Abstract
A new quantitative structure–property relationship (QSRR) eight parameter correlation (
R
= 0.998) of gas chromatographic retention times for a diverse set of 35 petroleum condensate components was developed by application of multiple linear regression analysis (MLR). The descriptors are all calculated directly from the molecular structure using Dragon software. A boiling point-based model (
R
= 0.999) of a subset of 19 components was used as assistance. The predictive ability of both models was tested for some unknown components. The obtained model was used for interpretation of the retention behaviour of the investigated petroleum components.