Development and simulation application of a reduced diesel/methane/hydrogen tri-fuel mechanism based on multi-objective optimization and multi-criteria decision-making

Abstract

Hydrogen, as a carbon-neutral fuel, makes diesel/methane/hydrogen tri-fuel blends a promising pathway for decarbonizing diesel engines. The combustion simulation of these engines demands high-precision, compact chemical kinetic mechanisms. To achieve automatic mechanism optimization, this study proposes a framework that combines Non-Dominated Sorting Genetic Algorithm II (NSGA-II) and Entropy Weight-Based Ideal Solution Similarity Sorting Technique (EW-TOPSIS). The framework produced a reconstructed n-dodecane mechanism achieving average absolute relative errors of 15.0 % for ignition delay time (IDT) and 6.9 % for laminar flame speed (LFS) across validation datasets. Coupled with methylcyclohexane and toluene sub-mechanisms, it formed a mechanism comprising 88 species and 443 reactions. The final mechanism in engine simulations across the 10–60 % HAR range, achieving errors within 3 % for peak pressure and 5 % for engine IDT, providing a predictive capability for combustion phasing and emission trends. Finally, kinetic analysis confirmed hydrogen's dual role: inhibiting auto-ignition at low-temperatures while promoting it at high-temperature.