Aerodynamic Characteristics Evaluation of a Cable-Stayed Bridge Section with a New-Type Hybrid Fairing

In this study, new-type hybrid faring is suggested to improve the aerodynamic performance of the long-span cable-stayed bridge. The proposed fairing is developed by applying the concept of the multibox section to the normal faring. The proposed faring has void regions inside the faring so that wind passes through the gaps in the faring. As a result, the wind ﬂow is changed and the forces to the bridge section are reduced. The eﬃciency of the proposed faring was veriﬁed by a series of wind tunnel test. From the test result, it can be found that aerodynamic performances, such as drag force and ﬂutter resistance, are enhanced.


Introduction
Aerodynamic forces are the main design load for the longspan cable-stayed bridge, and several studies have been conducted to improve the aerodynamic performance of such bridges. One of the effective methods to improve the aerodynamic performance of the bridge is to change the bridge section shape. ere are three different approaches to changing the bridge section shape. e first method is direct modification of the section dimensions, such as the height, width, and slope of the section change, or relocation of the girder and barrier. e second method is the installation of additional attachments, such as fairing and flap, to the existing bridge section. e third method is the development of a new concept bridge section. An example of the new concept bridge section to improve the aerodynamic performance is the multibox section. Stonecutters bridge (main span � 1,018 m) and Messina bridges (main span � 3,300 m) are designed with twin and triple-box sections, respectively. By applying the multibox section, the winds pass through the void region between the boxes, and the aerodynamic performance can be improved. Studies to improve the aerodynamic characteristics with new ideas and technologies have been conducted by previous researchers [1,2].
Among these methods, the focus of this study was made on the proposal of new-type hybrid faring by adding the concept of multibox section to the normal faring, as shown in Figure 1.
e proposed hybrid fairing is basically an additional attachment, and it is used for an existing bridge to improve the aerodynamic performance. In this study, the target bridge section is a single box section. By applying the concept of the multibox section to the normal fairing, void regions inside the faring were provided. As a result, winds are passed through a void in the fairing and the wind flow of the section is changed. To evaluate the performance of the proposed hybrid fairing, static and dynamic wind tunnel tests for the bridge deck were conducted. e main parameter is the wind penetration ratio, which is defined as the void area to the total area of the fairing. Wind penetration ratios of 0%, 25%, and 50% were selected for the tests, where 0% wind penetration ratio fairing represents the normal fairing without the void. e tests show that the aerodynamic characteristics of the section were improved by applying the proposed new-type hybrid fairing.

Previous Researches and the Proposed New-
Type Hybrid Fairing 2.1. Previous Researches. e cross section of the bridge is an important parameter that affects the aerodynamic characteristics of the long-span cable-stayed bridge. e static wind load is calculated by using the aerodynamic drag coefficient, which is determined by the section shape of the bridge. e vortex characteristic is also closely related to the section shape of the bridge. e Strouhal number, which varies depending on the section shape, affects the vortex-induced vibration speed and amplitude. Furthermore, the section shape affects the flutter speed. Vortex-induced vibration is motions induced on structures interacting with an external wind flow and it occurs when the vortex shedding and the structural vibration synchronize at common frequency. Flutter is one of the aeroelastic instabilities, and it can lead the collapse of the bridge. e flutter occurs when the total damping of the structure becomes zero by the structurewind flow interaction. As a result, the displacement (or rotation) diverges. In consequence, several previous studies have been conducted for the effect of the cross section shape of the bridge on the aerodynamic characteristics, such as the static wind load, vortex-induced vibration, and flutter.
As mentioned before, there are three different methods to change the bridge cross section to enhance the aerodynamic performance, as shown in Figures 2(a)-2(c). e first method is direct modification of the dimensions of the ordinary section. e second method is the use of attachments to the section. e third method is to develop a new concept section. In the case of the first method, the following recent research efforts have been conducted. Larsen and Wall [3] investigated the cross section shape of the box girder bridge to prevent the vortex shedding response. ey reported that when the angle of the lower face is less than 16°, the vortex shedding response is reduced. Fransos and Bruno [4] studied the aerodynamic characteristics of the trapezoidal box girder section by using computational fluid dynamics (CFD) analysis. e analysis results showed that the aerodynamic behavior is affected by the lower corner degree of sharpness of the trapezoidal box section. Lee et al. [5] studied the effect of the edge-box girder section shape on the static wind load coefficients through a wind tunnel test and CFD analysis. From the results, the optimum slope angle of the section was proposed. Also, Daito et al. [6] conducted similar research for the edge-box girder section focusing on the dynamic characteristics.
When the shape modification of the section is difficult, additional aerodynamic attachments are used to improve the aerodynamic performance of the bridge. Although these attachments are nonstructural members, they change the wind flow near the bridge section. Fairings, flaps, edge plates, side plates, baffle plates, and gratings are used as additional attachments [7]. Among these attachments, the fairing is widely used, since it is easy to handle. A triangular shape fairing is generally used since it segregates the streamline near the section, and several studies have been conducted for it. Nagao et al. [8] investigated the effect of the size of the triangular shape fairing on the aerodynamic behavior of the box girder. Similar studies were performed for the edge-box girder and trapezoidal box girder by Sukamta et al. [9] and Hanque et al. [10], respectively. Also, there are several types of fairing, such as the half-circular and the propeller type. e research into these fairings has been conducted by Tanaka et al. [11].
Since the mid-2000s, to overcome the length limit of the cable-stayed bridge, the multi-section was proposed [Refer en, aerodynamic lift force and moment can be reduced, without changing the B/D ratio [12,13]. e multisection is called by several different names as the slotted box girder [14], flat box girder [15], tandem box [16], twin box [17], and two separated steel boxes [18]. Sometimes, the multisection consists of more than two separated boxes. For example, Messina bridge has a triple deck system, and Diana and Fiammenghi [19] conducted a wind tunnel test for it. Recently, Wang and Dragomirescu [20] proposed a fourth deck system, and Zasso et al. [21] investigated the effect of the number of boxes (from one to three) in multisection on the aerodynamic behavior.
In addition to the methods mentioned above, several challenging studies have been conducted to improve the aerodynamic performance of the bridge. For example, Li et al. [22] and Gouder et al. [23] proposed an active wind load feedback control method by sensors and actuator to change the wind flow near the bridge and improve the flutter resistance.

