A generalized correlation for the prediction of phase behaviour in supercritical systems

A generalized two parameter model is presented which uses molecular properties of pure fluids to predict interaction parameters in the attractive term of the Peng-Robinson equation of state. The predictive method is based on a consideration of London (dispersive) forces and includes attractive forces between polar and non-polar molecules that result from induction. Polar-polar effects and quantum forces are omitted. A large body of experimental vapor-liquid equilibrium data measured at near and supercritical conditions (that excludes quantum components and polar-polar systems) was used to calibrate the generalized model. Overall bubble point pressure deviations calculated using the proposed generalization were 5.66% for 3240 data points which compares to average deviations of 3.27% obtained by using regressed binary interaction parameters. Average vapor mole fraction deviations were just under 0.01 using both the generalized and the regressed interaction parameters. The sensitivity of predicted phase envelopes to dispersive and inductive term in the generalized correlation is shown graphically for several systems. Comparisons are also made to another recent interaction parameter generalization presented by Nishiumi et al. (1988).