Construction and Building Materials 469 (2025) 140448

Contents lists available at ScienceDirect

Construction and Building Materials

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

Research on dynamic strain distribution patterns of asphalt pavement

under the influence of hydro-thermo-mechanical interactions
Pengjun Dong a , Xuejuan Cao b,* , Boming Tang a,*
a
School of Civil Engineering, Chongqing Jiaotong University, Chongqing 400074, China
b
School of Materials Science and Engineering, Chongqing Jiaotong University, Chongqing 400074, China

A R T I C L E I N F O A B S T R A C T

Keywords: Strain is a crucial indicator of the design of the asphalt pavement. To qualitatively and quantitatively study the
Asphalt pavement distribution laws of dynamic strains under the coupling of multiple factors, this paper constructs single stress,
Strain component hydro-mechanical, thermo-mechanical and hydro-thermo-mechanical models, which study the state and nu­
Principal strain
merical value differences of dynamic strains in various physical fields, and explores the distribution character­
Numerical simulation
Spatial distribution
istics of dynamic strain components within the pavement zone under coupling of hydro, thermal, and mechanical
processes. The results show that under different physical fields, the absolute values of the peak dynamic strains
(lateral, longitudinal and vertical) at the wheel load center decrease as pavement depth increases, but the nu­
merical values and waveform change laws of dynamic strains in different directions are different. The peak
absolute value of the εyy is the highest, followed by the εxx, and the εzz is the smallest. Among the three shear
strain components, εyz has the largest peak absolute value, followed by εxy, and εxz has the smallest peak absolute
value. The numerical differences of the ε1 peak in various physical fields are large, and the dominant factors of
negative and positive peak are different. At Y= 0 m, the negative peak of the ε1 under hydro-thermo-mechanical
coupling is 2.4 times that under the action of hydro-mechanical field. The negative peak of the ε1 under the
action of thermo-mechanical field is 2.5 times that under single stress field. Along the road horizontally and
longitudinally, the εxx and εyy both show alternating positive and negative changes, suggesting that the road
material alternates between compression and tension, which is easy to cause cracks. With the increase in
pavement depth, the superposition phenomenon of the dynamic strain peak values is more obvious, and the
extent of dynamic strain action in the lateral and longitudinal directions of the pavement is widening. The
research results have certain reference value for the design of long-life asphalt pavement.

1. Introduction When designing asphalt pavements, the strain at the base of the asphalt
mixture layers is frequently utilized as a design indicator for asphalt
Asphalt pavement experiences the simultaneous influences of tem­ pavement, which is crucial for accurately predicting the longevity of the
perature, moisture, and load during its service life, which leads to pavement under repeated stress. However, due to factors such as the test
extremely complex strain behavior in the pavement structure under the method, specimen size, and stress level, even for the same pavement
interaction of various factors. However, the accurate calculation of material, the modulus values in pavement design in different countries
strain has a vital impact on forecasting the pavement’s design lifespan. are different [1–3].
In recent times, with the accelerated construction of green, low-carbon, Many scholars tested the modulus of bituminous mixtures and
durable, and long-life pavement, the analysis and forecasting of the discovered that the direct tensile modulus was lower than the uniaxial
mechanical behavior of bituminous pavements under the influence of compression dynamic modulus [4–6]. Islam et al. [7] conducted uniaxial
various factors have emerged as a significant research focus in road compression tests on on-site drill core samples of asphalt pavement and
engineering. found that porosity and asphalt content have a substantial effect on the
The stiffness of each structural layer in the pavement is closely uniaxial compression modulus. Subsequently, numerous researchers
associated with the calculated value of the pavement dynamic response. [8–13] examined how strain levels, temperature, loading frequency,

* Corresponding authors.
E-mail addresses: caoxuejuan2023789@163.com (X. Cao), Tangboming6210@163.com (B. Tang).

https://doi.org/10.1016/j.conbuildmat.2025.140448
Received 11 October 2024; Received in revised form 18 December 2024; Accepted 14 February 2025
Available online 21 February 2025
0950-0618/© 2025 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 1. Road structure model.

Table 1
Foundation soil parameters.
Name Thickness Modulus Poisson’s ratio Density Damping ratio Thermal conductivity Specific heat capacity Thermal expansion coefficient
(m) (MPa) (kg/m3) (J.m− 1.s− 1.℃− 1) (J.kg− 1.℃− 1) (10− 5.℃− 1)

Clay 10 60 0.4 1800 0.05 1.30 860 0.45

Table 2
Pavement material parameters.
Layers Thickness Modulus Poisson’s Density Damping Thermal Specific heat Thermal expansion Permeability coefficient
name (m) (MPa) ratio (kg/ ratio conductivity capacity coefficient （m/s）
m3) (J.m− 1.s− 1.℃− 1) (J.kg− 1.℃− 1) (10− 5.℃− 1)
3
Upper 0.04 7200 0.40 2526 0.05 1.82 1105 3.34 1.80 × 10−
layer
6
Middle 0.06 10500 0.25 2480 0.05 1.95 815 3.34 1.20 × 10−
layer
6
Lower 0.08 11000 0.25 2510 0.05 1.99 880 3.34 1.38 × 10−
layer
Base 0.38 10000 0.25 2400 0.05 1.20 895 1.00 ​
layer
Subbase 0.20 3000 0.25 2200 0.05 1.30 910 1.10 ​
layer
Subgrade 1.24 60 0.40 1800 0.05 1.30 860 0.45 ​
Solar radiation absorption 0.90
Pavement surface emissivity 0.81
Absolute zero temperature (℃) − 273
Stefan-Boltzmann constant (W⋅m− 2⋅K− 4) 5.6697 × 10− 8

Fig. 2. Vehicle load diagram.

