The effect of filler type and shape on HMA energy dissipation performance.

The Effect of Filler Type and Shape on HMA Energy Dissipation Performance

Yo han n e s Ge b r e me sk e l k i f l at

Degree Project in highway and railway engineering

stockholm, sweden 2013

TSC-MT 13-009

www.kth.se

The Effect of Filler Type and Shape on HMA Energy Dissipation Performance

Master ’s Thesis

Yohannes Gebremeskel Kiflat

Division of Highway and Railway Engineering Department of Civil and Architectural Engineering

Royal Institute of Technology SE-100 44 Stockholm

TSC-MT 13-009 Stockholm 2013

TSC-MT 13-009

© Yohannes G. Kiflat

The Effect of Filler Type and Shape on HMA Energy Dissipation Performance

Yohannes Gebremeskel Kiflat Graduate Student

Infrastructure Engineering

Division of Highway and Railway Engineering School of Architecture and the Built Environment Royal Institute of Technology (KTH)

SE- 100 44 Stockholm

Kiflat@kth.se

Abstract: Hot mix asphalt pavements require adequate compaction to achieve the required density to resist rutting. The amount of energy required to achieve the optimum degree of compaction depends on the type of gradation, bitumen content, filler type and shape, type of compaction equipment etc. In this study, the net energy required to reduce the specimen volume (size) after each gyration of the superpave gyratory compactor is used as the compaction energy index (CEI) to measure the compactability of the samples. Samples with different filler types and content are used for the analysis.

Effect of fillers on the viscosity of the mastic has been studied previously. Viscosity of mastics in return affects the compactability of the mix in general. In this regard this paper tries to study the effect of fillers on the compaction of hot mix asphalt with the help of the superpave gyratory compactor. Moreover, resistance of the asphalt mix samples against rutting is evaluated using the simple performance test. In this test, the sample is subjected to a hydraulic loading while strain transducers attached to the sample measure the displacement. A computer program receives the displacement data at various frequencies and calculates the dynamic modulus and flow number which are used for the evaluation of the pavement performance.

KEY WORDS: Hot mix asphalt; mastic; fillers; compaction energy index; superpave gyratory compactor; simple performance tester; dynamic modulus; flow number.

ii

Acknowledgement

First of all I would like to thank the almighty God for making me who I am today. I take this opportunity to thank my family for their unconditional love, support and encouragement. I wouldn’t have achieved this without them.

I would like to thank my supervisors Dr.Alvaro Gurain and Ebrahim Hesami for their assistance and indispensable comments in the making of this thesis. I am also thankful to Associate Professor Nicole Kringos for her encouragement and technical advice.

Finally, I would like to thank all who in one way or another contributed to the success of this thesis.

Yohannes Gebremeskel Kiflat Stockholm, June 2013

iv

Biographical Sketch

Yohannes Gebremeskel Kiflat was born on February 14, 1981 in Addis Ababa (Ethiopia) to Gebremeskel Kiflat and Tsega Paulos. He earned his bachelor’s degree in civil engineering from University of Asmara (Eritrea) in 2005.

Yohannes worked as dam engineer, resident road engineer, quality control and quality assurance engineer for a number of contractors and consultants in his native Eritrea, Uganda and the U.A.E

Upon completion of his master’s program, Yohannes plans to work as pavement design consulting engineer.

vi

Dedication

To my parents Gebremeskel Kiflat and Tsega Paulos and all family.

viii

List of Symbols

ߩ Material density

ௗ௨

ௗ௧

Rate of change of the internal energy ߪ

_௜௝

Stress

ܦ

_௝௜

Deformation

݀ݑ Change in internal energy

݀ݒ Change in internal energy due to volumetric change

݀ݏ Change in internal energy due to shear strain ܥܧܫ

_{௦௧௔௕௜௟௜௧௬}

Contact Energy Index

݀

_௘

Change in height

ܵ

_ఏ

Shear Force

ܰ

_ீଵ

Ƭܰ

_ீଶ

Starting and Ending Number of Gyration

E

* Dynamic Modulus

M Phase angle

߳

_଴

Strain amplitude ߪ

_଴

Stress amplitude

ܶ

_௜

Time lag between stress and strain

ܶ

_௣

Period of applied stress

߳

_௣

Permanent Strain οܰ Sampling interval

ܷ Total Mechanical Energy

ܲ Applied Load

ܲ

_{஼௢௡௧௔௖௧}

Contact Load

ܲ

_{஽௬௡௔௠௜௖}

Dynamic Load ܣ

_଴

Area

ܮ

_଴

Length

ܩ

_௠௠

Theoretical Maximum unit weight

ܰ

_{௜௡௜௧௜௔௟}

Initial Number of Gyrations

ܰ

_{ௗ௘௦௜௚௡}

Design Number of Gyrations

ܰ

_௠௔௫

Maximum Number of Gyrations

x

List of Abbreviations

AASHTO - American Association of State Highway and Transportation

Officials

ESALs - Equivalent Single Axle Load HMA - Hot Mix Asphalt

NCHRP - National Cooperative Highway Research Program PG - Performance Grade

SHRP - Strategic Highway Research Program

SGC -Superpave Gyratory Compactor CEI -Compaction Energy Index SPT -Simple Performance Tester

