KSCE Journal of Civil Engineering (2022) 26(12):5146-5159 pISSN 1226-7988, eISSN 1976-3808

DOI 10.1007/s12205-022-1768-8 www.springer.com/12205

Structural Engineering

CFD Numerical Simulation of Wind Field and Vehicle Aerodynamic

Characteristics on Truss Bridge Deck under Crosswind

Mingjin Zhang a,b,c, Jinxiang Zhang a

, Junting Long d
, Yongle Li a,b
, Yulin Zou e
,
and Dianguo Yin f
a
Dept. of Bridge Engineering, Southwest Jiaotong University, Chengdu 610031, China
b
Wind Engineering Key Laboratory of Sichuan Province, Southwest Jiaotong University, Chengdu 610031, China
c
T.Y. Lin International Engineering Consulting (Chongqing) Co., Ltd., Chongqing 400031, China
d
China Railway Eryuan Engineering Group Co., Ltd., Chengdu 610000, China
e
Sichuan Yanjiang Panning Expressway Co., Ltd., Xichang 615000, China
f
General Research Institute of Architecture & Planning Design Co., Ltd., Chongqing University, Chongqing 400030, China

ARTICLE HISTORY ABSTRACT

Received 28 September 2021 Due to the complexity and changeability of the wind field in deep-cut gorges, the vehicles on
Revised 18 February 2022 the bridge deck are easily affected by a strong crosswind. Thus, to accurately evaluate the
Accepted 22 July 2022 wind field characteristics of a suspension bridge deck, the wind profiles of wind speed and
Published Online 7 October 2022 angle of attack (AoA) and the vehicle aerodynamic parameters were investigated by the CFD.
The results show that the shape of wind speed profile is mainly controlled by the AoA but less
KEYWORDS affected by the Reynolds number. The main girder's shielding effect can accelerate the local
wind field, and the closer to the windward lane, the less the interference; thus, a suitable
CFD Numerical simulation location of measuring points to represent the incoming flow is found. Furthermore, the
Wind field characteristics equivalent wind speed based on the equivalent side force is generally larger than the value
Vehicle based on the rolling moment, and the responding value is greatly affected by the AoA. In
Bridge deck addition, the vortex is a time-dependent phenomenon, the averaged flow field produced less
Equivalent wind speed
force in the wake, but the high local wind speed variations may affect the traffic unfavorably.
Aerodynamic coefficient
The results provide an essential reference significance for studying the local wind field
characteristics of the bridge deck.

1. Introduction deck focus on the local wind field on the bridge deck, that is, the
wind environment on the bridge deck near the bridge tower and
Long-span bridges located in mountainous terrain need to cross pier (Salati et al., 2018; Ni et al., 2019). Yuan et al. (2018) conducted
complex terrain. The terrain fluctuation and climate variability numerical simulation of the wind field above the highway bridge
lead to an extremely complex wind environment in mountainous deck around the bridge tower based on CFD method, and
areas. Therefore, to evaluate the driving stability of vehicles measured the wind speeds at the center height of three typical
passing through mountain bridges in windy weather, it is necessary road vehicles and trains via the wind tunnel test of large local
to analyze the wind field characteristics of their bridge decks rigid tower beam model. The results show that there is a certain
(Wang et al., 2021). There have been many outstanding contributions wind speed acceleration effect in the flow field near the bridge
to the wind environment of bridges, among which there are three tower area, and the wind speed on the windward side is greater
main research and analysis methods: numerical simulation using than that on the downwind side, but the wind speed gradient on
computational fluid dynamics (CFD) (Zhang et al., 2019; Guo et the downwind side is larger. At the same time, Wang et al. (2020)
al., 2021), wind tunnel test (Chen et al., 2017; He et al., 2020), also used experimental and CFD methods to study the wind
and field measurement (Peng et al., 2018; Zhang et al., 2021). environment near the inverted Y-shaped bridge tower. Through
Most of the researches on the wind environment on the bridge the visualization and numerical simulation of the flow field in the

CORRESPONDENCE Jinxiang Zhang [email protected] Dept. of Bridge Engineering, Southwest Jiaotong University, Chengdu 610031, China
ⓒ 2022 Korean Society of Civil Engineers
 KSCE Journal of Civil Engineering 5147

