A spark energy deposition model in mixture fraction space for simulations of turbulent non-premixed flame ignition

Abstract

Predicting the ignition probability remains important for designing reliable combustors. A spark ignition model in mixture fraction space is proposed and used in a Large Eddy Simulation (LES)-Conditional Moment Closure (CMC) simulation of initiation of a n-heptane spray swirl flame. The model is based on including source terms for the enthalpy and species that mimic the effect of plasma kinetics on the gaseous thermochemical state, in contrast to previous approaches that included only a heat source to the energy equation or a burning distribution in mixture fraction space as the initial condition. The model is evaluated based on its prediction of the ignition probability against experimental data. In laminar non-premixed counterflow flames, failed ignition case with low energy deposition have been found to successfully ignite when portion of the deposited energy has been assigned for the oxygen dissociation. In the turbulent spray swirl flame, the results reveal a tendency towards a successful ignition when the spark is subjected to a higher probability of finding stochiometric mixture fraction values, lower axial velocity and higher probability of finding negative axial velocities (pointing towards the bluff-body). The terms budget of the CMC equation is investigated for the successful and failed ignition events. The sum of convection and dilatation remains the dominant term to suppress the spark for the investigated realisations, and a tendency towards a failed ignition is observed when the spark energy assumes comparable magnitudes compared to the sum of convection and dilatation. In the vicinity of the spark, convection and turbulent diffusion remain of equal importance, with the latter dominating at later sparking time instants. The present approach quantitatively captures the ignition probability spatial distributions compared to the experiment. The proposed spark ignition model can improve the spark representation in CMC-based simulations, thereby allowing more reliable simulations in realistic combustors.