Construction and Building Materials 216 (2019) 379–389

Contents lists available at ScienceDirect

Construction and Building Materials

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

Fatigue cracking performance evaluation of laboratory-produced

polymer modified asphalt mixture containing reclaimed asphalt
pavement material
Zhou Zhou a, Xingyu Gu a,b,⇑, Jiwang Jiang a, Fujian Ni a, Yanxu Jiang a
a
School of Transportation, Southeast University, Nanjing, Jiangsu Province 211189, China
b
College of Engineering, Tibet University, Lhasa, Tibet 850000, China

h i g h l i g h t s

The fatigue cracking performance of polymer modified asphalt mixture containing RAP material was evaluated.

The repeated loading semi-circular test was utilized.

The mixture fatigue cracking performance was analyzed by energy consumption approach.

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

Article history: This study investigated the effect of reclaimed asphalt pavement (RAP) on the rheological properties of a
Received 26 February 2018 polymer modified binder and the fatigue cracking performance of polymer modified asphalt mixture.
Received in revised form 28 April 2019 Performance grading, master stiffness curve and black space analysis were conducted on the asphalt bin-
Accepted 5 May 2019
ders at different laboratory ageing conditions. The repeated loading semi-circular (R-SCB) test was con-
Available online 9 May 2019
ducted on the mixtures. Test results of the asphalt binder showed that the addition of RAP increased the
stiffness and the Glove-Rowe (G-R) Parameter value of asphalt binder. Higher RAP binder could be sus-
Keywords:
ceptible to non-load-associated cracking after long-term ageing or in a cold climate as binders with more
Asphalt mixture
Reclaimed asphalt pavement
than 30% RAP after the PAV 40 h ageing condition had a G-R value greater than 180 kPa. The fatigue dam-
Fatigue cracking performance age in R-SCB test were analyzed by energy consumption approach and can be divided into three stages
SCB test including the steady and cumulative damage stage and jump damage. The cumulative energy increased
linearly with the loading cycles and the cumulative energy evolution meets Linear Miner’s Law in R-SCB
test, then the average consumption per cycle (ECPCAVE) was selected as the index to evaluate the fatigue
cracking performance of asphalt mixtures. It is found that higher ECPCAVE, faster cumulative rate of fati-
gue damage causing a lower fatigue life and poor fatigue cracking performance of asphalt mixtures. The
addition of RAP contributed to the increase of ECPCAVE and the cumulative rate of fatigue damage. The
higher RAP contents, the faster the cumulative rate of fatigue damage was, which caused smaller fatigue
life. This paper suggests that ECPCAVE is a useful indicator to evaluate the fatigue cracking performance of
asphalt mixture with and without RAP.
Ó 2019 Elsevier Ltd. All rights reserved.

1. Introduction ture performance, especially cracking resistance [2–5]. To address

this concern, several methods have been used to improve the
There is a trend towards utilizing more reclaimed asphalt pave- cracking performance of RAP mixtures. The most popular and
ment (RAP) in asphalt mixtures because of economic and environ- widely used method is using a softer binder [6,7]. However, in
mental benefits over the past few years. However, the use of RAP in China, the SBS modified binder is usually utilized in the construc-
surface courses is relatively low in China [1]. The main concern is tion of surface course of highway and the performance grade is
that the aged RAP binder would cause negative effect on the mix- usually PG70-xx or PG76-xx. So, when using high RAP (usually
higher than 25%) in the surface course, the rejuvenator is often
⇑ Corresponding author at: School of Transportation, Southeast University, used to soften the RAP binder [8–11]. So far, there are limited
Nanjing, Jiangsu Province 211189, China.
E-mail address: [email protected] (X. Gu).

https://doi.org/10.1016/j.conbuildmat.2019.05.031
0950-0618/Ó 2019 Elsevier Ltd. All rights reserved.
380 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389

