10th joint ERCOFTAC (SIG-15) -IAHR-QNET/CFD Workshop on Refined Turbulence Modelling

 
 
 

CASE 9.4
Flow around a simplified car body (Ahmed body)

Database: Experiments by S. Becker, H. Lienhart, C. Stoots


Fig. 1: Illustrative views of the flow around the Ahmed body (Exp. Becker, Lienhart, Stoots)









Flow description:

Measurements of Becker, Lienhart, Stoots (2000) have been performed within the European project MOVA (Models for Vehicle Aerodynamics 1998-2001).   The authors kindly provided their measurements for the preceding workshop held in Darmstadt in October 2001 and for the present workshop.

The experiments were performed in the LSTM low-speed wind-tunnel. Cross-section of the tunnel is 1.87x1.4 m2 (width x height).  Fig. 2  illustrates the investigated Ahmed Body.
 
 

Fig. 2:  Dimensions of the Ahmed Body (from Ahmed et al.,1984)

The diameter of the Stilts is 30 mm.  The distance between the body and the plate representing the ground (not shown in Fig. 2) is 50 mm.
Two slant angles of 25° and 35° are considered.  The slant length is the same for both angles (222 mm).

The front part of Ahmed Body is given with all details in the file Ahmed_Front_Geo.txt (465k).
 

Flow Parameters:

Kinematic viscosity of air: 15x10-6 m2 s-1
Bulk velocity: Ub=40 m s-1
Height of the body: h=288 mm
Reynolds number based on the height of the body: Re=768,000
 

Database:

The origin of the coordinate system is located at the end of Ahmed Body (x=0 end of the model, y=0 symmetry plane, z=0 ground plane).

LDA measurements (first, second and third moments) are available in the following planes (coordinates are given in mm):

25° case:
y=0 (whole flow)
y=0 (focus on the slant part)
y=100
y=180
y=195
y=-195

x=-178
x=-138
x=-88
x=-38
x=0
x=80
x=200
x=500

35° case:
y=0 (whole flow)
y=0 (focus on the slant part)
y=100
y=180

x=-88
x=0
x=80
x=200
x=500

For both slant angles, boundary layer profiles are available in the symmetry plane at the locations (only u-w-components):

x=-243; -223; -203; -183; -163; -143; -123; -103; -83; -63; -43; -23; -3
 

The wall pressure distribution at the rear of the body is also available for both slant angles.

Hot wire measurements of the boundary layers on the central part of the body are also available at the locations described in Fig. 3.  They were performed in the 25° case, but are probably not very sensitive to the slant angle of the rear part.


Fig. 3: Description of the locations of the boundary layer measurements on the central part of the body
(Click here to download a postscript version of the image (166k))

Download the database:

        - LDA Measurements (planes and boundary layer in the symmetry plane)
                25° case:     Excel format (2.3M) or ascii format (912k)
                35° case:     Excel format (1.6M) or ascii format (626k)

        - Pressure measurements on the rear of the body
                25° case:     Pressure coefficient (17k)
                35° case:     Pressure coefficient (18k)

        - Hot wire measurements of the boundary layer on the central part of the body (coordinates are given in m)
                           Boundary layer (22k)
 

Instructions for calculations:

Profiles for velocity and Reynolds stresses are provided at three y-locations (y=0, 200, 400 mm) in the plane x=-1444 mm (i.e., 400 mm upstream of the front end of the body).  This profiles can be used to generate the inlet boundary conditions.  Measured inlet viscosity ratio is approximately 10 and this value should be used to recalculate a dissipation rate.  Calculations are not sensitive on small variations of this ratio.

Download the Inlet profiles (14k)










As can be seen in this file, the velocity field depends on the location on the y-axis, which means that the flow already "feels" the presence of the body.  Therefore, it can be preferable to apply the inlet boundary conditions far upstream (several body length is recommended), and to use a constant velocity U=40 m s-1 in the inlet plane or boundary layer profiles.
In that case, it is must be checked that the boundary layer thickness at x=-1444 mm is of the same order as in the measured profiles, i.e.,=30 mm.
 

