unbound base course (“obundet bärlager”); unbound subbase course (“obundet förstärkningslager”); and the subgrade “terrass” is not shown. Source: Trafikverket (2014).

(20)

The design of asphalt pavements is mainly governed by factors such as the expected traffic, functional requirements, design service life and prevailing climatic condition. These factors control and guide, among others, the structure design, material selection, reliability level, performance requirements and construction techniques. The construction is usually performed by placing and compacting each layer individually, providing an appropriate level of cohesion within and adhesion between layers.

Appropriate drainage is required for the unbound granular layers, as their mechanical properties are highly sensitive to moisture. Factors that may have negative impacts on the long-term pavement performance, such as workability and segregation, require extra attention during the construction stage (NCHRP, 2004).

The material for the bound layer of asphalt pavement, namely the asphalt mixture, is a heterogenous mixture of mineral aggregates, bituminous binder, and air voids. Figure 2 presents an X-ray CT (computed tomography) slice from a scanned asphalt mixture sample, showing the typical internal structure of asphalt mixtures (Onifade et al., 2016). As the bound layer in contact with traffic, it is expected to sustain high stress and strain, which primarily governs the way it is designed and constructed. The design of asphalt mixtures is primarily performed using the volumetric-based approach where the proportion of its constituents (i.e. aggregates, air voids and binder) are optimized based on volume and weight. Volumetric parameters such as the effective binder content, air voids, voids in mineral aggregate (VMA) and voids filled with asphalt (VFA, for which the “asphalt” means bituminous binder) are commonly used to optimize asphalt mixtures for a given design condition. In addition, special attention is given to factors such as the (nominal) maximum aggregate size, aggregate gradation and binder type and content, as these properties play a critical role in the long-term

performance of mixtures (AASHTO, 2003). Mixture design approaches that use performance as a design criterion have also been employed for the optimization of asphalt mixtures. Mechanical properties such as modulus and creep compliance, which can be correlated with the long-term performance, are primarily used in performance-based designs. The introduction of new technology and materials, such as the use of additives, modifiers, rejuvenators, and extenders to address

performance-, economy- and environment-related problems, has further complicated the asphalt mixture design process (Ksaibati and Butts, 2003).

Figure 2. An X-ray CT slice from a scanned asphalt mixture sample. Source: Onifade et al. (2016).

2.1.2. Traffic loading-induced distresses in asphalt pavement

Asphalt pavements are designed and constructed to sustain and endure various types of traffic- and environment-related distresses. To effectively achieve the purposes they are built for, asphalt pavements are required to incorporate some performance characteristics and qualities. Flexibility, which is the ability to bend or deform without fractures, and stability, which is the resistance against shear flow, respectively allow asphalt mixtures to resist the brittle fracture and permanent deformation. Fatigue resistance allows asphalt mixtures to bend repeatedly without cracking while tensile strength provides resistance to fractures due to various loadings. Impermeability allows asphalt mixtures to

(21)

hinder the flow of fluids, both air and water, through the layers and thus protects the integrity of the pavement structures. Durability governs the mixture’s resistance to aging and weathering, thus allowing asphalt pavements to last longer. A smooth riding surface with appropriate textures allows asphalt pavements to retain adequate friction between the road surface and tires (NCHRP, 2004). Asphalt pavements fail due to a multitude of failure modes that occur either independently or

concurrently. Most of which are a direct result of the traffic loading, which induces three-dimensional (3D) contact stresses in the pavement body: vertical, transverse, and longitudinal. The vertical stress is non-uniform and is primarily caused by the bending stiffness of the tire structure. The transverse stress is mainly caused by the restricted inward movement of the tire ribs while the longitudinal stress is primarily attributed to the tire-pavement friction force (De Beer et al., 1999). The complex 3D stress state near the pavement surface increases the potential for pavement damages and failures. Distresses such as top-down cracking, near surface cracking and rutting, can be directly linked to the stress distribution pattern in the pavement. This stress distribution is also influenced by moving loadings that can induce stress rotations and loading rates. These may accelerate the pavement deterioration

(Hernandez et al., 2013).

Rutting, or permanent deformation, is caused by a complex combination of densification and shear flow and manifests itself as vertical depressions along the wheel path on the pavement surface. The lateral movement of pavement materials or consolidation under traffic, post compaction during service life, and excess fines in the mixture are some of the factors to which the permanent deformation are attributed. Cracking, which takes different forms depending on the driving failure mechanism, can be categorized into top-down (usually longitudinal), bottom-up, transverse, thermal, and frost heave cracking. Cracking is mainly associated with factors such as weak surface, base or subgrade and poor drainage. Ravelling is a surface failure, caused by the lack of integrity, e.g. weak adhesion/cohesion, poor compaction, and dirty materials (Hernandez et al., 2013). Figure 3 shows the rutting and top-down cracking distresses in asphalt pavements.

