Ahmed Fluent 2020
Ahmed Fluent 2020
April 2020
1 Introduction and problem definition 1
“Most of the drag of the body is due to pressure drag, which is generated at the
rear end. The structure of the wake is very complex, with a separation zone and
counter-rotating vortices coming off the slant side edges, whose strength is mainly
determined by the base slant angle.” [3] (See Figure 2).
Once the characteristics of the problem have been defined as in Table 1, where air at T = 20◦
C is the working fluid, our main goal is to compare the computed drag coefficient CD to
experimental data:
2Fx
CD = (1)
ρu2∞ Ax
where Fx is the computed x-component of the force acting on the body (drag force) and
Ax is the projected body area along the x-axis.
Since we’ll use a RANS (time-averaged) approach, we can take advantage of the symmetry
of the problem around the x-y plane (see Figure 1) for both geometry and boundary conditions
to simulate half the space around the Ahmed’s body.
Figure 1: Ahmed’s body with a rear vehicle slant angle of 25◦ (dimensions in mm).
2 Workbench project
− Start WB and drag&drop a Fluid Flow (Fluent) component into the main window as in
Figure 3; rename it Ahmed_body and save this new project: File → Save as...
Figure 4: SpaceClaim.
− Click on Sketch mode icon and move the mouse in the main window to position the
sketch on the x − y plane (or click on the z axis), then click on Plan view icon to
have an orthogonal plan view (Figure 4).
− Using Line, define the simplified 2D profile of the Ahmed’s body reported in Figure 5:
− Click on View icon and select the Trimetric view (Figure 6):
− Now we can extrude the profile using the Pull command: click on the profile and insert
194,5mm by keyboard to pull the profile along the +z direction (Figure 7):
− Use middle mouse button to rotate the view, mouse wheel to zoom the view, Ctrl + middle
mouse button to move the view.
− To create the double roundings on the vehicle front, select Pull command and click on one
of the front edges, then click on Round icon (Figure 8):
Figure 8: Roundings.
Figure 9: Roundings.
− Select the Sketch mode icon and select the bottom plane of Ahmed’s body (Figure 10), then
click Plan View.
− Use Construction line and Offset Curve to draw the construction geometry of Figure 11 that
will be required to define the supporting feet of Ahmed’s body:
− Using Circle, draw the base of cylindrical feet of Ahmed’s body, specifying 30mm by
keyboard as diameter (Figure 12):
− Use Pull to extrude both circles with an extruding length of 50mm, defined by keyboard
(Figure 13):
Now the geometry of the Ahmed’s body is completed, but we need to define an external
enclosure to simulate the external aerodynamics around the body. This enclosure starts 2L
before the body and extends 5L behind it; its height is 1.4m and its width is 2.5W (these di-
mensions are recommended for the 10th joint ERCOFTAC Workshop on Refined Turbulence
Modelling).
− Click on Prepare tab and then click on Enclosure icon; select Box as Enclosure type and
untick Symmetric dimensions (Figure 14):
− Select the body and insert the dimensions reported in Figure 15, then click on icon:
− In the Structure tab, right click on Solid (the Ahmed’s body) → Suppress for Physics since
we don’t need to mesh it nor to perform calculations inside it (as it would in the case of a
two-way fluid-structure coupling) (Figure 16):
Figure 16: Ahmed’s body volume is not required for our CFD analysis.
Since we want to increase the mesh resolution around the body, we need to define a Body
of influence: click on Sketch icon under Design tab and select the symmetry plane of the
Ahmed’s body (Figure 17):
− Click on Plan View and define the rectangular geometry of Figure 18 (the x length of the
body is 1044mm, the width of the rectangle is 2544mm):
− Click on Pull to extrude the rectangular profile in the +z direction for 400mm (Figure 20):
− Under Outline tab → Project → Model → Geometry right click on FFF\Size_box and
select Hide Body (Figure 23):
− Click the Face icon and select the inflow face (left click), then right click on it → Create
Named Selection and specify the name Inflow (Figure 24):
− Repeat this operation for all the faces of the enclosure (Top, Ground, Side, Symmetry,
Outflow, Ahmed_surface). For the surface of the Ahmed’s body: click on the z axis for a side
view, click on Select Mode → Box Select to easily select all its patches (Figure 25):
− Under Outline tab → Project → Model → Mesh, in the Details of "Mesh" tab, specify all
the marked properties as in Figure 26:
− Since we want an extruded mesh with prisms in the boundary layers, we must define their
locations; under Outline tab → Project → Model → Named Selections, click on Ground and
set Include under Program Controlled Inflation in the Details of "Ground" tab (Figure 27):
− Outline → Project → Model, right click on Mesh → Insert → Sizing (Figure 28):
− Under Outline tab → Project → Model → Geometry, right click on FFF\Size_box → Show
Body; click the Single Select icon and Body icon ;
− Repeat the insertion of a new sizing: in Details of "Sizing" - Sizing tab, for Scoping Method
select Geometry Selection, for Geometry select the whole enclosure, for Type select Body
of Influence, for Bodies of Influence select the size box and for Element Size insert 0,030m
(Figure 29).
− Right click on Mesh → Generate Mesh to create the mesh (Figure 30):
− Outline → Project → Model → Geometry, click on Mesh to display the mesh; click on the
top Mesh tab → Insert → Section Plane to insert a section plane for exploring the volume
mesh (Figure 31). Click on Show Whole Elements icon to display whole mesh elements
and not their projection on the section plane:
Figure 31: Mesh prism layers around the body surface and the ground.
