Discrepancy in results mp1.py when calculating g via state() vs using g().

[jranalli]

Some of the mp1 tests are failing when calculating g using the subst.state() function, but do fine when calculating subst.g(). The value returned by g() is consistent with the theoretical h-Ts value.

It happens to be for the test case in the multiphase region. I can see that you note in there # Gibbs energy is constant across an equilibrium phase transition. I think it would just boil down to a decision about how you'd really want that value to be produced. I have no opinion, because I am not well versed on the underlying theory here. The only thing I would have strong feelings about is that subst.state()['g'] should match subst.g().

Also of note this only fails for R-134 in the test code but passes for water, but that's only because of level of the tolerances. There is a small variance in the actual water data as well. In both cases the discrepancy is quite small though.

Code to demonstrate:

import pyromat as pm
subst = pm.get('mp.C2H2F4')
state = subst.state(p=5, x=0.5)
state['g']
# -89.32931659
state['h']-state['T']*state['s']
# -89.30886541
subst.g(p=5, x=0.5)
# -89.30886541