CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   How do turbulence viscocity ratio (http://www.cfd-online.com/Forums/fluent/62969-how-do-turbulence-viscocity-ratio.html)

Daisy Duck March 24, 2009 20:21

How do turbulence viscocity ratio
 
Hi there. I was just running a quite big case file and after it was going just fine for about 150 iterations I started getting the message: turbulence viscosity ratio limited to 1e+005 in 1 cells....2 cells.. and the cell kept of increasing. I read a couple of similar topics but they all said that it happened in the first couple iterations and it disappeared. Anyone knows why is this happening? I tried plotting the ratio, but I can't show the contours everywhere because fluent freezes ( probably not enough memory). I did however see them on the sides of the cabinet, and the components inside. They looked just fine. So i am assuming this is happening somewhere in the fluid around it, but i can't see where. also the highest ratio on the scale was about 3.5e+4, which is not so close to 1e+5.

edit: this is about air flow over a race car. I am using Reynolds stress model(7 eqs). This is what my teammate told me to use ( i am still learning). I though k-e is the one I should have used. Can anyone tell me what is the difference?

zhaopeng March 25, 2009 05:16

try to improve mesh quality,it may be effective.

you can find the difference between k-e model and Reynolds stress model in FLUENT 6.3 User's Guide 12.2

Daisy Duck March 25, 2009 07:57

The mesh is already pretty fine, but I will see what can be done. thanks fore replying

Freeman March 25, 2009 15:44

Which is your flow Regime (i.e. Reynolds number)? I have also experienced that when Re is high and there are strong recirculations in your problem (high or extreme flow separation), turvulent viscosity ratio (TVR) keeps also high, despite your mesh is fine. If the TVR is in a no. of cells reasonablely low with respect the total number of your domain's cells (let's say 3%-7%), I would not be very worried about it...

But of course, I agree also that refining mesh may improve the TVR.

You can also try to switch on the Multigrid solver and try to do some VCycles-type iterations in the k and e variables, also enabling BCGSTAB smoother and setting up max. 2 pre-sweeps. This may improve your TVR, but it will also make a bit slower your computation (specially if you make this pre-sweeps).

Hope this helps,

Freeman


All times are GMT -4. The time now is 09:34.