See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/373082039

Assessing permanent deformation potential of asphalt mixtures based on

viscoelastic characteristics

Article in International Journal of Pavement Engineering · August 2023

DOI: 10.1080/10298436.2023.2240472

CITATION READS

1 69

5 authors, including:

Meng Ling Luis Fuentes

Texas A&M University Universidad del Norte (Colombia)
21 PUBLICATIONS 517 CITATIONS 91 PUBLICATIONS 1,171 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Meng Ling on 27 August 2023.

The user has requested enhancement of the downloaded file.

 International Journal of Pavement Engineering

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/gpav20

Assessing permanent deformation potential

of asphalt mixtures based on viscoelastic
characteristics

Meng Ling, Sang Ick Lee, Jie Ji, Luis Fuentes & Lubinda F. Walubita

To cite this article: Meng Ling, Sang Ick Lee, Jie Ji, Luis Fuentes & Lubinda F. Walubita (2023)
Assessing permanent deformation potential of asphalt mixtures based on viscoelastic
characteristics, International Journal of Pavement Engineering, 24:1, 2240472, DOI:
10.1080/10298436.2023.2240472

To link to this article: https://doi.org/10.1080/10298436.2023.2240472

Published online: 11 Aug 2023.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=gpav20
INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING
2023, VOL. 24, NO. 1, 2240472
https://doi.org/10.1080/10298436.2023.2240472

Assessing permanent deformation potential of asphalt mixtures based on

viscoelastic characteristics
Meng Linga, Sang Ick Leeb, Jie Ji a
, Luis Fuentes c
and Lubinda F. Walubitad
a
School of Civil and Transportation Engineering, Beijing University of Civil Engineering & Architecture, Beijing, People’s Republic of China;
b
Texas Department of Transportation (TxDOT), Austin, TX, USA; cCivil and Environmental Engineering Department, Universidad del Norte,
Barranquilla, Colombia; dTexas A&M Transportation Institute (TTI), The Texas A&M University System, College Station, TX, USA

ABSTRACT ARTICLE HISTORY

In this study, permanent deformation potential of different asphalt mixtures subjected to the same aging Received 11 November 2022
condition was assessed, based on viscoelastic characteristics from the dynamic modulus (DM) test. First it Accepted 19 July 2023
was shown that the HWTT and RLPD test results were not well correlated, which might be due to the
KEYWORDS
differences in the boundary condition and the impacts of moisture damage and stripping in the Asphalt mixtures; permanent
HWTT. The analysis results suggested that the shape parameter (γ) of the DM master-curves could be deformation; viscoelastic
utilized to assess the rutting resistance potential of the asphalt mixtures from the RLPD test, characteristics; dynamic
compared to other rheological parameters such as the inflection point frequency and β. The |E*| at the modulus (|E*|, DM); Hamburg
RLPD test condition was well correlated to the rutting performance compared to the asphalt mixture’s wheel tracking test (HWTT);
rutting parameter |E*|/sinδ and the |E*| at the inflection point frequency. A novel rutting parameter repeated load permanent
(VERIndex) was developed to assess the rutting resistance potential of the asphalt mixtures, which deformation (RLPD) test;
simultaneously accounts for the |E*| and slope of the DM master-curves. Thereafter, the cracking and balanced mix design (BMD)
rutting performance thresholds were proposed, and, correspondingly used in the performance space
diagram to aid with implementing the balanced mix design approach using the viscoelastic
characteristics of asphalt mixtures.

Introduction
design procedures, particularly mechanistic-empirical (M-E)
Permanent deformation (PD) or rutting is one of the primary methods (Solatifar et al. 2019, Bueno et al. 2022). Amongst
distresses occurring in hot-mix asphalt (HMA) pavements. It its treasured characteristics, it is sensitive to temperature,
typically manifests as longitudinal depressions on the pave- loading time, aging, moisture, and mix-design volumetrics.
ment surface along the wheel paths, which inherently The DM test itself is also repeatable with reasonably accep-
increases the roughness of the pavement surface, reduces table variability to be reliably used for asphalt mixture prop-
the driving comfort, and even introduces motorist-safety erty characterisation and performance prediction (Kassem
hazards such as hydroplaning (Laukkanen et al. 2015, Tong et al. 2020).
et al. 2022). Rutting performance is interactively affected by It is widely accepted that higher modulus or viscosity is
several factors including asphalt mix-design, pavement layer beneficial to rutting resistance (Sivasubramaniam and Had-
properties, pavement structure, climate, and traffic loading. dock 2005, Apeagyei 2011). However, several researchers
Rut depth (RD) is accumulated in most of the pavement have noted that using modulus as a single factor is not
layers including the asphalt layers, base course, and subgrade sufficient to adequately predict the rutting resistance of asphalt
(Wang et al. 2011, Lytton et al. 2019). In addition to other mixtures. For instance, it has been shown that the RD
factors such as traffic loading and environment (e.g. tempera- measured from the asphalt pavement analyzer (APA) was
ture and moisture), the asphalt mixture’s rutting performance not able to be differentiated by the DM at elevated tempera-
is also influenced by the asphalt mixture properties including tures and loading frequencies, respectively (Birgisson et al.
viscoelastic properties, which are fundamental properties for 2004). The dynamic modulus (|E*|) at a single loading fre-
constitutive modelling and are essential inputs for most quency and temperature has also been questioned for ranking
asphalt pavement designs (Solatifar et al. 2019, Walubita the rutting resistance incorrectly as compared to some rutting
et al. 2019). performance tests such as the flow number (FN) and repeated
As widely reported, the viscoelastic properties are com- load permanent deformation (RLPD) tests (Walubita et al.
monly used to characterise the time and temperature depen- 2012). Moreover, some studies (Walubita et al. 2019, Jiang
dence of the stiffness and rheological characteristics of et al. 2016) have also suggested that strain in the DM test is
asphalt binders and asphalt mixtures (Ling et al. 2017, not large enough (as evaluated in the elastic zone) to represent
Polo-Mendoza et al. 2022). Most importantly, the dynamic the real characteristics of PD, and hence, the need for appro-
modulus (|E*|) measured from the dynamic modulus (DM) priate interpretation of the DM results. Likewise, an asphalt
test is a routinely required property for most pavement mixture rutting parameter, namely E*/sinδ, which is analogous

CONTACT Meng Ling menglingup@hotmail.com

© 2023 Informa UK Limited, trading as Taylor & Francis Group
2 M. LING ET AL.

