Novel method for prediction of normal boiling point and enthalpy of vaporization at normal boiling point of pure refrigerants: A QSPR approach

In the present study, new quantitative structure–property relationships (QSPR) were presented to predict the normal boiling point (Tb), and enthalpy of vaporization of pure refrigerants at Tb (ΔHvb). For developing these models, the experimental data of Tb for 192 pure components and the experimental data of ΔHvb for 180 pure components were used. For each component, 1650 molecular descriptors were determined. Enhanced replacement method (ERM), as an effective tool for subset variable selection, was used. The obtained models are multivariate linear equations with five parameters for prediction of Tb and six parameters for prediction of ΔHvb. The parameters of models are calculated only from chemical structure of refrigerants. The average absolute relative deviation (AARD, %) and squared correlation coefficient (R2) of the obtained models over all experimental data are 3.42%, and 0.95 for prediction of Tb, and 6.83%, and 0.96, for prediction of ΔHvb, respectively.