Converge
 

hi,
I'm study on external flow in sand domain, does i always need to wait the solution until converge. As it take longer time to converge for external flow and in transient state..

Dear tang
if you want final solution, you must use steady-state solver not transient.

Regards

when i use steady state, error will be detected:

divergence detected in AMG solver: k increasing relation sweep

Why this error will appear??

If all settings are ok, usually this error appears when initial conditions for k and epsilon are incorrect. Plz change initial conditions for k & epsilon.

i'm setting :
Inlet. Turbulent intensity = 5%. Turbulent viscosity ratio= 10
Outlet turbulent intensity=1% turbulent viscosity ratio= 10

Any problem in my setting ?

its your boundary condition and seems ok! you should change INITIAL CONDITION. By default initial condition for k and epsilon is 1 in fluent. Plz see
http://www.cfd-online.com/Forums/ope...urbulence.html

Regards

So, I should lower my k and epsilon value. Does I need to change relaxation values?

Why should I change my k and epsilon value? Actually what value is the most possible for k and epsilon

Yes. you should select lower value for k and epsilon. initial value of k and epsilon are depended to your problem. but initial values havent any effect on final solution. so change value of k and epsilon to 0.001 and 0.0001 respectly.

Regards

I'm trying to lower my k and epsilon in my solution control, however the error (divergent detected) become more faster (in lesser iteration calculation).

Is your mesh dimension correct?

how to check whether why mesh is correct?? There is no warning on my mesh,,, Do u mind send me your email...i will send my case data for u, to better understanding on my fluent setting.

I send my mail for you (private message). please send your geometry dimension too.

tnx

i already send to u past two week..

help
 

do you have converge sofeware ?Can you give me a Decoding software?