32bits and 64bits FMUs for all examples

[MaximeBaudette]

All examples should be delivered with binaries for both 32 bits and 64 bits.

[lvanfretti]

Actually, we need to generate binaries from different tools - because we have seen with Michael Tiller that there are issues with FMUs depending on tool.

I recommend that we generate the FMUs using:

• OpenModelica
• MapleSim
• JModelica

We can't generate them using Dymola because we lack a license that would allow us to do this kind of job.

[MaximeBaudette]

I thought the only problem was that FMUs generated by Dymola required a dymola license to execute, or that we buy a more expansive Dymola license that allow standalone FMUs. @tinrabuzin can you confirm whether there are specific issues on FMUs generated with JModelica?

@lvanfretti it would be surprising that FMI being a standard, you'd have software dependent issues. Now I have checked on https://www.fmi-standard.org/tools, they have done a lot of cross plateform checks, we should take these tests into account, so we don't need to do them ourselves. Note that Jmodelica and OpenModelica can't generate 64bits FMUs, but MapleSim can.

[tinrabuzin]

@MaximeBaudette Yes, we can't generate FMUs with Dymola because we lack the license for Binary Code Export. However, what I noticed is that, when the FMU created by Dymola without license is used, FMI Toolbox looks for the license on our server. Since, for I don't know which reason, the connection to it is not good, the simulation and estimation process is interrupted. I should note here that this was happening with the quite old version of RaPId so I don't know if it's the same now. But, in any case, those kinds of FMUs (without the license) we should not include in the repo.

Regarding the FMUs created with JModelica, I haven't experienced any problems. The only thing that might cause some issues is that JModelica is much stricter when it comes to the rules of Modelica language so I had to change some models a little bit. Also, there were some problems with Boolean variables which was noticed by Michael Tiller and he promised to look into that. I'm not sure what's going on with that now.

I should also note that we haven't tried running RaPId simulations with FMUs generated with MapleSim.