The Seventh Asia-Pacific Conference on Wind Engineering, November 8-1, 009, Taipei, Taiwan WIND TUNNEL TEST WITH MOVING VEHICLE MODEL FOR AERODYNAMIC FORCES OF VEHICLE-BRIDGE SYSTEMS UNDER CROSS WIND ABSTRACT Yong-le Li 1, Peng Hu 1, Ming-jin Zhang 1, Hai-li Liao 1 1 Professor, Department of Bridge Engineering, Southwest Jiaotong University Chengdu 610031, Sichuan, P.R. China, lele@swjtu.edu.cn The complicated interactions among wind, vehicle and bridge should be considered in the wind-vehicle-bridge (WVB) systems. The aerodynamic forces of vehicle and bridge are the important parameters in studying the coupled vibration of the WVB systems. These parameters are usually obtained by theoretical analysis or computational fluid dynamics (CFD) simulation or wind tunnel model test. Most researches often neglect the relative motion between vehicle and bridge. In this paper, a new experimental setup is presented with considering the movement of vehicles to measure the aerodynamic forces of a moving vehicle-bridge system under the cross wind. Various cases are carried out to demonstrate more realistic roles of each factor in the system, such as vehicle speed, vehicle location on different tracks, the effect of head car and tail car, and etc. KEYWORDS: AERODYNAMIC FORCES, MOVING VEHICLE, WIND-VEHICLE-BRIDGE SYSTEM, WIND TUNNEL TEST, WIND LOAD Introduction The coupled vibration between moving rail vehicles and a bridge under cross wind would occur due to the interaction among wind,vehicle and bridge, which represents the special features of the wind-vehicle-bridge (WVB) system [1]. Since wind loads on vehicles will be affected by the existence of the bridge deck and vice versa, effects of static and dynamic wind loads acting on vehicles may be of great significance in the WVB system, especially when train vehicles run at high speed. Interaction between wind and vehicles running on the ground was investigated often through either theoretical analysis or computational fluid dynamics (CFD) simulation or wind tunnel model test.[1] Nevertheless, due to its great complexity and difficulty, the wind tunnel experiment is still the major and most reliable tool for this problem[]. In order to obtain aerodynamics on vehicles, conventionally, section model test are conducted by static methodologies with the models of vehicles fixed in the wind tunnel statically [4]. These static tests do not simulate the relative motion between vehicles and bridge deck or ground, thus do not take into account the effect of vehicles relative motion [3]. In the early 1980 s, to overcome shortcomings of static methodologies, C.J. Baker [3] carried out some experiments with a train model of 1/50 th scale running across the wind tunnel by some well-designed spring devices. His experiments took into account the effects of both vehicles motion and atmospheric turbulence through simulated boundary layer, determining aerodynamic forces and moments of the train involved. Baker s experiments started an attempt of moving model test on the simulated ground. However, currently, there are very few similar studies, not to mention systemic analyses for aerodynamic interaction of vehicle motion and its supporting structure, the bridge deck. This
is mainly attributed to the so small model scale and the so short valid test time for the model moving through the tunnel section owing to the limited wind tunnel size, which leads to the very considerable difficulties about these experiments conducting and uncertainties of the results concerned. Fortunately, a newly built industry wind tunnel XNJD-3 at Southwest Jiaotong University, P.R. China facilitates the moving model test. The tunnel spans up to 4 m at the test section with a height of 4.5m. Take advantage of the experimental condition, a new experimental setup is presented with considering the movement of vehicles to measure the aerodynamic forces of a moving vehicle-bridge system under the cross wind. A series of advanced and detailed investigation on the vehicles motion effects is conducted in this paper. Moving Vehicle Model and Testing Methods A set of devices were conceived and manufactured. Generally, the devices constitute 4 cooperating component systems: the models of vehicles and bridge deck, the supporting frame, the launch & recovery system, and the measurement system. To get an overall understanding about the experimental system, see the general arrangement of the moving vehicle model as shown in Fig.1. Motor Guide