AMPT -Asphalt Mixture Performance Tester JMF -Job Mix Formula

MDL -Maximum Density Line

SEM -Scanning Electronic Microscope FN -Flow Number

TDI -Traffic Densification Index

LVDT -Linear Variable Displacement Transducers

TRVKB Trafikverket (Swedish Traffic administration)

xii

Table of Contents

Abstract . . . . i

Acknowledgement . . . iii

Biographical Sketch . . . v

Dedication . . . vii

List of Symbols . . . . ix

List of Abbreviations . . . . xi

Table of Contents . . . xiii

1. Introduction . . . 1

1.1 Background . . . 1

1.2 Objective . . . 3

1.3 Scope . . . 4

2. Literature Review . . . 5

2.1 Effect of Fillers on HMA . . . 5

2.2 Compaction . . . 6

2.2.1 Compaction of HMA . . . 6

2.2.2 Superpave Gyratory Compactor (SGC) . . . 7

2.2.3 Compaction Energy Index (CEI) . . . . 9

2.3 Performance Tests . . . 13

2.3.1 Dynamic Modulus . . . 13

2.3.2 Flow Number . . . 16

2.3.2 Energy Dissipation . . . 19

3. Methodology . . . 21

3.1 Laboratory Testing . . . 21

3.1.1 Sample Preparation . . . . 21

3.1.2 Test Plan . . . 26

4. Results and Discussion . . . 33

5. Conclusion . . . 51

xiv

Bibliography . . . 53

Appendix . . . 56

1. Introduction

1.1 Background

Good road infrastructure has long been considered one of the foremost indicators for measuring the development of a nation. It is hence not surprising in witnessing most of the developed part of the world being notable for possessing the best road infrastructures when compared to the developing and underdeveloped world. Pavement engineering has existed as part of an engineering discipline for a long time , however, its advancement in technology have not been boosted by new discoveries in method of construction as well as construction materials in comparison to the other branches of engineering. The roads we see today do not differ much from those in the past 70 to 80 years. However, recently advancement in construction equipment, construction materials, additives, testing equipment and last but not least new design philosophies is contributing towards a new era of advancement in the road construction industry.

Roads consist of the subgrade, base, sub base and wearing courses.

Compaction in these different layers of the road has been one of the most important stages in highway construction works. It is highly essential to compact asphalt mixes in the wearing course to obtain the target density level and air void content in road construction works. Achieving a correctly compacted asphalt layer can reduce to a considerable extent the premature distresses that can occur on the pavement surface. Suitable compaction during the construction of pavement layers can be realized only if in design procedure, all parameters that have effect on the compactibility of the asphalt mixtures are taken into consideration. Hence, understanding the effects of all these parameters on the compactability of the asphalt mixture is the first step in having a successful mix design.

Hot mix asphalt (HMA) is generally composed of aggregates, asphalt binders and air voids. Among these components, aggregates comprise the highest share amounting 90-96 % by weight and provide the skeleton for the mix. In a way, most of the traffic load is carried by this skeleton of aggregate structures. It is also wise to note that asphalt binders amount 4-10 % by weight

2

of the total mix. These components provide the adhesive property in the total mix helping in their action as visco-elastic material. Fillers are fine minerals passing the no. 200 sieve (0.63 mm sieve), in the American standard sieve these materials are those that pass the 75 µm sieve. Fillers are fine material and are usually considered as modifiers and are not considered in the gradation of aggregates (Anggraini et al. 2012). Fillers in HMA have a lot of advantages. In addition to filling the voids, they reduce moisture susceptibility (hydrated lime filler), increase bond between aggregate and asphalt and increase the stiffness by adding of rigid materials in less rigid matrix (Buttlar et al. 1999).

Fillers play an important role in reducing the optimum asphalt content (Brown et al.1989; Kandhal et al. 1998; Tayebali et al. 1998). However, having too much filler in hot mix asphalt can reduce the interaction between aggregates and binder as coating of the aggregates by fillers will increase the amount of binders hence weakening the mix. (Elliot et al.1991; Kandhal et al.

1998). High content of fillers will stiffen the mix to a great extent that the workability is reduced.

There are a number of filler types used in the pavement industry which include fly ash, hydrated lime, rock flour, volcanic ash, silt, Portland cement, mineral sludge and recycled brick powder to name few. These fillers differ in surface texture, shape, surface area, void content, mineral composition and other petrochemical properties (Bahia et al 2011). Hence their influence on the performance of asphalt concrete mixtures also differs accordingly. Asphalt binders together with the fillers constitute “mastic asphalt”. The physiochemical interaction between fillers and binders has an effect on the eventual performance of the HMA.

Performance of HMA is usually indicated by its resistance to failure or distress when subjected to traffic loading or other environmental factors such as moisture susceptibility or the freeze/thaw cycle. The two most prominent forms of distress in asphalt pavements are the fatigue cracking and rutting (permanent deformation). Stiffening the asphalt mix in such a way as to minimize the fatigue and rutting distresses has been the goal of the pavement engineers for a long time. In this regard, the fillers which are considered as modifiers by themselves are studied for their effect on the pavement performance from the compactibility and resistance to rutting point of view.

Hot mix asphalt mixtures need to be compacted adequately to achieve optimum density capable of withstanding early rutting, moisture damage or else fatigue cracking of the pavement structure. The term compactibility refers to the ease with which the different components of mix (aggregates, binder and fillers) are packed together to give the required level of density.

So far, a number of researches have been made on evaluating the effects of different parameters on the compactibility of asphalt mixture; however, due to complexity and the vastness of the parameters and their effects, further work still need to be done by isolating the specific parameters for study. This is done, for example, by keeping constant all other parameters which could have an effect on the compactibility and studying the effect of a single parameter.

In this study, the effect of fillers on the performance of HMA is evaluated by preparing laboratory samples with different filler types and contents. These samples are studied while being compacted by the superpave gyratory compactor (SGC). The SGC simulates the field compaction in such a way that it introduces the shearing action of compaction equipment by introducing a tilting angle of around 1, 2 degrees while the specimen is being subjected to a constant vertical loading. In addition, the same number of samples are also tested for their performance using the simple performance tester (SPT) to find the dynamic modulus and flow number. These parameters give an insight into the resistance of the samples against rutting.

1.2 Objective

The overall objective of this study is to evaluate the compactibility and resistance against rutting of HMA with respect to changes both in filler type and content. This was done by closely observing how the compaction energy index changes with respect to the filler content and type and then finding the dynamic modulus and flow number of the samples from the SPT.

4

The following tasks were carried out in achieving the overall objective of the study;

1. Study the rate of change in the height of the specimen for each filler content with respect to the number of gyrations during compaction.

2. Estimate the compaction energy index for specific filler content and type in the mix samples and evaluate the compactibility.

3. Use the simple performance tester to evaluate the resistance to rutting of the mix by determining the dynamic modulus and flow number.