to the asphalt binder rutting parameter (G*/sinδ), has been DM test, on the other hand, is still needed and routinely used to
shown by different researchers (Bhasin et al. 2003, Walubita characterise the viscoelastic properties of asphalt mixtures that
et al. 2016, Alimohammadi et al. 2020, Nemati et al. 2020) are required critical inputs for asphalt pavement M-E model-
to yield conflicting results with respect to predicting the ling, numerical analysis, structural design, and performance
asphalt mixtures’ rutting performance. prediction (Ling et al. 2020, Bai et al. 2021, Ling et al. 2021).
Compared to the cracking resistance evaluation using the DM Thus, it is envisioned that it would be beneficially useful, prac-
test setup, the literature is scantly limited with respect to research tically convenient, and cost-effective if the DM test results
studies conducted on the prediction of the rutting resistance of could be considered as an alternative option to aid with the
asphalt mixtures with the same aging level using the viscoelastic implementation of the performance-based BMD methods as
properties solely (e.g. using dynamic modulus master-curve well as quantifying the cracking and rutting resistance potential
characterisation and development of associated rutting perform- of asphalt mixtures without performing independent standa-
ance parameter and threshold). One recent study reveals that the lone cracking and rutting tests, respectively.
rutting performance measured using the Hamburg wheel track-
ing test (HWTT) could be satisfactorily correlated to the slope of
the DM master-curve (Nemati et al. 2020). However, water is Study objectives
involved in the HWTT, which could also introduce moisture Based on the above background, this study was undertaken
damage in the tertiary phase (i.e. stripping strain) to the output with the primary goal of evaluating the rutting performance
in addition to the accumulated PD. It is also not simple to directly of different asphalt mixture types at the same aging condition
exclude the stripping strain introduced by moisture damage based on the viscoelastic properties of asphalt mixtures.
from the total strain without the aid of mechanistic modelling Additionally, the secondary goal of the study was to explora-
(Martin et al. 2019; Zhang et al. 2022). tory use the generated DM (|E*|) master-curves and the cor-
Like the DM test, the RLPD test is also a widely used M-E per- responding DM-based parameters including the cracking
formance test for quantifying the PD of asphalt mixtures in the performance index and rutting performance index for the
laboratory. In most traditional test setups, the RLPD measures BMD methods. The detailed objectives were as follows:
and uses the repeated load cycles to failure as a quantitative indi-
cator of an asphalt mixture’s resistance to PD. The test has also o Comparative evaluation of the HWTT and RLPD tests and
been adapted as a standard rutting test and incorporated in identifying the most appropriate DM (|E*|) master-curve
many M-E pavement design methods and software programmes, parameter(s) for assessing the rutting resistance potential
such as the AASHTOWare Pavement ME and Texas Pavement of asphalt mixtures.
ME (Hu et al. 2014, Walubita et al. 2020, Seitllari et al. 2021). o Development of a rutting performance index based on the
On this basis, it was therefore considered beneficial in this DM master-curves and the associated threshold for com-
study to include the RLPD test for the PD characterisation of paratively evaluating the rutting resistance and screening
asphalt mixtures in addition to the DM test and HWTT. asphalt mixtures.
In general, there is a need to further investigate how to better o Comparatively assessing the implementation of the rutting
utilise the DM test results including the viscoelastic character- performance index with threshold and the cracking per-
istics to predict the rutting resistance of asphalt mixtures. Fur- formance index with threshold in the BMD approach.
thermore, a DM-based rutting index could be explored to
quantitatively identify the rutting resistance potential of asphalt This paper is organised as follows: details of the asphalt mix-
mixtures during the linear viscoelastic characterisation of tures used in this study, including the data source and sample
asphalt mixtures. On the other hand, asphalt mix-design nowa- preparation, are discussed first followed by the discussion of
days is gradually shifting from the traditional volumetric mix- the DM, HWTT, and RLPD tests, respectively. The data analyses
design approach to the performance-based balanced mix-design and discussions, development of rutting performance index, and
(BMD) methods (Walubita et al. 2012, Elias et al. 2021). The tra- synthesis of the results are thereafter presented followed by the
ditional volumetric mix-design has been shown to have chal- implementation of the rutting and cracking performance indices
lenges and limitations with respect to simultaneously meeting for the BMD approach. The paper then concludes with a sum-
the long-term cracking and rutting performance requirements, mary of key findings and recommendations.
particularly for asphalt mixtures with modification additives
such as polymers, recycled materials, etc. (Coleri et al. 2020,
Al-Khayat et al. 2021, Zhou et al. 2021). Thus, another goal of Study plan and experimental design
this study was to propose a BMD approach based on the visco-
Laboratory data source
elastic properties of asphalt mixtures.
Presently, most of the current asphalt mixture mix-design The asphalt mixtures and corresponding laboratory test per-
methods are largely based on two independent cracking and formance data for this study was selected from the Texas
rutting laboratory tests. In the State of Texas (U.S.A.), for flexible pavements and overlays database, namely the Texas
example, these standalone laboratory tests include the HWTT Data Storage System or Texas (DSS) (Walubita et al. 2017).
or indirect tensile rutting test for rutting performance evalu- The DSS is managed in the Microsoft Access®, comprising of
ation with the indirect tensile cracking test (i.e. IDEAL cracking 115 asphalt pavement sections including laboratory perform-
test) or overlay tester (OT) used for crack-resistance evaluation ance test results for each layer and field performance data.
of asphalt mixtures (Walubita et al. 2012, Zhou et al. 2021). The The viscoelastic performance from the DM test and the rutting
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 3

performance from the HWTT and RLPD tests of the asphalt then compacted in the Superpave gyratory compactor and
mixtures are the subject of this study. then cored and trimmed to a target air void (AV) content of
7 ± 1% except for the PFC at 20 ± 2% in accordance with
Texas laboratory and field requirements (TxDOT 2014). The
Volumetric properties of asphalt mixtures diameter and height of the DM samples were 100 and
As summarised in Table 1, 12 asphalt mixtures with various 150 mm, respectively.
mixture properties (e.g. different performance-grade (PG) After gluing the linear variable displacement transducers
high temperature, aggregate gradations, asphalt-binder con- (LVDTs), the DM samples were then conditioned in a temp-
tent, recycled materials content, etc.) were comparatively eval- erature chamber for 2 h to reach the temperature equilibrium
uated. These asphalt mixtures were plant mixes collected from prior to DM testing. The asphalt mixture performance tester
the field construction sites (Walubita et al. 2017). (AMPT) in strain-controlled mode was used for the DM test
In general, the gradation of Type C is coarser than that of at test temperatures of −10, 4, 21, 38, and 55°C and the test fre-
Type D, which is beneficial for rutting resistance since the coarse quencies of 0.1, 0.5, 1, 5, 10 and 25 Hz, according to AASHTO
gradation provides a tighter stone-on-stone contact. Theoreti- TP 62-03 (Walubita et al. 2017). The sigmodal function in
cally, the higher the PG high temperature, the better the high- Equation 1 was used to generate the DM (namely |E*|) mas-
temperature rutting resistance for the asphalt mixture and vice ter-curves of the asphalt mixtures with respect to the reduced
versa (26). Based on the volumetric properties presented in loading frequency.
Table 1, it would be theoretically expected that the asphalt mix- a
tures on SH 121 and US 271 would yield poor rutting resistance log |E∗ ( fr )| = d + (1)
1
performance since their gradations are finer or porous without
(1 + leb+g log fr )l
RAP, and have relatively higher asphalt-binder contents – volu-
metric characteristics that are however ideal for cracking resist- In Equation 1, fr is the reduced loading frequency whilst d, a,
ance as the name of the CAM (Crack Attenuating Mix) mixture b, g, and l are the regression parameters. The lower asymptote
on SH 121 infers (Walubita et al. 2021, Ling et al. 2022). By con- and upper asymptote of the DM master-curve are 10d and
trast, the asphalt mixtures on SH 358, US 83, and SH 304 should 10a+d , respectively. The inflection point frequency of the mas-
be more rut resistant than the others due to their relatively lower ter-curves was determined using the b parameter whilst the
asphalt-binder contents and higher RAP contents, respectively. relaxation spectrum width and the master-curve shape were
On the other hand, the SMA, which is superiorly known for characterised using the g parameter (Ling et al. 2022).
being both cracking and rutting resistant, is expected to have bet- In a recent study (Ling et al. 2022), it has been shown that
ter rutting performance than the other asphalt mixtures with the inflection point frequency had a stronger correlation to the
similar cracking resistance potential. fatigue cracking resistance than the g parameter. In this study,
one of the major targets was to determine the parameter(s) or
the combinations derived from the viscoelastic properties (i.e. |
DM preparation and testing
E*| master-curves) with higher accuracy to adequately assess
The plant mixes were reheated at the compaction temperature the rutting performance of the asphalt mixtures. As illustrated
in an oven for 2 h without long-term aging. The samples were in Equation 2, the William-Landel-Ferry (WLF) model

