Rep17 04 PDF

NCAT
Report 17-04

HIGH-MODULUS ASPHALT
CONCRETE (HMAC)
MIXTURES FOR USE AS
BASE COURSE

By
Fabricio Leiva-Villacorta
Adam Taylor
Richard Willis

June 2017

HIGH-MODULUS ASPHALT CONCRETE (HMAC) MIXTURES FOR USE AS BASE COURSE

NCAT Report 17-04

By:

Fabricio Leiva-Villacorta, PhD
Assistant Research Professor
National Center for Asphalt Technology at Auburn University

Adam Taylor, P.E.
Assistant Research Engineer
National Center for Asphalt Technology at Auburn University

Richard Willis, PhD
Director of Pavement Engineering & Innovation
National Asphalt Pavement Association

Sponsored by
Federal Highway Administration

June 2017
ii
ACKNOWLEDGEMENTS
This project was funded by the Federal Highway Administration (FHWA). The authors would like to
thank the many personnel who contributed to the coordination and accomplishment of the work
presented herein.
DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible for the facts and
accuracy of the data presented herein. The contents do not necessarily reflect the official views or
policies of the sponsor(s), the National Center for Asphalt Technology, or Auburn University. This
report does not constitute a standard, specification, or regulation. Comments contained in this
paper related to specific testing equipment and materials should not be considered an
endorsement of any commercial product or service; no such endorsement is intended or implied.
iii
TABLE OF CONTENTS

Abstract ............................................................................................................................................ 5
1. Introduction .............................................................................................................................. 6
2. Objective ................................................................................................................................... 7
3. State-of-the-practice ................................................................................................................. 7
3.1 Mixture Design ................................................................................................................. 8
3.1.1 Aggregate Selection ...................................................................................................... 9
3.1.2 Designing a Gradation ................................................................................................ 11
3.1.3 Binder Selection and Richness Factor ......................................................................... 12
3.1.4 Performance Tests ...................................................................................................... 16
3.2 Pavement Design ............................................................................................................ 21
3.3 Performance ................................................................................................................... 23
3.4 Construction ................................................................................................................... 29
3.5 Summary of Current Practice ......................................................................................... 29
4. Experimental Design and Analysis .......................................................................................... 30
4.1 Laboratory Testing .......................................................................................................... 30
4.2 Mixture Design ............................................................................................................... 31
4.3 Dynamic Modulus ........................................................................................................... 33
4.4 Flow Number .................................................................................................................. 38
4.5 AMPT Cyclic Fatigue ....................................................................................................... 41
5. AASHTOWare Pavement ME Design Analysis ......................................................................... 47
5.1 Traffic ............................................................................................................................. 48
5.2 Climate ........................................................................................................................... 49
5.3 Estimated Performance .................................................................................................. 50
6. Conclusions and Recommendations ....................................................................................... 56
7. Recommended Mixture Design Procedure ............................................................................. 57
References ...................................................................................................................................... 58

iv
Leiva-Villacorta, Taylor, and Willis
ABSTRACT
Recent studies on long-life flexible pavements indicate that it may be advantageous to design
and construct asphalt mixtures comprising the underlying layers in such a manner that very high
modulus mixtures are produced. The French have been experimenting with and designing
pavements with high-modulus bases since the 1980s. This study considered the engineering
properties of asphalt mixtures produced using a European specification for high-modulus asphalt
concrete (HMAC) mixtures and used as base course. This specification includes volumetric
requirements such as asphalt content and air voids, but there are also requirements for
engineering parameters that address performance requirements such as rutting and fatigue
cracking. The purpose of this study was to investigate the design of asphalt mixtures having
higher modulus. The study was limited to a laboratory performance evaluation and a theoretical
modeling component where the results were used to indicate potential field performance.
A comprehensive literature study was performed to assess the current state-of-the-practice on

HMAC mixture design, pavement design, laboratory performance tests, and full scale pavement
performance. The experimental plan included a variety of mixtures with different material and
binders such that higher moduli were obtained compare to conventional mixtures. The plan
included a French mixture with a stiff binder (PG 88-16), two mixtures containing 35% RAP both
with polymer-modified binders but one high polymer content (HiMA), another mixture
containing 25% RAP and 5% RAS with a polymer-modified binder, and finally, a 50% RAP mixture
with a polymer-modified binder. The laboratory testing program evaluated binder performance
grade, mixture stiffness over a wide temperature range, fatigue cracking, and permanent
deformation. In addition, AASHTOWare Pavement ME Design software was used to determine
how a high-modulus base would affect predicted performance of asphalt pavements.
The results of this study indicated that European mix design standard methods and specifications
were successfully implemented on local (U.S.) virgin and recycled materials. In addition,
increased stiffness of high modulus mixtures improves mechanistic-empirical predicted
performance of pavement in rutting, fatigue cracking, and ride quality. However, it was
determined that performance of new materials cannot be reliably modelled with the current
transfer functions and further field validation is required.
5
1. INTRODUCTION
Currently in the United States, asphalt paving mixtures are primarily designed using the
Superpave system where the proportioning of components relies mainly on volumetric
properties. Early Superpave implementation focused primarily on rutting resistance. Mixture
designs for moderate and high traffic pavements were designed for improved rutting resistance
by specifying a higher grade of asphalt binder and higher quality aggregates. Most highway
agencies now report that rutting problems have been virtually eliminated. However, there have
been growing concerns that the primary mode of distress for asphalt pavements is cracking of
some form or another. There are several possible contributing factors to increased cracking,
including issues with mixture designs, increased use of recycled materials, problems with the
quality of construction, and failure to adequately address underlying pavement distresses during
pavement rehabilitation. It is now well recognized that current mixture design practices have
some shortcomings.
Most state departments of transportation (DOTs) currently utilize volumetric criteria for asphalt
mixture designs that follow the Superpave mixture design methods of AASHTO M 323 and R 35
with some modifications. In response to pavement durability issues, many DOTs have modified
their design and acceptance requirements to obtain more durable and high crack-resistant
mixtures by increasing the asphalt content of the lower layer of hot-mix asphalt, commonly
referred to as rich-bottom mixtures. Rich-bottom mixtures are made with the same grade of
asphalt binder but are designed at a lower air void content so as to increase the design asphalt
content by 0.6% to 0.8%.
This study considered the engineering properties of asphalt mixtures produced using a European
specification for high-modulus asphalt mixtures and used as base course. This specification
includes volumetric requirements such as asphalt content and air voids, but there are also
requirements for engineering parameters that address performance requirements such as
rutting and fatigue cracking. High-modulus asphalt is routinely produced with hard asphalt
binders, PG 88 or higher, for critical high temperature properties. In this study, an effective hard
asphalt binder was obtained by combining polymer-modified asphalt with several contents of
RAP (between 25% and 50%) and utilized in a high-modulus mixture produced following French
standard procedures.
In the 1980s, the French Public Works Research Institute or Laboratoire Central des Ponts et
Chaussées (LCPC) developed high-modulus mixtures, referred to as Enrobé à Module Élevé (EME).
The objective for this type of new mixture was improved mechanical properties to include high-
modulus, good fatigue behavior, and excellent resistance to rutting. High stiffness and improved
fatigue resistance allow a decrease in pavement thickness for both new construction and for
rehabilitation. One goal of the EME developers was to reduce geometric constraints (overhead
clearance constraints) during rehabilitation. Early trials of EME occurred in the mid-1970s and by
the early 1980s had developed into a new mixture type. A specification was set by the early 1990s
and in the late 1990s the mixture had become part of the standard catalog of mixtures used in
6
pavement structural design for high traffic pavements, 20 million equivalent single axle loads
(ESALs) or greater.
In the United States, agencies are beginning to adopt the Mechanistic Empirical Pavement Design
Guide (MEPDG) for determining pavement structural thickness. Unlike the previous structural
design method, typically AASHTO 1993, which only considered mixture properties indirectly, the
MEPDG makes direct use of mixture properties as part of the structural pavement design. This
study compares an MEPDG pavement structural design using only Superpave mixtures and
several high-modulus mixtures for the base course.
2. OBJECTIVE
The objective of this project was to evaluate the current mixture design methodology of high-
modulus base layers and evaluate potential effects on performance. In order to successfully
analyze the effect of high-modulus mixtures as base courses, the following tasks were completed
on this project:
1. Literature review: A literature study was performed to assess the current state-of-the-
practice. This included information from published journals, technical reports, articles,
presentations, as well as personal communications and interviews with contractors and
agencies that have successfully implemented high-modulus mixtures in their pavement
structures.
2. Provide recommendations of mixture design and structural design procedures: Data and
resources gathered during the literature review were used to develop material selection
and mixture design procedures for a high-modulus mixture that would be resistant to the
tensile strains at the bottom of the asphalt pavement structure.
3. Laboratory study: Engineering properties of high-modulus mixtures were determined and
design procedures were assessed using laboratory performance tests.
4. Pavement design and analysis: AASHTOWare Pavement ME Design software was used to
determine how a high-modulus base would affect predicted performance of asphalt
pavements.
3. STATE-OF-THE-PRACTICE
High-modulus asphalt concrete (HMAC) was originally developed in the 1980s at a time when
France was looking to design high performance asphalt mixtures to increase the life span of
conventional asphalt pavements or reduce the necessary thickness required to carry the
increasing loads seen on European highways. Although these mixtures were designed to serve as
either asphalt base or binder courses, they were eventually also used in wearing courses in the
mid-1980s, but these are outside of the scope of this report (EAPA, 2005; Nkgapele et al., 2012;
Corte, 2001).
In the 1990s, the French developed a standard for EME mixtures (Denneman, 2011; Petho and
Denneman, 2013). This standard had two classes of EME mixtures. Class 1 was a low fatigue
resistance mixture while Class 2 was a higher fatigue resistant mixture. The main difference
7
between these two classes was the binder content of the mixtures. In 2007, a European Standard
(EN 13108-20) was developed (Brosseaud, 2012; Guyot 2013).
To date, numerous European countries such as France, Austria, the Czech Republic, Denmark,
Ireland, Italy, Netherlands, Portugal, and the United Kingdom have all had positive experiences
using this material; however, each country has a slightly different approach to mixture design
and performance criteria, as expected. Generally, these mixtures have been successfully
incorporated at times when low quality aggregates are available to reinforce the mixtures,
industrial areas are subjected to heavy loads, and when existing pavements need to be reinforced
during rehabilitation or reconstruction (Brosseaud, 2012; EAPA, 2005; Petho and Denneman,
2013). Case studies at airports have also been conducted to improve runway and taxiway
durability (EAPA, 2003; Guyot, 2013).
3.1 Mixture Design
As with most asphalt mixtures, asphalt and aggregate are the two primary constituents used in
HMAC. However, unlike Marshall mixture design or Superpave mixture design, the mix design is
not driven by volumetric properties as much as it is driven by trying to pass performance-based
specifications. This method of mixture design is actually developed to assess performance in
relation to the loading and environmental conditions the mixture may experience. This type of
design methodology reduces barriers to innovation, promotes mixture performance, and
encourages the efficient use of resources (Denneman et al., 2011). Figure 1 provides a flowchart
of the basic process of developing an HMAC mixture. This section of the literature review will
explain each portion of the flowchart. Since the French have the most experience with developing
HMAC mixtures, this literature review will follow the French method and show how South Africa
has taken the European standard and adopted it to follow ASTM and AASHTO methods.
8

Figure 1 HMAC Mixture Design Process (Denneman et al., 2012)
3.1.1 Aggregate Selection
In many European countries, South Africa, and in Australia, HMAC incorporates fully crushed
aggregate due to the importance of both surface area and texture. These two properties aid in
increasing the voids in the mineral aggregate (VMA), which must be sufficient to accommodate
the higher asphalt binder content in the mixtures. Aggregate selection guidelines have been
developed to aid in properly choosing the skeletal structure. Table 1 shows an example of HMAC
aggregate selection criteria in South Africa.
9
Table 1 HMAC Aggregate Selection Criteria (Denneman et al., 2011)

