REVIEW

published: 14 April 2022

doi: 10.3389/fmats.2022.879245

Literature Review on the Discrete

Element Method in Asphalt Mixtures
Hui Yao 1*, Mei Xu 1, Junfu Liu 1, Yu Liu 2, Jie Ji 3 and Zhanping You 4
1
Beijing Key Laboratory of Traffic Engineering, College of Metropolitan Transportation, Faculty of Architecture, Civil, and
Transportation Engineering, Beijing University of Technology, Beijing, China, 2Professor, School of Highway, South Erhuan Middle
Section, Chang’an University, Xi’an, China, 3Professor, School of Civil Engineering and Transportation, Beijing University of Civil
Engineering and Architecture, Beijing, China, 4Professor, Department of Civil and Environmental Engineering, Michigan
Technological University, Houghton, MI, United States

Asphalt mixtures are commonly used in pavement engineering, especially for highway
construction. The mechanism exploration and analysis of pavement distress are the main
challenges for researchers and industry managers. Thereby, it is important to understand
their properties and interaction mechanisms in asphalt mixtures. It is difficult to conduct
some sophisticated or microscale tests in the laboratory, and numerical simulation and
virtual tests can be solutions for these cases with low costs. The Discrete Element Method
(DEM) is a promising tool for researchers to undertake these tasks. This paper mainly
summarized and analyzed the research progress and development prospect of DEMs in
Edited by: asphalt mixtures from a series of technical sections. The laboratory test results were often
Antonio Caggiano,
used to calibrate the DEM simulations as well as Two-dimensional (2D) and Three-
Darmstadt University of Technology,
Germany dimensional (3D) modeling. Several modeling methods were developed to generate
Reviewed by: digital samples, like user-defined, image-based, random-modeling. In addition, the
Tao Ma, conclusions can be referenced by researchers for the development of numerical
Southeast University, China
Tuo Huang,
simulations.
Changsha University of Science and
Keywords: pavement engineering, asphalt, asphalt mixtures, discrete element, modeling
Technology, China
Augusto Cannone Falchetto,
Aalto University, Finland
*Correspondence: 1 INTRODUCTION
Hui Yao
[email protected] Generally, the behaviors of materials can be evaluated from the performance of asphalt mixtures,
such as high-temperature stability, low-temperature crack resistance, durability, workability, etc.,
Specialty section: The laboratory test method is a good way to understand the properties of asphalt mixtures. Also,
This article was submitted to another analysis way is the numerical simulation method to get the fundamental views and insights.
Structural Materials, Different numerical simulation methods are gradually emerged to study the properties of asphalt
a section of the journal mixtures, such as the Finite Element Method (FEM), Discrete Element Method (DEM), Molecular
Frontiers in Materials
Dynamics (MD) method, etc., Three numerical methods focus on the different scales. The time scale
Received: 19 February 2022 is from femtosecond to second, and the length scale is from meter to Angstrom. The time scale and
Accepted: 25 March 2022
length scale of the DEM is microsecond and micrometer, respectively. The DEM is a suitable
Published: 14 April 2022
numerical simulation method to analyze the interactions in asphalt mixtures at the microscale. Many
Citation: tests cannot deal with the interaction between materials at the microscale, but the DEM is a good way
Yao H, Xu M, Liu J, Liu Y, Ji J and You Z
to explore mechanisms at the microscale. Therefore, many scholars used the DEM to research the
(2022) Literature Review on the
Discrete Element Method in
properties of asphalt mixtures in recent years. The following discussions relate to applications of the
Asphalt Mixtures. DEM in asphalt mixtures.
Front. Mater. 9:879245. Generally speaking, the research on the performance of asphalt mixture mainly starts from the
doi: 10.3389/fmats.2022.879245 macroscopic and microscopic aspects. Firstly, from the macroscopic perspective, the experimental

Frontiers in Materials | www.frontiersin.org 1 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

method is a more traditional method to study the performance of

asphalt mixtures, and it is also the most widely used method. It is
controlled by humans in the laboratory, and its advantage is
strong intuitiveness. 1) Although some errors cannot be avoided
during the test process, the test results are still generally more
reliable from realistically thinking. However, the costs of some
tests are high and the operation process is also complicated and
time-consuming. 2) During data collection in the tests, indirect
and direct measures can be collected, like stress and strain data.
However, the stress and strain values of each point in the entire
material cannot be obtained directly and the route of the force
transfer cannot be directly demonstrated. It indicates that the
mechanisms or operation ways are not visible for the researchers
during laboratory tests. Therefore, different numerical methods
are invented and developed for different purposes.
Compared with the experimental method, the advantages of
numerical simulations are 1) the cost is low and most cases in the
tests can be simulated using the computer resource; 2) any type or
amount of loading can be applied without worrying about the FIGURE 1 | Trends for the discrete element method (DEM) in the
actual loss caused by the excessive load; 3) some extreme literature review.
situations cannot be achieved and done by the experimental
method and they can also be applied in the numerical
methods. On the other hand, the numerical simulation (2017). The results showed that the DEM was an effective way
method can analyze the stress of each area and measuring to simulate elastic-plastic properties of the asphalt mixtures, and
point of the test materials, and get the change of stress and simulation results were also of great help to improve the
strain, as well as the force and displacement distributions. The test durability and drainage performance of the asphalt pavement.
results can be obtained quickly through certain operation settings, Hou et al. (2015) established discrete element specimens for
but the numerical simulation method is not perfect, and there are asphalt mastic macadam mixtures, SMA-13, and applied
some shortcomings, such as abstraction, intuition, credibility, vehicle loads to study its meso-mechanical response under
etc., The accuracy of the simulation can make certain predictions loading. Huang et al. (2008) carried out a series of biaxial
about the experiment results. The numerical simulation method compression tests on sand granular materials under different
mainly refers to the finite element method and DEM, which are densities and confining pressures, and further analyzed their
the most commonly used. The DEM is used to solve the problems mechanical properties through the response of sand
of discontinuous media by analyzing the interlocking contacts of characteristics from simulation tests. Compared with
discrete elements. The constitutive relation of the contact is found laboratory test methods, the DEM simulation can provide the
out and the physical and mechanical model of contact is details of meso-structural changes in the loading process. The
established. Then the discontinuous and discrete elements are researchers simulated the Compaction Flow Test (CFT) of asphalt
simulated according to Newton’s second law of motion. The mixtures based on the 3D DEM (Roozbahany et al. (2019) and
DEM can better explain the stress or loading types, deformation analyzed several factors, including the fluidity of asphalt mixture
mechanism, and failure processing of the asphalt mixtures from particles, mold size, loading strip geometry and loading rate, etc.,
the meso and micro perspectives. It can also be used to predict Kusumawardani and Wong (2020) carried out the DEM
rutting initiations, fatigue mechanisms, and engineering simulation and laboratory tests on the mechanical properties
evaluation indexes for asphalt mixtures. However, the finite of Porous Asphalt Mixtures (PAM) with three aggregates shapes,
element method is not good for large and discontinuous and also analyzed the influence of the aggregate shape on the
deformations and is more suitable for the study of continuum accumulation of mixtures. Other studies tried to simulate the
problems. Therefore, the researchers tend to use the DEM to rutting test of asphalt mixtures with 3D DEM and analyze the
simulate the stress-strain relationships, damage situations, and influence of internal structural characteristics of the mixtures on
fracture processing of the asphalt mixtures. The examples of the rutting behavior (Zhang et al., 2018). The three-stage
specific studies on discrete elements are discussed as follows: simulations (gravity drop, static, and rotary compaction) were
The parameters of the generalized Burgers model fitted by performed on three different asphalt mixtures to analyze the
Mao (2018) were used to simulate the uniaxial test of asphalt internal consistency of the mixtures (Zhou et al., 2020). The
mortars with a three-dimensional (3D) DEM. Although the randomly generated algorithm was used to establish a model to
results had deviations from the tests, this simulation provided simulate the aggregate penetration test and the uniaxial creep test.
a good attempt to understand the asphalt mortars and their Compared with the experimental tests, it was proved that the
behaviors. The scholars used the 3D DEM to simulate the established microstructure models were correct and simulation
porosity changes of asphalt mixtures under the dynamic results are feasible (Ma et al., 2016b). The static indirect tensile
loading and summarize the changing laws Zhang and Li test of the asphalt mixture was simulated, and the laboratory