wind tunnel test, the flow direction and wind speed within the different lanes.
vehicle height range near the bridge tower are observed and Besides, many studies have focused on the field measurement
analyzed. The analysis of the wind environment on the bridge of the wind characteristics at the bridge site (Zhang et al., 2022).
deck is mainly aimed at studying the distribution of the wind Kim et al. (2016) combined the wind tunnel test and long-term
field above the bridge deck, so as to better evaluate the driving measured wind data at the bridge site to analyze the influence of
performance of vehicles in windy weather. Therefore, many vehicle type, lane position, and vehicle speed on vehicles' driving
CFD studies on the wind field on the bridge deck directly focus performance under crosswind. Wang et al. (2015) measured the
on the analysis of the aerodynamic coefficient and driving wind field characteristics above the bridge deck of a cable-stayed
performance of vehicles (Charuvisit et al., 2004). Salati et al. bridge through four measuring points (MPs) arranged at two
(2018) used CFD method with dynamic grid technology to towers. In the wind resistance design stage of a long-span bridge,
simulate the relative speed between vehicle and bridge, and it is necessary to obtain wind field data at the bridge site.
studied the aerodynamic force received by the vehicle when it However, when the monitoring system is arranged on the bridge
passed through the tower wake area generated by crosswind deck, the selection of monitoring points and instrument height
during the overtaking process of a heavy truck. Ma et al. (2018) has a great impact on the accurate acquisition of the local wind
used the large eddy simulation method to study the transient field on the bridge deck; thus, it is necessary to explore the
aerodynamic interference mechanism of vehicles in the process measured points. In order to further understand the local wind
of overtaking in the same direction and in the opposite direction field on the bridge deck and provide guidance for the actual
in the crosswind environment. Considering that different driving measurement of the bridge deck, this study investigated the wind
lanes may be affected differently by the wind field, the current field characteristics (the profiles of wind speed, angle of attack
research mainly focuses on the wind field characteristics in (AoA)), vehicle aerodynamic parameters, and the location of
specific areas of the bridge deck, and there is a lack of research measuring points (MPs) on a suspension bridge deck by CFD
on the wind field characteristics of different driving lanes on the and wind tunnel tests. In Section 2, the methods used in this
bridge deck. study, including the setting of CFD and wind tunnel test, were
As a necessary tool to obtain wind field parameters directly, introduced. Section 3 mainly discusses the influence of different
wind tunnel test is widely used to study driving safety, comfort factors on the local wind field characteristics and the installation
and local wind field characteristics of bridge deck. Kramer et al. position of measuring points. Section 4 is the main conclusion of
(1991) analyzed various vehicle parameters in crosswind and the this paper.
impact of driver operation on vehicle driving performance based
on a wind tunnel test with a scale of 1:10. In addition, Charuvisit 2. Background and Study Methods
et al. (2004) and Cheli et al. (2011) used the wind tunnel test
methods to determine different vehicle types and vehicle shapes' 2.1 Description of the Bridge Site
influence on vehicle aerodynamic coefficients. Based on this, Li As shown in Fig. 1(a), the Xingkang Bridge on Luding Dadu
et al. (2013) studied the aerodynamic coefficients of trains passing River, located in the mountainous region of southwestern China,
through pylons and double-vehicle intersections and discussed is a single-span suspension bridge with a main span of 1,100 m,
the effects of different factors, including wind speed, train speed, and spans a typical high-altitude deep U-shaped canyon of more
train track position, and train type. Kim et al. (2016) also analyzed than 1,500 m. The anchor spans are 220 m and 253 m on the
the influence of vehicle type, lane position and vehicle speed on West and East sides, respectively, and the deck of the bridge is
vehicle performance under cross wind by combining wind tunnel 238 m from the bottom of the canyon. Fig. 1(b) shows the overview
segment model test and long-term measured wind data at the of the bridge with a detail of the bridge deck, and Fig. 1(c) shows
bridge site. Meanwhile, Wang and Xu (2015) studied the impact of the local topography at the bridge site. It can be seen from the
bridge tower shielding effect on vehicle aerodynamic characteristics figure that the average altitude of the ridge of the East side is
through wind tunnel tests. The results show that the bridge tower 2 km. In addition, the elevation of the mountain peaks on both
has a significant effect on the vehicle yaw moment coefficient sides of the river valley in the bridge site is more than 5,000 m,
and rollover moment coefficient. At the same time, it is concluded and the elevation difference between the mountain peak and the
that when the time-varying aerodynamic coefficient caused by the bottom of the river valley is more than 3,500 m. The climate in
bridge tower is ignored, the rollover risk of vehicles passing the gorge is a typically dry, hot valley. The temperature difference
through the bridge tower is underestimated. Han et al. (2013, between the bottom and the top of the gorge and between day
2014) also discussed the influence of the turbulence intensity of and night at the bridge site is significant.
the incoming flow, the mutual interference between vehicles and
the position of vehicles on the bridge on the vehicle aerodynamic 2.2 Study Methods
coefficient by using the wind tunnel test method. The test results
show that all the factors studied have a significant impact on the 2.2.1 CFD 3D Model
aerodynamic coefficient of the vehicle. Therefore, it is very In order to catch the data of the wind speed and angle of attack
important to study the aerodynamic coefficients of vehicles in (AoA) above the bridge deck, the numerical simulation software
5148 M. Zhang et al.

Fig. 1. The Diagram of Xingkang Bridge on Luding Dadu River Including: (a) The Elevation of the Bridge, (b) View of the Bridge with a Detail of the Bridge
Deck, (c) The Local Topography at the Bridge Site

FLUENT (Canonsburg, Pennsylcommercial vania, U.S.) was the car body into chamfers to reduce the number of facets. The
used for the CFD. A 3D model of the main girder with 70 m × 28 m vehicle was placed in the middle of the main girder section in the
× 11.6 m was built for further analysis. Furthermore, considering calculation. Fig. 2(a) shows the simplified model of the vehicle
the large number of meshes generated by the 3D model, some (commercial van) and the main girder for the sake of assessing
model components were simplified to save computer resources the hazards when vehicles cross the bridge deck in the crosswind.
and improve the quality of mesh generation, as shown in Fig. 2(a). The size of the calculation domain directly affects the accuracy
The principles of the simplification are as follows: 1) The cross- and efficiency of the results, and many scholars have done much
section of the I-shaped members is simplified as a rectangular research. It is confirmed that the size of the computational
section with the equal windward area; 2) The railing of the domain is determined by the purpose of the experiment, the size
maintenance track is simplified as a single rectangular section of the model, and the computational resources. It can be seen
based on the same windward area; 3) The anti-collision railing of from Fig. 2(b) that the size of the calculation domain is set to 810 m
the bridge deck is simplified from a grid shape to a rectangle with × 70 m × 140 m. The distance between the main girder and the
the equal ventilation rate; 4) The influence of some unnecessary velocity entrance and upper boundary is 250 m and 66 m,
components such as the periphery of the bridge deck on the wind respectively, and the main girder intersects with the left and right
field characteristics of the bridge deck is ignored. At the same boundary. As plotted in Fig. 2(c), the three-dimensional numerical
time, the 3D vehicle model was also simplified in calculating the model is meshed by ICEM-CFD software. Due to the complex
vehicle aerodynamic coefficient by changing the arc angles of structure of truss members, unstructured tetrahedral and prismatic
 KSCE Journal of Civil Engineering 5149

