Novel combustion modes are becoming an important area of research with emission regulations more stringent than ever before, and with fuel economy being assigned greater importance every day. Homogeneous Charge Compression Ignition (HCCI) and Premixed Compression Ignition (PCI) modes in particular promise better fuel economy and lower emissions in internal combustion engines.Multi-zone combustion models have been popular in modelling HCCI combustion. In this work, an improved multi-zone model is suggested for PCI combustion modelling. A new zoning scheme is suggested based on incorporating the internal energy of formation into an earlier conventional HCCI multi-zone approach, which considers a two-dimensional reaction space defined by equivalence ratio and temperature. It is shown that the added dimension improves zoning by creating more representative zones, and thus reducing errors compared to the conventional zoning approach, when applied to PCI simulation.

The spray-interactive flamelet and extended multi-zone combustion models coupled with multi-dimensional computational fluid dynamics are applied to investigate the effects of charge stratification in a direct-injection compression ignition engine under low load conditions. A parametric study was carried out in order to compare the two approaches for early and late fuel injection timings. Comparison of numerical results with available experimental data shows that for early fuel injection, both models predict the auto-ignition and combustion characteristics with comparable fidelity. As the fuel injection timing is delayed, however, the spray-interactive flamelet model is found to capture the onset of combustion and subsequent heat release with greater accuracy. Further investigation reveals that the better performance of the spray-interactive flamelet model over a wider range of mixture-stratified conditions is mainly attributed to its ability to capture the diffusive transport resulting from small-scale mixing and turbulence-chemistry interaction, which becomes more important when significant mixture inhomogeneities exist in the engine cylinder.

It has been reported that early combustion in a spark-ignition engine determines the subsequent combustion. Also, the early combustion has a very strong correlation with cycle-to-cycle variability, which limits engine operating range. As such, accurate modeling of the early flame development is very important in accurate simulation of spark-ignition engine combustion. During the early flame development, the flame kernel, initiated by spark, grows initially at laminar flame speed. As the kernel grows, the flame surface wrinkles due to surface instability and interacts with the flow turbulence as the flame transitions from laminar to turbulent flame. In this study, a semi-empirical model is proposed to simulate the laminar-to-turbulent flame transition process during early spark-ignition combustion. A hyperbolic tangent function was used to emulate the laminar-to-turbulent flame speed transition process. The proposed transition function was evaluated during early flame kernel development for both Reynolds-averaged Navier–Stokes and large eddy simulation models against combustion analysis data from high-speed optical particle image velocimetry. Difference in Reynolds-averaged Navier–Stokes and large eddy simulation transition function was analyzed and discussed.

No abstract

No abstract