[mixer]

Hi

I would really appreciate if someone can give me any suggestions. The attached result figure shows a shoot-up in velocity in the outlet region of a Mixer while using MRFSimpleFoam k-epsilon model. It could probably be due to my tetrahedral mesh as the figure shows, but my checkMesh says everything is okay ( Max aspect ratio = 8.56327 OK, Mesh non-orthogonality Max: 68.285 & Average: 19.9527, Non-orthogonality check OK, Max skewness = 1.02244 OK). Surprisingly, the CFX has no problems running and giving good results with this mesh.

a) Is this really due to mesh at outlet region? Kindly suggest what changes should I do to mesh, if it is indeed due to mesh.

b) On looking at Mesh can anyone suggest me what schemes or limited correction should be tried. I have tried upwind scheme, increasing from 1 to10 non-orthogonal correctors, changing outlet bc's from basic to inletoutlet, underrelaxaing k,eps to 0.005 at start, initializing with potentialfoam, but still i get the velocity overshoot at outlet shown in figure. In fact with each iterations, its magnitude increases.

Can anyone suggest any BC at outlet or anything in Schemes to help me get good results.

c) Should i go for transient study. I mean is it not possible to have a SS solution for this kind of problem.


Thanks a lot for your suggestions, Mandar

[josp]

Can you post the entire case?

[mixer]

Hi

Can anyone please suggest me why the reverse flow is happening in the mixer. Except potentialfoam , all other solvers after sometime gives reverseflow and continuity goes on increasing with each iteration (see attached result figures, Figure 2 is potentialfoam result and figure 3 shows reverse flow). I tried changing of mesh and geomeotry (attached figure 1, checkMesh says everything is OK) , and have relooked many many times at all BC's, but still not able to figure things out. I have tried starting with low k and epsilon Relxation, and all the tricks mentioned in the forum...just nothing works..where i am going wrong.

SXMixer.zip

geomeotry.jpg

potentialfoam1.jpg

AfterSomeTime.jpg

Thank you for your help.

[jaswi]

Hi Mandar

Can you try to run it without the turbulence model, I mean run your case within the range of Reynolds number 1-10 and see if you can get some convergence or atleast get rid of the reverse flow. If that works then the problem may not be your grid.

My experience with SimpleFOAM has been similar to yours . I have a Rotor in a pipe which blows up as soon as I switch on the k-epsilon turbulence model.

Best Regards
jaswi