Hi,
I have had the same problem and spent much time trying to come around it.
Manual discretisation of the problem finally led to models that the solver was not able to solve at t=0 (note that the equations were not that linear...).
I finally opted for external functions, with clean runge-kutta-like methods, and will soon try to use the external functions in Modelica.
JB

Hello,
I have thought of that problem for some time, and am looking for a few hints.
I know how to model a batch process, that is time dependant; for example: Progress=f(time, temperature, ...) The end-time progress is of main interest. Now, I would like to model an entire plant, where every worshop is working as a continuous process. The variables are not (necessarily) time dependant, but location dependant; and for every workshop, mainly the exit properties are of interest, to connect them to the next workshop.
I have thought of a few way to come around that issue, but I do not know if they can be done in Modelica, or I am not really satisfied of them:
- 'Hand-discretise' the model, along the space variable, and hand-integer the variables. From an algorithm point of view, it works quite well, but the solver tend to fail much to often to my linking (and inevitably at zero time).
- Make a time dependant model in Modelica, externalise it one way or another, and call for results in the plant model. I do not know if this can be done.
- Make a time dependant model in another langage (C), and call for end-time results in the plant model in Modelica. This is quite deceptive a solution when I decided to work under Modelica.
I would really appreciate a few hints or advice, if someone who has already come upon that kind of problem and would share a few ideas.
Thanks,
JB.

Hello,
I am no great specialist, but this syntax worked for me:

    connect(Vap_in,exchanger[1].Vap_in);
    connect(exchanger[end].Vap_out,Vap_out);
    connect(LN_in,exchanger[1].LN_in);
    connect(exchanger[end].LN_out,LN_out);
  for i in 1: (N-1) loop
    connect(exchanger[i].Vap_out,exchanger[i+1].Vap_in);
    connect(exchanger[i].LN_out,exchanger[i+1].LN_in);
  end for;

Good luck,
JB.

Yes, it is working now !
Thanks for the correction, working with OMEdit will be possible again.
JB.

Hello,
Is it possible to declare a new enumeration as being a supertype of an existing enumeration ? So that when the base enumeration is modified, the second takes it into account.
I can't find a syntax that would work, nor any help on the subject.
Thanks,
JB.

For example:
type BaseEnumeration=enumeration(base1,base2,base3);
type CompletedEnumeration=enumeration(base1,base2,base3,comp1,comp2);  //as =BaseEnumeration added of comp1,comp2

Hello,
I'm new with OMEdit, but I have not been able to resolve this:
I create a new package, save it, close it (no error message, everything seems to be working fine) and close OMEdit. I open the package again, only heading and closing remains.
It is the same problem with class definition in models, the class definitions disappear but not the rest...
It looks like this bug: Bug [# 1404] that has been resolved recently as WORKSFORME.
I work under Window XP, I don't know if it helps.
Thanks.


for example, I save:

package Package0
record Rec1
  Real A1;
  Real A2;
end Rec1;
record Rec2
  Real B1;
  Real B2;
end Rec2;
end Package0;

When I open it again, only remains:

package Package0
end Package0;