|
[Sponsors] |
April 22, 2005, 18:15 |
Stopping and starting unsteady cases
|
#1 |
Guest
Posts: n/a
|
Hi guys, I am running some unsteady cases - and I have a question. When I stop (or cancel) and restart an unsteady case during a timestep, and the residuals have not converged during that timestep - Fluent restarts with a "new time step" i.e. it does not continue with the momentum, continuity etc. residuals from the previous time step.
Any way to make Fluent finish the timestep i.e. continue the residuals before moving to a new time? Thanks -Riaan |
|
April 22, 2005, 19:27 |
Re: Stopping and starting unsteady cases
|
#2 |
Guest
Posts: n/a
|
Good question. I wonder why you would bother about that so much though?
|
|
April 22, 2005, 21:41 |
Re: Stopping and starting unsteady cases
|
#3 |
Guest
Posts: n/a
|
I don't know if it's possible, and definetly would like to know too.
Surely it's not a good approach because it helps to spread errors if you stop your calculation. A similar behaviour happens when you use the adaptive time stepping. If a time step doesn't reach convergence, it doesn't repeat it. Regards, ap |
|
April 25, 2005, 06:39 |
Re: Stopping and starting unsteady cases
|
#4 |
Guest
Posts: n/a
|
Just set the number of timesteps to 0 and restart iterating. Fluent will now stay in the current time step and hopefully reduce your residuals. After that you can finish/continue with your remaining time steps. If you do it in the tui, use /solve/iterate xxx.
But I have not found any possibility to get fluent to do my auto-execute-commands and/or time-step based monitors associated with each time-advance for the timestep in which I cancelled the iterations (even not if I restart - so the line in the monitor file is lost)... |
|
|
|