Predictive Modelling of NOx and Unburned Hydrogen Emissions in a Direct-Injection Hydrogen SI Engine

Hydrogen is emerging as a compelling energy carrier for future transportation due to its potential to enable fully decarbonised operation and near-zero tailpipe pollutant emissions. Realising this potential in reciprocating internal combustion engines requires a detailed understanding of the complex interactions governing hydrogen combustion and emissions formation. In this context, physics-based emission predictive modelling offers a powerful means to accelerate the development and optimisation of hydrogen-fuelled engines by enabling rapid evaluation of operating strategies without the need for extensive experimental campaigns. This study investigates the simulation of pollutant emissions from a 0.5L single-cylinder research engine equipped with side-mounted direct hydrogen injection and spark ignition. The work focuses on the development and assessment of predictive models for nitrogen oxides (NOx) and unburned hydrogen (uH₂) emissions within a 1D–0D simulation framework. For NOx prediction, a simplified kinetic mechanism is coupled with both a 0D two-zone combustion model and a multi-zone in-cylinder representation, enabling assessment of the need to account for mixture inhomogeneity for accurate prediction. For uH₂ emissions, phenomenological submodels describing wall quenching processes and top-land crevice flow are implemented and calibrated to capture the dominant sources of hydrogen escape during combustion. The models are validated against an extensive experimental dataset spanning a wide range of engine conditions, including variations in indicated mean effective pressure, engine speed, relative air–fuel ratio from stoichiometric to ultra-lean combustion, and dilution via exhaust gas recirculation (EGR). The comparison highlights the models' ability to reproduce observed physical trends across different engine operating conditions. Overall, the study provides insights into the key modelling requirements for accurate pollutant simulation in hydrogen engines and supports the broader use of predictive tools to guide the design of clean hydrogen combustion systems.