surface flow and wake structure of a …bbaa6.mecc.polimi.it/uploads/validati/tr04.pdfvortex wrapped...

1

BBAA VI International Colloquium on: Bluff Bodies Aerodynamics & Applications

Milano, Italy, July, 20-24 2008

SURFACE FLOW AND WAKE STRUCTURE OF A REAR VIEW MIRROR OF THE PASSENGER CAR

Jeong-Hyun Kim∗, Yong Oun Han∗, Min Hwa Lee∗, In Ho Hwang†, and Ui Hun Jung†

∗School of Mechanical Engineering Yeungnam University, Gyeongsan, Gyeongbuk 712-749, Korea

e-mail: [email protected], [email protected], [email protected]

† Technical Research Laboratory SL Corporation, Gyeongsan, Gyeongbuk 712-837, Korea

e-mails: [email protected], [email protected]

Keywords: Rear View Mirror, Wake Flow, Surface Pressure, Hot Wire Anemometry, Oil Film Visualization, Vortex Envelop

Abstract. In order to investigate the wake structure and the vortex body frame interaction near the real scale rear view mirror of a passenger car, the velocity vector fields in wake and pressure distribution over the mirror skin have been measured by use of the hot-wire anem-ometry and the pressure scanning system, respectively in the blow down wind tunnel at the Reynolds number of 5105.2 × . The boundary layer flow over the mirror housing and the mir-ror surface was visualized by the oil film technology as well. The vertical velocity vector fields to the main stream in the mirror wake showed that within the half distance of the mirror span the recirculation zone appeared and the intrinsic vortex was shed with the frequency of 19.6 Hz, and also showed that a conical envelop of the vortex sheet was developed with its center trailing towards the body frame. The minimum pressure zone on the mirror surface appeared near the upper tip corner, which seemed to be the origin of the vortex envelop. The separation focus was found on the upper tip face of the mirror housing and a surface separa-tion line was also distributed over the edge line of the mirror housing in the visualization re-sults. Such flow characteristics of the single mirror were compared to those of the mounted mirror to the real scale car as well.

Jeong-Hyun Kim, Yong Oun Han, Min Hwa Lee, In Ho Hwang and Ui Hun Jung

2

1 INTRODUCTION

Streamlined body design in a passenger car helps reducing the aerodynamic drag and even-tually improves the engine mileage. On the contrary, accessories attached on the body skin of a car cause the unfavorable aerodynamic effects. The rear view mirror either side of a driver or an accompanier is one of unfavorable aerodynamic examples. In order to obtain the rear sight, unfortunately the mirror does not pay only the aerodynamic penalty which increases body form drag, but also causes the acoustic noise and the mirror fluctuation to the cabin crews. While the aerodynamic body styling of the passenger car has been concerned with con-siderable efforts, rather ignored have been defects caused by such accessory, the rear view mirror. The main stream meets a side flow which has the flow direction tangent to the wind shield surface near the A-pillar. And a conical vortex sheet is generated along the pillar and merges into the main stream. Therefore, very complicate flow pattern appears by combining these flow patterns near the driver side window. Moreover, since the side mirror is mounted on the driver door near hinge, the wake flow behind this obstacle become much complicated [1] [2]. When the vehicle has a negative yaw angle, defined as the rear view mirror is in a lee-ward situation relative to the vehicle, the A-pillar vortex changes considerably, and large gra-dients in mean velocity and turbulent intensity are apparent. It gives the worst circumstance to induce aerodynamic noise easily [3].

The rear view mirror as one of blunt obstacle generates an intrinsic periodic wake pattern which composes several vortex wraps depending on its geometry. Such a spatially periodic wake causes an intrinsic acoustic noise as well as a flow induced vibration making the mirror surface tremble resulting unclear vision to the driver. Worse thing happens when such peri-odic vortex wraps breaks down making vortex body frame interaction noise (VBIN). There-fore, it could be a good strategy to make the vortex trail not meet the door panel as possible by path control [4] [5].

One of the intrinsic noises was experienced in the mirror mount. When the transition to tur-bulent boundary layer over the mirror housing is not complete, pressure fluctuations can cause rapidly changing flow patterns generating whistle sound to the driver. This noise can be de-tected near the gap between the mirror housing and the mirror mount. It is known that heavy surface graining, bumps and grooves on the mirror surface or the mount are effective to re-duce whistle noise by promoting the turbulent boundary layer easily on these devices [6].

Another intrinsic noise is also generated by the mount device. There are two basic configu-ration of rear view mirror, door mounted mirror and sail-panel mounted mirror, respectively. Intrusion of the mirror support in sail-panel mounted mirror deflects approaching flow along the body and eventually forms a disk-like vortex underneath the mirror housing. It causes a strong vortex shedding, which stimulates instability within the body frame shear layer [7].