Property Test Method Criteria
Fines aggregate crushing test: 10% FACT TMH1, B1 ≥ 160 kN
Hardness
Aggregate crushing value ACV TMH1, B1 ≤ 25%
Flakiness index test SANS 3001 ≤ 25%
Particle shape
>15
& texture Particle index test ASTM D3398
10-15 (Delorme, 2007)
Water absorption coarse aggregate
Water TMH1, B14 ≤ 1.0%
(>4.75 mm)
absorption
Water absorption fine aggregate TMH1, B14 ≤ 1.5%
Cleanliness Sand equivalency test TMH1, B19 ≥ 50
While these criteria were developed to ensure high quality aggregates are used in HMAC,
countries such as Latvia were not permitted to use dolomite aggregate in high modulus mixtures.
A small-scale research project was conducted using locally available aggregate in Latvia to see if
either using a polymer or hard binder in conjunction with dolomitic aggregate might equal or
better the performance of a reference mixture that represents a typical mixture (Haritonovs et
al., 2014).
Mixtures were designed using the Marshall mixture design method and subjected to TSR, rutting,
and fatigue performance testing. The reference mixture was not designed using the HMAC
methodology; thus, it had a lower asphalt content. When compared to the HMAC mixtures, the
reference mixture performed better with respect to rutting, but the lower binder content caused
reduced fatigue capacity. The HMAC with a polymer-modified binder performed better than the
hard binder in rutting. Overall, the study results showed that using a local aggregate that might
not be considered high quality might be acceptable in an HMAC mixture if specification
requirements are still met (Table 2) (Haritonovs et al., 2014).
Table 2 Compliance with Sustainable Pavement for European New Member States (SPENS)
Requirements (Haritonovs et al., 2014)
Mixtures
Parameter PMB 10/45-65 B20/30
Requirement
HMAC-1/1 HMAC-1/2 HMAC-2/2 HMAC-2/3
Void content, % 3.9 3.7 3.9 3.7 3.0 – 5.0
Rut resistance, mm/1000 cycles 0.04 0.04 0.14 0.22 0.03-0.25
Stiffness (10°C, 10 Hz) MPa 16700 16100 17100 17900 Min 14000
Min. 130 at 1
Fatigue (10°C, 10Hz), µmm/mm 130 130 130 130
million Cycles
Water Sensitivity, TSR, % 100 100 98 94 TSR 80
During an HMAC implementation study that consisted of testing HMAC with off-scale limestone,
granite, crushed cobblestone, steel slag, and basalt, it was determined that lower quality
aggregates could be used because HMAC mixture design moves away from empirical mixture
10
design and progresses to competent mixture design. This provides the opportunity to pair weaker
or lower quality materials with higher quality materials to ensure pavement performance
(Bankowski et al., 2009).
A common concern with HMAC relates to the inclusion of reclaimed asphalt pavement (RAP).
Early experiments containing 30% RAP showed issues with compaction, resistance to water, and
fatigue damage (Des Croix, 2004). Since that time, other limited laboratory studies have shown
that RAP can be included up to 40%, and 25% RAP has been included in field studies with success.
However, much of this research caveats the conclusions by stating that not all binders and RAP
sources are equivalent, and this must be evaluated on a mixture by mixture basis to ensure
performance of the mixture, which is the primary goal (de Visscher et al., 2008; Bueche et al.,
2008). Additionally, past work has shown that improper characterization of RAP might influence
low field binder contents compared to target contents during production (Nkgapele et al., 2012).
One study recently tested mixtures that had been placed in the field using 0%, 50%, and 65%
RAP. When the mixture was tested in the laboratory, all of the performance requirements for
HMAC mixtures were met even at higher RAP contents. While this work was preliminary, it did
suggest that RAP could be used to produce an EME mixture; however, like any high recycle
mixture, it is important to have good homogeneity, control, and material characterization
(Brosseaud et al., 2012).
Another study compared HMAC mixtures with 0, 15, 30, and 50% RAP. These mixtures were
analyzed mechanically for toughness (Fenix test), stiffness, rutting, moisture damage, and fatigue
resistance. The study concluded that increasing RAP content using a softer binder did not have a
negative impact on mechanical mixture performance; however, the research stated that plant
logistics may make reaching 50% RAP unattainable. When the mixtures were attempted at a local
batch plant, the 50% RAP mixture could not be produced because the non-heated RAP would not
thoroughly mixture with the heated aggregate and bitumen for material transfer of the aged
binder on the RAP (low RAP binder activation); thus, plant considerations may limit RAP use in
HMAC similar to standard asphalt mixtures (Miro et al., 2011).
3.1.2 Designing a Gradation
When considering the gradation requirements used in numerous countries, deviations occur. For
example, the grading envelopes in France (Delorme et al., 2007) are different than those of the
United Kingdom (Sanders and Nunn, 2005). Other challenges in directly translating specifications
from the French versions are differences in nomenclature and European sieve sizing.
South Africa developed a table for targeted grading curves that included both European and
metric sieve sizes (Table 3). This allowed development of gradations like those in Europe;
however, they could use their own equipment. Additionally, one should note that the French
designs were based on the maximum sieve size with requiring 100% passing at 2D, 98-100%
passing at 1.4D, and 85-98% pass at D (D = max sieve size).
11
Table 3 Target Grading Curves and Envelopes for HMAC Base Courses (Denneman et al., 2011)
Percent Passing D = 10 mm D = 14 mm D = 20 mm
Sieve Size Min Target Max Min Target Max Min Target Max
6.7 mm 47 56 68 52 54 72 46 54 66
6.3 mm 45 55 65 50 53 70 45 53 65
4.75 mm - 53 - 43 49 63 42 49 62
4.0 mm - 52 - 40 47 60 40 47 60
2.36 mm 32 36 44 28 26 42 28 36 42
2.0 mm 28 33 38 25 33 38 25 33 38
0.075 mm 6.4 6.9 7.4 5.5 6.9 7.9 5.5 6.7 7.9
0.063 mm 6.3 6.7 7.2 5.4 6.7 7.7 5.4 5.7 7.7
D = max sieve size
3.1.3 Binder Selection and Richness Factor
European Standard EN 13924 governs binder selection for HMAC mixtures. Typically, 10/25 or
15/25 pen binders have been used in Europe (Denneman et al., 2011). In the 1980s, France began
designing and producing the hard binder needed for these mixtures. In 1990, France produced
39,000 tons of the binder. By 2000, that value had grown to over 100,000 tons. The binder was
originally developed through an air blowing process; however, this increased the brittleness of
the binders and made them more susceptible to fatigue cracking. Since that time, vacuum
distillation and propane-precipitated asphalt has been used to produce the stiff asphalt needed
(Corte, 2003). Most of these binders had penetrations between 10 and 30 and softening points
greater than 60 or 70°C (EAPA, 2005). Examples of typical hard asphalt characteristics before
aging are given in Table 4. Rheological properties of common asphalts used in HMAC are provided
in Table 5.
Table 4 Typical Hard Asphalt Characteristics (Before Aging) (Corte, 2001)

Penetration Grade 15/25 10/20 5/10
R&B Softening point (°C) 66 62 to 72 87
Pfeiffer IP (Penetration Index) +0.2 +0.5 +1.0
2
Dynamic Viscosity at 170°C (mm /sec) 420 700 980
Complex Modulus at 7.8 Hz, IE*I, (MPa)
0°C 425 700 980
10°C 180 300 570
20°C 70 110 300
60°C 0.4 0.7 7

12
Table 5 Rheological Characteristics of Seven 10/20 Asphalts and a 35/50 Asphalt (Corte, 2001)

It should be noted that this hard binder is not available in all locations (Denneman et al., 2011)
and might require innovative binder modification techniques to achieve the desired results.
When Korea experimented with HMAC, they mixed a high boiling point petroleum with a
conventional asphalt to increase binder stiffness. At that point, 4% styrene-butadiene-styrene
was introduced so that the binder could maintain some ductility. Despite the addition of polymer,
the binder developed for the HMAC was still more brittle at low temperatures than conventional
or polymer-modified asphalt when determining Frass temperature (Lee et al., 2007). Thermal
distress is not common in many European countries; however, this becomes a more critical
property to monitor as HMAC is adopted in countries with colder climates (EAPA, 2005).
Today, polymer-modified asphalts (PMA) are part of the available binder selection for HMAC.
Polymer-modified asphalts resist rutting in summer months and provide flexibility to resist tensile
stresses. Some research suggests that even the incorporation of fibers as part of the PMA can
provide additional durability to reduce cracking; however, additional information is needed to
ensure these findings (Montanelli, 2013).
A similar experiment to the Korean study was conducted in Lithuania that compared HMAC with
crushed granite, crushed dolomite, and crushed gravel. Additionally, two polymer-modified
binders and a traditional HMAC hard binder were used to give the study nine mixture iterations.
Mixtures were designed and tested for stiffness modulus (LST EN 12697-26), resistance to rutting
(LST EN 12697-22), fatigue resistance four-point bending (LST EN 12697-24), and stability and
flow (LST EN 12697-34). While some differences were seen between aggregate types, the type
and amount of binder made the most difference. The study recommended that polymer-
modified binders be recommended for HMAC base mixtures and only polymer-modified binders
be used in the binder layers. This recommendation was based off of fatigue results, which showed
that the mixtures with polymer-modified binders (PMB) performed better than those without
polymer (Vaitkus, 2013).
Chappat et al., (2009) proposed similar findings that binder source is a crucial component of
producing a high performing HMAC. When comparing modulus values to fatigue strength for
mixtures using different binder sources, the research team could see tendencies for some binders
to outperform others. For example, in Figure 2, one sees that from source A, the mixture will get
the modulus but have trouble attaining the fatigue requirements. On the other hand, source B
13
produces mixtures that are too soft but have adequate fatigue performance. Knowing the
materials is vital for producing mixtures that will be judged on performance.

Figure 2 Combined Presentation of the Modulus and Fatigue Strength Results According to
Binder Source (Chappat et al., 2009)
Increased binder rigidity is commonly balanced by using higher binder contents in the mixtures
(Guyot, 2013). Minimum binder contents are given based on the class of the mixture and the
maximum aggregate size. The French have seen a continual increase in binder content and
increased binder stiffness to help produce high-modulus mixtures since the 1970s (Figure 3)
(Distin et al., 2006).

Figure 3 Evolution of Base Course Mixtures in France (Distin et al., 2006)
14
The Class 1 mixture is less fatigue resistant and designed for lower traffic volumes while the Class
2 mixture is designed for higher volumes with additional resistance to fatigue. Table 6 provides
the minimum binder contents based on aggregate density (ρ), class, and maximum aggregate
size. Binder content is calculated not through volumetric properties like in the U.S., but by
calculating a richness factor, K. However, the Asphalt Institute binder film thickness equation
seems better because is based on the actual measure of asphalt absorption; ie effective binder
film thickness. This factor is calculated through the following method (Denneman et al., 2011;
Denneman and Nkagdme 2011).
1. Calculate the specific surface area of the aggregate (Σ) using Equation 1.
𝟏𝟎𝟎𝚺 = 𝟎. 𝟐𝟓𝑮 + 𝟐. 𝟑𝑺 + 𝟏𝟐𝒔 + 𝟏𝟓𝟎𝒇 (1)
where
G = proportion of aggregate retained on and above the 6.3 mm sieve;

S = proportion of aggregate retained between the 0.25 mm and 6.3 mm sieves;
s = proportion of aggregate retained between the 0.063 and 0.25 mm sieves; and
f = percent passing the 0.063 mm sieve.
2. Calculate a correction coefficient (α) for the relative density of the aggregate (RDA) using
Equation 2 (in this case RDA = Gse).
𝟐.𝟔𝟓
𝜶= (2)
𝑹𝑫𝑨
3. Calculate the binder content of the mixture by mass of total aggregate using Equation 3.
𝟓
𝑻𝑳𝒆𝒔𝒕 = 𝑲𝜶 𝚺 (3)
4. Calculate the percent binder by mass of total mixture using Equation 4.

𝟏𝟎𝟎𝑷𝒃
𝑻𝑳𝒆𝒔𝒕 = (4)
(𝟏𝟎𝟎;𝑷𝒃 )
Table 6 Typical Values for Minimum Binder Content and Target Richness Factor (Denneman et
al., 2011)
HMAC Base Course

Class 1 Class 2
D (mm) 10, 14, 20 10, 14 20
Pb min ρ=2.65 g/cm3 3.8 5.1 5.0
3
Pb min ρ=2.75 g/cm 3.8 4.9 4.9
Richness factor, K 2.5 3.4 3.4

15
3.1.4 Performance Tests
Once the binder content is determined, the final phase of mixture design is to undergo a series
of performance tests to ensure the mixture will be durable in the field. The French suite of tests
revolves around five standards (Table 7).
Table 7 French Performance Tests (Modified from Denneman et al., 2011)

Parameter French Test Method
Workability EN 12697-31: Gyratory Compactor
Durability EN 12697-12: Duriez test
Permanent Deformation EN 12697-12: Wheel Tracker
Dynamic Modulus EN 12697-26: Flexural beam
Fatigue test EN 12697-24: Prism
The workability of the asphalt mixtures is assessed by ensuring that the mixture has less than the
maximum void content after 100 gyrations in the European gyratory compactor. South Africa
conducted a study using the Superpave gyratory compactor to assess what deviations occur when
switching from testing using the European standard to the Superpave method. The study showed
that the equivalent gyrations and average air voids were reduced for the Superpave method
(Figure 4 and Table 8).