Fig. 2. CFD Simulation Settings: (a) 3D Model with the Vehicle, (b) Global Grid, (c) Boundary Conditions of the Computational Domain, (d) Local
Grid, (e) Local Grid Detail

meshes are used near the main beam, while hexahedral structured symmetry of the parts outside the two sides for the truss main
meshes are used far away from the main beam, as shown in beam section, the boundary conditions on the left and right sides
Fig. 2(d). The advantage of using this hybrid grid division form are considered the symmetrical boundary. It can ensure that the
is that the tetrahedral grid can adapt well to the complex geometry of flow field on both sides of the segment is more accurate and
the truss section, and the peripheral structured grid can effectively avoid solving the redundant calculation domain outside the two
control the number of grids and improve computational efficiency. sides to save calculation resources. The boundary condition at
Fig. 2(e) shows the local grid detail, where the thickness of the the horizontal outlet is always taken as the pressure outlet boundary
adherent layer grid is taken as 0.016 m, and its growth rate is condition. The calculation basin's bridges, railings, and vehicles
taken as 1.1. The number of grids under each working condition take the solid wall boundary. At the same time, to better predict
is 20 − 22 million. 90% of the overall grid quality is above 0.3. the flow around the wall and reflect the wind field characteristics
By checking various indicators, the grid quality can meet the of the bridge deck, the turbulence calculation model is set as SST
conditions of steady calculation. k-ω (Launder and Spalding, 1974). For saving computing resources,
the model adopts steady calculation with the SIMPLEC algorithm,
2.2.2 Calculation Settings and the solving equations adopt second-order discretization.
Figure 2(c) shows the boundary conditions of the computational In order to investigate the wind field characteristics and drive
domain. As can be seen from the figure that the velocity inlet the safety of bridge deck driving, the 3D model establishes the
boundary condition is adopted at the inlet side of the horizontal bare main beam model and the main beam model with three
flow field. In order to coordinate with wind tunnel test, the different vehicle models located in different lanes. The bare main
incoming flow is uniform with the turbulence intensity of 0.5% girder model is mainly used to analyze the bridge deck driving
and the viscosity of 2. The upper and lower boundaries adopt the wind field at different angles of attack, and the girder model,
solid wall boundary to simulate the wall. As a result of the including vehicles, is used to analyze the aerodynamic coefficient
5150 M. Zhang et al.

Table 1. Size Parameters of the Articulated Car, the Traveling Car, and
the Commercial Van
Projection area (m2) Height of the
Types of Length
chassis from
vehicles Width-As Vertical-Af (m)
the ground (m)
Articulated car 44.045 8.319 15.410 0.808
Commercial van 29.868 7.567 10.420 0.495
Travelling car 19.894 6.722 7.670 0.620

the bridge deck. Considering the bridge site is in a complex wind

field, the angle between the incoming flow direction and the
bridge deck is large. Therefore, the AoA of wind is taken as -7°,
-5°, -3°, 0°, +3°, +5°, and +7° with a wind speed of 10 m/s, in
order to investigate the influence of the AoA on the wind field
Fig. 3. Locations of Wind Speed Measurement Points in the CFD characteristics of the bridge deck. The calculation case is shown
(units: cm)
in Table 2, where the bold black part (wind speed of 10 m/s with
an angle of attack of 0°) represents the incoming wind speed and
the corresponding angle of attack setting when there is a vehicle.
It is worth noting that the monitoring points of the test are
arranged in the center of the model along the bridge to monitor
the wind speed distribution above the four lanes. Each monitoring
point is 0.25 m, 0.5 m, 0.75 m, 1 m, 1.5 m, 2 m, 2.5 m, 3 m, 3.5 m,
4 m, 5 m, 7 m, and 10 m away from the bridge deck. There are
52 monitoring points in total, and the wind profile of the lane is
drawn through the monitoring points. The measurement points
(MPs) in vertical is plotted in Fig. 3.

2.2.3 Numerical Model Validation

In order to validate the accuracy of the CFD results, large-scale
wind tunnel tests were carried out in the XNJD-3 wind tunnel at
Southwest Jiaotong University. The test section of the wind
tunnel is 22.5 m width, 4.5 m height and 36 m length, which is
the world's largest boundary layer wind tunnel (Ren et al., 2019).
Fig. 4. The Size of Different Types of Vehicles, Including: (a) Articulated According to the professional testing, the maximum wind speed
Car, (b) Traveling Car, (c) Commercial Van
can reach 16.5 m/s, and the minimum wind speed can be
controlled at 1.0 m/s. At the same time, the turbulence intensity
when vehicles are in different lanes. Three types of vehicle of the incoming flow was less than 1.5%, and the thickness of the
models (Zhang et al., 2020a), including the articulated car, the boundary in the tunnel was less than 0.2 m. The scale ratio of the
traveling car, and the commercial van, were investigated, as bridge section model was set to be the same as that of the CFD.
described in Fig. 4. Specifically, the heights of the three vehicles Fig. 5 shows the model placed in the wind tunnel, and the size of
are 3.78 m (articulated car), 3.35 m (traveling car), and 3.3 m the model is 3,460 mm (long) ×406.9 mm (high) × 1323.5 mm
(commercial van), respectively. The specific dimensions are (wide). It is observed from the figure that the Cobra probe with a
illustrated in Table 1. speed accuracy of 0.3 m/s was used to measure the wind speed
In addition, the inlet wind speed is set to be 5 m/s, 8 m/s, and above the bridge deck of the truss section at a sampling rate of
10 m/s to investigate the influence of the Reynolds number (The 1,000 Hz (Hargreaves and Wright, 2007; Dong et al., 2021).
corresponding Reynolds numbers are 9.3800 × 106, 1.5008 × Therefore, wind tunnel test results are used to verify the wind
107, and 1.8760 × 107, respectively) on the local wind field above speed profile of the CFD when the wind tunnel test is consistent