Frontiers in Materials | www.frontiersin.org 2 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

experiment was carried out to verify the model. The crack on the 2D plane. The 3D program was not developed because
propagation mechanism of asphalt mixtures was explained of the complexity and advanced computer technology. After
through simulation and test results (Dan et al., 2018). There the development of the computer algorithms, the DEM
are many scholars to publish more manuscripts related to the was gradually accepted to solve the problem of
DEM and they are trying to solve many kinds of engineering discontinuous media. One of DEM companies is called
problems through numerical simulations. According to the “ITASCA,” and it developed many discrete element
currently statistical survey and literature review (main programs including the 2D discrete element
databases), research on the application of the DEM showed an program—UDEC (Universal Distinct Element Code), and
increasing trend, and also the topic of the literature mainly 3D discrete element program—3DEC (3D Distinct Element
focused on asphalt mixtures. This trend can be seen from Code). These programs aimed to solve problems from rock
Figure 1. mechanics and mining engineering.
This paper briefly introduced the historical status and Wang (1986a) introduced the basic principle and
development of the DEM, and demonstration of its advantages application cases of DEMs to the field of rock mechanics
and disadvantages, as well as differences between the laboratory and engineering in China at the first National Symposium.
test method and DEM in performance analysis of asphalt The DEM can also be generalized for other research areas
mixtures. The key point was to summarize characteristics of (Wang, 1986b), like mining or other engineering fields (Wang
the DEM and learn from others. The DEM applications and and Xing, 1986; Wang et al., 1987; Xing and Wang, 1988;
shortcomings were also mentioned, as well as prospects of its Wang and Xing, 1991; Wang and Xing1993a; Wang and
future development in civil engineering. Xing1993b). The DEM progress was summarized from
application examples (Wang and Xing, 1989; Wang, 1990).
The 2D and 3D DEMs were also promoted for the simulation
2 HISTORICAL OVERVIEW AND DISCRETE of civil engineering materials (Wang and Xing, 1990; Xing
ELEMENT METHOD and Wang, 1990). One of the discrete element programs was
called “MatDEM” based on MATLAB and it was developed by
2.1 Basic Overview of Discrete Element Liu et al. (2020) for 3D DEM. This program can build the
Method virtual particles of asphalt mixtures with 2D and 3D models.
The numerical simulation is a more advanced analysis method Another discrete element simulation software “StreamDEM”
based on computer resources and it is different from the was developed and it improved the simulation speed and
traditional laboratory test method. The numerical simulation efficiency based on the GPU operation. It met the calculation
mainly included the finite element method and boundary requirement of the large-scale and large number of particles
element method before the invention of the DEM (Wang and (million levels), and can be commonly used in the
Li, 2005). As a new numerical simulation method, the DEM was engineering fields (Zhang and Zhang, 2016; Zhang et al.,
applied to various studies by researchers. Compared with the 2017).
former two methods, the DEM focused on solutions of
discontinuous media based on Newton’s second law of motion
(F = ma) (Wang et al., 2019). The DEM software PFC (Particle 3 RESEARCH METHODS FOR ASPHALT
Flow Code) was more suitable and widely used (Liu et al., 2020). MIXTURES
The PFC was divided into two-dimensional (2D) and three-
dimensional (3D) PFC programs. The PFC was based on the 3.1 Analysis of the Laboratory Test Method
theories of Newton’s second laws and relationship between force The traditional tests were employed to evaluate the properties of
and displacement, when simulating the related model of asphalt the asphalt binders and mixtures with different modifiers at
materials (You and Liu, 2010). different conditions based on the Marshall and Superpave
systems. The recycled materials are the hotspot for the
2.2 Historical Overview development of asphalt mixtures. Recently, Li et al. (2021)
The DEM was originally proposed by Prof. Cundall (1971) as a recycled SMA-13 asphalt mixtures for the pavement
method for the analysis of granular discrete materials. It was reconstruction project. Song (2021) utilized waste materials to
originally used to analyze geotechnical problems. He also reduce emissions and increase waste consumption, as well as the
realized the purpose of interactive simulations after polypropylene fibers and glass. Yao et al. (2021) used comparative
inputting and outputting geometric data on the computer experiments to analyze performance with other waste materials,
(Cundall, 1974), and developed a 2D plane “Ball” program such as red mud. It was mixed into asphalt mixtures by the
to simulate the behavior of granular media (Cundall, 1978). replacement of the mineral powder. Based on the test results, it is
Cundall and Strack (1979) carried out the numerical a good way to use the waste resources in the pavement
simulation in 1979 and compared them with the real engineering. In addition, different modifiers were used for
experimental data. The use of the DEM can carry out more performance improvement of asphalt mixtures. The polyester
realistic simulations of specimens. However, due to the fibers can also effectively improve performance of asphalt
immaturity of the technology and the limitations of the mixtures with the Reclaimed Asphalt Pavement (RAP) (Zhong
conditions at that time, the research was carried out based and Zhang, 2021). The comparative tests were also used to

Frontiers in Materials | www.frontiersin.org 3 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 2 | Laboratory tests and processing for the dynamic modulus of the asphalt mixtures (Jia et al., 2021; You et al., 2009; Adhikari and You, 2010). (A) the
dynamic modulus test samples (Jia et al., 2021); (B) changes in the stress and strain with loading frequencies (Jia et al., 2021); (C) the dynamic modulus tests with the
sand mastic and asphalt mixtures (You et al., 2009; Adhikari and You, 2010); (D) stress and Strain response during the tests(You et al., 2009; Adhikari and You, 2010).

determine the optimum content of these modifiers, like the dynamic modulus and phase angle) are measured at different
Freeze-Thaw Test, Four-Point Bending Beam, Dynamic Shear temperatures and frequencies. Before testing, the samples are
Rheometer (DSR) and Bending Beam Rheometer (BBR) test mixed by asphalt binders and aggregates with the certain
(Gao, 2021). Meng et al. (2021) studied the effect of the snow gradation. The mixtures are compacted by the Superpave
melting agent on the adhesion of asphalt mixtures. The use of Gyratory Compactor (SGC) with a targeted air void. The
the snow melting agent (chloride) did not affect the adhesion samples are compacted and cut with the specified
between asphalt mixtures, but the negative effect on dimension (100 mm in diameter and 150 mm in height);
performance of asphalt mixtures may occur by the increase The detailed procedure can be found from the specification
of the agent content. The Styrene-butadiene-styrene (SBS) AASHTO T378. The indirect tensile test is measured by the
modified asphalt binders were used for the preparation of the Universal Testing Machine (UTM), and this test is destructive
composite modified asphalt binders to enhance the low- and used to evaluate the moisture susceptibility. This test
temperature performance (Xia et al., 2021). Figures 2A,B results can also be used to predict the performance of anti-
show the dynamic modulus test samples and curve of stress stripping and fracture toughness. The range of the air void in
and strain changes with time in asphalt mixtures (Jia et al., the asphalt mixture is around 6%–8%. The dimension of the
2021). sample is 100 mm in height and 63.5 mm in diameter. One
The dynamic modulus test is conducted under the haversine group of the samples is cured with different conditions and
or sinusoid loading conditions. The output parameters (the another group is stored in the dry environment. The procedure

Frontiers in Materials | www.frontiersin.org 4 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 3 | 2D DEM samples with different gradation and air void distributions (Nian et al., 2021; Zhang et al., 2018; Peng and Sun, 2015a; You et al., 2010). (A)
2Ddiscrete element samplesof the asphalt mixtureswith different sizes (Nian et al., 2021; Zhang et al., 2018). (B) 2D DEM model generation for the AC13 asphalt mixture
(Peng and Sun, 2015a). (C) Air void distribution in the 2D DEM models and 3D DEM models from 16 slices of 2D models (You et al., 2010).

