Assessment of disparate strategies for octane prediction

Includes bibliographical references.Octane quality is a key factor in determining the profitability in a modern refinery. The final commercial product is defined by the combined blend of various gasoline component streams which are produced from different units within the refinery. The accurate prediction of the octane numbers of these blends enables the economic optimization of the production process. Currently, empirical octane models are used exclusively for this purpose. Octane is a measure of the spontaneous autoignition propensity of a fuel-air mixture and it is quantified using a specific engine-based test method. This research project was founded on the premise that an improved octane prediction model could be harvested from building blocks that included a fundamental understanding of autoignition, appropriate choices of autoignition models and an engine model. This objective was pursued in this work by investigating detailed and reduced kinetic mechanisms for the oxidation of selected fuel molecules using various modeling techniques. Empirical octane models and semi-chemical models of autoignition were also investigated. All of these methodologies were assessed as possible strategies towards octane prediction. In this study it was observed that both detailed and highly reduced kinetic models could describe the oxidation behaviour of pure fuel components and predict their subsequent ignition delays