Available online at www.sciencedirect.

com

Construction
and Building

MATERIALS

Construction and Building Materials 22 (2008) 13231330

www.elsevier.com/locate/conbuildmat

Laboratory evaluation of fatigue characteristics of recycled

asphalt mixture
Xiang Shu, Baoshan Huang *, Dragon Vukosavljevic
Department of Civil and Environmental Engineering, The University of Tennessee, Knoxville, TN 37996, USA
Received 6 February 2007; received in revised form 12 April 2007; accepted 12 April 2007
Available online 19 June 2007

Abstract
This paper presents the results of a laboratory study of evaluating the fatigue characteristics of hot-mix asphalt (HMA) mixtures using
dierent testing methods. In this study, the fatigue performance of HMA mixtures was evaluated with the Superpave indirect tension
(IDT) tests and beam fatigue test. The HMA mixtures containing 0%, 10%, 20%, and 30% of recycled asphalt pavement (RAP) were
plant prepared with one source of aggregate, limestone, and one type of binder, PG 6422. The fatigue properties tested included indirect
tensile strength (ITS), failure strain, toughness index (TI), resilient modulus, DCSEf, energy ratio, plateau value, and load cycles to failure. The results from this study indicated that both Superpave IDT and beam fatigue tests agreed with each other in ranking the fatigue
resistance of mixtures when proper procedures were followed.
 2007 Elsevier Ltd. All rights reserved.
Keywords: Hot-mix asphalt; Fatigue; Dissipated energy; Indirect tensile strength

1. Introduction
Fatigue cracking is one of the three major distresses
(fatigue cracking, low temperature cracking, and rutting)
of exible pavements. Fatigue cracking is mainly caused
by repeated trac loading and it can lead to signicant
reduction in the serviceability of exible pavements. The
cracking resistance of hot-mix asphalt (HMA) mixtures is
directly related to the fatigue performance of exible pavements. Therefore, the laboratory characterization of the
fatigue behavior of HMA mixtures has been a topic of
intensive study for many years.
Many laboratory testing methods are available to characterize the fatigue behavior of HMA mixtures. Probably
the one that possesses the most similar stress condition to
HMA eld mixtures under trac loading is the repeated
exural test (also called beam fatigue test) [1]. This test
was developed under SHRP-A-003A to evaluate the fatigue
*

Corresponding author. Tel.: +1 865 974 7713.

E-mail address: bhuang@utk.edu (B. Huang).

0950-0618/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2007.04.019

response of HMA mixtures and to summarize what is

known about the factors that inuence pavement life using
a third point loading. The exural beam fatigue test was
later modied in SHRP-A-404 to improve its simplicity
and reliability.
This test uses a digitally controlled, pneumatic beam
fatigue equipment, which subjects a beam specimen under
repeated stress or strain controlled loading, which is
applied at the center of the beam until failure occurs. The
failure of the exural fatigue test can be dened as a 50%
reduction in initial stiness, which is measured from the
center point of the beam after 50th load cycle [1].
Recently, a new way to determine the failure of the exural fatigue test was proposed by Carpenter et al. based on
the dissipated energy [24]. In this new method, the ratio of
dissipated energy change (RDEC) is dened as a ratio of
the change in dissipated energy between two neighboring
cycles divided by the dissipated energy of the rst cycle.
A plateau value (PV), or the nearly constant value of
RDEC, can be determined and it represents a period where
there is a constant percent of input energy being turned

