Numerical Investigation of NOx emissions in a gas turbine burner using hydrogen-ammonia and partially cracked ammonia fuel blends: A combined LES and CRN approach

The increasing interest in ammonia as a carbon-free fuel alternative underscores the need for accurate numerical models capable of predicting the complex combustion chemistry and associated NOx emissions. This study presents a detailed numerical investigation of NOx emissions in a gas turbine burner operating with hydrogen-ammonia and cracked ammonia fuel blends, utilizing Large Eddy Simulation (LES) and Chemical Reactor Network (CRN) methodologies. The primary objective is to validate a Computational Fluid Dynamics (CFD) model against experimental data collected under atmospheric conditions. The experimental campaign involved a non-premixed burner with ammonia concentrations up to 70%, generating a NOx emission database useful for model validation. LES were performed using a tabulated chemistry model, using a detailed chemical kinetic scheme alongside additional transport equations for the main species responsible for the formation of pollutants, to better capture the combustion characteristics and emissions. Furthermore, a CRN model was developed based on time-averaged LES data. This approach facilitated a more detailed and wide examination of NOx formation mechanisms and pathways. The results indicate that both LES and CRN models predict NOx emissions with an accuracy within 10% of experimental measurements, although LES slightly underestimates NOx levels and overestimates outlet temperatures by 3%. The CRN model, derived from LES data, offers a computationally efficient means for analyzing key emission pathways. Furthermore, a comparison was conducted between the combustion characteristics of the hydrogen-ammonia blend and a mixture resulting from an 80% ammonia cracking process. This was achieved through the utilization of both LES and CFD-CRN methodologies, with the objective of analyzing the impact of cracking on NOx emissions, while maintaining a constant burner power and equivalence ratio. In conclusion, the study demonstrates the effectiveness of combining LES and CRN methodologies in predicting NOx emissions and analyzing NO formation pathways from NH₃/H₂/N₂ combustion. The utilization of a cracked-derived mixture resulted in a 25% reduction in NOx emissions. The findings provide valuable insights for optimizing gas turbine operation while addressing NOx emission concerns, contributing to the development of cleaner combustion technologies.