researches regarding the fatigue cracking resistance of high RAP 50%). A rejuvenator was used to soften and partially recovered
surface course mixtures with modified binder. the aged binder when introducing more than 30% RAP. The virgin
There are many studies focusing on the fatigue cracking perfor- aggregate used throughout this study was basalt stone from a con-
mance of RAP mixtures and binders. Huang et al. evaluated the struction agency in Jiangsu Province of China. RAP was collected
fatigue performance of HMA mixtures containing RAP by conduct- from the top layer of the Huning Highway in Jiangsu Province. A
ing semi-circular bending (SCB) fatigue tests and beam fatigue SBS modified binder meeting the superpave PG 70-22 was selected
tests. Results indicated that the addition of RAP increases the fati- because it is commonly required by Chinese transportation agen-
gue resistance of the mixtures [12]. Shu et al. performed strain con- cies for use in surface courses. Assessment of the effects of RAP
trolled beam fatigue tests and used the ratio of dissipated energy on fatigue cracking performance was made at both binder and mix-
change (RDEC) to determine failure to evaluate the fatigue charac- ture levels.
teristics of RAP mixtures. The authors found the addition of RAP
resulted in higher PV values, which corresponds to lower fatigue
life. Then the addition of RAP reduces the fatigue life of the mix- 3. Materials and experimental plan
tures [13]. Mogawer et al. evaluated the cracking resistance char-
acteristics of plant produced hot-mix asphalt containing up to 3.1. Materials
40% RAP. The recovered binders were also evaluated. His
researches showed that both the recovered binder and mixture 3.1.1. RAP characterization
stiffness increased with increasing RAP content and that if using RAP binder was extracted and recovered from the RAP mixes in
softer binder as the virgin binder, the overall stiffness would accordance with AASHTO T 164 and ASTM D 5404, respectively. A
decrease. Cracking resistance was measured using the Overlay Tes- rejuvenator was then added to the RAP binder to restore physical
ter (OT) device at 15 °C and the results showed decreasing cracking and chemical properties of the binder. The rejuvenator is a mix
resistance (lower number of cycles to failure) with increasing RAP of chemicals and consists of modified polyamines. The rejuvenator
contents [14]. McDaniel et al. performed push-pull fatigue test to dosage was determined according to the research by Arámbula-
study the effect of RAP contents and virgin binders on the fatigue
life of the mixes and he found that the increasing RAP contents
increased the fatigue life of the mixtures [15]. Li and Gibson
employed the simplified viscoelastic continuum damage (S-
VECD) to evaluate RAP fatigue performance. Results showed that
increasing RAP percentage reduced the fatigue performance of
the mixtures [16].
In summary, the conclusions regarding the effect of RAP on the
fatigue cracking performance of asphalt mixtures are conflict with
each other. In particular, there are few studies focusing on the fati-
gue cracking performance of polymer modified asphalt mixture
containing RAP materials. As the polymer modified asphalt mix-
tures are usually used in surface course in China, it is necessary
to study the effect of RAP on the fatigue performance of surface
course mixtures.

2. Objectives and scope

The primary objectives of this paper are as follows:

(1) Assess the influences of RAP contents on the intermediate

fatigue cracking performance of polymer modified asphalt Fig. 1. Recovered RAP material aggregate gradation.
binder;
(2) Assess the sensitivity of the fatigue cracking performance of Table 2
polymer modified asphalt mixtures to RAP contents; Recovered RAP binder properties.
(3) Compare the fatigue cracking performance of asphalt bin- RAP type RAP (＞ RAP RAP
ders and mixtures. 9.5 mm) (4.75  9.5 mm) (0  4.75 mm)
Asphalt content 3.05 3.59 5.63
To achieve these objectives, one AC-13C surface mixture was (%)
selected to incorporate various RAP contents (0, 15, 30, 40, and PG true grade (°C) 84.1–15.1 84.3–14.8 84.0–15.0

Table 1
Superpave Test Results for blended binder for different rejuvenator dosages.

Tested Binder High Temp. Inter. Temp. Low Temp. PG Grade

Origin RTFOT PAV (20 h)
S/MPa m-slope
RAP 84.1 — 22.1 16.7 15.1 84–15
RAP + 2%RA 77.0 76.6 22.6 20.1 19.2 76–19
RAP + 4%RA 73.9 73.0 20.4 –23.2 –22.6 73–23
RAP + 6%RA 69.8 69.1 14.2 26.1 25.8 69–25

Note: RA = rejuvenator,  = Data Not Applicable.

 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389 381

Table 3
Characteristics of Laboratory-Designed Mixtures.

Mixture Property V70-0 V70-15 V70-30 V70-40 V70-50

RAP content, % 0 15 30 40 50
Rejuvenator dosage, % 0 0 4 4 4
Optimum AC, % 5.0 5.0 4.9 4.9 4.8
Air Voids, % 4.0 4.0 4.0 4.0 4.0
VMA, % 15.0 14.9 14.7 14.3 13.9
VFA, % 73.3 73.0 72.8 72.5 72.0
DB ratio 1.2 1.3 1.4 1.6 1.9
Sieve Size, mm
16.0 100.0 100.0 100.0 100.0 100.0
13.2 95.0 96.3 95.7 95.4 95.5
9.5 76.5 77.5 76.4 76.5 76.4
No. 4 53.0 52.5 53.0 53.0 53.0
No. 8 37.0 36.6 37.1 36.9 37.0
No. 16 26.5 25.6 26.5 26.7 26.7
No. 30 19.0 17.8 18.8 18.9 19.0
No. 50 13.5 12.4 13.3 13.3 13.3
No. 100 10.0 9.0 9.7 9.7 9.7
No. 200 6.0 6.2 6.6 6.8 7.0

Note: 1. V70-0, V70-15, V70-30, V70-40, and V70-50 indicate Virgin PG 70–22 SBS modified binder with different percentages of RAP materials.
2. AC = asphalt content, including virgin binder, aged binder from RAP (100% blending between virgin and recycled binder was assumed) and rejuvenator.
3. VMA = Voids in aggregate, VMA (%).
4. VFA = Voids filled w/asphalt, VFA (%).
5. DB ratio = Dust-binder ratio.

Fig. 2. Half-sinusoidal load in R-SCB test.

Fig. 3. R-SCB test configuration.

