Closed FWuellhorst closed 2 years ago
While further investigating the issue I found ATransparent
to be a redundant parameter. AWin
is used instead. ATransparent
only occurs in the OneElement
model.
I also temporarily fixed the issue in the AixLib. However, I still think the underlying issue is the ROM
model in the IBPSA.
https://github.com/RWTH-EBC/AixLib/issues/1195
and the PR: https://github.com/RWTH-EBC/AixLib/pull/1196/files
@mwetter : Do you have any idea on why these warnings occur or how we can avoid them?
It looks to me like Dymola gives up when it parses
Real y=0 if sum(ATransparent) > 0;
For what it is worth, the model looks valid to me and simulates with OpenModelica.
Thanks!
I will still remove ATransparent
as it's of no use.
Nevermind. I did not saw this comment:
ATransparent and AWin are equal. When merging windows and exterior walls, AWin can be set to zero while ATransparent still represents the actual window area for solar radiation calculations.
When using the
ReducedOrder
model, propagation of the parameternOrientations
leads to the following warnings (on Dymola 2020x)If I set the parameter directly and don't propagate it, the warnings disappear. Adding
Evaluate=true
to the declarations does not solve the issue.The problem seems to be the conditional declaration of additional parameters, e.g.
One solution would be to remove
nOrientations
and declare the parameters using[:]
notation. However, this then raises warnings when checking theOneElement
model. Another solution would be to specify booleans, e.g.use_traAreas
or similar. What also works is to setnOrientations=size(ATransparent, 1)
on the upper-level model. Does anyone know another solution to this issue directly in theOneElement
model? I can work with the last solution (specifying thesize
). But in my opinion, this should not occur in the first place.I also added a small package to reproduce the warnings.