4. Relate filler content and type with performance indicators such as dynamic modulus and flow number.

1.3 Scope

This study involves a comprehensive literature review on compaction of hot mix asphalt using the gyratory compactor and compaction energy index as well as the use of the simple performance tester to determine the performance parameters such as dynamic modulus and flow number. Further, the project also includes the preparation of twelve samples from three types and four different contents of fillers. The samples were compacted using the IPC superpave gyratory compactor to a predetermined number of gyrations as per the superpave design procedures. The bulk density (Gmb) as well as the maximum density (Gmm) of the samples are determined in the lab .The compaction energy index as recommended by Bahia (Bahia et al. 1998) was calculated and analyzed. Finally the samples are tested using the IPC simple performance tester to determine the dynamic modulus and flow number.

2.0 Literature Review

2.1 Effect of Fillers on HMA

Fines or mineral fillers are those materials which are passing 75 µm (0, 63 mm BS standard) sieves in a mix. The two most important properties of mineral filler are geometry and composition. Filler geometry can be defined by size, shape, angularity and texture. However, it is considerably difficult to measure the later three. Asphalt filler interaction is affected by a number of chemical compounds. The two main properties of these interactions are the reactivity (calcium compound and water solubility) and the harmful fines (active clay content and organic content) (Bahia et al 2011).

Mineral fillers which are used in the pavement industry can be divided into two groups namely, the natural fillers and imported fillers. The natural fillers include andesite, basalt, caliche, dolomite, granite and limestone while the imported fillers include fly ash, slag, hydrated lime (NCHRP 9-45). Fillers cannot be regarded as aggregates, however, their interaction with bitumen results in the formation of mastics.

The effect of fillers on the properties of hot mix asphalt has been studied by a lot of researchers since the beginning of the 19^th century. Richardson pointed out that the larger surface area exhibited by mineral fillers contributes to a larger surface energy associated with it and permits the use of more bitumen which in turn contributes to more cementing power (Richardson 1915). The presence of filler in another case is associated with reduced optimum asphalt content (Brown et al 1989).

The concept of free volume in the characterization of fillers was first developed by P.J Rigden (Rigden 1947). He postulated that the free asphalt which is the asphalt in excess of the fixed asphalt used to fill the void in a dry compacted bed is the main factor which affects the flow properties of the mastic. Rigden voids’ content describes the volume percentage of voids in a dry compacted filler sample. Higher Rigden voids leads to higher stiffening of binder.

Fillers tend to increase the resilient modulus of asphalt mixes (Anderson 1987) .However, most of the strength of HMA is attributed to the surface

6

contact between the aggregate particles. Excessive amount of fillers will result in an increase in the bitumen content and can result in a weak asphalt mix (Kandhal et al 1998).

The most important effect of filler is its stiffening effect. Mineral fillers are also important in changing the viscosity of the binder. They tend to make the binder less viscous at elevated temperatures in such a way they can be regarded as modifiers. Having a high content of filler, however, will result in undesirable stiffness which can affect the workability of the mix. The stiffening effect of fillers tends to decrease at lower temperatures instead the mineral fillers tend to improve the fracture properties of the asphalt binder (Lee et al 1995).

Different agencies have different recommendation of the dust to filler ratio. The SHRP volumetric mix design criteria recommends, for example , a dust to filler ratio between 0,6 to 1,2 percent by volume (SHRP).There is a need for further research to determine what optimum amount of filler content should be used to give a better pavement performance .

2.2 Compaction

2.2.1 Compaction of HMA

Compaction is the process by which the volume of air void in a HMA mixture is reduced through the application of external forces. The most important purpose of compaction is to increase the unit weight of the material under compaction. Compaction of HMA is one of the most important processes in the construction of flexible pavements. HMA should be compacted adequately to increase the particle to particle contact between aggregates and reduce the air void content. However, sufficient air voids should be present in the compacted mixture so that the binder should be able to expand and contract during temperature changes and to allow for further densification during the first few years of service by traffic.

For dense graded asphalt mixtures the air void content should not be allowed to drop below 3-4 percent. If it drops below 3 percent there is a high risk for significant amount of deformation. The pavement life is reduced by

about 10 percent for each percent air void content increase above 7 percent (Linden et al 1989). The right amount of air void content in a mixture for best performance has not yet been scientifically determined, however, it is claimed from experience that 4 percent air void ratio is optimum (Bahia et al 1998).

The densification process of HMA mixture is affected by number factors which include temperature, mix properties such as gradation and viscosity of binder, fine content and method of construction. The effect of temperature on compaction is the direct result of its effect on the viscosity of binder. Increase in temperature results in decreasing the viscosity of the binder in the HMA mix. A binder with higher viscosity is more resistant to deformation. This leads to the addition of more compactive effort to reduce the air voids in the mix.

Gradation affects how the aggregate particles will pack under the application of compactive effort. In addition surface texture, angularity and particle shape have considerable effect on the compaction. Binder grades affect compaction because of their different viscosities. Moreover, binder ageing during production results in resistance to deformation.

2.2.2 Superpave Gyratory Compactor (SGC)

There have been various methods of compaction used in the laboratory to simulate the field compaction of HMA. A study carried out on five types of laboratory compaction methods which include Marshall automatic impact compaction, Marshall manual impact compaction, California kneading compaction , gyratory shear compaction (angle of gyration 1,25°) and gyratory shear compaction (angle of gyration 6°) shows that the gyratory shear compaction (angle of gyration 1,25°) method best represented the field compaction (Khan et al 1998).

The superpave gyratory compactor (IPC Servopac) is a servo-controlled multi axis pneumatic loading system for the laboratory production of asphalt specimens having characteristics that closely resemble that of the field condition. The gyratory angle 2 Ф can be changed for a specific test as needed. The mould is hold at an angle of 2 Ф degrees from the center line as

displacement of each gyratory motion actuator and this signal is used to accurately set and maintain the gyratory angle (1,25°) during compaction.

There is a possibility to use two mould sizes of 100 and 150 mm inside diameter and a mould height of 270 mm. The maximum number of gyration to be applied to the system can be selected manually from the control pendant or from the connected PC. Specimen data measured during testing is displayed by a PC in real time as the test proceeds (Servopac manual).