Mercado, Edith, et al. [17] and this method is based on the perfor-
mance grading test according to AASHTO T 315. The dosage was
cal specifications for construction of highway asphalt pavements
selected as 2%, 4% and 6% by the weight of the RAP binder.
(JTG F40-2004) [18], was produced in the laboratory and served
The results of blended binder performance for different rejuve-
as the control mixture. Then the mixtures incorporating various
nator dosages are shown in Table 1.
percentages of RAP materials were produced. The control mixture
From Table 1, it can be seen that a dosage rate of 4% of the reju-
design was modified to include various contents of RAP material.
venator by the weight of the RAP binder could be used to restore
Virgin aggregate was adjusted as needed to maintain the gradation
the continuous grade of the RAP binder from 84.1 to 15.1 to
of all RAP mixtures as close as possible to the control mixture. The
73.0–22.6. Then the dosage rate of 4% of the rejuvenator was used
rejuvenator with 4% of the weight of aged binder was added in
for mixture performance evaluation as the target performance
mixtures which contain more than 30% RAP contents to evaluate
grade of virgin binder is 70–22.
the effect of rejuvenator on the fatigue cracking resistance of
RAP materials were divided into three parts by the aggregate
asphalt mixtures containing RAP. To produce the consistent virgin
sizes: 0–4.75 mm, 4.75–9.5 mm and 9.5–16 mm. Fig. 1 shows the
aggregate gradation for all RAP mix designs, the virgin basalt
recovered aggregate gradation after extraction of the divided three
aggregates were divided into sieve sizes, which were then recom-
parts RAP materials. Table 2 shows the asphalt content and PG true
bined to produce the desired aggregate gradation. Determination
grade for RAP materials of each part.
of asphalt content for the control and all RAP mixtures was based
on results in 4% air voids at N design ¼ 75 using Superpave Gyratory
3.1.2. Mix design Compactor (SGC). More details of the design of high percentage
A dense-graded surface course mixture with nominal maximum of RAP mixtures can be found in the research by Yu et al. [8]. Table 3
aggregate size (NMAS) of 13.2 mm, designed based on the techni- summarizes the mix design results for all mixtures.
382 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389

Table 4
Performance Grading Results for tested binders.

Tested Binder High Temp. °C Inter. Temp. °C Low Temp. °C Continuous Grade, °C PG Grade
Origin RTFOT PAV(20 h)
S/MPa m-slope
V70-0 74.1 73.9 24.7 25.6 26.0 73.9–25.6 70–22
V70-15 75.5 75.2 25.8 24.3 25.2 75.2–24.3 70–22
V70-30 75.0 74.0 24.7 24.1 25.2 74.0–24.1 70–22
V70-40 75.4 75.0 24.0 23.9 25.0 75.0–23.9 70–22
V70-50 74.1 74.5 24.8 23.9 25.3 74.5–23.9 70–22

Table 5 controlled frequency sweep tests were conducted on the binders

Binder rheological properties. using DSR. The frequencies for the shear oscillating loading were
Binder Type Ageing Condition R xc Td
over a frequency range from 100 to 0.1 rad/s at each temperature
and strain level. The temperatures were 5, 15, 25, 35, 46, 58, 64,
V70-0 Original 1.668 73.16 9.36
70, and 76 °C. 2 different geometries binders were used:
RTFO 1.682 6.34 7.21
PAV 20 h 1.757 1.15 3.86
PAV 40 h 1.911 0.43 2.48 (1) 8 mm diameter with 2 mm gap for T = 5 °C to 35 °C and
V70-15 Original 1.700 49.96 8.45 (2) 25 mm diameter with 1 mm gap for T = 46 °C to 76 °C.
RTFO 1.726 6.27 5.20
PAV 20 h 1.886 1.14 2.95 The amplitude strain was set at 0.1% for intermediate tempera-
PAV 40 h 2.133 0.28 1.21 tures (5–35 °C) and 1% for high temperatures (46–76 °C) to ensure
V70-30 Original 1.709 26.43 7.72
the binders were loaded in a linear viscoelastic range. For asphalt
RTFO 1.806 5.29 4.94
PAV 20 h 1.960 0.61 2.65 binder frequency sweep testing, original, RTFO, 20 h PAV ageing
PAV 40 h 2.189 0.13 0.37 and 40 h PAV ageing conditions were utilized.
V70-40 Original 2.078 20.86 6.93
RTFO 2.165 4.46 4.76
PAV 20 h 2.305 0.14 0.70
3.2.2. Asphalt mixture testing program—repeated loading SCB test
PAV 40 h 2.425 0.046 1.22
V70-50 Original 2.268 10.95 5.65 The Repeated Loading Semi-Circular Bending (R-SCB) test was
RTFO 2.333 2.63 3.82 performed as the fatigue cracking test for RAP mixtures in this
PAV 20 h 2.371 0.11 0.76 study.
PAV 40 h 2.565 0.002 2.08 The R-SCB test was used to evaluate the effect of RAP contents
on the fatigue fracture property of polymer asphalt mixtures. The
semi-circular specimen was loaded under a three-point dynamic
compressive loading by a stress control mode. In the repeated
loading mode, the specimen was loaded with a half-sinusoidal load
at a frequency of 10 Hz with no rest time, as shown in Fig. 2. Three
levels of load, 0.15, 0.2 and 0.25 times of the average SCB specimen
strength of each mixtures were applied to the specimens. The SCB
specimen strength was determined by the SCB strength test under
50 mm/min loading rate [9]. This specific loading rate is consistent
with the indirect tensile strength test procedure ASTM D6913-12

Fig. 4. Glove-Rowe Parameter Results of all tested binders.

