A Predictive Extended Corresponding States Model for Pure and Mixed Refrigerants

Vle have developed a predictive corresponding states model for the thermo· physical properties of pure refrigerants and refrigerant mixtures. The bulk phase properties such as the density, enthalpy and entropy are predicted using the prin· ciple of extended corresponding states. This theoretically based model uses shape factors to insure conformality among the various components and R134a is used as the reference fluid. The shape factors are found by mapping saturation boundaries of the fluids of interest onto the reference fluid. In the case where no saturation data are available, the shape factors are predicted. In addition to equilibrium properties, a one-fluid corresponding states model is used to predict the viscosity and thermal conductivity. Phase equilibria is predicted using a Peng-Robinson equation of state. The corresponding states model is contained in an interactive microcomputer program, PROZPER, (PRoperties of OZone Protecting Environmentally acceptable Refrigerants). The program performs various vapor-liquid equilibria calculations-dew point and bubble point pressures and temperatures, isothermal, isenthalpic and isentropic flashes-and gives the bulk phase properties (such as density, enthalpy, entropy, Joule-Thomson coefficient, Cp, Cp/Cv, sound speed, viscosity and thermal conductivity) of the coexisting phases. Calculations may be performed on pure fluids or on mixtures of up to 20 components. A built-in database contains many common refrigerants including fully-halogenated CFC's (for example Rll, Rl14, R12 and R13) as well as environmentally acceptable alter· natives (such as R123, R134a and others). In addition, the program can "learn" new components using only a minimum amount of information-the molecular weight, the normal boiling point and the critical parameters of the chemical species. Additional information such as vapor pressures, saturated liquid densities, saturated liquid thermal conductivities and viscosities may be input to improve the predictive capability of the model.