(A) (B)

Figure 3. Traffic load-induced distress in asphalt pavement: (A) Rutting. Source: Verhaeghe et al. (2007). (B) Top-down cracking. Source: Harmelink and Aschenbrener (2003).

2.1.3. Stress distribution in asphalt pavement

Accurate understanding of the mechanical response of pavements under traffic and environment loadings is important for the accurate prediction of pavement distress potential and thus estimation of pavement design life as well as life cycle cost. Traditionally, pavement analysis and design are performed assuming vertical uniform circular contact stress, induced by traffic. However, the actual tire-pavement interaction is very complex and involves stresses of various magnitudes in three dimensions. The 3D stress distribution is highly influenced by factors such as pavement stiffness,

(22)

pavement structure, loading rate, and tire material and geometry. Figure 4 presents the 3D contact stress distribution for a free rolling truck tire of 25 kN wheel loading and 500 kPa tyre inflation pressure. In addition, environmental loadings such as temperature and moisture, can also induce complex responses in the pavement, which is governed by the nature and characteristics of bound and unbound materials and the spatial distribution of the environmental inputs (De Beer et al., 2012).

(A) (B) (C)

Figure 4. 3D contact stress distribution for a free rolling truck tire of 25 kN wheel loading and 500 kPa tyre inflation pressure: (A) Vertical; (B) Transverse; (C) Longitudinal. The arrows indicate the direction of traffic. Source: Woldekidan (2011).

Previous studies have shown the relevant impacts that the 3D contact stress distribution can have on the pavement response and distress development. One important aspect of how the 3D contact stress differs from a conventional assumption is on the near-surface stress states, where contact stresses are larger in magnitude and more localised in the 3D reality. A higher tangential contact stress, which increases the potential for the development of top-down cracking, primary rutting and fatigue damage, has been observed when the measured contact stresses of a moving load is simulated in a validated 3D finite element pavement model (Wang and Al-Qadi, 2010; Hernandez et al., 2013). As there is an established link between the transverse traction and shear flow (unrecoverable deformation), the magnitude of the transverse contact stress is also relevant in addition to the vertical compressive stress. Furthermore, the strain in the transverse direction is more sensitive to the nonuniformity in the stress that might increase the potential of bottom-up fatigue cracking. In Figure 5, it is also observed that surface tensile stress, which results in top-down cracking, reaches its peak at some distance away from the load centre (Gu et al., 2018). The perpetual pavement design that aims to achieve long-lasting structures through thick and binder-rich asphalt layers has also been linked with the development of shear strain on the surface near the tire edge. Figure 6(A) presents the shear strain distribution in an asphalt pavement along the longitudinal direction, considering the impact of speed. This is related to the deformation (leading to unevenness) along the traffic direction and corrugation failures. Figure 6(B) presents the shear stress distribution in the vertical-transverse plane. Another more detailed study on computed distribution of shear stress and strain in asphalt pavement was reported by Wang and Al-Qadi (2009), where more information about their computation can be found.

(23)

(A) (B)

Figure 6. Shear distribution in asphalt pavement: (A) Shear strain along longitudinal direction. Source: Wang and Al-Qadi (2010). (B) Shear stress in vertical-transverse plane. Source: NCAT (2016).

The stress distribution in an asphalt mixture layer is highly dependent on the stiffness and thickness of the layer. Past research has shown that the bending-induced tensile strain at the bottom of the asphalt mixture layer is usually greater in thin pavements than in equivalent thick pavements. On the other hand, computation shows that the approximate pavement shear centre (i.e. the concentrated strain near the loading edge) is closer to the road surface in thicker pavements (Roque et al., 2010). Overall, the transverse and longitudinal strain distribution with depth has been shown to be compressive in the upper half and inverted tensile in the lower part of the layer. As for the vertical strain, Said et al. (2011) reported the distribution of permanent vertical strains in a semi-infinite asphalt mixture layer as illustrated in Figure 7 where the upper illustration represents the strain due to compressibility while the lower one represents the strain due to flow deformation.

Figure 7. Permanent vertical strain in micro strain (με) in a semi-infinite asphalt mixture layer without lateral wander: The upper represents the strain due to compressibility while the lower represents the strain due to flow deformation. Source: Said et al. (2011).