3.2. Laboratory testing program

3.2.1. Asphalt binder testing program

The asphalt binder used in the laboratory-prepared mixtures
was extracted and recovered. After the recovery process, the true
performance grade properties of the binder was firstly evaluated
by using the Dynamic Shear Rheometer (DSR) and Bending Beam
Rheometer (BBR). Next, G* master curves were measured and uti-
lized to evaluate the overall stiffness properties of the asphalt bin-
ders as well as their relative ageing characteristics. Strain Fig. 5. Energy Consumption in one cycle in R-SCB test.
 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389 383

Fig. 6. ECPC evolution in R-SCB test for V70-30 mixture: (a) 15 °C: (b) 25 °C.

[19]. Two test temperatures—15 °C and 25 °C were selected in this supports is 120 mm. The failure standard for R-SCB test is the bro-
study to correspond to the intermediate temperature range at ken of the specimen. More details of R-SCB test can be found else-
which most of the traffic load is applied. The load and deformation where [20–22].
were continuously recorded in R-SCB test. Fig. 3 presents the R-SCB The R-SCB test was conducted using a UTM-25 system. For R-
test configuration in this study. The distance between two bottom SCB test, at least three replicates of each mixture were evaluated.
384 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389

The temperature-shift factor used in the CA model was

expressed through WLF equation:

C 1 ðT  T d Þ
log aðT Þ ¼  ð2Þ
C 2 þ ðT  T dÞ

where, log aðTÞ = shift factor; T = temperature; Td = defining tem-

perature; and C1 and C2 are constants.
Many studies have proved that the CA model parameter xc, R
and Td have specific significance. xc is defined as the crossover fre-
quency and larger crossover frequency indicates lower degree of
ageing. R-value is the Rheological Index. R-value is higher when
the asphalt binder is oxidized or aged [14]. The methodology to
establish the master curve is based on the method proposed by
Mogawer et al. [24]. In this paper, all the master curves were
shifted to the reference temperature of 15 °C. Table 5 gives the rhe-
ological indices using CA model. When the binder is aged, the R-
Fig. 7. Evolution of ECPC in R-SCB test.
value and Td values increases while the xc values decreases. From
Table 5, it can be found that the addition of RAP binder increases
3.2.2.1. Specimen preparation. The R-SCB test specimens in this the R-value and Td values while decreases the xc values of the bin-
study were obtained by slicing the SGC compacted cylindrical der, so addition of RAP could increase the hardness of asphalt bin-
specimens along their central axes. Short Term Oven Ageing der and with the increasing amounts of RAP binder, the hardness of
(STOA) was done by putting the loose mix in the oven with the the binder also increases.
compaction temperature of 135 °C for 4 h. Then the loose mixture
was compacted to a height of 180 mm and 150 mm in diameter
4.1.3. Glover – Rowe (G-R) parameter
with an air void content of 7 ± 0.5%. After that, the specimens were
The Glover–Rowe parameter, which was originally developed
cut from the middle of the specimen into three slices with a thick-
by Glover et al. [25] and then reformulated by Rowe [26] in a dis-
ness of 50 mm for each slice. Finally, each slice was cut into two
cussion to the work of Anderson, King, Hanson, and Blankenship
semi-circular halves. For each semi-circular specimen tested, a
[26], was determined with Eq. (1). The Glover–Rowe parameter
notch with 15 ± 1 mm depth and 2 ± 0.5 mm width was cut at
considers both binder stiffness and embrittlement and offers an
the center of the specimen bottom. The actual air voids for all spec-
indication of the cracking resistance at intermediate temperature.
imens are in the range of 0.5% ± target air voids.
In this paper, the G-R parameter was calculated based on establish-
ment of a master curve from frequency sweep testing at 5 °C, 15 °C,
4. Results and discussion 25 °C, and 35 °C and the value of G-R at 15 °C and 0.005 rad/s. A
higher G-R value indicates a severer brittleness of binder.
4.1. Asphalt binder performance results
jG jðcosdÞ2
4.1.1. Performance grading
GR¼ ð1Þ
sind
A summary of the performance grade properties is shown in
Table 4. The PG grade determination at high and low temperature where, G* – complex shear modulus (kPa).
was characterized according to AASHTO M320. The data in Table 4 d – phase angle (degree).
indicates that the addition of SBS modified binder/rejuvenator The results of the G-R parameter from the tested binders after
changed the performance grade of RAP binder to PG70-22. All PAV 40 h ageing are shown in Fig. 4. As the RAP contents increase,
the tested binders achieved the same target Performance Grade the G-R parameter values increase, indicating that incorporating
(PG). The addition of 15% RAP binder made the high-temperature RAP would cause the mixture more prone to cracking. And it is
grade and low-temperature grade increased indicating that the found that the G-R parameter value increases dramatically when
addition of RAP binder can make the virgin SBS binder more stiff. the RAP content is more than 30%. This phenomenon suggests that
However, as the addition of rejuvenator, the high temperature the cracking of the binder may be poor when introducing higher
grade and low-temperature grade of binders containing more than amounts of RAP binder. Based on the research by Rowe [26], bin-
30% RAP binder were lower than that of binder with 15% RAP ders with G-R parameter values greater than 180 and 600 kPa are
which is attributed to the undesirable softening effect caused by prone to the onset of cracking and significant cracking, respec-
the rejuvenator. tively. As shown in Fig. 4, binders with 30%, 40% and 50% RAP after
the PAV 40 h ageing condition failed the ‘‘Damage Onset” value,
indicating that higher RAP binder could be susceptible to non-
4.1.2. Rheological indices using Christensen-Anderson model (CA)
load-associated cracking after long-term ageing or in a cold
The Christensen-Anderson Model (CA) is used to build the bin-
climate.
der master curve as given by Eq. (1) [23].
"
ðlog2=R #R=ðlog2Þ
xc 4.2. R-SCB test results
G ðxÞ ¼ Gg 1 þ ð1Þ
xr
Several researches have analyzed R-SCB results from displace-
where, G ðxÞ = complex shear modulus; G(g) = glass modulus ment evolution or strength degradation [20,21,27–29]. None of
assumed equal to 1 GPa; xr = reduced frequency at the defining them studied the test results from the energy and dissipated
temperature (rad/s); xc = crossover frequency at the defining tem- energy approach. This paper developed a method to analyze the
perature (rad/s); x = frequency (rad/s); and R = rheological index. R-SCB test results from energy consumption approach.
 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389 385