1324

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

into damage. This PV can be used to characterize the fatigue life of HMA mixtures. For a strain-controlled test, the
lower the PV, the longer the fatigue life for a specic HMA
mixture [4].
Since the 1970s, fracture mechanics theory has been
used to analyze the fatigue behavior in HMA mixtures
[5]. In recent years, comprehensive laboratory and eld
studies were conducted by Roque et al. at the University
of Florida to characterize the crack growth rate of HMA
mixtures using the Superpave indirect tension (IDT) tests
[68]. They used the three Superpave IDT tests (IDT
strength test, resilient modulus test, and creep test) and
developed a viscoelastic fracture mechanics-based crack
growth law for HMA mixtures. In addition, they introduced two thresholds, the dissipated creep strain energy
(DCSE) limit and the fracture energy (FE) limit, to account
for the crack development and propagation in HMA mixtures. When these two thresholds are not exceeded, only
healable micro-damage occurs. Non-healable macro-damage appears unless one of the thresholds is exceeded. This
suggests that the higher the values of DCSE or FE, the
longer the fatigue life of HMA mixtures [7].
In addition to the traditional fatigue approach and fracture mechanics approach, damage mechanics is also
applied to HMA mixtures to characterize their fatigue
behavior. Kim et al. developed a fatigue model for HMA
mixtures using the elasticviscoelastic correspondence principle and continuum damage mechanics [9,10]. This model
has been successfully used to predict the fatigue life of
HMA mixtures with multiple rest periods based on the
materials viscoelastic properties, loading conditions, and
damage and micro-damage healing characteristics.
The objective of this study was to evaluate and compare
the fatigue performance of HMA mixtures based on the
results of dierent laboratory fatigue testing. In this study,
HMA mixtures were plant prepared with one source of
coarse aggregate (limestone), four percentages of recycled
asphalt pavement (RAP) (0%, 10%, 20%, and 30%), and
one asphalt binder (PG 6422). The fatigue properties of
HMA mixtures were evaluated using the Superpave IDT
tests and beam fatigue test.
2. Laboratory experiments

Table 1
Asphalt binder properties
Binder status

Binder test

Original binder

Rotational viscosity at
0.52
135 C, Pa s
70 C 0.78
DSR, G*/sin d,
kPa
64 C 1.63

RTFO aged
binder

Specication
3 Pa s max
1.00 kPa min

70 C 1.66
64 C 3.54

PAV aged binder DSR, G*sin d MPa, 25 C

BBR creep stiness S,
MPa
BBR creep slope, m value
PG grading

2.20 kPa min

3725
238

5000 kPa max

300.0 MPa
max
0.300 min

0.310
6422

Table 2
Properties of aggregates
Sieve
size

Limestone Drock

No. 10
screening

Natural
sand

Manufactured
sand

5/800
1/200
3/800
#4
#8
#30
#50
#100
#200

100%
97%
70%
21%
7%
4%
3%
2.0%
1.8%

100%
100%
100%
92%
61%
29%
21%
20.0%
16.0%

100%
100%
100%
98%
93%
63%
13%
2.0%
1.0%

100%
100%
100%
99%
82%
28%
17%
9.0%
5.0%

Gsb

2.524

2.424

2.501

2.476

Note: Gsb bulk specic gravity of aggregate.

that was comparable to the ne aggregate. RAP gradation

was determined on the bare aggregate after the binder was
extracted from RAP, as shown in Fig. 1. The asphalt content of RAP was 5.5% and its maximum specic gravity
(Gmm) was 2.412.
2.2. Mixture design
The Marshall mix design procedure was employed to
design the control mixture. For the HMA mixtures, 50%
limestone D-rock, 15% No. 10 screenings, 25% natural

2.1. Materials

100
Percent Passing (%)

One type of asphalt binder, PG 6422, was chosen in the

study. Its properties are presented in Table 1.
The coarse aggregates selected in this study were
crushed limestone with a nominal maximum size of
12.5 mm. The ne aggregates consisted of No. 10 screenings, natural sand, and manufactured sand. Their gradations and other properties are presented in Table 2. All
the aggregate properties meet the specication of the Tennessee Department of Transportation (TDOT) [11].
The RAP used in this study was screened through the
No. 4 sieve (4.75 mm) to acquire a consistent gradation

DSR, G*/sin d,
kPa

Test
results

80
60
Aggregate

40

RAP
20

Design Envelope
Maximum Density Line

0
0

2
3
Sieve Size (0.45 Power)

Fig. 1. Aggregate and RAP gradations.

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

1325

sand, and 10% manufactured sand were selected. The

aggregate gradation in the experiment is shown in Fig. 1.
The optimum asphalt content was 5.0%, namely asphalt
from RAP together with the virgin asphalt was 5.0% by
the weight of the mix. Since mixtures contained dierent
contents of RAP, the asphalt contribution from RAP and
virgin asphalt was dierent for each mixture, as presented
in Table 3. Mixture volumetric properties are listed in
Table 4.