The wheel loading and tire configuration (i.e. dual tire assembly or wide base) play important roles in the development of near-surface stress distributions. It was observed that a higher wheel loading mainly increases the tensile strain at the vicinity of the tire edges while a higher tire inflation pressure primarily increases the shear strain under the tire centre. This can be explained by the fact that the wheel loading level is associated with vertical contact stress under the tire edge ribs whereas the tire pressure mainly affects the vertical and transverse contact stresses under tire centre ribs (Wang et al., 2012).

(24)

2.2. Shear-related distresses in asphalt pavement

The impact of shear loading on pavement performance evaluation has been overlooked in most pavement analysis and design tools. Nevertheless, several studies have shown that shear stress and strain play a critical role in the development of many distress types (Hernandez et al., 2013). The vertical shear stress has been linked to major failure modes such as permanent deformation

(densification), top-down cracking, bottom-up cracking and reflective cracking, while loading-induced horizontal shear stress has been connected to distresses such as permanent deformation (shear flow), corrugation, shoving and inter-layer failures.

Shear flow, namely unrecoverable deformation without volume change as shown in Figure 8, and densification (i.e. volume change) are the two major causes to permanent deformation of asphalt pavement. It has been shown that shape distortion induced by shear flow contributes immensely to rutting, especially in the tertiary stage (Figure 9) of permanent deformation. Therefore, shear properties such as shear strength and shear modulus play an important role in the development of permanent deformation (Li et al., 2011).

(A) (B)

Figure 8. Permanent deformation of asphalt pavement: (A) Shear flow-induced rutting. Source: FHWA (1998). (B) Computed directions of permanent displacements. Source: Sousa et al. (1991).

(25)

The shear strength of asphalt mixtures is mainly derived from two sources: cohesion and friction angle. Cohesion (𝑐𝑐) reflects the adhesion or binding mechanism of the binder whereas friction angle (𝜙𝜙) describes the interlocking capability of the aggregate matrix from the applied loading. This behaviour can be analysed using the Mohr-Coulomb failure theory as follow:

𝜏𝜏𝑓𝑓 = 𝑐𝑐 + 𝜎𝜎𝑓𝑓 𝑡𝑡𝑡𝑡𝑡𝑡𝜙𝜙 (1)

where 𝜏𝜏𝑓𝑓 and 𝜎𝜎𝑓𝑓 are respectively the shear strength and normal stress at failure. After the failure

envelope is determined, it can be readily derived how far a stress state is from the failure, as shown in Figure 10. The shear stress to strength ratio 𝜏𝜏/𝜏𝜏𝑓𝑓 has been related to asphalt mixture rutting

performance and employed for permanent deformation modelling (Li et al., 2010, 2011).

Figure 10. Representation of stress state using Mohr-Coulomb failure theory. Source: Li et al. (2010). The shear modulus of asphalt mixtures, also called modulus of rigidity, is a measure of its resistance to shear strain. The following relationship exists between shear modulus and the currently more

commonly used dynamic modulus:

|𝐺𝐺

∗

_{| =}

|𝐸𝐸∗|

2(1+𝜈𝜈) (2)

where |𝐺𝐺∗_{| is (dynamic) shear modulus; |𝐸𝐸}∗_{| is dynamic (Young’s) modulus; and 𝜈𝜈 is the Poisson’s}

ratio. At low temperatures and high frequencies, asphalt mixture’s shear modulus is about 1/3 of the dynamic modulus with the Poisson’s ratio at approximately 0.5 (Harvey et al., 2001). As the

temperature and strain amplitude increase, however, the Poisson’s ratio becomes more time dependent. The concept of asphalt mixture viscosity based on time-dependent shear modulus has been discussed in the context of asphalt mixture rutting prediction (Said et al., 2011, 2013).

In addition to rutting, asphalt pavement cracking is associated with shear loading as well. Hernandez et al. (2013) have shown that the maximum shear strain is greater than the tensile strain at the bottom of a typical asphalt mixture layer for both uniform one-dimensional (1D) and non-uniform 3D contact stresses. Moreover, for thick flexible pavements at intermediate to high temperatures, the vertical shear strain at shallow depths, up to 10 cm below the surface, is significantly greater than the critical transverse and longitudinal tensile strain at the surface and bottom of the layer. This shear strain can result in the initiation and propagation of near-surface fatigue cracking. For a dual wheel configuration with 1D and 3D loading conditions, the maximum shear strain is concentrated at the upper part of the asphalt mixture layer, at the tire edge. It can easily be concluded that the shear strain and low

confinement at the tire edge can result in the formation of significant cracking or shear flow, or both, at the pavement near surface (Al-Qadi et al., 2018).