2.2.3 Compaction Energy Index (CEI)

The first law of thermodynamics which states that the total work done on a system by all external forces must equal to the rate of increase of the total energy of the system is the basis for applying the energy principle to the gyratory compaction.

Dessousky et al. (2002) pointed out that if the deformation during the compaction of a specimen by SGC is assumed to occur under constant temperature, the equation of conservation of energy can be written as:

ߩ כ^ௗ௨

ௗ௧ ൌ ߪ_௜௝ܦ_௝௜ [1]

Where ρ = material density; du/dt = rate of change of the internal energy per unit volume and ߪ_௜௝ܦ_௝௜= represent the mechanical work done by the external forces not converted into kinetic energy. Here the time increment used in the above equation is taken as the time needed to complete one gyration. Assuming a plastic deformation then the change in internal energy for each gyration (du) is equivalent to the dissipated energy due to volumetric and shear strain.

݀ݑ ൌ ݀ݒ ൅ ݀ݏ [2]

Where dv and ds = changes in the internal energy due to volumetric change and shear deformation respectively. The exact compaction energy given in equation 2 is difficult to account for due to the complexity of the applied stresses and the deformation associated with them instead it is possible to account for an index based on the number of gyration and the vertical deformation and the applied vertical stress.

The Ninitial and Ndesign values are the other parameters used to describe the compactibility of HMA. These are the superpave design gyratory compactive efforts for the specific mix. They are dependent on the number of ESALs and the average design high air temperature of the area (Asphalt institute manual).

According to the research done by Birgisson et al. (2004), the mixture rut resistance can be described by the gyratory shear slope calculated between air void content of 4 and 7 percent if the maximum gyratory shear strength is not reached prior to 4 percent air voids and from 7 percent to the point of maximum gyratory shear strength otherwise i.e. if the gyratory shear strength is reached prior to the 4 percent air voids. Accordingly it is shown that the higher the gyratory shear slope, the greater the resistance to deformation.

2.3 Performance Tests

Performance tests are required to check how well HMA performs against rutting. These tests are use both in the mix preparation stage as well as for quality control in field. The national cooperative highway research program (NCHRP) took the initiative to develop advanced pavement technology for the purpose of improving the longevity of pavement structures. One of these technologies is the simple performance tester (SPT) or as is called more recently asphalt mixture performance tester (AMPT).

AMPT is testing equipment which is designed to perform the measurement of engineering properties of HMA. It is the product of the NCHRP Projects 9- 19 and 9-29. As per the recommendation of these projects three candidate tests were selected for a thorough validation process in the fields. These tests are:

x The Dynamic modulus x Flow Number and x Flow Time

2.3.1 Dynamic Modulus Test

In the SPT dynamic modulus testing, the specimen is subjected to a controlled sinusoidal (haversine) compressive stress loading at various test frequencies between 0.01 and 25 Hz. The specimen is kept at specific test

14

temperature and confining pressure. The three on specimen displacement transducers are used to measure the applied stress, confining pressure, temperature and the resulting axial strain as a function of time. From these measurements the dynamic modulus, phase angle, average temperature and average confining pressure are calculated.

The Control and Data Acquistion System (CDAS) provides loading control and transducer data acquision and timing functionality. Control and measurements are done from the load cell (axis 1), pressure transducer (axis 2) and three Linear Variable Displacement Transducers (LVDT) which are mounted on the test specimen.

As the test progresses the data from the CDAS is transferred to the PC this continues to be updates for each cycle of loading. Test results are archived and can be transferred to spread sheets for further analysis.

The dynamic modulus is defined by the equation:

o

E o

H

* V

[4]

( 360 )

p i

T

I

T [5]

Where:

_E*_ = dynamic modulus I = phase angle, degree V^o = stress amplitude H^o = strain amplitude

Ti = time lag between stress and strain Tp = period of applied stress

For a pure elastic material I = 0 and for a pure viscous material I = 90°.The dynamic modulus test can be performed both in confined mode (with the application of confining pressure) and unconfined mode. In addition to its use in the construction of the master curves which are used as input values for the mechanistic-empirical design guide (MEPDG) in the design of flexible

0.00 0.05 0.10 0.15

TIME, SEC

LOAD AXIAL STRAINTIME LAG, T_I

V

O

H

O PERIOD, T_P

2 V

_R

H

_R

Figure 5. Schematics of Dynamic Modulus Test Data (NCHRP Project 9-29)

16

pavements, the dynamic modulus is a useful tool to compare stiffnesses of different asphalt mixtures.

2.3.1 Flow number Test

One other way to determine the permanent deformation properties of an HMA is to apply a repeated dynamic load and observe the relative deformation characteristics of the material as a function of the number of cycles (repetitions).This test is very similar to the repeated load creep test.

In this test, the specimen, at a specific test temperature, is subjected to a repeated haversine axial compressive load pulse of 0.1 sec every 1.0 sec in other words there is a 0,9 sec rest period. This is a simulation of the traffic load that the pavement is subjected under its operation.

As in the repeated load creep test, there are three zones to describe the permanent strain of the specimen under a test as shown in the figure below. In the initial stage (Primary zone) permanent strain accumulates rapidly, the incremental permanent deformation decreases slowly reaching a constant value in the secondary zone. In the final section (Tertiary zone) the permanent strain starts to increase again rapidly. This point or cycle where the tertiary zone begins is termed as the flow number. In other words this is the point where minimum rate of strain is recorded (Witczak et al. 2002).

Figure 6. Cumulative strain vs. No. of load cycles

The test may be conducted with or without confining pressure. The resulting permanent axial strains are measured as a function of time and numerically differentiated to calculate the flow number. The flow number is defined as the number of load cycles corresponding to the minimum rate of change of permanent axial strain. (NCHRP Project 9-29)

The procedure for calculating the flow number involves three steps. First, the rate of change of permanent axial strain is H^p, with respect to the number of load cycles is estimated by the use of finite difference formula (NCHRP Project 9-29).

