期刊名称：Computational Methods in Science and Technology
印刷版ISSN：1505-0602
出版年度：2016
卷号：22
期号：3
页码：133-141
DOI：10.12921/cmst.2016.0000022
出版社：Poznan Supercomputing and Networking Center
摘要：The quantitative structure-property relationship (QSPR) method is used to develop the correlation between structures of crude oil hydrocarbons and their physical properties. In this study, we used VolSurf descriptors for QSPR modeling of the boiling point, Henry law constant and water solubility of eighty crude oil hydrocarbons. A subset of the calculated descriptors selected using stepwise regression (SR) was used in the QSPR model development. Multivariate linear regressions (MLR) are utilized to construct the linear models. The prediction results agree well with the experimental values of these properties. The comparison results indicate the superiority of the presented models and reveal that it can be effectively used to predict the boiling point, Henry law constant and water solubility values of crude oil hydrocarbons from the molecular structures alone. The stability and predictivity of the proposed models were validated using internal validation (leave one out and leave many out) and external validation. Application of the developed models to test a set of 16 compounds demonstrates that the new models are reliable with good predictive accuracy and simple formulation.
其他关键词：boiling point, water solubility, Henry’s law constant, crude oil hydrocarbons, volsurf descriptors
