Modelling and Optimization of Narrow-Throat Pre-chamber Engines

This chapter summarizes the modelling effort to unravel pre-chamber combustion developed under the FUELCOM 3 project at KAUST. CONVERGE™ CFD solver was utilized to understand combustion in a narrow-throat pre-chamber. Extensive assessment of geometries (pre-chamber, piston, and their correlation), combustion models, and semi-empirical flame speed correlations were performed. In most cases, the pre-chamber was fuelled (active operation mode), which allows to regulate the pre-chamber composition, while the main chamber was operated with excess air–fuel ratio (lean conditions). The model was successfully validated against engine experiments (both metal and optical engines). Aiming at a drop-in design in existing diesel engines, the pre-chamber features a narrow and long channel (called throat). Two combustion models, G-equation and the multi-zone well-stirred reactor (MZ-WSR) were assessed and later compared; when using the former model, both laminar and turbulent flame speeds were from a look-up table and Peters’ correlation, respectively. The modelling results are well-aligned with the experiments and reveal that the narrow-throat and the jet-piston interaction have a significant influence on combustion and flow development. Towards fundamental combustion aspects, the accurate prediction of the laminar flame speed of lean charges was found to be critical and more relevant than empirical turbulence corrections for high Karlovitz regimes. To connect the fundamentals of combustion and practically relevant engine metrics, the Borghi-Peters diagram provided additional insights into the turbulent combustion regimes observed in pre-chamber engines. Finally, the pre-chamber optimization using machine learning is discussed.