Frontiers in Materials | www.frontiersin.org 5 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

of this test can be found from the specification AASHTO T283. gradations were selected (Yu, 2015) for laboratory tests and
These are the details for the dynamic modulus and indirect simulations of the splitting test with the 2D DEM. The
tensile test. feasibility of discrete element simulation was studied and the
influence of aggregates with different properties on the cracking
3.2 Discrete Element Analysis of asphalt mixtures was analyzed. The results demonstrated that
3.2.1 Two-Dimensional Discrete Element Method the DEM simulation may provide more intuitive and effective
In many early studies of the discrete element modeling method, details compared to the laboratory test processing, like loading
the 2D model was commonly used and the advantage of the 2D changes, displacements, and cracking propagations during
model is high calculation speed for the complex system. testing. However, this study also revealed the limitations of the
Sometimes, although simulation results of some relatively 2D DEM, such as the lack of geometric information. Figure 4
complex experiments are not consistent with the actual shows the loading-displacement relationship of the AC-16
measured data, the 2D model program is still advantageous gradation section at different temperatures when simulating
for the experiments with the simple internal structure of the the splitting test of the asphalt mixture by a 2D DEM (Peng
mixture and the analysis of the general trend without requiring and Sun 2015b). In addition, the 2D DEM can be also used in
accurate results. The asphalt mixtures with a 2D discrete element many other studies, like other Marshall system tests of asphalt
program as an example (Nian et al., 2021; Peng and Sun 2015a) mixtures.
are shown in Figure 3.
Liu et al. (2009) established a 2D discrete element model to 3.2.2 Three-Dimensional Discrete Element Method
focus on the viscoelastic properties of asphalt mixtures with the From the morphology perspective, the effect of the 3D model is
burger model. Through the DEM simulation of the asphalt close to the real experiment compared to the 2D one and the 3D
mixtures, the dynamic modulus and phase angle of the asphalt model has more complex structures and interactions. During the
mixtures were predicted under the sinusoidal loading, and the generation and optimization of 3D models, the calculation time is
viscoelastic models were verified by the measured results. There is longer than that of the 2D model and the time depends on the
only a little research on the influence of aggregate uniformity on element number in the simulation system. Sometimes, the
the splitting strength test of asphalt mixtures. The simulation computing time can be greatly reduced using the temperature-
results showed that whether the aggregates are uniform or not has frequency superposition method, but the computation time is still
little effect on the average splitting strength. However, it was relatively long compared to the 2D model. The temperature-
closely related to the variability of splitting strength, which frequency superposition method specifically is to amplify or
showed that the coefficient of variation of the splitting adjust the normal loading frequency from the available test
strength decreases with a better aggregate uniformity (Peng Y. results and convert Burger’s model parameters under the
and Sun L., 2015a). The variation relationship of the splitting normal frequency into parameters under the virtual frequency.
strength under different loading rates was that the higher the The model can be established with the magnified virtual
loading rate was, the higher the splitting strength was with a frequency and is matched by the modified Burger’s model
logarithmic growth (Pei et al., 2010). The cracking and failure parameters. The comparative test results of the dynamic
properties of asphalt mixtures with three different gradations modulus of asphalt mixtures at different temperatures and
were analyzed and compared under the splitting test simulations frequencies showed that the simulation results were close to
using the 2D DEM. The simulation results demonstrated that the measured results under the temperature-frequency
large nominal maximum particle size and huge porosities of superposition method. However, the calculation time was
aggregates with low asphalt contents caused low splitting greatly reduced. When the computation time took months or
strength in asphalt mixtures. The crack resistance of the even years in the traditional method, it can be reduced to only
OLSM-25 (Open-graded Large Stone Asphalt Mixes) asphalt hours, and the calculation speed increased rapidly (Liu and You,
mixture was better than that of AC-16 and AC-25 (Asphalt 2011a). It has been used most by scholars in recent years.
Concrete) (Guo and Chen, 2014; Guo and Chen, 2015). The Due to the limitation of the computer technology in the early
splitting test process of four different aggregates was simulated by stage, the research with the 3D DEM is far less than the 2D one.
the 2D DEM and the influence of aggregate properties on splitting Dr. You’s research team members conducted much research on
strength was also analyzed with laboratory tests. It was concluded the 3D DEM, like a 3D DEM model to predict the dynamic
that the high stiffness and strength of aggregates lead to a strong modulus and phase angle of asphalt mixtures, and the relative
bearing capacity of the asphalt mixture models (Xiong, 2013). errors between the laboratory and simulation results were within
The indirect tensile test of the asphalt mixture was simulated by an acceptable range (You and Liu, 2010). Another 3D DEM
2D DEM to analyze the influence of the bond strength ratio and model was established to simulate the creep test of the asphalt
loading rate on the splitting strength of the asphalt mixture. The mixture to study the creep stiffness and response (You et al.,
results showed that the splitting strength decreased with the 2011), and also, it was used to analyze whether the element size
increase of the bond strength; the splitting strength increases affected the performance of the asphalt mixtures (especially in
with the increase of loading rate, and it is proved that the mechanics) (Liu et al., 2012). The dynamic modulus of the asphalt
performance of asphalt mixtures plays the most important role mixture was predicted and the influence of the aggregate size was
in enhancing the crack resistance of asphalt mixture (Chang et al., analyzed by comparing the predicted results with the measured
2011). Four kinds of aggregates with different properties and data (Adhikari and You, 2011). The fitting curves of the dynamic

Frontiers in Materials | www.frontiersin.org 6 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 4 | Load-displacement curves of the splitting test (indirect tensile test) and contact distributions (Peng and Sun, 2015b; Huang et al., 2008; Nian et al.,
2021). (i) force-displacement distributions in the DEM models, AC13 and AC16. (ii) Distributions of contacts with different shearing statuses (A) 0; (B) 10%; (C) 20%; (D)
40% (Huang et al., 2008); (iii)Scanning 2D samples of the asphalt mixtures(Peng and Sun, 2015b); (iv) the contact responses under the loading (Nian et al., 2021).

Frontiers in Materials | www.frontiersin.org 7 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 5 | Fitting curves between the simulation data and measured values in the laboratory with the 2D and 3D DEMs, (A) the dynamic modulus; (B) the phase
angle (Liu and You, 2011b).

modulus and phase angle of asphalt mixtures are shown in and 3D DEM models (Peng et al., 2016). The accuracy of the
Figure 5, and the relationship between the 2D and 3D DEM simulation model was also improved from the 2D model to
simulation values and laboratory-measured values was the 3D model (Peng and Bao, 2018). The influence of the
demonstrated (Liu and You, 2011b). Both 2D and 3D DEM aggregate gradation and temperature on the splitting strength
coincide with the test results. Zhang et al. (2015) modeled rubber of asphalt mixtures was also analyzed with the 3D DEM by
particles and aggregates, respectively, to simulate the splitting test Peng et al. (2019). The DEM simulation data trend of the
of the asphalt mixture by the 3D DEM. The simulation results asphalt mixtures was consistent with the laboratory test
were in good agreement with the test results. It can be seen that it results. However, minor errors can be found in the
was successful to use the DEM to simulate the test processing simulation data. Further studies were needed for the
from the initial loading and cracking to the failure of the asphalt improvement method and fundamental interactions.
mixture. In addition, the contact model in the DEM is composed Figure 6 shows the 3D asphalt mixture DEM model (Shan
of three built-in parts: “stiffness model, slip model and bonding et al., 2019; Liu and You, 2011b; Ma et al., 2016b; You et al.,
model.” These models constitute other complicate models. The 2008; Zhou et al., 2020; Liu et al., 2012).
slip model in the DEM means the frictional properties between The aggregate uniformity of asphalt mixtures was also
contact points. The bonding model represents shear and normal investigated with the 3D DEM through the low-temperature
strength between contact points. The stiffness model addresses splitting strength, and the Cohesive Zone Model (CZM) was
the relationship between the contact forces and displacements of also used in the DEM model to explore the relationship between
elements. The viscoelastic models or properties of asphalt the aggregate uniformity and splitting strength of asphalt
mixtures can be expressed by the Burger’s model and this mixtures. In the study of Liu et al. (2018), the same
model can be constituted from three parts (stiffness model for conclusion was obtained as the previous study (Yu, 2015) did.
asphalt mastics and aggregates or within asphalt mastics; slip In recent years, the 3D DEM and the cohesive zone model were
model for all contacts; bonding model for all contacts except also used by many scholars to study the splitting strength of
between aggregates). Thereby, different viscoelastic models can asphalt mixtures. The asphalt and aggregate models were
be further formed from three contact parts in the DEM (Liu et al., reconstructed by Liang et al. (2019) to simulate the splitting
2012). test with the 3D DEM and MATLAB, and the cohesive model in
The Digital Image Processing Technology and this model was applied. The cohesive model was also combined
Computerized Tomography (CT) Scanning Technology with the finite element analysis in six kinds of asphalt mixtures.
were also employed and assisted for the DEM simulation The results showed that the splitting strength trends of the asphalt
to explore the interaction mechanisms in the asphalt mixtures were the same as the experimental results roughly. The
mixtures. Different factors were analyzed through the DEM simulation was feasible for the asphalt mixtures and
splitting test of the asphalt mixture with the 3D DEM by provided a reference for the fundamental research. The
Wan (2016), including the aggregate gradation, asphalt applicability of the contact bond and parallel bond models was
contents, temperatures, and loading rates. The research explored by Shan et al. (2019) and Sheng et al. (2018) through the
results were consistent with the previous analysis results. A 2D model and 3D DEM models with different parameters. The
possible reason for the error between simulation and results showed that it was reasonable to adapt the parallel bond
measured results is the selection of parameters in the DEM model when simulating the mechanical properties of asphalt
models. It can be concluded and referred that 3D DEM mixtures with the load-displacement behavior. The splitting
simulation is more stable and reliable than the 2D one test is commonly used for the evaluation of water
through the comparison between microstructures of the 2D susceptibility and permanent deformation resistance of asphalt

Frontiers in Materials | www.frontiersin.org 8 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

mixtures. The dynamic modulus test of the asphalt mixture is also

used for the evaluation of the high- and low-temperature
performance. The 3D DEM is a promising way to understand
the microscopic behavior of the asphalt mixtures, as well as the
splitting test and dynamic modulus test of asphalt mixtures. More
fundamental understanding between the materials may be
revealed and induced from the DEM simulations in the near
future.

