id	summary	reporter	owner	description	type	status	priority	milestone	component	version	resolution	keywords	cc
5823	Verification of FMI models has interpolation issues	Francesco Casella	Lennart Ochel	"As Lennart pointed out in the Annual Workshop, the reported verification performance of FMI on MSL 3.2.3 is abysmal: the [https://libraries.openmodelica.org/branches/master-fmi/Modelica_3.2.3/Modelica_3.2.3.html verification success ratio] is currently 68/424.

I checked some of the results. In many cases they are just plain wrong, see, e.g., [https://libraries.openmodelica.org/branches/master-fmi/Modelica_3.2.3/files/Modelica_3.2.3_Modelica.Electrical.PowerConverters.Examples.DCDC.HBridge.HBridge_RL.diff.inductor.i.html this one].

However, in some cases I understand this mismatch is just an artifact. Consider for example [https://libraries.openmodelica.org/branches/master-fmi/Modelica_3.2.3/files/Modelica_3.2.3_Modelica.Mechanics.MultiBody.Examples.Elementary.PendulumWithSpringDamper.diff.prismatic.s.html this report]. The FMI output is piecewise linear; the corners are fitting exactly with the reference results, but of course errors in-between those points can be quite large. I guess the problem here is just that the FMI is not correctly interpolating the results on the same grid as the reference results, or maybe just some problem with how the CSV compare is working with the FMI simulation results."	defect	new	critical		FMI				Andreas Heuermann