Machine Learning Driven Chemiluminescence-Based Modelling of Combustion Parameters in Premixed Swirling NH3/H2 Flames - Data
Ammonia (NH₃) is a carbon-free fuel with high energy density and an established global transport infrastructure, making it a promising candidate for clean energy applications. However, ammonia-only combustion faces significant challenges due to its low flame speed, high ignition temperature, and tendency to form harmful nitrogen oxides (NOx). These issues make stable combustion and emissions control difficult in engines and turbines. To address this, researchers are exploring fuel blending strategies, particularly with hydrogen (H₂). This work develops predictive models for estimating equivalence ratio (), ammonia fraction (xNH3) and noxious emissions (NOx) from the flames of turbulent premixed NH3/H2 fuel blend stabilised using a tangential swirl burner. Bayesian Regularisation Artificial Neural Network (BR-ANN) is utilised to estimate both and xNH3 with excited ratios NH*/OH*, violet/OH* and NH2*/NO2* as inputs. NOx was predicted with NO*, OH*, NH*, NH2* and NO2* as input variables.
The developed models operate within the ranges of ammonia-hydrogen blend (0.55xNH30.90), Reynolds numbers (4000Re7000), equivalence ratios (0.601.40), room temperature and atmospheric pressure.
The data for the NOx emissions and chemiluminescence ratios covering the above listed conditions are presented in a .xlsx file. Columns A to C contain the data for ammonia fraction, equivalence ratio, Reynolds number respectively. Columns D to F have the data for NO, NO2 and N2O all in ppmv, dry 15% O2 respectively in them. Columns G, H and I contain the ratios NH*/OH*, violet/OH*, NH2*/NO2* respectively. Finally, columns J, K,L, M and N contain the data for NH2*, OH*, NH*, NO2* and NO*.