Table 2. Arrangement of the Calculation Case

Case no 1 2 3 4 5 6 7 8 9 10 11 12 13
Wind speed (m/s) 5 8 10 10 10 10 10 10 10 10 10 10 10
AoA (°) 0 0 -7 -5 -3 -2 -1 0 1 2 3 5 7
 KSCE Journal of Civil Engineering 5151

addition, the main reason for the difference in the low altitude
area lies in the certain simplification of the components by the
CFD, such as simplifying the railing to a rectangular length and
using a rectangular section instead of an I-section. Therefore, in
order to quantify the difference of results between the CFD and
wind tunnel tests, Pearson correlation coefficient (as expressed
by Eq. (2)) was introduced to express the correlation of the wind
profile between each lane and to explore the difference of each
lane affected by the bridge deck facilities. From Table 3, the
Fig. 5. Scale Model of the Bridge in the Wind Tunnel Test
windward lane has the most significant influence on the adjacent
lane, and the influence effect is related to the distance. This
with the incoming flow conditions of CFD (wind speed 10 m/s phenomenon is reflected in the CFD and wind tunnel tests
with an angle of attack of 0°). Considering the difference between simultaneously. In addition, the correlation coefficient in the
the CFD calculation model and the wind tunnel test model, the CFD is more significant than that in the wind tunnel test, which
normalized ratio k to gradient wind speed is used to compare the is not only the reason for the simplification of components but
wind speed profiles (Zhang et al., 2020b), and the ratio k can be also the influence of the turbulence model used in CFD simulation
calculated by Eq. (1). Fig. 6 compares the conversion wind speed (Yang et al., 2009; Rezaeiha et al., 2019).
profile between the CFD and wind tunnel test. It can be seen
U
from the figure that the CFD results are slightly different from k= (1)
wind tunnel test results within the range of fewer than 5 m from U0
the bridge deck. When the height is more than 5 m, the two here k represents the ratio to gradient wind speed; U is the wind
results tend to be consistent gradually with height increase. In speed of the measuring point, and U0 is the wind speed of the
general, the CFD results are highly consistent with the wind tunnel incoming flow, where the U0 is taken as 10 m/s in the wind
calculation; thus, the CFD results are reasonable and reliable. In tunnel test and the CFD.

Fig. 6. Comparison of the Conversion Wind Speed Profile between the CFD and Wind Tunnel Test: (a) In Lane 1, (b) In Lane 2, (c) In Lane 3, (d) In
Lane 4

Table 3. Correlation Coefficient between Lanes

Wind tunnel test CFD
Lane no
1 2 3 4 1 2 3 4
1 1 0.9687 0.9148 0.8518 1 0.9914 0.9667 0.9490
2 0.9687 1 0.9718 0.9328 0.9914 1 0.9809 0.9757
3 0.9148 0.9718 1 0.9880 0.9667 0.9809 1 0.9943
4 0.8518 0.9328 0.9880 1 0.9490 0.9757 0.9943 1
5152 M. Zhang et al.

Fig. 7. Wind Speed Profiles on the Bridge Deck at Different Incoming Wind Speeds: (a) In Lane 1, (b) In Lane 2, (c) In Lane 3, (d) In Lane 4

n
the wind speed of the incoming flow in the CFD. The incoming
E[( X − μ X )(Y − μY )] ∑ ( X i − X )(Yi − Y )
i =1
wind speeds are set as 5 m/s, 8 m/s, and 10 m/ s, respectively, as
ρ ( X ,Y ) = = plotted in Fig. 7, which shows the comparison of the wind profile
σ XσY n n
∑
i=
1
( X i − X ) ∑ (Yi − Y )
2