additives, and interlayer contact conditions influence the modulus, of­ avoiding the artificial selection of material modulus design parameters
fering valuable insights for determining modulus values in the devel­ during pavement structure design. The tensile modulus and compressive
opment of asphalt pavement. To obtain a uniform characterization of the modulus of pavement materials have different characteristics [15]. To
dynamic modulus of asphalt mixtures across various testing conditions. examine the disparity between the tensile modulus and the compressive
Lv et al. [14] characterized the material modulus under strength, Jin et al. [16] determined the compressive and tensile moduli
three-dimensional stress conditions based on the equivalent stress ratio, of asphalt mixtures using bimodulus theory. They subsequently explored

2
P. Dong et al. Construction and Building Materials 469 (2025) 140448

speeds. Tan et al. [30] considered the nonlinear gradient distribution of

temperature when modeling and predicted the distribution law of strain
in the pavement. To further clarify the damage of the critical speed to
the pavement, the critical stress-strain behavior at the base of the
asphalt layer under different vehicle speeds was simulated [31]. Assogba
et al. [32] examined the dispersion of three-dimensional strain in the
road across the lateral and depth orientations. The roughness of the
pavement will affect the structural strain. Liu et al. [33] discovered that
lengthwise, lateral strain, and deflection all decrease with higher vehicle
speeds and greater pavement roughness grades. Additionally, the dy­
namic response of the entire pavement system is influenced by the
constitutive model of the base and subgrade materials. Saad et al. [34]
examined how subgrade strength, thickness, and quality affect fatigue
and rutting strains. Temperature is essential in the formation and
development of pavement diseases. Ren et al. [35] used ANSYS to
numerically simulate the deformation of the real-world environmental
factors in Xinjiang under dynamic load and pavement coupling fields,
and explored the damage mechanism of temperature on the pavement.
Fig. 3. Tire pressure time history curve.
Li et al. [36] constructed the two-dimensional finite element
thermo-hydro-mechanical coupling model. By studying the dynamic
response of the pavement under multi-field coupling, they found that
compared with dry asphalt pavement, pore water can reduce the effec­
tive stress caused by vehicle load, but increase the residual stress when
the vehicle is loaded. Liu et al. [37] found that frost causes the tem­
perature and moisture in the pavement in cold areas to be redistributed,
thereby affecting the mechanical response of the pavement. Sun et al.
[38] constructed the three-dimensional multi-field coupling model to
study the dynamic response of asphalt pavement under dynamic load
Fig. 4. Vertical schematic diagram. and temperature load. The study found that pore water pressure plays an
important role in the destruction of asphalt pavement, and found that
temperature has a greater impact on the lateral dynamic stress of the
the alterations in relation to the tensile and compressive moduli, as well
upper layer. Dong et al. [39,40] constructed the three-dimensional
as their fatigue performance characteristics, using indirect tensile tests.
hydro-thermo-mechanical fully coupled models, analyzed the differ­
The material modulus is affected by factors such as testing methods,
ences in stress components under different physical fields, and explained
force modes, and boundary conditions, and cannot fully reflect the
the mechanism of temperature on excess pore water pressure. Finally,
actual mechanical state of asphalt pavement. Therefore, Wang et al. [17]
the formation mechanism of pavement cracks under multi-field coupling
measured it using the FWD test, the pavement surface deflection is uti­
was analyzed from the perspective of stress principal axis rotation.
lized to reverse-engineer the structural layer modulus. The aforemen­
In summary, existing studies have amassed extensive findings
tioned research on modulus provides valuable reference for the
regarding the mechanical action of road structures under vehicle loads,
calculation of strain in asphalt pavement design.
providing valuable insights for asphalt pavement design. However,
To calculate the mechanical response of asphalt pavement from the
current studies have insufficiently addressed the temporal and spatial
structural perspective, numerous domestic researchers have embedded
variations of dynamic strain in asphalt pavement structures due to the
various types of sensors within the road structure to examine the effects
combined effects of hydro-thermo-mechanical conditions and structural
of factors such as vehicle speed, axle weight, and temperature on
self-weight. Therefore, this paper aims to qualitatively and quantita­
pavement structure strain. Their findings indicate that the strain peak of
tively examine the distribution characteristics of strain components and
the pavement structure rises with increasing load and diminishes with
principal strains influenced by temperature, moisture, load, and struc­
higher speeds [18–22]. The mechanical response of the pavement in­
tural self-weight. This paper focuses on the commonly used semi-rigid
creases when the temperature rises [23,24]. The University of Waterloo
base asphalt pavement in China as its research subject. First, taking
[25] discovered through field measurements that the damage resulting
into account the preliminary static stress field generated self-weight and
from pavement strain due to temperature may exceed that caused by
temperature, the single stress, thermo-mechanical, hydro-mechanical,
load. Pan et al. [26] selected different asphalt pavement structures and
and hydro-thermo-mechanical models are constructed. Then, the dif­
utilized fiber Bragg grating optical sensing technology to perform in-situ
ferences in time-history distribution of dynamic strain in the pavement
strain tests under both static and dynamic loading conditions, examining
structure across various physical fields are analyzed both qualitatively
the characteristics of strain field variations in the horizontal and depth
and quantitatively. Finally, the distribution characteristics of the dy­
directions of the pavement. Wang et al. [27,28] conducted continuous
namic strain components throughout the pavement space, influenced by
tracking and observation of the mechanical responses of road structures,
hydro-thermo-mechanical effects and structural self-weight, are exam­
including internal stress and strain. The strain-temperature relationship
ined. The findings will provide valuable insights for the construction of
curve showed that there was a tension-compression transition point. As
durable pavements.
the temperature increased, the position of the compression-tension
transition would gradually move away from the road surface and
2. Model overview
downward. On-site testing is costly and has the characteristics of high
randomness of vehicle loads. Based on this, researchers also use nu­
2.1. Model introduction
merical methods to study the mechanical distribution of the pavement.
Lytton et al. [29] developed finite element model and noted that the
Using finite element software, the three-dimensional numerical
longitudinal strain at the base of the asphalt layer rises with increased
analysis model of the asphalt pavement, as shown in Fig. 1(a), was
axle loads and temperatures, while it decreases with higher loading
established. In this numerical model, the X-axis, Y-axis, and Z-axis