4 DISCRETE ELEMENT MODELING

METHOD AND SIMULATION PROCESS
4.1 Modeling Method
Generally speaking, there are three modeling methods to
construct the microstructure of asphalt mixtures with DEM,
including the idealized model, randomly generated irregular
shape model and image-based model. The idealized model
cannot represent the geometric properties of aggregates and
mixtures, and does not have authenticity; the image-based
model can truly simulate the characteristic structures of
asphalt mixtures with the help of computer scanning
technology. However, it is dependent on the laboratory
conditions. The scanning equipment is expensive and the cost
is relatively high. The testing process is long and complex; the
randomly generated irregular shape model is independent of the
laboratory, which is more convenient and faster than the former
two modeling methods and consumes less manpower and
material resources. The above has been emphatically
mentioned in many kinds of literature. The idealized model is
idealized, which is not consistent with the real experimental
results, the latter two models are mostly used in general
research. Therefore, only the latter two modeling methods are
introduced in this article.

4.1.1 Based on Scanning and Image Processing

Technology
A 3D viscoelastic model of the asphalt mixture was established
by You et al. (2008) based on X-ray image scanning technology
to simulate and restore the creep flexibility test. A time-
temperature superposition principle was developed to
improve the calculation time. The microstructure
reconstruction of asphalt mixtures with the 3D DEM was
completed by You and Liu (2010) based on images from the
X-ray computed tomography. The frequency-temperature
superposition principle was used to shorten the simulation
time. The advantage of this image processing technology is that
it can truly show most of the structural characteristics of
asphalt mixtures. The dynamic modulus and phase angle of
FIGURE 6 | Three-dimensional discrete element model and DEM the asphalt mixtures were predicted and compared with the
viscoelastic configurations of the asphalt mixtures (Shan et al., 2019; Liu and laboratory test results. The results showed that most of the
You, 2011b; Ma et al., 2016b; Zhou et al., 2020; You et al., 2008; Liu et al.,
predicted values were close to the measured values. The 3D
2012). (A) 3D DEM models for the asphalt mixtures with different sizes
and gradations (Shan et al., 2019; Liu and You, 2011b; Ma et al., 2016b; Zhou discrete element model of the asphalt mixture was
et al., 2020; You et al., 2008; Liu et al., 2012). (B) The configuration of 3D DEM reconstructed through the scanned 2D model of the
models with the viscoelastic properties (Liu and You, 2011b). (C) 3D DEM aggregates and the related research was carried out to
models of the asphalt mixtures from 2D DEM models(You et al., 2008). (D) analyze the influence factors of the modulus (You et al.,
Three basic contact models in the DEM.
2011). Also, the 3D structured light scanning technology

Frontiers in Materials | www.frontiersin.org 9 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 7 | Virtual samples of the asphalt mixtures (Liu, 2011; You et al., 2009; Adhikari and You, 2010). (A) 3D virtual samples from X-ray computedimages (You
et al., 2009; Adhikari and You, 2010); (B) aggregates, mastic,and air void distributions in the 2D models (You et al., 2009; Adhikari and You, 2010). (C) 2D DEM samples
with loading platen(Liu, 2011); (D) 2D samples with different vertical cuts(You et al., 2009; Adhikari and You, 2010).

was used to scan the aggregate particles and the aggregate information acquisition software with the help of CT scanning.
information was compiled with the help of the MATLAB The aggregate characteristics were imported into PFC3D to
program. The 3D model of asphalt mixtures was established generate a 3D discrete element model of the asphalt mixture to
based on the aggregated information in PFC5.0 to study the study the effect of the gap distribution on the shear fatigue life.
multiscale influence factors of volume characteristics (Li et al., the relevant information of aggregate particles was obtained by
2020). X-ray CT technology was used to scan aggregates for the Shi (2020) from the laser scanning and established a PFC3D
microstructure of aggregates. The MATLAB program and model of the asphalt mixture. The model was verified with the
PFC3D were employed to establish a 3D model of the help of the laboratory experimental data and it was used to
asphalt mixture to simulate the uniaxial penetration test to analyze the aggregate characteristics. The particle images of
explore how aggregate affected the shear performance (Wu aggregates were intercepted by Lu (2020) with CT scanning
et al., 2021). The scanned internal images of aggregates were technology and processed them with the MATLAB program to
processed by Guo (2020) and Peng et al. (2021)with image establish a 3D model of the asphalt mixture in PFC3D. The

Frontiers in Materials | www.frontiersin.org 10 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 8 | 3D DEM scanning samples and digital generation specimens (Ge et al., 2021; Adhikari and You, 2010). (A) 3D scanning processingand model
generation of the asphalt mixtures (Ge et al., 2021). (B) X-ray CT images, Mastic, aggregate, andair void samples (Adhikari and You, 2010).

virtual splitting test (indirect tensile test) and uniaxial research. The 2D mixture image and the scanning process
penetration test of the asphalt mixtures were carried out of 3D specimens obtained with the help of scanning technology
and verified with real experimental data for further are shown in Figure 7 (Liu, 2011) You et al., 2009; Adhikari

Frontiers in Materials | www.frontiersin.org 11 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 9 | SCB digital samples, loading setting and deformations of the asphalt mixtures (Liang et al., 2021; Nian et al., 2021).

FIGURE 10 | Splitting test (indirect tensile test) and three-point bending test samples of the asphalt mixtures with the DEM models (Nian et al., 2021).

Frontiers in Materials | www.frontiersin.org 12 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 11 | Dem modelling steps and interaction behaviors between balls (Liu, 2011; You et al., 2009; Adhikari and You, 2010; Cundall and Strack,1979; Ding
et al., 2019; Ma et al., 2018). (A) Steps for the discrete element models with different methods(Liu, 2011; You et al., 2009; Adhikari and You, 2010; Ma et al., 2018). (B)
The force-displacement law in the DEM(Cundall and Strack,1979).

and You, 2010) and Figure 8 (Ge et al., 2021) (Adhikari and creep experiments to study the influence of void-related factors at
You, 2010). different temperatures (Ma et al., 2016a). A 3D user-defined
model was established by Du et al. (2021) to simulate the indirect
4.1.2 User-Defined Method tensile test to analyze the gradation effect on the crack
The randomly generated irregular shape model refers to the propagation at low temperatures, as well as the virtual
aggregates of the asphalt mixtures represented by irregular performance test of different gradation asphalt mixtures.
polygonal particles in the discrete element model established Different mixture gradations were adapted for the
by a random polygon algorithm (Liu and You, 2008). A self- performance analysis and improvement measurements (Ling
defined discrete element model was established by Chen et al. et al., 2020). the user-defined method was used by Wang et al.
(2012) and the cyclic loading uniaxial compression tests of (2021) to establish PFC2D and PFC3D models of asphalt
asphalt mixtures were carried out. The model was verified to mixtures and simulated the dynamic creep and rutting tests.
predict the dynamic modulus and phase angle of asphalt mixtures Other tests were modeled with the PFC2D program to study the
and analyze factors for the viscoelastic characteristics of asphalt factors of asphalt mixtures for the low-temperature performance,
mixtures. four kinds of typical aggregate gradations were selected such as the Semicircular Bending Test (SCB) (Figure 9), Splitting
by Zhou (2016) to carry out the dynamic modulus test and Test (indirect tensile test), and Three-point Bending Test
generate the numerical model to analyze the DEM feasibility for (Figure 10) (Liang et al., 2021; Nian et al., 2021).
the virtually mechanical test of asphalt mortars and mixtures. The
classically viscoelastic model (Burger’s self-defined constitutive 4.2 Simulation Process
contact model) was adopted in the discrete element model. The 4.2.1 Modeling
digital cylinder specimen was composed of multiple discrete The basic steps of the discrete metamodels (Randomly Generated
elements with different sizes and the discrete element model Models) for asphalt mixtures in PFC3D were roughly divided into
was used for the dynamic modulus test. The randomly generated the following steps (Liu, 2011) (You et al., 2009; Adhikari and
PFC3D models of asphalt mixtures were used to simulate the You, 2010). 1) the digital size and aggregate gradations were