Figure 4 Gyratory Compaction Curves for Two Mixtures using European and Superpave
Configuration (Denneman et al., 2011)
16
Table 8 Summary of Gyratory Compaction Study (Denneman et al., 2011)

HMAC Design Class 1 Class 2
Specification EN 12697-31 Superpave EN 12697-31 Superpave
Number of Specimens 9 8 5 5
Average Voids (%) after 100
4.8 3.0 3.5 2.2
Gyrations
Standard Deviation 0.8 0.9 0.8 0.4
Coefficient of Variation (%) 16.0 30.0 22.1 19.8
Equivalent Gyrations 100 43 100 46
The Duriez test is the French equivalent of AASHTO T283 for moisture susceptibility. In South
Africa, a modified Lotmann test (ASTM D4867) is used to assess durability. Most countries that
use HMAC do not differentiate tensile strength ratio (TSR) requirements between asphalt
mixtures and HMAC (Denneman et al., 2011).
EN 12697-22 is the standard in Europe for assessing rutting resistance of mixtures by wheel
tracking on an asphalt slab. The wheel tracking apparatus consists of a loaded wheel, which bears
on a sample held on a moving table. The table reciprocates with simple harmonic motion through
a distance of 230 ± 5 mm with a frequency of 53 passes (± 1%) per minute. For research purposes,
the test speed can be adjusted by inverter control. The wheel is fitted with a solid rubber tire of
outside diameter 200 mm. The wheel load under standard conditions is 700 ± 10 N. The wheel
tracker is fitted with a temperature controlled cabinet with a temperature range from
environment to 65°C ± 1.0°C. The sample may be either a 200 mm diameter core or a 300 x 400
mm slab of asphaltic mixture from 25mm to 100 mm thick. A 25 mm stroke LVDT transducer is
included for monitoring rut depth in the center of a sample during a test to better than 0.1 mm.
The deformation and sample temperature is recorded by the internal data acquisition and control
system and is then sent to the Windows® compatible software.
Other countries use their own standard rutting tests such as AASHTO T320, the Repeated Simple
Shear Test at a Constant Height. Research has consistently shown that despite the higher asphalt
content in HMAC, the stiffer binder allows the mixture to resist rutting more than a standard
bituminous base material (BTB) (Figure 5) (Denneman et al., 2011). This is an especially important
consideration in warmer climates due to the extra binder, which adds to the richness of the
mixture (Capitão and Picado-Santos, 2006).
17

Figure 5 Permanent Deformation HMAC Compared to BTB (Denneman et al., 2011)
In Europe, dynamic modulus is measured using standard EN 12697-26 (equivalent to AASHTO TP

62) where the stiffness of the asphalt mixture beam is determined (AASHTO T321). HMAC must
then have a stiffness greater than 14,000 MPa at 10°C and a frequency of 10 Hz. While just
ensuring that the mixture can surpass this stiffness at one temperature provides an easy check,
it limits the data available to the practitioner. Additionally, choosing 10°C might be an
appropriate check in the temperate climates of Europe; however, other countries such as South
Africa have decided to use a higher temperature of 15°C and still require a stiffness greater than
14,000 MPa due to hotter climates. South Africa has also implemented AASHTO TP 62 as the
method for ascertaining mixture stiffness, which is more comparable to what might be done if
HMAC was to become a design methodology in the U.S. (Denneman et al., 2011). While choosing
one temperature and frequency to check the stiffness, Figure 6 shows that HMAC (EME) mixtures
are typically stiffer than conventional mixtures across the temperature-frequency spectrum
(Petho and Denneman, 2013).
18

Figure 6 Dynamic Modulus of High-Modulus Compared to Common European Mixtures (Petho
and Denneman, 2013)
The final performance test for HMAC mixtures is fatigue. While the mixture is stiffer, it is still
important for it to retain some elasticity and resistance to fatigue cracking. The Europeans test
fatigue through the bending of a prism, which is not common in the United States. South Africa
uses a four-point bending fatigue test on beams following AASHTO T321. South African tentative
performance criteria require Class 1 mixtures to have no greater than a 70% stiffness reduction
at 310 microstrain for 10,000,000 repetitions. For a Class 2 mixture, this requirement is set for a
strain load of 410 microstrain (Denneman et al., 2011).
Some concerns have been raised about fatigue test repeatability and results. A laboratory and
full scale pavement testing study was conducted at Nantes, France to assess four different
mixtures’ behaviors to fatigue in various laboratory and field conditions. The circular test track of
LCPC used four different mixtures to make up each quadrant of the track: (1) an asphalt mixture
with 50/70 pen asphalt from one source, (2) an asphalt mixture with 50/70 pen asphalt from a
second source, (3) an HMAC with 10/20 pen asphalt, and (4) a road base asphalt with 50/70 pen
asphalt. In addition to the field work, each mixture was subjected to fatigue testing using the
following procedures: (1) two-point bending fatigue tests on trapezoidal samples with controlled
strain, with and without rest periods; (2) two-point bending fatigue tests on trapezoidal samples
with controlled stress, without rest periods; and (3) three-point bending fatigue tests on
parallelepiped-like samples, with control stress, with and without rest periods.
The laboratory rankings of the materials depended greatly on the testing procedure, showing
that choosing the correct testing protocol is critical for ensuring that the right mixture properties
19
are being analyzed. However, for the field experiment, 1.0 cm of very thin asphalt concrete
(VTAC) was placed over 7.7 cm of HMAC, and it was compared to 1.5 cm of VTAC over 10 cm of
road base asphalt. All asphalt materials were placed over a softer base than is typically found in
France. The study used a 65 kN dual wheel at 10 rpms (70 kph linear due to 19 m radius of testing
device) to load the pavements for 2,665,000 load repetitions. Figure 7 shows less deflections in
the typical road base (GB3) compared to the HMAC (EME) mixtures; however, the authors also
note this was not the typical condition for EME mixtures due to the softer subgrade. As for
cracking, the HMAC was the last mixture to exhibit fatigue cracking; however, once it exhibited
cracking, it tended to progress faster, showing the more brittle nature of the material (Figure 8).
It should be noted that these experiments were conducted before polymers were commonly
used in HMAC binders (de La Roche et al., 1994).

Figure 7 Deflections Referred to 20°C (de La Roche et al., 1994)
20

Figure 8 Cracking (de La Roche et al., 1994)
3.2 Pavement Design
The National Asphalt Pavement Association recognized the potential for HMAC to be an integral
part of long-life or perpetual pavement design. The organization recognized early that the French
had used HMAC to justify thinner sections for their long-life pavement designs (Newcomb and
Hansen, 2004; Newcomb et al., 2010).
The French have developed an analytical pavement design method that can capture the effects
of HMAC. LCPC has developed Alize-LCPC software, which uses a mechanistic-empirical approach
where traffic, material properties, and performance coefficients are used to predict pavement
performance (Guyot, 2013), but other design programs can be used that incorporate material
properties into design.
21
One case study that incorporated HMAC as a design option was the 2011-12 runway overlay and
new parallel taxiway at the Sir Seewoosagur Ramgoolam International Airport in Mauritius in the
Indian Ocean. When comparing the typical runway buildup from the FAARFIELD Airport
Pavement Design Software sponsored by the FAA, the use of HMAC reduced the necessary
runway thickness by 105 mm. Not only did this save money and natural resources, but a study
showed that it also saved approximately 13% in greenhouse gas emissions (Guyot, 2013).
As Thailand was considering using HMAC in pavement structures, Alize-LCPC was used to conduct
structural analyses given an expected modulus of 14,000 MPa and fatigue resistance (106 cycles)
at 130 microstrain for the HMAC. The road base asphalt was expected to have a modulus of 9,300
MPa and a fatigue resistance (106 cycles) at 90 microstrain. The design software showed that for
similar wearing courses, 7 cm less structure was needed for the HMAC mixtures, thus, reducing
the required pavement structure by 20% (Lefant, 2012). Other sources cite that some
governments have seen a 30% reduction in needed pavement structure due to the increased
stiffness of the pavement and added fatigue resistance (Corte, 2003).
Carbonneau et al. (2008) conducted an experiment where they compared the reference cross-
section of the Herning bypass to two cross-sections containing HMAC mixtures, called a HMA
GAB II mixture in Denmark. When the mechanical properties of the two mixtures were placed
through a mechanistic-design program, the use of the HMAC mixture allowed the bypass cross-
section to be reduced by 25 mm in thickness. The two final cross-sections are given in Figure 9.
This analysis resulted in the HMAC mixture being used to reduce material quantities, and the
roadway is currently being monitored (Carbonneau et al., 2008).

Figure 9 Roadway Structure of Mixture Design Based on Characteristic Mechanical Gain of
HMA GAB II (Carbonneau et al., 2008)
22
Weilinski and Huber (2011) showed these mixtures could be incorporated into the current
version of the Mechanistic-Empirical Pavement Design Guide (MEPDG). The design incorporated
American materials from Indiana that included recycled asphalt shingles (RAS), RAP, binder, and
aggregate. The team found that binder stiffness had little to do with performance in the MEPDG
on rutting, IRI, or fatigue cracking. These results were similar when comparing stiff mastercurves
like the PG 64-22 binder with RAS compared to the PG 76-22 binder mastercurve. The increased
mixture stiffness did improve fatigue performance and ride in the MEPDG. Additionally, using the
HMAC reduced the pavement thickness by 16% to achieve similar performance.
3.3 Performance
A small-scale laboratory experiment was conducted in China to assess the impact of high-
modulus asphalt on rutting resistance when compared to polymer-modified and conventional
softer binders when used in the binder course of a pavement. Using finite element modeling, the
shear strains and compressive stresses within the middle of the pavement were calculated
through a typical pavement cross-section for the area. The results (Figure 10) show that using the
high-modulus materials reduced shear strain. These results also showed that despite the higher
binder content used in an HMAC design, the rutting resistance of the pavements using the HMAC
was increased (Wei et al., 2010).

Figure 10 Tendency of Shear Strain (Wei et al., 2010)
A field experiment was developed outside of Brussels, Belgium to assess how aggregate skeleton,
recycled materials, binder content, and grade all influenced field performance of pavement
structures with HMAC binder courses. Each 140-m test section was constructed with a 9 cm
variant of an HMAC with a 3 cm stone matrix asphalt (SMA) or porous asphalt surface. During
construction, cores were taken from each test section and tested for rutting using EN 12697-22
at 50°C. The results (except for the porous asphalt sections) showed that using HMAC improved
rutting resistance. The differences in aggregate skeleton (stony versus sandy) did not impact the
rutting values of the mixtures, nor did binder type. The primary driver for rutting resistance was
binder content, as mixtures with lower binder contents had less rutting (Figure 11). After a year
in the field, the high-modulus test sections all out-performed the control test section (DeBacker
et al, 2008; DeBacker, deVisscher et al., 2008).
23

Figure 11 Rut Depth as a Function of Binder Content for Mixture with Sand Skeleton
(deVisscher et al., 2008)
The country of Korea wanted to evaluate the use of HMAC for long-life pavements. After
developing an HMAC mixture design, the HMAC was compared to a conventional asphalt mixture
in performance tests before being introduced at an accelerated loading facility. The mixtures
were evaluated for stiffness (Figure 12), fatigue in indirect tension (Figure 13), rutting via KS
F2374 test procedure, and moisture damage using ASTM D4867 (Table 9). In all cases, the HMAC
was shown to perform better than the conventional mixture.

Figure 12 Dynamic Modulus Mastercurves at 15°C Reference Temperature (Lee et al., 2007)
24

*RP = rest period
Figure 13 Results of Fatigue for the Conventional Mixture and HMAC (HMAM) (Lee et al.,
2007)
Table 9 Performance Test Results from Korean Laboratory Study (Data from Lee et al., 2007)
Mixture Dry Strength, kPa Wet Strength, kPa TSR Rut Depth, mm
Conventional 1070.9 948.3 88.54 7.28
HMAC 1515.1 1489.6 98.32 2.79
Due to the success in the laboratory, these mixtures were then compared in the field at the
Hanyang University Accelerated Pavement Tester. In this procedure, an 11-ton load (maximum
axle load in Korea is 10 tons) is applied to 12.5 m of pavement at a maximum speed of 17 km/h.
Two lanes of these mixtures were produced. The first lane was designed to study fatigue cracking
and the experimental mixtures were constructed at 94 and 83 mm in thickness for the
conventional and HMAC mixtures, respectively. The second lane was designed to study rutting
and the mixtures were constructed thicker. The conventional mixture was 268 mm thick while
the HMAC mixture was 215 mm thick. Strain gauges were placed at the bottom of the asphalt
layer to characterize the pavement response.
The results showed that despite having a thinner cross-section, the HMAC could reduce the
tensile strains in the test sections except for the thin test section at the lowest wheel loading
(Figure 14). The preliminary results suggested that this fatigue performance would, in fact, allow
Korea to use these mixtures as part of a long-life pavement concept with additional validation as
no fatigue cracking was noticed in either test section at the end of the experiment. The high-
modulus mixtures also performed better than the conventional mixtures in terms of rutting
25
(Figure 15). The HMAC mixture had less than half of the rutting in the conventional mixture
despite being constructed on a thinner asphalt cross-section (Lee et al., 2007).

Figure 14 Tensile Strain with Change of Dual Wheel Load for (a) Thin Pavement Section and
(b) Thick Pavement Section (Lee et al., 2007)
26

Figure 15 Comparison of Rut Depths for Conventional Mixtures and HMAC Pavement (Lee et
al., 2007)
As part of the implementation effort in Europe, two HMAC mixtures were compared to two basic
asphalt mixture designs in the laboratory and the field. Field testing was conducted under a heavy
vehicle simulator (HVS). Structural cross-sections of the four test sections are provided in Table
10. The HVS applied 60 kN via a single axle loading with a tire pressure of 800 kPa at a speed
ranging from 10-12 kph.
Table 10 Structural Build-Ups (Information from Barkowski et al., 2007)

Pavement
A B C D
Layer
Thickness: 4 cm
Thickness: 2 cm Thickness: 4 cm
Wearing Mixture type:
Mixture type: SMA Mixture type: SMA
Course porous
Binder: PmB 45/8065 Binder: PmB 45/80-65
Binder: 50/70
Thickness: 10 cm Thickness: 10 cm Thickness: 8 cm Thickness: 7 cm
Binder Mixture type: HMAC Mixture type: AC Mixture type: AC Mixture type: HMAC
Course 16 16W 16W 15
Binder: 20/30 Binder: 35/50 Binder: 35/50 Binder: 20/30
Unbound Thickness: 20 cm
Materials Aggregate: dolomite
Additionally, modeled fatigue life and damage were determined based on the Asphalt Institute
transfer function considering a 60 kN with the HVS wheel configuration. The two HMAC test
sections showed the best performance (Figure 16). It is interesting to note that section D had
higher measured and modeled strains than Section A (Figure 17); however, it had better expected
27
performance. This may have been due to the differences in fatigue performance of the HMAC in
the binder/base course (Bankowski et al., 2009).