2.4.1.1. Resilient modulus test. This test was performed on

the cylindrical samples by applying a repeated peak-load
resulting in horizontal deformations within the range of
200300 microstrains. Each load cycle consists of 0.1-s load
application followed by a 0.9-s rest period. The load and
deformation were continuously recorded and the resilient
modulus can be calculated as follows [12,13]:

2.3. Sample preparation

where,

In the mixing drum, virgin aggregate was superheated

and used to heat RAP added in the middle of the mixing
process before both were mixed with liquid asphalt. The
plant-prepared HMA mixtures were collected at the mixing
plant and then taken to the University of Tennessee laboratory for specimen preparation and fatigue testing. Cylindrical 150-mm samples were compacted with the Superpave
gyratory compactor (SGC) and then sliced into about
50 mm thick specimens for the Superpave IDT tests. The
air voids for the Superpave IDT tests were 4 0.5%.
2.4. Mixture performance testing
2.4.1. Superpave IDT Tests
The Superpave IDT tests include the resilient modulus,
creep, and indirect tensile strength tests and they were performed following the procedures developed by Roque and
Buttlar [12,13]. Fig. 2 shows the test setup of the Superpave
IDT tests. The testing system and associated analysis procedures are described in detail by Roque and Buttlar
[12,13]. It should be pointed out that the gage length was
not exactly one quarter of the specimen diameter as suggested by Roque and Buttlar [12,13]. However, it is still
reasonable to use the equations developed by them to compare the relative fatigue performance of HMA mixtures in
this study. The tests were performed at 25 C compared to
the 10 C temperature used by Roque et al. [14].

MR

P  GL
DH  t  D  C cmpl

MR = resilient modulus;
P = maximum load;
GL = gage length;
DH = horizontal deformation;
t = thickness of specimen;
D = diameter of specimen;
Ccmpl = nondimensional creep compliance factor,
Ccmpl = 0.6354(X/Y)1  0.332;
(X/Y) = ratio of horizontal to vertical deformation.

2.4.1.2. Creep test. The creep compliance test was

performed on the same specimen used for the resilient modulus test. After allowing the specimen to re-stabilize (5
10 min) the creep compliance test was performed. During
this test the specimen was loaded with a constant load
for 1000 s. The constant load was chosen such that it
produced a horizontal deformation within the range of
200750 microstrains after 1000 s of loading. The creep
compliance is calculated as follows [12,13]:
Dt

DH  t  D  C cmpl
P  GL

Mixture

Asphalt content
from RAP (%)

Asphalt content from

virgin asphalt (%)

0 % RAP added
10 % RAP added
20 % RAP added
30 % RAP added

0
0.55
1.11
1.66

5.5
4.45
3.89
3.34

AC (%)

Gmm

Gmb

Air voids

VMA

Stability (kN)

Flow (mm)

5.0

2.456

2.356

4.0

16

11.6

2.77

Note: AC, asphalt cement content; VMA, voids in mineral aggregate;

Gmm, maximum specic gravity of mixture; Gmb, bulk specic gravity of
compacted mixture.

where, D0, D1 and m = parameters obtained from the creep

test.
With these two parameters, D1 and m, Roque et al. proposed to use the term, DCSEmin, the minimum dissipated
creep strain energy to characterize the cracking performance of HMA mixtures. DCSEmin is expressed as follows:
DCSEmin

Table 4
Volumetric properties of mixture

where, D(t) = creep compliance at time t;

P, GL, DH, t, D, and Ccmpl are the same as described
above.
The creep compliance D(t) can be represented using the
following power function [7]:
Dt D0 D1 tm

Table 3
Asphalt contribution from RAP and virgin asphalt for the mixtures

m2:98  D1
A

The parameter A is a function of tensile strength and tensile

stress in the asphalt pavement as follows:
A 0:0299r3:10
6:36  S t 2:46  108
t

where, rt = applied tensile stress of asphalt layer, St = indirect tensile strength.

1326

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

Fig. 2. The Superpave IDT test setup.

2.4.1.3. Indirect tensile strength test. The IDT strength test

was used to determine tensile strength and strain of the
mixture specimens compacted to 4 1% air voids. Cylindrical specimens with 152.4 mm in diameter and 50.8 mm
in thickness were monotonically loaded to failure along
the vertical diametric axis at the constant rate of
76.2 mm/min. The indirect tensile strength can be calculated as follows:
St

2  P  C sx
ptD

where,
St = indirect tensile strength;
P = failure load;
Csx = horizontal stress correction factor;
Csx = 0.948  0.01114 (t/D)  0.2693 m + 1.436 (t/
D) m;
m = Poissons ratio, m = 0.1 + 1.480 (X/Y)2  0.778
(t/D)2 (X/Y)2; and
t, D, and (X/Y) are the same as described above.