(26)

Surface shear stress, besides tensile stress, can also contribute to the initiation and propagation of top-down cracks, which may develop at various locations on the surface in different orientations with respect to the traffic. Top-down cracks in asphalt pavement often initiate because of severe aging or thermal fatigue damage of the asphalt mixture near the surface. They may then propagate due to tensile and shear stresses induced by vehicle wheels (Ameri et al., 2011). Despite this evidence, shear-mode or mixed-shear-mode fracture behaviours have received little attention for top-down cracks.

Furthermore, the propagation of bottom-up and reflective cracks is partially due to shear stress as well. Figure 11 illustrates the stress distribution causing the propagation of an underlying crack.

Figure 11. Stress distribution at the tip of an underlying crack during a wheel passing, causing crack propagation. Source: Lytton (1989).

2.3. Shear-based models and tools for predicting distresses

Efforts have been made to incorporate shear properties for the prediction of rutting in asphalt

pavements. Fwa et al. (2004) developed a model for asphalt mixture rutting prediction based on shear properties, including cohesion and friction angle. Some other shear-based rutting prediction models incorporate the shear stress to strength ratio to represent the rutting resistance of pavement materials (Su et al., 2008; Li et al., 2010, 2011; Kim et al., 2017). Most of these models were based on the power law. However, a power law model can only predict the secondary stage of permanent deformation while ignoring the tertiary shear flow. Moreover, the power law model coefficients are usually sensitive to the data range selection (Diaz and Archilla, 2008). These aspects may limit the application of these models in practice. Nevertheless, the model by Li et al. (2010, 2011), based on the Hoerl model, could predict the permanent deformation all the way up to the tertiary flow, exceeding the minimum permanent strain rate. These previous studies have shown the potential of using shear-related properties for the prediction of permanent deformation in asphalt pavements, although it remains questionable to use strength parameters for predicting the performance of viscoelastic asphalt mixtures under repeated loading.

As for shear-based mechanistic-empirical tools, the CalME program (Ullidtz et al., 2010) assumes the elastic theory and predicts the rutting development in asphalt mixture layers using the permanent strain determined by the repeated simple shear test at constant height and considering factors such as shear stress, elastic shear strain and loading number (Deacon et al., 2007). It should be noted that CalME

(27)

estimates the rut depth based on the shear stress at only one point in the structure at 50 mm under the tyre edge. Said et al. (2011, 2013) proposed a linear viscoelastic model and developed a program, named PErmanent Deformation of asphalt concrete layer for ROads (PEDRO), for the calculation of permanent deformation of an asphalt mixture layer over the entire area that is important for rutting estimation. It employs the concept of viscosity for asphalt mixtures in the linear viscoelastic range and can separately predict the effects of post-compaction (volume change) and shear flow on the asphalt mixture rutting development. The input material properties for the PEDRO model include the viscosity at the maximum phase angle of the asphalt mixture determined by a dynamic shear modulus test (Said and Hakim, 2016; Said and Ahmed, 2018; Jelagin et al., 2018; Said et al., 2020).

Shear-related aspects have also been considered for prediction models of crack initiation and propagation in asphalt pavements. Experimental findings and material mechanics were employed in the development of these models. The mechanics-based models rely on theories such as fracture mechanics and viscoelastic continuum damage (Masad et al., 2008; Sabouri and Kim, 2014). Furthermore, advanced modes for material behaviour characterization and response prediction were incorporated into these models. For example, for top-down cracking initiation in thicker pavements, a shear-induced near-surface tension is incorporated as a driving mechanism (Roque et al., 1999, 2002).

(28)

3. Shear behaviour characterization of asphalt mixtures

Many different test methods and setups have been proposed for characterizing the shear behaviour of asphalt mixtures. In early days, the related methods and setups were originally derived from the testing in soil mechanics and adapted for asphalt mixtures, e.g. the axial compression tests, direct and simple shear tests. As the recent development in this area, some new methods and setups have been developed especially for asphalt mixtures. This chapter will review these test methods and setups in the context of asphalt mixture characterization. They are categorised into two groups: axial compression tests with a shear component and shear tests. Most of these tests can be carried out under static, repeated, and dynamic loading modes. However, the focus of this chapter is placed on the dynamic loading mode, as it can better reflect the complex stress-strain response within the viscoelastic asphalt mixture under traffic loading. In addition, the simulative tests for correlation with pavement performance are discussed as well. As most past studies have correlated the shear resistance of asphalt mixtures to the potential of permanent deformation, the discussed simulative tests mainly regard the rutting failures of asphalt pavement at high temperatures. At the end of this chapter, the effects of raw materials on shear resistance of asphalt mixtures are analysed briefly.