Frontiers in Materials | www.frontiersin.org 13 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

FIGURE 12 | 2D and 3D DEM simulations for the splitting test (indirect tensile test) of the asphalt mixtures (Shan et al., 2019; Peng and Bao, 2018; Ma et al., 2016b).
(A) 2D DEM model with simple ball systems (Shan et al., 2019); (B) 3D DEM model with simple ball systems (Shan et al., 2019). (C) 2D DEM model with different shapes of
aggregates(Peng and Bao, 2018); (D) 3D DEM model with different shapes of aggregates(Peng and Bao, 2018). (E) the coarse aggregates in the 3D DEM models (Ma
et al., 2016b).

defined, and the asphalt content and air voids were set based on System (AIMS) was used to measure the aggregate particles to get
the volume characteristics of asphalt mixtures; 2) balls with the the image characteristics. Three methods were adopted to form
same size were created and a polyhedron was randomly created; the virtual aggregate samples, including the Radius Expansion
3) A two-phase system of the aggregate and mortar was (RE), Interior Filling (IF), and Contour Filling (CF) Methods
generated; 4) the volume characteristics of the two-phase (Ding et al., 2019). The relationship between the contact points
system were verified. The air voids were randomly generated was also demonstrated and defined in the literature (Ding et al.,
by deleting the balls in the two-phase system. Finally, a digital 2019) (Cundall and Strack (1979). Three parts (asphalt mastic,
PFC sample of the asphalt mixture was generated with the air void aggregates and air voids) was combined to generate the asphalt
by the random method. The Aggregate Imaging Measurement mixture samples in PFC3D with the user-defined procedure. The

Frontiers in Materials | www.frontiersin.org 14 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

asphalt mixture samples were made for the wheel tracking test problems and it is hard to deal with or figure out them
(Ma et al., 2018). The complex processes and interaction through the laboratory test. Most importantly, the
behaviors of these methods are shown in Figure 11 (Liu, simulations can avoid most laboratory risks during real
2011; You et al., 2009; Adhikari and You, 2010; Cundall and testing in the laboratory, and it is safe that the hazardous
Strack,1979; Ding et al., 2019; Ma et al., 2018). materials are also processed with the simulations. Thereby,
more scholars tried to use different simulations to solve the
4.2.2 Loading Model problems with different scales. In addition, the required
The shapes and sizes of the asphalt mixture models were relatively simulation time is shorter than that used in the
different from the test samples. The dynamic modulus test laboratory. Some tests in the laboratory may take as long
models of asphalt mixtures were cylindrical; the splitting as several years, like the creep test for concretes. For instance,
strength test models of asphalt mixtures were cylindrical with during the dynamic modulus simulations of asphalt
the same size as the actual Marshall specimen, and the four-point mixtures, the calculation time in the DEM simulation can
bending beam test model is rectangularly trabecular. It is be shortened to several days or even hours using the
necessary to apply the load in the model with “force” after the temperature-frequency superposition method (Liu and
generation of the mixture model. The specific operation was You, 2011a). It can be found that most DEM simulations
summarized as follows: setting several different gradations; are verified by real experimental data from different studies.
writing the servo control program in the Fish language; and The computational and numerical simulations may be the
achieving the function of applying load to the top and bottom of promising way to understand the behavior of materials based
the model specimen. It is worth noting that the force on the “wall” on the fast development of computer technologies. Some
cannot be directly controlled in the program. Instead, it is advantages are highlighted compared to laboratory testing,
necessary to control the speed of the “wall” through like the fast calculation, low cost, less labor, less risk, no
commands. The “wall” at the bottom of the specimen is not damage to human health, and so on. Most importantly, the
moving and a downward speed is applied to the “wall” at the top numerical simulation can explore the interactions between
of the specimen to realize the loading of the model, and the materials, which cannot easily realize by the experiments,
control of the speed of the “wall” is realized using the servo such as the interaction at the nano- and micro-scale between
control program to apply the periodic load to the model materials compositions. In addition, experimental data is still
specimen. The PFC2D and PFC3D models of the asphalt needed in the DEM simulations of asphalt mixtures to ensure
mixtures are generated as shown in Figure 12 (Shan et al., the relative accuracy of the results.
2019; Peng and Bao, 2018; Ma et al., 2016b). (2) Generally speaking, there are two ways to simulate different
tests or models of asphalt mixtures with discrete element
programs, PFC2D for 2D mixture models and PFC3D for 3D
5 DISCUSSIONS mixture models. It also means that two aspects and
perspectives are available to explore the digital materials.
(1) The DEM was designed for the analysis of rock problems and The PFC3D program is developed and extended from the
it was extended to solve the pavement problems later, as well PFC2D. It is difficult to reflect the simulation authenticity
as to explore the mechanical behaviors of asphalt mixtures or with the PFC2D under different conditions and research
cement. The difference between the continuum-based proposals. Thereby, the PFC3D programs are produced and
method and DEM is the system topology and construction introduced with more accurate simulation results, and the
manner, as well as the solution strategies and material programs restore the actual actions between materials. The
behaviors. During the development of discrete elements, 2D and 3D discrete element models of asphalt mixtures are
the DEM algorithms were invented and developed for established, respectively, and the two-type models are
different scenarios of materials at the micro scales, such as compared from these aspects, such as the calculation
the boundary types, contact-point constitutions, force speed and simulation effect compared with the laboratory-
transfers, and calculation, sample generation, particle measured data. Some example models can be described as
shape and types, fluid flow mechanics, geomaterial follows: Prediction for the dynamic modulus of the asphalt
micromechanics, etc. Many algorithm blocks and methods mixture with DEMs (You et al., 2009; Adhikari and You,
were investigated to fully satisfy different research proposes 2010); Influence of voids and its distribution laws in asphalt
and objectives, such as the discrete fracture network, the mixtures with DEMs (You et al., 2010). After comparison, the
Navier-Stokes equation, deformation analysis, stiffness effect of 3D model simulation is better than that of the 2D
matrix, heat conduction, stress wave propagation, various model in most cases. It is relatively close to the laboratory
moduli, tensile test, etc. There are many advantages and results.
disadvantages for the laboratory test method and the discrete (3) The generation methods or algorithms of DEM samples were
element numerical simulation method to evaluate the discussed here, like the image-based, user-defined, idealized
performance and analyze the interaction between and random models. Three methods had their advantages
elements. The virtual laboratory test plan is the main and the application scenarios depend on the research
reason for us to use the simulation to conduct the test; purposes. Integrating the current simulation databases and
Another reason is that the simulation can solve the computer resources, the image-based method seemed to be

Frontiers in Materials | www.frontiersin.org 15 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