Ap = area under the normalized stressstrain curve up

to strain ep;
e = strain at the point of interest; and
ep = strain corresponding to the peak stress.
This toughness index compares the performance of a
specimen with that of an elastic perfectly plastic reference
material, for which the TI remains a constant of 1. For
an ideal brittle material with no post-peak load carrying
capacity, the value of TI equals zero. In this study, the values of indirect tensile toughness index were calculated up to
tensile strain of 10%.
With the stressstrain response from the IDT strength
test, the dissipated creep strain energy threshold (DCSEf)
was determined by Roque et al. as follows (Fig. 4) [7]:
DCSEf FE  EE

where, FE = fracture energy, it is dened as the area under

the stressstrain curve to the failure strain ef, and
EE = elastic energy.
1.2

TI

Ae  Ap
e  ep

ITS Normalized

Toughness index (TI), a parameter describing the toughening characteristics in the post-peak region, was also calculated from the indirect tensile test results. Fig. 3 presents
an example of normalized indirect tensile stress and strain
curve. A dimensionless indirect tensile toughness index, TI
is dened as follows:

0.8
0.6
0.4

Ap

0.2

where,
0

TI = toughness index;
Ae = area under the normalized stressstrain curve up to
strain e;

Strain
Fig. 3. An example of normalized IDT curve for TI calculation.

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

ER

St

2.4.2. Beam fatigue test

The exural beam fatigue test was a strain-controlled
test to determine the fatigue life of 38.1 cm long by
5.08 cm thick and by 6.35 cm wide beam specimens sawed
from laboratory compacted samples subjected to repeated
exural bending until failure (AASHTO T321-03).
Beam specimens were compacted using the vibratory
compactor to 7 1% air voids and tested at 25 C according to AASHTO T321-03. Specimens were placed in a
beam fatigue xture (Fig. 5) that would allow four-point
bending with free rotation and horizontal translation at
all load and reaction points using an MTS closed loop
computer controlled data acquisition system.
A strain level of approximately 600 microstrains and a
loading frequency of 10 Hz were used such that the specimen will undergo a minimum of 10,000 load cycles. During
each load cycle beam deections were measured at the center of the beam to calculate maximum tensile stress, maximum tensile strain, phase angle, stiness, dissipated energy,

MR
DCSEf

EE
0

Fig. 4. Determination of creep strain energy threshold (DCSEf).

FE

DCSEf
DCSEmin

1327

ef

Se de

1
EE S t ef  e0
2

10

where, e0 can be found in Fig. 4.

With DCSEf and DCSEmin, energy ratio (ER) was
dened as follows [13]:

Fig. 5. Beam fatigue xture.

11

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

Flexural Stiffness (lbf/in^2)

1328

5.000E+05
4.000E+05
3.000E+05
2.000E+05
1.000E+05
0.000E+00
100

50,100

100,100

150,100

200,100

250,100

Loading Cycles
Fig. 6. Flexural stiness versus loading cycles.

RDEC

II

III

Plateau Value (PV)

IDT strength (MPa)

a 1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0

Load cycles

0%

30%

IDT failure strain (%)

b 2.5
2.0
1.5
1.0
0.5
0.0
0%

10%

20%

30%

RAP content

1.0
0.8

IDT TI

3. Discussion of test results

20%

RAP content

Fig. 7. Typical RDEC plot with three behavior zones (after [2]).

and cumulative dissipated energy. Fig. 6 represents a typical stiness versus load cycle plot using an automated fatigue software.
For the beam fatigue test, fatigue life is traditionally
dened as the number of cycles corresponding to a 50%
reduction in initial stiness and initial stiness was measured at the 50th load cycle (AASHTO T321-03). Recently,
Carpenter et al. proposed to use RDEC to determine the
fatigue life [24]. Fig. 7 presents a typical RDEC plot. As
seen from Fig. 7, the curve can be divided into three dierent zones. RDEC value decreases with the load cycle in
zone 1. RDEC value is approximately constant in zone 2,
representing a period where there is a constant percent of
input energy turned into damage. In zone 3, RDEC value
increases with the load cycle, indicating that more and
more input energy are turned into damage and ultimately
the mixture loses the load carrying capability.