Table 6
Results of the R-SCB test.

Temp (°C) Mixture Stress ratio Nf Mean Nf COV (%) ECPCAVE(J) Mean ECPCAVE(J) COV (%)
15 V70-0 0.15 313900 412,800 26.5 0.0129 0.0138 5.3
568,450 0.0138
375,500 0.0146
393,350 0.0140
0.20 329,874 271,920 21.4 0.0199 0.0211 6.3
303,742 0.0223
256,462 0.0200
197,602 0.0222
0.25 166,885 140,010 16.2 0.0344 0.0337 2.8
146,416 0.0332
134,005 0.0326
112,734 0.0346
V70-15 0.15 447,982 359,480 20.8 0.0184 0.0179 5.4
380,566 0.0180
339,692 0.0165
269,637 0.0187
0.20 223,145 172,480 24.8 0.0329 0.0316 7.4
184,553 0.0317
161,155 0.0283
121,053 0.0335
0.25 112,182 124,250 6.5 0.0390 0.0368 10.9
127,492 0.0370
128,123 0.0311
129,203 0.0401
V70-30 0.15 184,492 240,570 15.6 0.0229 0.0230 6.4
255,635 0.0244
258,565 0.0210
263,587 0.0237
0.20 120,975 96,700 16.8 0.0429 0.0431 8.8
90,179 0.0467
88,910 0.0379
86,736 0.0449
0.25 39,522 47,700 22.8 0.0575 0.0528 6.7
37,504 0.0527
59,643 0.0489
54,131 0.0522
V70-40 0.15 187,064 234,090 26.4 0.0250 0.0259 5.9
176,867 0.0242
302,487 0.0270
270,045 0.0274
0.20 71,465 75,610 7.2 0.0455 0.0480 8.9
70,567 0.0433
81,633 0.0511
78,775 0.0521
0.25 36,120 42,570 19.9 0.0555 0.0533 4.3
34,722 0.0545
51,943 0.0528
47,495 0.0503
V70-50 0.15 184,233 206,920 14.4 0.0309 0.0291 6.2
179,317 0.0301
239,886 0.0287
224,244 0.0268
0.20 69,331 61,510 16.7 0.0547 0.0506 8.2
71,026 0.0528
50,145 0.0497
55,538 0.0452
0.25 30,313 25,320 25.9 0.0664 0.0625 6.4
31,395 0.0647
18,065 0.0616
21,507 0.0573
25 V70-0 0.15 292,513 271,510 18.9 0.00799 0.00756 4.4
291,634 0.00721
306,660 0.00764
195,233 0.00743
0.20 145,140 132,200 24.3 0.0138 0.0132 3.6
144,447 0.0127
84,526 0.0133
154,687 0.0130
0.25 56,186 69,450 26.2 0.0175 0.0163 5.7
90,589 0.0153
78,541 0.0165
52,484 0.0159
V70-15 0.15 223,546 252,320 15.2 0.00756 0.00824 6.5
283,260 0.00882

(continued on next page)

386 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389

Table 6 (continued)

