Optimization of direct injection parameters for lean-burn hydrogen-ammonia engines using Kriging surrogate model coupled NSGA-II

Lean-burn hydrogen-ammonia engines face challenges of high nitric oxide (NO) emissions and insufficient peak cylinder pressure (Pmax). Research and optimization of fuel injection strategies, particularly fuel composition and injection timing, are crucial for improving engine performance. This study employs a three-dimensional simulation model of a spark-ignition hydrogen-ammonia engine, combined with a Kriging surrogate model and Non-dominated Sorting Genetic Algorithm II, to optimize the fuel injection strategy. The objectives are to reduce NO emissions while increasing Pmax and indicated thermal efficiency (ITE). The optimal condition selected from Pareto solutions using TOPSIS. The results demonstrate that there exists a linear trade-off relationship among low NO emissions, high Pmax, and high ITE in the Pareto optimal solution set. The optimized condition exhibits more concentrated heat release with 1–4 °CA shorter combustion duration than the initial condition. Compared to the initial condition with Pmax of 4.13 MPa, ITE of 46.57 %, and NO emissions of 21.99 g/kW·h, the optimized case demonstrates significant performance improvements: Pmax reaches 5.95 MPa (44.1 % improvement), ITE increases to 47.15 % (1.2 % enhancement), and NO emissions are reduced to 8.07 g/kW·h (63.3 % reduction).