Page 2 of 2
Re: Strange Result using Diffuse Radiation Boundary Condition
Posted: 14 Dec 2012, 13:19
by carstenp
Hi Peter,
I prepare a small example and upload it as soon as possible.
carsten
Re: Strange Result using Diffuse Radiation Boundary Condition
Posted: 15 Dec 2012, 14:17
by carstenp
Hi Peter,
her is the small test case for the "NaN in the first step" problem. Simply extract an run with ElmerSolver.
carsten
Re: Strange Result using Diffuse Radiation Boundary Condition
Posted: 20 Dec 2012, 23:38
by carstenp
Hi Peter,
did you have any time to take a look on the NaN example?
carsten
Re: Strange Result using Diffuse Radiation Boundary Condition
Posted: 21 Dec 2012, 01:25
by raback
Sorry, too busy these last days before Xmas. I can't much promise to look it during the holiday season either. Definately NaN should never occur at the first iteration. -Peter
Re: Strange Result using Diffuse Radiation Boundary Condition
Posted: 04 Feb 2013, 16:33
by raback
Hi Carsten
Better late than never, I hope. I just committed a fix for the problem (rev. 6024 in the new repo).
The problem was that sometimes the visible fraction, Fsum, was a little bit larger than one. Then when at some stage you take a root of 1-Fsum the result is NaN.
-Peter
Re: Strange Result using Diffuse Radiation Boundary Condition
Posted: 04 Feb 2013, 18:37
by carstenp
Hi peter,
thank you very much for fixing the problem!
carsten