CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   flat plate with turbulent model (http://www.cfd-online.com/Forums/openfoam/87633-flat-plate-turbulent-model.html)

danielcebrian April 25, 2011 09:51

flat plate with turbulent model
 
Hello everyone.

I write because I am studying a turbulent model in a flat plate with an inclination of 30 degrees, but I canīt get the values ​​of Cd and Cl found in the literature.

http://s2.subirimagenes.com/otros/pr...5clipboard.jpg

I think it might be because it fits well the values ​​of k and epsilon RAS model, looking for the forum I found a thread that talks about how to get it and do not know if my conclusion is wrong, for my model:

http://www.cfd-online.com/Wiki/Turbu...ary_conditions

k = 1/2U '· U', where U '= 0.05Uo, where Uo is the inlet velocity of the fluid.

epsilon = ( Cn ^0.75 ° k ^ 1.5) / l, with Cn = 0.09, l = 0.2 D, and D is the hydraulic diameter.

In my example:
Uo = 0.01
D = 1
kinematic viscosity = 1e-6
Re = 1E4

k = 3.75 e-7
epsilon = 1.886 e-10

The result that i got of Cd was:

http://s3.subirimagenes.com:81/image...1image2993.png

I think it's wrong, because it greatly increases the value of Cd in time

Is that correct? Did i enter either the values ​​of k and epsilon?

thank you.

danielcebrian June 7, 2011 14:59

could someone answer me please? I've been trying for months to find out how
adjust the parameters k and epsilon.

Thanks.

fusij June 7, 2011 21:31

Hey,

First I want to ask, are you calculating the coefficients yourself from the forces or are you using the forceCoeffs tool to get them?

If you are using the latter, are you sure that everything is defined as it should be?

Concerning the exact kinetic turbulent energy and dissipation values, their order of magnitude should be in the range of 10^-4 or lower.

Can you please give me a reference for the drag coefficients for an inclined plate?

Thank you.

danielcebrian June 8, 2011 07:10

Hi Vigfús,

I am calculating the aerodynamic coefficients for myself, but I compared the results with OpenFOAM tool, and have obtained the same results, so I guess the fault is not the way to get Cd and Cl.

I think my problem may be to not use a proper value of k epsilon. When I change these values ​​and put those that come by default in the OpenFOAM tutorial, as is the case of incompressible/pisoFoam/ras/cavity, the results change significantly.

For example the cavity use:

k=0.00325
epsilon=0.000765

If I use these values ​​in my example, the problem converges, but I donīt get the desired values ​​of Cd and Cl according to the literature that it must be:

Cl = 3.02
Cd = 1.76

And I get:

Cl=2.5
Cd=1.45

Thank you very much for answering,

Sincerely.

fusij June 8, 2011 07:30

You said that Cd and Cl were not constant with time, you should see some vortex shedding? There should be fluctuation in the coefficients, have you simulated for longer time? Is your flow in the laminar or turbulent regime?

danielcebrian June 8, 2011 10:19

I'm doing the simulations with turbulent flow, so I am using RAS model with k and epsilon. The flow is turbulent Reynold's 10 ^ 4.

The Cd and Cf are average values.

In the figure below, you can see from the shaft and the value of Cd, and the x-axis the number of iterations. For this I used the RAS method, but the model kepsilon, and RNGkepsilon.

http://s3.subirimagenes.com:81/otros...24grafica1.jpg

Why is there such a difference using one method or another?.

The CD comes kepsilon means = 1.5
The CD means RNGkepsilon = 1.7

I used to both k and epsilon values ​​that came in the tutorial for the case of cavity. I do not know if this would go well and if for any case of flat plate

kiran June 10, 2011 13:30

hi
 
Hi daniel

I think you choose a wrong turbulence model to calculate Cl and Cd values.

i think you should simulate the tubulence properties near the wall(surface of flat plate)
so better use K-W model.

if you still want to continue with k-e then he is my suggestion.

i think you forgot the tubulent intensity(I) to multiply and take inlet velocity as U.
and calulate tubulent kinetic energy using the below formulae

K= 1.5 (U*U*I*I)

then you get k value more than 1
similarly calculate Epsilon that you will get more than 1000.

then reload you case.

All the best

danielcebrian June 13, 2011 10:29

Hi Kiran,

I used the turbulent intensity of 0.05, for kepsilon, because I used the same formula k = 1.5 U · U · I · I = 1.5 · 0.01 · 0.01 · 0.05 · 0.05=3.75e-7, but you made me think I should use the model komega, the problem is I donīt know what values ​​should be put to k and omega.

In the model k-omega could I use the same equation for kepsilon?:

k=1,5 ·U·U·I·I

And omega?

Thank you very much,

Sincerely.

danielcebrian June 16, 2011 07:51

Hi Kiran,

I've been trying to simulate the case with komega, but I could not because he always showed me error when simulating.

I upload the case to the next link because I do not know what wrong, if you can help.

http://www.megaupload.com/?d=QWN8M7NI

I used the following input values:

U=0.2921m/s
Leght flatplate=1m
Re=2000
kinematic viscosity=1.46E-05 m^2/s

Thus calculated the following values:

I(turbulent intensity)=13,3%
k=3/2*U*U*I*I=2.26E-03
omega=k/kinematic_viscosity=155

Thanks in advance.

Sincerely.

Bollonga October 20, 2012 07:58

Hi Danielcebrian

I'm trying to solve a similar case and I'm facing similar problems. Did you get to resolve it? Which turbulence model did you use and which boundary conditions? What are your literature references?

Thanks a lot. Regards.

mertg November 13, 2014 16:47

Hi,
I am trying to run kepsilon case for flat plate with pimpleFoam solver in openfoam. Can anyone tell me the importance of the turbulent length scale. It looks like openFoam is very sensitive.
thank you


All times are GMT -4. The time now is 15:18.