e Proposed
New-Type Hybrid Fairing. As a part of such challenging studies mentioned in Section 2.1, a new-type hybrid fairing is proposed by mixing the concept of normal fairing and multi-section, as shown in Figure 1. Basically, the proposed attachment is the fairing, and the focus is made on the improvement of the static and dynamic aerodynamic performance of the normal fairing, by adding the advantage of the multisection. e performance of the proposed newtype hybrid fairing was evaluated by a series of wind tunnel tests. e main test parameter is the wind penetration ratio, which is defined as the ratio of void area to total area of the fairing. us, the wind penetration ratio (WPR) is defined as where the definition of the void and the total area of the fairing is shown in Figure 3. By applying the void region to the normal fairing, it is expected that the drag force is reduced, due to the wind that is passing through the void region. Also, the increase of flutter speed is expected by reducing the pressure difference between the upper and lower parts of the bridge section.

Test Bridge and Section.
To verify the performance of the proposed fairing, static and dynamic wind tunnel tests were conducted. Figure 4(a) shows the target cable-stayed bridge where the main span of the bridge is 1,200 m. Figure 4(b) shows the cross section of the bridge. e single steel box section is considered where the total width and the depth of the section are 29 m and 4 m, respectively, with 4 lanes. e width of the steel box section is 24.8 m. At the left and right ends of the steel box, the triangular shape fairing with the width of 2.1 m is installed, as shown in Figure 4(b). us, the total width is equal to 29 m.
In this study, the main parameter is the wind penetration ratios (WPR) and WPR of 0% (the normal fairing), 25%, and 50% were considered. With increasing WPR, the effect of WPR on the aerodynamic characteristics may be improved. However, the structural rigidity is decreased. us, the maximum WPR was set as 50% so that the fairing can have enough structural rigidity to handle. 2D wind tunnel test was conducted for each case. A 1 : 70 scale model was used for the tests. Figure 5 shows the example of the test model used in this study.

