CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   [FLUENT/ICEM] Transition modeling (http://www.cfd-online.com/Forums/fluent/98040-fluent-icem-transition-modeling.html)

robboflea March 1, 2012 16:04

[FLUENT/ICEM] Transition modeling
 
2 Attachment(s)
Hi all,

I'm trying to model transition on a 2D turbine airfoil by using the mesh in the picure. Basically two of the blocks in the boundary layer have been moved to a "LAMINAR" part in ICEM while the remaining part is in the "FLUID" part. The idea is to have the laminar zone to be set-up as a "laminar zone" in the cell zone conditions tab in Fluent, while running the simulation with fully turbulent Spalart Allmaras model. (I know I could use a transition model but I really want to try to tackle the problem in this way and I read somewhere in the forum that it was possible to do it but it was not explained well how).

All mesh check in ICEM works fine and I can import the mesh successfully in fluent. together with the normal inlet, outlet, blade and periodic boundaries I get also a boundary between the turbulent and laminar zones (set as an internal boundary) and a laminar wall boundary on the blade which I set as a normal wall.

The problem is that I can initialize the solution but I cannot start the solver. Using the same exact mesh (with the laminar patch being part of the fluid and the solver fully turbulent) I could run it without problems. Normally I used fmg initialization and then I run the case. Now when I run the fmg i get the following:

>> Reordering domain using Reverse Cuthill-McKee method:
zones, cells, faces, done.
Bandwidth reduction = 159/159 = 1.00
Done.

Creating multigrid levels...
Grid Level 0: 49126 cells, 98574 faces, 49900 nodes; 2 clusters
Grid Level 1: 12217 cells, 49704 faces, 0 nodes; 2 clusters
Grid Level 1: 12217 cells, 27317 faces, 0 nodes
Grid Level 2: 3532 cells, 27978 faces, 0 nodes; 2 clusters
Grid Level 2: 3532 cells, 9694 faces, 0 nodes
Grid Level 3: 1165 cells, 18093 faces, 0 nodes; 2 clusters
Grid Level 3: 1165 cells, 3500 faces, 0 nodes
Grid Level 4: 391 cells, 12094 faces, 0 nodes; 2 clusters
Grid Level 4: 391 cells, 1238 faces, 0 nodes
Grid Level 5: 135 cells, 8227 faces, 0 nodes; 2 clusters
Grid Level 5: 135 cells, 450 faces, 0 nodes
Done.


FMG: Converge FAS on level 5


FMG: Converge FAS on level 4


FMG: Converge FAS on level 3


FMG: Converge FAS on level 2


FMG: Converge FAS on level 1
0.->1.->2.->3.->4.->5.<<<<<

Error: Divergence detected in AMG solver: nut
Error Object: #f

Am I doing something wrong in the mesh generation or in Fluent setup?

Sorry for the long post.

Rob

robboflea March 7, 2012 06:35

Apparently it was a Fluent bug. When restarting Fluent the day after everything was working fine.
If somebody needs help with this strategy for transition modeling I'll be happy to help.

Rob


All times are GMT -4. The time now is 12:58.