Temp (°C) Mixture Stress ratio Nf Mean Nf COV (%) ECPCAVE(J) Mean ECPCAVE(J) COV (%)
287,467 0.00812
215,006 0.00846
0.20 115,973 100,500 10.4 0.0137 0.0144 3.8
93,566 0.0150
97,252 0.0143
95,208 0.0146
0.25 52,538 55,090 14.4 0.0168 0.0176 3.6
66,741 0.0183
52,145 0.0174
48,936 0.0179
V70-30 0.15 153,594 222,320 23.6 0.00804 0.00952 12.3
266,077 0.0108
209,270 0.00925
260,341 0.0100
0.20 74,285 71,770 15.7 0.0137 0.0161 12.0
55,210 0.0182
79,711 0.0157
77,876 0.0169
0.25 29,007 47,250 26.5 0.0186 0.0196 4.0
56,116 0.0202
49,204 0.0203
54,673 0.0194
V70-40 0.15 123,582 192,040 23.8 0.0113 0.0110 4.2
217,170 0.0106
216,050 0.0114
211,358 0.0106
0.20 54,250 67,380 13.6 0.0184 0.0166 14.5
74,390 0.0143
72,854 0.0190
68,027 0.0147
0.25 26,654 38,090 21.3 0.0234 0.0215 11.8
38,053 0.0191
44,491 0.0240
43,162 0.0195
V70-50 0.15 187,463 159,190 22.3 0.0121 0.0122 4.0
191,706 0.0128
122,619 0.0117
134,972 0.0122
0.20 74,708 62,720 24.0 0.0197 0.0198 3.4
76,508 0.0207
47,213 0.0191
52,451 0.0198
0.25 19,063 23,850 25.2 0.0242 0.0245 6.1
18,345 0.0265
30,042 0.0228
27,950 0.0245

4.2.1. Energy consumption evolution in R-SCB test of all tested samples are listed in Table 6. As presented in Table 6,
The authors found that in R-SCB test, the load-displacement the coefficient of the variance (COV) values for ECPCAVE in the R-
curve in one cycle can form a loop as shown in Fig. 5. With the SCB test were relatively low, proving that the energy consumption
increasing loading cycles, the energy will accumulate until the
specimen is broken. The loop area is defined as the energy con-
sumption per cycle. The authors calculated the energy consump-
tion per cycle values of the tested mixtures and Fig. 6 shows the
typical energy consumption per cycle evolution trends of asphalt
mixture in R-SCB test. It can be found that the fatigue damage evo-
lution of mixture in R-SCB test consists of three stages as shown in
Fig. 7: the unsteady stage of short duration cause of UTM setup, the
steady and cumulative damage stage, and jump damage. In the
steady and cumulative damage stage, the energy consumption
per cycle kept stable and this stage is the primary fatigue damage
stage. In the last stage, which is defined as jump damage, the
energy consumption per cycle increases dramatically and the spec-
imen is broken quickly. This stage is also of short duration. Based
on the energy consumption per cycle evolution presented in
Fig. 6, the fatigue life Nf in R-SCB test is defined as the terminal
point loading cycle of the steady and cumulative damage stage,
which is not the loading cycle when the specimen is failure.
The mean fatigue life (Mean Nf) and average energy consump-
tion per cycle (Mean ECPCAVE) values in the steady stage results Fig. 8. Relationship between Nf and ECPCAVE.
 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389 387

Fig. 9. Relationship between cumulative energy and loading cycles.

per cycle in the steady stage kept stable. It is also found that there Nf and ECPCAVE can be expressed in Eq. (2). The higher the ECPCAVE,
is a good linear relationship between the logarithm of Nf and the lower the fatigue life Nf is.
ECPCAVE, as shown in Fig. 8. The relationship of the logarithm of
LOG10 ðNf Þ ¼ k  ECPC AVE þ b ð2Þ
388 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389

Fig. 10. Effects of RAP contents on the ECPCAVE: (a) 15 °C: (b) 25 °C.
Fig. 11. Asphalt binder rheological property results vs Asphalt Mixture ECPCAVE in
R-SCB test: (a) master curve crossover frequency, (b) Log (G-R).
where, k, b—derived parameters from the test results.
with the cumulative energy increasing rate, it can be used to pre-
dict the mixture fatigue cracking performance. The higher ECPCAVE,
4.2.2. Understand ECPCAVE—Cumulative energy evolution in R-SCB test the higher cumulative increasing rate and faster cumulative rate of
The cumulative energy is the sum of the energy consumption fatigue damage would result in a lower fatigue life and poor fatigue
per cycle. As shown in Fig. 9, the relationship of cumulative energy cracking performance of mixture.
with loading cycles is examined. It is indicated that the cumulative
energy increases linearly with the loading cycles and this trend 4.2.3. The effects of RAP on the fatigue cracking performance
changes and cumulative energy increases faster only when the Fig. 10 shows the effects of RAP contents on the ECPCAVE results.
specimen is close to broken. Then it can be assumed that the cumu- As Fig. 8 demonstrates, the ECPCAVE values increase with the RAP
lative energy evolution meets Linear Cumulative Damage (Miner’s contents. The ECPCAVE represents the cumulative rate of fatigue
Law) in R-SCB test [30,31]. According to linear Miner’s law, mixture damage. So, the addition of RAP increases the ECPCAVE, which also
cracks when cumulated damage exceeds unity. In R-SCB test, the means increasing cumulative rate of fatigue damage. The higher
damage accumulated linearly until the specimen is encountering RAP contents, the faster the cumulative rate of fatigue damage is,
failure. Fig. 10 also presents the fitting results of cumulative energy which results in smaller fatigue life. Incorporating RAP decreases
(CE) versus loading cycle (N). The coefficient of determination R2 the fatigue cracking performance of the mixture. The ECPCAVE
values are all greater than 0.98. It is also found that the slope of index is a useful indicator to evaluate the fatigue cracking perfor-
the fitting curve is similar to ECPCAVE. Through ECPCAVE, it is conve- mance of asphalt mixture with and without RAP.
nient to get the slope of the fitting curve of cumulative energy ver-
sus loading cycle. 4.3. Relationships between binder and R-SCB test results
The slope of the cumulative energy vs. loading cycle curve rep-
resents the increasing rate of cumulative energy and it implies the The relationships between the asphalt binder and mixture
cumulative rate of mixture fatigue damage. Higher cumulative intermediate fatigue cracking measured in this study were evalu-
energy increasing rate and lower cumulative energy mean faster ated. The proposed ECPCAVE index was compared with the results
cumulative rate of mixture fatigue damage and the fatigue life of from asphalt binder rheological test (crossover frequency xc and
mixture is relatively low. Since the ECPCAVE is positively correlated G-R parameter). The performance comparisons were made using
 Z. Zhou et al. / Construction and Building Materials 216 (2019) 379–389 389

