Ammonia blends for gas-turbines: preliminary test and CFD-CRN modelling

Ammonia, with its high hydrogen density, cost effectiveness and ease of storage, is emerging as a potential fuel to meet decarbonisation targets in various sectors. However, the combustion of pure ammonia poses challenges due to its low calorific value and reactivity. Existing gas turbine combustion systems, refined over decades with conventional fuels, struggle to maintain comparable flue gas emission concentrations when burning ammonia. Overcoming this challenge requires a paradigm shift in gas turbine combustion, with implications for chemistry and combustor design methodology. Research into the chemical kinetic mechanism for ammonia combustion worldwide reveals differences in flame radical pools compared to methane combustion. The dominant NOx formation mechanism shifts to a "fuel-bound nitrogen" type, which differs from the "thermal" mechanism in hydrocarbon and hydrogen combustion. The integration of chemical reaction rates into Computational Fluid Dynamics (CFD) models with transport equations results in impractical computational times for industrial applications. Therefore, streamlined approaches such as chemical reaction reduction methods are being developed. This involves the construction of lumped parameter models (Chemical Reactor Network - CRN) to identify key parameters that affect species production. A reliable CRN implementation involves the identification of critical volumes using a preliminary CFD approach (CFD-CRN). This paper analyses preliminary experimental tests by Baker Hughes on an industrial burner using different ammonia, hydrogen, and methane fuel mixtures. The NOx results show a different behavior between methane and hydrogen. While the detailed mechanisms behind these results remain open questions, the CFD-CRN modelling approach is adopted. A proposed CRN is investigated for methane, hydrogen, and ammonia blends, providing insight into key parameters and a reasonable physical explanation for the different NOx results. A single calibration parameter () is used to predict accurately NOx emissions using experimental data available from experimental campaign.