3
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 5. Transverse dynamic strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

symbolize the lateral, depth, and lengthwise orientations of the road, for the model are presented in Table 1 and Table 2. The wind speed,
respectively. solar radiation intensity and pavement surface temperature are taken
The pavement structure is examined using the structure 9 of the from reference [39], the pavement temperature, emissivity, solar radi­
RIOHTrack established in China. The road structure form is illustrated in ation intensity, and convective exchange coefficient are set in temper­
Fig. 1(b). The road features two lanes, with each lane measuring 3.5 m in ature, radiation, distributed heat flux, and convection loads,
width, resulting in a total road surface width of 7.5 m. The width along respectively. The pavement material is assumed to be the porous me­
the driving direction is 20 m. The thickness of the structure in the road’s dium material. The materials of the other structural layers are assumed
depth direction matches the dimensions of the actual road structure. The to be continuous media and simulated using the linear elastic constitu­
model grid division is consistent with the division method of the liter­ tive model.
ature [41]. The following assumptions are applied to the model: the
materials of each structural layer are considered isotropic. Binding
2.3. Vehicle load simulation
constraints are applied between the foundation and the subgrade, while
continuous contact is used between the other structural layers of the
This study used RIOHTrack’s vehicle load for simulation. The
pavement, with the surface layer treated as a porous medium material.
loading test vehicle was a steyr heavy truck [49] as shown in Fig. 2(a).
The author provides a detailed explanation of the model’s boundary
The vehicle’s travel speed was 45 km.h− 1. In this study, in order to
conditions in the literature [39]. Based on Biot theory [42,43], ther­
better reflect the waveform of the vehicle load, the vertical dynamic
modynamics theory [30,44] and hydro-thermo-mechanical coupling
stress curve measured by the soil pressure gauge at the bottom of the
theory of saturated porous media [45], the dynamic response of asphalt
surface layer was used as the reference curve to further obtain the tire
pavement under multifield coupling was solved.
pressure curve waveform of the tire to the road surface. A large number
of pressure gauges are arranged in the RIOHTrack full-scale road test
2.2. Calculation parameters loop structure. Among them, for structure 9, the sensor closest to the
road surface is buried at 0.18 m. According to existing literature [50],
The material parameters of the model encompass temperature, force, for vehicle loads, when the distance between the two wheels is close, the
and humidity field material properties. The modulus, damping ratio, rear wheel pressure of the dual rear axle is generally larger because the
poisson’s ratio, and density for the model are sourced from literature dynamic response cannot be dissipated in time. Based on this, this study
values, and the temperature field material parameters are also based on uses the contact area and tire pressure corresponding to wheel 3 as the
referenced literature [3,39,46]. The permeability coefficients are standard, maps the vertical dynamic stress of the bottom surface of the
derived from values in the literature [47,48]. The material parameters bottom layer to the road surface, and obtains the tire pressure curve. The

4
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 6. Vertical dynamic strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

axle load distribution of the vehicle load is shown in Fig. 2(b), the represent the first, second, and the third principal strains, respectively.
literature [51] measured the tire contact area corresponding to wheel 3 To verify the correctness of the numerical model, the author constructed
on site to be 416 cm2 and the tire pressure to be 0.82 MPa. Therefore, the three-dimensional road structure model that is completely consistent
the peak value of the tire curve corresponding to wheel 3 is calibrated to with reference [52], with the same material parameters, boundary
0.82 MPa, and the ground pressures of wheel 1 and wheel 2 are conditions and loads. Then, the longitudinal dynamic strain was used as
0.45 MPa and 0.79 MPa respectively. The tire action curve is shown in a comparison index for comparative analysis. At the pavement depth of
Fig. 3. To simplify the modeling, it is assumed that the ground contact 0.04 m, the error between the two is within 2 %, indicating that the
area of wheel 1, wheel 2 and wheel 3 is 400 cm², and the ground contact model constructed in this paper is correct and can better predict the
length of the wheel in the longitudinal and transverse directions is dynamic strain distribution law of asphalt pavement. The peak values in
20 cm. In the finite element software, the vehicle load is simplified to a the following text are explained as follows, the positive peak value refers
moving load, and the moving load is simulated using the pressure to the corresponding maximum peak value among all peak values, and
function. The amplitude of the load is controlled by the time function. the negative peak absolute value refers to the corresponding maximum
peak value among all negative peak values.
3. Results and discussion
3.1. Time history distribution of normal strains
To examine how the strain in the pavement structure changes over
time and space when heavy-loaded trucks operate under various phys­
In various physical fields, along the pavement depth direction, the εxx
ical conditions, the vertical path shown in Fig. 4 was selected, where the
has the similar transmission laws in Fig. 5. With the enhancement of the
load investigation was conducted at Z= -8.6 m, with the extraction
pavement structure’s depth, the εxx tends to decrease, and the wave­
points for the vertical path (L1) data located at the road surface (Y=0 m),
forms of εxx under different physical fields are similar. At Y= 0 m, the
at the upper layer’s bottom (Y=0.04 m), at the middle layer’s bottom
absolute values of the negative peaks of the εxx under the thermo-
(Y=0.1 m), and at the lower layer’s bottom surface (Y=0.18 m). The
mechanical coupling field, the hydro-mechanical coupling field, and
time-history distribution characteristics of strain components and prin­
the hydro-thermo-mechanical coupling field are 34.7με, 33.2με, and
cipal strains were studied. To examine the spatial arrangement laws of
29.6με, respectively. From the above data, we can see that the εxx is
dynamic strain, the data of lateral and lengthwise data points in the xz
greatest under the thermo-mechanical coupling, and the εxx is minimal
plane were extracted at Y= 0 m, Y= 0.04 m, Y= 0.1 m, and Y= 0.18 m.
under thermo-hydro-mechanical coupling. Among them, the negative
εxx, εyy, and εzz represent horizontal, vertical, and longitudinal dynamic
peak value of the εxx under the thermo-mechanical coupling is 1.2 times
strains, respectively. εxy, εxz, and εyz are shear strains. ε1, ε2, and ε3
that under the hydro-thermo-mechanical coupling, suggesting that

