Construction and Building Materials 101 (2015) 918–925

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Glass fiber reinforced asphalt membrane for interlayer bonding between

asphalt overlay and concrete pavement
Zhesheng Ge a, Hao Wang b,⇑, Qingshan Zhang a, Chunlong Xiong a
a
School of Civil Engineering and Transportation, South China University of Technology, Wushan Road No. 381, Guangzhou 510640, PR China
b
Rutgers, The State University of New Jersey, New Brunswick, NJ, USA

h i g h l i g h t s

Evaluate interface shear and fatigue behavior of asphalt concrete overlay on concrete pavement.

Conduct laboratory performance testing using emulsified asphalt and asphalt membranes modified with glass fiber.

Determine the effect of interlayer types on shear toughness and fatigue cracking.

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this study is to evaluate the shear and fatigue behavior of asphalt concrete (AC) overlay on
Received 28 February 2015 Portland cement concrete (PCC) pavement with different interlayer bonding materials, including
Received in revised form 4 October 2015 modified emulsified asphalt (MEA) membranes, modified asphalt (MA) membranes, glass fiber reinforced
Accepted 19 October 2015
modified emulsified asphalt (GFMEA) membranes and glass fiber reinforced modified asphalt (GFMA)
membranes. An in-house shearing device was used to evaluate the shear behavior between asphalt layer
and PCC layer. An orthogonal experimental design was performed to determine the impact of each
Keywords:
influential factor on the interlayer shear behavior, including the testing temperature, the bonding
PCC pavements
AC overlay
material, and the confinement pressure applied during the tests. The results indicate that AC overlay
Bonding on PCC pavement can reach its optimal performance in terms of shear toughness when the GFMA is used
Reflection cracking at the interlayer. Four-point bending fatigue tests were conducted on the AC beam with interlayers using
GFMA membranes the Material Test System (MTS). The fractal dimension (FD) of fatigue cracks in AC overlay was analyzed
using image measurements. The results show that the AC beam with the GFMA has a longer fatigue life.
Therefore, it is concluded that the GFMA can strengthen bonding and mitigate reflection cracking of AC
overlay on PCC pavements.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction layer thickness. In practice, most of the delamination distresses

can be attributed to either improper construction techniques or
A thin bonded asphalt concrete (AC) overlay is an economical inadequate choice of tack coat [1].
way of pavement rehabiliation to restore the riding quality of dete- Slippage is a visible indicator of inadequate bonding between
riorated Portland cement concrete (PCC) pavement. However, the AC overlay and PCC pavement where high horizontal stresses
most common problems occurring with rehabilitation are debond- occur. This distress develops in areas where braking, accelerating,
ing of the overlay and reflection of cracking through the overlay. or turning wheels move and deform the surface layer of the hot
At locations where poor interface bonding or debonding occurs, mix asphalt (HMA) structure [2,3]. Slippage can be seen on the
premature failure such as fatigue cracking will quickly appear that pavement surface in the form of half-moon-shaped cracks with
is initiated at the bottom of surface layer. The following conditions the ends pointing toward the traffic direction.
may contribute to the occurrence of delaminations: improper Reflective cracking is the main distress in flexible overlays on
consideration of temperature and field conditions, excessive load composite pavements. Reflective cracking can be defined as cracks
repetitions and dynamic impact loading, and a very thin surface in the flexible overlay that are directly above joints or above
existing cracks in the underlying concrete pavement. Reflective
⇑ Corresponding author. cracking is a premature distress because it usually occurs within
E-mail address: hwang.cee@rutgers.edu (H. Wang). only a few years after overlay construction. These cracks accelerate

http://dx.doi.org/10.1016/j.conbuildmat.2015.10.145
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925 919

