CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   OpenFOAM Running, Solving & CFD (http://www.cfd-online.com/Forums/openfoam-solving/)
-   -   Convergence problem with simpleFoam solver on tilted channel (http://www.cfd-online.com/Forums/openfoam-solving/83334-convergence-problem-simplefoam-solver-tilted-channel.html)

daev December 23, 2010 16:14

Convergence problem with simpleFoam solver on tilted channel
 
5 Attachment(s)
Hello,

I would like to solve flow in complex duct where the flow conditions are determined by pressure at the inlet and at the outlet.

Firtsly I am trying to solve flow in a simple rectangular channel using simpleFoam solver, no turbulence, where the 'system' directory is taken from tutorial simpleFoam/pitzDaily (everything only 2D). The channel is made by blockMesh, 0.1 m long and 0.02 m height (I will upload the case in next post)

Attachment 5851

At the inlet and at the outlet pressure is specified while velocity is prescribed as pressureInletVelocity and pressureInletOutletVelocity, respectively for inlet and outlet. This runs without problems, the results match perfectly with analyticaly calculated laminar velocity profile. (see results in attached straightU.jpg)

Attachment 5852

Secondly I tried the same scenario but now the geometry (same lenght and channel height) is slightly skewed, so the channel is tilted (again I will upload the case in next post)

Attachment 5856

This solution does not converge and eventually crashes as the 'time step continuity errors' explodes to astronomical values. Mesh and results are in attached tiltedMesh.jpg, tiltedU.jpg and tiltedP.jpg

Attachment 5857

Attachment 5858

I also tried to engage nOrthogonalCorrections but without any change.

Does anybody have idea what's wrong?

Thanks and Marry X-mas

daev December 23, 2010 16:20

2 Attachment(s)
Here are the cases.
I also forgot to mention, both results are after 1000 iterations (not converged). Straight case was ran up to 20,000 iterations, when solution was compared with analytical, while tilted one started to boost 'time step continuity errors' after approx 2000 iterations and crashed after 13,000 iterations.
checkMesh is in both cases fine
OpenFoam 1.7.1 was operated.

Attachment 5859

Attachment 5860


All times are GMT -4. The time now is 16:22.