CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   how to get parabolic velocity distribution in a channel? (http://www.cfd-online.com/Forums/openfoam/91083-how-get-parabolic-velocity-distribution-channel.html)

F1r3 July 29, 2011 20:42

how to get parabolic velocity distribution in a channel?
 
3 Attachment(s)
Hi,

I am new to openfoam and I want to simulate a channel flow with simpleFoam. The simulation runs but i have no parabolic velocity distribution when I look at my results. It seems that everywhere is the same velocity. I want to use the kOmegaSST model.

I'm searching in this forum but I didn't find a similar problem.
Could anybody help me please?

Here are my case folder and pictures of the velocity and pressure of my channel.

greetz Roland

Amir July 30, 2011 04:38

Dear Roland,

you are using turbulent model and in turbulent flow the velocity profile is not parabolic. (it's not laminar)
Indeed, your velocity magnitude is very high, consequently, you achieved small boundary layer. Also, to reach such profiles, the flow has to be fully developed.

Amir

F1r3 July 30, 2011 08:10

Hi Amir,

thanks for your reply.
So the kOmegaSST model is only for turbulent flow. When I change it to laminar I would get a parabolic velocity? Our could you tell me what to change in my case to reach a parabolic velocity? I've tried a simulation with a laminar model and set in the RASProperties turbulence to off but I don't reach a parabolic velocity. Furthermore I made my channel longer so that the flow can fully develop.
To your information I've never worked with CFD before...
Later on I want to simulate the flow through a hydraulic valve. There I think I have to simulate with the kOmegaSST model...

Roland

Amir July 30, 2011 08:56

To reach parabolic profile, use laminar model and turn off other turbulent parameters; also, it doesn't need to extend geometry to achieve fully developed region, just set boundary condition of velocity to zeroGradient at both inlet and outlet of your pipe.


Amir

F1r3 July 30, 2011 09:42

2 Attachment(s)
Thanks for your quick reply.

Now I've reached a parabolic velocity. I've changed my viscosity to a higher value (8.95e-04) only for testing. With the higher viscosity I see the parabolic profil. I think with a lower viscosity I have to simulate with a fine mesh cause the change in velocity is only at the wall and in a very thin region.

Now when I set the velocity for the beginning to (0 0 0) I get bad results. What do I have to change to get better results (faster convergence) cause I can't calculate the velocity for my valve later on...

A colleague told me that he reaches good results (good convergence) within 20 iterations with ANSYS CFX. Is it possible to reach that with OpenFOAM too or do have to run a lot more iterations?

Roland

Amir July 30, 2011 11:34

Dear Roland,
Obviously, the results which you attached are incorrect and the most probable reason is that it's not converged yet.
Many factors can affect rate of convergency such selecting proper solver,schemes,good initialization, good solution control and so on. For your case, I think that it's more appropriate to initialize with average velocity.

PS: I have used both commercial software (FLUENT) and OpenFOAM; it can be proved that if the case is managed precisely, openFOAM would be much more efficient in rate of convergency.


Regards,


Amir

F1r3 July 30, 2011 12:52

Hi Amir,

thanks a lot for your help. So I've to try with diffrent solvers and schemes for simulating my valve. Do you have any suggestions which solvers and schemes are good for hydraulic flow?

greetz Roland

Amir July 30, 2011 15:18

Quote:

Originally Posted by F1r3 (Post 318157)
Hi Amir,

thanks a lot for your help. So I've to try with diffrent solvers and schemes for simulating my valve. Do you have any suggestions which solvers and schemes are good for hydraulic flow?

greetz Roland

Hi,

I believe that it depends on your specific case (flow and geometry); there are many solvers which you can use (refer to manual 3.5) and also different schemes, since this is not my field I prefer to let this discussion open for more experienced users.


Bests,


Amir

F1r3 July 30, 2011 19:26

Dear Amir,

thanks for your help!
I will try with my case and if I don't find a solution I think it's best to start a new thread for a new topic.

greetz Roland

m2montazari August 5, 2011 06:19

hi Roland,
getting parabolic velocity profile is very simple! first you should specify what is the case you want to run? if you want to model a pipe with a fixed velocity at inlet and have a parabolic velocity at outlet, pipe length should be long enough to have fully develope flow at outlet. in this case you should specify a small velocity value (to have reynolds number less than 2300 to have laminar flow) at inelt and zerogradient at outlet; for pressure just use zerogradient at inlet and fixedvalue,value=0 at outlet.
but if you want to specify pressure gradient in pipe and have parabolic velocity profile, as amir said, just set zerogradient for velocity in both inlet and outlet, and fixedvalue for pressure(pressure of inlet should be a bit more than outlet).
you can use simplefoam solver for steady state and icofoam for unsteady(laminar) and piso foam for unsteady(turbulent and laminar). shemes should be linear or upwind and you should achieve converged solution in a minute!

I hope it helps,
Mohammad


All times are GMT -4. The time now is 18:07.