     

N dN

d p i p i N p i N

'

# ^'  ^'

2

H H

H

[6]

Where:

d(H^p)i/dN = rate of change of permanent axial strain with respect to cycles or creep rate at cycle i, 1/cycle

(H^p)i-'N = permanent strain at i-'N cycles (H^p)i+'N = permanent strain at i+'N cycles 'N = sampling interval

The next step is smoothing of the creep rate data obtained in equation 6 by calculating the running average at each point, by adding to the derivative at that point the two values before and two values after that point, and dividing the sum by five:

¸¸¹·

¨¨©§







 ^{' '  '}

'



dN d dN d dN d dN d dN d dN

d _{p i} ( _p)_{i 2N} ( _p)_{i N} ( _p)_i ( _p)_{i N} ( _p)_{i 2 N} 5

' 1 )

(

H H H H H H

[7]

18

Where:

d(H^p)’i/dN = smoothed creep rate at i sec, 1/cycle d(H^p)i-2'N/dN = creep rate at i-2'N cycles, 1/cycle d(H^p)i-'N/dN = creep rate at i-'N cycles, 1/cycle

d(H^p)i/dN = creep rate at i cycles, 1/cycle d(H^p)i+'N/dN = creep rate at i+'N cycles, 1/cycle d(H^p)i+2'N/dN = creep rate at i+2'N cycles, 1/cycle

The Flow Number is reported as the cycle at which the minimum value of the smoothed creep rate occurs.

Figure 7. Flow Number Schematic (NCHRP Project 9-29)

2.3.2 Energy Dissipation

The area under the stress strain curve is used to describe the mechanical energy dissipated during the dynamic loading of the specimen under the specified frequency. This is easily shown by the following equation.

ܷ^כൌ ^ଵ

௏׬ ܲ݀ܮ ൌ ׬ _஺^௉

బ^ௗ௅

௅బ

௅

଴  ൌ  ׬ ߪ݀߳_଴^ఢ [8]

Where: U = Total mechanical energy V = Volume of the specimen P = Applied load

A0 = Original Area of the specimen L0 = Original Length of the specimen

During loading the area under the stress-strain curve is the strain energy per unit volume absorbed by the material while the area under the unloading curve is the energy released by the material. For elastic materials the areas are equal and no net energy is absorbed. For visco- elastic materials like HMA the difference between the energy absorbed and the energy released is the dissipated energy. The binder in HMA is the leading media of energy dissipation.

The dissipated energy approach is mostly used for fatigue analysis in pavement engineering in recent years. Studies by SHRP showed that the relationship between the cumulative dissipated energy and the number of loading cycles to failure are temperature and mixture dependent. The relationship used for predicting the fatigue behavior using the dissipated energy assumes that all dissipated energy represents damage done to the material.

20

Figure 8. Energy Dissipation loop

Materials with low level of loading amplitude or having high fatigue resistance will produce lower amount of relative energy dissipation (Shihui et al 2006). The change in the dissipation energy from one cycle to the next cycle is in fact the one that causes cracks in the sample.

At low strain or damage levels the fatigue behavior cannot clearly be represented and samples should be subjected to higher levels of strain to inflict damage or failure. This is one of the main disadvantages of the energy dissipation model for fatigue analysis (Shihui et al 2006).

3.0 Methodology

To perform this study, 15 laboratory samples were prepared as per the superpave mix design procedures using the superpave gyratory compactor.

The compaction data collected from the gyratory compactor is then analyzed in detail. Possible concluding remarks are then presented based on the results obtained. Finally the samples are tested using the simple performance tester to determine their dynamic modulus and flow number.

3.1 Laboratory Testing 3.1.1 Sample preparation

x Gradation

The gradation for the samples was chosen as per the Swedish specification given in TRVKB 10 (Trafikverket- Swedish traffic administration). According to this specification, the gradation requirement for ABT 16, dense asphalt concrete, with nominal size of aggregate 16 mm is selected. This type of gradation is commonly used for the interstate highway section of the roads in Sweden. The table below shows the particle size distribution table as well as the 0,45 power gradation curve.

Table 1. Particle size Distribution

Sieve Size (mm)

Requirement JMF

(percent passing)

Max Min

22,4 100 100 100

16 100 90 93

11,2 88 71 83

8 73 57 66

2 47 26 38

0,5 30 13 25

0,06 9 6 6

22

A 0,45 power chart is used to evaluate the gradation of the aggregate blend. As can be seen from the particle size distribution curve, the gradation satisfies the requirement stipulated in the specification. Moreover, the curve is close to the maximum density line (MDL), which clearly shows it is a dense or well graded mix. In addition, since the gradation curve falls above the maximum density line (MDL), it can be said that fines are more prominent in the mix (Fine gradation).

Figure 9. Gradation Chart (ABT 16) x Binder Content and Type

The binder type used for producing the samples is 70/100 penetration grade bitumen manufactured by Nynas AB.TRVKB 10 recommends a minimum binder content by weight of 6%.However, due to the variability of the aggregate property the 6% binder content isn’t applicable for this project.

In the planning of the test, it was required to study the effect of the filler on the asphalt mix by keeping all parameters constant except the filler content and type. A binder content of 4 percent was selected so as to get the higher effect of the filler’s shape and type under the performance tests after the preparation of several trial mixes and inspecting the sample properties.

0 20 40 60 80 100

0 1 2 3 4 5

% Passing

Sieve Size^0,45

0.45 Power chart Dense Asphalt (ABT 16)

ABT 16 MDL Upper Limit Lower Limit

x Mixing and Compaction Temperature

From Temperature-Viscosity charts from the asphalt institute mixing and compaction guidelines, the compaction range is found to be between 137,5°C to 141,5 °C . For the project a compaction temperature of 140°C is selected.

The mixing range is between 148,5°C to 153,5 °C. The mixing temperature for the production of the samples is selected to be at 150°C.

x Types of Fillers and Content

Three types of fillers are used for preparation of the test samples. The filler types are M600, M10 which are silica based fillers and fly ash.M600 and M10 have almost the same type of chemical composition but their physical properties such as particle size, size distribution, specific surface area and density. As characterized by the scanning electronic microscope (SEM), both M600 and M10 have angular shapes, however, because of their different particle sizes, their interaction is different. Looking at the filler size distribution, the M600 particles are finer than M10 particles and fly ash particles are slightly coarser than M10 (Hesami et al. 2012).