5
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 7. Longitudinal dynamic strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

temperature has a considerable effect on the εxx at the road surface. positive peak value relating to wheel 2 is the highest, while the absolute
When assessing the strain in the pavement structure, the effect of tem­ value of the negative peak corresponding to wheel 3 is the greatest.
perature on the εxx should be taken into account. At 0.04 m, the εxx Within the range of Y= 0.1 m to 0.18 m, the absolute value of the
under hydro-mechanical field is greater than the peak value of the εxx negative peak of εyy under thermo-mechanical coupling is the largest.
under thermo-mechanical coupling. At Y= 0.1 m and below, the abso­ From the distribution pattern of εzz in Fig. 7, it is clear that at
lute values of the positive peak and negative peak of the εxx under the Y= 0 m, the absolute value of the negative peak of εzz exceeds that of the
action of hydro-thermo-mechanical coupling are the largest, and during positive peak. Specifically, the absolute value of the negative peak under
the process of load movement, the εxx has an alternating transformation thermo-mechanical coupling is the largest, while the smallest value
of positive strain and negative strain. As the depth of the road surface occurs under hydro-thermo-mechanical coupling. During the passage of
structure increases, the positive peak of the εxx increases, indicating that each wheel through the inspection section, the εzz experiences alter­
the tensile strain of the material increases, making it more likely to suffer nating tension-compression changes, which can readily result in cracks
fatigue damage under repeated action. in the pavement structure. Within the range of Y= 0.1 m to 0.18 m, the
From the variation laws of εyy in Fig. 6, it can be seen that at Y= 0 m, absolute value of the negative peak of εzz under hydro-mechanical
the absolute value of the negative peak of εyy under hydro-mechanical coupling is the greatest, and the wave patterns corresponding to
field is the highest, However, the positive peak value of εyy under wheels under different physical fields are relatively complex. At
hydro-thermo-mechanical coupling is the highest. As wheel 1 ap­ Y= 0.18 m, the absolute value of the negative peak of εzz for wheel 2 is
proaches, εyy first increases positively, and when the load arrives at the greater than that for wheel 3.
inspection section, the positive peak of εyy reaches its maximum. After As illustrated in Figs. 5, 6, and 7, the absolute values of the peak
the wheel passes this section, the maximum and minimum values of εyy dynamic strains (lateral, longitudinal, and vertical) diminish with
change instantaneously, transitioning from the positive peak to the increasing pavement depth across various physical fields. However, the
negative peak. As the wheel moves away, εyy gradually decreases. When values, waveform characteristics, and strain modes of dynamic strains in
wheels 2 and 3 pass the inspection section, εyy follows a similar trend to different directions vary. The peak absolute value of εyy is the highest,
that of wheel 1. At Y= 0 m, the negative peak values of the εyy under followed by εxx, with εzz being the smallest. Asphalt pavement materials
hydro-mechanical and hydro-thermo-mechanical coupling are − 47.1με in actual use are porous, thus, when calculating dynamic strain in high-
and − 38.0με, respectively. The εyy under hydro-mechanical increases by temperature and rainy regions, it is essential to treat the pavement
24 % compared with the εyy under hydro-thermo-mechanical coupling. material as a porous medium and analyze its mechanical properties. This
The maximum and minimum values of εyy occur at different times, the approach aligns more closely with the real stress state of asphalt

6
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 8. εxy: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

pavement. Additionally, when assessing dynamic strain in rainy areas, that under hydro-mechanical coupling, suggesting that the effects of
the impact of temperature on pore water pressure and dynamic strain temperature and humidity on shear strain should be considered when
should be considered, which is crucial for accurately predicting asphalt calculating shear strain.
pavement design in regions experiencing high temperatures and heavy
rainfall.
3.3. Time history distribution of principal strains

