I agree with acar,
if you plan to utilize Prode Properties multiphase flash operations with solid models
you should do data regression for best results,
differently you get the equilibrium values based on VLE BIPs