Table 1 The basic mechanism of reflection cracking is the propagation of

Properties of SBS modified emulsified asphalt. cracks through the flexible overlay due to horizontal and vertical
Experiment project Testing result movements in the vicinity of cracks and joints in the underlying
The remain of the 1.18 sieve/％ 0.1 concrete pavement. These movements are caused by vehicular
Bitumen standard viscosity C25, 3/s 43 loading, temperature and moisture variations, or the combination
Evaporation residue Content/% 60 of both. It has been found the bending of asphalt overly under
Penetration degree 65 vehicular loading is the main driving mechanism to induce
(100 g, 25 °C,5 s)/dmm
Softening point/°C 53
reflective cracking [4]. Previous researchers have found that the
Ductility(5 °C)/cm 20 interface bonding condition affects significantly the initiation and
Solubility (C2HCl3)/% 97.5 propagation of reflective cracking [5,6]
In general, an interlayer system is a supplementary layer to con-
trol reflective cracking. The interlayer system is very thin and is not
Table 2
considered in an overlay design. Typical interlayer systems in this
Properties of SBS modified asphalt. category are geogrid, geocomposite, interlayer stress absorbing
composite (ISAC), strain absorbing membrane interlayer (SAMI),
Experiment project Testing result
and steel netting [7,8]. A mixture-type interlayer has also been
Penetration degree (25 °C, 100 g, 5 s)/0.1 mm 54 used as either a supplementary layer or an alternative layer of
Penetration index +0.05
Ductility (5 °C, 5 cm/min)/cm 33
the leveling binder of HMA overlay. This mixture type usually con-
Softening point TR & B/°C 80.0 sists of smaller, sand-size aggregates and modified rich binder to
Flash point/°C 323 have more crack resistance than normal HMA overlays [9].
Solubility/% 98.5 To prevent reflection cracking in AC overlays, an interlayer
Elastic recovery (25 °C)/% 91.0
stress-absorbing composite (ISAC) was developed and evaluated
Tests after rolling thin Quality changes/% 0.01
film oven test(RTFOT) Ductility 25 for its effectiveness against reflection cracking [10]. A laboratory
(5 °C, 5 cm/min)/cm pavement section with an AC overlay over a jointed PCC slab was
Penetration ratio/% 80.0 constructed and placed in an environmental chamber. The results
indicate that the ISAC layer is highly effective in preventing reflec-
tion cracking in a 2.5-in. AC overlay. When compared to a control
test section and a section using commercially available reflection
Table 3
cracking controlling material, the ISAC layer provided superior
Passing rate of aggregate in each sieve hole.
performance.
Sieve size (square hole sieve)/mm Passing rate/% In another field study, seven test sections (six alternatives and
26.5 100.0 one control section) were constructed to compare pavement
19 96.6 performance under mixed traffic and environmental conditions.
16 87.2 It was found that little reflection cracking occurred on the interlay-
13.2 76.0
9.5 59.4
ered sections during the first 3-year observation period [11]. How-
4.75 32.9 ever, limited studies have been conducted to study the effect of
2.36 23.5 interlayer system on interface bonding and reflective cracking
1.18 18.0 together using laboratory performance testing.
0.6 12.4
0.3 7.8
0.15 6.0 2. Objectives
0.075 4.6
The objective of this paper aims to evaluate shear and fatigue
behavior of AC overlay on PCC pavement with different interlayer
pavement deterioration because it allows moisture to penetrate bonding materials using laboratory performance testing. The shear
the flexible overlay and thus causes further damage to the interface testing evaluates the interface bonding behavior; while fatigue
and possibly the underlying pavement layers. testing evaluates the potential of reflective cracking in the

shearing force
locking nut
bottom guide rod
plate

Top plate

bonding layer 1 structural layer 3

50mm

p ro lo a d in g fo rc e p ro lo a d in g fo rc e
bonding layer 2

structural layer 1 structural layer 2

shearing baseboard

Fig. 1. Side-view of shear testing device.

