A G-equation-based multidimensional combustion model incorporating detailed chemical kinetics was developed and implemented in KIVA-3V for spark-ignition (SI) engine simulations. The integrated model simulates the turbulent flame propagation, pollutant formation, flame quenching and knocking combustion processes in SI engines, and was applied to both homogeneous charge and direct-injections (DI) SI engines.
The G-equation method is employed to track the position of the mean turbulent flame front surface. A progress variable concept is introduced into the turbulent flame speed correlation to account for the laminar to turbulent evolution of the spark kernel flame. The laminar flame speed correlation was also updated to take the mixture stratification effect into account. A new method based on tracking the subgrid-scale burnt/unburnt volumes of the flame-containing cells is proposed for the primary heat release calculation. In the post flame zone, detailed hydrocarbon fuel oxidation mechanisms coupled with a reduced NOx mechanism are applied for modeling secondary heat release and NOx formation. The chemical kinetic mechanisms are also applied in front of the flame front for engine knock modeling.
The detailed iso-octane mechanism used to simulate gasoline fuel combustion was validated using ignition delay test data in a shock tube and measured pressure data of a gasoline homogeneous charge compression ignition (HCCI) engine. The SI engines used for model validation include a Caterpillar propane engine, a two-stroke marine GDI engine and a Ford engine operating in both port-fuel-injection (PFI) mode and DI mode. Good agreement between simulated and measured in-cylinder pressure traces and engine-out emissions was obtained over a wide range of operating conditions.
Based on model validation, knocking combustion processes in the Ford engine under boost conditions was modeled. Knock mitigation strategies using cooled EGR and split-injection were assessed based on the numerical study.
As an independent topic, a semi-implicit numerical solver for detailed chemistry was developed aiming at improving computational efficiency. The present solver was successfully applied to diesel engine simulations and 40–70% CPU time savings were achieved compared to the standard CHEMKIN.