Thermodynamic efficiency, emissions, and economic performance optimization of dual-fuel (diesel/methanol) HCCI engines: A multi-objective approach using detailed reaction mechanism and intelligent modeling

• Modeled dual-fuel HCCI engine with 27,829-reaction detailed mechanism. • Developed Deep ANN to reduce simulation time while maintaining accuracy. • Used NSGA-II and GA to optimize efficiency, cost, and multi-pollutant emissions. • Identified 195 optimal engine setups balancing thermodynamics and emissions. • Revealed strong link between carbon emissions and work production cost. We develop a dual-fuel HCCI engine (diesel/methanol) based on a validated detailed kinetic mechanism (27,829 reactions, 621 species) and couple it to a deep artificial neural network surrogate for rapid multi-objective optimization. The surrogate (trained on zero-D OpenSMOKE++ simulations) predicts thermodynamic efficiency, work cost, and emissions (UHC, N-based, C-based, VOC) as functions of speed, compression ratio, equivalence ratio, intake state, and fuel split. The Pareto Front is mapped using NSGA-II, delineating trade-offs among efficiency, economic, and environmental objectives and identifying 195 Pareto-optimal operating points. Single-objective optima include 0.537 thermodynamic efficiency and emissions per cycle down to 2.92 × 10 -12 g (UHC), 6.51 × 10 -13 g (N-based), 1.12 × 10 -11 g (C-based), and 1.23 × 10 -14 g (VOC). Results demonstrate that fuel composition and equivalence ratio dominate the trade-space and optimizing them yields large simultaneous gains. The approach gives fast and accurate design guidance for large-scale HCCI over diverse potential policy and cost assumptions.