920 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925

(a) PCC slab (b) Spreading modified emulsion asphalt

(c) Spreading glass fiber (d) Spreading modified asphalt

e Spreading macadam
Fig. 2. Procedure of preparing sample with GFMA.

composite pavement. The interlayer bonding materials considered 3.1.3. Glass fiber
include the modified emulsified asphalt (MEA), the modified As a non-metallic material, fiber has broad application fields due to its excellent
performances, such as good insulation, heat resistance, and corrosion resistance.
asphalt (MA), the glass fiber reinforced modified emulsified asphalt The main components of fiber are inorganic and metal oxides, such as silica,
(GFMEA) and the glass fiber reinforced modified asphalt (GFMA) alumina, calcium oxide, and boron oxide. Alkali content determines different types
membranes. of fiber, such as non-alkali fiber, alkali fiber, and high-alkali fiber. The general
properties of glass fiber were listed as follows, the melting point is 680 °C, the
boiling point is 1000 °C, and the density ranges from 2.4 g/cm3 to 2.7 g/cm3.
3. Materials and test methods
3.1.4. Hot mix asphalt (HMA)
3.1. Materials The asphalt pavement layer was prepared with an AC-20 type asphalt mixture.
The aggregate gradation was shown in Table 3. The mass ratio of asphalt to
3.1.1. Modified emulsified asphalt (MEA) aggregate was 4.4%.
The mechanical properties of SBS modified emulsified asphalt were listed in
Table 1. 3.1.5. Portland cement concrete (PCC) slab
The Portland cement concrete has design strength of 30 MPa. The main compo-
nents of PCC were Portland cement, granite gravel, medium sand, and water with a
3.1.2. Modified asphalt (MA) mass ratio of 0.38:1:1.1:2.72. The bulk densities of Portland cement, granite gravel,
The mechanical properties of SBS modified asphalt were listed in Table 2. medium sand and water were 461, 1252, 512, and 175 kg/m3, respectively.
 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925 921

Table 4 composed of two PCC slabs butted with epoxy. Four types of interlayer membranes
Orthogonal experimental design factors. were used in the bonding layers between HMA and PCC. The composite pavement
samples were prepared with roller compactor and then cut into 50  50  50 mm
Interlayer Testing temperature/°C Confinement pressure/MPa cubes using a high-precision double-sided saw tooth.
MEA + Macadam 20 1.0 The procedure of sample preparation with the MEA was spreading the half
40 0.7 dosage of MEA with an area density of 0.8 kg/m2 first, and then spreading the
20 0.7 remainder with the same area density until the primary half was demulsified.
40 1.0 The final step was applying fine aggregates of 4.75–9.5 mm sizes on the MEA.
MA + Macadam 20 1.0 The procedure of sample preparation with the MA was spreading modified
40 0.7 emulsified asphalt with an area density of 0.3 kg/m2 first, and then spreading the
20 0.7 MA with an area density of 1.0 kg/m2. The final step was applying fine aggregates
40 1.0 of 4.75–9.5 mm sizes on the MA.
GFMEA + Macadam 20 1.0 The procedure of sample preparation with GFEMA was spreading half dosage of
40 0.7 MEA with an area density of 0.8 kg/m2 and glass fiber (before the MEA demulsified),
20 0.7 and then spreading the remainder with the same area density. The final step was
40 1.0 applying fine aggregates of 4.75–9.5 mm sizes on the GFEMA.
GFMA + Macadam 20 1.0 The procedure of sample preparation with GFMA was spreading modified emul-
40 0.7 sified asphalt with an area density of 0.3 kg/m2 first (Fig. 2b), then spreading glass
20 0.7 fiber evenly right before the modified emulsified asphalt was demulsified (Fig. 2c).
40 1.0 The following step was spreading the MA with an area density of 1.0 kg/m2 (Fig. 2 d)
and applying fine aggregates of 4.75–9.5 mm sizes on the GFMA (Fig. 2e).

