Soot Optical Diagnostic Development and Application to Turbulent Non-Premixed Buoyant Flames

This thesis details the development and application of laser-based optical diagnostics capable of measuring soot volume fraction, mean aggregate radius of gyration and primary particle size in turbulent flames. Research focussed on techniques suitable for making instantaneous measurements in turbulent flames, initially investigating a 2D auto-compensating laser-induced incandescence (2D-AC-LII) diagnostic prior to developing a unique combined AC-LII and elastic light scattering (ELS) system, which was used in subsequent experiments. Comprehensive Monte Carlo-based uncertainty analyses performed on each system indicated that the imprecise knowledge of soot properties was the largest source of uncertainty for all measurements. Use of a newly-developed ELS calibration method allowed a reduction in uncertainties resulting from calibration and instrument noise. The AC-LII/ELS technique was applied to turbulent buoyant non-premixed flames relevant to solution gas flares used throughout the upstream energy industry, resulting in several important insights. First, the results demonstrate that decreases in soot volume fraction seen near the flame tip are attributable to increased flame intermittency rather than decreases in soot volume fraction within soot-bearing structures, providing support for a recent suggestion from the literature that this could occur in momentum-dominated flames and further extending it to the buoyancy-dominated flames studied here. Secondly, and contrary to a suggestion in the literature for momentum-dominated flames, the current results indicate that soot-bearing structures oxidize very rapidly or not at all, rather than being preferentially oxidized in structures with low soot volume fraction. Thirdly, soot aggregate size was found to vary linearly with residence time, and trends for a wide range of flames collapsed when residence times were offset to account for implied variations in soot inception height. Finally, considering the large range of flow rates and burner diameters investigated, it is significant that measured quantities among all turbulent buoyant flame conditions were well-correlated when scaled in the axial direction by either flame length or residence time. The body of work presented here has provided new insights into the sooting behaviour of turbulent buoyant non-premixed flames and has made significant contributions to the development of diagnostic tools that will facilitate future investigations in a range of flame configurations.