I am currently studying on the compression/expansion of a phase-change fluid (CO2) in a piston cylinder using Fluent v19.2. The 2D axisymmetric model of the cylinder is as shown below.

Before jumping into the phase-change fluid, I tested my model first using N2 gas from the NIST real gas models. The N2 gas is initiated at P = 20bar and T = 294 K.

I've created a simple profile for the moving piston to simulate the following cycle:
1) Compression: Piston moves to the right, compressing the gas up to ~70 bar

2) Dwell: Piston stays stationary, allowing the compressed gas to cool down to ambient temperature

3) Expansion: Piston moves back to original position, expanding the gas back to ~20 bar

4) Dwell: Expanded gas warms up to ambient temperature

Figure below shows the result for the Pressure-Time graph. The model works beautifully as predicted.

Now going into the complicated part where I have to model the condensation-evaporation of CO2 with the same compression/expansion cycle. As I am new to the multiphase and using real gas model in Fluent, I would really appreciate any inputs for me to model this problem the best way possible. Here are some of my thoughts on this:

1) I am mainly interested in the vapor-liquid saturation of CO2 in the compression/expansion cycle. The model is created in a way that the CO2 will not go into supercritical.

1) From my understanding, is it true that I can't use the cubic equation and NIST real gas models to model the phase-change of CO2 (vapor-liquid)? Therefore the only way is to create my own UDRGM?

2) Is there any way to simplifly this problem without using the UDRGM? Maybe using ideal gas instead if there is no other way.

3) In the multiphase model, would the mixture model with Lee evaporation-condensation model be suitable?

4) Since there's no liquid CO2 in the Fluent database, do I have to create my own?

Thank you.