1

Let's say I export a model in Dymola with FMI-1 standard:

translateModelFMU("SomeModel", false, "", "1", "me", false)

and then I export the same model with FMI-2 standard:

translateModelFMU("SomeModel", false, "", "2", "me", false)

I noticed that the model statistics (Sizes of nonlinear systems of equations, Initialization problem) changes.

Can anybody explain why?

AdamKe
  • 63
  • 4

1 Answers1

1

The FMI standard version 2.0 has a number of additions to the FMI standard that can allow a tool to more efficiently simulate the exported model. These changes include:

  • Support for Jacobian matrices

  • New classification of interface variables

  • Continuous state variables are explicitly listed as FMU variables

More information about the changes can be found in this paper from the Modelica'2012 conference: Functional Mockup Interface 2.0: The Standard for Tool independent Exchange of Simulation Models

jrhodin
  • 659
  • 5
  • 14
  • 1
    Thanks for the reply. Yes, I'm aware of (most) changes that FMI 2.0 introduced. Currently I'm focusing on performance of my model, as it has to be real-time capable. And as exporting it with 2.0 the non-linear system increased, my alarms went off. And what efficiency is conserned, so far, the models exported with FMI 1.0 are runing faster (by not much, but enough to notice) – AdamKe Dec 15 '16 at 16:54
  • 1
    My answer is to the question "Why did the model statistics change?". It sounds like your *real* question is how to get your model to run in real-time. I suggest opening a new question for that, preferably with the model code. – jrhodin Dec 15 '16 at 17:12