A modelling study of acetylene oxidation and pyrolysis

Abstract

This study initiates the gradual upgrade of the DLR reaction database The upgrade plan has two main steps: an optimisation of the C1–C4 oxidation chemistry and a revision of the polyaromatic hydrocarbon (PAH) formation sub-mechanism based thereupon. The present paper reports the main principles applied to model improvements and results obtained for the acetylene (C2H2) oxidation sub-mechanisms. The principle acetylene oxidation reactions have been revised as well as the detailed chemistry of important intermediates, i.e. methylene, ethynyl, vinylperoxy radical and also diacetylene, vinylacetylene and higher diacetylenes, important for PAH formation. The uncertainty intervals of the studied reactions were statistically evaluated, providing general bounds for the performed modifications to reaction rate coefficients. The first stage of the presented update was performed through revision of the thermochemical data and model optimisation on ignition delay data and laminar flame speed data, since they exhibit lower uncertainty in comparison to species profile data. The final model optimisation was obtained through simulations of concentration profiles measured in shock tubes and laminar flames for improvement of the reaction paths and rate coefficients related to acetylene pyrolysis and PAH precursor formation. Approximately 500 data points were analysed. The updated reaction mechanism predicts all simulated experimental data, also not included in the optimisation loop data prom plug flow and jet-stirred reactors, either with good or satisfactory agreement. It was found that the vinylperoxy radical formation and consumption dictate the reaction progress at low temperatures. The performed study clearly determined that acetylene combustion proceeds through the strongly coupled reaction paths of fuel oxidation and PAH precursor formation; the same species are involved in these parallel processes. Therefore, the self-consistent reaction model for acetylene combustion could be obtained only by an optimisation performed on the experimental dataset encompassing both processes.