iterations in FLUENT CFD

Here are our CFD links and discussions about aerodynamics, suspension, driver safety and tyres. Please stick to F1 on this forum.
weis
weis
0
Joined: 13 Sep 2005, 20:32

iterations in FLUENT CFD

Post

Hi all
i am doing some CFD simulation using FLUENT.
However, i cannot manage to get converged results after 1000 times of iterations. The are errors 'divergengence' displayed. And 'turbulence viscosity ratio limited to 1.0e+05'
I have tried to refine my model many times, meshing in Gambit was success but iterations in FLUENT still failed :cry:
Can anyone experienced with FLUENT suggest what is actually causing the problem?
appreciate your help. thanx in advance.

Guest
Guest
0

Post

You'll need to tell a whole lot more info before anyone can make any usefull comment on the reasons of divergence.... What cell type(s) are you using? What is the maximum skewness in the domain? What discretisation, coupling and URs are you using? What Turbulence model? What kind of scales are involved? etc etc etc.
Out of interest... did you follow any courses on CFD at Uni? Are you using it at Uni?

Guest
Guest
0

Post

Anonymous wrote:You'll need to tell a whole lot more info before anyone can make any usefull comment on the reasons of divergence.... What cell type(s) are you using? What is the maximum skewness in the domain? What discretisation, coupling and URs are you using? What Turbulence model? What kind of scales are involved? etc etc etc.
Out of interest... did you follow any courses on CFD at Uni? Are you using it at Uni?
Thanx for the response. :wink:
I am using at my own PC. I learn to use the software myself by following the tutorials. I have minimal knowledge on the CFD software, i read most of them from help files but not quite understand all. i am running analysis on a simplified race car model at 120kmh

Wind tunnel setting: according to Ahmed LES

505893 tetrahedral cells
minimum skewness 0.8 (could not find the maximum skewness)
solver: segregated
time: steady
formulation: implicit
Turbulence model: K-epsilon (2eqn)
k epsilon model: realizable
near wall treatment: non-equilibrium wall functions

Boundary condition
turbulence specification method: intensity and length scale
turbulence intensity: 10%
turbulence length scale: 8% height of car

Discretization
Pressure:standard
Momentum: Second order upwind
Pressure velocity coupling: SIMPLE
Turbulence Kinetic energy: Second order upwind
Tubulence Dissipation rate: Second order upwind

Solutions limit and others are set as default

I have tried with others type of turbulence model but the outcome is still the same. :( i suppose i am not using the appropriate turbulence model or configurations.

Thanx again for the patience. Your advice is much aprreciated.

sasquatch
sasquatch
0
Joined: 22 Apr 2003, 03:31
Location: Melbourne, Australia

Post

use grid inspection to ensure the equi-angle grid skewness is less than 0.9

Run your solution in 1st order options for the first few 100 iterations and then switch to 2nd order if model is stable.

Chances are..if it the solution always diverges for all the options you have tried, then you have something wrong with your grid

if you are not using BL on the wall surfaces, to reduce the Y+, then don't use non-equilibrium wall functions

When you solution diverges, inspect the velocity field and see where it is bad. might help you determine where bad regions in your grid are

WhiteBoy
WhiteBoy
0

Post

I agree with what was said above, but I'd also advise you that, since you appear to be modeling fairly complex geometry (even a simple car is complex) you're very likely to have convergence problems. Before getting too bogged down looking for tiny issues, try using standard k-epsilon instead of k-e Realisable.
By the way, if you're modelling wheels touching the ground, make sure you don't have a hideously skewed grid element down where the wheel touches the ground. Very easy to miss and stuffs your model up good and proper :)
WhiteBoy