i= 1
2 of the bridge deck under different incoming wind speeds. It can
be concluded that the normalized wind speed profile does not
change with the change of wind speed. Therefore, the Reynolds
(2)
number only affects the value of the wind speed but has little
where ρ(X, Y) is the Pearson correlation coefficient of random effect on the shape of the wind profile on the bridge deck, so the
variables X and Y; Xi and Yi mean wind speed profile samples of i wind speed can not be discussed in subsequent studies.
(i = 1, 2, 3, 4); X and Y are the average value of the Xi and Yi,
respectively. 3.2 Influence of the AoA of Wind
As described in Section 2.2.2, the AoAs of -7°, -5°, -3°, 0°, +3°,
3. Results and Discussion +5°, and +7° with a wind speed of 10 m/s were analyzed to
investigate the influence of the AoA on the wind field characteristics
3.1 Influence of the Reynolds Number of the bridge deck. Fig. 8 shows that the wind speed at the same
In order to explore the influence of the Reynolds number on the height decreases with the increase of the AoA within a certain
wind field characteristics of the bridge deck, the CFD results height range (see the shaded part in the figure). With the increase
need to be verified first. It is worth noting that lanes 1 and 2 are of AoA from -7° to +7°, the height needed for the measuring
about 3.5 m away from the influence of the bridge deck, while point (MP) to reach the velocity of the incoming flow increases
lanes 3 and 4 need about 5 m. In addition, the wind speed at the gradually. The wind field on the bridge deck is affected by the
height away from the influence of the bridge deck is slightly flow around the main girder components and the ancillary structures
higher than the incoming wind speed. This is mainly because the above the bridge deck, such as the barrier of railings and central
main girder affects the wind field, making the fluid in an extensive separator. In addition, the boundary layer is generated when the
range above the bridge deck accelerate. Thus, this phenomenon wind flows through the bridge deck, which makes the wind
can provide some guidance for field measurement. Specifically, speed change in a certain height range along the bridge deck.
compared with the lane on the leeward side, the wind speed at the With a further study for this phenomenon, it is found that the
lower position on the windward side can relatively accurately blocking effect of the main girder causes the vortex at the back of
represent the wind speed of the incoming flow. It is recommended the incoming flow. As described in Fig. 9, the vortex above the
that the measuring point (MP) used to record the data of the bridge deck gradually increases with the increase of the AoA so
incoming flow in the field measurement be placed in the windward that the influence range of the bridge surface on the wind field
measurement as far as possible, and the height should not be less becomes larger. Combined with Figs. 8 and 9, it can be concluded
than 3.5 m away from the bridge deck. that the wind speed at the height of 5 m in lane 1 corresponding
Furthermore, the influence of different Reynolds numbers on to different AoAs is stable and close to the incoming wind speed.
the wind speed profile of the bridge deck is studied by changing It also verifies the rationality of using the samples from the
 KSCE Journal of Civil Engineering 5153

Fig. 8. Wind Speed Profiles of Each Lane at Different AoAs: (a) In Lane 1, (b) In Lane 2, (c) In Lane 3, (d) In Lane 4

Fig. 9. Section Streamlines at the Center of Segment Corresponding Different AoAs: (a) In Lane 1, (b) In Lane 2, (c) In Lane 3, (d) In Lane 4
5154 M. Zhang et al.

Table 4. Equivalent Wind Speed of Different Lanes at Different Angles of Attack (AoA)

AoA Lane 1 Lane 2 Lane 3 Lane 4

(°) VeqS VeqR VeqS VeqR VeqS VeqR VeqS VeqR
-7 6.699 5.418 6.240 5.137 4.911 4.515 4.556 4.162
-5 6.625 5.399 6.052 5.020 4.631 4.257 4.138 3.779
-3 6.504 5.357 5.790 4.855 4.302 3.961 3.694 3.374
-2 6.439 5.335 5.638 4.758 4.106 3.789 3.436 3.148
-1 6.396 5.327 5.524 4.681 3.920 3.627 3.187 2.930
0 6.462 5.458 5.430 4.657 3.804 3.528 3.017 2.802
1 5.516 4.856 4.102 3.682 2.770 2.562 2.223 2.081
2 5.388 4.766 3.893 3.509 2.460 2.289 2.005 1.877
3 5.139 4.579 3.504 3.174 2.143 1.994 1.574 1.462
5 5.012 4.419 3.346 2.998 1.862 1.725 1.212 1.055
7 4.074 3.728 1.830 1.699 0.687 0.647 0.663 0.462

Fig. 10. Equivalent Wind Speed in Different Lanes Based on the: (a) Equivalent Side Force, (b) Equivalent Rolling Moment

windward side to represent the incoming flow. At the same time, result in the wind speed above the bridge deck significantly
it can be found from Fig. 9 that the vortex near the bridge deck lower. In addition, it can be noted from the results that the
makes the wind speed of the bridge deck decrease significantly equivalent wind speed based on the equivalent side force is
under the large AoA, and the energy contained in the incoming generally more significant than the value based on the rolling
flow is dissipated in the vortex. Therefore, a vortex on the bridge moment, with the same variation trend. It shows that the vehicle
deck is beneficial to the traffic safety on the bridge, and the is more vulnerable to the side force on the bridge deck. Therefore,
negative AoA should be noted when analyzing the wind-induced two conclusions can be drawn as follows: 1) the equivalent wind
traffic safety problems on the bridge deck. speed based on the equivalent side force should be used to analyze
To further study the influence of the variation of the local the vehicle sideslip stability when the road is slippery and the vehicle
wind field at the bridge deck for the driving performance, the is prone to sideslip; 2) the equivalent wind speed based on the
equivalent wind speed was introduced to make it more intuitive equivalent rolling moment should be used to analyze the vehicle
to evaluate the wind field within the driving height range. sideslip stability when the vehicle is high and no-load.
Generally, there are two principles to derive the equivalent wind Moreover, the reduction coefficients of wind speed in different
speed, one is based on the equivalent side force (see Eq. (3)), and lanes are illustrated in Table 5, and the difference of different
the other is based on the rolling moment (see Eq. (4)). The lanes at different angles of attack (AoA) are shown in Figs. 11(a)
equivalent wind speed in different lanes based on two principles and 11(b). It can be seen from the diagrams that the reduction
is listed in Table 4, and Fig. 10 can be obtained. The results show coefficient of the lane on the leeward side is greater than the
that the larger the AoA, the smaller the equivalent wind speed, as value on the windward side. It is primarily due to the combined
a result of the influence of the blocking effect of the main girder effect of the obstruction of ancillary facilities, the bridge deck's
 KSCE Journal of Civil Engineering 5155

Table 5. Reduction Coefficients of Different Lanes at Different Angles of Attack (AoA)