Fly ash particles have a round shape which makes them different from the other two. The specific surface area of M10, M600 and fly ash as measured by Micrometrics Gemini 2360 Surface Area Analyzer gave 0.93, 4.0 and 1.5 m²/g respectively. The densities as determined from the helium pycnometer measurement are 2.79, 2.75 and 2.41 kg/cm³ for M10, M600 and fly ash respectively. Filler concentrations of 10, 20, 30 and 40 percent by volume are used. Figures 10-12 show photographic images from scanning electronic microscope (SEM) for M10, M600 and fly ash.

24

Figure 10. Photograph of M10 filler under SEM

Figure 11. Photograph of M600 under SEM

Figure 12. Photograph of Fly ash under SEM x Compactive Effort

In selecting the amount of compactive effort to be used for the test, an average design high air temperature of less than 39°C is assumed.

This is done taking into consideration of the average high air temperature in the region (Sweden).

The three critical gyration numbers established by SGC are the Ninital , Ndesign and Nmax. The Ninitial is used to indicate the compactability of mixtures during the construction process. A sample which compacts too quickly has a low air void content at Ninitial . This is an indication of tender mixes which are unstable when subjected to traffic.The Ndesign is used to indicate the design number of gyrations which are required to produce a sample having the same density requirements as that of the field after the design amount of traffic has passed.

26

The Nmax is the number of gyrations which indicate the plastic threshold of the mix. This threshold should normally be not exceeded otherwise excessive potential rutting shall occur. The air void content at the Nmax should not be allowed to fall below 2 percent. Table 2 which is adapted from the asphalt institute Manual SP-2 is used to select the superpave design gyratory compactive effort.

Table 2. Superpave Design Gryatory Compactive Effort (Asphalt Institute Manual SP-2)

Design ESALs(millions)

Average Design High Air Temperature

< 39°C 39-40°C 41-42°C

Nini Ndes Nmax Nini Ndes Nmax Nini Ndes Nmax

<0,3 7 68 104 7 74 114 7 78 121

0,3-1 7 76 117 7 83 129 7 88 138

1-3 7 86 134 8 95 150 8 100 158

3-10 8 96 152 8 106 169 8 113 181

10-30 8 109 174 9 121 195 9 128 208

30-100 9 126 204 9 139 228 9 146 240

>100 9 143 233 10 158 262 10 165 275

3.1.2 Test Plan

A total of 15 test samples were evaluated to see how the level of compaction energy index varies with the increase in the content of the fillers.

The first four samples are prepared by using the M600 silica based filler, the next four samples with M10 silica based filler and the last four with fly ash filler. The filler content was varied by increments of 10 percent from 10 to 40 percent by volume. In addition, three extra samples of 40 percent filler content by volume of M600, M10 and fly ash samples were tested to check if the results obtained from the previous samples can be repeated.

Controlled parameters

Before testing the specimen, it is worth noting that the following variables are controlled in the production and testing of the specimen.

x Design ESALs, which describes the level of traffic is assumed to be 10-30 millions. This is used to determine the amount of the compactive effort (Nini, Ndesign and Nmax).

x The binder type is the same for all samples (70/100 from Nynas AB).

This binder has viscosity of 330 centistokes at 135°C.

x The aggregate source is the same for all the samples. The aggregates are washed and oven dried before mixing is done. The aggregate gradation is also the same for all the samples (ABT 16).

x The binder content for all the samples is fixed at 4 percent by weight.

x The mixing and compacting temperatures are also controlled at 140°C and 150°C respectively.

Superpave Gyratory Compactor (SGC)

The superpave gyratory compactor (SGC) provides number of gyration versus height of specimen data for a test sample when the maximum number gyrations, diameter of the specimen, number of gyrations per minute is selected by the personnel in charge of the testing operation. It is possible to select either the maximum number of gyrations or the required height of the sample as the termination criteria when running the SGC. For the project case, the maximum number of gyration which is 174 is selected as the termination criteria. This means once the number of gyration reaches 174 the machine terminates compaction. The mould size for the test was a 100 mm diameter by 150 mm height. Typical SGC equipment is shown in figure 13.

The vertical stress applied to the specimen is kept at 600 kPa while the gyratory angle is fixed at 1,25°. There is a constant measurement of the height of the specimen after the completion of each gyration. The recorded data is transferred to a PC connected to the SGC.

28

Figure 13. Servopac Superpave Gyratory Compactor (SGC).

Simple Performance Tester (SPT)

The IPC SPT is a computer controlled hydraulic loading machine designed to perform simple performance tests such as the dynamic modulus and flow number as recommended by the NCHRP Project 9-19 .These tests are conducted to evaluate the resistance of superpave designed HMA mixes to permanent deformation and fatigue cracking.

The IPC SPT machine has two heavy duty circular cross heads and 3 vertical columns which are large enough to facilitate triaxial testing of 100

mm diameter by 150 mm tall samples. A confining cell surrounding the frame acts as a pressure cell and temperature controlling environment.

Loading force during testing is applied through the shaft of an actuator attached to the base. The shaft has a displacement transducer which is mechanically attached to the actuator. A strain gauge force transducer attached to the loading shaft measures the force applied to the specimen. The specimen is mounted between loading platens that allow axial compressive loading. The cell surrounding the frame allows confinement pressure to be applied simultaneously during testing. The environmental chamber is capable of applying a confining pressure of 350 kPa.

The HPS (hydraulic power supply) is the energy source for the servo actuator (loading) using high pressure oil. At the high pressure setting a 160 bar pressure is used.

A PC software machine control panel (virtual pendant) provides a convenient and easy way to operate the machine. Typical IPC simple performance tester (SPT) equipment is shown in figure 14 below. A steel cylindrical device called proving ring is used to verify the test equipment.

30

Figure 14. IPC Simple Performance Tester Dynamic Modulus

In this test the specimen is tested at 15°C without any confining pressure.

