Add FMU participants to oscillator tutorial#466
Conversation
The HTML variant is used in the CMakeLists.txt, maybe it is used by other tools
|
I can now run FMI2 and FMI3 with the newest version of FMpy 0.3.20 and also all plotting options. Ready to merge from my perspective. |
MakisH
left a comment
MakisH
left a comment
There was a problem hiding this comment.
I can now also run and plot the results. Almost ready, I just thought we could remove a few potentially unused files and options (see comments).
|
Note that if I execute the left with FMI and right with Python, visualizing with gives me this plot:
If I execute the left with Python and right with FMI, visualizing with gives me this plot:
fmi-fmi gives me the same as python-python:
I have not looked into the details, I am just raising this in case it looks strange. |
|
The different trajectories look like numerical features to me. Depends on which timestepping methods are used and combined. Playing around with WI, we should be able to remove the instability, but that's beyond this PR. |
4 participants
v3-compatible replacement of #322
Using my own branch to have write access, instead of #465
Ready for review. Should work with the
mainbranch of the FMI Runner (v0.2.0 will be released before merging here).Does still ignore #461. I would suggest to do the
oscillatorafter merging here. Otherwise, things get complicated here.I ran the cases with the precompiled FMU, I did not try to recompile on any other system. @LeonardWilleke compiled the FMU.
The FMU cases give a higher error than the Python implementation (but converge for smaller time step sizes). We still don't understand why exactly IIRC, but this is consistent with the thesis and paper from @LeonardWilleke.