CFD Online Discussion Forums

CFD Online Discussion Forums (
-   Main CFD Forum (
-   -   RSM+2nd ordedr causes turbulent viscosity limited to viscosity ratio of.... (

khsiavash July 29, 2009 12:44

RSM+2nd ordedr causes turbulent viscosity limited to viscosity ratio of....
Dear friends
I have modeled "opposed coollant jets injected normally into a heated confined crossflow".
Problem statement: -fluid: incomp ideal gas - Ujets=75m/s -U(crossflow)=13.7m/s
-Symmetry planes are considered D(jets)=8mm Inlet Area=0.02*0.1m2 Total length of Crossflow=0.7m
Also I`ve checked the inlet values and also Ypluse (for proper WF), in addition I`ve maked 640000 hexahedral mesh in reasonable way.

My Problem: when I have used RSM model + second order upwind order of descretization, I have received this error: " turbulent viscosity limited to viscosity ratio of 1.000000e+05 in 2 cells--continue to 20000 cell or more up to "floating point!!!
Notice:But in other turbulence models(ske-rke-RNG ke-v2f) even with second order, I reached to reasonable answers.I`ve started with ske then I`ve switched on RSM or I`ve started with 1st order then I`ve switched on 2nd order (in RSM).In addition I`ve reduced the URF of Reynolds Stresse.nothing happened, again I faced with that error.
What is wrong with RSM+2nd order?
Do you have any suggestion?

pz1337 April 4, 2011 08:04

I have read through many posts, this one describes my problem very good. I have exactly the same problem with 2nd order RSM.

Increasing Turb Visc ratio to 1e20 => floating point error
All URF => 0.1 => still the same problem.

I have already read through fluent manual, CFD Wiki und over 50 threads on the forum.
Thank you very much!

pz1337 April 4, 2011 14:34

Oh well, this should be in ANSYS/FLUENT ... if a moderator/admin would be so kind, thank you very much!

Assen October 19, 2011 03:10


I am in a similar situation, did you ever solve the case?.

Thank you.


pz1337 October 19, 2011 03:14

I did not and used k-e Realizable in the end....

Assen October 19, 2011 03:37

Ok, thanks. The only way I found to solve the problem by adapting the mesh based on the isovalue of turbulent viscosity ratio. The drawback is that after adaptation resulting number of cells is very high, multiplying the original number of cells per 5.

Far October 19, 2011 05:38

whats the ASPECT RATIO of mesh? Because I have similar problems and I have concluded that it is due to the aspect ratio. Typically aspect ratio should be atleat 100 and better is even less. With aspect ratio 100 one should also switch to the double precision solver

Assen October 19, 2011 06:21

Hi Far,

Double precision mode was activated. The aspect ratio is 57 and skewness 0.51. I think they are acceptable values but perhaps not good enough for an enhanced wall treatment.


gfoam January 6, 2016 09:27

I know this is an old post, but I have similar problems with a cyclone. I'm trying to simulate it with RSM turbulence model, and I have tested everything. I'm using a full hexaedral mesh, and the AR is in average 25, but it has a maximum of 150 near the walls. I've seen that the cells having tvr limited are on the lower cone where near the exit to the dust discharger. The mesh has over 9M cells and every test I do takes days to get nothing but floating point error! Can anyone help me with an advice? Thanks in advance.

CFDYourself January 7, 2016 10:39

This message normally occurs when the solution is diverging.

Grid quality (such as aspect ratio) will affect the solution stability, if your mesh is already appropriate there are other things you can do to help the solution converge.

gfoam January 7, 2016 13:30

Yes I know that my problem is diverging, the question is Why. Ive read the posts of Far and I'm trying to implement the recomendation posted there. Regards

All times are GMT -4. The time now is 11:14.