Figure 16 Comparison of Fatigue Life for Each Individual Section (Bankowski et al., 2009)

Figure 17 Comparison of Measured and Calculated Strains at the Bottom of the Asphalt Layers
(Bankowski et al., 2009)
In 2010, the Virginia Transportation Research Council (VTRC) reported the results of a study of
field trials of high-modulus high-binder-content base layer hot-mix asphalt mixtures
28
(Diefenderfer and Maupin, 2010). Three locations where deep rehabilitation or new construction
were selected, and HMA base mixtures were used at designed asphalt content, designed asphalt
content plus 0.4% additional asphalt, and/or designed asphalt content plus 0.8% additional
asphalt. Two of the field trial locations had no construction-related issues; difficulties during
compaction occurred at the third. The results of this study indicated that the binder stiffness for
an HMAC mixture should be at least equivalent to that of a PG 70-22 binder to guard against
potential rutting and addition of RAP may be necessary.
3.4 Construction
While little has been published regarding the construction of these mixtures, the Belgian study
did gain some insights to the approach. At times, conventional methods were not appropriate
either due to the design asphalt content or binder stiffness; thus, the research team made note
of the following items: (1) contractors must use binder producer’s recommended temperatures
during production; (2) common aggregate size to lift thickness ratios did not apply, as HMAC was
placed 9 to 10 cm thick without any problem; (3) mixtures were easy to compact with traditional
equipment; (4) compaction temperatures were commonly about 10°C higher than conventional
mixtures; (5) the fatty look of the mixture does not indicate over compaction; and (6) voids ratios
were similar to those in conventional binder courses (DeBacker et al., 2008). Denneman (2011)
observed that these mixtures commonly require higher mixing temperatures.
Nicholls et al. (2008) conducted an experiment to monitor the durability and buildability of HMAC
on five different sites in the UK. During the pilot projects, some instances were noted when the
level of compaction was not achieved (Site D); however, on other trial projects, the in-place air
voids were extremely low (1% at site B). Falling weight deflectometer and laboratory tests
showed that with few exceptions, the mixtures were designed, produced, and constructed well.
Michaut (2014) provided the following recommendations for producing and laying HMAC:
• Mixing temperature should be between 160 and 180°C and always less than 190°C.
• The minimal laying temperature for this mixture is 145°C, but this will depend on binder
properties.
• Granular base must be compacted well to ensure high in-situ density of HMAC.
Jamois et al. (2000) notes that sometimes these temperatures can be exceeded if material
properties dictate the need. Mixtures placed at the circular test track in France were produced
between 200 and 210 °C and placed at 195°C.
3.5 Summary of Current Practice
A comprehensive literature study was performed to assess the current state-of-the-practice on

HMAC mixture design, pavement design, laboratory performance tests, and full scale pavement
performance. The majority of the observed experience comes from Europe. The French HMA
mixture design method is the most commonly used methodology and has some variations in the
design procedure compared to conventional Superpave design mixtures commonly utilized in the
29
United States. The first main difference is the method of compaction. The determination of
minimum binder content in the French design method is also quite different from Superpave
design. The French method calls for a minimum asphalt binder content based on the richness
factor, surface area, and specific gravity of the aggregates. However, as mentioned before, the
AI Hveem-Edward binder film thickness calculation is as good. Asphalt mixture design procedures
include performance testing requirements for moisture damage, a rut test for rutting, complex
modulus for structural stiffness, and fatigue testing for fatigue cracking.
Pavement design and analysis of HMAC mixtures is conducted using mechanistic-empirical

approaches. This practice includes evaluation of potential field performance and reduction in
needed pavement structure due to the increased stiffness of the pavement and added fatigue
resistance. Full scale performance testing has been used to validate this added fatigue resistance.
Overall, there is an expected difference in the laboratory performance of HMAC mixtures when
compared to traditional asphalt mixtures. This performance difference is expected to translate
into the field where pavements can either be designed thinner with the same expected life or
designed at the same thickness with long-life performance as a viable expectation. Table 11
exhibits an example of change in specifications for HMAC mixtures compared to conventional
asphalt concrete mixtures in Europe.
Table 11 Roadbase High-Modulus Asphalt Concrete versus Traditional AC (EAPA, 2005)

Roadbase High-Modulus Asphalt Concrete versus Traditional AC
Test HMACR1 HMACR2 AC
Inmersion Copression test at 18°C >0.7 >0.75 >0.7
Rutting test at 60°C 30,000 cycles <8% <8% <10%
Stiffness modulus at 15°C and 10Hz >14,000 MPa >14,000 MPa >9,000 MPa
Allowed microstrain from fatigue law at
6 >100 >130 >90
10°C and 25 Hz and for 10 cycles
Void content for laying thickness <10% <6% <10%
4. EXPERIMENTAL DESIGN AND ANALYSIS
4.1 Laboratory Testing
There were two objectives addressed in the laboratory experimental plan: (1) determine the
engineering properties of high-modulus mixtures, and (2) determine whether or not the
recommended design procedures were appropriate. The information to accomplish both
objectives was obtained from European experience. This section details the approach adopted
to address the two objectives of the laboratory research.
To assess statistical differences among mixtures, the general linear model (GLM) (α = 0.05) was
conducted. Overall comparisons of such properties were made using Tukey-Kramer analysis with
the results from all laboratory performance tests. The results of the laboratory testing were also
used to determine if the current testing procedures could adequately predict the performance
30
of pavements containing these types of mixtures in the field. The laboratory testing program
evaluated binder performance grade, mixture stiffness over a wide temperature range, fatigue
cracking, and permanent deformation, as follows:
• Volumetric mixture design and material characterization,

• Mixture stiffness: dynamic modulus (AASHTO TP 79-13),
• Flow number (AASHTO TP 79-13), and
• AMPT cyclic fatigue (AASHTO TP 107-14).
4.2 Mixture Design
The French asphalt mixture design method has some variations in the design procedure
compared to conventional Superpave design mixtures commonly utilized in the United States.
The French gyratory compactor uses an internal angle of 0.82 degrees, whereas the Superpave
gyratory compactor employs an internal angle of 1.16°. Wielinski and Huber (2011) used in their
research study the results of a comparison study of the LCPC gyratory compactor and a Superpave
gyratory compactor for EME mixture design that was conducted by the Jiangsu Transportation
Research Institute. This comparison work determined that 80 gyrations in the Superpave
compactor produced the same compaction as 100 gyrations in the LCPC gyratory compactor.
For this study, samples were compacted at 80 gyrations in the Superpave compactor, and the
target design air voids at Ndes was set from 3.0 to 6.0% (European specification requires design
air voids to be less than six percent). For dynamic modulus testing, the range of allowable air void
content was also 3.0 to 6.0% with a minimum dynamic modulus at 15 °C and 10 Hz of 14,000
MPa. In addition, the gradation selected for each trial dictated the minimum binder content
required in the design according to the French methodology.
The experimental plan included a variety of mixtures with different material and binders such
that higher moduli were obtained compare to conventional mixtures. The plan included a French
mixture with a stiff binder (PG 88-16), two mixtures containing 35% RAP both with polymer-
modified binders, but one high polymer content (HiMA), another mixture containing 25% RAP
and 5% RAS with a polymer-modified binder, and finally, a 50% RAP mixture with a polymer-
modified binder.
Table 12 shows the aggregate gradations and blend formulas for the four mixtures that were
produced. Table 13 shows the volumetric properties of each mixture determined during the
design phase.
31
Table 12 Aggregate Gradations of Mixtures

Sieve Size French 35% RAP 25% RAP, 5% RAS 35% RAP 50% RAP
(in.) EME 14 PG 76-22 PG 76-22 HiMA PG 76-22
2" 100.0 100.0 100.0 100.0 100.0
1.5" 100.0 100.0 100.0 100.0 100.0
1" 100.0 99.3 98.6 99.3 98.8
3/4" 100.0 95.5 91.0 95.5 92.4
1/2" 88.9 89.3 81.8 89.3 85.5
3/8" 79.7 79.5 72.6 79.5 78.5
#4 58.9 54.9 51.5 54.9 51.1
#8 37.7 42.7 41.5 42.7 40.7
#16 26.6 32.5 32.0 32.5 31.2
#30 19.3 22.5 21.5 22.5 21.3
#50 14.0 12.1 11.1 12.1 10.6
#100 8.8 7.1 6.4 7.1 5.9
#200 7.9 4.7 4.2 4.7 3.9

32
Table 13 Mixture Design Properties

French 35% RAP 25% RAP, 5% RAS 35% RAP 50% RAP
Mixture Designation
EME 14 PG 76-22 PG 76-22 HiMA PG 76-22
Gyration level 80 80 80 80 80
NMAS (U.S. sieves) 19 19 19 19 19
Binder content (%) 5.70 5.12 5.01 5.12 5.04
Base binder content (%) 5.70 3.24 2.73 3.24 2.98
PG 76-22 PG 76-22 PG 88-22 PG 76-22
Base binder grade PG 88-16
(SBS) (SBS) (High SBS) (SBS)
Percent RAP 0 35 25 35 50
Percent RAS 0 0 5 0 0
RAP AC% n/a 5.37 5.37 5.37 4.12
RAS AC% n/a n/a 18.69 n/a n/a
RAP binder ratio n/a 0.367 0.268 0.367 0.409
RAS binder ratio n/a n/a 0.187 n/a n/a
Gsb 2.751 2.716 2.730 2.716 2.690
Gmm 2.478 2.542 2.543 2.542 2.530
Design air voids (%) 1.5 2.0 3.0 2.3 2.6
Gmb design 2.441 2.491 2.467 2.484 2.464
VMA 15.0 13.0 14.2 13.2 13.0
VFA 90.0 84.6 78.8 82.6 80.0
Gb 1.028 1.028 1.028 1.028 1.028
Gse 2.709 2.761 2.757 2.761 2.743
Pba n/a 0.62 0.37 0.62 0.73
Pbe n/a 4.53 4.66 4.53 4.34
Dust proportion n/a 1.03 0.91 1.03 0.90
E* at 15°C and 10 Hz, MPa 17,506 14,519 15,753 14,457 17,137
Air voids - E* specimen (%) 3.4 3.6 4.0 4.00 3.8
4.3 Dynamic Modulus
A single point measurement (E* at 15°C and 10 Hz) cannot be expected to describe material
behavior across all possible loading temperatures/frequencies; therefore, Dynamic Modulus
testing was performed for all mixtures according to AASHTO TP 79-13. In addition, these results
were used to estimate pavement performance using the AASHTOWare Pavement ME Design
software.
Samples were compacted to a height of 175 mm and a diameter of 150 mm and prepared to
meet the tolerances outlined in AASHTO PP60-14. Dynamic modulus testing was performed in an
IPC Global Asphalt Mixture Performance Tester (AMPT), shown in Figure 18. Dynamic modulus
testing is performed in order to quantify the stiffness of the asphalt mixture over a wide range of
testing temperatures and loading rates (or frequencies). The temperatures and frequencies used
for testing these mixtures are those recommended by AASHTO PP61-13. For this methodology,
33
the high test temperature is dependent on the high PG grade of the base binder utilized in the
mixture being tested.

Figure 18 IPC Global Asphalt Mixture Performance Tester
Dynamic modulus testing was performed in accordance with AASHTO TP 79-13 in an unconfined
condition. Unconfined data is most commonly used for dynamic modulus testing since current
mechanistic design software packages were calibrated using unconfined dynamic modulus data.
Unconfined testing is also significantly easier to perform than confined testing and ME packages
were calibrated using unconfined results.
The collected data were used to generate a mastercurve for each individual mixture. The
mastercurve uses the principle of time-temperature superposition to horizontally shift data at
multiple temperatures and frequencies to a reference temperature so that the stiffness data can
be viewed without temperature as a variable. This method of analysis allows for visual relative
comparisons to be made between multiple mixtures.
Generation of the mastercurve also allows for generation of the dynamic modulus data over the
entire range of temperatures and frequencies required for mechanistic-empirical pavement
design. By having an equation for the curve describing the stiffness behavior of the asphalt mix,
both interpolated and extrapolated data points at various points along the curve can then be
calculated. The general form of the mastercurve equation is shown as Equation 5. As mentioned,
the dynamic modulus data are shifted to a reference temperature by converting testing
frequency to a reduced frequency using the Arrhenius equation (Equation 6). Substituting
Equation 6 into Equation 5 yields the final form of the mastercurve equation, shown as Equation
7. The shift factors required at each temperature are given in Equation 8. A reference
temperature of 20oC was used for this analysis. The limiting maximum modulus in Equation 8 is
calculated using the Hirsch Model, shown as Equation 9. The Pc term, Equation 10, is simply a
34
variable required for Equation 9. A limiting binder modulus of 1 GPa is assumed for this equation.
Non-linear regression is then conducted using the ‘Mastersolver.exe’ program to develop the
coefficients for the mastercurve equation. Typically, these curves have an Se/Sy term of less than
0.05 and an R2 value of greater than 0.99. Definitions for the variables in Equations 5 to 10 are
given in Table 14.
(𝑴𝒂𝒙;𝝏)
𝑳𝒐𝒈 𝑬∗ = 𝝏 + (5)
𝟏E𝒆𝜷G𝜸𝒍𝒐𝒈𝒇𝒓
∆𝑬𝒂 𝟏 𝟏
𝒍𝒐𝒈𝒇𝒓 = 𝒍𝒐𝒈 𝒇 + − (6)
𝟏𝟗.𝟏𝟒𝟕𝟏𝟒 𝑻 𝑻𝒓
(𝑴𝒂𝒙;𝝏)
𝒍𝒐𝒈 𝑬∗ = 𝝏 + ∆𝑬𝒂 𝟏 𝟏 (7)
𝜷G𝜸 𝒍𝒐𝒈𝒇G Q
𝟏E𝒆 𝟏𝟗.𝟏𝟒𝟕𝟏𝟒 𝑻 𝑻𝒓
∆𝑬𝒂 𝟏 𝟏
𝐥𝐨𝐠 [𝒂 𝑻 ] = − (8)
𝟏𝟗.𝟏𝟒𝟕𝟏𝟒 𝑻 𝑻𝒓
𝑽𝑴𝑨 𝑽𝑭𝑨∗𝑽𝑴𝑨 𝟏;𝑷𝒄

|𝑬∗ |𝒎𝒂𝒙 = 𝑷𝒄 𝟒, 𝟐𝟎𝟎, 𝟎𝟎𝟎 𝟏 − + 𝟒𝟑𝟓, 𝟎𝟎𝟎 + 𝑽𝑴𝑨 (9)
𝟏𝟎𝟎 𝟏𝟎,𝟎𝟎𝟎 𝟏Q
𝟏𝟎𝟎 𝑽𝑴𝑨
E
𝟒,𝟐𝟎𝟎,𝟎𝟎𝟎 𝟒𝟑𝟓,𝟎𝟎𝟎(𝑽𝑭𝑨)
𝟒𝟑𝟓,𝟎𝟎𝟎 𝑽𝑭𝑨 𝟎.𝟓𝟖

