Optimization strategy for hydrogen/ammonia/diesel blended fuels using ELM-MOPSO-RSR with respect to ammonia/hydrogen energy substitution rate and injection timing

Ternary ammonia/hydrogen/diesel blended fuels represent a compelling alternative to conventional fuels. However, ammonia-based fuels face inherent challenges such as low laminar flame speeds, narrow flammability ranges, and poor combustion stability. This study developed a three-dimensional CONVERGE simulation model to investigate the effects of the ammonia energy substitution rate (AESR), hydrogen energy substitution rate (HESR), and injection timing on the combustion and emissions of the blended fuel. The results indicate that increasing the AESR reduces the heat release rate (HRR), cylinder pressure (CP), NO x , and CO 2 emissions, while increasing unburned ammonia emissions (UAE). Conversely, the addition of hydrogen yields a significant improvement in combustion efficiency, alongside an increase in NO x emissions and a substantial reduction in UAE. Based on these findings, a multi-objective optimization model integrating an extreme learning machine (ELM), multi-objective particle swarm optimization (MOPSO), and the rank sum ratio (RSR) method was developed. The integrated framework identified a statistically optimal compromise strategy within the evaluated parameter space (AESR: 40.299%, HESR: 8%, injection timing: 12 °CA BTDC), which reduced UAE by 62.413%. This study provides a robust theoretical basis for optimizing the combustion performance and emissions of ternary blended fuels.