There have been observed typical strategies to reduce such the traditional problems of the rear view side mirror as noise and mirror surface vibration. Geometric shape of the mirror housing should be primarily concerned to reduce the strength of intrinsic shedding vortex as possible, such as elimination of sharp edges on all corner, tube typed extension of the mirror housing which keeps away the flow recirculation [8]. As another strategy, the path control of vortex wrapped sheet could be used to put away from the door window or panel by changing the mounting location in order to avoid VBIN.

In this paper, the wake structure of the rear view mirror and the pressure profile over both the mirror housing and the mirror surface will be measured directly to understand the flow structure and its interference to the body of the passenger car with the real scale side mirror. Also the path control of the vortex trail generated by the rear view mirror will be observed based on data obtained.


3

2 DESCRIPTION OF THE EXPERIMENT

A blow down type wind tunnel which has the dimension of 0.27.09.0 ×× [m] was used to make parallel flow around a real scale rear view mirror only. To observe the body frame in-teraction of the mirror wake, one of the side walls of the wind tunnel test section was modi-fied to fit to the driver side skin of the real scale car by locating the mirror center mounted on it at the test section center of the wind tunnel. The real scale mirror of New Pride manufac-tured by KIA motor Korea was chosen for the experiments of both the single model which used the real scale mirror and mounting zig only and the real scale model which mounted on the real scale car. The experimental set up of the mirror in the test section is shown in figure 1.

(a) (b)

Figure 1: Photographs of the rear view mirror installed in the test section of the wind tunnels for (a) the single model and (b) the mounted model, respectively.

For convenient experiment, a pressure model was fabricated by using the composite mate-

rial under RP process. Figure 2 shows the real scale pressure model. Static pressure holes were made with 0.8 [mm] diameter at 135 points over the housing surface and at 137 points on the mirror surface, respectively. The space between two holes on the mirror surface was 10[mm] horizontally and vertically. Each hole was connected by tubing to the pressure trans-ducer with the scanni-valves. The analog signal measured by pressure transducers was digi-tized by 16 bit A/D converter and it was stored in personal computer.

Figure 2 : The real scale pressure model

Tip

Root


4

To obtain the velocity vector field of the wake, the x-configuration hot wire was utilized. Wake velocity fields were measured at the several vertical sections to the main flow direction. In recirculation zone of the wake behind the mirror the two dimensional LDV system used to get the negative velocity.

To visualize the movement of the surface vortex on the boundary layer over the mirror housing surface including the mounting device and the mirror surface, an oil film visualiza-tion technique was utilized. This oil film technique was also used to observe the body frame interaction of the vortex wrap generated by the mirror by tracing where the oil left from the mirror arrives over the body frame.

3 RESULTS AND DISCUSSION

3.1 Pressure distribution Static pressure profiles over the surface of the mirror housing were obtained along the geo-

metric latitude lines; middle span which is most convex, upper span where the housing has almost right turn to the main flow direction, and lower span where the housing has mild turn to the main flow, and three vertical lines. Figure 3 shows representation of data points over the surface of mirror housing. Plots of pressure distribution for the speed of 25[m/s] are shown in the figure 4. Pressure distribution has the peak near the mounting root even though the whole housing front was inclined as 30 degrees to the main flow. Vertical profile of the total pressure at the middle line shows mild peak at the little upward position which matches well to the housing contour. These profiles were not changed even when the single mirror model was mounted to the real car.

Figure 3 : Representation of data points over the surface of mirror housing.

(a) (b)

Figure 4 : Pressure distributions over the surface of mirror housing measured on (a) the middle span line and (b) the middle vertical line, respectively.


5

Considerable static pressure measurement was carried out over the whole mirror surface using 137 static holes. Figure 5 shows static pressure distributions on the mirror surfaces both single and mounted models, respectively. It is important where the minimum static pressure locates because the origin of the vortex sheet wrap may be considered to locate at the mini-mum point. The minimum pressure appeared near upper root corner for the single model, but it was shown up lower tip region for the case of the mirror mounted to the real car. Based on this observation, this mirror was optimized geometrically to induce the vortex center located far from the car body, which is the one of main objectives to control the path of the vortex wrap.

(a) (b)

Figure 5 : Pressure distributions on the mirror surface for (a) the single model and (b) the mounted model.

3.2 Wake structure To find the wake structure behind the rear view mirror, the velocity vector fields were ob-

tained on the vertical planes; 0.5d, 1.0d, 1.5d and 2.0d (d is based on the span length of the mirror) downstream to the main flow direction by using the x-configuration hot wire and 2-D LDV system. The mirror surface was inclined with 30 degrees for the driver to retain the rear view to the main direction, note that the test sections were not arranged as parallel to the mir-ror surface.