It is recommended that the domain at least extends to five body lengths behind Ahmed Body.

Other dimensions are defined by the dimensions of the channel: width (y-direction) = 1.87 m; height (z-direction) =1.4 m (but the test section is 3/4 open).

All walls should be treated as fixed walls (including the ground).

The stilts can be omitted (please give this information is the headings of the result files).

In the case that transient calculations are performed, a full body should be calculated.  Otherwise, a symmetry boundary condition can be employed through the symmetry plane (y=0).
 

References:

H. Lienhart, C. Stoots, S. Becker (1.3M), Flow and Turbulence Structures in the Wake of a Simplified Car Model (Ahmed Model), DGLR Fach Symp. der AG STAB, Stuttgart University, 15-17 Nov., 2000
A. Spohn, P. Gilliéron(259k), Flow Separations Generated by a Simplified Geometry of an Automotive Vehicle, IUTAM Symp., 2002
P. Gilliéron, F. Chometon, Modelling of Stationary Three-Dimensional Separated Air Flows around an Ahmed Reference Model, ESAIM Symp., 1999
B. Basara, Numerical simulation of turbulent wakes around a vehicle, FEDSM 99-7324, 1999
S.R. Ahmed, G. Ramm, G. Faltin, Some salient features of the time-averaged ground vehicle wake, SAE Paper 840300, 1984
 

Required results:

Participants are invited to submits the following files (coordinates in mm, velocities in m s-1, Reynolds stresses in m2 s-2):

25° case:
- Variables in measurement planes:
 

file
plane
window
Ahmed_25_y=0
y=0
-1442<x<638; 0<z<738
Ahmed_25_y=100
y=100
-243<x<87; 28<z<388
Ahmed_25_y=180
y=180
-243<x<87; 28<z<388
Ahmed_25_y=195
y=195
-243<x<87; 28<z<388
Ahmed_25_x=-178
x=-178
-250<y<250; 28<z<538
Ahmed_25_x=-138
x=-138
-250<y<250; 28<z<538
Ahmed_25_x=-88
x=-88
-250<y<250; 28<z<538
Ahmed_25_x=-38
x=-38
-250<y<250; 28<z<538
Ahmed_25_x=0
x=0
-250<y<250; 28<z<538
Ahmed_25_x=80
x=80
-400<y<400; 28<z<538
Ahmed_25_x=200
x=200
-400<y<400; 28<z<538
Ahmed_25_x=500
x=500
-400<y<400; 28<z<538

These files consist of 12 columns for x, y, z, U, V, W, k, urms,vrms, wrms, uv and uw, respectively.

For k-epsilon models, the Boussinesq equation is to be used for the evaluations of the Reynolds stresses.

Each file is to start with 8 comment lines beginning by a "#", consisting of:

#Case
#Name
#Affiliation
#Code/Numerical method
#Mesh type
#Turbulence model
#Any other useful information
#Column headings for the data

Example: File "Ahmed_25_y=0":

#Case9.4
#Rémi Manceau
#Laboratoire d'études aérodynamiques, université de Poitiers
#Finite volumes
#Block structured
#Reynolds stress model (Speziale, Sarkar, Gatski, 1991)
#Without non-linear slow term, fine mesh, stilts omitted
# x    y    z    U    V    W    k    urms    vrms    wrms    uv    uw
 
 

- Wall pressure:

File Ahmed_25_Cp containing 4 columns for x, y, z and the pressure coefficient Cp on the rear part of the body (end of the roof, slant part and base)
(-215<x<0; 0<y<194.5; 50<z<338 as in the database).

