Development of a Global Mechanism for Ammonia Combustion Assisted by Machine Learning

To achieve carbon neutrality, sustainable energy solutions like ammonia are being explored. Ammonia, with its high energy density and existing infrastructure, is promising as a hydrogen carrier. This study aims to reduce the computational cost of ammonia combustion simulations by developing a global mechanism using a chemical reaction neural network (CRNN). The CRNN was trained using data from the detailed mechanism by Okafor, resulting in a global mechanism with 4 reactions and 7 species. Performance evaluations showed that the developed global mechanism accurately predicts temperature, species concentrations and ignition delay times, achieving high prediction accuracy for steady-state conditions. The results indicate that the CRNN-based global mechanism significantly enhances computational efficiency and accessibility for ammonia combustion research.