Katsuya Saijyo

Computational fluid dynamics analysis of the effect of mixture heterogeneity on combustion process in a premixed charge compression ignition engine

Abstract We analyzed the interrelationships between mixture heterogeneity and reaction in a premixed charge compression ignition (PCCI) combustion, using large eddy simulation (LES) in conjunction with a reaction kinetics model. The aim of this analysis is to find the statistical characteristics of the mixture heterogeneity in a turbulent flowfield for moderating the PCCI combustion and for increasing an output limit, which is restricted by a severe knock. Several different initial conditions of heterogeneity of an air-fuel or air-fuel-EGR gas mixture were given at the intake valve closing time by a new method, which generated statistically reasonable turbulent fluctuations in both velocity and fuel mass fraction fields. The autoignition and combustion behaviours were analysed for several different sets of the r.m.s. and the length scale of the fluctuations in the fuel mass fraction. The analyses show that the combination of a larger r.m.s. value and a longer-length scale of the fluctuations in fuel mass fraction is effective to slow the combustion in a hot flame reaction phase and to avoid knocking. The analytical results also show that the heterogeneous distribution of an EGR gas has a considerable effect in making the combustion slower, even when a fuel-air mixture is homogeneous.

A Numerical Prediction Method for the Auto-Ignition Process in a Homogeneous Charge Compression Ignition Engine

An auto-ignition process in a homogeneous charge compression ignition engine has been numerically solved by the Very Large Eddy Simulation (VLES) which integrates a reduced kinetic model for the low temperature oxidationof hydrocarbon fuels. We employ a new method to set the initial turbulent velocity field, reforming the velocity field so that the turbulence dissipation process may fit the results of the K-e model simulation. The phase-averaged quantities of the VLES agree well with those of the K-e model simulation. The VLES exhibits the spatially random appearance of auto-ignition sites, which is similar to experimental observations shown in the reference literature.