Multi-objective optimization of an ammonia-cracking process for hydrogen production using NSGA-III: Balancing economy with NO_x and CO₂ emissions

The transition from fossil fuels to clean energy is essential to mitigate air pollution. Ammonia is a promising hydrogen carrier with high hydrogen density and storage stability; however, its decomposition process generates climate impacts and pollutants through energy consumption and wastewater treatment. Moreover, fuel cell-grade hydrogen requires ammonia concentration below 0.1 ppm, which commercial simulators struggle to achieve. This study developed a multi-objective optimization approach integrating artificial intelligence-enhanced simulation with the non-dominated sorting genetic algorithm III (NSGA-III) to simultaneously optimize the levelized cost of hydrogen (LCOH), NO_x emissions, and global warming potential (GWP). CatBoost machine-learning models were combined with multi-objective optimization of reactor length, temperature, and pressure. The AI-enhanced simulation achieved 100% convergence with R2 exceeding 0.973. The optimization yielded 66 nondominated solutions, with 80.3% achieving LCOH below USD 6.0/kg-H₂. Four distinct operational strategies emerged: economy-focused, NO_x-focused, GWP-focused, and compromise. This approach provides quantitative guidance for designing economically viable and environmentally sustainable ammonia cracking processes for industrial hydrogen production and clean-energy transition.