Table 1. Mixture properties of selected asphalt materials.

Hwy Mixture type PG grade Binder content (%) Aggregate properties Recycled material information
US 059 Type D 64-22 5.1 Quartzite 20.10% RAP
(Fine-graded)
IH 035 SMA 76-22 6.0 Limestone/Dolomite 12.1% RAP
(Gap-graded)
SH 121 CAM 76-22 7.0 Igneous/Limestone/ Dolomite –
(Fine-graded)
US 271 PFC 76-22 6.0 Sandstone –
(Porous-graded)
Loop 480 Type C 64-22 5.2 – 20% RAP
(Dense-graded)
US 082 Type D 64-22 5.3 – 15% RAP, 5% RAS
(Fine-graded)
FM 1887 TOM 76-22 6.7 Granite/Quartzite/ Limestone 16% RAP, 1% RAS
(Fine-graded)
SH 358 Type C 70-22 4.9 Limestone/Dolomite 19.8% RAP
(Dense-graded)
SH 095 Type D 64-22 5.2 Gravel/Limestone/ Dolomite 3% RAS
(Fine-graded)
SH 007 Type C 64-22 4.9 Limestone/Dolomite –
(Dense-graded)
US 083 Type C 64-22 4.8 Limestone/Dolomite 17% RAP, 3% RAS
(Dense/coarse -graded)
SH 304 Type C 64-22 4.8 Limestone/Dolomite 17% RAP, 2.5% RAS
(Dense-graded)
Hwy = Highway; CAM = Crack Attenuating Mix; SMA = Stone Mastic (Matrix) Asphalt; PFC = Porous Friction Course mix; TOM = Thin Overlay Mix; Type C = Texas dense-
graded mix; Type D = Texas dense-fine graded mix; RAP = Recycled Asphalt Pavement; RAS = Recycle Asphalt Shingles.
4 M. LING ET AL.

(Walubita et al. 2011) was used as the time-temperature shift water and laterally confined. These HWTT conditions, par-
function. ticularly water submergence, have the undesired potential to
introduce additional permanent strain arising from moisture
−C1 (T − TR )
damage and stripping of the asphalt mixture that ultimately
aT = 10 2 + (T − TR )
C (2) contributes to the RD accumulated by the samples (Walubita
In Equation 2, aT is the time-temperature shift factor; C1 and et al. 2019). Therefore, these effects and differences need to
be cautiously considered when analyzing, synthesising, and
C2 are the regression parameters; T is the test temperature; and
TR is the reference temperature, which was 21°C in this study. interpretating the test results, particularly in comparative PD
studies of this nature.

RLPD preparation and testing

Laboratory results and analysis
The DM sample preparation procedures were applied when
fabricating the RLPD samples. The final AVs and dimensions DM test results and |E*| master-curves
are the same as those of the DM sample (TxDOT 2014). Note The DM master-curves based on an average of three sample
that having the same DM and RLPD sample dimensions could replicates, are presented in Figure 1, with the correspond-
partly help to simplify the sample preparation/setup and ing regression coefficients and ANOVA table are listed in
reduced the undesired variations when comparatively analyz- Tables 2 and 3, respectively. As theoretically expected of
ing the two independent tests in this study. The samples the viscoelastic materials, Figure 1 shows a progressive
were preconditioned in a temperature-controlled chamber decay in the DM as the frequency decreased whilst the
for 2 h to reach the test temperature prior to the RLPD testing temperature was increased, i.e. moving towards the left
(Walubita et al. 2017). side of the figure from 1 Hz. With moving to the lower
A repeated compressive haversine loading waveform was temperature domain and higher frequency regime (i.e.
applied in the RLPD test to measure the accumulative PD of right side of 1 Hz in the figure), respectively, a stiffening
the asphalt mixtures. For the work presented in this paper, effect is visually noted with a significant increase in the
the RLPD test was conducted in a stress-controlled loading |E*| of the asphalt mixtures.
mode with a 69 kPa load using the Universal Testing Machine Looking at the high-frequency low-temperature domain
(UTM) without any confining pressure at a test temperature (i.e. right side of the figure from 1 Hz), Figure 1 shows
and loading frequency of 50°C and 1 Hz, respectively, based that apart from the SH 121 and US 271 asphalt mixtures,
on the procedure suggested (Walubita et al. 2017). The test the |E*| values of the asphalt mixtures are comparable with
was stopped when the repetitive-cyclic loading reached their master-curves nearly overlapping each other. In the
10,000 cycles or the accumulated permanent strains reached low-frequency high-temperature domain, however, the
25,000 microstrains, whichever came first. In this study, the difference in the master-curves is distinctively evident, with
key output obtained from the RLPD test was the accumulative US 83 and SH 121 quantitatively exhibiting the highest
PD versus load cycles. and lowest modulus values, respectively. This |E*| behav-
ioural response-trend is consistent with the mix-design volu-
metrics listed in Table 1. For instance, both SH 121 (fine-
HWTT preparation and testing
graded) and US 271 (porous-graded) comprised of a high
The HWTT was comparatively used in this study to evaluate content PG 76-22 asphalt binder and coupled with the low
the rutting resistance of the asphalt mixtures based on the density of the PFC mixture (US 271), partly contributed to
Tex-242-F test procedure (TxDOT 2021). A 705 N vertical their viscoelastic behaviour and the |E*| response trends
load at a wheel speed of 52 passes/min (i.e. 0.87 Hz) up to observed in Figure 1. The Type C mixture for US 83, on
20,000 load cycles, in a 50 ± 1°C water bath, was applied to the other hand, contained 17% RAP and 3% RAS with a
150 mm diameter and 62.5 mm thick samples. Like the DM relatively low asphalt binder content and dense/coarse aggre-
and RLPD samples, the same plant-mix reheating process gate gradation that possibly contributed to its high |E*|
was followed, the moulding and sample fabrication was values (Walubita et al. 2017).
adhered to Tex-242-F (TxDOT 2021). The HWTT was termi- To comparatively evaluate the effects of the DM master-
nated when either the accumulative rut depth (RD) reached curve on the rutting performance, two DM parameters, namely
12.5 mm or the maximum load passes of 20,000 cycles was β and γ that controls the rheological properties, were of interest
reached, whichever came first. The key parametric output gen- in this study. The inflection point frequency 10(-β/γ) has been
erated form the HWTT is the accumulative RD as a function of shown to have a high correlation with fatigue cracking resist-
load cycles. ance, which is the boundary line between the viscous and elastic
As can be inferred from above, one key difference between part of asphalt mixtures (Ling et al. 2022). The greater the
the HWTT and the other two laboratory tests are the test con- inflection point frequency and the β parameter normally
ditions (e.g. dry versus water submergence, boundary con- means more viscous components and better fatigue resistance
ditions, sample configuration/sizes, etc.). When comparing potential for the asphalt mixture. On the other hand, the par-
these three tests, one of the most critical differences is that ameter g controls the shape of the master-curve and width of
unlike the DM and RLPD tests, the HWTT is conducted in the relaxation spectrum, and quantitatively increases as the
wet conditions with the samples completely submerged in aging level increases or the content of the recycled material
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 5

