Probability Density Function (PDF) is often selected to couple chemistry with turbulence
for complex reactive flows since complex reactions can be treated without modeling
assumptions. This paper describes an investigation into the use of the particles
approximation of this transport equation approach applied to Homogeneous Charge
Compression Ignition (HCCI) combustion. The model used here is an IEM (Interaction by
Exchange with the Mean) model to describe the micromixing. Therefore, the fluid within the
combustion chamber is represented by a number of computational particles. Each particle
evolves function of the rate of change due to the chemical reaction term and the mixing
term. The chemical reaction term is calculated using a reduced mechanism of n-heptane
oxidation with 25 species and 25 reactions developed previously. The parametric study with
a variation of the number of particles from 50 up to 104 has been investigated
for three initial distributions. The numerical experiments have shown that the hat
distribution is not appropriate and the normal and lognormal distributions give the same
trends. As expected when the number of particles increases for homogenous mixture (i.e.
high turbulence intensity), the in-cylinder pressure evolution tends towards the
homogeneous curve. For both homogeneous and inhomogeneous (i.e. low turbulence intensity)
cases, we have found that 200 particles are sufficient to model correctly the system, with
a CPU time of a few minutes when a restriction of initial distribution is adopted.