Experimental and numerical analysis of pre-chamber combustion systems for lean burn gas engines 2019-01-0260

The current trend in automobiles is towards electrical vehicles, but for the most part these vehicles still require an internal combustion engine to provide additional range and flexibility. These engines are under stringent emissions regulations, for the reduction of CO2 in particular. Gas Engines which run lean burn combustion systems provide a viable route to these emission reductions, however designing these engines to provide sustainable and controlled combustions is challenging. One method to address these problems is to use a scavenged pre-chamber system which creates more favourable conditions for ignition close to the spark plug. The lean charge in the main combustion chamber is then ignited by flame jets which propagate through nozzles from the pre-chamber into the main combustion cylinder. To design and optimise such a pre-chamber in industrial time frames, simulations have been performed using RANS turbulence model combined with a newly developed spark ignition model which allows for the simulation of the early flame kernel development in the pre-chamber. A novel spark model has been developed allowing for an accurate specification of the spark-ignition process. The model covers all stages of spark discharge from breakdown and the formation of the initial kernel and includes a predictive model for the initial flame kernel size. The flame kernel evolution is computed via a 1D variable temperature model incorporating plasma physics with two-way coupling with 3D CFD. This model has then been validated against experimental data Following this initial verification, the model is then applied to the analysis of a novel pre-chamber ignition system developed within Horizon 2020 GASON project. Several pre-chamber configurations have been simulated to fine the best configuration for mixing in the pre-chamber and then the final flame propagation into the main chamber The results demonstrate good accuracy of the developed model using both cut-cartesian style grids and the results are obtained in time frames suitable for industrial CFD usage.