Figure 6 : Contour plot of the axial velocity measured by LDV system for the single model with the inflow ve-locity of 25 [m/s].


6

From the test results until downstream at 1.0d there appeared recirculation zone in the axial flow velocity components which was observed by the smoke visualization. LDV result shows the recirculation area in the U-contours as shown in the figure 6.

Figure 7 shows the two-dimensional time-averaged vector fields for the single model. The vortex pair was found to form at a mounting zig and then moved towards tip region in down-stream. Also, vortex center generated on the root region of the mirror surface moved in the same direction during this process. The vertical velocity fields show the strong upwash veloc-ity behind the neck of the mirror.

(a) (b)

(c) (d)

Figure 7 : Vertical velocity vector fields for the single model measured at the vertical section of (a) 0.5d (b) 1.0d (c) 1.5d and (d) 2.0d downstream to the main flow with the inflow velocity of 25[m/s].

The corresponding axial vorticity fields at various downstream locations are plotted as

shown in figure 8. The strong vorticity was shown behind upper root corner of the mirror in-cluding vortex pair induced from a mounting zig. The location of the maximum axial vorticity gradually moved towards tip region in downstream. These results show that influence of con-


7

traction between side face of the mirror and mounting zig was dominant in wake flow for the single model.

(a) (b)

(c) (d)

Figure 8 : Vorticity contour plots for the single model at the vertical section of (a) 0.5d (b) 1.0d (c) 1.5d and (d) 2.0d downstream to the main flow with the inflow velocity of 25[m/s].

Wake survey results for the mounted model from the x-configuration hot-wire anemometry

are shown in figure 9 and 10. Figure 9 shows the time-averaged vector fields for the mounted mirror on the vertical planes; 0.5d and 2d, respectively. Compared with the single model, dif-ferent wake structure was found for each section. A vortex center was formed behind tip re-gion of the mirror surface and moved more adjacent to the body skin creating a large rotating vortex in the downstream of the model. With the inflow velocity increase the vortex sheet wrap was found to come closer to the body frame which may result in the increase of noise. The vertical velocity vector fields show vortex sheet wraps in downstream which does not only explain how frequently the periodic conical vortex sheet appears in the downstream di-rection, but also shows where the path of the vortex center establishes.

The corresponding vorticity plots are shown in figure 10. Strong vorticity was shown be-hind the housing edge of the mirror at the vertical section of 0.5d downstream and it was ob-served near the body frame at the vertical section of 2d downstream because of a counter rotating vortex.


8

(a) (b)

(c) (d)

Figure 9 : Vertical velocity vector fields for the mounted model measured at the vertical section of (a) 0.5d and (b) 2.0d downstream to the main flow with the speed of 10[m/s] and (c) 0.5d and (d) 2.0d downstream to the

main flow with the speed of 25[m/s].

3.3 Surface flow visualization

To look into details for such vortex wrap control on the boundary layer, streamline over both the housing and mirror surfaces must be observed. For this purpose the oil film technique was utilized for dynamic visualization over the mirror skin in both the single model and the mounted model with the inflow velocity of 25[m/sec]. Figure 11 shows a flow pattern of oil-film surface for the single mirror. The nodal attachment point was observed near the root of the housing surface because rear view mirror had a yaw angle to a mounting zig. Surface flow spread out over the housing surface from that point and formed a separation line at a little bit ahead the edge line of the housing, which explains the lower pressure zone was established before the apex line in the boundary layer flow. A huge separation focus appeared at the upper corner on the mirror housing near the tip region. On the mirror surface, the surface streamline was tended to make circle in clockwise direction. This implies that the lower pressure zone


9

appeared near the root corner. Therefore, oil film visualization may be used to expect the pressure distribution on the mirror surface in the future work.

(a) (b)

(c) (d)

Figure 10 : Vorticity contour plots for the mounted model at the vertical section of (a) 0.5d and (b) 2.0d down-stream to the main flow with the speed of 10[m/s] and (c) 0.5d and (d) 2.0d downstream to the main flow with

the speed of 25[m/s].

Figure 11 : Surface flow visualizations for the single model.


10

Compared with results of the single model, the mounted mirror model showed difference for the location of the separation and the minimum pressure zone. Figure 12 shows that the separation focus was become weaker and drifted toward to the tip face so that it implies the resultant strength of the vortex wrap generated by the mirror housing becomes favorable for the vortex control. On the mirror surface, the surface streamline was tended to make circle in counter clock wise direction, implying the lower pressure zone appeared near the tip corner and therefore, forming the vortex center at the upper tip corner on the flat mirror surface. In addition, we found that flow separated from the lower tip face arrived over the body frame. These results show the consistency to the vertical velocity vector measurements.

Figure 12 : Surface flow visualizations for the mounted model.