more likely to be real samples. However, the image-based (2) The user-defined and random models also have more
model was restricted by the image capturing technology and potential to be developed for virtual laboratory tests, 1)
image numbers or databases. More images of samples needed The difficulty of the methods is the useful and universal
to be captured and processed under different conditions and databases for the virtual sample generation; 2) The shape and
these are not transferable. Compared to the idealized models, texture properties of aggregates can be precisely described
the other two methods (image-based and user-defined and added in the virtual samples, as well as other properties
models) are commonly used in 2D or 3D discrete element in the models; 3) The interaction between materials, like
programs. Through the study of the literature (Yang et al., asphalt and aggregates in asphalt mixtures, can be accurately
2016), the image-based DEM models are inseparable from expressed in the models; 4) The physical and chemical
the scanning equipment and image processing technology, reactions can be strictly presented in the models; 5) The
and these are needed to be completed in the laboratory with DEM calculation speed and element numbers also affect the
the high cost and long time-consuming processing. The simulation effect and generalization. 6) the contact model
advantage of this method is that the simulation effect is between the elements may be updated for the constitutive
close to the real status for materials, which can accurately models in DEM model, and this update may provide an
present the shape, position, and distribution of aggregates, option for the solution of the plastic deformation problems in
and also reflect the real structures of the asphalt mixtures. In the DEM simulation. It helps to understand behaviors of
contrast, the user-defined model method is used to build the visco-elastic plastic materials. These points are needed to be
model according to the basic data of the simulation test in the considered for us to generate good models, as well as other
discrete element program with the help of the “Fish” possible factors.
language in the program. It is not expensive and also not
time-consuming, as well as the independence of the Therefore, the DEM is a promising tool for us to explore the
laboratory restriction. However, sometimes, the accuracy micro behaviors of asphalt mixtures. It is also affordable for
of the simulation data is less than that of the image-based researchers to conduct the tests in the virtual laboratory. It can
model method. It is difficult to fully show the internal achieve different test simulations at different dimensions and
structures of the aggregates in the asphalt mixtures. explore mechanisms under extreme conditions. These operations
can not be easily realized in the laboratory or fields. The
researchers benefit from these advantages and program the
6 CONCLUSION simulation with their purposes. The research results can
directly guide the industry application and production, and
After the above discussions, we can understand that: also, accelerate the research development for the
engineering field.
(1) The application of DEMs in pavement engineering, especially
in asphalt mixtures, was also studied from the function
realization to the mechanism discussion. Many contact/ AUTHOR CONTRIBUTIONS
connectivity patterns and behaviors were generated and
developed for the interaction discovery of asphalt and HY and MX wrote the manuscript; JL and YL helped revise the
aggregates. Many tests of asphalt mixtures were simulated manuscript; JJ and ZY provided the technical support for the
to explain and investigate the performance by the DEM, like manuscript.
the splitting test (indirect tensile test), dynamic modulus test,
three-point, and four-point beam test, compressive test, etc.,
Those were demonstrated partially in this manuscript. There FUNDING
are still more questions and mechanisms needed to be
discovered and explored in the DEM simulation, for The authors appreciate the financial support of the Hunan
instance, 1) interaction between the water and mixtures Department of Transportation (No. 202152) and Hunan
under different conditions, 2) damage patterns of different Expressway Group Co. Ltd. The authors also appreciate
shapes and forces under different conditions in the the funding support from the National Natural Science
pavement, 3) contribution rates of different aggregates and Foundation of China (No. 51778038), Beijing Key
asphalt in asphalt mixtures under different conditions, 4) Laboratory of Traffic Engineering (No. 2018BJUT-
void distributions in asphalt mixtures and formation laws, 5) JTJD007), and Beijing Municipal Commission of
water propagations in asphalt mixtures, etc., Sometimes, Transportation (Grant Number: 2018-kjc-01-213). Any
these questions may be divided into different pieces and opinion, finding, and conclusion expressed in this paper
can be solved from different views and sights. In addition, it is are those of the authors and do not necessarily represent
common that the DEM simulation explores from 2D to 3D the view of any organization. The funder was not involved in
for the reality demonstration and links the microscale to the study design, collection, analysis, interpretation of data,
macro scale. Also, there are some different advantages in the the writing of this article or the decision to submit it
2D and 3D simulations and researchers need to integrate for publication. All authors declare no other competing
your aims and thoughts to make decisions. interests.

Frontiers in Materials | www.frontiersin.org 16 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

REFERENCES Kusumawardani, D. M., and Wong, Y. D. (2020). The Influence of Aggregate Shape
Properties on Aggregate Packing in Porous Asphalt Mixture (PAM).
Construction Building Mater. 255, 119379. doi:10.1016/j.conbuildmat.2020.
Adhikari, S., and You, Z. (2010). 3D Discrete Element Models of the Hollow 119379
Cylindrical Asphalt concrete Specimens Subject to the Internal Pressure. Int. J. Li, Q., Wu, C., Fang, Y., Zhong, M., Huang, H., and Han, Q. (2021). Performance
Pavement Eng. 11. doi:10.1080/10298436.2010.489114 Research and Application of SMA-13 Asphalt Mixture with Plant Mixed Hot
Adhikari, S., and You, Z. (2011). Investigating the Sensitivity of Aggregate Size Recycling[J/OL]. Highw. Eng., 1–12.
within Sand Mastic by Modeling the Microstructure of an Asphalt Mixture. J. Li, Z., Jing, H., and Wang, Z. (2020). Multi-factor Interaction of Asphalt Mixture
Mater. Civil Eng. 23. doi:10.1061/(asce)mt.1943-5533.0000212 Volume Characteristics Based on Discrete Element Method[J]. J. Guangxi Univ.
Chang, M., Pei, J., and Zhang, J. (2011). “Influences of Mesoscopic Parameters on (Natural Sci. Edition) 45, 1301–1314.
Crack Resistance of Asphalt Mixture,” in ICCTP 2011: Towards Sustainable Liang, H., Shi, L., Wang, D., Xiao, X., and Deng, K. (2021). Influence of Graded
Transportation Systems, 3324–3332. doi:10.1061/41186(421)329 Coarse Aggregate Content and Specific Surface Area on the Fracture Properties
