Calculation stops before end

[jacques.l]

Hello,

I have a problem when I run a simulation of 2 years or more on WUFI 2D.
The calculation always stops after 14271 time steps. I get a message that the calculation is finished and the results are available, but not over the whole period that I would like.

Any idea where the problem might come from?

Thank you in advance for your help!

Best regards,
Lionel

[Christian Bludau]

Please see this topic:
viewtopic.php?f=15&t=1920

[jacques.l]

Hello Christian,

Thank you for your reply.

Indeed, it is the same problem that I have. But the strange thing is that I received the file from a former colleague where the simulation was running correctly with the same grid.
Could it be that the simulation is computer dependent? (related to computer memory or something like that?)

I have tried it again with maximum 'Increased Accuracy' (5E-5) and maximum 'Adapted Convergence' (1500). The calculation goes a little further, but still stops before the end (see attachment).

[Christian Bludau]

Hi Lionel,
the result of WUFI 2D should not depend on the computer. If your colleague had an older version of WUFI 2D, the grid may be have changed during loading. Did you change anything else but the grid? Because that also can be affected by eg. different boundary conditions, materials....
At the time step the calculation break down, does there a peak in water content appear somewhere in the construction?
Christian

[jacques.l]

Hello Christian,

No, I didn't change anything from the file that I received.

Indeed, there is a very high water content peak (up to +/- 400 kg/m³!), but it appears in the whole insulation components, not only in a specific area.

When I refine the grid, the calculation runs to the end but I still get strange results with a very high water content peak. See attachments.

When I make the grid even finer, then the calculation doesn't run to the end again.

2023-03-08 16_17_38-Convergence Analyzer.png (56.81 KiB) Viewed 4456 times

2023-03-08 16_17_00-Window.png (23.93 KiB) Viewed 4456 times

[jacques.l]

Up

[Christian Bludau]

This problem is somehow related to convergence problems. If reducing the grid does not help, maybe change in material, sources if included, boundary conditions or also simplifying the construction may. If the condensation occurs in air layers, splitting may help (see here: viewtopic.php?f=32&t=1777&p=5226#p5226). Without knowing what you like to simulate, it is not possible to give more specific tips.
Christian