𝟐𝟎E
𝑽𝑴𝑨
𝑷𝒄 = 𝟒𝟑𝟓,𝟎𝟎𝟎(𝑽𝑭𝑨) 𝟎.𝟓𝟖
(10)
𝟔𝟓𝟎E
𝑽𝑴𝑨
Table 14 Mastercurve Equation Variable Descriptions

Variable Definition
|E*| Dynamic modulus, psi
δ, β, and γ Fitting parameters and parameters describing the shape of sigmoidal function
Max Limiting maximum modulus, psi
fr Reduced frequency at the reference temperature, Hz
f The loading frequency at the test temperature, Hz
ΔEa Activation energy (treated as a fitting parameter)
T Test temperature, oK
Tr Reference temperature, oK
a(T) The shift factor at temperature, T
|E*|max The limiting maximum HMA dynamic modulus, psi
VMA Voids in mineral aggregate, %
VFA Voids filled with asphalt, %
Figure 20 exhibits the mastercurves for all mixtures including the 19.0mm NMAS base course
(control mixture) from the 2009 NCAT Test Track cycle. It can be observed that at the low
temperature, high frequency end of the curve, all of the mixtures tended to have similar E*
values. However, when moving towards the opposite range of temperatures and frequencies,
slight differences can be observed, especially for the mixture containing recycled asphalt
shingles. These trends can also be observed when analyzing the mastercurve regression
35
coefficients. Table 15 gives a summary of the mastercurve regression coefficients that were
generated using the modified MEPDG mastercurve model. Goodness of fit parameters are also
shown in Table 15.
Maximum E* values were similar; however, minimum E* values did show significant differences.
In terms of the steepness of the curve given by the parameter -γ, the 25% RAS and 5% RAS
mixture showed the lowest slope (less susceptible to changes in frequency), and the 50% RAP
mixture showed the highest slope (most susceptible to changes in frequency). The inflection
point frequency parameter -β/γ was the highest for the 25% RAS and 5% RAS mixture (4.55 Hz),
followed by the French EME mixture (3.61 Hz); the remaining mixtures had similar inflection
points around 2.85 Hz. The activation energy term is best regarded as an experimentally
determined parameter that indicates the sensitivity of the shift factors to temperature, and
consequently affects the shape of the mastercurve. In this case, all mixtures had similar activation
energy terms, but the 25% RAP-5% RAS mixture had the highest term producing a wider range of
reduced frequencies and a more flattened curve.
10,000
Dynamic Modulus, ksi
1,000
100
French EME 35%RAP PG 76-22

25% RAP, 5% RAS 35%RAP HiMA
50% RAP PG 76-22 NCAT Control
10
1.0E-06 1.0E-04 1.0E-02 1.0E+00 1.0E+02 1.0E+04
Reduced Frequency, Hz

Figure 20 Dynamic Modulus Mastercurves
Table 15 Mastercurve Coefficients

Max E* Min E*
Mixture ID Beta Gamma EA R2 Se/Sy
(Ksi) (Ksi)
French EME 3240.96 8.57 -2.023 -0.560 207939.5 0.991 0.066
35%RAP PG 76-22 3417.23 17.05 -1.697 -0.595 197985.7 0.993 0.057
25% RAP, 5% RAS 3389.56 10.40 -1.972 -0.433 244676.9 0.985 0.088
35%RAP HiMA 3423.82 26.10 -1.548 -0.556 215061.3 0.997 0.041
50% RAP PG 76-22 3446.47 18.40 -1.778 -0.610 206143.0 0.993 0.058
36
In an attempt to identify testing variability and/or non-linearity in the material behavior due to
non-compliance to the recommended micro-strain levels, the dynamic modulus and phase angle
were averaged for each laboratory’s data and plotted in Black Space (Airey, 2002; Christensen et
al, 2003). Figure 21 contains the Black Space plots for all the different mixes including the 19.0
mm NMAS control mixture from the 2009 NCAT Test Track cycle. It should be noted that all plots
show good uniformity in their respective Black Space diagrams, as noted with their R2 values
being greater than 0.94 for a 4th-order polynomial fitted function. Due to the interaction of the
asphalt binder with aggregate, the Black Space diagram for a mixture shows a peak phase angle
value at intermediate dynamic modulus. At high temperatures, the aggregate structure begins to
dominate behavior of the mixture while at lower temperatures volumetric properties and binder
stiffness control the behavior. This peak value is associated with the inflection point in the
mastercurve using the terms described earlier (-β/γ).French EME and 50% RAP mixtures had
similar peak phase angles around 33 degrees but different peak dynamic modulus of 220 ksi and
237 ksi, respectively. For the 35% RAP mixture, this peak occurs around dynamic modulus of 230
and for the 25% RAS-5% RAS and 35% RAP HiMA mixtures, this occurs around dynamic modulus
of 205 ksi and 218 ksi, respectively.
Additional analysis of the Black Space diagram indicates that mixtures with lower phase angle
values are more elastic (25% RAS-5% RAS, 35% RAP HiMA mixtures). On the other hand, if the
phase angle is high, the mixture is more viscous (French EME and 50% RAP mixtures) (Rahbar-
Rastegar and Daniel, 2016). In addition, stiffer mixtures at lower phase angles are more
susceptible to cracking (Anderson et al., 2011) In this case, the 35% RAP and 50% RAP mixtures
have slightly higher moduli at low phase angles than the other mixtures.
35
30
Phase Angle, degrees
25
20
French EME R² = 0.99871
15 35% RAP PG 76-22 R² = 0.9974
25%RAP 5%RAS R² = 0.9976
10
35%RAP HiMA
R² = 0.99646
5 50%RAP PG 76-22
R² = 0.99757
NCAT Control Base Course
R² = 0.99673
0
4 4.5 5 5.5 6 6.5 7
Log E*, psi
Figure 21 Black Space Diagrams
To assess statistical differences, a general linear model (GLM) (α = 0.05) was conducted on the
test data measured at 4°C, 20°C, and 45°C, and at two frequencies: 10 Hz and 1Hz. Thus, the GLM
37
was completed four times to assess statistical differences at each temperature. The Tukey-
Kramer test (α = 0.05) was used to determine where these statistical differences occurred and
how the mixtures grouped within each project. Table 16 shows the results of the Tukey-Kramer
test on E* values for 20°C and 45°C only (results at 4°C followed similar statistical trend of results
at 20°C). Mixtures given the same letter in the table were statistically grouped together (no
statistical difference among mixtures at α = 0.05). As expected, at high temperatures and low
frequencies, statistical differences were obtained for some of the mixtures. The results indicated
that the mixture with 25% RAP and 5% RAS had the highest E* values while the mixture with 35%
RAP HiMA had the lowest E* values. On the other hand, no statistical differences were obtained
among mixtures at 20°C for a significance level α = 0.05.
Table 16 E* Statistical Grouping

20 °C, 10 Hz 20 °C, 1 Hz 45 °C, 10 Hz 45 °C, 1 Hz
Mixture ID
Mean Group Mean Group Mean Group Mean Group
French EME 2,103.7 A 1,517.8 A 615.8 A B 293.1 A B
35% RAP PG 76-22 1,982.9 A 1,388.3 A 545.7 A B 264.1 B
25% RAP, 5% RAS 2,086.0 A 1,544.7 A 673.0 A 373.2 A
35% RAP HiMA 1,910.1 A 1,375.6 A 505.6 B 261.5 B
50% RAP PG 76-22 2,119.8 A 1,503.8 A 555.1 A B 265.5 B
4.4 Flow Number
The Flow number test is a rutting resistance test that is performed using the AMPT. It applies a
repeated compressive loading to an asphalt specimen while the AMPT records the deformation
of the specimen with each additional loading cycle. The user defines the temperature, applied
stress state (deviator stress and confining stress), and number of cycles at which the test is
performed. The loading is applied for a duration of 0.1 seconds followed by a 0.9 second rest
period every 1 second cycle. Flow number data is commonly modeled with the Francken model,
shown as Equation 11 (AASHTO TP 79-13). An example of unconfined flow number test data is
shown in Figure 22.
ε p ( N ) = aN b + c(e dN − 1)
(11)
where
εp (N) = permanent strain at ‘N’ cycles,

N = number of cycles, and
a, b, c, d = regression coefficients.
The flow number is defined as the number of cycles at which the sample begins to rapidly fail and
coincides with the minimum rate of strain accumulation measured during the test. This is more
properly defined as the breakpoint between steady-state rutting (secondary rutting) and the
more rapid failure of the specimen (tertiary flow). Figure 22 demonstrates this concept
graphically. If the samples do not exhibit tertiary flow (common for confined samples), then the
38
amount of deformation at a specified loading cycle can still be used to give a relative ranking of
tested mixtures with respect to rutting susceptibility.

Figure 22 Typical Flow Number Test Data
Flow number testing for this project was performed in accordance with AASHTO TP 79-13 in an
unconfined state with a deviator stress of 87 psi. The tests were run until either the samples
reached 5% axial strain (7.5 mm of deformation on a 150 mm sample) or the test went the full
20,000 cycles. Samples were prepared in accordance with AASHTO PP 60-14 to a target air void
level of 3.5 ± 0.5 percent on the final cored and trimmed specimen. By AASHTO TP 79-13, the
flow number test temperature is selected based on the LTPPBind 50% reliability high pavement
temperature at the project location adjusted for a 20 mm depth in the pavement structure. The
Auburn, Alabama climate region was assumed to generate the flow number test temperature.
The temperature data from the LTPPBind v3.1 software for Auburn is shown in Figure 22 below.
Based on these criteria, the temperature of 59.5°C was selected for this project (LTPPBind
temperature rounded to the nearest 0.5°C). While AASHTO TP 79-13 does contain traffic level
criteria for mixtures based off their flow number results, these criteria are not completely
applicable for this study since the specimens were fabricated to a different target air void content
(these criteria are for specimens fabricated to 7.0 ± 0.5 percent air voids).
39

Figure 22 LTPPBind v3.1 Output for Auburn, AL Area
Table 17 shows the results of the flow number test performed on all the mixtures and the
statistical grouping. The GLM (α = 0.05) showed no statistical difference between the 35% RAP
and the 50% RAP mixtures. On the other hand, the remaining mixtures were statistically different
from each other. The 35% RAP HiMA mixture showed the highest resistance to permanent
deformation followed by the 25%-5% RAS mixture. Similar strain values were obtained for all
mixtures but the French EME mixture (more ductile). All of the mixtures exhibited flow number
values well in excess of the 740 recommended for a greater than 30 million ESAL design
pavement by AASHTO TP 79-13. While these criteria are not completely applicable given the
aforementioned air void level of the specimens, it does give some frame of reference for the high
level of rutting resistance offered by these high-modulus mixtures.
40
Table 17 Flow Number Test Results
Mixture Air Voids (%) Francken Flow Number Francken Microstrain at FN FN Statistical
ID Average Average Standard CV Average Standard CV (%) Group
Deviation (%) Deviation
French 3.5 4,665 716 15.3 29,381 2,271 7.7 A
EME
35% RAP 3.2 1,910 577 30.2 17,648 1,365 7.7 B
PG 76-22
25% RAP, 3.5 8,229 676 8.2 15,128 1,367 9.0 C
5% RAS
35% RAP 3.3 18,374 1,807 9.8 16,362 3,122 19.1 D
HiMA
50% RAP 3.1 1,337 485 36.3 15,987 1,406 8.8 B
PG 76-22
4.5 AMPT Cyclic Fatigue
Fatigue testing for this project was performed using the uniaxial tension fatigue method available
in the Asphalt Mixture Performance Tester (AMPT). This method is summarized in AASHTO TP
107-14. This methodology utilizes the simplified viscoelastic continuum damage (S-VECD) model
(Hou et al., 2010). Hereafter, this testing protocol will simply be referred to as AMPT cyclic fatigue
S-VECD testing. AMPT cyclic fatigue analysis on a given mixture requires both dynamic modulus
(|E*|) testing as well as uniaxial fatigue testing. S-VECD is a mode-of-loading independent,
mechanistic model that allows for the prediction of fatigue performance parameters at different
temperatures and loading conditions (Jacques et al., 2016; Daniel and Kim, 2002; Underwood et
al., 2012).
Specimen preparation for cyclic fatigue testing is identical to that for the AMPT specimens
required for the dynamic modulus test (AASHTO PP 60-14), with the exception that the
specimens are trimmed to a height of 130 mm tall instead of the standard 150 mm tall. Specimens
for the high-modulus mixtures were prepared to a target air void content of 3.5 ± 0.5% after
trimming. A minimum of four specimens were tested per unique mixture. Guidance in AASHTO
TP 107-14 was used to select strain levels for testing that provide a range of cycles to failure (Nf).
To conduct this test, an AMPT sample is glued with a steel epoxy to two end platens. The sample
and end platens are then attached with screws to the actuator and reaction frame of the AMPT
prior to installing on-specimen LVDTs. A photo of this test setup is shown in Figure 23.
41
An AMPT sample
glued with steel
epoxy to rigidly
mounted end platens.