Figure 1. DM (|E*|) master-curves of asphalt mixtures at 21°C.

increases; with a corresponding decrease in the cracking resist- et al. 2017). Like the flow number test, it should be noted that
ance of the asphalt mixture (Ling et al. 2022). the purpose of the RLPD test in this study was to quantify the
However, it has been shown that the g parameter solely PD resistance of the asphalt mixtures. For this study, the accu-
could not adequately screen and distinctively evaluate the mulated PD was used as a parametric indicator of the RD and
cracking resistance potential of asphalt mixtures subjected to was quantitatively measured in terms of the accumulated per-
the same aging condition (Ling et al. 2022). Likewise, it was manent microstrain, namely εp. The generated RLPD test
still unknown as to which viscoelastic parameter and/or the results based on the average of three replicate samples are pre-
combination is most effective at distinguishing, screening, sented in Table 4.
and discreetly ranking the rutting resistance potential of asphalt Based on the DSS protocol (Walubita et al. 2017), the
mixtures subjected to the same aging condition. As discussed in suggested maximum εp for satisfactory resistance against PD
the subsequent sections of this paper, the DM master-curve is 10,000 microstrains. From Table 4, it is noted that the
parameters, |E*| at multiple loading frequencies, and the rutting asphalt mixtures on US 271 and SH 95 exceeded this limit,
parameter |E*|/sinδ were comparatively assessed for their rut- whilst the rest of the other asphalt mixtures are satisfactorily
ting performance prediction of asphalt mixtures. below the threshold. The asphalt mixture on US 271 is a
PFC mixture, which is porous graded and has an AV content
of about 20 ± 2%. It is theoretically expected that the asphalt
RLPD test results and permanent deformation (PD) mixtures with low density would have relatively poor rutting
resistance due to the low stone-on-stone contact matrix,
The RLPD test was performed to determine the accumulative which along with its high AV and densification under RLPD
PD to 10,000 load cycles following the DSS protocol (Walubita repeated compressive loading partly contributed to its poor
rutting performance (i.e. high εp value exceeding 10,000 micro-
Table 2. Regression coefficients of the DM (|E*|) master-curves.
strains) (Walubita et al. 2017, Walubita et al. 2019). In prac-
Pavement section δ α β γ R 2 (%)
tice, however, PFC mixtures are rarely used as pavement
US 59 1.44 3.02 −0.78 −0.63 98
structural layers; their primarily function is serving as sur-
IH 35 2.01 2.46 −0.65 −0.59 97 face-water draining, skid-resistant enhancing, and noise-
Loop 480 1.7 2.73 −0.89 −0.56 98 reduction surface layers (TxDOT 2014). For the Type D mix-
SH 121 1.31 3.16 −0.19 −0.6 97
US 82 0.81 3.78 −1.11 −0.39 99
ture on SH 95, its mix-design volumetrics, namely fine aggre-
FM 1887 2.18 2.2 −0.56 −0.71 99 gate gradation, low PG high-temperature (i.e. PG 64-22), and
SH 358 1.47 2.94 −1.23 −0.43 99 relatively low recycled materials content (i.e. 3% RAS), are
US 271 1.97 2.03 −0.25 −0.74 98
SH 95 1.54 2.76 −0.67 −0.65 98
assumed to have partly contributed to its poor PD perform-
SH 7 1.37 3.13 −0.73 −0.49 99 ance, leading to a high εp value exceeding 10,000 microstrains
US 83 1.93 2.57 −0.95 −0.38 99 (Walubita et al. 2017, Walubita et al. 2019). Note that all the
SH 304 1.62 2.92 −0.77 −0.43 98
asphalt mixtures in this study were subject to 10,000 load
6 M. LING ET AL.

Table 3. ANOVA results for the DM master-curve parameters.

SUMMARY
Group Count Sum Average Variance
δ 12 19.35 1.6125 0.142366
α 12 33.7 2.808333 0.218942
β 12 −8.78 −0.73167 0.093942
γ 12 −6.6 −0.55 0.015382
ANOVA
Source of variation SS df MS F P-value F crit
Between Groups 106.3336 3 35.44455 301.2503 2.45E-29 2.816466
Within Groups 5.176958 44 0.117658
Total 111.5106 47

