Start values lost in big power plant initialization

[casella]

When trying to simulate FlexiCaL.Plants.Test.TestPlantDeltaOffDesign with -d=evaluateAllParameters, I get a solver failure at the first iteration of the big nonlinear system of equations for lambda = 0.

This is most likely due to wrong start values, since the model is successfully initialized by Dymola. Some analysis reveals that the two most probable causes are

• loss of start values for fixed = false parameters
• loss of start values for some enthalpy variables.

The former issue is addressed in see ticket:5807#comment:9.

The latter issue is reported here. For example, the flattened model contains

Real Plant.Carbonator.SH.htilde[10](
  quantity = "SpecificEnergy", unit = "J/kg",
  min = -10000000000.0, max = 10000000000.0,
  start = 3331000.0, nominal = 500000.0,
  stateSelect = StateSelect.prefer) "Enthalpy state variables";

but then the nonlinear solver diagnostic message reports

proper start-values for some of the following iteration variables might help
...
[306] Real Plant.CalcinerSection.a_SH2_1.fluidFlow.htilde[4](start=100000, nominal=500000)
...

with the start value of 100000 being the default start value for the type used to define htilde.

[casella]

See #5813 for a discussion of the root cause of the issue