Problems with ModelicaTest.Fluid.TestComponents.Vessels.TestSimpleTank

[casella]

As reported here, ModelicaTest.Fluid.TestComponents.Vessels.TestSimpleTank is failing because the temperature of the fluid in the tank is getting out of range.

I investigated the issue in depth. The fundamental problem is that the base class Modelica.Fluid.Interfaces.PartialLumpedVolume is governed by a dynamic mass and energy balance, with a state variable change to more suitable state variables than mass and internal energy, in this case level and fluid temperature.

It is clear that the energy balance equation becomes singular when the level (hence the mass) becomes zero. Don't even mention what happens when the mass becomes negative (anti-matter anyone?). Some solver could miraculously handle the singular case without producing utter garbage, but I don't think that the MSL should rely on miracles, it should only contain solid, well-posed models that can safely be handle by all MLS-compliant tools.

Bottom line: the tank should never get completely empty.

[casella]

See PR #4527. If successful, will need to be back-ported to maint/v4.1.x