OK, Thanks, I’ll explore that!
I think that appart from accuracy there is another point of debate.
When the user set up the iteration to default, is the time period set up to default 1. or does it consider the overall time introduced in the amplitude card?.
I understand default value is 1. but in case the user has set up an Amplitude there shouldn’t be a default value as there is a user input value. ¿isn’t it?
Could you please take a look at the file before I post in ccx?. I would not want the report something is just a wrong prepomax setup from my side.
It’s still 1, the time covered by the amplitude is not considered. Also because it may not cover the whole analysis (the rest of the amplitude is just extrapolated).
Looks good. But let’s post the .inp file on the CalculiX forum so that users of other preprocessor can also take a look.
Something should be done because “Default” analysis in thermal transient makes amplitude definition completely useless.
Maybe the solution is to remove the amplitude definition option when analysis is set to “Default”.
We know ccx allows the user to do many things , but that doesn’t mean it is right. The preprocessor could help here because sometimes the error es not so evident.
The disadvantage would be that the users could have the impression that amplitudes are not supported. Perhaps a warning (in the documentation or maybe even in GUI) would be sufficient for now.
Amplitudes work perfectly with Automatic and Direct. The problem is Default. If something doesn’t produce reliable results, I would not allow it.
This is like with the Only Compression. Amplitude is used typically to define Nonlinear behaviors. If default use an incrementation as large as possible , the details on the Amplitude card are lost.
I would give a warning in GUI at least.
Is there an idea of how this could be improved in PrePoMax?
I would suggest a warning for now. There’s also a discussion about this on the CalculiX forum since I think that it could be improved in the solver in some way. I also raised my doubts there and highlighted the fact that IMO it should be an optional change to keep consistency with Abaqus, among the others.
Yep, warning or disable the Default iteration scheme option for transient problems until this is clearer? This way the user should enter a period and find a suitable maximum time step.
Now I don’t think the Default option is reliable for capturing nonlinearities (included those introduced through an amplitude card).
I wouldn’t go that far. After all, even Abaqus has the same (not so intuitive but documented in that case) behavior.