3.1. Axial compression tests with a shear component

An axial compression test can be either the unconfined uniaxial test or confined triaxial test. In such a test, the applied compressive stress generates a shear stress in the tested sample and can cause its failure along the shear plane, as illustrated in Figure 12. The setup applies relatively uniform stress states and holds the possibility of varying the axial loading and confining pressure to represent

different stress states, with which a shear stress component is associated. These aspects have made the axial compression test one of the most extensively used methods in the past (Sousa et al., 1994a). In Europe, most of these axial compression methods are standardised in EN 12697-25 and EN 12697-26.

Figure 12. Shear stress (τ) generated in the confined triaxial compression test, adapted from Zhang et al. (2017).

3.1.1. Unconfined uniaxial compression test

The unconfined uniaxial compression test is simple to conduct (ranked the highest for simplicity in the SHRP-A-415 report), as there is no lateral pressure applied to the sample, which is usually cylindrical. Hence, it normally takes shorter time than the test with confinement. However, the lack of confining forces limits the uniaxial test to be conducted at much lower stress levels than it is under real field conditions. A typical test setup is shown in Figure 13.

(29)

Figure 13. Unconfined uniaxial compression test setup. Source: Ramirez Cardona et al. (2015). The uniaxial loading can be in static, repeated, and dynamic modes. The static loading mode leads to a static creep test, which is one of the most common tests for asphalt mixtures due to its simplicity and widespread equipment availability in most laboratories. The repeated loading cycles can better simulate traffic conditions with commonly block or haversine pulse loadings and thus enable more reasonable test results as compared with the static creep test. The dynamic loading that is applied on unconfined samples, often in the linear viscoelastic range, allows the determination of dynamic modulus |𝐸𝐸∗_{| and phase angle 𝜑𝜑, despite the complexity of measurement.}

In the linear viscoelastic range, the stress-strain relationship of asphalt mixtures under dynamic loading is defined by the complex modulus 𝐸𝐸∗_{. It is a complex number with the real and imaginary}

portions written as:

𝐸𝐸∗_{= 𝐸𝐸}′_{+ 𝑖𝑖𝐸𝐸}′′ ₍₃₎

where 𝐸𝐸′_{is the real portion often called storage or elastic modulus; and 𝐸𝐸}′′_{is the imaginary portion}

often called loss or viscous modulus. The absolute value of the complex modulus, |𝐸𝐸∗_{|, is defined as}

the dynamic modulus. It represents the ratio between the maximum dynamic stress 𝜎𝜎0 and the

maximum recoverable axial strain 𝜀𝜀0:

|𝐸𝐸

∗

_{| =}

𝜎𝜎0

𝜀𝜀0 (4)

In addition, the phase angle 𝜑𝜑 represents the phase difference between the dynamic stress and strain, as presented in Figure 14. It is a measure of the viscous properties. The tangent of phase angle equals to the ratio between the viscous modulus 𝐸𝐸′′_{and the elastic modulus 𝐸𝐸}′_{, written as:}

tan(𝜑𝜑) =

𝐸𝐸_𝐸𝐸′′_′ (5)

For a purely elastic material, 𝜑𝜑 = 0°. For a purely viscous material, 𝜑𝜑 = 90°. For asphalt mixtures, however, the phase angle is between 0° and 90° depending on temperature and loading frequency.

(30)

Figure 14. Dynamic stress and strain of the dynamic modulus and phase angle test. Source: Witczak et al. (2002).

Related to the measurement of dynamic modulus and phase angle, a biaxial test method has also been widely used, i.e. the diametrical indirect tension (IDT) test with dynamic loading. This type of test is carried out in the diametrical plane of a typically cylindrical asphalt mixture sample, as shown in Figure 15. Since the IDT test requires a much smaller sample size (thinner), it is more appropriate than the axial compression test for the evaluation of cores from the field, despite the complexity of

measurement with dynamic loading. In such a biaxial test within the linear viscoelastic range, the complex modulus 𝐸𝐸∗_{of asphalt mixtures represents the ratio between the biaxial stress (𝜎𝜎}

𝑥𝑥− 𝜈𝜈𝜎𝜎𝑦𝑦) and

horizontal strain 𝜀𝜀𝑥𝑥, written as follow:

𝐸𝐸

∗

₌

𝜎𝜎𝑥𝑥−𝜈𝜈𝜎𝜎𝑦𝑦

𝜀𝜀𝑥𝑥 (6)

where 𝜎𝜎𝑥𝑥 is the complex stress in horizontal direction (perpendicular to loading); 𝜈𝜈 is the Poisson’s

ratio; and 𝜎𝜎𝑦𝑦 is the complex stress in vertical (loading) direction. Kim et al. (2004) compared the axial

compression and IDT tests with 12 asphalt mixtures and concluded that the dynamic modulus and phase angle results from the two methods are in good agreement.

