CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   false time step implementation for steady state (http://www.cfd-online.com/Forums/cfx/25965-false-time-step-implementation-steady-state.html)

Kushagra June 17, 2008 23:34

false time step implementation for steady state
 
Hello All,

Convergence for the steady state simulations can be obtained by decreasing the 'false time step'.

Keeping the default RMS convergence criteria (10^-4) for the steady state simululation, convergence in my simulation became difficult to achieve after a while and hence I changed the 'Timescale Factor' from 1 to 0.01 [in the pannel Solver Control -> Basic Settings -> Convergence Control ->Fluid Timescale Control].

It resulted in decreasing the auto time scale from 2*10^-2 to 2*10^-4 and I achieved the convergence.

I do not have a understanding of how the steady solver works (using a 'false' time step)? Could you please help me in understanding the affect/effect of high/low false time step on the accuracy of results.

And what convergence criteria should I use for the problems like cyclone?

With very kind regards, Kushagra

Glenn Horrocks June 22, 2008 20:06

Re: false time step implementation for steady stat
 
Hi,

How timestepping works is described in the documentation.

As for what convergence criteria you need, that is problem dependant and you should do a sensitivity study to check it. There is also a guide on convergence levels in the documentation.

Glenn Horrocks


All times are GMT -4. The time now is 03:55.