Modelling of ignition and early stages of combustion in spark-ignited internal combustion engines

The increasing global energy demand and the detrimental effects of climate change have intensified the need for innovative technologies, among which spark-ignition (SI) engines powered by sustainable fuels such as hydrogen and ammonia are emerging as a promising alternative. However, their use is hindered by several issues, especially related to the control of the flame kernel ignition and expansion. Critical is also the modelling of flame stretch, for which currently available literature models fail in cases of high stretch intensity. Addressing this topic, the present Ph.D. thesis proposes a novel and more advanced model for the prediction of the flame kernel initiation and early expansion after the spark ignition in combustible mixtures. The research begins by considering the plasma generated by the spark discharge, which initiates the combustion process. The determination of the plasma properties in thermodynamic equilibrium at elevated temperatures, up to 100,000 K, is essential to describing the initial plasma expansion, caused by the electrical breakdown. These plasma properties are used for the flame kernel initiation, simulated with a literature model based on the shock wave expansion theory adapted to the simulation of combustible mixtures, which is shown to provide a more accurate representation of the conditions that control the flame kernel initiation. Concerning the flame kernel expansion, the key idea of this work is coupling the thermodiffusive theory of flames, which enables very good theoretical description of the flame stretch, with the convective framework of actual flames. An initial attempt is proposed in this thesis with a preliminary flame kernel expansion model, but its foundations turn out to be shaky despite some acceptable results. However, subsequently a novel and vastly improved model for predicting the flame kernel expansion is proposed based on transient thermodiffusion, which allows for a consistent mathematical development that results in a simple yet robust two-equation kernel expansion model. After calibration, this novel model is validated successfully against experimental data on propane/air flame kernels from literature, demonstrating satisfying predictive capabilities. In particular, the model is able to capture the highly non-linear effects of flame stretch, represented by medium-to-high Karlovitz numbers (Ka > 0.03), for small kernels (radius up to 5 mm). Following this key advancement, this novel expansion model is applied to the prediction of the MIE (minimum ignition energy) of combustible mixtures, which is conducted after modifying the model to account for electrode heat transfer and plasma geometry. The MIE predictions are generally good (around 0.01 to 1 mJ) for propane and hydrogen across a wide range of equivalence ratios except for lean hydrogen/air mixtures, suggesting the need for future developments in this regard. Finally, further validation of the novel kernel expansion model is conducted through a comparison with new and recently made available experimental data, obtained through a collaboration between engine manufacturer Wärtsilä and the University of Udine. The data involve both methane/air and propane/air mixtures ignited at compressed conditions up to 10 bar, and the validation was successful, finding very good agreement between the experimental data of flame kernel radius over time and the model predictions, further confirming the strength of the novel kernel expansion model. A key future development is the inclusion of the effects of turbulence into this novel model, which will prompt its application to the simulation of combustion in SI engines.