Figure 15. Indirect tension test setup. Source: Kim et al. (2004).

However, the biaxial nature of the IDT test does not lead to uniform stress distribution in the sample, especially when considering the shear stress associated with permanent deformation issues. In an IDT test, the only state of stress that is relatively uniform is the horizontal tension along the loading line, i.e. 𝜎𝜎𝑥𝑥(𝑦𝑦) as shown in Figure 16. The other states of stresses are not uniform, which can cause great

difficulties in correlating the test results with the field conditions. In addition, the stress state created by the IDT test is very sensitive to the shape of the sample. Brown et al. (2001) claimed that the test result can be highly influenced at high temperatures or loading levels as the shape of the sample is

(31)

subjected to change, because this change of shape can significantly affect the state of stress during the test. Furthermore, due to the tensile nature of the IDT test, the result is largely a reflection of the binder properties while the aggregate has less influence than in the axial compression test (Sousa et al., 1994a). One last concern about the IDT test is that its loading direction is usually perpendicular to the sample compaction direction, which is not the case for the axial compression test or the real pavements.

Figure 16. Stress distribution by the indirect tension test. Source: Fettahoglu and Sel (2016).

3.1.2. Confined triaxial compression test

Similar to the unconfined uniaxial compression test, the confined triaxial compression test is also carried out on cylindrical samples of asphalt mixtures. This test setup is more sophisticated than the uniaxial one, as it considers the confining pressure around the sample and is thus a better simulation of the field conditions. A typical test setup is shown in Figure 17. Because of the presence of confining pressure, the results of this test can be better correlated to the real pavements. However, the

application of lateral pressure makes it more difficult and time-consuming to perform this test as compared with other tests. The triaxial test can be carried out under static, repeated, and dynamic loading modes and holds the possibility of varying the axial loading and confining pressure to

represent different stress states. As mentioned previously, a shear stress component is associated with the generated state of stress in the sample.

(32)

The static loading mode with confining pressure leads to a triaxial creep test, which has been proven to be a better method than the uniaxial creep for predicting the field performance of asphalt mixtures (Roberts et al., 1996). The triaxial test with repeated loading is more complex but has a better

representation of the traffic loading in real field conditions. The dynamic loading applied on confined samples enables the determination of dynamic modulus |𝐸𝐸∗_{| and phase angle 𝜑𝜑 in the linear}

viscoelastic range.

Recent developments in asphalt mixture testing equipment has made it relatively simple to perform the triaxial tests. The asphalt mixture performance tester (AMPT), earlier also called simple performance tester (SPT), is one of the examples that have made the test preparation and application of confining pressure easier than it used to be in older test setups. Two typical AMPTs are shown in Figure 18. They are often used to perform the flow time test with static loading, flow number test with repeated loading and dynamic modulus test with dynamic loading (Irfan et al., 2018).

(A) (B)

Figure 18. Asphalt mixture performance testers: (A) By Interlaken Technology Corporation; (B) By IPC Global. Source: FHWA (2013).

Moreover, triaxial tests are also used to determine the shear strength of asphalt mixtures and develop the Mohr-Coulomb failure envelope (discussed in Section 2.2) by applying different confining pressures. Such a measurement is often in the displacement-controlled mode with a constantly increasing loading. Three or more confinement levels are usually needed to accurately determine the failure envelope. According to Witczak et al. (2002), the cohesion (𝑐𝑐) values for dense-graded asphalt mixtures are typically in the range of 34.5 kPa to 241.3 kPa while the friction angle (𝜙𝜙) values usually range from 35° to 48°. Li et al. (2010, 2011) investigated the effects of binder content and temperature on shear properties of asphalt mixtures (Figure 19) and incorporated the shear stress to strength ratio 𝜏𝜏/𝜏𝜏𝑓𝑓 into a permanent deformation model for rutting performance prediction.

(33)

(A) (B)

Figure 19. Effects of (A) binder content and (B) temperature on shear properties of asphalt mixtures. Source: Li et al. (2010, 2011).

3.1.3. Partial triaxial compression test

The conventional uniaxial and triaxial tests were the inspirations for developing the partial triaxial compression test (PTCT). This test has been suggested as a method for evaluating the dilatancy of asphalt mixtures. Not like in conventional triaxial tests, however, the PTCT test avoids applying the confining pressure but uses pedestals of smaller area than the sample diametrical surface (Figure 20), i.e. only as big as three times of the nominal maximum aggregate size of the asphalt mixture (Zhang et al., 2016). It was claimed that the unloaded area of sample in PTCT test could provide the required confinement without imposing external confining pressures. The test was carried out at 55 °C on cylindrical samples with diameter of 100 mm and thickness of 80 mm. It gave the results of volume strain and its relationship with the stress. Since this test method is developed very recently and not so many studies have been conducted so far, it is difficult to evaluate its effectiveness for predicting the field performance of asphalt mixtures. It is also noted that the PTCT is in a similar form as the so-called double punch test (Wen et al., 2013), which is shown in Figure 21 and reportedly an alternative test method to IDT.

