CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM (http://www.cfd-online.com/Forums/openfoam/)
-   -   simple 3D pipe flow diverges with icoFoam (http://www.cfd-online.com/Forums/openfoam/89844-simple-3d-pipe-flow-diverges-icofoam.html)

Rol June 23, 2011 09:24

simple 3D pipe flow diverges with icoFoam
 
1 Attachment(s)
Hello!

I am trying to get grips with OpenFoam. I set up a really EASY case. It is some kind of straight tube. The simulation should be easy, but after 6 iterations, icoFoam diverges. I use the solver settings from the elbow testcase.
Possible problems i thought on:
mesh: it is really fine, much too fine (just to be sure)
time step: is also set really small (dx=2/40 m, U=1m/s -> dt=0.05 should be sufficient, i set it to 0.01)
boundary conditions: inlet with Velocity set to 1 m/s; p at outlet set to 0 Pa. The rest is wall. I checked BC several times, and i find no mistake

I include the test case.

the case is ready to

blockMesh

and

icoFoam

you will see, it diverges after 6 time steps.

any hint is really appreciated!

Roland

Rol June 23, 2011 13:33

I think my test case was not as simple as i thought. The geometry of the tube leads to secondary flows. however, i found out convergence with upwind divScheme is much better.

mechy June 24, 2011 09:55

Quote:

Originally Posted by Rol (Post 313306)
I think my test case was not as simple as i thought. The geometry of the tube leads to secondary flows. however, i found out convergence with upwind divScheme is much better.


I think the length of your tube in y direction is very small and you can not impose zeroGradient at the outlet I suggest you, to increase your length ten times of square root of section area

Rol June 24, 2011 12:03

mechy, thanks, although i checked the solver conditions several times i payed less attention on the geometry itself. It makes sense that the short tube interferes with the zeroGradient boundary condition. Thanks!

mechy June 25, 2011 04:28

Quote:

Originally Posted by Rol (Post 313425)
mechy, thanks, although i checked the solver conditions several times i payed less attention on the geometry itself. It makes sense that the short tube interferes with the zeroGradient boundary condition. Thanks!


is your problem solved with increasing length of tube ?

Rol June 27, 2011 17:02

yes, increasing the length solves the problem. Also i set the time step very low. maybe setting pressure boundary condition instead of velocity at inlet would also help convergence. but it works for me. thanks


All times are GMT -4. The time now is 02:06.