Add `isCompressible()` method to fluid properties

[joshuahansel]

Reason

Some equations of state do not make sense to be used in certain flow models. For instance, incompressible equations of state do not make sense to be used with flow models where the density and specific internal energy are solved independently, like the Euler equations.

Design

This would add a method isCompressible() to SinglePhaseFluidProperties. Then flow models can query this method if they need to.

Impact

New virtual method in SinglePhaseFluidProperties.

[joshuahansel]

@GiudGiud

[GiudGiud]

This would be cleaner than erroring on missing T_from_v_e

which is problematic because we (my fault mostly) have been adding many of these T_from_v_e, as long as it was analytically tractable to do so

[joshuahansel]

This also suggests that maybe there's a problem with the design of SinglePhaseFluidProperties. Maybe it should be something like

SinglePhaseFluidPropertiesBase->CompressibleSinglePhaseFluidProperties
SinglePhaseFluidPropertiesBase->IncompressibleSinglePhaseFluidProperties

[GiudGiud]

could be SinglePhaseFluidPropertiesBase->CompressibleSinglePhaseFluidProperties->IncompressibleSinglePhaseFluidProperties

I think the compressible ones work in incompressible solvers as long as the mach number is not too high

[joshuahansel]

Incompressible flow models always confuse me. So there, "incompressible" means more than actually being incompressible - it means constant density, right? But it doesn't require the equation of state to have a constant density?

[GiudGiud]

yeah it depends who you talk to. In my mind now, incompressible means rho(T). not rho constant

as in you cannot compress it, with pressure, or maybe compressure it

[lindsayad]

I do agree that incompressible could be a sub-class of compressible

[joshuahansel]

After more thought, I think the current design plus the method in the description is the right way to go.

[joshuahansel]

The ideal solution is something like:

FluidBlah : public FlowModelACompatible, public FlowModelBCompatible

where FlowModelACompatible would have pure virtual methods for all of the interfaces needed by the flow model A. I don't know off the top of my head the rules of multiple inheritance of virtual methods, since there's overlapping interface needs. If the overlapping works fine, it'd be a nice way to enforce compatibility.

[GiudGiud]

There's definitely overlapping interface needs. diamond inheritance isnt well supported

[joshuahansel]

Yeah, I thought as much. In any case, what we have now is fine. Just need to add isCompressible().