3.2. Time history distribution of shear strains The principal strain refers to the strain produced along the direction
of principal stress when the material is subjected to stress. Compared
It can be observed from the shear strains in various directions in with the strain in a certain direction, the principal strain can provide
Figs. 8, 9 and 10 that during the vehicle loading process, the shear strain more comprehensive strain information. Therefore, in order to consider
waveforms under different physical fields have similar distribution the deformation of the material in all directions, the change laws of the
characteristics, however, their numerical values are different. Among principal strain under different physical fields is analyzed, which helps
the three shear strains, the peak absolute value of εyz is the largest, to understand the overall deformation characteristics of the material
followed by εxy, and the peak absolute value of εxz is the smallest. From more accurately.
the shear strain εxy across various physical fields, at Y= 0 m, the peak Figs. 11, 12 and 13 are schematic diagrams of the change law of ε1, ε2
absolute values of εxy and εxz are the greatest under thermo-mechanical and ε3 respectively. As shown in Fig. 11, along the pavement depth di­
coupling, and the peak absolute value of εyz is the topmost under hydro- rection, under different physical fields, the absolute value of the nega­
thermo-mechanical coupling. There is an alternating phenomenon in εyz tive peak of the ε1 under thermo-mechanical coupling is the largest,
at different layers of the pavement, and the dominant factors of εyz are followed by hydro-thermo-mechanical coupling, and the smallest under
different at different depths. At Y= 0.04 m, the peak absolute value of hydro-mechanical coupling field. According to Biot theory, pore water
εyz under hydro-thermo-mechanical coupling is greater than the peak and asphalt mixture jointly bear the vehicle load, so ε1 under thermo-
absolute value of εyz under other physical fields. The negative peak mechanical coupling is greater than ε1 under hydro-mechanical
values of εyz under thermo-mechanical, hydro-mechanical and hydro- coupling. As the temperature rises, the modulus of the material de­
thermo-mechanical coupling are − 4.2με, − 3.5με and − 4.7με, respec­ creases, so it can also be seen that ε1 under thermo-mechanical coupling
tively. The εyz under hydro-thermo-mechanical coupling is 1.3 times is greater than ε1 under single stress. For positive peaks, At Y= 0 m,

7
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 9. εxz: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

0.04 m and 0.1 m, the positive peak of the ε1 under hydro-thermo- thermo-hydro-mechanical coupling is 2.4 times that under hydro-
mechanical field is the largest. At Y= 0.18 m, the positive peak under mechanical coupling, and the negative peak value of the ε1 under
the action of hydraulic coupling is the largest. The absolute value of the thermo-mechanical coupling is 2.5 times that under single stress field.
peak of the ε1 decreases as the pavement depth increases. In contrast, the From the above analysis, it can be seen that when the effect of tem­
distribution of the ε2 along the depth of the pavement is more intricate. perature on the ε1 is considered, the absolute value of the negative peak
As shown in Fig. 12, at Y= 0 m, the absolute value of the negative peak of the ε1 increases significantly, showing that temperature has a greater
of the ε2 under hydraulic coupling is the highest. At Y= 0.04 m, the influence on the ε1, and by comparing the ε1 under hydro-mechanical
largest absolute value of the negative peak of the ε2 occurs under a single coupling and hydro-thermo-mechanical coupling, it can be seen that
stress field. Meanwhile, at Y= 0.1 m and Y= 0.18 m, the greatest ab­ the negative peak value of the ε1 under hydro-thermo-mechanical
solute value of the negative peak of the ε2 is observed under thermo- coupling is much larger than the negative peak value of the ε1 under
mechanical coupling. The variation of the ε3 is also quite complex. As hydro-mechanical coupling. This indicates that when calculating asphalt
shown in Fig. 13, at Y= 0 m, 0.1 m and Y= 0.18 m, the negative peak pavement in hot and rainy areas, the coupling effects of temperature,
absolute value of the ε3 under the single stress field is the largest. humidity and vehicle load should be considered, otherwise the value of
However, at z = 0.04 m, the negative peak absolute value of the ε3 the ε1 will be underestimated.
under the hydro-mechanical coupling field is the largest. The existing asphalt pavement design specifications [3] use the
Based on the peak values of the ε1 in Fig. 14, it can be observed that tensile strain design index of the bottom surface of the asphalt layer
the numerical differences in the peak values of the ε1 under different along the vehicle direction and estimate the fatigue life. However, it can
physical fields are significant, and the dominant factors for negative and be seen from the actual service status of the asphalt pavement that the
positive peak values differ. At Y= 0 m, the positive peak values of the ε1 asphalt pavement is in a three-dimensional strain state under service,
under the single stress, thermo-mechanical, hydro-mechanical and and according to the calculation results, the tensile strain corresponding
thermo-hydro-mechanical coupling are 4.6με, 4.5με, 4.2με and 7.7με to the ε1 is much larger than the tensile strain along the vehicle direc­
respectively. The ε1 under thermo-hydro-mechanical coupling is 83 % tion. In other words, if the fatigue life is estimated based on the hori­
larger than that under hydro-mechanical coupling and 71 % larger than zontal strain in the driving direction, the service life of the asphalt
that under thermo-mechanical coupling. At Y= 0 m, the negative peak pavement will be overestimated. It is recommended to consider the
values of ε1 under single stress, thermo-mechanical, hydro-mechanical strain state of the asphalt pavement under the three-dimensional stress
and thermo-hydro-mechanical coupling are − 17.4με, − 43.4με, − 16.1με state, which is more in line with reality. When predicting the fatigue life
and − 38.0με respectively. The negative peak value of the ε1 under of asphalt pavement in rainy areas, the tensile strain corresponding to

8
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 10. εyz: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