Figure 23 IPC Global® AMPT S-VECD Fatigue Test Setup
The recommended temperature for the cyclic fatigue test is the average temperature of the high
and low PG grade of the base binder, minus three degrees Celsius. The maximum allowable
temperature according to AASHTO TP 107-14 is 21°C. The maximum allowable temperature of
21°C was necessary in order to test the high-modulus mixtures from this study. The reason for
the maximum temperature is to avoid viscoplastic effects during the test. This results in a simpler
model because strain decomposition is not needed
The fatigue test is performed at a frequency of 10 Hz and consists of two phases. First, a small
strain (50 to 75 on-specimen microstrain) test is performed to determine the fingerprint dynamic
modulus of the sample. This is conducted to determine the ratio of the finger-print dynamic
modulus (|E*|FP) of the testing sample to the dynamic modulus determined from AMPT dynamic
modulus testing (|E*|LVE). This value is known as the dynamic modulus ratio (DMR) and is
recommended to be between 0.9 and 1.1 (Equation 12) (Hou et al., 2010). This ratio is used for
controlling the quality of the fatigue testing and is incorporated into the S-VECD fatigue model
(Hou et al., 2010).
Secondly, the sample is subjected to a fatigue test in which the AMPT actuator is programmed to
reach a constant peak actuator displacement with each loading cycle. During this test, the
dynamic modulus and phase angle of the sample are recorded. Failure of the sample is defined
as the point at which the phase angle peaks and then drops off (Hou et al., 2010). This concept is
demonstrated graphically in Figure 24.
𝐄∗ 𝐅𝐏
𝐃𝐌𝐑 = (12)
𝐄∗ 𝐋𝐕𝐄
42

Figure 24 Determination of Cycles to Failure for S-VECD Fatigue Test
S-VECD analysis was performed using the ALPHA-Fatigue (v 3.1.5) analysis software developed
by Underwood and Kim. The ALPHA-Fatigue software produces two outputs from the input
dynamic modulus and fatigue tests: the damage characteristic curve and the energy-based failure
criterion. The damage characteristic curve (or C vs. S curve) plots the pseudo secant modulus (C)
of the mixture against its damage parameter (S). Practically, this illustrates how fatigue damage
evolves in a unique asphalt mixture (Jacques et al., 2016). For this study, this model was
generated using an exponential function (Equation 13).
𝐛
𝐂 = 𝐞𝐚𝐒 (13)
where
C = pseudo secant modulus,

S = damage parameter, and
a, b = model coefficients.
The second output from the ALPHA-Fatigue software is energy-based failure criterion, or GR
method (Sabouri and Kim, 2014). The GR term is defined as the rate of change of the averaged
released pseudostrain energy (per cycle) throughout the test (Sabouri and Kim, 2014). The GR
term characterizes the overall rate of damage accumulation through fatigue testing (Jacques et
al., 2016). A plot of GR versus cycles to failure (Nf) can be generated from the ALPHA-Fatigue
analysis, and the slope and position of these curves can be used to gage the relative fatigue
resistance of one mixture to another (Jacques et al., 2016).
43
A summary of the results from the individual S-VECD tests is included in Table 18 below. Of the
24 individual specimens tested, 16 specimens had a DMR value in the recommended range (0.9
to 1.1). The remaining eight specimens had borderline DMR values of between 0.84 and 0.90,
indicating a maximum 16% disconnect between the mixture E* tested during dynamic modulus
and E* verified during the fatigue testing. These results were not excluded for two reasons. First,
the testing was being performed on unique (high-modulus) materials at a non-standard air void
content. Secondly, these specimens were not detrimental to the quality of the GR versus Nf model
discussed hereafter.
The damage characteristic (C vs. S) curves for this project are shown in Figure 25 while the energy
release (GR vs. Nf) curves are shown in Figure 26. A power model of standard form was fit to the
GR versus Nf curves, with the model coefficients summarized in Table 19. Figure 25 shows three
of the mixtures (35% RAP PG 76-22, 50% RAP PG 76-22, and 25% RAP-5% RAS PG 76-22) to have
virtually identical damage characteristic curves, while the EME 14 and 35% RAP HiMA mixtures
have the greatest stiffness as additional damage is applied to the specimens. The energy release
curves all had power model R2 values of 0.94 or above, indicating a good model fit. The curve
with the highest slope and highest intercept was the EME 14 mixture. This indicates that at low
energy release rates (10 or 100), this mixture has poor fatigue resistance relative to the other
mixture designs. Three of the mixtures (50% RAP PG 76-22, 25% RAP-5% RAS PG 76-22, and 35%
RAP HiMA) had virtually identical slopes at the low end of the spectrum, indicating improved
fatigue resistance relative to the other two mixtures. The 35% RAP HiMA mixture had the highest
intercept of this grouping of three mixtures and is further to the right of the plot in Figure 26,
indicating it would be the most fatigue resistant mixture in this grouping.
44
Table 18 Summary of S-VECD Individual Test Results

Specimen Air Voids |E*|LVE |E*|FP
Mixture ID DMR Nf GR
ID (%) (MPa) (MPa)
EME 14 9 3.0 14,694 14,715 1.001 7,795 270.9
EME 14 11 3.8 14,643 14,102 0.963 47,682 12.9
EME 14 12 3.9 14,694 14,172 0.964 2,835 489.7
EME 14 13 3.7 14,694 14,614 0.995 73,284 6.8
35% RAP PG 76-22 326 3.2 13,954 11,904 0.853 47,685 20.0
35% RAP PG 76-22 327 4.0 13,954 13,314 0.954 1,175 2,037.9
35% RAP PG 76-22 328 4.0 13,954 11,952 0.857 11,500 217.0
35% RAP PG 76-22 329 4.0 13,899 11,992 0.863 39,060 33.1
35% RAP PG 76-22 330 4.0 13,954 11,795 0.845 6,035 282.6
50% RAP PG 76-22 221 3.5 14,939 14,374 0.962 3,275 654.0
50% RAP PG 76-22 222 3.2 14,939 15,161 1.015 3,795 333.3
50% RAP PG 76-22 223 3.3 14,882 12,484 0.839 214,224 4.7
50% RAP PG 76-22 224 3.7 14,882 13,001 0.874 40,900 34.6
50% RAP PG 76-22 225 3.2 14,939 14,021 0.939 69,991 36.4
25% RAP-5% RAS PG 76-22 128 3.1 14,815 13,458 0.908 795 2,644.8
25% RAP-5% RAS PG 76-22 131 3.8 14,815 12,501 0.844 1,195 747.2
25% RAP-5% RAS PG 76-22 132 3.9 14,764 12,532 0.849 4,195 317.5
25% RAP-5% RAS PG 76-22 134 3.9 14,764 13,282 0.900 78,330 14.0
25% RAP-5% RAS PG 76-22 135 3.4 14,764 13,233 0.896 4,075 610.7
25% RAP-5% RAS PG 76-22 136 3.8 14,764 13,303 0.901 1,655 1,385.5
35% RAP HiMA 421 3.6 13,454 12,869 0.957 55,031 33.5
35% RAP HiMA 422 3.2 13,563 13,014 0.960 1,635 1,730.3
35% RAP HiMA 423 3.4 13,454 13,788 1.025 7,295 521.0
35% RAP HiMA 425 3.5 13,398 12,694 0.947 3,255 462.4
45

Figure 25 S-VECD: C versus S Curves

R
Figure 26 S-VECD: G versus Nf Curves
46
Table 19 Summary of S-VECD GR vs. Nf Power Model Coefficients

Mixture ID α1 α2 R2
French EME -1.422 5.512E+07 0.973
35% RAP PG 76-22 -1.207 1.135E+07 0.981
50% RAP PG 76-22 -1.053 2.695E+06 0.963
25% RAP-5% RAS PG 76-22 -1.087 3.229E+06 0.951
35% RAP HiMA -1.059 3.925E+06 0.940
5. AASHTOWARE PAVEMENT ME DESIGN ANALYSIS
One of the objectives of this study was to use the AASHTOWare Pavement ME Design software
to determine how a high-modulus base can affect the performance of asphalt pavements. To
achieve this objective, the Pavement ME Design file for section S9 in the 2009 NCAT Test Track
was utilized to perform the simulation. Level 1 input was used in the Pavement ME simulations
for all layers. The measured dynamic modulus in Section 4.3 was used in the simulation of high-
modulus base layers in Pavement ME Design software. Five years of design life were used in the
simulation. Pavement construction information is shown in Table 20.
Six simulated scenarios were used to determine the effect of different high-modulus base
mixtures on the performance of asphalt pavements. Each scenario is explained as follows:
1. Simulation 1 utilized material properties from the 2009 Test Track Cycle Section S9
(control section).
2. Simulation 2 was designed to determine how a high-modulus mixture designed based on
a French mixture design procedure can affect the performance predicted by Pavement
ME Design.
3. Simulation 3 was designed to determine how a 35% RAP mixture can affect the
performance predicted by Pavement ME Design. The binder grade was PG 76-22 and the
mixture was labeled 35% RAP PG 76-22 No Lime.
4. Simulation 4 was planned to determine how a 25% RAP + 5% RAS mixture can affect the
performance predicted by Pavement ME Design. The binder grade was also PG 76-22 and
the mixture was labeled 50% Aged Binder PG 76-22 No Lime.
5. Simulation 5 was planned to determine how a 35% RAP with high polymer-modified
asphalt binder (HiMA) mixture can affect the performance predicted by Pavement ME
Design. The mixture utilized SBS from Kraton with No Lime.
6. Simulation 6 was designed to determine how a 50% RAP mixture can affect the
performance predicted by Pavement ME Design. The binder grade was PG 76-22 and the
mixture was labeled 50% RAP PG 76-22 No Lime.
47
Table 20 Simulation Plan

Simulation Simulation Simulation Simulation Simulation Simulation

1 2 3 4 5 6
Control French 25% RAP, 35% RAP
Name 35% RAP 50% RAP
base EME 5% RAS HiMA
Surface AC: 1.2 in 9.5 mm PG 76-22
Binder AC: 2.8 in 19 mm PG 76-22
50% Aged
Unmodified 35% RAP 35% RAP 50% RAP
French Binder PG
Base AC: 3.0 in Mixture PG PG 76-22 Kraton No PG 76-22
EME 14 76-22 No
67-22 No Lime Lime No Lime
Lime
Base Binder PG PG 67-22 PG 88-16 PG 76-22 PG 76-22 PG 94-28 PG 76-22
Granular base: 5.8 in Crushed stone granular base
Subgrade Test Track subgrade
5.1 Traffic
The truck fleet at the NCAT Test Track runs at a target speed of 45 mph, and operates 16 hours
daily, six days a week for each two-year cycle. Each of the trucks completes about 680 miles per
day so as to apply 10 million ESALs collectively in two years. Thanks to simple truck patterns and
running schedules, input Level 1 for traffic information was precisely characterized for the
MEPDG analysis. Traffic information is displayed in Table 21. Trafficking at the 2009 NCAT Test
Track was conducted using four triple flat-bed trailer trucks (Figure 27) and one triple box trailer
loaded the pavement from Monday to Saturday. Table 22 provides the axle weights for each of
the five trucks under normal loading conditions.

Figure 27 Triple Flat-Bed Trailer Truck at NCAT Test Track
48
Table 21 Traffic Information

Age (year) Heavy Trucks (cumulative)
2009 (initial) 3,082
2011 (2 years) 2,814,250
2014 (5 years) 5,628,500
Table 22 Axle Weights (lbs) for Trucking Fleet at NCAT Test Track
Front Drive Rear Drive Single Single Single Single Single
Truck # Steer
Tandem Tandem # 1 # 2 # 3 # 4 # 5
1 9,400 20,850 20,200 20,500 20,850 20,950 21,000 20,200
2 11,200 20,100 19,700 20,650 20,800 20,650 20,750 21,250
3 11,300 20,500 19,900 20,500 20,500 21,000 20,650 21,100
4 11,550 21,200 19,300 21,000 21,050 21,000 20,750 20,800
5 11,450 20,900 19,400 20,100 20,450 21,000 20,050 20,650
Average 11,450 20,900 19,400 20,100 20,450 21,000 20,050 20,650
In order to represent a triple trailer, two fictitious vehicle classes were used together with five
single axles and one tandem axle from the Class 13, and the remaining one single axle from the
Class 12. The average axle width was 8.5 ft, the dual tire spacing was 13.5 in, and the tire pressure
was approximately 100 psi. Other traffic inputs (i.e., lateral traffic wander) were assumed to be
routine design values, and they were left as the defaults provided by the MEPDG. There was no
annual traffic growth.
5.2 Climate
The climatic data required in the MEPDG is used by the Enhanced Integrated Climate Model
(EICM) to calculate changes in the temperature and moisture profile throughout the pavement
cross section. The climatic input for the MEPDG is actually a file that contains a recorded history
of temperature, rainfall, wind speed, humidity, and sunlight conditions for a specific area. There
are two ways to prepare the climatic inputs for the MEPDG, either by selecting a climatic data file
for representative areas or by preparing a new climatic data file based on a local weather station.
The latter was adopted in this study because the Test Track has an on-site weather station (Figure
28), which is responsible for collecting environmental information on an hourly basis. The Test
Track is at a geographic coordinate of 32°59´N, -85°30´W, and an elevation of 600 ft. The next
section will cover the method to prepare a climate file for a particular condition.
49

Figure 28 Test Track On-Site Weather Station
It is noted that two formats of files function in the MEPDG: the ICM file and the hourly climatic
database file. The ICM file was generated by the MEPDG calculation based on an hourly climatic
database file. In fact, the hourly climatic database file was either given for those representative
areas or can be self-developed.
AASHTOWare Pavement ME Design software was used to compute the following distresses to
simulate the pavement performance over a five-year period of analysis:
• International Roughness Index (IRI),