Chen, J., Pan, T., Chen, J., Huang, X., and Lu, Y. (2012). Predicting the Dynamic of Asphalt Mixtures Based on Discrete Element Simulations and Indoor Tests.
Behavior of Asphalt concrete Using Three-Dimensional Discrete Element Construction Building Mater. 299, 123942. doi:10.1016/j.conbuildmat.2021.
Method. J. Wuhan Univ. Technol. Mater. Sci. Edition 27. doi:10.1007/ 123942
s11595-012-0470-y Liang, H., Wang, D., Deng, Z., Liang, X., and Xu, C. (2019). Application of
Cundall, P. (1974). A Computer Model for Rock Mass Behavior Using Interactive Cohesive Zone Model in Discrete Element Simulation of Foamed Asphalt
Graphics for the Input and Output of Geometrical Data. A Report Prepared Mixture[J]. J. Hunan Univ. (Natural Sciences) 46, 115–123.
under Contract Number DACW 45-74-C-006. Washington, D.C.: The National Ling, T., Yang, Y., Zhao, C., Zhu, L., and Chen, Q. (2020). Design of Asphalt
Academies of Sciences, Engineering, and Medicine. Mixture with Low Voidage Based on PFC3D[J]. J. Chongqing Jiaotong Univ.
Cundall, P. A. (1971). “A Computer Model for Simulating Progressive, Large-Scale (Natural Science) 39, 73–80.
Movement in Blocky Rock System,” in Symposium of the International Society Liu, C., Le, T., Shi, B., and Zhu, Y. (2020). Discussion on Three Major Problems of
for Rock Mechanics. Engineering Application of the Particle Discrete Element Method. Chin. J. Rock
Cundall, P. A. (1978). BALL–A Program to Model Granular media Using the Mech. Eng. 39, 1142–1152.
Distinct Element Method. London, United Kingdom: Dames & Moore. Liu, G., Dai, J., Liu, Y., Gao, H., and Peng, Y. (2018). Numerical Simulation of
Advanced Technology Group". Technical Note No. TN-LN. Aggregate Uniformity Influence on Low Temperature Splitting Strength of
Cundall, P. A., and Strack, O. D. L. (1979). A Discrete Numerical Model for Asphalt Mixtures[J]. Low Temperature Architecture Tech. 40, 12–16.
Granular Assemblies. Géotechnique 29, 47–65. doi:10.1680/geot.1979.29.1.47 Liu, Y., Dai, Q., and You, Z. (2009). Viscoelastic Model for Discrete Element
Dan, H.-C., Zhang, Z., Chen, J.-Q., and Wang, H. (2018). Numerical Simulation of Simulation of Asphalt Mixtures. J. Eng. Mech. 135. doi:10.1061/(asce)0733-
an Indirect Tensile Test for Asphalt Mixtures Using Discrete Element Method 9399(2009)135:4(324)
Software. J. Mater. Civ. Eng. 30, 04018067. doi:10.1061/(asce)mt.1943-5533. Liu, Y. (2011). Discrete Element Methods for Asphalt concrete: Development and
0002252 Application of User-Defined Microstructural Models and a Viscoelastic
Ding, X., Ma, T., and Huang, X. (2019). Discrete-Element Contour-Filling Micromechanical Model. Houghton, MI: Michigan Technological University.
Modeling Method for Micromechanical and Macromechanical Analysis of Liu, Y., and You, Z. (2011a). Accelerated Discrete-Element Modeling of Asphalt-
Aggregate Skeleton of Asphalt Mixture. J. Transp. Eng. Part. B: Pavements Based Materials with the Frequency-Temperature Superposition Principle. J.
145, 04018056. doi:10.1061/jpeodx.0000083 Eng. Mech. 137, 355–365.
Du, J., Ren, D., Ai, C., and Qiu, Y. (2021). Study on Low Temperature Crack Liu, Y., and You, Z. (2011b). Accelerated Discrete-Element Modeling of Asphalt-
Propagation and Evolution of Asphalt Mixture[J/OL]. J. Building Mater., 1–10. Based Materials with the Frequency-Temperature Superposition Principle. J.
Gao, Y. (2021). Experimental Study on Rheological Properties of Sulfur SBS Eng. Mech. 137, 355–365. doi:10.1061/(asce)em.1943-7889.0000234
Modified Asphalt and Pavement Performance of its Mixture[J]. Synth. Liu, Y., and You, Z. (2008). “Simulation of Cyclic Loading Tests for Asphalt
Mater. Aging Appl. 50, 67–69+22. Mixtures Using User Defined Models within Discrete Element Method,” in
Ge, H., Quezada, J. C., Houerou, V. L., and Chazallon, C. (2021). Three- GeoCongress 2008: Characterization, Monitoring, and Modeling of Geosystems,
dimensional Simulation of Asphalt Mixture Incorporating Aggregate Size 742–749.
and Morphology Distribution Based on Contact Dynamics Method. Liu, Y., You, Z., and Zhao, Y. (2012). Three-dimensional Discrete Element
Construction Building Mater. 302, 124124. doi:10.1016/j.conbuildmat.2021. Modeling of Asphalt concrete: Size Effects of Elements. Construction
124124 Building Mater. 37, 775–782. doi:10.1016/j.conbuildmat.2012.08.007
Ghafoori Roozbahany, E., Partl, M. N., and Jelagin, D. (2019). Modelling the Flow Lu, J. (2020). Study on Numerical Simulation of Asphalt Mixture Based on CT
of Asphalt under Simulated Compaction Using Discrete Element. Construction Technology and Three-Dimensional Discrete Element Method[D]. Master.
Building Mater. 227, 116432. doi:10.1016/j.conbuildmat.2019.07.158 Guangzhou, China: South China University of Technology.
Guo, H., and Chen, S. (2014). Discrete Element Numerical Analysis of Splitting Ma, T., Zhang, D.-y., Zhang, Y., Zhao, Y.-l., and Huang, X.-m. (2016b).
Test of Open-Graded Large-Size Asphalt Macadam Mixture[J]. J. Highw. Microstructure Modeling and Virtual Test of Asphalt Mixture Based on
Transportation Res. Dev. 31, 22–26+40. Three-Dimensional Discrete Element Method. J. Cent. South. Univ. 23,
Guo, H., and Chen, S. (2015). Numerical Analysis on the Splitting Test of Open- 1525–1534. doi:10.1007/s11771-016-3204-z
Graded Large Stone Asphalt Mixture Based on Discrete Element Method. J. Ma, T., Zhang, D., Zhang, Y., Wang, S., and Huang, X. (2018). Simulation of Wheel
Highw. Transportation Res. Dev. 9. doi:10.1061/jhtrcq.0000434 Tracking Test for Asphalt Mixture Using Discrete Element Modelling. Road
Guo, Z. (2020). Influence of Void Distribution Based on Discrete Element Method Mater. Pavement Des. 19, 367–384. doi:10.1080/14680629.2016.1261725
on Shear Fatigue Life of Asphalt Mixture[J]. Low Temperature Architecture Ma, T., Zhang, D., Zhang, Y., Zhao, Y., and Huang, X. (2016a). Effect of Air Voids
Tech. 42, 11–14. on the High-Temperature Creep Behavior of Asphalt Mixture Based on Three-
Hou, S., Zhang, D., Huang, X., and Zhao, Y. (2015). Investigation of Micro- Dimensional Discrete Element Modeling. Mater. Des. 89, 304–313. doi:10.1016/
mechanical Response of Asphalt Mixtures by a Three-Dimensional Discrete j.matdes.2015.10.005
Element Model. J. Wuhan Univ. Technol.-Mat. Sci. Edit. 30, 338–343. doi:10. Mao, Z. (2018). The Research on Creep of Asphalt Mortar Based on Three-
1007/s11595-015-1150-5 Dimensional Discrete Element[J]. Low Temperature Architecture Tech. 40,
Huang, Z.-y., Yang, Z.-x., and Wang, Z.-y. (2008). Discrete Element Modeling of 4–6+13.
Sand Behavior in a Biaxial Shear Test. J. Zhejiang Univ. Sci. A. 9, 1176–1183. Meng, L., Zhang, W., and Wang, X. (2021). Research on the Effect of Chlorine Salt
doi:10.1631/jzus.a0720059 Snow-Melting Agent on Asphalt concrete Performance[J]. J. Qingdao Univ.
Jia, H., Sheng, Y., Lv, H., Kim, Y. R., Zhao, X., Meng, J., et al. (2021). Effects of Tech. 42, 45–51.
Bamboo Fiber on the Mechanical Properties of Asphalt Mixtures[J]. Nian, T., Ge, J., Li, P., Wang, M., and Mao, Y. (2021). Improved Discrete Element
Construction Building Mater. 289. doi:10.1016/j.conbuildmat.2021.123196 Numerical Simulation and experiment on Low-Temperature Anti-cracking