AoA Lane 1 Lane 2 Lane 3 Lane 4

(°) λVeqS λVeqR λVeqS λVeqR λVeqS λVeqR λVeqS λVeqR
-7 0.33 0.46 0.38 0.49 0.51 0.55 0.54 0.58
-5 0.34 0.46 0.39 0.50 0.54 0.57 0.59 0.62
-3 0.35 0.46 0.42 0.51 0.57 0.60 0.63 0.66
-2 0.36 0.47 0.44 0.52 0.59 0.62 0.66 0.69
-1 0.36 0.47 0.45 0.53 0.61 0.64 0.68 0.71
0 0.35 0.45 0.46 0.53 0.62 0.65 0.70 0.72
1 0.45 0.51 0.59 0.63 0.72 0.74 0.78 0.79
2 0.46 0.52 0.61 0.65 0.75 0.77 0.80 0.81
3 0.49 0.54 0.65 0.68 0.79 0.80 0.84 0.85
5 0.50 0.56 0.67 0.70 0.81 0.83 0.88 0.89
7 0.59 0.63 0.82 0.83 0.93 0.94 0.93 0.95

Fig. 11. Reduction Coefficients of Wind Speed in Different Lanes Based on the: (a) Equivalent Side Force, (b) Equivalent Rolling Moment

roughness, and the vortex's scale and location. 3.3 Vehicle Aerodynamic Coefficient
Zr
The driving safety of vehicles under crosswind is significantly
1
∫U (3) affected, so it is essential to evaluate its aerodynamic parameters
2
VeqS = ( z )dz
Zr 0 (Guo et al., 2020; Zhang et al., 2020a). Thus, to better understand
the aerodynamic coefficients of vehicles in different lanes more
Zr clearly, three vehicle models (commercial van, traveling car, and
1
∫ z ⋅U
2 (4)
VeqR = ( z )dz articulated car) are placed in the center of each lane to obtain the
Zr 2 0
aerodynamic coefficients in each lane. The aerodynamic coefficients
include side (CS), lift (CL) force coefficients, rolling (CR), yaw
U 0 − Veq
λs = (5) (CY), and pitching (CP) moment coefficients, and the related
U0
formula are as Eqs. (6) − (10). The definitions of the coordinate
here VeqS and VeqR are the equivalent wind speed based on the system and the direction of aerodynamic forces are shown in Fig.
equivalent side force and equivalent rolling moment, respectively; 12, where the angle of attack (AoA) α is defined as the angle
U(z) means the normalized wind speed at a height z above the between the incoming flow and the vertical direction, and the AoA
bridge deck; λs is the reduction coefficient of wind speed, and U0 is positive when the vertical component of the incoming flow is
is the wind speed of the incoming flow; Zr is the corresponding upward. In addition, the calculation adopts the CFD results when the
height above the deck of the equivalent wind speed. Considering incoming wind speed is set as 8 m/s with the AoA of 0°.
the height of a high-sided vehicle, the value of Zr corresponding
2 FL
to three vehicle models can refer to Section 2.2.2. CL = (6)
ρUAs
5156 M. Zhang et al.

resulting in the decrease of the wind speed at the bottom of the

vehicle and the increase of pressure at the bottom of the vehicle.
Thus, the lifting force reverses. At the same time, it can be found
that the absolute value of the aerodynamic coefficients of the
articulated car, traveling car, and commercial van gradually
decreases, which indicates that the larger the side windward area
is, the larger the side force coefficient is. Therefore, under the
condition of the same wind speed, the side force on the vehicle is
more remarkable, which is not conducive to its driving safety.
Fig. 12. Definition of Vehicle Aerodynamic Forces The rollover moment coefficients of the traveling car and the
commercial van are relatively similar and less than the value of
the articulated car. It is mainly because the similar vehicle height
2 FS
CS = (7) makes the distribution of wind force on the car similar and
ρUAs
makes the rollover moment coefficient similar. Specifically, the
2M R vehicle height of the traveling car and commercial van is 3.30 m
CR = (8) and 3.35 m, respectively, while the height of the articulated car is
ρUAs hv
3.78 m.
2M Y
CY = (9) 4. Conclusions
ρUAs hv

2M P The wind field characteristics above the bridge deck of the

CP = (10) Xingkang Bridge on Luding Dadu River have been investigated
ρUAs hv
in this paper by the CFD. On this basis, the CFD results are
where FL, FS, MP, MY, MR represent the mean values of lift force, discussed, including the wind field characteristics (the profiles of
side force, pitching moment, yaw moment, and rolling moment, wind speed & angle of attack), vehicle aerodynamic parameters,
respectively, and CL, CS, CP, CY, CR are the corresponding and the location of measuring points (MPs), and four conclusions
coefficients; U is the wind speed of the incoming flow, and ρ is can be drawn as follows:
the air density; As is side area of the vehicle, and hv is the height 1. Reynolds number has little effect on the shape of the wind
of the vehicle gravity center from the bridge deck. profile on the bridge deck, but has a great effect on the
Table 6 lists the aerodynamic coefficients of different vehicles, wind speed. In addition, Due to the influence of the main
based on this, the result can be obtained and shown in Fig. 13. It girder, the fluid in an extensive range above the bridge deck
can be seen from the figure that the CS of different cases are produces an acceleration effect, and the wind speed is
positive, which indicates that the side force on the vehicle has slightly higher than the incoming wind speed.
always remained in the same direction as the flow direction. In 2. With the increase of the angle of attack, the height required
addition, the lift coefficients CL in lanes 1 − 3 are negative, but for the wind field to separate from the influence of the
the value in lane 4 is positive. It is due to that the wind speed near bridge deck will gradually increase. On the windward side,
the bridge deck is deficient when the vehicle is located in lane 4, when the height is greater than 5 m, the wind speed reaches

