Processing Error

[guyEANZ]

I'm struggling to get usable results out of a foundation detail.
The construction is a SIP panel on a concrete waffle slab (on ground) with a floating floor over.
The calculation runs for about 5 seconds at a normal speed and then suddenly speeds up and completes 99% in a further second or two.
The results, unsurprisingly just show a straight line in all graphs.
I've tried deleting the moisture sources, air cavity and membranes to see if this is the problem, but the calculation still doesn't return anything useful.

Am I missing something obvious?

[Christian Bludau]

Dear Guy,
thats convergence problems you have.

The main problem in your construction is the numerical grid. Your grid is much too coarse. You have to use a steady-going grid as this shows better convergence behavior, than a size-jumping-grid. Please check the expansion factor. It should not be higher than two respectively lower than 0.5.

Grid left is much too coarse, can cause convergence problems. Right is a nice, steady-going grid.

grid2d.png (84.28 KiB) Viewed 13217 times

Further your rain water sources can cause problems, if a material gets supersaturated during the rain event.

More information on such convergence problems you can find in the following topics:
http://www.wufi-forum.com/viewtopic.php?t=332
http://www.wufi-forum.com/viewtopic.php?t=494
http://www.wufi-forum.com/viewtopic.php?t=789

Best regards,
Christian

[StrawConstruction]

Hello Christian,
I am also having problems with the grid: is there a way to understand clearly which grid is giving me the lower error? I am working on a simple wall, trying out a 75x75; 100x100; 112x112;130x130;150x150 grid. I would like to confront them to chose the proper oneto use a basis for my following calculations. However I cannot chose a value to confront all of them (I expect to have smaller error with finer grid of course).
Thank you in advance
G.

[Christian Bludau]

Hi G.
that question can not be answered universal. That is always depending on the construction and layers. Generally spoken a finer grid should produce more accurate results.
Show us a picture of your construction and the grid you selected.
Christian

[StrawConstruction]

Geometry

1.PNG (12.66 KiB) Viewed 10552 times

Here is the geometry. It's a simple wall with brics and EPS insulation.

Grid 75x75

Griglia 75x75.PNG (21.7 KiB) Viewed 10552 times

I tried with many grids, in order to define the most optimized one in terms of time of calculation and error. However I cannot chose the right value to confront all of them and justify my choice.
I attache here all the grids i tried out.
Thank you again
G.

[StrawConstruction]

It didn't work with all the pictures. Sorry.

Grid 100x100

Griglia 100x100.PNG (22.18 KiB) Viewed 10552 times

Grid 112x112

Griglia 112x112.PNG (19.15 KiB) Viewed 10552 times

Grid 130x130

Griglia 130x130.PNG (21.7 KiB) Viewed 10552 times

[Christian Bludau]

First one (75x75) looks sufficient for me.
BTW: I would not optimizing the time of calculation. Let the computer do the work during night.
Christian

[StrawConstruction]

I also think 75 is sufficient... but as I am using it for my thesis I need to justify my choice somehow with numbers... isn't there a way? I tried first to confront the biggest theta and phi values, but the results were not as expected (getting smalle with the finer grid).
Thank you again for your help.
G.