Frontiers in Materials | www.frontiersin.org 17 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

Performance of Asphalt Mixture Based on PFC2D. Construction Building Mining Engineering——Proceedings of the Academic Conference on Rock
Mater. 283, 122792. doi:10.1016/j.conbuildmat.2021.122792 Mechanics and Engineering Applications in North China, 502–508.()
Pei, J., Chang, M., Chen, S., and Wang, B. (2010). Numerical Simulation of Indirect Wang, Y., and Xing, J. (1989). “Improvement and Application of Discrete Element
Tensile Test for Asphalt Mixture[J]. J. Chang’an Univ. (Natural Sci. Edition) Method[A],” in Engineering Application of Numerical Methods in Rock and Soil
30, 6–10. Mechanics——Proceedings of the Second National Symposium on Numerical
Peng, Y., and Bao, J. (2018). Comparative Study of 2D and 3D Micromechanical Calculation and Model Experiments in Rock Mechanics, 254–259.
Discrete Element Modeling of Indirect Tensile Tests for Asphalt Mixtures. Int. Wang, Y., and Xing, J. (1991). “New Progress of Numerical Methods and Their
J. Geomechanics 18. doi:10.1061/(asce)gm.1943-5622.0001155 Engineering Applications[A],” in New Progress in Rock Mechanics, 168–192.()
Peng, Y., Bao, J., and Wang, Z. (2016). A Comparison of Two-Dimensional and Wu, W., Si, L., and Lu, J. (2021). Influence of Aggregate on Study on Shear
Three-Dimensional Micromechanical Discrete Element Modeling of the Performance of Asphalt Mixture Based on Virtual Uniaxial Penetration Test[J].
Splitting Tests for Asphalt Mixtures. DEStech Trans. Eng. Tech. Res. J. Highw. Transportation Res. Dev. 38, 1–8.
Peng, Y., Gao, H., Wan, L., and Liu, G. (2019). Numerical Simulation of Influence Xia, Q., Gao, J., Zhang, Q., Li, Z., Yang, F., and Luo, H. (2021). Low-temperature
Factors of Splitting Strength of Asphalt Mixtures[J]). J. Jilin Univ. (Engineering Performance of Composite Modified Hard Asphalt Used in High Modulus
Tech. Edition) 49, 1521–1530. Asphalt concrete[J/OL]. Changchun, China: Journal of Jilin University
Peng, Y., and Sun, L. (2015a). Aggregate Distribution Influence on the Indirect Engineering and Technology Edition, 1–9.
Tensile Test of Asphalt Mixtures Using the Discrete Element Method. Int. J. Xing, J., and Wang, Y. (1988). Discrete Element Simulation of Ore Drawing by
Pavement Eng. 18. doi:10.1080/10298436.2015.1121778 Caving Method in Mining[J]. Shenyang, China: Journal of Northeast University
Peng, Y., and Sun, L. J. (2015b). Aggregate Distribution Influence on the Indirect of Technology, 148–153.
Tensile Test of Asphalt Mixtures Using the Discrete Element Method [J]. Int. J. Xing, J., and Wang, Y. (1990). Improvement of Discrete Element Method and its
Pavement Eng. 18, 1–14. doi:10.1080/10298436.2015.1121778 Applications in Granular Media[J]. Chin. J. Geotechnical Eng., 51–57.
Peng, Y., Yang, H., Lu, X., and Li, Y. (2021). Effect of Void Characteristics on Xiong, Y. (2013). Analysis of the Influence of Aggregate Properties on Splitting Test
Virtual Shear Fatigue Life of Asphalt Mixtures Using Discrete Element Method of Asphalt Concrete Based on Discrete Element Method[J]. J. China Foreign
[J]. J. Jilin Univ. (Engineering Tech. Edition) 51, 956–964. Highw. 33, 318–322.
Shan, J., Du, Y., Fan, D., Guo, L., and Yang, X. (20192019). Determination of Yang, X., You, Z., Wang, Z., and Dai, Q. (2016). Review on Heterogeneous Model
Parameters of the Discrete Element Bond Model for Asphalt Mixture Based on Reconstruction of Stone-Based Composites in Numerical Simulation. Construction
Splitting Test. Adv. Civil Eng. doi:10.1155/2019/5047214 Building Mater. 117, 229–243. doi:10.1016/j.conbuildmat.2016.04.135
Sheng, F., Hou, L., Wang, Y., Li, Z., and Shan, J. (2018). Studying on the Parameters Yao, L., Gao, W., Tao, F., and Wang, H. (2021). Effects of Bayer Red Mud as Filler
of Bond-Model in DEM for Asphalt Mixture. IOP Conf. Ser. Mater. Sci. Eng. on Properties of Asphalt Mixture[J/OL]. Highway 66, 61–68.
392. doi:10.1088/1757-899x/392/2/022004 You, Z., Adhikari, S., and Dai, Q. (2010). Air Void Effect on an Idealised Asphalt
Shi, Y. (2020). Simulation Study on Split Test of Asphalt Mixture with Flat- Mixture Using Two-Dimensional and Three-Dimensional Discrete Element
Elongated Particles Using Discrete Element Method[D]. Master. Harbin, Modelling Approach. Int. J. Pavement Eng. 11. doi:10.1080/10298436.2010.
China: Harbin Engineering University. 488737
Song, W. (2021). Influence of Polymer Fiber and Glass on the Mechanical You, Z., Adhikari, S., and Dai, Q. (2008). Three-Dimensional Discrete Element
Properties of Asphalt Mixture[J/OL]. Highway 66, 29–33. Models for Asphalt Mixtures. J. Eng. Mech. 134. doi:10.1061/(asce)0733-
Wan, L. (2016). Study on Asphalt Mixture Splitting Test Using Cohesive Zone Model 9399(2008)134:12(1053)
and Three-Dimensional Discrete Element Method[D]. Master. Zhejiang, China: You, Z., Adhikari, S., and Kutay, M. E. (2009). Dynamic Modulus Simulation of the
Zhejiang University. Asphalt concrete Using the X-ray Computed Tomography Images. Mater.
Wang, H., Huang, W., Cheng, J., and Ye, G. (2021). Mesoscopic Creep Mechanism Structures 42. doi:10.1617/s11527-008-9408-4
of Asphalt Mixture Based on Discrete Element Method. Construction Building You, Z., Liu, Y., and Dai, Q. (2011). Three-Dimensional Microstructural-Based
Mater. 272, 121932. doi:10.1016/j.conbuildmat.2020.121932 Discrete Element Viscoelastic Modeling of Creep Compliance Tests for Asphalt
Wang, W., and Li, X. (2005). A Review on Fundamentals of Distinct Element Mixtures. J. Mater. Civil Eng. 23. doi:10.1061/(asce)mt.1943-5533.0000038
Method and its Applications in Geotechnical Engineering[J]. Beijing, China: You, Z., and Liu, Y. (2010). Three-Dimensional Discrete Element Simulation of
Geotechnical Engineering Technique, 177–181. Asphalt Concrete Subjected to Haversine. Loading an Application of the
Wang, X., Zheng, D., and Yu, C. (2019). Application of Discrete Element in the Frequency-Temperature Superposition Technique. Road Mater. Pavement
Teaching of "Asphalt and Asphalt Mixture. Education Modernization 6, Des. 11, 273–290. doi:10.3166/rmpd.11.273-290
218–220. Yu, D. (2015). Splitting Test Modeling of Asphalt Mixture Using Particle Flow
Wang, Y. (1990). “Development of Discrete Element Method and its Application in Code[J]. Mod. Transportation Tech. 12, 1–4.
Geotechnical Engineering[A],” in Engineering Application of Numerical Zhang, D., and Li, Y. (2017). Numerical Simulation of Asphalt Concrete’s Porosity
Methods in Rock and Soil Mechanics——Proceedings of the Second National Variation under Dynamic Loading[J]. J. Xuzhou Inst. Tech. (Natural Sci.
Symposium on Numerical Calculation and Model Experiments in Rock Edition) 32, 73–76.
Mechanics, 44–54. Zhang, H., Lian, H., and Yang, D. (2015). Split Test and Numerical Simulation of
Wang, Y. (1986b). Discrete Element Method and its Application in Rock mechanics Crumb Rubber-Modified Asphalt Mixture[J]. Mater. Rep. 29, 145–148.
[J]. Shenyang, China: Metal Mine, 13–17+15. Zhang, Y., Ma, T., Ding, X., Chen, T., Huang, X., and Xu, G. (2018). Impacts of Air-
Wang, Y. (1986a). Discrete Element Method—A Numerical Method Suitable for Void Structures on the Rutting Tests of Asphalt concrete Based on Discretized
Analysis in Jointed Rocks. Proc. first Natl. Symp. Numer. calculation Model Test Emulation. Construction Building Mater. 166, 334–344. doi:10.1016/j.
Rock Mech. 6. conbuildmat.2018.01.141
Wang, Y., Liu, G., and Xing, J. (1987). Research on Applications of Discrete Element Zhang, Y., and Zhang, B. (2016). “Blast Furnace for Bell Distribution Chute
Method in Ore Drawing for Caving system[J]. Beijing, China: Non-ferrous Research Based on StreamDEM[A],” in Proceedings of the 12th China CAE
metal, 20–26. Engineering Analysis Technology Annual Conference, 141–143.
Wang, Y., and Xing, J. (1993a). Discrete Element Method and its Application in Zhang, Y., Zhong, X., Zhong, Y., and Zhang, B. (2017). “Study on the First Layer of
Coal Mining (Lecture 1)[J]. Coal Mining Tech., 55–58. Blast Furnace Fabric Based on StreamDEM[A],” in Proceedings of the 11th
Wang, Y., and Xing, J. (1993b). Discrete Element Method and its Application in China Iron and Steel Annual Conference-S01 (Beijing, China: Ironmaking and
Coal Mining (Lecture 2)[J]. Coal Mining Tech., 56–58. Raw Materials), 830–834.
Wang, Y., and Xing, J. (1986). Discrete Element Method and its Applications in Zhong, X., and Zhang, Y. (2021). Influence of Fiber Type on Road Performance of
Mine Pressure research[J]. Fuxin, China: Journal of Fuxin Institute of Mining High RAP Content Asphalt Mixture[J]. Highways & Transportation in Inner
Technology, 1–12. Mongolia, 29–33.
Wang, Y., and Xing, J. (1990). “Fully Deformed Discrete Element Method and its Zhou, J. (2016). Laboratory Test and Numerical Simulation for Dynamic Modulus
Application in Mining[A],” in Rock Mechanics Issues in Hydropower and of Asphalt Mortar[J]. Mod. Transportation Tech. 13, 1–5.

Frontiers in Materials | www.frontiersin.org 18 April 2022 | Volume 9 | Article 879245

Yao et al. Discrete Elements for Asphalt Mixtures

Zhou, X., Chen, S., Ge, D., Jin, D., and You, Z. (2020). Investigation of Asphalt the publisher, the editors and the reviewers. Any product that may be evaluated in
Mixture Internal Structure Consistency in Accelerated Discrete Element this article, or claim that may be made by its manufacturer, is not guaranteed or
Models. Construction Building Mater. 244, 118272. doi:10.1016/j. endorsed by the publisher.
conbuildmat.2020.118272
Copyright © 2022 Yao, Xu, Liu, Liu, Ji and You. This is an open-access article
Conflict of Interest: The authors declare that the research was conducted in the distributed under the terms of the Creative Commons Attribution License (CC BY).
absence of any commercial or financial relationships that could be construed as a The use, distribution or reproduction in other forums is permitted, provided the
potential conflict of interest. original author(s) and the copyright owner(s) are credited and that the original
publication in this journal is cited, in accordance with accepted academic practice.
Publisher’s Note: All claims expressed in this article are solely those of the authors No use, distribution or reproduction is permitted which does not comply with
and do not necessarily represent those of their affiliated organizations, or those of these terms.

Frontiers in Materials | www.frontiersin.org 19 April 2022 | Volume 9 | Article 879245