the ε1 is used as the fatigue life prediction index. ε3, it will underestimate the maximum value of the ε3 while over­
Based on the peak values of the ε2 shown in Fig. 15, it is apparent that estimating the minimum value. A comparison of the data from the
the positive peak values across different physical fields are relatively thermal coupling field and the uniaxial stress coupling field reveals that
similar, while the negative peak values vary significantly. At Y= 0 m, when temperature effects are not considered, the positive peak value of
the negative peak value of the ε2 in the uniaxial stress field is 2.5 times the ε3 is underestimated, whereas the negative peak value is
greater than that under the thermo-mechanical coupling effect, which overestimated.
means that the damage of the ε2 to the pavement will be overestimated if
the temperature effect is not considered. The minimum value of the ε2 3.4. Spatial distribution of dynamic strains under hydro-thermo-
under the hydro-mechanical coupling field is 3.8 times that under the mechanical coupling field
hydro-thermo-mechanical coupling effect, which means that the ε2 will
decrease when the temperature, humidity and load coupling effects are To further examine the distribution of dynamic strain along the road
considered. As the pavement depth increases, the difference in the value surface during vehicle movement under the hydro-thermo-mechanical
of the negative peak value under different physical fields will decrease. field, data on dynamic strain components in both the lateral and
The above analysis shows that if the coupling of multiple factors is not lengthwise directions of the pavement were extracted for discourse. As
considered when calculating the ε2 of the road surface, large data errors shown in Fig. 17, the εxx is roughly symmetrically distributed in the
will be caused on the road surface. transverse direction of the pavement. As the depth of the pavement in­
From the peak value of the ε3 in Fig. 16, it can be seen that at creases, the εxx increases along the width in the transverse direction. In
Y= 0.1 m, the difference in the positive peak value of the ε3 in various the Y= 0 m plane, the εxx is predominantly negative, suggesting that the
physical fields is the largest, and at Y= 0.04 m, the ε3 positive peak road surface material is primarily experiencing compression. It can be
value under different physical fields has the largest difference. At clearly seen that the strain corresponding to wheel 1 is the smallest, and
Y= 0.1 m, the positive peak value of the ε3 under the hydro-thermo- the strain corresponding to wheel 3 is the largest. At Y= 0.04 m, it can
mechanical coupling field is 2.2 times that under the hydro- be found that the εxx peaks generated by wheels 2 and 3 overlap, and the
mechanical coupling. This shows that when calculating the ε3, if the width of the εxx along the transverse direction increases significantly. At
influence of temperature is not considered, the ε3 at the base of the Y= 0.1 m and Y= 0.18 m, at the center of the wheel, the εxx is positive,
middle surface layer will be underestimated. At Y= 0.04 m, the negative indicating that the road surface material is under tension. The strain
peak value of the ε3 under the hydro-mechanical coupling field is 1.7 state and magnitude of the material vary at different positions across the
times that under the hydro-thermo-mechanical coupling. Through road’s width, which can easily cause uneven deformation of the pave­
analysis, it can be concluded that if Biot’s theory is used to calculate the ment material.

9
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 11. First principal strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

10
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 12. Second principal strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

11
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 13. Third principal strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

Fig. 14. Peak value of first principal strain: (a) Positive peak values, (b)Negative peak values.

12
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 15. Peak value of second principal strain: (a) Positive peak values, (b)Negative peak values.

Fig. 16. Peak value of third principal strain: (a) Positive peak values, (b)Negative peak values.

From the distribution of εyy along the road space in Fig. 18, it can be differences of pavement structure dynamic strains in various physical
observed that in the load application area, the εyy is predominantly fields, and studies the distribution characteristics of dynamic strain
negative throughout the road’s width. Outside the load application area, components along the pavement space under the hydro-thermo-
positive peak values of εyy occur. The vertical dynamic strain experi­ mechanical coupling. Under this research condition, the following
ences a complex loading process of tension-compression-tension along main conclusions are drawn.
the width, with the increase in pavement depth, the εyy gradually in­
creases along the transverse width of the pavement. Along the length of (1) Under different physical fields, the absolute values of the peak
the road, the three negative peak values of εyy correspond to the three dynamic strains (lateral, longitudinal, vertical) decrease as the
wheels. pavement depth increases. However, the values and waveforms
From the distribution of εzz in both the horizontal and longitudinal of dynamic strains in different directions are different. The peak
directions of the road surface shown in Fig. 19, it is evident that as the absolute value of εyy is the largest, followed by εxx, and εzz is the
depth of the pavement increases, the influence width of εzz in both di­ smallest. Among the three shear strains, εyz has the largest peak
rections expands. At Y= 0.1 m and Y= 0.18 m, the peak values associ­ absolute value, followed by εxy, and εxz has the smallest peak
ated with the wheel are more intricate. When a wheel passes over, two absolute value.
peaks are generated along the length of the road, with the significant (2) The peak values of the ε1 differ significantly in various physical
tensile strain occurring between wheel 1 and wheel 2. Moreover, there is fields, and the dominant factors of the negative and positive peak
a clear superposition of peak values of εzz along the longitudinal direc­ values are different. At Y= 0 m, the negative peak value of the ε1
tion of the pavement for wheel 2 and wheel 3. under hydro-thermo-mechanical coupling is 2.4 times that under
hydro-mechanical coupling. The negative peak value of the ε1
4. Conclusions under thermo-mechanical coupling is 2.5 times greater than that
under the single stress field. This analysis indicates that when
By constructing the multifield coupling models, the paper qualita­ calculating the ε1 of the structure, it is crucial to consider the
tively and quantitatively analyzes the time-history distribution

13
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 17. Lateral dynamic strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

14
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 18. Vertical dynamic strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

15
P. Dong et al. Construction and Building Materials 469 (2025) 140448

Fig. 19. Longitudinal dynamic strain: (a) Y= 0 m, (b) Y= 0.04 m, (c) Y= 0.1 m, (d) Y= 0.18 m.