10%

0.6
0.4

3.1. Superpave IDT test results

0.2

Fig. 8 presents the strengths, failure strains, and toughness index (TI) values from the IDT strength test. It can be
seen that mixtures containing higher percentages of RAP
exhibited higher indirect tensile strength (ITS), lower strain
at peak-load, and lower toughness index than the control

0.0

0%

10%

20%

30%

RAP content
Fig. 8. Results from IDT strength test. (a) IDT strength, (b) IDT failure
strain, and (c) IDT TI.

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

sonable to characterize the cracking resistance of HMA

mixtures than DCSEf because it takes into account both
the energy required to fracture HMA mixtures and the dissipated energy accumulation in HMA mixtures under certain loading condition. The energy ratios from this study
were usually lower compared to typical values obtained
by Roque et al. [13]. The reason is that a higher tempera0.30
0.25

Energy Ratio

mixture (0% RAP mixture). These phenomena can be

attributed to the aged, stiened, and embrittled asphalt
binder in RAP due to the aging process. The test results
showed that incorporation of RAP increased the strength
of HMA mixtures. However, due to the increase in the brittleness (decreased failure strain) and the decrease in TI, the
fatigue life of HMA mixtures may still be compromised.
These test results are consistent with those from laboratory-prepared HMA mixtures [15].
Fig. 9 presents the resilient modulus test results. As seen
from Fig. 9, the resilient modulus of HMA mixtures
increased with the increase in the percentage of RAP, which
means that incorporation of RAP increased the elastic component of the viscoelastic material HMA mixtures. On the
other hand, HMA mixtures containing RAP became more
brittle (Fig. 8b). The overall eect of RAP was disadvantageous to the fatigue performance of HMA mixtures.
Fig. 10 presents the test results of the dissipated creep
strain energy threshold (DCSEf). It can be seen that with
the incorporation of RAP, DCSEf values became lower,
which means that the energy required to fracture HMA
mixtures decreased as RAP percentage increased. This
clearly indicated that the fatigue behavior of HMA mixture
was compromised by the incorporation of HMA mixtures.
Fig. 11 presents the energy ratios calculated from the
Superpave IDT tests. The energy ratio concept is more rea-

1329

0.20
0.15
0.10
0.05
0.00
0%

10%

20%

30%

RAP content
Fig. 11. Energy ratios from Superpave IDT tests.

MR (GPa)

0
0%

10%

20%

3.00

Plateau Value ( 10-5)

2.50

30%

2.00
1.50
1.00
0.50
0.00

RAP content

0%

10%
20%
RAP content

0%

10%

DCSEf (kJ/m3)

20.0

15.0

10.0

5.0

140000

Load cycles to failure

Fig. 9. Resilient modulus results.

120000

30%

100000
80000
60000
40000
20000
0

0.0
0%

10%

20%

30%

20%

30%

RAP content

RAP content
Fig. 10. DCSEf results.

Fig. 12. Results from beam fatigue test. (a) Plateau value and (b) load
cycles to failure.

1330

X. Shu et al. / Construction and Building Materials 22 (2008) 13231330

ture (25 C) was used in this study compared to the 10 C

temperature used by Roque et al. [13]. The energy ratio values from this study are presented in Fig. 11. It can be seen
that the energy ratio of HMA mixtures decreased with the
incorporation of more RAP, which means that HMA mixtures containing higher content of RAP were more likely to
fracture than control mixture.
3.2. Beam fatigue test results
Fig. 12 presents the plateau values and load cycles to
failure based on the 50% reduction in stiness. It can be
seen that HMA mixtures incorporating more RAP exhibited higher plateau values (Fig. 12a), which indicated that
they experienced more damage and would result in shorter
fatigue life. However, from Fig. 12b mixtures containing
higher percentages of RAP appeared to experience longer
fatigue life. Although the load cycle results were consistent
with the authors pervious study, it is still doubtful that
incorporation of RAP will increase the fatigue life of
HMA mixtures due to the controversy about the test
method and the failure criterion. The plateau value method
seems more reasonable in evaluating the fatigue life of
HMA mixtures.
4. Summary and conclusions
A laboratory study was conducted to evaluate the fatigue characteristics of HMA mixtures containing RAP
using dierent testing methods. HMA mixtures were plant
prepared with one source of aggregate limestone and one
type of binder PG 6422, containing 0%, 10%, 20%, and
30% of RAP. The fatigue properties tested included indirect tensile strength (ITS), failure strain, toughness index
(TI), resilient modulus, DCSEf, energy ratio, plateau value,
and load cycles to failure. Based on the results from the
study, the following conclusions can be summarized:
 The inclusions of RAP into HMA mixtures in this study
generally increased the tensile strength and reduced the
post-failure tenacity in indirect tensile strength test.
 The inclusions of RAP also generally decreased the dissipated creep strain energy threshold and energy ratio,
which may result in the short fatigue life of HMA
mixtures.
 Based on the failure criterion of 50% reduction in stiness, incorporation of RAP increased the fatigue life
of HMA mixtures. Whereas using the plateau values
from the beam fatigue test, inclusion of RAP would
make more input energy turn into damage, which may
result in the shorter fatigue life.
 The plateau value failure criterion appeared more reasonable in evaluating the fatigue performance of HMA
mixtures in this study.
 The energy ratio concept seemed eective in the evaluation of fatigue cracking behavior of HMA mixtures in
this study.

