- Index
- » Usage and Applications
- » OpenModelica Usage and Applications
- » ModelicaML - workflow questions
ModelicaML - workflow questions
ModelicaML - workflow questions
Hello everyone,
I have a few questions regarding ModelicaML and Papyrus. I originally used Papyrus for modeling of a waste treatment facility. In a second step, I now wanted to run simulations on that (especially, calculate probable outputs depending on uncertain inputs). As Papyrus itself does not have a solver I am now looking into ModelicaML.
As far as I can see from the tutorial and playing around with ModelicaML a bit, I will have to re-build my model in ModelicaML again - the funtional and structural composition, requirements, use cases,... It is not possible to import SysML projects and 'convert' parts of them into a modelica model, or is it?
The modeling of the project (including requirement analysis, use cases, functional decomposition,...) will be done quite as I would have done it with Papyrus in SysML, no? Only that a few names are different (SysML's Internal Block Diagram becomes ModelicaML's Composite Structure Diagram).
And, on another note: ModelicaML seems to be very flexible in the way I describe a model's behaviour. So, I can not only use mathematical equations, but also a script-based approach ("if.. then..." relationships and such), isn't it? (I'm sorry if I'm inaccurate here, but I hope you get the point?). I am asking because I am afraid that using a mathematical description of the flows of matter through the system could be problematic, as the inputs will probably be highly uncertain - so using thresholds or binary decisions here and there might be more suitable.
Thanks for the help already! Please feel free to point out any misconceptions on my side, and also link any tutorials you find useful!
Best,
Adrian.
Re: ModelicaML - workflow questions
Hi Adrian
in the ModelicaML prototype there is a feature for applying (or removing) ModelicaML stereotypes to UML/SysML models.
This feature is just a helper. It maps as good as possible. After you will still need to check, validate and possibly modify the model.
Regarding behavior modeling: ModelicaML is good for dealing with dynamic behavior described using differential equations (you can use Modelica equations or algorithm for describing the class behavior). ModelicaML also allows translating state machines to algorithmic code. However, there are many limitations (e.g., w.r.t. using events, etc.). If you only have discrete or event-based behavior there are probably better tools for executing the model.
I hope it helps,
Wladimir
- wladimir
- 37 Posts
- Index
- » Usage and Applications
- » OpenModelica Usage and Applications
- » ModelicaML - workflow questions