CFD Simulation of Combustion in an Optically Accessible Hydrogen Engine: Comparison between Lean and Diluted Stoichiometric Operations

The reduction of CO2 emissions from anthropogenic activities is pushing the green energy transition, prompting the search for alternative and more environmental-friendly solutions compared to traditional technologies based on fossil fuels. One of the most affected sectors is transportation, which is undergoing a significant change to increase sustainability. To achieve this goal, development of hybrid and electric propulsion systems has taken hold over the past decade, but electrification is proceeding slower than expected due to many challenges related to charging infrastructure, cars range and cost, thus pushing the European automotive sector into a potential crisis. To reverse this trend and simultaneously accelerate the transition to sustainable transportation, further development of ICEs technology aimed at enhancing efficiency when using alternative fuels like hydrogen, is staging a comeback. Thanks to the possibility to retrofit existing units, benefiting from a strong know-how and a developed supply chain, hydrogen fuelled ICEs are an attractive mid-term solution. The application of reliable numerical models can support the retrofitting, providing a deep insight into in-cylinder processes such as injection and combustion, enabling virtual optimizations and reducing development costs and time to market. Comprehensive experimental campaigns on research engines are necessary to develop and validate advanced numerical models, identifying potential weaknesses in the modelling approach. In the current work, experimental results on the Darmstadt optical-accessible engine are applied to validate a 3D CFD framework for the simulation of hydrogen combustion in ultra-lean and stoichiometric ultra-diluted conditions. A detailed analysis of the experienced combustion regime is provided to address the validity of the adopted modelling assumptions as well as the possible occurrence of flame instabilities. This lays the foundation for future model development. The modelling of quenching and a detailed representation of piston crevices contribute to improved accuracy in both operating conditions.