Modeling Continuous Dynamics in SysML



By paul ~ December 9th, 2008. Filed under: Industry, Systems Engr..

An area of great interest for me is hybrid continuous-time / discrete time simulations.  A very sophisticated player in this arena is the Modelica language (also see the informative Wikipedia article.)  If you haven’t heard of it, I’d definitely recommend checking it out.  I first became aware of Modelica while creating a proposal in response to an SBIR topic for hybrid simulation capability.  One of the things that is so appealing about Modelica is the declarative, bi-directional nature of the language and simulation.  It reminds me of Prolog, but for math.

As you know, I’m also following the SysML effort.  (If you’re interested in conversing with others regarding SysML, I would like to recommend the SysML Forum Google Group.)  A recent post, directed me to a very interesting paper that I wanted to share with you folks.  It’s called Integrating Models and Simulations of Continuous Dynamics into SysML (PDF) by Thomas Johnson and Christiaan J.J. Paredis of the G.W. Woodruff School of Mechanical Engineering at Georgia Tech. and Roger Burkhart of Deere & Company.  In this paper, the authors show how they expressed a continuous dynamics model and simulation in SysML then provided translation to the Modelica language for simulation.

What appealed to me was not just this specific integration, but the fact that the authors had “shown the way”, in a sense, for SysML to be used as a means for integrating models of different kinds required for systems engineering.  This was a bit of a paradigm shift for me, personally.  I had been looking at SysML from the perspective of “How does it allow me to specify simulatable models of my systems?” and I was having some difficulty in making the mapping.  Now, I’m looking at SysML a little differently.  I’m not sure I can express it, but I think I’m appreciating SysML more for its ability to express a lot of different artifacts of the systems engineering process as a kind of integration repository.

I’ll be the first to admit that I’ve got a lot more thinking to do on this one.  Take a look at the paper and let us know what you think.

Comments are closed.