rail Train models Bear frame 6-Component balance Bridge deck Guide rail 5-Component balance Fig. 1: Overall configuration of the experiment system The model system includes bridge deck models and train vehicle models. According to the span of test section of the wind tunnel, the model scale of bridge deck and train is determined as 1/45. The deck is simulated by rigid section model which made of hard wood. In order to explore the change of flow characteristics with relative locations of the train on the deck, the tracks may be installed on either the side or middle of the deck, which are marked by the letters of A, B, C, D respectively, as shown in Fig.. The overall length of the deck model is 13.07m long. The train models are made of hard wood, too, according to the geometrical scale, the head car, the intermediate car and tail car represent the locomotive, the carriage and the tail of the train respectively, as shown in Fig.3. The wheels were not reproduced, but the clearance between the vehicle bottom and the tracks is simulated. The support frame consists Flow direction of 4 spans of steel continuous D C B A beam bellow which a notch shape Vehicle Vehicle Vehicle Vehicle guide rail was fixed. Fig.4 shows a sketchy layout of the models with the support frame. It is very Bridge important to ensure the train models get a smooth motion; therefore, the guide rail was Fig. : The location of the tracks manufactured as smoothing as possible, since a sliding block that links the train vehicles by a vertical rigid rod needs to travel along the guide rail. Besides, the friction and the rail irregularities are reduced to the minimum and the deflection of the steel continuous beam is checked up by FEM analysis. The overall length of the guide rail is 18m, which identified with 3 functional units. The first 4.5m
is used to accelerate for the train, 4.5-13.5m serves as uniform motion segment keeping the train motion at a certain speed, and 13.5-18m is the deceleration segment as buffer distance. The train models can be droved by a servo motor (refer to Fig.1) fixed at the end of the guiding rail, which is specially designed to drag the models at a speed varied from 0 to 10m/s with a specific acceleration and the corresponding duration time. The motor serves fully automatically with a programed parameter set; consequently the models would have a desired resultant journey. With respect to the recovery process, it is supposed for the models to skid to halt with a buffer. The motor power would be cut off automatically as soon as the models reach the last 4.5m journey, thus the models skid on, decelerate gradually and halt finally owing to extension reaction from the four elastic rubber bands along the last 4.5m journey. A six-component balance iss fixed on the train carriage to measure its aerodynamic forces and moments at 6 directions (refer to Fig.1 and Fig.4): the drag force, side force, lift force, rolling moment, pitching moment and yawing moment, respectively. Besides, the bridge measurement section with 0.5m length (refer to Fig.4), which is discontinuous but close to the adjacent decks, are measured with another five-component balance (refer to Fig.1 and Fig.4), but only 3 components are concerned: the side force, lift force and longitudinal moment. The force signal of balances should be amplified first prior to being collected by an A/D board. With a short sampling period of 100μs, thedynamic and instantaneous force signals that associated to the models are recorded. In order to obtain adequate data sampling resolution, choose 909Hz as the sampling frequency for the measurement. T ransition section Measurement section Transition section 6-Component balance Fig. 3: Train vehicle models used in case study Measurement section Driving direction Bear frame 6-Component balance Vehicle G liding block Bridge deck 5-Component balance Fig. 4: Local view of the mock-up Aerodynamic Coefficients The aerodynamic coefficients of vehicles and bridge deck could be defined with the same formulas as follows: Drag Coefficient (side force) CH = FH ( 1 ρv HL) (1) Lift coefficient CV = FV ( 1 ρv BL ) () Moment coefficient CM = FM ( 1 ρv B L) (3) In the above equations, where 1 ρ V is the dynamic pressure of wind; H, B and L are the height, width and the length of the section model respectively. F H, F V and F M are the drag force, lift force and moment, C H, CV andc M are the drag coefficient, lift coefficient and the moment coefficient, respectively.
