Numerical investigation of the NOX emissions of a perfectly premixed NH3-H2 flame at moderate pressure levels

• A novel skeletal chemical kinetic scheme for the NH3-H2 combustion has been presented. • The change in the flame shape induced by the pressure change has been optically measured. • The TFM have been validated for the prediction of NH3-H2 flame topology. • The TFM has been validated for the estimation of the emissions of NH3-H2 mixtures. • The role of the differential diffusion on NO formation has been discussed. Blends of green hydrogen-ammonia for Gas Turbines (GT) are gaining more and more interest, as their carbon neutrality can abate CO 2 emissions. Such fuel mixture could be particularly effective since the low reactivity of ammonia can compensate the aggressive properties of hydrogen in terms of both flame speed and low heating value, reducing the re-design effort of the traditional dry-low emission combustors. From the environmental perspective, the main drawback is represented by the NO x emissions, mostly related to the fuel-bound pathway activated by the cracking of NH 3 : its mitigation requires the proper control of the equivalence ratio of the blend along the selected ratio of ammonia and hydrogen in the fuel mixture. Additionally, at relevant GT conditions, a moderate benefit for NO x reduction can be played by the pressure rise. In this work, the effect of the operating pressure onto the NO x emission is investigated numerically in the range 1.1–2 bar along a perfectly premixed mixture of 25 % NH 3 – 75 % vol. H 2 at constant equivalence ratio, leveraging the corresponding experimental data. The tests employ a radial swirler whose performances are measured not only in terms of NO x but also through detailed flame imaging. The latter is used to evaluate the accuracy of a Thickened Flame Model (TFM) in predicting the flame shape and position. The TFM is based on a skeletal mechanism consisting of 27 species that embeds the OH* to directly compare the numerical line of sight with the images from the experiment and the NO x chemistry as well. Regarding the NO x , the numerical results show a reasonable accuracy: not only can the overall flame length be captured but also a quantitative estimation can be retrieved from the numerical model. Lastly, the impact onto NO x of the hydrogen preferential diffusion related to the flame curvature is discussed.