Identification of well-parameterised reaction steps in detailed combustion mechanisms – A case study of ammonia/air flames
András Gy. Szanthoffer, Máté Papp, Tamás Turányi
Abstract
• A new algorithm is proposed for developing better detailed combustion mechanisms. • It can be applied using readily available data from the ReSpecTh website. • Creation of “mosaic mechanisms” from existing mechanisms. • Has been successfully applied to a large amount of experimental NH 3 /air LBV data. • The best mosaic mechanism reproduces these data better than the previous mechanisms. Testing detailed combustion mechanisms typically concludes that some mechanisms reproduce the experimental data well at most conditions but are inaccurate at other conditions. However, other mechanisms may perform well under these conditions. A better mechanism (“mosaic mechanism”) may be obtained by identifying the overall best-performing mechanism and loaning the most important reaction steps and their rate parameters from another mechanism with good performance at the conditions where the overall best model is ill-performing. A new algorithm based on this approach is presented here, which is successfully applied using a comprehensive collection of NH 3 /air laminar burning velocity data (348 data points in 61 data series) and eight recent detailed NH 3 combustion mechanisms. The suggested new mosaic mechanism is an improved version of the CEU-2022 mechanism and provides a better reproduction of the utilised data than the previously published mechanisms. The proposed algorithm can be applied to any chemical kinetics system and any other types of experiments. All data needed to apply the algorithm to various combustion systems are already available or can be generated with minimal human effort using the experimental data files, mechanisms, and codes available on the ReSpecTh ( https://ReSpecTh.hu ) website.