15 °C mixture R-SCB results and 40 h PAV aged asphalt binders [3] R.C. West, J.R. Willis, M.O. Marasteanu, Improved mix design, evaluation, and
materials management practices for hot mix asphalt with high reclaimed
results, which is shown in Fig. 11. The results show a good correla-
asphalt pavement content, Transp. Res. Board (2013).
tion between the ECPCAVE and the Crossover Frequency. The [4] G. Valdés, F. Pérez-Jiménez, R. Miró, A. Martínez, R. Botella, Experimental study
ECPCAVE value increases with the decreasing Crossover Frequency. of recycled asphalt mixtures with high percentages of reclaimed asphalt
There is also an excellent correlation between the ECPCAVE and the pavement (RAP), Constr. Build. Mater. 25 (3) (2011) 1289–1297.
[5] F. Xiao, N. Su, S. Yao, S. Amirkhanian, J. Wang, Performance grades,
G-R parameter. As G-R parameter increases, the ECPCAVE also environmental and economic investigations of reclaimed asphalt pavement
increases. These two trends are reasonable and prove that ECPCAVE materials, J. Cleaner Prod. 211 (2019) 1299–1312.
can be used to evaluate asphalt mixture fatigue cracking [6] A. Copeland, Reclaimed asphalt pavement in asphalt mixtures: State of the
practice, 2011.
performance. [7] I.L. Al-Qadi, Q. Aurangzeb, S.H. Carpenter, W.J. Pine, J. Trepanier, Impact of high
RAP contents on structural and performance properties of asphalt mixtures,
2012.
5. Findings and conclusions
[8] B. Yu, X. Gu, M. Wu, F. Ni, Application of a high percentage of reclaimed asphalt
pavement in an asphalt mixture: blending process and performance
Based on the test results, the following findings and conclusions investigation, Road Mater. Pavement Des. 18 (3) (2017) 753–765.
[9] J. Li, F. Ni, J. Jin, Z. Zhou, A comparison of rejuvenator and sryrene–butadiene
can be obtained:
rubber latex used in hot in-place recycling, Road Mater. Pavement Des. 18 (1)
(2017) 101–115.
(1) With the increasing of RAP contents, the Glove-Rowe Param- [10] N. Guo, Z. You, Y. Zhao, Y. Tan, A. Diab, Laboratory performance of warm mix
eter value increased especially when the RAP amount is asphalt containing recycled asphalt mixtures, Constr. Build. Mater. 64 (2014)
141–149.
more than 30% after PAV 40 h ageing condition. And binders [11] Q. Li, X. Chen, G. Li, S. Zhang, Fatigue resistance investigation of warm-mix
with 30%, 40% and 50% RAP after the PAV 40 h ageing condi- recycled asphalt binder, mastic, and fine aggregate matrix, Fatigue Fract. Eng.
tion failed the ‘‘Damage Onset” value, indicating that higher Mater. Struct. (2017).
[12] B. Huang, W. Kingery, Z. Zhang, Laboratory study of fatigue characteristics of
RAP binder could be susceptible to non-load-associated HMA mixtures containing RAP, International Symposium on Design and
cracking after long-term ageing or in a cold climate. Construction of Long Lasting Asphalt Pavements, 2004, Auburn, Alabama, USA,
(2) The fatigue damage in R-SCB test can be divided into three 2004.
[13] X. Shu, B. Huang, D. Vukosavljevic, Laboratory evaluation of fatigue
stages from energy consumption approach: the unsteady characteristics of recycled asphalt mixture, Constr. Build. Mater. 22 (7)
stage of short duration, the steady and cumulative damage (2008) 1323–1330.
stage, and jump damage. In the steady and cumulative dam- [14] W. Mogawer, T. Bennert, J.S. Daniel, R. Bonaquist, A. Austerman, A.
Booshehrian, Performance characteristics of plant produced high RAP
age stage, the energy consumption per cycle kept stable.
mixtures, Road Mater. Pavement Design 13 (sup1) (2012) 183–208.
(3) The cumulative energy increased linearly with the increas- [15] R.S. McDaniel, A. Shah, G. Huber, Investigation of low-and high-temperature
ing loading cycles and the cumulative energy evolution properties of plant-produced, RAP mixtures (2012).
[16] X. Li, M. Marasteanu, R. Williams, T. Clyne, Effect of reclaimed asphalt
met Linear Miner’s Law in R-SCB test.
pavement (proportion and type) and binder grade on asphalt mixtures, Transp.
(4) The slope of the cumulative energy vs. loading cycles curve Res. Rec.: J. Transp. Res. Board 2051 (2008) 90–97.