• Top-down cracking,
• Bottom-up fatigue cracking,
• Thermal cracking,
• Total pavement rutting, and
• Rutting in the asphalt concrete layer.
5.3 Estimated Performance
Figure 29 shows the estimated layer moduli for all six scenarios. As expected, all of the high-
modulus base courses exhibit higher layer moduli throughout the five-year performance period
of analysis. The results indicate that high-modulus base courses can have higher layer moduli
ranging from 1.5 to 2.0 times the layer modulus of the control section. However, no significant
differences among the high-modulus mixtures can be observed. Further analysis indicates that
the 35% RAP HiMA mixture exhibits lower layer moduli, especially at high temperatures,
compared to the remaining high-modulus mixtures. This behavior is in agreement with the
dynamic modulus test results for the high temperature low frequency range. However, the 25%
RAP-5% RAS base course does not exhibit higher modulus at the low temperature high frequency
as expected from the E* test results.
50
4000 Control French EME 35% RAP

25% RAP 5% RAS 35% RAP HiMA 50% RAP
3500
3000
AC MODULUS, KSI
2500
2000
1500
1000
500
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
PAVEMENT AGE, YEARS
Figure 29 Estimated Base AC Layer Modulus
Cracking of the asphalt layer was expected to decrease by the use of high-modulus base courses.
Figures 30 and 31 show estimated results of cracking. It was observed that the use of high-
modulus base courses could reduce the bottom-up cracking, which was reasonable since high-
modulus base may reduce the tensile stress and strain at the bottom of the binder layer. This
reduction in cracking could range from 20% to 25%. Moreover, it can be observed that the effect
on top-down cracking can be more significant with a decrease in cracking ranging from 28% to
35%. In this case, the 35% RAP HiMA mixture seems to be the least resistant to fatigue cracking
of all HMAC mixtures and the EME mixture seems to show the best performance, contrary to the
results obtained with the S-VECD test results. However, the observed trends in fatigue cracking
performance for the five high-modulus designs can be considered similar for practical purposes.
45 Control French EME 35% RAP

40 25% RAP 5% RAS 35% RAP HiMA 50% RAP
35
BOTTOM-UP CRACKING, %
30
25
20
15
10
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
PAVEMENT AGE, YEARS
Figure 30 Estimated Bottom-Up Fatigue Cracking
51
180
Control French EME 35% RAP
160 25% RAP 5% RAS 35% RAP HiMA 50% RAP
140
TOP-BOTTOM DAMAGE, %
120
100
80
60
40
20
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
PAVEMENT AGE, YEARS
Figure 31 Estimated Top-Down Fatigue Damage
Figure 32 shows the estimated results of permanent deformation of the asphalt concrete layer.
It was observed that using high-modulus base layer materials would have no significant effect on
the rutting of the entire asphalt concrete section. This can be explained by the small observed
differences among high-modulus base courses as well as the use of the same top and
intermediate AC layers in the simulations. A reduction in less than 4% of the AC layer rut depth
was estimated. When comparing the simulated permanent deformation and the results from the
flow number testing, no correlation in the results was evident.
0.45
0.40
0.35
0.30
AC RUT DEPTH, IN
0.25
0.20
0.15
0.10
0.05 Control French EME 35% RAP

0.00
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
PAVEMENT AGE, YEARS
Figure 32 Estimated Permanent Deformation of the Entire AC Layer
Figure 33 shows the predicted performance of the six pavement structures in terms of IRI. It was
observed that the use of a high-modulus base can reduce the IRI from 5.6% to 6.7% relative to
52
the control section, and no significant difference in IRI was found in the simulations among high-
modulus base courses.
140
130
120
IRI, IN/MILE
110
100
90
Control French EME 35% RAP
80
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
PAVEMENT AGE, YEARS
Figure 33 Estimated Surface Roughness
Finally, Table 23 contains a summary of thermal cracking performance and percent change with
respect to the control section. As shown in this table, high-modulus base courses may have no
effect on the thermal cracking. This is expected since all the simulations had the same surface
and binder layers whose properties mainly affect the development of thermal cracking.
53
Table 23 Summary of Estimated Performance

Predicted
Distress Type Target French 35% 25% RAP 35% RAP 50%
Control
EME RAP 5% RAS HiMA RAP
Terminal IRI
172 134.49 125.51 126.56 126.37 127.02 126.13
(in/mile)
Permanent deformation:
0.75 0.57 0.55 0.55 0.55 0.55 0.55
total pavement (in)
AC bottom-up fatigue cracking
25 38.21 28.61 29.81 29.61 30.31 29.31
(% lane area)
AC thermal cracking
0 27.17 27.17 27.17 27.17 27.17 27.17
(ft/mile)
AC top-down fatigue cracking
2000 10304 6695 7202 7099 7419 7006
(ft/mile)
0.25 0.46 0.44 0.44 0.44 0.44 0.44
AC only (in)
Percent Reduction vs. Control
Terminal IRI
6.7% 5.9% 6.0% 5.6% 6.2%
(in/mile)
3.5% 3.5% 3.5% 3.5% 3.5%
total pavement (in)
AC bottom-up fatigue cracking
25.1% 22.0% 22.5% 20.7% 23.3%
(% lane area)

AC thermal cracking
0.0% 0.0% 0.0% 0.0% 0.0%
(ft/mile)
AC top-down fatigue cracking
35.0% 30.1% 31.1% 28.0% 32.0%
(ft/mile)
3.6% 3.6% 3.6% 3.6% 3.6%
AC only (in)
Figure 34 measured distresses on section S9 at the NCAT Test Track from 2009 to 2014. It can be
observed that performance was highly overestimated for section S9. The maximum measured
rut depth was 0.34 inches of the entire structure, while the estimated maximum rut depth was
0.56 inches. Measured IRI started at 60 in/mile and did not change significantly over time with a
final IRI of 80 in/mile. On the contrary, the initial estimated IRI was 94 in/mile, which was
expected to increase over time to reach a final IRI of 135 in/mile. Finally, it can be observed that
cracking showed the largest offset. Initial measured cracking was observed after 11 million ESALs
or more than two years of truck trafficking and reaching only 10%, while cracking is expected to
appear during the first year and significantly increase over time. These results can be explained
due to the application of nationally calibrated transfer functions or default functions in the
employed software. Therefore, not only is local calibration required for section S9, but calibration
of the HMAC mixtures used in this study may be needed to further reflect their benefit as base
courses.
54
a. 0.4
0.3
Rut Depth, in
0.2
0.1
0
0 5 10 15 20
ESALs, Millions

b. 100
80
IRI, in/mile
60
40
20
0
0 5 10 15 20
ESALs, Millions

c. 12
10
Cracking, % lane
8
6
4
2
0
0 5 10 15 20
ESALs, Millions

Figure 34 Measured Pavement Performance on Section S9: a. Rutting, b. Roughness, c.
Cracking
In summary, the use of high-modulus base courses could improve the overall performance of an
asphalt concrete layer and the entire flexible pavement structure. The type of distress that may
be the most affected is top-bottom cracking followed by bottom-up fatigue cracking. Ride quality
can also be slightly improved (lower IRI), and no significant effect on rutting and thermal cracking
should be expected.
In Pavement ME, the performance prediction uses transfer functions that are calibrated with the
existing performance data in which stiffer layers correspond to better pavements. Figure 29
showed lower modulus for the 35% RAP HiMA mixture, and as calibrated, Pavement ME
predicted least resistant to fatigue cracking (higher cracking). However, the uniaxial fatigue test
showed that the 35% RAP HiMA mixture is the most fatigue resistant, even though the layer
moduli is lower.
55
With the empirical nature of the Pavement ME transfer functions, performance of new materials
cannot be reliably modelled with the current transfer functions. A high modulus brittle material
will have different fatigue behavior than a high modulus ductile material. Therefore, the current
transfer functions in Pavement ME should be calibrated with laboratory and field performance
of HMAC mixtures to have reliable predictions. The better Pavement ME predicted performance
of the new material is likely a combination of its higher modulus and also an artifact of not so
applicable transfer function. Until the latter is resolved, quantifying the field performance of the
former through Pavement ME would be somewhat inconclusive.
6. CONCLUSIONS AND RECOMMENDATIONS
The objective of this project was to develop and validate mixture designs and evaluate predicted
performance effects of high-modulus base layers. Based on experimental results and structural
analysis, the following conclusions and recommendations are made:
• European mix design standard methods and specifications were successfully

implemented on local (U.S.) materials. The Level 3 requirement for dynamic modulus,
14,000 MPa at 15°C and 10 Hz, was met for all HMAC mixtures.
• Reclaimed asphalt pavement can be used to stiffen the asphalt binder sufficiently for
high-modulus asphalt mixtures. The minimum RAP content utilized in this investigation
was 35%.
• No significant differences in dynamic modulus were obtained for all HMAC mixtures at
low temperature and high frequency. However, on the opposite side of the
temperature/frequency spectrum, the 25% RAP, 5% RAS mixture provided significantly
higher E* values due to the inclusion of RAS.
• Flow number test results were significantly greater than those of the conventional
mixture.
• Based on AMPT cyclic test results, fatigue properties among HMAC mixtures seemed to
improve for the high polymer-modified mixtures and seemed to decrease for the EME
French mixture, which has a stiffer virgin binder.
• Increased stiffness of HMAC mixtures improves MEPDG predicted performance of
pavement in rutting, fatigue cracking, and ride quality compared to conventional base
courses. A great improvement in fatigue cracking (top down and bottom up) and some
improvement in rutting and ride quality.
• Correlations between laboratory performance trends among HMAC mixtures and
predicted structural performance were not obtained. This was attributed to the
significant role that other materials/properties have on pavement responses and
performance. For instance, lower modulus for 35% RAP HiMA mixture and as calibrated
Pavement ME predicted higher cracking. However, the uniaxial fatigue test showed that
the 35% RAP HiMA mixture is the most fatigue resistance even though the layer moduli is
lower.
• Calibration of MEPDG transfer functions applicable to HMAC mixtures is recommended
to obtain more representative performance predictions.
56
• Based on the results of this study, a detail cost and benefit assessment is recommended
in order to further quantify the effect the HMAC mixtures have on potential long-lasting
perpetual-type flexible pavements.
• This study was based on a laboratory experiment. The ability to produce HMAC mixtures
through a hot mix plant and to successfully lay and compact them in the U.S. has not been
demonstrated. Mixing and compacting in the laboratory suggest that field operations
would be more difficult with HMAC mixtures; the EME’s design and testing has to be
adapted to U.S. standards and conditions.
7. RECOMMENDED MIXTURE DESIGN PROCEDURE
Based on the results of this research study and the current state-of-the-practice, the following
steps are recommended as HMAC mixture design procedure:
1. Determine the aggregate trial blend for the HMAC mixture.

2. Determine the minimum asphalt binder content using the French method based on the
aggregate surface area (Equations 1 to 4). However, the Asphalt Institute Hveem-Edward
equation can be used successfully here.
3. Set Ndes with the Superpave gyratory compactor to 80 gyrations and compact design
samples to target air voids lower than 6%.
4. Prepare three trial dynamic modulus samples compacted to 3.0 - 6.0% air voids according
to the French methodology and test at 15°C and 10 Hz.
5. Select optimum binder content to meet E* = 14,000 MPa (at 15°C and 10 Hz) to meet the
minimum asphalt content from step 2 and to meet Ndes specimens target air voids lower
than 6%.
a. Adjusting the gradation or mixture components (additives, recycled material,
binder grade, etc.) may be necessary to meet E* and air voids requirements.
b. For each gradation adjustment, the minimum AC required will need to be
recalculated.
6. Select laboratory performance tests and criteria (rutting, cracking, and moisture damage)
for further verification and conduct AASHTO TP 79-15 to determine dynamic modulus to
be used in ME simulations.
ME simulations should be used for relative comparison purposes and not for structural pavement
design until field validation has been performed. Pilot projects are a proven tool for validating
and fine-tuning new practices resulting from research. Using traditional projects as a benchmark,
pilot projects or programs have been used extensively to measure the relative success of new
specifications and test methods. The results of pilot projects have served to effectively promote
the long-term implementation of new industry practices. It is recommended that an agency
champion the use of the proposed standard methodologies for design, analysis, construction,
and specifications related to HMAC impact on pavement performance.
57
REFERENCES
Airey, G. D. Use of Black Space Diagrams to Identify Inconsistencies in Rheological Data. Road
Materials and Pavement Design, Vol. 3, No. 4, 2002, pp. 403-424.
Anderson, R. M., G. N. King, D. I. Hanson, and P. B. Blankenship. Evaluation of the Relationship

between Asphalt Binder Properties and Non-Load Related Cracking. Journal of the Association of
Asphalt Paving Technologists, Vol. 80, 2011, pp. 615-664.
Bankowski, W., M. Tusar, L. G. Wiman, D. Sybilski, M. Gajewski, R. Horodecka, M. Maliszewksi,

and K. Mirski. Laboratory and Field Implementation of High Modulus Asphalt Concrete.
Requirements for HMAC Mix Design and Pavement Design. European Commission Research,
Paris, 2009.
Brosseaud, Y. Small Overview of High Modulus Asphalt Mixes EME. Presented at Australian Scan
Tour, June 2012.
Brosseaud, Y., F. Farcas, and V. Mouillet. High Modulus Asphalt Mixes with High Rate of RA: What
Does it Happen? In Congrès Eurobitume Eurasphalt, EE2012, Turkey, October 2012.
Bueche, N., A. G. Dumont, A. Vanelstraete, J. de Visscher, S. Vansteenkiste, F. Vervaecke, J. Maek,

