Thermally constrained hydrogen utilization in compression-ignition engines via reduced-order digital-twin Pareto calibration

Hydrogen enrichment can improve compression-ignition (CI) engine fuel-conversion performance, but its practical use is constrained by pressure-rise, peak-pressure, emissions, and component-temperature limits. This study develops a thermally constrained reduced-order digital-twin framework for identifying safe-beneficial hydrogen utilization in a mildly boosted direct-injection CI engine. A crank-angle thermo-combustion model with hydrogen-dependent burn phasing and duration modifiers is coupled to a transient six-node thermal network. Four calibration variables—injection-timing shift, air–fuel ratio (AFR), hydrogen blend fraction, and base load—are sampled at 1600, 2000, and 2400 rpm to train speed-specific surrogates for brake thermal efficiency, brake specific fuel consumption (BSFC), nitrogen oxides (NO x ) proxy, peak pressure, peak wall temperature, and peak pressure-rise rate. Feasibility is imposed using limits of 10.8 MPa, 4.9 bar/deg, and 660 K for peak pressure, pressure-rise rate, and peak wall temperature, respectively. Principal thermo-performance surrogates achieved R 2 = 0.992–0.998, while pressure-rise-rate and gated NO x models achieved R 2 = 0.972–0.975 and 0.945–0.976, respectively. Hydrogen enrichment increased brake thermal efficiency and reduced BSFC, but intensified NO x tendency, pressure-rise, and wall-temperature loading, especially under richer/high-load operation. The safe-beneficial envelope expanded under leaner operation and compressed rapidly at lower AFR and high load. The selected 2000 rpm balanced point was recovered by the full model with brake thermal efficiency of 0.436, BSFC of 196 g/kWh, peak pressure of 9.70 MPa, pressure-rise rate of 4.05 bar/deg, and peak wall temperature of 597 K, while retaining positive local robustness margins. The framework supports thermally constrained hydrogen-assisted CI pre-calibration.