(A) (B)

Figure 20. Partial triaxial compression test setup: (A) Photo of the tester; (B) The schematic diagram. Source: Zhang et al. (2016).

(34)

Figure 21. Double punch test setup. Source: Wen et al. (2013).

3.1.4. Simple punching shear test

The simple punching shear test (SPST) method was developed to evaluate the shear properties related to the rutting performance of hot-mixed asphalt mixtures in the field (Faruk et al., 2015). The SPST is performed on cylindrical samples with the diameter of about 150 mm (6 inches) and thickness of around 64 mm (2.5 inches). The loading strip, which is applied from the top of the sample, is a cylindrical block with the diameter of about 38 mm (1.5 inches). The fixture placed under the sample is a hollow cylinder as shown in Figure 22. The test is carried out in a displacement-controlled mode with a monotonic axial compressive loading. This test method has been carried out between 40 °C and 60 °C for measuring the shear strength, modulus, and strain energy of several asphalt mixture types (Walubita et al., 2018, 2019). It was claimed that the SPST is a viable laboratory test method for routine use to evaluate shear properties of asphalt mixtures related to rutting performance. However, further validation of this method is still needed as concluded by Walubita et al. (2019).

(35)

3.1.5. Uniaxial penetration test

The uniaxial penetration test (Bi, 2004; Bi and Sun, 2005) is a simple test method in which a steel rod with the diameter of 28.5 mm is used for imposing an either static or cyclic loading on a sample of 100 mm diameter (bigger sample is possible with bigger rod). The loading is applied by means of a

conventional loading frame from the top as shown in Figure 23. The static test is carried out in a displacement-controlled mode with the recommended loading rate of 1 mm/min, whereas the cyclic loading has been carried out in the stress-controlled mode with a sinusoidal compressive loading of 10 Hz. This test has been claimed to show promising results compared with other shear-related tests. However, when comparing the results of this test with the confined triaxial compression test, no quantitative conclusions could be drawn. The static version of this test has been standardised in the Chinese specification JTG D50-2017 while further development still goes on (Ji et al., 2018). More validation results from the field are still necessary before a wider use of this method in practice.

Figure 23. Uniaxial penetration test setup. Source: Chen et al. (2006).

3.2. Shear tests for asphalt mixtures

Due to the complexity of performing axial tests, especially the confined triaxial test, many test methods were developed in shear to evaluate the shear behaviour of civil engineering materials. The best known one is probably the direct shear test of soil, which allows to independently control the normal and shear stresses but forces the sample to fail along a predetermined plane (Figure 24). Some of these shear tests have been employed or adapted for characterizing the shear behaviour of asphalt mixtures, which has further inspired the development of new shear test methods especially for asphalt mixtures. Compared with the axial compression test, which represents only the state of stress directly beneath the tire but not anywhere else (Sousa et al., 1994a), the shear test can directly measure the effects of a specific stress state. This assists in better simulating the real field conditions of asphalt pavement.

(36)

3.2.1. Direct shear test

The shear behaviour of asphalt mixtures has been evaluated by various test methods. The direct shear test is a “guillotine” type of test mostly for interface bonding measurement, as in Figure 25(A). Different versions of direct shear devices have been widely used for characterizing the interface bonding between asphalt mixture layers as well as between the asphalt mixture and steel bridge deck (Raab et al., 2009; Yao et al., 2016). However, Yin et al. (2015) employed this test method for

measuring the shear stiffness and shear strength of asphalt mixtures and investigated the shear fracture behaviour at different temperatures. It is usual for this method that the loading is static and that no normal force is applied, although some complex and uncommon devices could have such options.

(A) (B)

Figure 25. Shear stress distribution in different shear tests: (A) Direct shear test; (B) Simple shear test. Source: Raab et al. (2009).

3.2.2. Simple shear test

In contrast to a direct shear test, the simple shear test applies a shear force from the front surface of the sample, generating a more even shear stress distribution within the whole sample as in Figure 25(B). This method was originally used in soil mechanics with a test setup as illustrated in Figure 26(A) leading to a shear deformation of the cylindrical sample as shown in Figure 26(B). It is simple to perform and does not have the complexity of the triaxial test procedure. This test method also allows the application of an independent normal force that is perpendicular to the shear direction. Thus, the simple shear test, after adapted for asphalt mixtures, holds the possibility of simulating more

representative stress states of asphalt mixtures and the unrecoverable shear flow related to the rutting performance of asphalt pavement.