coupling effects of temperature, humidity. Otherwise, the calcu­ Software, Methodology. Xuejuan Cao: Supervision, Investigation,
lation results may have significant errors. Funding acquisition.
(3) Under the action of hydro-thermo-mechanical coupling, the
strain state and magnitude of the pavement material in space are
relatively complex along the transverse and longitudinal di­ Declaration of Competing Interest
rections of the pavement, with alternating tension and compres­
sion and different values, which can easily lead to uncoordinated The authors declare that they have no known competing financial
structural deformation at different spatial positions of the pave­ interests or personal relationships that could have appeared to influence
ment, causing cracks and rutting damage to the pavement. the work reported in this paper
(4) As the road surface depth increases, the dynamic strain peaks
generated by wheels 2 and 3 appear superimposed along the Acknowledgements
lateral and longitudinal directions of the road surface. By
analyzing the variation laws of dynamic strains under multi-field The authors would like to acknowledge the financial support from
coupling, it has a certain reference value for revealing the dam­ the National Nature Science Foundation of China (Grant No. 52278441).
age mechanism of asphalt pavement under multifield coupling
and the design of long-life asphalt pavement. Data availability
(5) This study made some assumptions about the asphalt pavement
when calculating the dynamic strain of the asphalt pavement Data will be made available on request.
under multifield coupling, which led to some differences with the
actual asphalt pavement in service. Further improvements are References
needed to improve the prediction ability of the model.
[1] I.R. Congress. Guidelines for the design of flexible pavements[J]. Indian code of
CRediT authorship contribution statement practice, IRC, 2001, 37.
[2] X. Fan, S. Lv, N. Zhang, et al., Characterization of asphalt mixture Moduli under
different stress states, Materials 12 (3) (2019) 397.
Boming Tang: Formal analysis, Software, Supervision. Pengjun [3] China Communications Road and Bridge Technology Co., Ltd, Specifications for
Dong: Writing – review & editing, Writing – original draft, Validation, design of highway asphalt pavement: JTG D50-2017 [S], China Communications
Press, Beijing, 2017.

