Hello,

I am quite new to modelica, so I tryed to figure out
what corner cases of the modeling should be better
avoided or can lead to undesired results.

I could come up with 2 examples where the solution
would not be unique.

I admit the example is more a mathematical construction than
a physik model of a system.

Real a,b,c,x;
a*x^2+b*x+c = 0;

Quick test with dymola showed one solution, while
skiping the second one. (positiv discriminant)
In the case of negative discriminant the solver stops.

The other case that would lead to 2 solutions is ...
Real x,y;
x + y = f(time);
x = g(time);
y = .. unique

Chaning this into ..
x + abs(y) = f(time);
x = g(time);
y1 = .. the old one
y2 = -y1

Certainly there are more situations where one should
be careful .. and right now I don't have the whole picture.

I would like to learn from experts
What are the common pitfalls and how to avoid them?

Regards, Daniel

Are you asking how to limit the domain so that you get a unique solution?

If so, you could use if-then-else conditional expressions like

y=if v>limit then limit else v;


From memory, both of the modelica books (one by Fritzson the other by Tiller) use the parameterized diode model as an example of a conditional model, which you might find useful.

(not sure this answers your question though?)