3.2.3. Test procedure

3.2. Testing method
The testing temperatures were set at 20 °C and 40 °C. The sample temperatures
were achieved by submerging samples into a water bath at the target temperature
3.2.1. Test equipment
for one hour. To maintain the testing temperature of the samples, shear tests were
An electro-hydraulic static and dynamic, servo closed-loop control, multi-
conducted right after the assembled CCP samples were taken out of the water bath.
purpose, universal test machine was used. It provides accurate loading and
The preloading pressure levels used in the shear testing were 1.0 MPa and 0.7 MPa,
unloading program control in accordance with a scheduled load pattern using the
respectively, at 20 °C and 40 °C.
following specifications:
The assembled CCP samples were fitted to the top and bottom plates of the
preloading device using the servo-hydraulic tester. The locking nuts were then

host load capacity:±100 kN;
adjusted to transfer the preloading pressure to the holding rods until the servo-

hydraulic action stroke:±150 mm;
hydraulic pressure reading reduced to zero. The assembled CCP samples were held

load sensor precision: 0.5%: and
in the preloading condition for five minutes to eliminate creep deformation and

ambient temperature range: 129 °C to +540 °C.
stress relaxation [12].
Two CCP samples were placed symmetrically over the adjustable gap in the
An in-house shear testing device used in the study is shown in Fig. 1. The sym-
shearing device baseboard, with the center of PCC slabs coinciding with the center
metrical shear testing has been used to study the interfacial shear behavior
of the baseboard gap. The width of the shear baseboard gap was adjusted to lie
between the cubic composite pavement (CCP) samples [12]. The loading is applied
beyond the area of bonding material penetrating into the interlayer membranes
through the preloading force and the shearing force. The configuration of testing
and the PCC slabs [12]. The shear loading rate on assembled CCP samples was
specimen was composed of square-shape (100  100  10 mm) bottom and top
500 mm/min.
plates with four guide rods (8.5-mm diameter) and locking nuts with a center
distance of 80 mm between any two guide rods. The testing specimen is placed
on a shearing baseboard with an adjustable gap below the specimen. 3.2.4. Orthogonal experimental design
Adopting the orthogonal experimental design would reduce testing time and
simplify statistical data analysis because representative testing results subject to
3.2.2. Sample preparation various influencing factors can be easily selected from a comprehensive orthogonal
As shown in Fig. 1, the first and third structural layers were prepared with table of testing results [13]. In this study, each influencing factor has three measure-
hot-mix asphalt concrete (HMA), while the second (middle) structure layer was ment levels for comparison. The three influential factors considered were the type

100
50

280
15

Fig. 3. Fatigue testing setup with specimen dimensions (unit: mm).

922 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925

Load direction 3.3. Fatigue testing

3.3.1. Test device

The MTS mentioned above was used in the bending fatigue test. The picture of
testing setup is shown in Fig. 3. The fatigue loading was applied using a four-point
beam setup to create the pure bending in the middle of beam.
A stingray F-504B high-speed camera with fixed focus lens was used as image
50mm

acquisition device. The camera was connected directly through the computer by
setting collection frequency and resolution based on data collection needs. The syn-
chronous triggering software was used to guarantee the camera and the testing
device was triggered in a synchronized way. The image data were continuously
taken at different fatigue loading cycles until the sample failed.

Crack Crack 3.3.2. Sample preparation

The preparation of testing specimens was similar to the ones used in the shear
(a) MEA + Macadam test. The CCP samples were cut into 300  50  65 mm cubes using a high-precision
double-sided saw. In order to improve the accuracy of identifying crack propagation
with digital speckle correlation measurement (DSCM), the specimens were manu-
Load direction ally sprayed with white paint at one side.

3.3.3. Test procedure

Bean fatigue testing can be conducted using constant stress or constant strain
load applications. In constant strain mode, the strain is maintained constant and
the stress is allowed to vary, while in constant stress mode, the load is maintained
50mm

the same and the strain is allowed to vary. For the constant stress mode, the test is
continued until the crack appears or beam actually breaks. For the constant strain
mode, failure is normally defined as the point at which the stiffness reaches 50 per-
cent of the original value [14]. In this study, the stress controlled mode was selected
to observe the crack propagation at the interface between AC layer and PCC layer.
The stress ratio of 0.5 was used to cause the crack initiation and beam failure in
Crack Crack the relatively shorter test time.
The loading frequency was 10 Hz. Loading was applied on the beam specimen
(b) MA + Macadam using the sinusoidal function, in such way that 0.1s loading after the interval of
0.9 s unloading. The testing temperature in the fatigue testing was 20 °C. The spec-
Load direction imen temperature was achieved by submerging samples into the water bath at the
target temperature for ten hours.

