Predictive Flame Propagation Model for Stochastic Reactor Model Based Engine Simulations

Simon Bjerkborn, Cathleen Perlman, Karin Fröjd, Fabian Mauss

Abstract

A stochastic reactor model (SRM) based spark-ignition engine simulation code allowing for detailed chemistry based knock prediction [1] has been extended with a predictive flame propagation model. The previously used Wiebe function approach, has been replaced by a turbulent flame propagation model where the underlying laminar flame speeds rely on detailed chemistry calculations. This new model development allows for studying the effects of fuel chemistry on the flame propagation event, in addition to determining the knock limit. The turbulent flame propagation model derived by Kolla et al. was combined with a laminar flame speed library for iso-octane to obtain a predictive flame propagation model. The flame development was simulated for arbitrary cylinder shapes using a flexible Monte Carlo geometry
model, echoing and improving on a study published by Poulos & Heywood in 1983. The turbulent flame propagation model is based on a formula derived by Kolla et al. from
Kolmogorov-Petrovskii-Piskunov (KPP) analysis. It was simplified to better suit the application,
while retaining the features allowing for general application. Parameters which could be assumed
constant for a large spectrum of situations were replaced by three parameters. For cases with low ratio between the turbulent velocity fluctuation and the laminar flame speed, where the molecular diffusivity cannot be neglected as in KPP analysis, a simple turbulence model was used with suitable parameters. Validation was made qualitatively, against the predictions of Poulos & Heywood, and against experimental data.