Data Processing In the process of vehicle motion, due to the influence of the vibration of the vehicle and the bidirectional transformation linker, the inertia force caused by the vehicle speed variation, the irregularity of the guide rail, and etc, the testing signals of the aerodynamic forces will be obviously interference when the vehicle runs at a relatively high velocity, therefore, there will be some error in the calculation of the aerodynamic forces and moments if using the original signal. The time history curves of the vehicle drag coefficient at the cross wind speed of 8 m/s and the train speed of 6 m/s is shown in Fig.5 (a). It appears an obvious fluctuation with the interference signals. According to the spectrum analysis of the time history curve, it is shown that there are mainly two frequency bands. By studying the vehicle and bidirectional transformation linker,finding that the two frequency bands just correspond with the nature frequency of the vehicle and the bidirectional transformation linker, with the frequency ranges of 6~8Hz and 1~14Hz. Based on investigating the frequency spectrums of many signal curves, it is determined to adopt the low-pass filter with the cut-off frequency of 4 Hz to handle the original signal curves. The time history curve of the vehicle drag coefficient has been filtered as shown in Fig.5 (b). It is shown that the interference signals are obvious reduced, but there is still some fluctuation in the stationary section where vehicles run at uniform speed. According to the accelerating duration time which is preset, one entire time history curve could be divided into the acceleration section, the stationary section and the deceleration section after the aerodynamic signal are filtered(refer to Fig.5 (b)), and then take the average value of the stationary section as the representative value. In order to increase the stability of the data, each test condition is carried out for five times, and take the average value of the five time tests as the final representative value for each test case. 6 4 Vehicle drag coefficient 4 0 - Vehicle drag coefficient 3 1 Acceleration section Stationary section Deceleration section 0-4 -6 0 1 3 4 5 6 Time (s) Test Results and Analysis -1 0 1 3 4 5 6 Time (s) (a) (b) Fig.5 Time-history curve of vehicle drag coefficient (a) Before filtering (b) After filtering Effects of the Vehicle Speed When the train runs along the A track, in order to investigate the effects of the vehicle speed, train is driven at three different vehicle speeds, 4m/s, 6m/s, 8m/s, respectively. The vehicle speed just refers to the uniform speed at stationary section. The curves of the vehicle drag coefficient at different vehicle speed as shown in Fig.6. The curves of the vehicle drag coefficient become fluctuant when the velocity of the vehicle are increased. And the change of the vehicle drag coefficient is irregular at the different vehicle speeds, which
indicates that the vehicle drag coefficient is insensitive to the vehicle speed, so are the vehicle lift coefficient and moment coefficient. The possible reason is that the airflow field around the section models of vehicles and bridge deck becomes stable due to the existence of the transition section vehicles which is on the both sides of the measurement section vehicle (refer to Fig.3); in addition, the cross section of the vehicle is so regular that the influence of three-dimensional airflow field induced by the vehicle movement should be limited, therefore, the aerodynamic forces of vehicles have less change. The aerodynamic coefficients corresponding to the resultant wind velocity basically follow the Cosine rule. The Fig.7~Fig.9 show the aerodynamic coefficients of the bridge deck. The aerodynamic coefficients have less change at different vehicle speeds when the vehicle has not get to the measurement section of the bridge. While the location of vehicle is just on the measurement section of the bridge (at the range of 4.5m~5.5m of guiding rail), the drag coefficient of the bridge becomes smaller, but there is a little difference at the different vehicle speed. The lift coefficient and the moment coefficient of the bridge are sensitive to the vehicle speed, the absolute value of the lift coefficient become smaller while the absolute value of the moment coefficient increase slightly when the velocity of the vehicle increases. For the vehicle running along the bridge at a certain velocity under the cross wind, its real wind direction is a vector combining the incoming wind speed and vehicle speed which along the bridge axis. In the case of different wind speed (6m/s 8m/s 10m/s) and different vehicle speed (4m/s 6m/s 8m/s), the aerodynamic coefficients of the vehicle and bridge change with the resultant wind direction as shown in Fig10 and Fig11, both the aerodynamic coefficient of the vehicle and bridge are insensitive to the yaw angle of the wind. Fig.1 and Fig.13 show the aerodynamic coefficients of vehicle and bridge change with the resultant wind velocity, which reflects the effects of Reynolds number in some degree. the Reynolds number has some influence on the aerodynamic coefficients of vehicle and bridge, but not obvious. Effects of the vehicle location on different tracks When the vehicle is on the A, B, C and D track respectively, the aerodynamic coefficients of the vehicle and bridge are tested as shown in Fig.14 and Fig.15. During the vehicle changes from the track of A to D, the drag coefficient of the vehicle decreases gradually, while the drag coefficient of the bridge increases gradually. Vehicle drag coefficients.0 4m/s 6m/s 8m/s Bridge drag coefficients.0 4m/s 6m/s 8m/s Fig.6 Vehicle drag coefficients Distance (s) Fig.7 Bridge drag coefficients