3.3.4. Fatigue crack observation

The maximum resolution of Stingray F-504B CCD camera was 1224  1208
pixels and its highest data collection frequency was 10 Hz. During the fatigue test,
50mm

camera images were collected and processed for observing crack initiation, propa-
gation, cross link, and the final failure of testing specimen. Based on the images of
crack propagation, microscopic structural changes of asphalt concrete at different
loading cycles were identified. The loading frequency and the data acquisition fre-
quency were both 10 Hz. The camera was fixed and the main axis of the lens was
kept perpendicularly to the specimen surface. Spot metering mode was selected
Crack and two metering spots were aligned with the observation surface of each specimen
Crack
for obtaining accurate data. The digital speckle correlation measurement (DSCM)
system and the fractal theory were used for studying fatigue crack development
(c) GFMEA + Macadam in the asphalt mixture.

Load direction
4. Results and discussion

4.1. Interlayer shear strength test

4.1.1. Failure characteristics of specimen

50mm

Different failure patterns were observed in the assembled CCP

samples after shear tests. The failed samples with different inter-
layers were shown in Fig. 4. The results show the consistent failure
pattern at the interface between asphalt layer and PCC layer.

4.1.2. Shear toughness

Crack Crack
The displacement of these assembled CCP samples under the
shearing forces of the shear test indicates the process of energy
(d) GFMA + Macadam
dissipation due to interface debonding. When the shearing force
Fig. 4. The damage of samples containing different interlayer. reached the maximum value during the shear test, the energy
dissipation indicates the shear resistance before shear failure
occurs.
A mathematical model was used for fitting a polynomial
of interlayer, the testing temperature, and the confinement pressure used during equation to the relationship between shear force and shear
the shear test, as shown in Table 4. Totally 16 testing combinations were identified
in the orthogonal experimental design from Table 4. In order to improve the accu-
displacement. An example of the relationship between shear force
racy of test data, four assembled CCP samples were used for each testing and shear displacement and the resulting polynomial fitting is
combination. shown in Fig. 5. The shear force displacement curve was fitted
 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925 923

1400

1200

1000

Shear force / N
800

600 S
400

200

0
0 1 2
x1 3 4 5 6
Displacement / mm

Fig. 5. Shear force displacement relationship for a representative sample.

The results of shear toughness for each level of influence factor

Table 5 (Ki) were listed in Table 5. The testing levels for each influential
Shear test results (shear toughness) at different influential factors and testing levels factor were consistent with the orthogonal experimental design
(unit: Nmm).
shown in Table 4. For example, the four testing levels under the
Testing levels Interlayer Temperature Confinement fact of interlay bonding material represent modified emulsified
bonding material pressure asphalt (MEA), modified asphalt (MA), glass fiber reinforced mod-
K1 1864 2335 1985 ified emulsified asphalt (GFMEA), and glass fiber reinforced modi-
K2 1834 1815 1864 fied asphalt (GFMA) membranes. The results show that the
K3 1914 – –
composite specimen with the interlayer of GFMA has the greatest
K4 1990 – –
Impact Range R 156 520 121 shear toughness among four different interlay types. The shear
toughness increases as the temperature decreases and the
confinement pressure increases.
The impact range (R) was calculated as the difference between
using the polynomial equation, as shown in Eq. (1). As expected, the shear toughness at different testing levels under the same
shear force increases as the displacement increases until the peak influential factor. The larger the impact range was, the more influ-
force is reached. ence the influential factor had. It was found that the influential fac-
tor having the maximum impact on the shear test results was the
5 3
y ¼ ax6 þ bx þ cx4 þ dx þ ex2 þ fx þ g ð1Þ temperature and the influential factor having the minimum impact
was the confinement pressure.
where, x represents the shear displacement.
The results of the orthogonal experiment design were used to
The area formed with the shear displacement and the fitted
select the optimal combination of test conditions to have the max-
curve of shear force represents the dissipation energy to cause
imum shear resistance of the specimen. The optimal sample test
shear failure. It can be although defined as the shear toughness,
conditions: (1) bonding material of GFMA and macadam, (2) con-
which is similar to the concept of fracture toughness but describes
finement pressure of 1.0 MPa, and (3) testing temperature of 20 °C.
the ability of to resist shear failure. As shown in Fig. 5, the maxi-
mum shear displacement at the maximum shear force was round
2.3 mm. Therefore, the shear toughness (S) can be calculated as 4.2. Fatigue test results
the area under the curve of shear force as the shear displacement
increases from zero the maximum shear displacement. In the fatigue test, asphalt mixtures with different interlayer
In order to evaluate the impact of each influential factor on bonding materials were tested under the stress ratio of 0.5 for fati-
shear failure potential at the interface, the shear toughness (S) gue loading. The images of fatigue cracking were obtained and the
was calculated at different influential factors and testing levels. fractal characteristic of fatigue cracks was analyzed.

