DNS analysis of spark-ignited flame kernel growth in a multi-component gasoline surrogate mixture

Reliable spark ignition is a critical requirement on the development of highly-efficient, low emissions gasoline spark ignited IC engines. The low reactivity of lean/dilute mixtures associated with high levels of turbulence can lead to unsuccessful ignition events or flame extinction during the early kernel expansion. Cycle-to-cycle combustion variability is established during laminar-to-turbulent flame transition that occurs by the 1 % mass burn fraction.  Thermo-diffusive effects have an impact during the early flame evolution due to the coupling between flame stretch and burning velocity related to the Lewis number (Le) effects of the fuel component. The objective of this study is to characterize the Le number effects of a multi-component spark-ignited flame kernel evolution in a turbulent flow with different levels of exhaust gas recirculation (EGR). The computational domain is based on the cross-flow ignition experiment from SNL. PeleC and PeleLM are used to solve, respectively, the compressible and low-Mach reactive Navier-Stokes equations with Adaptive Mesh Refinement (AMR) and Embedded Boundary (EB) geometry treatment to account for the electrode geometries. Reaction rates are obtained using a 98-species multi-component gasoline surrogate mechanism. The results show a variation on heat release rate and mixture equivalence ratio caused differential diffusion effects and flame geometry. The Le number effects are investigated considering all the terms in transport equation contributing to the total flame displacement speed. Local extinction, caused by flame-flame interaction, is also observed in highly diluted mixtures.