Table 4. RLPD test results of asphalt mixtures at 10,000 load cycles. performance. Nonetheless, the Type C asphalt mixture on
Coefficient of Variation SH 304 with a RD of 11.34 mm still satisfactorily met the
Accumulated Permanent Deformation, εp (CoV) 12.5 mm pass-fail screening criteria (TxDOT 2021).
Highway (microstrains) (%) It was theoretically unexpected that the asphalt mixtures on
US 059 6485 3.38 SH 121 and SH 304 would succumb to the highest RD in the
IH 035 4428 47.85
Loop 480 3723 14.22 HWTT in Table 5, whilst, by contrast, the same SH 304 exhib-
SH 121 9536 29.86 ited the best PD resistance performance in the RLPD test in
US 082 1952 11.23 Table 4. It is noted that the stripping strain arising from the
FM 1887 3716 1.76
SH 358 2127 69.53 moisture damage is not detectable in the dry-conditioned
US 271 17188 33.68 RLPD test (i.e. no tertiary stage observed); however, these
SH 095 11663 27.88 were observed in the HWTT for the asphalt mixtures such
SH 007 4275 13.74
US 083 1938 29.87 as US 83, SH 121, SH 304, etc. Partially because of this, no sat-
SH 304 1771 43.69 isfactory correlations were observed between the HWTT and
RLPD test results from 5000 to 20,000 load cycles. Therefore,
when using the HWTT to comparatively assess the PD resist-
cycles in the RLPD test – that is the RLPD test was stopped and ance of asphalt mixtures, stripping strain induced by moisture
terminated at 10,000 load cycles. damage and the related boundary conditions should be well
accounted for.
To focus purely on the PD resistance evaluation (without
HWTT results and rut depth (RD) moisture effects) and simplify the analysis in this study, the
RLPD test was considered as the indicative rutting perform-
The HWTT results, namely the accumulated RDs at 5000, ance test for the subsequent analyses with the DM test in
10,000, 15,000, and 20,000 load cycles, are shown in Table 5. this paper. Conversely, another option for future comparative
Looking at the 20,000 load cycles, it is evident that the asphalt follow-up studies is to explore the possibility of conducting
mixture (Type D) on US 82 with a RD of 2.95 mm exhibited both the DM and RLPD tests using moisture preconditioned
superior rutting resistance performance whilst the CAM (SH samples for evaluation of moisture susceptibility and rutting
121) and Type C (SH 304) were the poorest performers, with resistance – or alternatively, running the HWTT under dry
double-digit RD values. The recycled materials, namely 15% conditions without water and modified boundary conditions.
RAP and 5% RAS shown in Table 1, were assumed to have
contributed to the US 82 asphalt mixture’s superior rutting
performance. In the case of the SH 121 and SH 304 asphalt
mixtures, however, moisture damage and stripping were Comparative evaluation of the DM-RLPD data
hypothesised to have contributed to their poor HWTT Different DM master-curve parameters were comparatively
assessed to determine how the parameters were related to
Table 5. HWTT results of the asphalt mixtures.
the accumulative RD measured from the RLPD test and there-
Rut depth at different load cycles (mm)
after, select the appropriate correlative parameter (s) or the
Highway 5000 10000 15000 20000
combinations based on their viscoelastic characteristics. It
US 059 2.60 3.40 3.90 4.30
can be seen from Figures 2 and 3 that β is positively correlated
IH 035 2.56 3.73 4.94 5.96 to εp and vice versa for γ, i.e. it is negatively correlated to εp,
Loop 480 3.19 3.93 4.38 4.76 respectively. The parameter γ indicates satisfactory correlation
SH 121 6.70 11.70 15.80 18.00
US 082 2.29 2.61 2.84 2.95
with a coefficient of correlation (R2) of 70.2%, which inciden-
FM 1887 3.10 3.96 5.84 8.38 tally is higher than the R2 value of about 27% that was pre-
SH 358 3.36 4.19 5.19 6.35 viously found by these authors when correlating the γ
US 271 2.18 2.77 3.24 3.79
SH 095 3.60 4.44 5.17 5.99
parameter to the fatigue cracking resistance of the same
SH 007 4.71 6.02 7.25 8.17 asphalt mixtures (Ling et al. 2022).
US 083 2.45 4.05 6.10 9.40 Based on the results in Figures 2 and 3, a larger and
SH 304 6.46 8.08 10.12 11.34
wider relaxation spectrum (indicated by the parameter γ)
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 7

Figure 2. Inflection point (β) versus the εp parameter.

Figure 3. Relaxation spectrum (γ) versus the εp parameter.

was more effective for PD resistance characterisation of the mixtures – and hence, the need for an integrated composite
asphalt mixtures than it does with respect to fatigue crack- viscoelastic parameter.
ing. The degree of aging and recycled material content Figures 5 and 6 graphically illustrates the relation of the |E*|
increase or the asphalt mixtures become less temperature values to the εp. Figure 5 comprises of the |E*| obtained at the
dependent, the relaxation spectrum (γ) becomes wider. inflection point frequency. Figure 6, on the other hand, rep-
This leads to an increase in the rutting resistance but vice resents the |E*| obtained at the reduced frequency correspond-
versa for the fatigue cracking resistance, respectively. ing to the RLPD test conditions.
In addition, the inflection point frequency of the |E*| The poor correlation and small R2 value of 4.82% in
master-curves in Figure 4 with an R2 value of 56.7% Figure 5 shows that there was no definitive correlation
appeared not as sensitive to rutting resistance as was between the |E*| values at the inflection point frequency
observed by these authors in their previous study (Ling (10-β/γ) from the DM test and the εp from the RLPD test
et al. 2022) on the fatigue cracking performance of the (Walubita et al. 2022). Therefore, the |E*| values correspond-
same asphalt mixtures that registered an R2 value of ing to the RLPD test conditions were instead explored. As
92.2%. Theoretically, asphalt mixtures with quantitatively shown in Figure 6, the regression relationship obtained was
greater inflection point frequency normally have larger vis- satisfactorily acceptable, with an R2 of 76.7%. This was theor-
cous components and better cracking resistance potential, etically expected, since it is well-known that stiffer asphalt
but poor PD resistance, respectively. Thus, considering the mixtures with higher modulus values generally have better
|E*| master-curve parameters discussed above, the γ par- PD resistance and vice versa. This means that the |E*| at
ameter or the relaxation spectrum width could be reasonably 50°C and 1 Hz from the DM test would yield reasonably
used to comparatively characterise the PD resistance poten- acceptable prediction accuracy for the εp at 50°C and 1 Hz
tial. However, a single parametric value is still not precise from the RLPD test. Furthermore, these correlative results
enough to adequately evaluate the PD resistance of asphalt highlight that in comparative studies of this nature, it is
8 M. LING ET AL.

Figure 4. Inflection point frequency versus εp at 10,000 load cycles.

Figure 5. |E*| at 10-β/γ (21°C) versus εp at 10,000 load cycles.

Figure 6. |E*| at RLPD test condition versus εp at 10,000 load cycles.

important to cautiously select an appropriate DM tempera- As illustrated in Figure 7, the |E*|/sinδ parameter at the
ture and loading frequency corresponding to the laboratory RLPD test condition was also evaluated for correlation to
rutting test conditions (e.g. RLPD test in this study) to obtain the εp parameter. In determining the |E*|/sinδ parameter,
reasonable correlations. the phase angle (δ) at the reduced loading frequency was
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 9

Figure 7. |E*|/sinδ at RLPD test condition versus εp at 10,000 load cycles.

