A CFD dataset for supporting machine-learning-based models of hydrogen-ammonia combustion.

This dataset contains the results of 175 Reynolds-Averaged Navier–Stokes (RANS) simulations of ammonia–hydrogen combustion in a staged Stagnation-Point Reverse-Flow (SPRF) combustor, as presented in the paper: Piscopo et al., Burning ammonia–hydrogen mixtures in a staged combustor with high efficiency and low pollutant emissions, International Journal of Hydrogen Energy, 2025. Additional information can be found in the provided ReadMe. Compared to the previous version, this dataset also contains the various components of the velocity and the pressure. 

The simulations cover a wide operating space with variations in:

• Equivalence ratio in the first rich stage (1-1.6)

• Hydrogen molar fraction in the fuel (0-75) [%v]

• Operating pressure (1-3) [bar] values are in the P_175.npy file, in columns. 

The CFD model was built using ANSYS Fluent 22.1 with:

• Realizable k–epsilon turbulence model

• Partially Stirred Reactor (PaSR) combustion model

• Shrestha chemical mechanism for NH3/H2 oxidation

• Discrete Ordinates radiation model with adapted WSGG coefficients

A total of 175 CFD simulations were performed using a combination of:

• Latin Hypercube Sampling

• Boundary sampling

• Adaptive sampling

Outputs provided in this dataset are collected in X_175.npy and include:

• Operating conditions for each simulation

• Thermochemical variable fields (N2, NH2, NO2, H2O, OH, O2, H2, NH3, NO, and T) as well as velocity components and pressure fields. 

Please cite the paper when using this material.