Knowledge generation through data research: New validation targets for the refinement of kinetic mechanisms

Abstract

Abstract Chemical structures of low-pressure premixed flames are routinely used as species-resolved validation targets for the development of chemical kinetic combustion mechanisms. Species identities, for example based on flame-sampling mass spectrometry, and species-specific mole fractions as function of distance to the burner are especially important targets. This work, which is based on a compilation of the chemical structures of 55 low-pressure premixed flames from various laboratories, unravels previously unexplored validation targets. It is shown here that such a compiled data set allows for the extraction of fuel-specific chemistry and can be used to identify inconsistent data, thus potentially improving the accuracy of the validation set. Most importantly, this large data set of flame structures was essential to discover that, despite the different fuels and flame conditions used in the individual flame studies, the maximum mole fractions of some intermediate species are correlated to each other in a systematic way across a large number of flames. Some examples, including the correlations between maximum mole fractions of benzene (C6H6) vs. toluene (C6H5CH3) and the correlation between the isomeric allene (CH2CCH2) and propyne (CH3CCH), are discussed and it is shown that these obtained correlations can be used as targets for model refinement.