computed based on the procedure derived from the visco- smaller in magnitudes than those at the inflection point
elastic theory (Ling et al. 2022). However, the poor corre- frequency, except for the asphalt mixture on US 82.
lation with an R2 of about 20% for the |E*|/sinδ parameter As shown in Figure 8, the regression relationship
was lower than that for |E*| in Figure 6, indicating that between the VERindex and εp (i.e. R2 = 82.3%) was relatively
the parameter |E*|/sinδ might not a good rutting index strong and satisfactorily acceptable (40). With 10,000
for the asphalt mixtures (Walubita et al. 2022). That is, microstrains as the terminal PD strain level (Walubita
the |E*|/sinδ parameter is statistically associated et al. 2017, Walubita et al. 2019), the area above the red
with some uncertainties and low accuracy in terms of pre- line in the figure is the ‘fail zone’ and the area below the
dicting the PD and rutting performance of the asphalt red line is the ‘pass zone’. On this basis, the minimum
mixtures. threshold for VERindex was therefore determined to be
From the above results (i.e. Figures 3 and 6), the two most 55.2. For practical purposes, however, 55 was used, i.e.
sensitive parameters to the rutting resistance of the asphalt VERindex ≥ 55. Theoretically, a higher VERindex value means
mixtures were observed to be the γ parameter (R2 = 70.2%) that the asphalt mixture has accumulated a lower RD
and |E*| at 50°C and 1 Hz (R2 = 76.7%), respectively. This depth and has better rutting resistance potential and vice
means that the slope of the |E*| master-curve and |E*| at the versa.
RLPD test conditions should be mathematically integrated On a comparative note, Figure 8 shows that the
and merged to form a single composite rutting parameter to asphalt mixtures on US 271 and SH 95 did not meet the pro-
achieve a higher prediction accuracy. After several mathemat- posed VERindex ≥ 55 criteria. For US 271, the same reasons
ical formulations, a novel composite rutting parameter, named previously advanced with respect to the |E*| master-curves
as the Viscoelastic Rutting Index (VERindex), was derived as in Figure 1 still holds here as well, namely porous aggregate
expressed in Equation 3. gradation, low density, and high asphalt-binder content; and
would be subjectively recommended for redesigning.
VERindex Other than in the HWTT (Table 5), the Type D mixture
on SH 95 had also performed relatively poorly in both DM
|E∗|RLPDloading condition × (|E∗|inflectionpoint − |E∗|1/1000∗inflectionpoint ) (Figure 1) and RLPD (Table 4) tests, being the third
=
(finflectionpoint − f1/1000∗inflectionpoint ) × 10000 and second poorest performer, respectively. Likewise,
(3) the same reasons previously advanced for the SH 95
asphalt mixture’s poor performance in the RLPD test
In Equation 3, |E∗|inflection point and |E∗|1/1000∗inflection point are the holds, namely fine aggregate gradation (Type D), low PG
DM (|E*|) at the inflection point frequency and 1/1000*inflec- asphalt binder (i.e. PG 64-22), and relatively low recycled
tion point frequency, respectively, finflection point and materials content (i.e. 3% RAS).
f1/1000∗inflection point are the reduced loading frequencies at the
inflection point and 1/1000* inflection point, respectively.
Application of performance index from viscoelastic
The purpose of using 1/1000*inflection point frequency was
characteristics for BMD approach
to characterise the slope of master-curve, which plays an
important role in the rutting resistance evaluation. In In this section of the paper, the viscoelastic characteristics of
Equation 3, the numerical factor 10,000 was arbitrarily used the asphalt mixtures were explored for direct application and
to mathematically normalise the VERindex for the practical sim- implementation in the BMD approach. For characterising
plicity of interpretation, screening, and ranking of the asphalt and quantifying the asphalt mixture’s performance, key dis-
mixtures. Note that the corresponding reduced loading tresses addressed in most BMD methods are rutting and crack-
frequency and |E*| values at the RLPD test conditions were ing (mainly fatigue cracking). Conceptually, one of the
10 M. LING ET AL.

Figure 8. VERindex versus εp at 10,000 load cycles.

primary goals of the BMD method is to optimally balance the From Figure 9, it can be observed that five asphalt mixtures
rutting and cracking resistance of an asphalt mixture by are above the NOT ≥ 150 threshold line. The CAM (SH 121),
adjusting the asphalt-binder content and/or other volumetric SMA (IH 35), and TOM (FM 1887) are traditionally designed
mix-design factors until both the rutting and cracking per- to be crack-resistant (26). Therefore, their superior cracking per-
formance criteria are simultaneously met (Walubita et al. formance was theoretically expected and is in line with their mix-
2012, Zhou et al. 2021). design volumetrics in Table 1. For the PFC (US 271) and Type D
For fatigue cracking resistance characterisation, the Visco- (US 59) mixtures, their mix-design volumetrics, namely high-
elastic Cracking Index (VECindex) developed in a previous quality sandstone aggregates, high asphalt-binder content
study was jointly used for the BMD application in this study (6.0%), and high PG asphalt-binder (PG 76-22), partially con-
(Ling et al. 2022). The VECindex was also derived from the tributed to their superior cracking resistance performance in
|E*| master-curves based on the number of repetitive load the OT as evident in Figure 9. By contrast, the rest of the asphalt
cycles to crack failure (NOT) measured in the OT (for the mixtures, mostly the coarse/dense graded Type C in Table 1, did
same asphalt mixtures listed in Table 1) as illustrated in not meet the tentatively adapted NOT ≥ 150 criteria, partly attrib-
Figure 9. The VECindex model, as formulated in (Ling et al. uted to their insufficient mix-design volumetrics including poor
2022), is shown in Equation 4. aggregate quality (i.e. limestone, dolomite, etc.), coarse-to-dense
aggregate gradation, low asphalt-binder content (4.8∼5.3%), low
b PG asphalt-binder (PG 64-22), and recycled materials (i.e. >15%
−
VECIndex = −10 g ∗g (4) RAP, ∼3% RAS, etc.).
After the rutting and cracking thresholds were tentatively
The VECindex versus NOT for the asphalt mixtures in Table 1 as determined, a performance space diagram (PSD) using the
graphically plotted in Figure 9 shows a strong linear correlation VECindex and VERindex, as exemplified in Figure 10, was then
and an R2 value of 95.3%. The traditional NOT pass-fail threshold used to diagrammatically characterise the cracking and rutting
for satisfactory cracking performance in the OT is 300 (TxDOT performance of the asphalt mixtures, with their corresponding
2014). However, NOT ≥ 300 was considered too conservative for minimum performance criteria.
this study, with some asphalt mixtures such as the Type D on In summary, the following logical basis was followed for the
US 59 with NOT = 255 exhibiting satisfactory cracking perform- BMD approach:
ance after over 5 years of field service. The SMA (IH 35), TOM
(FM 1887), Type D (NOT = 116 on SH 95), and Type C (NOT =
126 on Loop 480) mixtures also performed satisfactorily after 5 (1) The VECindex (cracking) developed in a previous study
years of field service (Walubita et al. 2017). On this basis and (34) was correlated to the Texas overlay test results and
in line with the findings/proposals (Walubita et al. 2012), NOT- failure criterion (i.e. number of load cycle to failure greater
≥ 150 was tentatively adapted as the pass-fail criteria for this than 150). From this, the failure criterion for VECindex was
study. The minimum VECindex corresponding to NOT ≥ 150 in determined to be VECindex ≥ 0.05, and this is how the ver-
Figure 9 was determined to be 0.04. For practical convenience tical red dash line was determined in Figure 10.
and as an extra precautionary safety net, 0.05 was selected and (2) The VERindex (rutting) developed in this study was corre-
used instead, i.e. VECindex ≥ 0.05. Quantitatively, the higher the lated to the repeated load permanent deformation (RLPD)
VECindex value in magnitude, the better the cracking resistant test and failure criterion (i.e. accumulative permanent
the asphalt mixture is and vice versa. strain greater than 10,000 microstrain). From this, the
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 11

Figure 9. VECindex versus NOT from the OT.

