Assessing the performance of recent detailed NH3 and NH3/H2 combustion mechanisms

There has been growing interest in applying ammonia (NH 3 ) as a fuel. It is often combined with hydrogen (H 2 ), valued for being carbon-free, to compensate for the poor combustion properties of neat NH 3 . Over the last decade, numerous detailed NH₃ combustion kinetic mechanisms have been published, but there is no consensus on the best available mechanism. This study quantitatively evaluates the performance of 36 recently published NH 3 combustion mechanisms based on how well they can reproduce indirect experimental data. The utilized experimental data collection includes shock tube ignition delay times (IDT), laminar burning velocities (LBV), and concentrations measured in flow reactors, jet stirred reactors, and shock tubes. With 17,242 data points in 1327 data series from 110 publications, this is, to the best of our knowledge, the most extensive data collection ever used for NH 3 /H 2 mechanism testing. The uncertainties of the experimental data were carefully evaluated, and they were considered in assessing the performance of the mechanisms. A novel, generally applicable, model-free approach was proposed to propagate the temperature uncertainty to the measured outlet concentrations, resulting in more realistic errors for the concentration data than in previous comprehensive mechanism testing studies. All utilized experimental data are available in the ReSpecTh database ( https://ReSpecTh.hu ). None of the mechanisms had satisfactory performance for all types of experimental data, but the NUIG-2024 mechanism and the Tsinghua-2024c mechanism demonstrated the best overall performance among the investigated models and were selected for further investigations. Missing third-body efficiencies for NH 3 were added to the third-body reactions of both mechanisms, and their H/O and NH 3 pyrolysis subsets were also replaced. The overall performance of NUIG-2024 could not be noticeably improved by these modifications, whereas modified versions of Tsinghua-2024c had slightly improved performance. However, none of the modifications resulted in a mechanism sufficiently accurate for all types of experimental data, which highlights the need for further NH 3 combustion mechanism development.