Separated steady_states

Many of the ThermoSysPro models have a steady_state parameter to define whether to initialize the model:

• Either with der(state) = 0;
• or state = state0 ← user defined state.

Depending on the model, steady_state acts on P, Q, h ...

It would probably be interesting to separately initialize some of these variables. For example, I think that sometimes one may want to ease the initialization by providing the initial h, while still letting the model to evaluate a proper P and Q distribution (der(Q)=0 and der(P)=0).

I also have some examples where setting steady_state=false everywhere may lead to unbalanced system (P and Q @ estremities may be setted twice in adjacient components)

changed milestone to %ThermoSysPro v5.0

added enhancement init labels

See issue about DynamicCondenser (package Water/Steam) (V=V0 and derV= 0 if steady state = true and false)

marked this issue as related to #69

Should be tested (DynamicCondenser could be an example) and also verified (physically) if it is possible to have different conditions on h, P...

assigned to @sarah.hocine-rastic

Hello, If a modification should happen on the initial equation section, then it could be nice to move the "state = state0 ← user defined state." part inside the declaration of the state.

For example (Nb: I am not sure that the param steadystate will be passed correctly) : Modelica.Units.SI.Pressure P(start=P0, fixed = not(steadystate)) "Fluid average pressure";

and Keep in the initial equation section : if steadystate = true der(P) = 0 etc.

Why this proposal? Because when you are using the "Save Start Value" on Dymola. It can produce an extend of the current model with all the start value set as modifier (ex: P=10). If you keep the P=P0 inside the initial equation section, your model will always start from P0. If you set "P(start=P0,..." , with steadystate = false, if you use the normal model, there is no problem if you use the "save start values", the model will start from the value of the modifier (here P=10) instead of P0.

This trick can help to initialize model faster and to produce (with the good manipulation) a working snapshot of the simulation saved in an extends. Regards

mentioned in issue #69

mentioned in merge request !14