Figure 10. PSD based on viscoelastic characteristics.

failure criterion for VERindex was determined to be o Zone B = flexible asphalt mixtures with acceptable labora-
VERindex ≥ 55, and this is how the horizontal red dash tory cracking resistance, but prone to rutting problems.
line was determined in Figure 10. o Zone C = stiff asphalt mixtures with adequate laboratory rut-
ting resistance, but susceptible to cracking issues.
o Zone D = poor asphalt mixtures needing to be modified and/
In the PSD in Figure 10, the general zonal description is as
or redesigned.
follows:

o Zone A = durable well balanced design asphalt mixtures with The asphalt mixtures in top right corner, namely Zone A,
satisfactory laboratory rutting and cracking resistance are the durable well balanced designed mixtures, with satisfac-
performance. tory cracking and rutting resistance performance. These are
12 M. LING ET AL.

superior asphalt mixtures that would be typically used on high cracking (VECindex) performance thresholds were proposed
traffic-volume highways and/or surface layers of pavements. and used in the PSD to comparatively evaluate the perform-
The asphalt mixtures that are resistant to cracking but poten- ance of asphalt mixtures and application of the BMD method.
tially susceptible to rutting issues would be in the bottom right The major conclusions and analytical efforts are summar-
corner of the performance space diagram, namely Zone ised as follows:
B. These asphalt mixtures, which are often with high
asphalt-binder content, would be typically used to mitigate o The HWTT and RLPD tests were performed and compared
cracking in pavement structures. The asphalt mixtures in the for PD performance. However, no strong correlations
top left corner of the PSD (namely Zone C) are the stiff rut- between these two tests were identified, suggesting that
resistant mixtures, but with potential susceptibility to cracking the boundary conditions and stripping strain induced by
issues. These asphalt mixtures are denoted as stiff mixtures and the moisture damage in the HWTT were possibly not neg-
would be typically used as structural support layers to mitigate ligible for the PD analysis. This highlights that in com-
rutting in a pavement structure. Lastly, the asphalt mixtures in parative studies of this nature, the test conditions need
the left bottom corner of the PSD, namely Zone D, are the poor to be well accounted for when performing the compara-
mixtures with inadequate rutting and cracking resistance. tive analysis.
These asphalt mixtures should be modified, redesigned, or o The γ parameter from the DM master-curves was found to
otherwise used only on very low-traffic and/or temporary be more sensitive to the rutting performance of the asphalt
roads. mixtures than the β parameter and the inflection point
As evident in Figure 10, this preliminarily proposed DM- frequency; indicating that the relaxation spectrum width
based PSD could be used to simultaneously identify the fatigue or the slope of DM master-curve is sensitive to the rutting
cracking and rutting susceptibility of asphalt mixtures in the performance of the asphalt mixtures.
laboratory. The PSD could potentially be used to compare o The DM at the RLPD test temperature and frequency was
and screen asphalt mixtures with different volumetric mix- also found viable for characterising PD performance of
designs and laboratory aging levels solely based on the DM the asphalt mixtures at the RLPD test conditions.
test. That is the DM test alone could be promisingly used as o A novel viscoelastic rutting index (VERIndex) incorporating
an alternative test option for BMD approach with reasonably the slope of the |E*| master-curve and the |E*| at the
acceptable repeatability and relatively low data variability. RLPD test condition, was successfully developed. There-
Note however, the proposed PSD for the BMD method and after, the associated rutting (VERindex) and fatigue crack-
the accompanying tentative rutting-cracking pass-fail criteria ing (VECindex) thresholds were quantitatively proposed
in Figure 10 are only preliminary with the need for more fol- in terms of both laboratory rutting and cracking
low-up studies to supplement and substantiate the results pre- performance.
sented herein. More laboratory work including moisture o A PSD for implementing the BMD method was developed
sensitivity and water damage evaluation along with field vali- along with the accompanying rutting and cracking pass-
dation are still warranted. Ultimately, this means that the ten- fail criteria. This PSD, which is based solely on the DM
tatively proposed PSD criteria in this paper may still be subject test, could be potentially used to comparatively evaluate
to modifications, enhancements, and refinements in future fol- the rutting and cracking performance of asphalt mixtures
low-up studies. It should also be noted that this study does not for subsequent volumetric mix-design adjustments,
intend to replace the independent cracking and rutting tests modifications, and screening of the asphalt mixtures.
using the DM test but provides an alternative novel approach
to comparatively evaluate and quantify the cracking and rut- Generally, the VERIndex could be used at the viscoelastic
ting performance of asphalt mixtures using viscoelastic prop- properties characterisation stage to evaluate the rutting resist-
erties, and further implement the DM-based viscoelastic ance potential of asphalt mixtures for material screening and
parameters into the BMD method. selection. It would also be interesting to use complex viscosity
in future studies for rutting performance prediction. The
mechanistic DM-based PSD could be utilised to comparatively
Summary and future work
evaluate the rutting and cracking performance of asphalt mix-
This study developed a rutting performance parameter based tures and contribute to implementing the BMD method. Like-
on the DM (|E*|) master-curve and viscoelastic characteristics, wise, the DM laboratory test has shown promising potential to
namely the viscoelastic rutting index (VERindex), to quantitat- be used as the sole BMD test with respect to simultaneously
ively evaluate the rutting resistance potential of asphalt mix- characterise the rutting and cracking resistance of asphalt
tures. The DM master-curves were constructed for 12 mixtures.
asphalt mixture types subjected to the same aging level. The For future recommendations, more laboratory test data for
HWTT and RLPD tests were also conducted to evaluate the additional asphalt mixtures including the DM and RLPD, and
rutting resistance of the same asphalt mixtures. The accumu- other widely used tests in the BMD approach (e.g. IDEAL-Rut-
lated PD from the RLPD test was used to quantify the rutting ting, IDT, complex viscosity, etc.) are warranted to further
resistance of the asphalt mixtures. Thereafter, elaborate ana- substantiate and validate the two performance indices and cor-
lyses were performed to develop regressions between various responding thresholds proposed in this study. Additionally,
viscoelastic parameters and the rutting resistance potential of more field performance data is recommended to aid in refining
the asphalt mixtures. The rutting (VERindex) and fatigue the proposed cracking and rutting criteria. Moisture sensitivity
 INTERNATIONAL JOURNAL OF PAVEMENT ENGINEERING 13