0.10-8 Bridge lift coefficients 5 0-5 -0.10-0.15-0 -5 4m/s 6m/s 8m/s -0.30 1.5 Fig.8 Vehicle drag coefficients Bridge moment coefficients -0.10-0.1-0.14-0.16-0.18-0 4m/s 6m/s 8m/s Fig.9 Bridge drag coefficients Vehicle aerodynamic coefficients 0.9 0.3-0.3 Bridge aerodynamic coefficients 36 40 44 48 5 56 60 64 68-36 40 44 48 5 56 60 64 68 7 Yaw angle β ( ) Fig.10 Vehicle drag coefficients Yaw angleβ ( ) Fig.11 Bridge drag coefficients 1.5 Vehicle aerodynamic coefficients 0.9 0.3-0.3 Bridge aerodynamic coefficients 6 7 8 9 10 11 1 13 14 Resultant wind velocity(m/s) Fig.1 Vehicle drag coefficients - 7 8 9 10 11 1 13 Resultant wind velocity( m/s) Fig.13 Bridge drag coefficients Vehicle aerodynamic coefficients 1.50 5 0 0.75 0.50 5 0-5 -0.50 A B C D Location of tracks Fig.14 Vehicle aerodynamic coefficients Bridge aerodynamic coefficients - A B C D Location of tracks Fig.15 Bridge aerodynamic coefficients
Effects of the Head Car and Tail Car The high-speed train usually adopts the streamline shape for the head car and tail car. In order to investigate the aerodynamic characteristics of the intermediate car, therefore, installing the transition section (head car and tail car) on the both sides of the intermediate car. To investigate the effects of the head car and the tail car, carrying out the contrast test of single vehicle (the intermediate car) and three vehiles (head car, intermediate car and tail car) at the track A and D respectively. The vehicle drag coefficients at the wind speed of 8m/s and vehicle speed of 6m/s are shown in the Table 1. Regardless of the vehicle at the track of A or D, the drag coefficient at three cars case is smaller than at single car case. It indicates that the head car and the tail car have obvious effect on the vehicle drag coefficient. Fig.16~ Fig.18 show the time history curves of drag coefficient at single car case and three cars case at different vehicle speeds. The vehicle drag coefficient increases slightly with the vehicle speed despite of the single car case or three cars case. Besides, for single car case, vehicle speed has more significant effect on the vehicle drag coefficient. Table 1 Vehicle aerodynamic coefficient vehicle Vehicle aerodynamic coefficient Location of tracks condition Single car 11 0.7507-0.1677 A Three cars 1.1849 33-064 Single car 0.913 763-846 Vehicle drag coefficients.0 Single car Three cars D Three cars 0.7589 0.3887 039 Fig.16 Vehicle aerodynamic coefficients..0 Vehicle drag coefficients.0 Single car Three cars Fig.17 Vehicle aerodynamic coefficients Vehicle drag coefficients Single car Three cars Fig.18 Bridge aerodynamic coefficients Conclusions By the study on the moving vehicle model for aerodynamic forces of vehicle-bridge systems, come to some conclusions as follows:
1. Considering the vehicle motion and the interaction between vehicles and bridge, a moving vehicle model system in the wind tunnel test is developed. The system could be fit well for testing the aerodynamic forces of vehicles and bridge deck with different vehicle speeds, wind speeds and different combination forms of vehicle and bridge.. For the case of the three cars, the aerodynamic coefficients corresponding the resultant wind speed and wind direction basically obey the Cosine rule. When the location of vehicle is just on the measurement section of the bridge, the absolute value of the lift coefficient becomes smaller while the absolute value of the moment coefficient increases slightly when the velocity of the vehicle. 3. With the track moved from the windward side track to the leeward side track, the drag coefficient of the vehicle decrease gradually, while the drag coefficient of the bridge increases gradually. It indicates that the relative locations between the vehicle and the deck have an obvious effect on the aerodynamic characteristics of the vehicle-bridge system. 4. The drag coefficients of intermediate vehicle at three cars case are smaller than at the single car case, which indicates that the head car and the tail car have obvious effect on the vehicle drag coefficient. Besides, for single car case, vehicle speed has more significant effect on the vehicle drag coefficient. The moving vehicle model is installed in the wind tunnel. Although the tunnel span is up to.5 m at the test section, and the guiding rail is up to 18m long, the total duration time of the stationary section of vehicle movement is very short except the acceleration section and deceleration section, so is the duration time of the vehicle running through the measurement section of the bridge. Therefore, it will lead to some fluctuation in the test data. The problem of short valid test time and signal noise need to be further improved. References [1] Yongle Li, Shizhong Qiang, Haili Liao, and Y.L. Xu(005), Dynamics of Wind - Rail vehicle - Bridge Systems, Journal of Wind Engineering and Industrial Aerodynamics, 93(005), 483-507. [] R.K. Cooper (1993), Bluff-Body Aerodynamics as Applied to Vehicles, Journal of Wind Engineering and Industrial Aerodynamics, 49(1993), 1-. [3] C.J. Baker (1986), Train Aerodynamic Forces and Moments from Moving Model Experiments, Journal of Wind Engineering and Industrial Aerodynamics, 4(1986), 7-51. [4] Stephane Sanquer, Christian Barre, Marc Dufresne de Virel and Louis-Marie Cleon (004), Effect of cross winds on high-speed trains: development of new experimental methodology, Journal of Wind Engineering and Industrial Aerodynamics, 9(004), 535-545.