Remove time events in ETS model with heat recovery chiller

[mwetter]

This closes #4058
It also removes a duplicate signal connection.

[mwetter]

@AntoineGautier : Can you please have a brief look into this refactoring. It replaces a FirstOrderHold with 60 s sampling time with a TrueFalseHold and a Pre at its input. The TrueFalseHold avoids the frequent time events and the Pre is needed for OpenModelica to avoid a nonlinear equation in a when construct during the initialization.

[AntoineGautier]

The change looks good to me.
In addition, I suggest increasing the minimum cycle time from 60 s to 5 min. This has minimal impact on the supply temperatures and avoids cycling the two HX pumps every 60 s, which seems a little demanding.
I committed this modification via the last two commits. The comparison results can be downloaded from here:
https://drive.google.com/file/d/1uEiDJpI6ymrWTNUKxW2WFFKu6d5cGJZS/view?usp=drive_link
Extract archive and run python -c "import buildingspy.development.regressiontest as t; t.Tester(tool=\"dymola\").report()" to visualize.

[mwetter]

@AntoineGautier : Thanks, increasing it to 5 min is good.