Quantitative Relationship Research Between Hydrocarbons Structure And Octane Number

QSAR are widely used in drugs, biological, chemical, environmental science andmany other disciplines as an important rsearch method in contemporary chemistrybasic theory and application. Octane value is one of the most important indicators ofmeasuring gasoline performance. In the past years, people have been studying therelationship between hydrocarbon structure and corresponding to the the octanenumber in order to improve the combustion performance of gasoline and reduce theharm to the environment. This paper studied the QSAR of paraffin, naphthene andaromatics between quantum chemical parameters, topological index and the octanenumber by multiple linear regression method to provide a theoretical basis forexploring the high antiknock gasoline component.The quantitative paraments of alkane, cyclanes and aromatic hydrocarbon werecalculated by DFT/B3LYP/6-31g**method to acquire the quantum chemicalparameters of structure information, calculate their molecular connectivity index andatomic type electrical topology status index by the topological theory, and studyquantitative structure-activity relationship of its octane number with multiple linearregression method.Using electrotopological state indices for atom type as structure description,andstudying the QSAR of octane numbers of alkane, cyclanes, aromatic hydrocarbon andof all above compounds by multiple linear regression method, and verifying thismodel by combining both internal and external methods. Four ETSIAT forecastingmodels were obtained by the stepwise regression variables selectionfor alkanes, its R2=0.965, Q2LOO=0.942, Q2ext=0.924for RON, its R2=0.955, Q2LOO=0.931, Q2ext=0.897for the MON. Three ETSIAT prediction models were obtained by via stepwiseregression variables selection for naphthene, its R2=0.895for RON, its R2=0.870for MON. For aromatics, Two ETSIAT prediction models were obtained by stepwiseregression variable filter for MON, its R=0.718. Sum for each compound. FiveETSIAT forecasting models were obtained by the stepwise regression variablesselection, its R2=0.828for MON. The research results show that the atom typeelectrical topology status index can be well used in characterizing the octane numbersof alkanes, cyclanes and aromatic hydrocarbon, the models have good stabilityrobustness and external prediction ability.