and water damage evaluation tests along with field validation Ling, M., et al., 2017. Time-temperature-aging-depth shift functions for
are also warranted. Nonetheless, this study adds to the state- dynamic modulus master curves of asphalt mixtures. Construction
and Building Materials, 157, 943–951.
of-the-art literature enrichment by providing an analytical Ling, M., et al., 2020. Investigation of fracture behavior of asphalt mixture
basis for analyzing the DM (|E*|) master-curves for evaluating composite using energy-based approach. Composites Part B:
the rutting resistance based on the viscoelastic characteristics Engineering, 181, 107324.
and the VERIndex that inherently contributes to implementing Ling, M., et al., 2021. A mechanistic framework for tensile fatigue resist-
the BMD method. ance of asphalt mixtures. International Journal of Fatigue, 151, 106345.
Ling, M., et al., 2022. Evaluating fatigue resistance based on viscoelastic
properties of asphalt mixtures. International Journal of Pavement
Engineering, 1–16. DOI:10.1080/10298436.2022.2126976
Disclosure statement Lytton, R.L., Luo, X., Saha, S., Chen, Y., Deng, Y., Gu, F., & Ling, M.
No potential conflict of interest was reported by the author(s). (2019). Proposed enhancements to pavement ME design: improved con-
sideration of the influence of subgrade and unbound layers on pavement
performance (No. NCHRP Project 01-53).
Martin, A.E., et al., 2019. Evaluating the effects of recycling agents on
Funding asphalt mixtures with high RAS and RAP binder ratios. NCHRP
This work was supported by Texas Department of Transportation [grant Research Report, (927).
number 6658] and Beijing University of Civil Engineering and Architec- Nemati, R., Dave, E.V., and Sias, J.E., 2020. Development of complex
ture Fund [grant number X23011]. modulus-based rutting index parameter for asphalt mixtures.
Journal of Transportation Engineering, Part B: Pavements, 146 (2),
04020026.
Polo-Mendoza, R., et al., 2022. Ultraviolet ageing of bituminous materials:
ORCID a comprehensive literature review from 2011 to 2022. Construction and
Jie Ji http://orcid.org/0000-0003-2764-1986 Building Materials, 350, 128889.
Luis Fuentes http://orcid.org/0000-0002-7811-8821 Seitllari, A., Lanotte, M., and Kutay, M.E., 2021. Recommended cali-
bration procedure of MEPDG asphalt rutting model using repeated
load permanent deformation and confined and unconfined dynamic
modulus data. Journal of Transportation Engineering, Part B:
References Pavements, 147 (1), 04020079.
Al-Khayat, H., et al., 2021. Evaluation of the Minnesota asphalt mixtures Sivasubramaniam, S. and Haddock, J.E., 2005. Validation of Superpave
based on balanced mix-design approach. Journal of Transportation mixture design and analysis procedures using the NCAT test track.
Engineering, Part B: Pavements, 147 (3), 04021045. Publication FHWA/IN/JTRP-2005/24. Washington, DC: Federal
Alimohammadi, H., et al., 2020. Field and simulated rutting behavior of Highway Administration.
hot mix and warm mix asphalt overlays. Construction and Building Solatifar, N., et al., 2019. Development of dynamic modulus master curves
Materials, 265, 120366. of in-service asphalt layers using MEPDG models. Road Materials and
Apeagyei, A.K. (2011, November). Correlating rutting with dynamic Pavement Design, 20 (1), 225–243.
modulus of asphalt concrete. In: Proceedings of the institution of civil Tong, J., et al., 2022. A criterion of asphalt pavement rutting based on the
engineers-transport (Vol. 164, No. 4, pp. 241–249). London: Thomas thermal-visco-elastic-plastic model. International Journal of Pavement
Telford Ltd. Engineering, 23 (4), 1134–1144.
Bai, T., et al., 2021. Viscoelastic modelling of an asphalt pavement based TxDOT, 2014. Texas department of transportation. Standard specifications
on actual tire-pavement contact pressure. Road Materials and for construction and maintenance of highways, streets, and bridges.
Pavement Design, 22 (11), 2458–2477. Austin, TX.
Bhasin, A., Button, J.W., and Chowdhury, A., 2003. Evaluation of simple TxDOT, 2021. Texas department of transportation. Test procedure for
performance tests on HMA mixtures from the south central United the Hamburg wheel-tracking test, TxDOT designation: Tex-242-F.
States. Rep. No: FHWA/TX-03/9-558-1. Washington, DC: Federal Austin, TX.
Highway Administration. Walubita, L.F., et al., 2016. The Hamburg rutting test–effects of HMA
Birgisson, B., et al., 2004. The use of complex modulus to characterize the sample sitting time and test temperature variation. Construction and
performance of asphalt mixtures and pavements in Florida (No. Final Building Materials, 108, 22–28.
Report). Walubita, L.F., et al., 2017. Texas flexible pavements and overlays: year 5
Bueno, L.D., et al., 2022. Asphalt pavement design optimisation: a case report-complete data documentation (No. FHWA/TX-15/0-6658-3).
study using viscoelastic continuum damage theory. International College Station, TX: A&M Transportation Institute.
Journal of Pavement Engineering, 23 (4), 1070–1082. Walubita, L.F., et al., 2019. The Hamburg Rutting Test (HWTT) alterna-
Coleri, E., Sreedhar, S., and Obaid, I.A., 2020. Development of a balanced tive data analysis methods and HMA screening criteria. International
mix design method in Oregon (No. FHWA-OR-RD-21-03). Salem, OR: Journal of Pavement Research and Technology, 12 (1), 110–116.
Dept. of Transportation. Research Section. Walubita, L.F., et al., 2020. Correlating the asphalt-binder high-tempera-
Elias, N.G., et al., 2021. Local agency transition to balanced mix design. ture properties (DSR) to HMA permanent deformation (RLPD) and
International Journal of Pavement Engineering, 23, 1–11. field rutting: a laboratory-field study. Construction and Building
Hu, S., Zhou, F., and Scullion, T., 2014. Development of Texas mechanis- Materials, 262, 120761.
tic-empirical flexible pavement design system (TxME) (No. FHWA/TX- Walubita, L.F., et al., 2021. Correlations and preliminary validation of the
14/0-6622-2). laboratory monotonic overlay test (OT) data to reflective cracking per-
Jiang, J., et al., 2016. Developing an optional multiple repeated load test to formance of in-service field highway sections. Construction and
evaluate permanent deformation of asphalt mixtures based on axle Building Materials, 267, 121029.
load spectrum. Construction and Building Materials, 122, 254–263. Walubita, L.F., et al., 2022. Correlating the asphalt-binder MSCR test
Kassem, H., Chehab, G., and Najjar, S., 2020. Effect of asphalt mixture results to the HMA HWTT and field rutting performance. Journal of
components on the uncertainty in dynamic modulus mastercurves. Transportation Engineering, Part B: Pavements, 148 (3), 04022047.
Transportation Research Record, 2674 (5), 135–148. Walubita, L.F., Alvarez, A.E., and Simate, G.S., 2011. Evaluating and com-
Laukkanen, O.V., et al., 2015. Creep-recovery behavior of bituminous paring different methods and models for generating relaxation mod-
binders and its relation to asphalt mixture rutting. Materials and ulus master-curves for asphalt mixes. Construction and Building
Structures, 48 (12), 4039–4053. Materials, 25 (5), 2619–2626.
14 M. LING ET AL.

Walubita, L.F., Hoeffner, J., and Scullion, T., 2012. New generation mix- Zhang, Y., et al., 2022. Prediction and evaluation of rutting and moisture
designs: laboratory-field testing and modifications to Texas HMA mix- susceptibility in rejuvenated asphalt mixtures. Journal of Cleaner
design procedures (No. FHWA/TX-13/0-6132-3). Austin, TX: Dept. Production, 333, 129980.
of Transportation. Research and Technology Implementation Office. Zhou, F., Steger, R., and Mogawer, W., 2021. Development of a coherent
Wang, H., et al., 2011. Effect of mineral filler characteristics on asphalt framework for balanced mix design and production quality control
mastic and mixture rutting potential. Transportation Research and quality acceptance. Construction and Building Materials, 287,
Record, 2208 (1), 33–39. 123020.

View publication stats