Table 6. Aerodynamic Coefficients of Different Vehicles

Type Lane No. CS CL CMP CMR CMY
Commercial van 1 -0.8660 -0.1058 -0.0107 -0.0094 -0.0312
2 -0.6642 -0.0240 0.0001 -0.0281 -0.0402
3 -0.4301 -0.0007 -0.0019 -0.0721 -0.0131
4 -0.2241 0.0598 0.0034 -0.0741 -0.0184
Travelling car 1 -1.3372 -0.0792 0.0007 0.0052 0.0032
2 -1.0418 -0.0173 0.0002 -0.0250 0.0008
3 -0.7050 -0.0607 -0.0009 -0.0526 0.0046
4 -0.3502 0.0625 0.0007 -0.0706 0.0072
Articulated car 1 -2.3472 -0.1112 -0.0184 0.0766 -0.0081
2 -1.9192 -0.0849 -0.0122 0.0550 -0.0145
3 -1.4689 0.0027 -0.0068 -0.0188 -0.0049
4 -0.9887 0.0631 0.0008 -0.0541 -0.0076
 KSCE Journal of Civil Engineering 5157

Fig. 13. Aerodynamic Coefficients of Different Vehicles Corresponding to: (a) The Side Force Coefficient, (b) The Lift Coefficient, (c) The Pitching
Moment Coefficient, (d) The Yaw Moment Coefficient, (e) The Rolling Moment Coefficient

a fixed value close to the incoming wind speed; thus, it can time-dependent phenomenon, the averaged flow field
be used as the reference point of incoming wind speed and produced less force in the wake, but there are the high local
provide an important reference for the field measurement wind speed variations that may affect the traffic unfavorably.
on the truss bridge deck. 4. The equivalent wind speed based on the equivalent side
3. In order to analyze the wind-induced traffic safety problems force is generally more significant than the value based on
on the bridge deck, it should be noted that the vortex is a the rolling moment, with the same variation trend. Therefore,
5158 M. Zhang et al.