(A) (B)

(37)

The simple shear test setup can apply static, repeated, and dynamic loadings over a range of

frequencies. The outputs of this test in the creep mode are shear creep modulus versus time and shear permanent deformation versus time; whereas, applying repeated shear loading enables obtaining shear permanent deformation versus cycles as well as resilient shear modulus. The dynamic loading mode of the simple shear test can give shear permanent deformation versus cycles too; but it also enables acquiring the damping ratio and dynamic shear modulus (Sousa et al., 1994a). This test method was recommended in the SHRP-A-415 report as the best method to define the permanent deformation characteristics of asphalt mixtures and was the basis for the development of the Superpave shear tester (Sousa et al., 1994b).

3.2.2.1. Superpave shear tester

The Superpave shear tester (SST) is the analysis system developed by the Strategic Highway Research Program (SHRP) to perform all loading-related performance tests of asphalt mixtures. This equipment is named with “shear”, but it can perform volumetric and uniaxial tests as well (Chowdhury and Button, 2002). It uses a servo-hydraulic loading system to apply axial and shear loadings, either in a confined or unconfined condition inside a controlled temperature chamber. However, SST has been mostly used in the shear mode for evaluating the rutting resistance of asphalt mixtures. The shear configuration of the SST setup enables performing the simple shear test with normally a cylindrical asphalt mixture sample. For the test preparation, two platens are glued to the two sides of the sample as shown in Figure 27(A) and then the sample is attached to the actuator for applying the shear loading, which generates shear deformation in the sample body like in Figure 27(B).

(A) (B)

Figure 27. Superpave shear tester (SST): (A) Setup and a sample with top and bottom platens attached. Source: Bennert et al. (2004). (B) Simulated shear displacement distribution in an asphalt mixture during SST FSCH test. Source: Coleri et al. (2012).

In a range of temperatures, SST can apply various forms of shear loadings to the sample, including sinusoidal, repetitive, and static loadings. The four common tests that can be carried out with SST are: (1) repeated shear at constant height (RSCH) and (2) repeated shear at constant stress ratio (RSCSR), both measuring the cumulative permanent deformation as a function of the number of loading cycles at high temperatures, (3) simple shear at constant height (SSCH), which measures the sample’s creep properties to resist shear strain at low to moderate temperatures (lower than 40 °C), and (4) frequency sweep at constant height (FSCH), which measures a sample’s stress-strain relationship and thus the modulus under a dynamic shear loading. Most of these tests are normally performed under unconfined pressure conditions. The RSCH, SSCH and FSCH tests are standard methods as described in

AASHTO T 320.

Due to the relative simplicity and better correlation with the pavement performance, the FSCH and RSCH tests are the most popular procedures among the SST tests. Chowdhury and Button (2002)

Characterization of asphalt mixtures and bitumen to minimize shear-related distresses in asphalt pavement : state of the art

Characterization of asphalt

mixtures and bitumen to

minimize shear-related

distresses in asphalt pavement

State of the art

VTI rapport 1056A

Characterization of asphalt mixtures and

bitumen to minimize shear-related

distresses in asphalt pavement

State of the art

Jiqing Zhu

Ehsan Ghafoori

Yared Dinegdae

Publication Information – Publikationsuppgifter

Abstract

Referat

Foreword

Quality review

Kvalitetsgranskning

Table of Contents

Summary

Sammanfattning

1.

Introduction

1.1. Background

1.2. Objectives and scope

1.3. Report organization

2.

Shear-related distresses in asphalt pavement

2.1. Asphalt pavement distresses

2.1.1. Asphalt pavement structure and materials

2.1.2. Traffic loading-induced distresses in asphalt pavement

2.1.3. Stress distribution in asphalt pavement

2.2. Shear-related distresses in asphalt pavement

|𝐺𝐺

| =

2.3. Shear-based models and tools for predicting distresses

3.

Shear behaviour characterization of asphalt mixtures

3.1. Axial compression tests with a shear component

3.1.1. Unconfined uniaxial compression test

|𝐸𝐸

| =

tan(𝜑𝜑) =

𝐸𝐸

=

3.1.2. Confined triaxial compression test

3.1.3. Partial triaxial compression test

3.1.4. Simple punching shear test

3.1.5. Uniaxial penetration test

3.2. Shear tests for asphalt mixtures

3.2.1. Direct shear test

3.2.2. Simple shear test

3.2.2.1. Superpave shear tester

_{| =}

_{| =}

₌