Wind Tunnel Test and Parameters.
e wind tunnel test is often used to investigate the effect of wind moving past the structure. e test model is mounted in the middle of the wind tunnel and it instrumented with several sensors to measure the response of the test model. e static wind tunnel test, where the test model is fixed with load cells at their ends, is used to investigate the wind force acting on the model. e vortex-induced vibration and flutter response can be investigated by using the dynamic wind tunnel test, where the test model is mounted on the spring support system.
2D wind tunnel test has been widely used to investigate the aerodynamic behavior of the bridge as an alternative of full 3D wind tunnel test. In the case of 2D wind tunnel test, mass, stiffness, frequency, and damping ratio of whole model (3D model) are firstly determined. en, these factors are considered in 2D wind tunnel test setup to reflect 3D bridge behavior. Table 1 shows the unscaled properties of original bridge and scaled properties for the wind tunnel test. e unscaled vertical and torsional frequencies are 0.1947 and 0.5044 Hz, respectively. us, the ratio of vertical frequency to torsional frequency, V/T ratio, is 2.5910. e unscaled mass per unit length of the section is 17,906 kg/m. e unscaled moment of inertia is equal to 898.79 kN/g·m 2 . e damping ratio is determined as 0.4%. ese values were scaled for the wind tunnel test with 95% accuracy, as shown in Table 1.
e wind tunnel in Korea University was used for the test, where the size of wind tunnel is 1.0 m × 0.8 m (Width × Height). Figure 6 shows the test setup for the wind tunnel test. In the case of the static wind tunnel test, the drag, lift forces, and moment are measured from the 3-    Shock and Vibration components load shown in Figure 6(a). For dynamic wind tunnel test, the spring support system shown in Figure 6(b) was used. e blockage ratio represents the ratio of projection area of the model to total area of 2D wind tunnel section. is ratio should be small enough (usually less than 5%) so that the wind flow in 2D wind tunnel section is not affected by the test model. In this study, the blockage ratio was 3.64%.
where C d is the aerodynamic drag force coefficient, C l is the aerodynamic lift force coefficient, C m is the aerodynamic moment coefficient, F d is the drag force, F l is the lift force, F m is the moment in the model, ρ is the density of air, U is the wind velocity, and B is the reference length.
(C.2) Table 2 shows the static wind tunnel test cases. e test was conducted for both laminar and turbulent flows, where the turbulent intensity was 10%. e maximum wind speed  e dynamic wind tunnel test was also conducted to evaluate the dynamic stability of the bridge with proposed newtype hybrid fairing. e dynamic wind tunnel test was conducted by using the spring support system shown in Figure 6(b). Similar to the static wind tunnel test, laminar and   turbulent flow with intensity of 10% were considered. Also, the same WPRs were used for the dynamic test. In the case of the angle of attack, three cases (−3°, 0°, and 3°) were considered. us, a total of 18 cases are tested. Generally, AOA varies from −3°to 3°in natural condition. us, in this study, the minimum and maximum AOA were considered with reference AOA (�0°) for the efficiency of the dynamic wind tunnel test.

Static Wind Tunnel Test
Results. From the static wind tunnel test, the static aerodynamic coefficients (C d , C l , and C m ) were extracted for each laminar and turbulent flow. Figure 7 shows the variation in C d with angle of attack (AOA) for the cases with the normal and new-type hybrid fairing.
From Figure 7(a), it can be seen that the cases with the proposed fairing have lower C d comparing with normal fairing for most AOA ranges. Figure 7(b) shows the amount of C d reduction. e difference is calculated from (C d of the proposed fairing-C d of the normal fairing)/C d of the normal fairing (%). In most cases, except for the case with AOA of 6°, the reduction of C d is larger than 10%. In general, the practical range of AOA is varied from −3°to 3°. In this range, the C d reduction is approximately 15% to 23%. is means that the static aerodynamic drag force can be reduced up to 15%-23% by applying the proposed new-type hybrid fairing. e winds pass through the void region of the proposed fairing. As a result, the wind acting on the bridge section is reduced compared to the case with the normal fairing. is leads to reduced C d for the case with the new-type hybrid fairing. Generally, except for the AOA of −9°, PWR of 25%  Shock and Vibration 7 shows better performance compared to WPR of 50%, as shown in Figure 7(b). Similar tests were conducted for turbulent flow condition and the results are shown in Figure 8. e overall trends of the test results are similar to those of the laminar flow condition. C d with the proposed fairing is lower than that with the normal fairing, as shown in Figures 8(a) and 8(b), except for the case with AOA of 9°. e effect of WPR on C d is not large, except for AOA of ±9°. Similar to the laminar flow condition, 15%-23% C d reduction compared to the case with the normal fairing was observed in the range of practical AOA (from −3°to 3°).
In summary, the proposed fairing shows lower C d compared to the normal fairing for both laminar and turbulent flow conditions. Figures 9(a) and 9(b) show the variation in C l with AOA for the laminar flow and turbulent flow conditions, respectively. For the lower AOA region, such as −9°to −6°, C l value with the proposed fairing is closer to 0 compared to the case with the normal fairing. is means that the proposed fairing can reduce the lifting force of the section. In the case of the practical AOA range from −3°to 3°, C l values are similar to each other for all the considered cases for laminar flow condition [refer to Figure 9(a)], while C l with the proposed fairing is slightly lower than that with the normal fairing for the turbulent flow condition [refer to Figure 9(b)]. By applying the proposed fairing, it is expected that the wind resisting area of the fairing will be reduced. However, it should be noted that most of the lifting force is applied to the bottom of the main section, and the area of the fairing is not large compared to the bottom area of the section, as shown in Figure 10. us, in the practical range of AOA, the reduction in C l by applying the proposed fairing is small. Figures 11(a) and 11(b) show the variations in C m with AOA for laminar and turbulent flow condition, respectively. For all considered AOA, the proposed fairing shows larger absolute values of C m compared to the normal fairing. us, the larger torsional moment to the section is expected when the proposed fairing is used, and the normal fairing shows better performance. Figure 12 shows a schematic of the wind flows for the torsion of the section. From Figure 12, it can be seen that the proposed fairing increases the wind resisting area due to the walls in the fairing. is results in increased C m for the proposed fairing comparing to the normal fairing. In summary, for the practical range of AOA, it can be found that the proposed fairing can reduce the drag force of the section, while increasing the torsional moment. e lifting forces are similar to each other for the cases with the normal and proposed fairing.