16
P. Dong et al. Construction and Building Materials 469 (2025) 140448

[4] B.F. Kallas, Dynamic modulus of asphalt concrete in tension and tension- [29] R.L. Lytton, J. Uzan, E.G. Fernando, et al., Development and validation of
compression and discussion, Assoc. Asph. Paving Technol. Proc. (1970). performance prediction models and specifications for asphalt binders and paving
[5] H. Cheng, Y. Wang, L. Liu, et al., Estimating tensile and compressive moduli of mixes, Strategic Highway Research Program, Washington, DC, 1993.
asphalt mixture from indirect tensile and four-point bending tests, J. Mater. Civ. [30] O.C. Assogba, Y. Tan, X. Zhou, et al., Numerical investigation of the mechanical
Eng. 33 (1) (2021) 04020402. response of semi-rigid base asphalt pavement under traffic load and nonlinear
[6] P.P. Khanal, M.S. Mamlouk, Tensile versus compressive moduli of asphalt concrete, temperature gradient effect, Constr. Build. Mater. 235 (2020) 117406.
Transp. Res. Rec. 1492 (1995) 144–150. [31] O.C. Assogba, Y. Tan, Z. Sun, et al., Effect of vehicle speed and overload on
[7] M.R. Islam, S.A. Kalevela, G. Mendel, How the mix factors affect the dynamic dynamic response of semi-rigid base asphalt pavement, Road. Mater. Pavement
modulus of hot-mix asphalt, J. Compos. Sci. 3 (3) (2019) 72. Des. 22 (3) (2021) 572–602.
[8] H. Jin, X. Zhou, Z. Ju, et al., Study on dynamic modulus dependence model for [32] O.C. Assogba, Z. Sun, Y. Tan, et al., Finite-element simulation of instrumented
asphalt mixture based on temperature and strain parameters, J. Highw. Transp. asphalt pavement response under moving vehicular load, Int. J. Geomech. 20 (3)
Res. Dev. 14 (3) (2020) 70–77. (2020) 04020006.
[9] Q. Li, A. Shen, Y. Guo, et al., Review of evaluating asphalt pavement structure [33] X. Liu, X. Zhang, Asphalt pavement dynamic response under different vehicular
integrity and strength with Rayleigh wave Methods: techniques, applications, and speeds and pavement roughness, Road. Mater. Pavement Des. 22 (6) (2021)
trends, Measurement (2024) 114849. 1287–1308.
[10] S. Xiong, S. Zhao, D. Wang, et al., Study on fatigue parameters of high modulus [34] B. Saad, H. Mitri, H. Poorooshasb, Three-dimensional dynamic analysis of flexible
asphalt pavement, Case Stud. Constr. Mater. (2024) e03491. conventional pavement foundation, J. Transp. Eng. 131 (6) (2005) 460–469.
[11] G. Zheng, N. Zhang, P. Wang, et al., Study on the characterization method of [35] G. Ren, A. Shen, H. Wu, et al., Mechanical response of flexible asphalt pavement
dynamic modulus for asphalt mixture under multi-factor coupling conditions, under large temperature difference and temperatures during four seasons, Constr.
Constr. Build. Mater. 421 (2024) 135758. Build. Mater. 437 (2024) 136967.
[12] W. Zeiada, G. Al-Khateeb, I. Fattouh, et al., Impact of confinement condition of [36] L. Li, Z. Zhang, Z. Wang, et al., Coupled thermo-hydro-mechanical response of
dynamic modulus test on the performance of flexible pavement structures, Innov. saturated asphalt pavement, Constr. Build. Mater. 283 (2021) 122771.
Infrastruct. Solut. 9 (8) (2024) 1–16. [37] Z. Liu, X. Yu, Coupled thermo-hydraulic modelling of pavement under frost, Int. J.
[13] J. Zhang, J. Cui, X. Wang, A testing theory for simultaneously determining asphalt Pavement Eng. 15 (5) (2014) 427–437.
mixture’s dynamic modulus in compression and tension by use of IDT test, Constr. [38] Y. Sun, R. Guo, L. Gao, et al., Study on dynamic response characteristics of
Build. Mater. 440 (2024) 137334. saturated asphalt pavement under multi-field coupling, Materials 12 (6) (2019)
[14] S. Lv, N. Zhang, Z. Wu, et al., Normalization of dynamic modulus of asphalt 959.
mixture under three-dimensional stress state, Chin. J. Highw. 36 (4) (2023) 27–37. [39] P. Dong, Y. Yuan, X. Cao, Numerical investigation of dynamic stresses in asphalt
[15] H. Cheng, L. Sun, J. Zheng, et al., Dynamic compressive-tensile moduli of asphalt pavement under the combined action of temperature, moisture and traffic loading,
mixture and its applications to pavement response prediction, China Civ. Eng. J. 55 Constr. Build. Mater. 417 (2024) 135131.
(03) (2022) 105–116. [40] P. Dong, X. Cao, B. Tang, Analysis of the mechanical behavior of asphalt pavement
[16] T. Jin, J. Yuan, X. Peng, et al., A fatigue damage model of asphalt mixture under multiple coupling effects, Constr. Build. Mater. 449 (2024) 138409.
considering tensile and compressive modulus decay, Case Stud. Constr. Mater. 20 [41] X.H. Chen, H. Li, Research on three-dimensional finite element model of asphalt
(2024) e03133. pavement structure, J. Wuhan. Univ. Technol. 44 (3) (2020) 584–590.
[17] H. Wang, M. Li, P. Szary, et al., Structural assessment of asphalt pavement [42] M.A. Biot, Theory of propagation of elastic waves in a fluid-saturated porous solid.
condition using backcalculated modulus and field data, Constr. Build. Mater. 211 I. Low-frequency range, Repr. J. Acoust. Soc. Am. 28 (2) (1956) 179–191.
(2019) 943–951. [43] M.A. Biot, Mechanics of deformation and acoustic propagation in porous media,
[18] G.Y. Yang, J.S. Zhang, Q. Wang, et al., Field experimental study on the influence of J. Appl. Phys. 33 (4) (1962) 1483–1498.
vehicle speed and load on road structure, Geotech. Mech. 29 (11) (2008) 5. [44] X. Wang, Y. Zhong, Reflective crack in semi-rigid base asphalt pavement under
[19] X. Ma, Z. Dong, X. Yu, et al., Monitoring the structural capacity of airfield temperature-traffic coupled dynamics using XFEM, Constr. Build. Mater. 214
pavement with built-in sensors and modulus back-calculation algorithm, Constr. (2019) 280–289.
Build. Mater. 175 (JUN.30) (2018) 552–561. [45] D.W. Smith, J.R. Booker, Boundary element analysis of linear thermoelastic
[20] M. Cao, W. Huang, Z. Wu, Influence of axle load and asphalt layer thickness on consolidation, Int. J. Numer. Anal. Methods Geomech. 20 (7) (1996) 457–488.
dynamic response of asphalt pavement, Geofluids 2022 (2022). [46] A.Q. Shen, X.K. Jin, Y.C. Guo, Mechanical response analysis of semi-rigid asphalt
[21] W. Wang, K. Zhao, J. Li, et al., Characterization of dynamic response of asphalt pavement in Northern Shaanxi area under coupled fields, J. Chang’Univ. 42 (05)
pavement in dry and saturated conditions using the full-scale accelerated loading (2022) 1–11.
test, Constr. Build. Mater. 312 (2021) 125355 (1-4). [47] J. Chen, F.P. Li, J.A. Shen, et al., Study on permeability coefficient of asphalt
[22] X. Chen, J. Zhang, X. Wang, Full-scale field testing on a highway composite mixture, J. Highw. Transp. Res. Dev. 1 (1) (2006) 27–30.
pavement dynamic responses, Transp. Geotech. 4 (2015) 13–27. [48] B. Wang, Y. Zhang, X. Zhu, et al., Experiment investigation and influence
[23] Z. Dong, Q. Xu, P. Lu, Dynamic response of semi-rigid base asphalt pavement based evaluation of permeability ability attenuation for porous asphalt concrete under
on accelerated pavement test, China J. Highw. Transp. 24 (2) (2011) 1–6. repeated clogging conditions, Buildings 13 (11) (2023) 2759.
[24] J. Li, Y. Li, C. Xin, et al., Dynamic strain response of hot-recycled asphalt pavement [49] L. Zhang, X. Zhou, X. Wang, Research progress of long-life asphalt pavement
under dual-axle accelerated loading conditions, Coatings 12 (6) (2022) 843. behavior based on the RIOHTrack full-scale accelerated loading test, Chin. Sci.
[25] A. Bayat, M.A. Knight, H.R. Soleymani, Field monitoring and comparison of Bull. 65 (30) (2020) 3247–3258.
thermal-and load-induced strains in asphalt pavement, Int. J. Pavement Eng. 13 (6) [50] X. Cui, J. Zhang, D. Huang, et al., Analysis of vehicle-force-induced dynamic pore
(2012) 508–514. pressure in saturated pavement with LSPM drainage base, J. Test. Eval. 45 (1)
[26] Q. Pan, C. Zheng, S. Lv, et al., Field measurement of strain response for typical (2017) 294–302.
asphalt pavement, J. Cent. South Univ. 28 (2) (2021) 618–632. [51] X. Wang, X. Zhou, L. Zhang, et al., Full-scale Test Track(RIOHTrack) Construction
[27] X. Wang, X. Zhou, W. Guan, et al., Characteristics and analysis of the mechanical Record, People’s Communications Press, Beijing, 2023.
response inside the structure of asphalt pavement, Chin. Sci. Bull. 65 (2020) [52] Z. Dong, L. Cao, Y. Tan, et al., Analysis of the dynamic response of three directional
3298–3307. strains in asphalt pavement under moving vehicle loads, China Civ. Eng. J. 42 (4)
[28] X.Y. Zhou, X.D. Wang, W. Guan, et al., Analysis on response behavior of asphalt (2009) 133–139.
pavement structure in wide temperature range, Harbin Inst. Technol. 52 (2020)
63–69.

17