[evcelica]

FYI: I already have a workaround to my problem, but am curious about this.

I have a long Liquid Nitrogen line which I analyzing for relief purposes.
I am using a homogeneous binary mixture of N2 and N2Vap from an RGP file, so this is a two phase simulation.
The model starts as a saturated mixture at 160psi with 22% vapor quality (mass fraction).
Heat is added to the fluid due to loss of vacuum: this causes the pressure and temperature to rise, and also the liquid to boil into a gas. A relief valve on the far end of the line is venting pressure at the same time.
After some time, all the liquid has boiled away and the line is completely gas.
After some more time, the gas temperature rises up to the critical temperature of 126.19K, and the solver then fails due to having no properties for liquid above the critical temperature. It Fails because of missing liquid properties even though the mixture is now 100% gas and it shouldn't need liquid properties for anything!
I tried using N2Vap as both the constraint and equilibrium fraction, but get the same results.
My workaround is to pause the model in between the point where I have 100% gas, but before anything reaches critical temperature. I then make a new .def file with just N2Vap as the fluid instead of the two phase mixture. I then resume the simulation using the results of the 2 phase model as initial conditions. This seems to work pretty seamlessly, but it seems a bit odd that this has to be done. Is there any way around this?

[ghorrocks]

It looks like you have found an annoying quirk in CFX. Have you discussed this with ANSYS support? It might be a bug in CFX and then you can get it fixed.

[evcelica]

Hi Glenn,
No, I have not asked ANSYS support about this yet. They are usually pretty slow to respond.
Luckily my workaround was possible this time, but if I had a simulation which contained both a saturated mixture, and supercritical gas, I would not be able to simulate that.
I'll ask support, and see if they have any worthwhile response.