| View previous topic :: View next topic |
| Author |
Message |
svenoxtoby
Posts: 11
|
Posted: Tue May 31, 2011 1:34 pm Post subject: External CFD convergence (part 2) |
|
|
Hi
I previously consulted Steve on this project by email through our local distributor. I thought I should rather share these queries on the forum where possible.
I have a large mansion+sloped terrain external CFD model (the same one you looked at before Steve, if you are reading) which I am simulating for wind from 8 directions (compass NESW and diagonals). Some of the cases reach a satisfactory solution, some do not as yet.
My "From South" simulation does as per attached "FromSouthRetry2FJ.png" which was run overnight.
I might try to repeat the run above and stop it just before it goes wild. But it looks like some of my velocities (the actual values, not residuals) will not have settled by that point.
Question: If I pause it at that point, reduce the false time steps and resume, would DB use the new time steps from then on? Do you think it might help me get through the troublesome patch? Is it best to reduce only the velocity false time steps, or all?
I gather from an existing post on this forum that I should also try a uniform grid.
Thanks for any advice.
| Description: |
|
| Filesize: |
72.35 KB |
| Viewed: |
9243 Time(s) |

|
|
|
| Back to top |
|
 |
Steve Potter

Posts: 572 Location: newcastle |
Posted: Thu Jun 02, 2011 4:20 pm Post subject: |
|
|
Hi Sven,
Perhaps you could take a look at my reply to your previous posting? I suspect this behaviour may be due to the air flow becoming 'locked' at some point in the calculation domain.
Regards,
Steve
|
|
| Back to top |
|
 |
svenoxtoby
Posts: 11
|
Posted: Thu Jun 02, 2011 7:43 pm Post subject: Thanks again |
|
|
Hi Steve
Again, thank you. It does make sense in the contect of your last answer.
I am still curious: would DB use updated false time step values entered partway through a simulation? (Not that this would seem to be the issue in this case!)
Many thanks
Sven
|
|
| Back to top |
|
 |
Steve Potter

Posts: 572 Location: newcastle |
Posted: Fri Jun 03, 2011 8:11 am Post subject: |
|
|
Hi Sven,
Yes, if false time steps are modified part-way through a simulation, the updated values are then used when the calculation is resumed. However, if the calculation is unstable, it is often much more effective to re-start the calculations having reduced the time steps.
Regards,
Steve
|
|
| Back to top |
|
 |
svenoxtoby
Posts: 11
|
Posted: Fri Jun 03, 2011 1:09 pm Post subject: Thanks |
|
|
Hi Steve
Thanks again. I have have much to learn; thanks for sharing the knowledge.
Regards
Sven
|
|
| Back to top |
|
 |
|