1.60
GFMEA + Macadam
1.50
1.40 GFMA+Macadam
FD

1.30
1.20 MA+Macadam

1.10 MEA+Macadam
1.00
0 1000 2000 3000 4000 5000 6000 7000 8000 9000
Load cycling / times

Fig. 6. Relationship between FD and cyclic loading times.

924 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925

The fractal theory involves the study of irregular surfaces or FD of interlayer with GFMA is greater than other interlayer types.
structures having the property of invariance scale transformation, This indicates that fatigue crack propagation is slower for the
which indicates that the parts of surfaces or structures are similar composite beam with the interlayer of GFMA.
to the whole in successive range of view in all directions or at least The crack propagation patterns of damaged specimens were
one direction (self-similarity or self-affinity). A feature having a observed, as shown in Fig. 7. The results show that the beam with
fractal property can be characterized by its fractal dimension. GFMA has more cracks but they are hair-like cracks. These cracks
The irregularity of crack propagation and the roughness of crack are not distinct cracks; instead they developed in dispersed cracks
surface can be analyzed using the self-similar fractal [15]. The with the increase of cyclic loading time. Although many short
fractal dimension (FD) analysis of fatigue crack is to establish the cracks appeared at the interface region, no interface debonding
relationship between FD and fatigue crack propagation path. was observed. On the other hand, the beam with other interlayers
Box dimension is the basic methodology for fractal dimension showed larger and wider cracks at the interface along with the
measurement [16]. The side length of box was defined as n. The interface debonding.
number of boxes containing cracks was N(n). N(n) was counted
after the processed crack image was divided by the box with the
5. Conclusion
side length of n. The relationship between n and N(n) can be linked
using the power function and the parameter of FD, as shown in Eq.
The study aimed to evaluate shear and fatigue behavior of AC
(2). The value of FD reveals the curve degree and the bifurcation of
overlay on PCC pavement with different bonding materials using
fatigue crack propagation in the asphalt mixture under the cyclic
laboratory performance testing. An in-house shearing device
fatigue loading.
was used to evaluate shear strength of the composite specimen
with interfaces with an orthogonal experimental design. The
NðnÞ ¼ nFD ð2Þ
four-point bending fatigue tests were conducted using the asphalt
Four types of interlayer were used to study the fractal dimen- mixture beam strengthened with different interlayers. The
sion of fatigue cracks under fatigue loading. Images data were col- fractal dimension of fatigue cracks was analyzed using image
lected during the test and processed. The development curve of FD measurements.
with loading cycles is shown in Fig. 6. The results show that as the The testing results indicate that AC overlay on PCC pavement
loading cycles increase, FD of crack increases for all the interlayer have the optimal performances when the interface is bonded with
types. The FD increases as crack propagates, indicating the the glass fiber reinforced modified asphalt membranes (GFMA).
microstructure damage of material increases. However, the total The asphalt mixture beam with the GFMA interlayer has the grater
shear toughness and the longer fatigue life. Therefore, glass fiber
reinforced asphalt membranes can strengthen interface bonding
and have the potential to mitigate reflection cracking of AC overlay
on PCC pavements. Future work will be conducted to study
microscopic mechanism of the interaction between glass fiber
and asphalt, and the impact of fiber content.

