Phd in ENERGY AND ENVIRONMENT

Thesis title: Advancing NOx Emission Prediction for Hydrogen-Rich and Ammonia-Based Fuel Blends in Gas Turbine Burners

The global push toward decarbonization has catalysed research into alternative fuel sources for power generation. Among these, hydrogen (H2) and ammonia (NH3) are emerging as viable candidates for sustainable energy systems due to their high energy densities and carbon-neutral nature. However, their combustion characteristics, instabilities and significant nitrogen oxides (NOx) production, present challenges for practical application in industrial gas turbines. To date, the prediction of NOx emissions from the combustion of these alternative fuels remains a significant challenge that requires further analysis and resolution. The use of chemical kinetics models in conjunction with combustion models has been demonstrated to lack effectiveness and accuracy in validating flame morphology and predicting emissions when operating parameters such as equivalence ratio or hydrogen content in fuel mixture are varied, particularly at high pressures. This presents a significant challenge in the implementation of these fuels in industrial gas turbines. This work includes extensive experimental campaigns on ammonia-hydrogen, ammonia-methane and cracked ammonia (NH3/H2/N2) blends, focusing on the effects of ammonia fuel fraction, equivalence ratio and pressure on combustion stability and pollutant formation. In addition, the potential of chemiluminescence to develop non-intrusive sensors for the monitoring and control of turbulent ammonia–hydrogen flames is investigated using Machine learning-based methods, to refine diagnostic techniques for real-time NOx prediction. Based on the experimental data obtained, this thesis investigates Computational Fluid Dynamics (CFD) methodologies to predict NOx emissions in gas turbine burners using hydrogen and ammonia fuel mixtures. In particular, numerical methodologies, such as Reynolds-Averaged Navier-Stokes (RANS), hybrid CFD - Chemical Reactor Network (CRN) techniques and Large Eddy Simulation (LES), are explored to model flame behavior and emissions pathways. By combining experimental data, CFD modeling, and machine learning techniques, the study provides a comprehensive evaluation of ammonia and hydrogen combustion processes, particularly in swirl burners. It addresses critical aspects such as combustion fundamentals, flame morphology and dynamics, chemical kinetics and NOx formation pathways to achieve emission mitigation. Finally, this work also explores the thermoacoustic instabilities associated with the mixtures investigated. The results reveal useful insights by validating numerical models on experimental data analysing both laboratory-scale burners and industrial burners, highlighting optimal configurations to reduce NOx emissions while maintaining combustion efficiency.