A new QSPR-based prediction model for biofuel lubricity

摘要

A new quantitative structure-property relationships (QSPR) based prediction method for fuel lubricity is proposed to enable fast screening of potential biofuels. The lubricity of 75 substances is experimentally determined with a High-Frequency Reciprocating Rig (HFRR). The list of substances includes potential biofuels and fluids with similar molecular structures. Molecular descriptors for the model are selected by the elastic net regression method. The model predictions enable the discrimination of good and insufficient lubricants.