CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   ANSYS (http://www.cfd-online.com/Forums/ansys/)
-   -   the turbulent model K-E did not convergence (http://www.cfd-online.com/Forums/ansys/101492-turbulent-model-k-e-did-not-convergence.html)

happy May 5, 2012 04:04

the turbulent model K-E did not convergence
 
Hi all
I run my simulation in steady state mode but the residual is oscillatory so that I believe that buoyancy driven flow that I simualte is follow transient mode. I have encounter big proble through the transient simulation. through the first time step the K-E have never reach the criteris and after the first time step, the convergence is easy?? As I know that during transient simulation MUST all the residual fro all governing equations reach the convergence criteri. Is my simulation acceptable ? I decrease the time step without any benefit.Any comment can help:(
Thanks in advance

LuckyTran May 6, 2012 21:09

Quote:

Originally Posted by happy (Post 359429)
Hi all
I run my simulation in steady state mode but the residual is oscillatory so that I believe that buoyancy driven flow that I simualte is follow transient mode. I have encounter big proble through the transient simulation. through the first time step the K-E have never reach the criteris and after the first time step, the convergence is easy?? As I know that during transient simulation MUST all the residual fro all governing equations reach the convergence criteri. Is my simulation acceptable ? I decrease the time step without any benefit.Any comment can help:(
Thanks in advance

If only the first time-step is having trouble converging then I would just ignore it. As long as the 2nd, 3rd, 4th, etc time-steps all converge well.

You should run the transient simulation long enough for the effect of the initial condition to disappear (regardless of what the first few time-steps looked like). If you run it long enough, then it will not really matter if the first few time steps did not converge since you are going to throw those time-steps away anyway. But, you need it to converge every time-step after the first few. few can be interpreted as, as many iterations as you would like to throw away.


All times are GMT -4. The time now is 08:05.