3.4 Shedding frequency estimation To estimate the shedding frequency in the wake of the mirror, a commercial code was used

to calculate the periodic characteristics generated by the Karman vortex. Figure 13 shows pe-riodic characteristics of the static pressure on the mirror surface in the condition on the inflow velocity was 25[m/sec]. As shown in the figure 13, the shedding frequency of the alternating vortices was seemed to be about 19[Hz], calculated by Fluent.

Generally, it is known that the frequency of the alternating vortices is proportional to in-flow velocity for the given Strouhal number. Therefore, in the realistic flow velocity range; 25~40[m/s], corresponded to be 90~144[km/h], the shedding frequency aroused by such scale mirror seems to be ranged in 20~30[Hz]. This result gives the useful information to find the acoustic noise source generated by the mirror housing in VBIN problem.

Figure 13 : Time history of static pressure fluctuation on the mirror surface


11

3.5 Comparisons

Through various measurements the flow structure around rear view mirror has been un-veiled. Around the rear view mirror the intrinsic vortex is shedding across the vertical mirror surface and the resultant vortex wrap in the wake is also formed. The geometric shape of the mirror housing must be designed to reduce the vortex strength as well as the vortex path should be away from the body as possible. In this experiment it was turned out the trail of the vortex sheet wrap generated by such a bluff body tends to meet the body skin of a car, arous-ing the vortex-body frame interaction. The co-ander effect seems to accelerate such interfer-ence. This requires the path control for the vortex drift should consider rigorously more than expected.

It was also found that the boundary layer structure of the single model shows many differ-ences from those of the mounted mirror. This implies that there exists any possibility to make a misleading for the flow structure by observing only those of the single model when not looking into the interaction. The pressure distribution and the visualization show their differ-ences clearly.

4 SUMMARY

Observing the flow structure around the rear view side mirror of a passenger car by the hot wire anemometry, pressure measurement and oil film visualization, it can be summarized as follows ;

The flow around the mirror model mounted on the real car shows much different from those of the single model specially on the position of the minimum pressure location on the mirror surface, and the separation focus as well; the minimum pressure appeared at the lower tip region for the mounted model while it was at the upper root region for the single model, and the weaker separation focus of the mounted was located on the vertical tip face while that of the single model was on the top corner, respectively.

With the inflow velocity increase the vortex sheet wrap comes closer to the body frame which caused the interaction and resulted in the noise increase. The trail of the vortex center shows more adjacent to the body skin than expected, which implies that the co-ander effect was prevailed.

Based on the vertical velocity vector fields the vortex strength was attenuated as much as an half of that of the 0.5d plane within 2.0d. It turned out that the intrinsic vortex shedding was not much affected to the body frame interference, but to the vibration of mirror surface, while the vortex sheet wrap affects strongly the interaction noise in downstream.

ACKNOWLEDGEMENTS

This work was supprted by SL Corporation, Korea. Their supports are greatly acknowl-edged.

REFERENCES [1] W. Hucho, Aerodynamics of Road Vehicles 4th edition, Society of Automotive Engi-

neers, Inc., 1998.

[2] K. Ono, R. Himeno, T. Fukushima, Prediction of wind noise radiated from passenger cars and its evaluation based on auralization, Journal of Wind Engineering and Indus-trial Aerodynamics, 81, 403-419, 1999.


12

[3] S. Watkins, On The Causes of Image Blurring in External Rear View Mirrors, SAE Pa-pers 2004-01-1309, SP-1874, Detroit, Michigan, USA, March 2004.

[4] Y. O. Han. Research for the Low Noise Rear View Side Mirror of a Passenger Car, Report of Aerodynamic & Turbulence Research Laboratory, 2007-05, May 2007.

[5] Y. O. Han, J. H. Kim, I. H. Hwang, J. B. Seo, B. H. Lim and U. H. Jung. Wake Flow Characteristics around the Side Mirror of a Passenger Car, KSME Conference, BEXCO Building, Busan, Korea, 30 May-1 June, 2007.

[6] T. Lounsberry, M. Gleason and M. Puskarz, Laminar Flow Whistle on a Vehicle Side Mirror, SAE Paper 2007-01-1549, V 116-6, Detroit, Michigan, USA, April 2007.

[7] O. Dolek, G. Ozkan and I. B. Ozdemir, Structures of flow around a full scale side mir-ror of a car with relevance to aerodynamic noise, Journal of Automobile Engineering, Volume 218 Number 10, 1085-1097, 2004.

[8] R. Jaitlee, F. Alam and S. Watkins, Pressure Fluctuations on Automotive Rear View Mirrors, SAE Paper 2007-01-0899, SP-2066, Detroit, Michigan, USA, April 2007.

surface flow and wake structure of a …bbaa6.mecc.polimi.it/uploads/validati/tr04.pdfvortex wrapped...

Documents