A controlled haversine compressive stress loading at various test frequencies namely 0,1 ; 0,2 ; 1 ; 5 and 10 Hz is applied . As per the recommendation of the NCHRP project 9-29 report the average dynamic strain is kept between 75 to 125 micro strain. An estimated initial modulus of 2000 MPa is interred in to the system based on the tuning default values. Then the Contact stresses to be applied are calculated automatically by the software.

ߪ_் ൌ ߪ_௅൅^{ሺఙೠାఙಽሻ}

ଶ [9]

Where:

σL = Lower strain limit ( default 75 micro –strain) σu = Upper strain limit ( default 125 micro –strain) σT = Target micro strain

The Dynamic load to be applied (kN):

ܲ_{஽௬௡௔௠௜௖} ൌ ܯ݋݀ כ^ሺఙ೅ሻ

ଵாଽ כ ܣ [10]

Where:

Mod= Requested modulus in MPa.

A= Specimen Cross sectional area in m²

The contact load to be applied (kN):

ܲ_{஼௢௡௧௔௖௧} ൌ ܲ_{஽௬௡௔௠௜௖}כ ͷΨ [11]

And the maximum load to be applied (kN) is given as:

ܲ_ெ௔௫ ൌ ܲ_{஼௢௡௧௔௖௧}൅ ܲ_{஽௬௡௔௠௜௖} [12]

Three Linear Variable Displacement Transducers (LVDT) mounted on the specimen measure the resulting axial strain as a function of time during dynamic loading of the specimen. The LVDTs are placed both vertically and diametrically on opposite sides of the specimen by using 6 brass studs which are glued on the specimen face by a gauge point fixing jig.

32

Verification of the dynamic performance was done by using the proving ring and running dynamic tests in the tuning mode. The measurement values for dynamic modulus obtained after the test are then compared with the test certificate and are found to be within the allowed +/- 2% range of the dynamic modulus in the test certificate. Figure 15 shows a typical specimen under testing.

Figure 15. Sample under Dynamic Modulus Test

Flow Number

In conducting the SPT flow number test, the specimens are conditioned at 40^°C for four hours and a minimum of 20 minutes in the environment chamber until temperature equilibrium is achieved. As per the recommendations of the NCHRP 9-30 a repeated deviator stress of 600 kPa and a contact stress of 30 kPa are used.

The specimen is placed in the loading platens and the required input parameters are inputted. The required FN (Flow number) is automatically calculated by the AMPT software when the minimum rate of micro strain is reached. The termination criteria for the test can be selected by the operator either at the maximum number of cycles or the maximum accumulated micro strain.

Figure 16. Test Sample after Flow Number Test

34 4.0 Results and Discussion

Table 3 shows the computed compaction energy index results for the 12 samples. Rice test was used to determine the theoretical maximum density of the samples. The percent Gmm and the percent air void content at Ninitial, Ndesign, Nmax are calculated and included with data table (Table 4).

Compaction Energy Index

Table 3 below summarizes the trend of the compaction energy with respect to the filler content.

Table 3 Compaction Energy Trend with filler content

Sample No. % Filler Content

CEI

CEI

1 10 5,8

2 20 10,4

3 30 20,3

4 40 51,9 53,36

5 10 6,7

6 20 10,2

7 30 18,7

8 40 47,7 47,22

9 10 6,6

10 20 10,1

11 30 17,6

12 40 44,9 45,02

Since the gradation, aggregate type and source are the same for all the samples, the strength of the aggregate skeleton will be the same for all. This implies that the energy dissipated by friction will be almost the same for all the samples. The remaining energy will be effectively used in compacting the

mixture.Hence the graph plotted with the number of cycles versus the reduction in volume represents the amount of energy dissipated in achieving the change in density.

Figure 17 Compaction Energy vs. filler content 5

10 15 20 25 30 35 40 45 50 55

0 10 20 30 40 50

CEI in % Gmm*N

Percent Filler Content

M 600 M 10 Fly ash

36

Figure 18 CEI versus Filler Content (Extra samples)

Discussion of Results

x As can be seen in Fig.17, the compaction energy required to densify the mix increases steadily with increase in filler content. This is for the filler content between 10- 30 percent. This is attributed to the fact that as the filler content increases the viscosity of the mastic in the HMA mix also increases which makes it more resistant to shearing while rolling is done.

x There is a sharp increase in compaction energy beyond the 30 percent filler content for all the three type of fillers. One possible explanation for this type of behavior can be that the physical particle to particle contact among the fillers is so pronounced that a stiff media is created in the mastics. This media establishes a higher resistant to compaction increasing the compaction index.

x Higher level of CEI is observed in the M600 silica based filler than the other two types of fillers (M10 and fly ash).M600 and M10 fillers as discussed earlier have angular shapes. The particles have the

5 10 15 20 25 30 35 40 45 50 55

0 10 20 30 40 50

CEI in % Gmm*N

Percent Filler Content

M 600 M 10 Fly ash

capability to form a higher shear resistance than fly ash which has round shapes. Moreover, previous study (Hesami et al. 2012) has showed that M600 mastic have a higher viscosity as compared to the other two types of fillers namely M10 and fly ash. Higher viscosity will result in higher resistance to compaction.

x M600 fillers have larger specific surface area than M10 and fly ash and hence can have higher affinity to interact both among each other and the binder.

x Fly ash as described in the filler type and content section has a round particle shape and this can make it easier to facilitate the shearing action during rolling and hence is easier to compact than the other two types of fillers with lower compaction energy index.

x Three extra samples prepared at 40 percent filler content of M600, M10 and fly ash to check if the results can be repeated, show a close to the original values for each type of filler.

