SOOT FORMATION IN DUCTED TURBULENT DIFFUSION FLAMES

A vertical combustion wind tunnel was built in which confined turbulent diffusion flames were studied. Mea- surements were made of the mean soot volume fraction in turbulent ethylene flames using laser extinction. Velo- cities were measured with laser Doppler velocity, and radiation was detected with a wide-view-angle radiometer. The flames were burned under three conditions of axial pressure gradient: 0, - 32, and - 92 Pa/m. The acceler- ating flow was produced by a converging wind-tunnel working section. It was found that the latter pressure gra- dient caused the soot loading to fall by one-half, with a commensurate fall in the radiation emitted by the flame. The flame was shorter when it was accelerated. Turbulence intensities were also greater in the convergent duct. The results indicate the potential impact of pressure gradients on the relatively slow chemistry of soot formation. SOTHERMAL free shear layers such as turbulent jets have been studied for many years. More recently, the chemically reacting turbulent shear layer has been the object of consider- able attention. In almost all of the studies of both types of flows there were no mean pressure gradients. An axial pres- sure gradient may be imposed on a free shear layer if it is con- fined in a sufficiently large duct through which an external or outer flow is maintained. The duct may be either convergent or divergent. In the former case a favorable gradient is estab- lished and the outer flow accelerates; in the latter case the un- favorable pressure gradient slows the outer flow. Normally one would expect that a favorable pressure gradi- ent would tend to suppress turbulence in an isothermal flow. This is the concept behind the use of contractions on wind tun- nels. However, if combustion is occurring in the flow and there are significant density variations, then the response of the shear layer to the imposed mean pressure gradient may be different. In one of the few studies to look into this matter, Starner and Bilger1 performed measurements of the velocity field in a hydrogen air diffusion flame that burned in a wind tunnel. They made the interesting observation that the flame became shorter when either a favorable or unfavorable pres- sure gradient was imposed on the flow. Measurements of the velocities indicated that the favorable pressure gradient caused the turbulence intensities to be enhanced. The effect was as- cribed to a pressure-velocity correlation in the equation for turbulent kinetic energy. The experiment suffered from some flaws. The wind tunnel was horizontal, and, as a result, buoyancy forces tended to cause a circulation in the flow. In addition, the working sec- tion walls were adjusted in only one direction so that symme- try was not maintained. Consequently, the flames assumed an elliptical shape that gave rise to some difficulty in interpreting the experimental data. Nevertheless, the results from this ex- periment indicated that the confinement of a turbulent hydro- carbon flame in a convergent duct may have some impact on the slower chemistry in these systems. Among the chemical reactions in a hydrocarbon-air flame, those that lead to soot particles are probably the slowest. This has been demonstrated by Kennedy2 in a series of experiments