Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
VT1
2011
Contents
Contents..................................................................................................................................... 2
Abstract...................................................................................................................................... 3
Introduction................................................................................................................................ 3
Method....................................................................................................................................... 3
Mesh study............................................................................................................................. 4
Drag coefficient.................................................................................................................... 7
Lift coefficient..................................................................................................................... 7
Flow fields................................................................................................................................. 7
Conclusion.................................................................................................................................. 9
Abstract
The technical report is the last part of our project work, after the poster presentation. The main goal is
to compare our results with experimental data.
Our studies are based on a blunt body which is designed in two different slant’s positions : 25 and 35
degrees of angle.
The meshing has been the same for the two positions of the slant. On different body surfaces as front,
back, slant and symmetries, the mesh was really fine. And for the rest of the domain, we put a coarser mesh. We
reached the meshed independency at 400K cells for the 25° geometry and at 750K for the 35° geometry.
On the CFD post results, we get one main vortex and another one, a little bit smaller and located under
the main one. However, the 25° geometry has a smaller main vortex and a better position for it (on the back of body so
the slant has a linear flow)
The two flow remains similar but there is much more disorder in the main vortex of the 35° geometry
than in the 25°’s one. We see that the main vortex of the 35° geometry is created at the end of the slant and not on the
back of the body.
About the drag and lift coefficients, there are differences. For the drag and the lift, the 35° geometry
has better results (less drag and more downforces)
Due to the view of all these results, we can admit that the 25° geometry is quite better than the 35°’s
one even if the results about the coefficients show the contrary. We can finally say that with a 25° geometry, it will be
easier to drive and we will need less fluel.
Introduction
The Ahmed Body is the name of a simplified car model. It is made up of a round front part, a
moveable slant plane placed in the rear of the car body to study what happens at this location for different angles (in
that case, 25 & 35)
This report focus on explaining the body’s mesh, describing and discussing the velocities and the flow
fields of the two geometries and the drag & lift coefficients on the body and especially in the back, where changes will
be made.
At least, we will compare our results with experimental data.
Method
To reach our goals, we used a Fluid Flow Method (CFX) by the help of the ANSYS software. We used the SST-
turbulence and turned off the heat transfer.
We drew half of the body and used symmetry to get the entire body for our calculations. The Reynolds number is
near to the experimental one (768000, based on the height of the body on the experimental case). The inlet
velocity is 40m/s and we used the same coordinate system as the experimental results.
Number of cells
Convergence criteria
The convergence criteria is something very complicated to decided, that is why we have decided to
keep the same residual target as the first assignment (1e-5). We have obtained good result and keep a short calculation
time with this.
First, to get a nice meshing and earn some time about this part, we put the number of iterations to 50.
Then, when our mesh was fine, we upgrade the number of iteration to 200 for all our calculation.
So, 200 interations with a residual target of 1e-5 is a good compromise: we could do a lot of tests
thanks to an acceptable calculation time.
After the mesh independancy study, we have studied our velocity profiles. For this, we have compared some profiles
given by CFX and the profiles given by a wind tunnel measurement. This comparison is really important for our
modelisation: it permit us to know if our results are conform to reality or not.
To read the reults we have plot the results with EXCEL. The plots represent the velocity on each “line” of
measurement. The inlet velocity is 40 m/s. When the wave goes on the left that means that the velocity decrease and
increase on the right.
Figure
For this case 4: CFX
(φ=35°) we canand experimental
notice velocity
an error between the real profiles
case and our modelisation. In the vortex
comparison
behind the body, in the top of the vortex the for φ=35°
velocity decrease more in the real case than in the CFX modelisation. It
is due to the vortice which is bigger so there are much more turbulences.
There is a real difference between the two vortices of the different angles. With the angle of 25° there is
just one vortex on the lowest part of the wake. For the 35° there is two vortices one on the top and one on the bottom.
Drag coefficient
Thanks to our modeling we have calculate the different Cd of the body and compare then to
the real case results.
25° 35°
Drag
coefficients
CFX wind tunnel CFX wind tunnel
Cd 0,3 0,29 0,27 0,255
Ck 0,03 0,02 0,04 0,03
Cs 0,13 0,15 0,1 0,09
Cb 0,08 0,06 0,08 0,09
Cr 0,06 0,06 0,05 0,045
Figure 6: Comparison of drag coefficients
We can notice some difference between the real case and our results but this difference is
really small.
The little difference can be explained with the different simplifications like the material we
used, the size of our mesh. The results for 25° are better than for the 35°. The reason is that
in the second case we will have more turbulence and so the results are more complex to
obtained, we need more cells for it.
Lift coefficient
CL COEFF
Flow fields
Figure 2 : 25° geometry’s flow field
Figure 3 : 35° geometry’s flow field
Here are the results for the flow fields of the two geometries.
We can see that there is, for the two geometries, a main vortex located on the
back of the body (even if the one of the 35° geometry is created on the end of the slant). We
can also see another, smaller vortex under the main vortex. The flow looks steady on the
slant (until a certain location where the vortex is created for the 35° geometry).
Conclusion
If we compare the two screenshots, we can say that the 25 geometry get better
results than the other one. Indeed , the vortex are smaller that means that the turbulence are
less present. They both are located on the back of the body but the main vortex of the 35
geometry is a bit higher than the 25 geometry’s one.
We can also notice another vortex in the 35 geometry results which is located
under the main vortex. Due to our results about the 35 geometry, we can say that there is
much more disorder on the back of the body than on the back of the other geometry. So we
can say that the 25 geometry is quite better than the 35 geometry.
However, the drag and lift coefficients are better with the 35 geometry are lower
than the 25’s ones that means better performances about drivability and fuel consumption.