Acknowledgment

(a) MEA + Macadam This study was supported by the Fundamental Research Funds
for the Central Universities at China.

References

[1] Applied Research Associates Inc, Composite Pavement Systems vol. 1: HMA/
PCC Pavements, Strategic Highway Research Program 2, Transportation
Research Board, 2015.
[2] H. Wang, I.L. Al-Qadi, Evaluation of surface-related pavement damage due to
(b) GFMEA + Macadam tire braking, Road Mater. Pavement Des. 11 (1) (2010) 101–122.
[3] S.A. Romanoschi, J.B. Metcalf, Characterization of asphalt concrete layer
interfaces, Transportation Research Record: Journal of the Transportation
Research Board, 1778, Transportation Research Board of the National
Academies, Washington, D.C.,, 2001, pp. 132–139.
[4] R.L. Lytton et al., Models for Predicting Reflection Cracking of Hot-Mix Asphalt
Overlays, NCHRP Report 669, Transportation Research Board, Washington, D.
C., 2010.
[5] J. Baek, H. Ozer, H. Wang, I.L. Al-Qadi, Effects of interface conditions on
reflective cracking development in hot-mix asphalt overlays, Road Mater.
Pavement Des. 11 (2) (2010) 307–334.
(c) MA + Macadam [6] H. Ozer, I.L. Al-Qadi, H. Wang, Z. Leng, Characterization of interface bonding
between hot mix asphalt and concrete pavements, Int. J. Pavement Eng. 13 (2)
(2012) 181–196.
[7] M. Elseifi, R. Bandaru, Cost effective prevention of reflective cracking of
composite pavement, FHWA/LA.11/478, Final Report, Louisiana Department of
Transportation and Development, 2011.
[8] I.L. Al-Qadi, W.G. Buttlar, J. Baek, M. Kim, Cost-effectiveness and performance
of overlay systems in illinois, volume 1, Effectiveness Assessment of HMA
Overlay Interlayer Systems Used to Retard Reflective Cracking, FHWA-ICT-09-
044, Final Report, Illinois Department of Transportation, 2009.
[9] H. Sheng, F. Zhou, T. Scullion, Reflection Cracking-Based Asphalt Overlay
Thickness Design and Analysis Tool, Transportation Research Board,
(d) GFMA + Macadam Washington, D.C., 2010.
[10] M.T. Mukhtar, B.J. Dempsey. Interlayer Stress Absorbing Composite (ISAC)
Fig. 7. Surface of failed specimen with different interlayers. for Mitigating Reflection Cracking in Asphalt Concrete Overlays. Report
 Z. Ge et al. / Construction and Building Materials 101 (2015) 918–925 925

UILU-ENG-96-2006. Department of Civil Engineering, University of Illinois at [13] R. Dong, B. Xiao, Y. Fang, The theoretical analysis of orthogonal test designs, J.
Urbana-Champaign, 1996. Anhui Inst. Archit. Ind. 12 (6) (2004) 103–106.
[11] Y.H. Cho, C. Liu, T. Dossey, B.F. McCullough, Asphalt Overlay Design Methods [14] H. Di Benedetto, Fatigue of bituminous mixtures, Mater. Struct. 37 (3) (2004)
for Rigid Pavements Considering Rutting, Reflection Cracking, and Fatigue 202–216.
Cracking. Research Report 987–9. Texas Department of Transportation, [15] B.B. Mandelbrot, Fractals: Form Chance and Dimension, W.H. Freeman and
Research and Technology Transfer Section/Construction Division, 1998. Company, San Francisco, Calusa, 1997.
[12] Z.S. Ge, Y. Wang, M. Hang, New device and methodology for evaluating the [16] H. Xie, D.J. Sanderson, Fractal effects of dynamic crack propagation, Acta Mech.
shear behavior of steel bridge deck pavement pasted by GFRP sheets, J. Test. Sin. 27 (1) (1995) 18–27.
Eval. 42 (1) (2014) 93–108.