The meshing phase is now complete: the mesh is composed by 380k tetrahedrons (Tet4, 4
nodes) for the volume mesh and 100k triangular prisms (Wed6, 6 nodes) in the inflation layers
near the walls.
− File → Save Project... to save the mesh into the WB Project and close ANSYS Meshing.
− In the Fluent Laucher window, untick Double Precision Option (Figure 33):
− Under the Tree tab → Setup → General, click on Check and Report Quality to check the
mesh quality and set the marked settings as in Figure 34:
− Under Setup → Models double click on Viscous - SST k-omega to specify a Realizable k-ε
turbulence model with Non-Equilibrium Wall Functions as in Figure 35:
− Under Setup → Materials → Fluid, click on air to check the correct values of Density and
(Dynamic) Viscosity as in Figure 36:
− Setup → Cell Zone Conditions, double click on enclosure_enclosure to check our enclosure
is filled with Fluid air as in Figure 37:
− Under Setup → Boundary Conditions, right click on the chosen Named Selection → Type
to set the correct boundary condition (Figure 38): side, symmetry and top boundary condi-
tions must be set to symmetry type (no shear stress), ahmed_surface and ground boundary
conditions must be set to wall;
− For inflow, set it as velocity-inlet type with the marked settings of Figure 39:
− For outflow, set it as pressure-outlet type with the marked settings of Figure 40:
− Under Setup → Reference Values, specify inflow under Compute from and Ax = 0.057516
m2 as Area (Figure 41):
− Under Solution → Methods select the Coupled method for the Pressure-Velocity coupling
and First Order Upwind (for now) for Momentum, Turbulent Kinetic Energy and Turbulent
Dissipation Rate (Figure 42):
− Under Solution → Controls specify 0.3 for both Momentum and Pressure Pseudo Transient
Explicit Relaxation Factors (Figure 43):
− Under Solution → double click on Report Definitions to define new quantities to be re-
ported; select New → Force Report → Drag to define a drag report (Figure 44):
− Under Drag Report Definition specify the settings of Figure 45 to define a drag coefficient
report:
− Under Solution → Monitor → Residual tick Show Advanced Options and set the Conver-
gence Criterion to none since we’ll define a fixed number of iterations (Figure 46):
− Under Solution → Initialization select Hybrid Initialization and click Initialize (Figure 47):
− Under Solution → Run Calculation set 50 as Number of Iterations and click Calculate
(Figure 48); anyway, we can stop the calculation before the 50 iterations if we believe the
solution to be acceptable (checking residuals or report plots, for example).
− When the solution "converged" (we are employing first order upwind schemes), we can
go back to Solution → methods to select Second Order Upwind for Momentum, Turbulent
Kinetic Energy and Turbulent Dissipation Rate (Figure 49) to improve the calculations (less
numerical viscosity dissipation); run the simulation for 200 additional iterations for example.
The history plot of drag coefficient CD is reported in Figure 50; we can see that with First
Order Upwind scheme (=till 50 iterations) for all the transported quantities (velocities and tur-
bulent quantities), CD "converged" to 0.45, while the Second Order Upwind scheme (=above
200 iterations) decreased CD to 0.334. This is in perfect agreement with the upwind schemes:
first order scheme is more diffusive than second order and simulates a higher viscosity flow
with higher drag coefficients.
The comparison of the computed drag coefficient to other experimental and numerical
results is reported in Figure 51; as we can see our result is in good agreement with both exper-
imental and numerical data from literature, even if it is computed through a RANS simulation
with a not so refined mesh.
− In the Outline tab, under Cases → Ahmed_body → enclosure-enclosure, select only the
ahmed_surface and the ground components. Click on the Location icon and select
Line, then specify the marked settings as in Figure 53 → Apply:
− Click on the Streamline icon and specify the marked settings as in Figure 54 → Apply,
to view the streamlines in the symmetry plane in the rear separation zone:
A comparison between computed and experimental (mean) streamlines along the symmetry
plane in the rear separation zone is reported in Figure 55: although the relative position of re-
circulation bubbles is not exactly computed, we can see that an approximate width of 200mm
is computed with good agreement to the experimental width.
Figure 55: Comparison of streamlines in the rear separation zone (from [4]).
A comparison between computed and experimental (mean) streamwise velocity profiles along
the symmetry plane in the rear separation zone is reported in Figure 56: the computed profiles
show a discrete agreement with experimental ones.
Figure 56: Comparison of streamwise velocity profiles in the rear separation zone (from [3]).
CFD-Post is a powerful tool to analyze and display the results of a CFD simulation in a
graphical way; be careful to choose a meaningful way to present your results (see Figure 57).
References
[1] S.R. Ahmed, G. Ramm, and G. Faltin. Some salient features of the time-averaged ground
vehicle wake. In SAE Technical Paper. SAE International, February 1984.
[2] F.J. Bello, T. Makela, L. Parras, C. del Pino, and C. Ferrera. Experimental study on
ahmed’s body drag coefficient for different yaw angles. Journal of Wind Engineering and
Industrial Aerodynamics, 157:140 – 144, 2016.
[3] C. Hinterberger, M. Garcia-Villalba, and W. Rodi. Large eddy simulation of flow around
the Ahmed body. In R. McCallen, F. Browand, and J. Ross, editors, Lecture Notes in
Applied and Computational Mechanics/The Aerodynamics of Heavy Vehicles: Trucks,
Buses, and Trains. Springer, New York, 2004.
[4] H. Lienhart, C. Stoots, and S. Becker. Flow and Turbulence Structures in the Wake of
a Simplified Car Model (Ahmed Model), pages 323–330. Springer Berlin Heidelberg,
Berlin, Heidelberg, 2002.