the equivalent wind speed based on the equivalent side 95(5):355-369, DOI: 10.1016/j.jweia.2006.08.002
force should analyze the vehicle's sideslip stability when He X, Xue F, Zou Y, Chen S, Han Y, Du B, Xu X, Ma B (2020) Wind
the road is slippery and prone to sideslip. In addition, the tunnel tests on the aerodynamic characteristics of vehicles on
highway bridges. Advances in Structural Engineering 23(13):2882-
equivalent wind speed based on the equivalent rolling
2897, DOI: 10.1177/1369433220924791
moment should be used to analyze the vehicle sideslip Kim S-J, Yoo C-H, Kim H-K (2016) Vulnerability assessment for the
stability when the vehicle is high and no-load. hazards of crosswinds when vehicles cross a bridge deck. Journal of
Wind Engineering and Industrial Aerodynamics 156:62-71, DOI:
Acknowledgments 10.1016/j.jweia.2016.07.005
Kramer C, Grundmann R, Gerhardt HJ (1991) Testing of road vehicles
The work described in this paper was supported by National under cross wind conditions. Journal of Wind Engineering and Industrial
Natural Science Foundation of China (NSFC) under grant No. Aerodynamics 38(1):59-69, DOI: 10.1016/0167-6105(91)90027-T
Launder BE, Spalding DB (1974) The numerical computation of turbulent
U21A20154.
flows. Computer Methods in Applied Mechanics and Engineering
3(2):269-289, DOI: 10.1016/0045-7825(74)90029-2
ORCID Li Y, Hu P, Cai C, Zhang M, Qiang S (2013) Wind tunnel study of a
sudden change of train wind loads due to the wind shielding effects
Mingjin Zhang https://orcid.org/0000-0001-5681-5901 of bridge towers and passing trains. Journal of Engineering Mechanics
Jinxiang Zhang https://orcid.org/0000-0002-6198-4942 139(9):1249-1259, DOI: 10.1061/(ASCE)EM.1943-7889.0000559
Junting Long https://orcid.org/0000-0002-8288-5898 Ma L, Chen X, Wu J, Han W (2018) Aerodynamic interference mechanism
Yongle Li https://orcid.org/0000-0001-9879-0626 of moving vehicles on a bridge deck in crosswind environment.
Journal of Bridge Engineering 23(4), DOI: 10.1061/(ASCE)BE.
Yulin Zou https://orcid.org/0000-0003-4537-5124
1943-5592.0001194
Dianguo Yin https://orcid.org/0000-0003-1131-1464 Ni Z, Zhang C, Zhang L, Liu L (2019) Effects of the wide-body
suspension bridge auxiliary structure on flutter characteristics by
References CFD. Vibroengineering PROCEDIA 28:223-229, DOI: 10.21595/
vp.2019.21080
Charuvisit S, Kimura K, Fujino Y (2004) Experimental and semi-analytical Peng Y, Wang S, Li J (2018) Field measurement and investigation of
studies on the aerodynamic forces acting on a vehicle passing spatial coherence for near-surface strong winds in Southeast China.
through the wake of a bridge tower in cross wind. Journal of Wind Journal of Wind Engineering and Industrial Aerodynamics 172:423-
Engineering and Industrial Aerodynamics 92(9):749-780, DOI: 440, DOI: 10.1016/j.jweia.2017.11.012
10.1016/j.jweia.2004.04.001 Ren W, Duan Q, Ma C, Liao H, Li Q (2019) Experimental investigation
Cheli F, Ripamonti F, Sabbioni E, Tomasini G (2011) Wind tunnel tests of the protective effect of wind barriers on high-speed railway bridge
on heavy road vehicles: Cross wind induced loads — Part 2. Journal in inland strong wind area. Advances in Structural Engineering
of Wind Engineering and Industrial Aerodynamics 99(10):1011- 22(15):3306-3318, DOI: 10.1177/1369433219862939
1024, DOI: 10.1016/j.jweia.2011.07.007 Rezaeiha A, Montazeri H, Blocken B (2019) On the accuracy of
Chen Z-S, Zhang C, Wang X, Ma C-M (2017) Wind tunnel measurements turbulence models for CFD simulations of vertical axis wind turbines.
for flutter of a long-afterbody bridge deck. Sensors 17(2):335, DOI: Energy 180:838-857, DOI: 10.1016/j.energy.2019.05.053
10.3390/s17020335 Salati L, Schito P, Rocchi D, Sabbioni E (2018) Aerodynamic study on a
Dong L, Lio WH, Simley E (2021) On turbulence models and lidar heavy truck passing by a bridge pylon under crosswinds using CFD.
measurements for wind turbine control. Wind Energy Science 6(6): Journal of Bridge Engineering 23(9):04018065, DOI: 10.1061/(ASCE)
1491-1500, DOI: 10.5194/wes-6-1491-2021 BE.1943-5592.0001277
Guo P, Li S, Wang D (2020) Analysis of wind attack angle increments in Wang L, Chen X, Chen H (2021) Research on wind barrier of canyon bridge-
wind tunnel tests for the aerodynamic coefficients of iced hangers. tunnel junction based on wind characteristics. Advances in Structural
Advances in Structural Engineering 23(4):603-613, DOI: 10.1177/ Engineering 24(5):870-883, DOI: 10.1177/1369433220971730
1369433219876206 Wang H, Tao T, Wu T, Mao J, Li A (2015) Joint distribution of wind
Guo J, Tang H, Li Y, Wang Z (2021) Effects of guardrails on wind speed and direction in the context of field measurement. Wind and
environment for vehicles and aerodynamic stability for bridges with Structures 20(5):701-718, DOI: 10.12989/was.2015.20.5.701
box girders. Advances in Structural Engineering 24(3):453-469, Wang B, Xu Y-L (2015) Safety analysis of a road vehicle passing by a
DOI: 10.1177/1369433220956827 bridge tower under crosswinds. Journal of Wind Engineering and
Han Y, Cai C, Zhang J, Chen S, He X (2014) Effects of aerodynamic Industrial Aerodynamics 137:25-36, DOI: 10.1016/j.jweia.2014.11.017
parameters on the dynamic responses of road vehicles and bridges Wang D, Zhang Y, Sun M, Chen A (2020) Characteristics of the wind
under cross winds. Journal of Wind Engineering and Industrial environment above bridge deck near the pylon zone and wind
Aerodynamics 134:78-95, DOI: 10.1016/j.jweia.2014.08.013 barrier Arrangement Criteria. Applied Sciences 10(4):1437, DOI:
Han Y, Hu J, Cai C, Chen Z, Li C (2013) Experimental and numerical 10.3390/app10041437
studies of aerodynamic forces on vehicles and bridges. Wind and Yang Y, Gu M, Chen S, Jin X (2009) New inflow boundary conditions
Structures 17(2):163-184, DOI: 10.12989/was.2013.17.2.163 for modelling the neutral equilibrium atmospheric boundary layer in
Hargreaves DM, Wright NG (2007) On the use of the k–model in computational wind engineering. Journal of Wind Engineering and
commercial CFD software to model the neutral atmospheric boundary Industrial Aerodynamics 97(2):88-95, DOI: 10.1016/j.jweia.2008.12.001
layer. Journal of Wind Engineering and Industrial Aerodynamics Yuan D, Zheng S, Hong C, Zhu J (2018) Wind environment around the
 KSCE Journal of Civil Engineering 5159

tower of long span railway-highway combined cable-stayed bridge. Zhang J, Zhang M, Li Y, Jiang F, Wu L, Guo D (2021) Comparison of
Journal of Harbin Institute of Technology 50(9):19-24, DOI: 10.11918/ wind characteristics in different directions of deep-cut gorges based
j.issn.0367-6234.201803143 on field measurements. Journal of Wind Engineering and Industrial
Zhang J, Zhang M, Jiang X, Wu L, Qin J, Li Y (2022) Pair-Copula- Aerodynamics 212:104595, DOI: 10.1016/j.jweia.2021.104595
based trivariate joint probability model of wind speed, wind direction Zhang M, Yu J, Zhang J, Wu L, Li Y (2019) Study on the wind-field
and angle of attack. Journal of Wind Engineering and Industrial characteristics over a bridge site due to the shielding effects of mountains
Aerodynamics 225:105010, DOI: 10.1016/j.jweia.2022.105010 in a deep gorge via numerical simulation. Advances in Structural
Zhang J, Zhang M, Li Y, Huang X, Zheng Z (2020a) Aerodynamics of Engineering 22(14):3055-3065, DOI: 10.1177/1369433219857859
high-sided vehicles on truss girder considering sheltering effect by Zhang M, Zhang J, Li Y, Yu J, Zhang J, Wu L (2020b) Wind characteristics
wind tunnel tests. The Baltic Journal of Road and Bridge Engineering in the high-altitude difference at bridge site by wind tunnel tests.
15(2):66-88, DOI: 10.7250/bjrbe.2020-15.473 Wind and Structures 30(6):547-558, DOI: 10.12989/WAS.2020.30.6.547