Example: File "Ahmed_25_Cp"

#Case9.4
#Rémi Manceau
#Laboratoire d'études aérodynamiques, université de Poitiers
#Finite volumes
#Block structured
#Reynolds stress model (Speziale, Sarkar, Gatski, 1991)
#Without non-linear slow term, fine mesh, stilts omitted
# x    y    z    Cp
 

- "Visualizations":

In order to better visualize the structures of the flow (see, e.g., A. Spohn, P. Gilliéron), participants are asked to provide some additional material:
        * Friction lines:
 

file
plane
window
Ahmed_25_FL_R.ext
z=338 (roof)
-944<x<-201.2; -194.5<y<194.5
Ahmed_25_FL_SP.ext
slant part
-201.2<x<0; -194.5<y<194.5; 
Ahmed_25_FL_B.ext
x=0 (base)
-194.5<y<194.5; 50<z<244.18
Ahmed_25_FL_S.ext
y=194.5 (side)
-500<x<0; 50<z<338 (bounded by the slant part)

The file extensions ".ext" depend on the format of the image (.ps, .eps, .gif, .jpeg, .png, ...).
In order to obtain the best plotting quality, postscript formats (.ps, .eps) are preferable.
All the friction lines are supposed to be viewed from the direction normal to the plane in which they are plotted
(in particular for the slant part, the direction inclined at 25° from the vertical is to be used).









      * Variables in additional planes:
 

file
plane
window
Ahmed_25_x=-794
x=-794
-194.5<y<194.5; 338<z<500
Ahmed_25_z=360
z=360
-1044<x<500; -194.5<y<194.5

with the same format as, for instance, the file "Ahmed_25_y=0" shown above.
 

- Boundary layer in the symmetry plane:

Boundary layer profiles similar to those in the database are to be provided in the symmetry plane (y=0), at the following x-locations (in mm):

x=-243; -223; -203; -183; -163; -143; -123; -103; -83; -63; -43; -23; -3.

The files are to be named Ahmed_25_BL_x=?, where ? stands for the x-location and consist of 12 columns for x, y, z, U, V, W, k, urms,vrms, wrms, uv and uw, respectively.

Example: File "Ahmed_25_BL_x=-243"

#Case9.4
#Rémi Manceau
#Laboratoire d'études aérodynamiques, université de Poitiers
#Finite volumes
#Block structured
#Reynolds stress model (Speziale, Sarkar, Gatski, 1991)
#Without non-linear slow term, fine mesh, stilts omitted
# x    y    z    U    V    W    k    urms    vrms    wrms    uv    uw
 

- Boundary layer on the central part of the body:

Boundary layer profiles similar to those in the database are to be provided on the central part of the body (25° case), as shown in fig. 3, in the following ranges:

Top:  0<z'<550
Side:  0<y'<360
Down: 0<z<50

The files are to be named Ahmed_Central_???_X, where ??? stands for Top, Side or Down, and X for the point: X=A, A1, B, B1, C, C1, D, D1, E, E1, F, or F1,
and contain U and urms.

Example: File "Ahmed_Central_Top_A"

#Case9.4
#Rémi Manceau
#Laboratoire d'études aérodynamiques, université de Poitiers
#Finite volumes
#Block structured
#Reynolds stress model (Speziale, Sarkar, Gatski, 1991)
#Without non-linear slow term, fine mesh, stilts omitted
# z    U    urms
 
 

35° case:
The same files as for the 25° case are to be provided (with names Ahmed_35_*), except Ahmed_35_y=195, Ahmed_35_x=-38, Ahmed_35_x=-138 and Ahmed_35_x=-178.
 

Moreover, participants are invited to submit a short description of their computations (mesh, numerical method, turbulence modelling, etc.)
Each participant will have 5 minutes to present these details (no results presentation).
 

To know how to submit the results, click here.
 
 


For additional information, please contact R. Manceau (remi.manceau@lea.univ-poitiers.fr)


Back to the Announcement Page

Back to the main page