Modelling soot formation for turbulent flame prediction

Abstract

A strategy is described for the development of simplified models of soot production suitable for incorporation in turbulent flame predictions. Detailed measurements in laminar diffusion flames are used to establish critical parameters in a reduced model of soot formation and burn-out based on two soot variables, the soot volume fraction fv, and number density, N. This model incorporates representations of the key processes of nucleation, coagulation and surface growth. The govering state properties are related to mixture fraction, a conserved scalar, which permits the ready extension of the model to turbulent flame prediction. The soot model is interpreted as a development of the laminar flamelet approach in which the source terms in the balance equations for fv and N are closed over the pdf for mixture fraction. Measurements of mean mixture fraction by microprobe sampling and mass spectrometric analysis, temperature by fine wire thermocouple and soot volume fraction by laser extinction and tomographic inversion are reported in a confined turbulent jet flame burning at atmospheric and elevated pressures. Comparisons are made between these measurements and model predictions. Whilst the experimental data is plausibly reproduced, uncertainties in the modelling are identified in relation to turbulence interaction, particle size and soot oxidation.