L. Gaspar, and F. Thogersen. Laboratory and ALT-Evaluation of High Stiffness Underlayers with
High Percentage of Re-use as Developed in the NR2C-Project. Proceedings of the 4th Eurasphalt
and Eurobitume Congress, Copenhagen, Denmark, 2008.
Capitãio, S. D., and L. Picado-Santos. Assessing Permanent Deformation Resistance of High

Modulus Asphalt Mixtures. Journal of Transportation Engineering, Vol. 132, No. 5, 2006, pp. 394-
401.
Carbonneau, X., J. P. Michaut, T. Anderson, C. Thorup, and L. Ladenhoff. High Modulus GAB II: A
Danish Experiment. Proceedings of the 4th Eurasphalt and Eurobitume Congress, Copenhagen,
Denmark, 2008.
Chappat, M., X. Carbonneau, and Y. Lefuevre. Analysis of the Measured Performance of High
Modulus Asphalt (HMA) and Road Base Asphalt (RBA) and a Proposal for a Compensation
Principle. European Roads Review, 14, 2009, pp. 72-83.
Christensen Jr., D. W., T. Pellinen, and R. F. Bonaquist. Hirsch Model for Estimating the Modulus
of Asphalt Concrete. Journal of the Association of Asphalt Paving Technologists, Vol. 72, 2003, pp.
97 – 121.
Corté, J. Development and uses of Hard-Grade Asphalt and of High-Modulus Asphalt Mixes in
France. Transportation Research Circular, 503, 2001, pp. 12-31.
58
Daniel, J. S., and Y. R. Kim. Development of a Simplified Fatigue Test and Analysis Procedure using
a Viscoelastic Continuum Damage Model. Journal of the Association of Asphalt Paving
Technologists, Vol. 71, 2002, pp. 619–650.
DeBacker, C., L. Glorie, and R. Reynaert. Test sections in High-Modulus Asphalt: A Comparative
Experiment with Ten Variants. Proceedings of the 4th Eurasphalt and Eurobitume Congress,
Copenhagen, Denmark, 2008.
DeBacker, C., J. De Visscher, L Glorie, A. Vanelstraete, S. Vansteenkiste, and L. Heleven. A

Comparative High-Modulus Asphalt Experiment in Belgium. In Proceedings of Transport Research
Arena Europe, 2008, pp. 21-24.
De La Roche, C., H. Odeon, J. Simoncellli, and A. Spernol. Study of the Fatigue of Asphalt Mixes
Using the Circular Test Track of the Laboratoire Central des Ponts et Chaussées in Nantes, France.
Transportation Research Record 1436, TRB, National Research Council, Washington D.C., 1994,
pp. 17-27.
De Visscher, J., S. Vansteenkiste, and A. Vanelstraete. Test Sections in High-Modulus Asphalt: Mix
Design and Laboratory Performance Tests. Proceedings of the 4th Eurasphalt and Eurobitume
Congress, Copenhagen, Denmark, 2008.
Delorme, J. L., C. De La Roche, and L. Wendling. LPC Bituminous Mixtures Design Guide.
Laboratoire Central des Ponts et Chaussées, Paris, 2007.
Denneman, E., M. Nkgapele, J. Anochie-Boateng, and J. W. Maina. Transfer of High Modulus

Asphalt Mix Technology to South Africa. In Proceedings of the 10th Conference on Asphalt
Pavements for Southern Africa, Champagne Sports Resort, Central Drakensberg, South Africa,
2011.
Denneman, E. High Modulus Asphalt (HiMA) Trial: Mix and Pavement Design. Presented at the
22nd Meeting of the Roads Pavements Forum (RPF), Pretoria, November 2011.
Denneman, E., and M. Nkgapele. Interim Guide for the Design of High Modulus Asphalt Mixes and
Pavements in South Africa. Report CSIR/BE/IE/ER/2010/0042/B, 2011.
Des Criox, P., and L. Planque. Experience with Optimized Hard Grade Bitumens in High Modulus
Asphalt Mixes. In Proceedings of the 3rd Eurasphalt and Eurobitume Congress, Vienna, Austria,
2004.
Diefenderfer, B. K. and G. W. Maupin. Field Trials of High-Modulus High-Binder-Content Base

Layer Hot-Mix Asphalt Mixtures. Report No. FHWA/VTRC 11-R2. Virginia Transportation Research
Council, Charlottesville, Va., 2010.
Distin, T., L. Sampson, H. Marais, and B. Verhaeghe. High Modulus Asphalt: Assessment of
Viability Based on Outcomes of Overseas Fact Finding Mission. White Paper, 2006.
59
European Asphalt Pavement Association. Airfield uses of Asphalt. Netherlands, 2003.
European Asphalt Pavement Association. High Modulus Asphalt: A State of the Art Report.
Netherlands, 2005.
Guyot, X. Use of High Modulus Asphalt - “EME” Case Studies in the Indian Ocean Area. 15th AAPA
International Flexible Pavements Conference, Brisbane, Australia, 2013.
Haritonovs, V., J. Tihonovs, M. Zaumanis, and A. Krasnikovos. High Modulus Asphalt Concrete
with Dolomite Aggregate. Transport Research Arena. Paris, France. 2014.
Hou, T., B. S. Underwood, and Y. R. Kim. Fatigue Performance Prediction of North Carolina
Mixtures Using the Simplified Viscoelastic Continuum Damage Model. Journal of the Association
of Asphalt Paving Technologists, Vol. 79, 2010, pp. 35-80.
Jacques, C., J. S. Daniel, T. Bennert, G. Reinke, A. Norouzi, C. Ericson, W. Mogawer, and Y. R. Kim.
Effect of Silo Storage Time on the Characteristics of Virgin and Reclaimed Asphalt Pavement
Mixtures. Transportation Research Record: Journal of the Transportation Research Board, No.
2573, Washington, D.C., 2016, pp. 76-85.
Jamois, D., J. C. Vaniscote, Y. Jolivet, and M. Malot. Development of a Concept of Very High
Modulus Bituminous Macadam for Pavement Base Courses. Proceedings of the 2nd Eurasphalt
and Eurobitume Congress, Vol. 2, Barcelona, Spain, 2000.
Lee, H. J., J. H. Lee, and H. M. Park. Performance Evaluation of High Modulus Asphalt Mixtures
for Long Life Asphalt Pavements. Construction and Building Materials, Vol. 21, No. 5, 2007, pp.
1079-1087.
Lenfant, M. High Rutting Resistant Asphalt Pavement. Seminar for Thailand Department of
Highways, TIPCO Asphalt Public Company Limited, Bangkok, Thailand, 269-289.
Michaut, J. P. EME or High Modulus Asphalt Concrete. COLAS Presentation. March 2014.
Miro, R., G. Valdés, A. Martinez, P. Segura, and C. Rodriguez. Evaluation of High Modulus Mixture
Behavior with High Reclaimed Asphalt Pavement (RAP) Percentages for Sustainable Road
Construction. Construction and Building Materials, Vol. 25, No. 10, 2011, pp. 3854-3862.
Montanelli, E. F. Fiber/Polymeric Compound for High Modulus Polymer Modified Asphalt (PMA).
Procedia-Social and Behavioral Sciences, Vol. 104, 2013, pp. 39-48.
Newcomb, D. and K. Hansen. Mix Type Selection for Perpetual Pavements. Presented at the
Perpetual Pavement Conference, Columbus, Ohio, 2004.
Newcomb, D. E., D. H. Timm, and R. Willis. Perpetual Asphalt Pavements: A Synthesis. Asphalt
Pavement Alliance, Lanham, Md., 2010.
60
Nicholls, J. C., R. Elliott, N. Meite, R. Perera, A. Hunter, and J. Williams. Monitoring the
Introduction of Enrobe a Module Eleve Class 2 onto UK Roads. In Proceedings of the 4th
Eurasphalt and Eurobitume Congress, Copenhagen, Denmark, 2008.
Nkgapele, M., E. Denneman, and J. K. Anochie-Boateng. Construction of a High Modulus Asphalt

(HiMA) Trial Section Ethekwini: South Africa’s First Practical Experience with Design,
Manufacturing and Paving of HiMA. Presented at 31st Southern African Transport Conference,
Pretoria, South Africa, 2012.
Petho, L., and E. Denneman. High Modulus Asphalt Mix (EME) for Heavy Duty Applications and
Preliminary Laboratory Test Results in Australia. Presented at 15th AAPA International Flexible
Pavements Conference, Brisbane, Australia, 2013.
Rahbar-Rastegar, R., and J. S. Daniel. Laboratory versus Plant Production: Impact of Marerial
Properties and Performance for RAP and RAS mixtures. International Journal of Pavement
Engineering, 2016, pp. 1-12.
Sabouri, M., and Y. R. Kim. Development of a Failure Criterion for Asphalt Mixtures Under
Different Modes of Fatigue Loading. In Transportation Research Record: Journal of the
Transportation Research Board, No. 2447, Transportation Research Board of the National
Academies, 2014, pp. 117-125.
Sanders, P. J., and M. Nunn. The Application of Enrobe a Module Eleve in Flexible Pavements. TRL
Report TRL636, Transport Research Laboratory, Crowthorne, United Kingdom, 2005.
Underwood B., C. Baek, and Y. Kim. Simplified Viscoelastic Continuum Damage Model as Platform
for Asphalt Concrete Fatigue Analysis. In Transportation Research Record: Journal of the
Transportation Research Board, No. 2296, Transportation Research Board of the National
Academies, 2012, pp. 36-45.
Vaitkus, A., and V. Vorobjovas. Use of Local Aggregate in High Modulus Asphalt Concrete Layers.
Gradevinar, Vol. 65, No. 4, 2013, pp. 353-360.
Wei, O., F. Xinghua, and W. Lianguang. Research on Anti-Rutting Performance of High Modulus
Asphalt Concrete Pavement. Journal of Highway and Transportation Research and Development,
Vol. 4, No. 2, 2010, pp. 77-79.
Wielenski, J. C., and G. A. Huber. Evaluation of French High Modulus Asphalt (EME) in Pavement
Structural Design (MEPDG). Journal of the Association of Asphalt Paving Technologists, Vol 80,
2011, pp 697-718.
61

Rep17 04 PDF

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Rep17 04 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Rep17 04 PDF

Uploaded by

Copyright:

Available Formats

NCAT

A comprehensive literature study was performed to assess the current state-of-the-practice on

3.1 Mixture Design

3.1.1 Aggregate Selection

Table 1 HMAC Aggregate Selection Criteria (Denneman et al., 2011)

3.1.2 Designing a Gradation

3.1.3 Binder Selection and Richness Factor

Table 4 Typical Hard Asphalt Characteristics (Before Aging) (Corte, 2001)

𝟏𝟎𝟎𝚺 = 𝟎. 𝟐𝟓𝑮 + 𝟐. 𝟑𝑺 + 𝟏𝟐𝒔 + 𝟏𝟓𝟎𝒇 (1)

G = proportion of aggregate retained on and above the 6.3 mm sieve;

4. Calculate the percent binder by mass of total mixture using Equation 4.

3.1.4 Performance Tests

Table 7 French Performance Tests (Modified from Denneman et al., 2011)

Table 8 Summary of Gyratory Compaction Study (Denneman et al., 2011)

In Europe, dynamic modulus is measured using standard EN 12697-26 (equivalent to AASHTO TP

3.2 Pavement Design

Table 10 Structural Build-Ups (Information from Barkowski et al., 2007)

3.5 Summary of Current Practice

A comprehensive literature study was performed to assess the current state-of-the-practice on

Pavement design and analysis of HMAC mixtures is conducted using mechanistic-empirical

Table 11 Roadbase High-Modulus Asphalt Concrete versus Traditional AC (EAPA, 2005)

4. EXPERIMENTAL DESIGN AND ANALYSIS

4.1 Laboratory Testing

• Volumetric mixture design and material characterization,

4.2 Mixture Design

Table 12 Aggregate Gradations of Mixtures

Table 13 Mixture Design Properties

4.3 Dynamic Modulus

𝑽𝑴𝑨 𝑽𝑭𝑨∗𝑽𝑴𝑨 𝟏;𝑷𝒄

𝟒𝟑𝟓,𝟎𝟎𝟎 𝑽𝑭𝑨 𝟎.𝟓𝟖

Table 14 Mastercurve Equation Variable Descriptions

French EME 35%RAP PG 76-22

Table 15 Mastercurve Coefficients

Figure 21 Black Space Diagrams

Table 16 E* Statistical Grouping

4.4 Flow Number

εp (N) = permanent strain at ‘N’ cycles,

Table 17 Flow Number Test Results

4.5 AMPT Cyclic Fatigue

C = pseudo secant modulus,

Table 18 Summary of S-VECD Individual Test Results

Table 19 Summary of S-VECD GR vs. Nf Power Model Coefficients

5. AASHTOWARE PAVEMENT ME DESIGN ANALYSIS

Table 20 Simulation Plan

Table 21 Traffic Information

• International Roughness Index (IRI),

5.3 Estimated Performance

4000 Control French EME 35% RAP

45 Control French EME 35% RAP

0.05 Control French EME 35% RAP

Table 23 Summary of Estimated Performance

6. CONCLUSIONS AND RECOMMENDATIONS

• European mix design standard methods and specifications were successfully

7. RECOMMENDED MIXTURE DESIGN PROCEDURE

1. Determine the aggregate trial blend for the HMAC mixture.

Anderson, R. M., G. N. King, D. I. Hanson, and P. B. Blankenship. Evaluation of the Relationship

Bankowski, W., M. Tusar, L. G. Wiman, D. Sybilski, M. Gajewski, R. Horodecka, M. Maliszewksi,