CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)

 kongl1986 February 28, 2012 14:27

Converge problem for unsteady gas solid flow

I am simulating gas solid flow for euler euler model in 10m circulating fluidized bed.
When I using a 18000cells mesh ,it converged well. and i can get the final solution about 12s flow time.
After refined it to 36000 and 54000 cells mesh, there is a converge problem.
the refined mesh only got converge before 4m or 2m from the inlet of bed, after that, the values monitored in rest of bed are keeping fluctuating when the flow time is more than 30s, although each time step shows that it converged.

The time step i am using is 0.0001s, the absolute criteria is 0.0001

can any one help me, thank you

 Musa February 28, 2012 21:54

Rule of thumb for setting time step is del t< delx/v
where del t= time step,del x = smallest mesh size, v= velocity. So make sure your time step is small enough for your refined mesh size. Next you can try and lower the under relaxation factor of pressure, momentum, volume fraction and K-epsilon if you are you using a turbulent viscous model.

 kongl1986 February 29, 2012 11:59

for the 36000 cells mesh .the delx=0.00056m,and V=2.3m/s, so i think the time step is small enough, and i am using implicit scheme for volume fraction, and I check the manual of fluent, implicit scheme dont require the courant number

 Musa February 29, 2012 23:40

Okay, have you tried lowering the under relaxation factors?

 kongl1986 March 1, 2012 17:28

now i am using default underrelaxtion factors, cuz i found in each time step iteration, it got converge, if i change it, which one should i change?any clue to do this? thank you for your reply

 kongl1986 March 1, 2012 17:31

now ,i tried to use lower time step to see whether it works

 All times are GMT -4. The time now is 19:35.