Abstract
Large-Eddy Simulations (LES) of a gasoline spray, where the mixture was ignited rapidly during or after injection, were performed in comparison to a previous experimental study with quantitative flame motion and soot formation data [SAE 2020-01-0291] and an accompanying Reynolds-Averaged Navier–Stokes (RANS) simulation at the same conditions. The present study reveals major shortcomings in common RANS combustion modeling practices that are significantly improved using LES at the conditions of the study, specifically for the phenomenon of rapid ignition in the highly turbulent, stratified mixture. At different ignition timings, benchmarks for the study include spray mixing and evaporation, flame propagation after ignition, and soot formation in rich mixtures. A comparison of the simulations and the experiments showed that the LES with Dynamic Structure turbulence were able to capture correctly the liquid penetration length, and to some extent, spray collapse demonstrated in the experiments. For early and intermediate ignition timings, the LES showed excellent agreement to the measurements in terms of flame structure, extent of flame penetration, and heat-release rate. However, RANS simulations (employing the common G-equation or well-stirred reactor) showed much too rapid flame spread and heat release, with connections to the predicted turbulent kinetic energy. With confidence in the LES for predicted mixture and flame motion, the predicted soot formation/oxidation was also compared to the experiments. The soot location was well captured in the LES, but the soot mass was largely underestimated using the empirical Hiroyasu model. An analysis of the predicted fuel–air mixture was used to explain different flame propagation speeds and soot production tendencies when varying ignition timing.
Highlights
The present study extends our previous works [15,16] by sharing the same objective: providing a better prediction of the fuel–air mixture produced by the spray, to gain better understanding of the ignition, flame propagation, and soot formation/oxidation processes
To investigate the validity of our meshing strategy, we examine the Turbulent Kinetic Energy (TKE) ratio as well as mixing field predicted by current Large-Eddy Simulations (LES) setup
The modeling weakness in ReynoldsAveraged Navier–Stokes (RANS) is attributed to artificially high flame speed induced by very high turbulent kinetic energy, consequence of ignition in the turbulent
Summary
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Applications of fuel-lean operation can be problematic because of the intrinsically low flame speeds and the susceptibility to instabilities To overcome these difficulties, spark-assisted compression ignition (SACI) or mixed-mode combustion is a promising strategy. There are few prior studies available on quantitative mixture measurements in gasoline engines [13,14], a detailed understanding of the fuel-air stratification is still limited due to the complexity of engine flow fields, making the comparison between CFD and experiments most of the time only qualitative These challenges motivated our previous experimental work [15,16] as well as CFD simulations [16]. The present study extends our previous works [15,16] by sharing the same objective: providing a better prediction of the fuel–air mixture produced by the spray, to gain better understanding of the ignition, flame propagation, and soot formation/oxidation processes. A detailed analysis of the mixing is proposed to provide more information on the relation between the local fuel–air mixture and the soot production
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
