Numerical simulations of homogeneous charge compression ignition engines with high levels of residual gas

Abstract

In order to account for the effect of mixture inhomogeneity in HCCI engines utilizing high levels of residual gas and to estimate accurate initial conditions, a sequential numerical procedure was devised. A one-dimensional (1D) engine cycle simulation and a three-dimensional (3D) CFD analysis were used to calculate the residual gas overall level and its spatial distribution. A Monte-Carlo method for the probability density function (PDF) of the mass fraction and temperature, assuming negligible spatial inhomogeneity in the mean quantities but including finite small-scale fluctuations, was used to allow for micro-mixing in the evolution of the chemical reactions. The computational fluid dynamics (CFD) analysis confirmed that substantial scalar inhomogeneity persists up to top dead centre (TDC). The result from the procedure showed a close prediction of the pressure profile from the experiment. However, the u HC level is underpredicted, attributed to the assumed spatial homogeneity of the mean quantities in the Monte-Carlo simulation, which causes an overprediction of the scalar fluctuation decay. Parametric studies of the initial mixture inhomogeneity and turbulence timescale showed that they affect both the ignition timing and combustion duration. A comparison of three mixing models (IEM, modified Curl, and EIEM) showed that the EIEM model predicts a later ignition. The results suggested that accurate prediction of pollutant emission in HCCI engines with high levels of residual gas can be achieved only by a fully 3D calculation incorporating turbulence-chemistry interactions, although combustion phasing and duration can be predicted with adequate accuracy with a volume-averaged representation of the initial residual gas fluctuations.