remove stateSelect.prefer for T

[mwetter]

This change is to test the computing time without the annotation

T(stateSelect=if preferredMediumStates then StateSelect.prefer else StateSelect.default)

@Mathadon reports that this gives smaller equations and faster computing time for
Buildings.Examples.VAVReheat.

[mwetter]

Removing stateSelect.prefer for T gives slightly faster simulation.

[Mathadon]

Another illustration of the effect of StateSelect:
If you run Buildings.Fluid.Storage.Examples.Stratified and enable Output information when differentiating for state reduction you get:

Differentiated the equation
tanEnh.vol[8].dynBal.U = 298.67580000000004*tanEnh.vol[8].dynBal.medium.u;

giving
der(tanEnh.vol[8].dynBal.U) = 298.67580000000004*der(tanEnh.vol[8].dynBal.medium.u);

Differentiated the equation
tanEnh.vol[9].dynBal.medium.u = 4184*(tanEnh.vol[9].dynBal.medium.T-273.15);

giving
der(tanEnh.vol[9].dynBal.medium.u) = 4184*der(tanEnh.vol[9].dynBal.medium.T);

for each mixing volume.

These differentiations disappear when using StateSelect.default for T and preferredMediumStates = false in ConservationEquation(medium). The model seems to become marginally faster and Number of F-evaluations : 912 is reduced by five.

[mwetter]

@Mathadon : I run a few more tests and think that we should indeed remove stateSelect.prefer for T as it leads to smaller models.
In my tests, I was interested to see what the estimated operation count is that Dymola computes. I chose for sequential execution, as the parallel execution depends on the number of cores, and probably does not give much more insight.

I compiled the big models of the Buildings library using

openModel("package.mo");
Advanced.ParallelizeCode=true;
Advanced.TranslationInCommandLog=true;
translateModel("Buildings.Examples.ChillerPlant.DataCenterContinuousTimeControl");
translateModel("Buildings.Examples.DualFanDualDuct.ClosedLoop");
translateModel("Buildings.Examples.HydronicHeating.TwoRoomsWithStorage");
translateModel("Buildings.Examples.VAVReheat.ClosedLoop");

Putting the outputs in one file an running

grep '#\|transla\|after manipulation of the nonlinear\|sequential exe' statistics.log

gives the following summary

#################################################################
# master
#################################################################
translateModel("Buildings.Examples.ChillerPlant.DataCenterContinuousTimeControl");
  Sizes after manipulation of the nonlinear systems: {1, 0, 3, 1}
    Sizes after manipulation of the nonlinear systems: {1, 3, 1, 1, 17}
Estimated operation count for sequential execution: 17339
Estimated operation count for sequential execution: 39699
translateModel("Buildings.Examples.DualFanDualDuct.ClosedLoop");
  Sizes after manipulation of the nonlinear systems: {1, 1, 1, 1, 1, 1, 1, 1, 1, 8, 0, 0, 0, 0, 2, 1, 1, 1, 1, 1, 1, 14, 14, 14, 14}
    Sizes after manipulation of the nonlinear systems: {1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 8, 14, 14, 14, 14, 1, 1, 1}
Estimated operation count for sequential execution: 17837
Estimated operation count for sequential execution: 254814
translateModel("Buildings.Examples.HydronicHeating.TwoRoomsWithStorage");
  Sizes after manipulation of the nonlinear systems: {1, 1, 1, 2, 1, 1, 1, 19, 19}
    Sizes after manipulation of the nonlinear systems: {5, 5, 3, 1, 1, 1, 1, 19, 19, 1}
Estimated operation count for sequential execution: 17349
Estimated operation count for sequential execution: 114789
translateModel("Buildings.Examples.VAVReheat.ClosedLoop");
  Sizes after manipulation of the nonlinear systems: {1, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 14, 14, 14, 14, 8, 1, 1, 1, 1}
    Sizes after manipulation of the nonlinear systems: {2, 1, 1, 1, 1, 14, 14, 14, 8, 14, 1, 1, 1, 1, 1, 1, 1}
Estimated operation count for sequential execution: 17837
Estimated operation count for sequential execution: 162588
#################################################################
# issue354_stateSelectT
#################################################################
translateModel("Buildings.Examples.ChillerPlant.DataCenterContinuousTimeControl");
  Sizes after manipulation of the nonlinear systems: {1, 0, 3, 1}
    Sizes after manipulation of the nonlinear systems: {1, 3, 1, 1, 17}
Estimated operation count for sequential execution: 17339
Estimated operation count for sequential execution: 39690
translateModel("Buildings.Examples.DualFanDualDuct.ClosedLoop");
  Sizes after manipulation of the nonlinear systems: {1, 1, 1, 1, 1, 1, 1, 1, 1, 8, 0, 0, 0, 0, 2, 1, 1, 1, 1, 1, 1, 14, 14, 14, 14}
    Sizes after manipulation of the nonlinear systems: {1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 8, 14, 14, 14, 14, 1, 1, 1}
Estimated operation count for sequential execution: 18337
Estimated operation count for sequential execution: 250480
translateModel("Buildings.Examples.HydronicHeating.TwoRoomsWithStorage");
  Sizes after manipulation of the nonlinear systems: {1, 1, 1, 2, 1, 1, 1, 19, 19}
    Sizes after manipulation of the nonlinear systems: {5, 5, 1, 3, 1, 1, 1, 1, 19, 19}
Estimated operation count for sequential execution: 17349
Estimated operation count for sequential execution: 102929
translateModel("Buildings.Examples.VAVReheat.ClosedLoop");
  Sizes after manipulation of the nonlinear systems: {1, 8, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 2, 1, 1, 1, 1, 1, 14, 14, 14, 14}
    Sizes after manipulation of the nonlinear systems: {2, 1, 1, 1, 1, 14, 14, 1, 1, 8, 14, 1, 1, 1, 14, 1, 1}
Estimated operation count for sequential execution: 18337
Estimated operation count for sequential execution: 159172

Hence, the branch issue354_stateSelectT (on the Buildings repository) gives the better performance. Hence, I suggest to remove stateSelect.prefer for T.

I added #164 for this.
Could you please review the changes.

[mwetter]

Change accepted, process pull request.