 The results presented in this paper were only the preliminary ndings of a more complete study. Further studies would be needed before the relevant testing methods
can be recommended to evaluate the fatigue performance of HMA mixtures.
Acknowledgements
The authors would like to acknowledge the nancial
support from the Tennessee Department of Transportation. Thanks are also due to Mr. William Gibbons and
Mr. Chun-Yip Chan who helped prepare testing specimens.
References
[1] Roberts FL, Kandhal PS, Brown ER, Lee D-Y, Kennedy TW. Hot
mix asphalt materials, mixture design, and construction. Lanham,
Maryland: NAPA Education Foundation; 1996.
[2] Carpenter SH, Ghuzlan K, Shen S. Fatigue endurance limit for
highway and airport pavements. Transportation research record, vol.
1832. Washington DC: National Research Council; 2003. p. 131138.
[3] Ghuzlan K, Carpenter SH. Energy-derived/damage-based failure
criteria for fatigue testing. Transportation research record, vol. 1723.
Washington DC: National Research Council; 2000. p. 141149.
[4] Shen S, Carpenter SH. Application of dissipated energy concept in
fatigue endurance limit testing. Transportation research record, vol.
1929. Washington DC: National Research Council; 2005. p. 165173.
[5] Majidzadeh K, Kaumann EM, Saraf CL. Analysis of fatigue of
paving mixtures from the fracture mechanics viewpoint. ASTM Spec
Tech Publ 1971;508:6784.
[6] Roque R, Zhang Z, Sankar B. Determination of crack growth rate
parameters of asphalt mixtures using the Superpave IDT. J Assoc
Asphalt Paving Technol 1999;68:40433.
[7] Roque R, Birgisson B, Sangpetngam B, Zhang Z. Hot mix asphalt
fracture mechanics: a fundamental crack growth law for asphalt
mixtures. J Assoc Asphalt Paving Technol 2002;71:81627.
[8] Zhang Z, Roque R, Birgisson B, Sangpetngam B. Identication and
verication of a suitable crack growth law. J Assoc Asphalt Paving
Technol 2001;70:20641.
[9] Kim YR, Little DL, Benson F. Chemical and mechanical evaluation
on healing mechanism of asphalt concrete. J Assoc Asphalt Paving
Technol 1990;59:24076.
[10] Kim YR, Lee H-J, Little DL. Fatigue characterization of asphalt
concrete using viscoelasticity and continuum damage theory. J Assoc
Asphalt Paving Technol 1997;66:52069.
[11] TDOT. Standard specication for road and bridge construction, the
Tennessee Department of Transportation, Nashville, TN, March,
1995.
[12] Roque R, Buttlar WG. The development of a measurement and
analysis system to accurately determine asphalt concrete properties
using the indirect tensile mode. J Assoc Asphalt Paving Technol
1992;61:30432.
[13] Buttlar WG, Roque R. Experimental development and evaluation of
the new SHRP measurement and analysis system for indirect tensile
testing at low temperature. Transportation research record, vol. 1454.
Washington DC: National Research Council; 1994. p. 163171.
[14] Roque R, Birgisson B, Drakos C, Dietrich B. Development and eld
evaluation of energy-based criteria for top-down cracking performance of hot mix asphalt. J Assoc Asphalt Paving Technol
2004;73:22960.
[15] Huang B, Kingery III WR, Zhang Z. Laboratory study of fatigue
characteristics of HMA mixtures containing RAP. In: International
symposium on design and construction of long lasting asphalt
pavements, Auburn, 2004. p. 50122.