Dynamic Wind Tunnel Test Results.
e dynamic wind tunnel test was conducted for the section with the normal and proposed fairing for the practical range of AOA of from −3°to 3°and both laminar and turbulent flow conditions. Figures 13-15 represent the dynamic wind tunnel test results for laminar flow condition with −3°, 0°, and 3°AOA, respectively.
In Figures 13-15, the x-axis is the wind velocity, and the y-axis denotes the RMS value for vertical displacement or rotation angle. For AOA of −3°, when the normal fairing is used, the flutter was observed for both vertical and rotational directions, as shown in Figures 13(a) and 13(b), while for the proposed fairing, flutter does not occur. However, when the e vortex-induced vibration was observed for the case with WPR � 25% around wind velocity of 5.3 m/s. e larger vertical displacements occurred with increasing wind velocity for the case with the proposed fairing. In the case of torsion, torsional behaviors are similar to each other for all cases, except the vortex-induced vibration for the 25% WPR case. Figure 15 shows the test results for AOA of 3°. For the case with normal fairing, both the vortex-induced vibration and flutter occurred. Flutter was also observed for the case with the WPR of 25%. e flutter speed of the case with the normal fairing was 68.8 m/s for vertical and torsional directions. In the case of WPR of 25%, the flutter speed was 85.7 m/s for the vertical and torsional directions. us, it can be seen that the proposed fairing increases the flutter speed even if flutter occurs. When WPR is 50%, there is no vortexinduced vibration and flutter from the dynamic test with laminar flow condition.
Similar to the laminar flow condition, the dynamic wind tunnel test was conducted for the turbulent flow with intensity of 10%. In turbulent flow, both the magnitude and direction of the wind flow are continuously changed. us, overall RMS response in turbulent flow is usually lager than that of laminar flow. Also, vortex-induced vibration is reduced in turbulent flow comparing to the case with laminar wind flow since the turbulent flow contains many different-frequency components of the wind. e vortexinduced vibration occurs when the vortex shedding and the structural vibration synchronize at common frequency. Similar behaviors were observed for the dynamic test conducted in this study. Figures 16-18 show the dynamic test results. When AOA is −3°and 3°, the torsional flutter occurs for the case with the normal fairing, as shown in Figures 16(b) and 18(b), respectively. For AOA of 0°, when the normal fairing is applied, both vertical and torsional flutter were observed, as shown in Figures 17(a) and 17(b). However, when the proposed fairing is used, there were no considerable instabilities. e flutter did not occur with the wind velocity considered. From the dynamic test results, it can be found that the proposed new-type hybrid fairing shows better flutter resistance comparing with the section with the normal fairing. In some cases, especially for laminar flow condition, large vortex-induced vibration occurred. is vortex-induced vibration is often observed in the multibox section [19], and similar behavior occurs in the proposed new-type hybrid fairing.

Conclusions
is study proposes the concept of a new-type hybrid fairing. To improve the aerodynamic performance of the normal fairing, the concept of the multibox section is applied. en, the aerodynamic characteristics of the proposed fairing are Shock and Vibration evaluated from a series of wind tunnel tests. From the wind tunnel test, the major findings are as follows: (1) From the static wind tunnel test, it can be found that the proposed fairing shows lower C d comparing with the normal fairing for both laminar and turbulent flow conditions in the practical range of AOA of −3°t o 3°. is is because the wind passes through the gaps in the fairing, and the wind acting on the bridge section decreases. us, by applying the proposed fairing, reduction in drag force is expected.
(2) In the case of C l , the effect of the proposed fairing is negligible for the practical range of AOA. Also, by applying the proposed fairing, C m is increased due to the increase of the wind resisting area. us, when the proposed fairing is applied, the torsional behavior should be carefully examined. (3) From the dynamic wind tunnel test, it can be seen that the proposed fairing improves the flutter resistance for all the considered cases. us, the use of the proposed fairing is an effective way to increase the flutter speed of the long-span cable-stayed bridge. However, in some cases, large vortex-induced vibration is observed. is vortex-induced vibration should be carefully considered when the proposed fairing is used.
Data Availability e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.