38

Table 4 Sample characterstics Sample Type

GmbGmmHeight (mm) %Gmm% Air Void Content NiniNdesNmaxNiniNdesNmaxNiniNdesNmax M 600 Filler 10%1 2,36524,113172,9160,9159,3790,4097,1498,089,602,861,92 20%2 2,39124,311165,04151,97150,4789,6697,3798,3410,342,631,66 30%3 2,39324,458161,32147,17146,288,6797,2097,8411,332,802,16 40%4 2,39424,525172,68156,09154,2587,2196,4897,6312,793,522,37 M 10 Filler 10%5 2,3512,409169,180157,7156,2690,1696,7297,619,843,282,39 20%6 2,3752,446168,400157,12155,7389,7696,2197,0610,243,792,94 30%7 2,3782,461178,910166,78165,2689,2695,7596,6310,744,253,37 40%8 2,3832,488172,910161,17159,788,4594,9095,7711,555,104,23 Fly ash Filler 10%9 2,3332,401173,280161,73160,3689,9396,3597,1810,073,652,82 20%10 2,3672,435175,350163,77162,3489,9896,3497,1910,023,662,81 30%11 2,3722,446168,620156,52155,5289,4396,3496,9610,573,663,04 40%12 2,3822,490176,220164,98163,6388,8594,9095,6811,155,104,32

Compaction energy index is computed by calculating the area under the compaction curve between the 8^th gyration and the 92 percent Gmm. A typical curve for CEI is as shown in the figure below. The densification curves for the 12 samples are shown in the appendix.

Figure 19 Densification Curve for M600 at 10 % filler content

Dynamic Modulus Results

The dynamic modulus at 15° C measured using the SPT for the 12 specimens are summarized in table 5-7. The samples are identified as D1 – D12. For M600 D1 is at 10%, D2 at 20%, D3 at 30% and D4 at 40 % filler content respectively. The same trend is applied in naming for M10 (D5 – D8) and fly ash (D9 – D12).

y = 85.662x^0.0271 R² = 0.986

80 82 84 86 88 90 92 94 96 98 100

0 20 40 60 80 100 120 140 160 180

% Gmm

No. of Gyrations

40

Table 5 Dynamic Modulus vs. Frequency (M600) at 15°C Dynamic Modulus (MPa)

M600

Frequency (Hz)

D1 (10%)

D2 (20%)

D3

(30%) D4 (40%)

10 7109 8481 9527 10167

5 5756 7065 8020 8621

1 2815 3732 4431 4890

0,2 1248 1685 2048 2284

0,1 923 1217 1454 1608

Table 6 Dynamic Modulus vs. Frequency (M10) at 15°C

Dynamic Modulus (MPa)

M10

Frequency (Hz)

D5 (10%)

D6 (20%)

D7

(30%) D8 (40%)

10 7600 7288 7964 9741

5 6265 5955 6591 8211

1 3355 3199 3734 4809

0,2 1646 1606 1897 2443

0,1 1254 1252 1432 1797

Table 7 Dynamic Modulus vs. Frequency (Fly ash) at 15°C Dynamic Modulus (MPa)

Fly ash

Frequency (Hz)

D9 (10%)

D10 (20%)

D11 (30%)

D12 (40%)

10 8027 9204 8964 9575

5 6582 7524 7573 7942

1 3615 4030 4399 4382

0,2 1819 1885 2232 2089

0,1 1384 1344 1658 1509

Figures 20-25 show the filler content and filler type comparisons for the different samples under investigation. As can be seen from the graphs, there is a general trend of increasing modulus with increasing frequency for all type of fillers. For the three types of fillers, as the filler content increases, the dynamic modulus also increases. This can be explained from the view that fillers generally stiffen the mastics in asphalt mixtures.

Figure 19 shows this general trend of increasing dynamic modulus with increasing filler content for the three types of fillers and the different filler concentrations.

Figure 20 Dynamic Modulus vs. Frequency for all filler type and content 0

2000 4000 6000 8000 10000 12000

0 2 4 6 8 10

Dynamic Modulus,MPa

Frequency, Hz

10 % M600 20% M600 30% M600 40% M600 10% M10 20% M10 30% M10 40% M10 10% Fly ash 20% Fly ash 30% Fly ash 40% Fly ash

42

Figure 21 Dynamic Modulus vs. Frequency for M600 Filler at 15°C

A comparison was also made among the three filler types for different filler content as shown in Figure 16. At 10 percent filler content, fly ash shows higher dynamic modulus values in comparison to both M10 and M600. At a lower filler concentration the physical filler binder interaction is very minimal and interaction between the angular particles is difficult to achieve as the particles are spaced very far apart from each other. The pysio-chemical interaction property that caused the fly ash to have higher modulus at lower filler content of 10 and 20 percent is difficult to explain due to the complex nature of filler binder interactive chemistry.

0 2000 4000 6000 8000 10000 12000

0 2 4 6 8 10

Dynamic Modulus ,MPa

Frequency ,Hz

10%

20%

30%

40%

Figure 22 Dynamic Modulus vs. Frequency (10 percent filler) at 15°C

At 20 percent filler content, fly ash still constitutes higher dynamic modulus value. At this percentage of filler content, M600 filler shows higher value of dynamic modulus than M10.This is probably because of the higher surface area available for interaction in M600 than M10. At 30 percent filler content, M600 shows the higher value of dynamic modulus followed by fly ash and M10 and finally at 40 percent, the three types of fillers in increasing dynamic modulus values are ranked as M600, M10 and fly ash as shown in Figures 18 and 19.Here we can observe the effect of filler concentration starts

0 2000 4000 6000 8000 10000 12000

0 2 4 6 8 10

Dynamic Modulus ,MPa

Frequency , Hz

M600 M10 Fly ash

44

to appear. The filler particles are close with each other and try to form a stiff interlocked media within the mastic.

It is understood from the results of the tests that at higher filler concentration the stiffening effect of the fillers is clearly visible from the increase in dynamic modulus. It is also observed that M600 filler at higher filler concentration (30 and 40 percent) exhibits higher stiffening characteristics. At higher filler concentration the fillers particles are placed close to each other.M600 and M10 filler particles because of their angular shape show higher interlocking and hence stiffening characteristics than the round shaped fly ash particles at these filler concentrations.

Figure 23 Dynamic Modulus vs. Frequency (20 percent filler) at 15°C 0

2000 4000 6000 8000 10000 12000

0 2 4 6 8 10

Dynamic Modulus , MPa

Frequency , Hz

M600 M10 Fly ash