represents the increasing rate of cumulative energy. The [17] E. Arámbula-Mercado, F. Kaseer, A.E. Martin, F. Yin, L.G. Cucalon, Evaluation of
ECPCAVE was positively correlated with the cumulative recycling agent dosage selection and incorporation methods for asphalt
mixtures with high RAP and RAS contents, Constr. Build. Mater. 158 (2018)
energy increasing rate. The higher ECPCAVE, the higher 432–442.
cumulative increasing rate and faster cumulative rate of fati- [18] J. Sheng, F. Li, J. Chen, Technical specifications for construction of highway
gue damage results in a lower fatigue life and poor fatigue asphalt Pavement, Ministry of Communications of the People’s Republic of
China, Beijing, 2005.
cracking performance of mixture. [19] ASTM, Standard test method for indirect tensile (IDT) strength of bituminous
(5) The addition of RAP increased the ECPCAVE, which also mixtures, (2007).
means an increasing cumulative rate of fatigue damage. [20] F. Ni, B. Yu, X. Gu, K. Xu, Characterization of Fatigue Damage of Hot Mix
Asphalt Based on Residual Strength, J. Mater. Civ. Eng. 29 (8) (2017) 04017075.
Higher RAP contents meant faster cumulative rate of fatigue [21] J. Jiang, F. Ni, Q. Dong, F. Wu, Y. Dai, Research on the fatigue equation of asphalt
damage. mixtures based on actual stress ratio using semi-circular bending test, Constr.
(6) ECPCAVE can be used to evaluate asphalt mixture fatigue Build. Mater. 158 (2018) 996–1002.
[22] J. Jiang, F. Ni, Q. Dong, Y. Zhao, K. Xu, Fatigue damage model of stone matrix
cracking performance and it has an excellent correlation
asphalt with polymer modified binder based on tensile strain evolution and
with the intermediate cracking performance indices of residual strength degradation using digital image correlation methods,
asphalt binders. Measurement 123 (2018) 30–38.
[23] J. Wang, T. Wang, X. Hou, F. Xiao, Modelling of rheological and chemical
properties of asphalt binder considering SARA fraction, Fuel 238 (2019) 320–
330.
Declaration of Competing Interest [24] A. Booshehrian, W.S. Mogawer, R. Bonaquist, How to construct an asphalt
binder master curve and assess the degree of blending between RAP and virgin
binders, J. Mater. Civ. Eng. 25 (12) (2012) 1813–1821.
None. [25] C.J. Glover, R.R. Davison, C.H. Domke, Y. Ruan, P. Juristyarini, D.B. Knorr, S.H.
Jung, Development of a new method for assessing asphalt binder durability
Acknowledgments with field validation, Texas Dept Transport 1872 (2005).
[26] R.M. Anderson, G.N. King, D.I. Hanson, P.B. Blankenship, Evaluation of the
relationship between asphalt binder properties and non-load related cracking,
This work was supported by the Basic Research Plan (Natural J. Assoc. Asphalt Paving Technol. 80 (2011).
Science Foundation) in the Jiangsu Province of China (Grant No. [27] Q. Li, H.J. Lee, T.W. Kim, A simple fatigue performance model of asphalt
mixtures based on fracture energy, Constr. Build. Mater. 27 (1) (2012) 605–
BK20151413) and National Natural Science Foundation of China 611.
(Grant No.51878162). The authors appreciate the funding [28] G. Saha, K.P. Biligiri, Fracture properties of asphalt mixtures using semi-
supports. circular bending test: a state-of-the-art review and future research, Constr.
Build. Mater. 105 (2016) 103–112.
[29] J. Wang, J. Yuan, F. Xiao, Z. Li, J. Wang, Z. Xu, Performance investigation and
References sustainability evaluation of multiple-polymer asphalt mixtures in airfield
pavement, J. Cleaner Prod. 189 (2018) 67–77.
[1] J. Xu, S. Huang, Y. Qin, Asphalt Pavement Recycling in Mainland China, [30] R.M. Christensen, An evaluation of linear cumulative damage (Miner’s Law)
Application of Reclaimed Asphalt Pavement and Recycled Asphalt Shingles in using kinetic crack growth theory, Mech. Time-Dependent Mater. 6 (4) (2002)
Hot-Mix Asphalt 51 (2014). 363–377.
[2] M. Zaumanis, R.B. Mallick, Review of very high-content reclaimed asphalt use [31] L. Sun, W.R. Hudson, Probabilistic approaches for pavement fatigue cracking
in plant-produced pavements: state of the art, Int. J. Pavement Eng. 16 (1) prediction based on cumulative damage using Miner’s law, J. Eng. Mech. 131
(2015) 39–55. (5) (2005) 546–549.