Measurements and Prediction of Sooting Tendencies of Pure Hydrocarbons

The amount of soot formed in combustion devices such as engines depends strongly on the chemical structure of the fuel. Over the past 10 years we have measured quantitative sooting tendencies for hundreds of pure hydrocarbons representative of those in practical fuels. The sooting tendencies have been characterized in terms of Yield Sooting Index (YSI), which is based on the amount of soot formed in a methane/air reference flame when a small amount of the test compound is doped into the fuel. In this new collaborative project we are developing a Quantitative Structure-Activity Relationship (QSAR) model from these results; this model can be used to predict YSI for new hydrocarbons of interest and to understand the details of the underlying chemistry. The model uses molecular descriptors to describe the structure of each compound in the database, and it predicts YSI with a feed-forward artificial neural network trained with the measured values. It has a rigorously defined applicability domain and its performance has been tested by both internal and external validation. New measurements have been performed to extend the applicability domain towards advanced biofuels and for external validation.