834707673

IV3-4

PHOTOGRAPHIC MEASUREMENTS

When there are solid or liquid particles in the flow, velocities can be measured by using high-speed photography. Local gas stream velocities were measured in a diffusion flame in the laminar boundary layer over a fiat piąte by Hirano and Kanno¹⁶. Magnesium oxide particles were suspended in the flow and, by using a repetitive illumination system, velocities were measured from the particie tracks. These velocity measurements are shown in Figurę 7. The velocity distribution across the boundary layer with fuel injection indicated that the thickness of the boundary layer with a diffusion flame was large compared with that without a flame, while the velocity gradient normal to the piąte, at the piąte surface in the presence of a diffusion flame, was found to be steep compared with that without a flame. A marked feature of the velocity distributions across the boundary layer with a diffusion flame was the appearance of a maximum velocity, which exceeded the air stream velocity. Evcn at a point only a few millimetres from the leading edge of the porous piąte a velocity over-shoot was observed near the flame zonę. Once a diffusion flame was established the streamline and velocity profiles changed considerably compared with those without a flame. These aerodynamic changes were attributed by Hirano and Kanno¹⁶ to the pressure distribution changes due to the flame reaction. Pressure distributions were calculated from the measured velocity and temperaturę distributions. A higher pressure region was found at the air stream side of the leading edge of the flame zonę. At the same time, a lower pressure region appeared at the fuel side of the flame zonę. The stream near the lower wali was decelerated and turned away from the lower wali until it attained the higher pressure region. After passing through the higher pressure region the stream was accelerated and turned towards the lower wali. The aerodynamic flow fields of the air stream side of the flame zonę were found to be distorted, as if an object had been placed on the porous piąte instead of a diffusion flame.

The effects of pressure changes, acceleration and increase of the width of the flow Fields as a direct consequence of combustion has not yet been clearly established in turbulent diffusion flames. Measurements madę recently at the University of Sheffield using a laser anemometer in flames have shown elear indications of accelerations in the vicinity of flame fronts.

Measurements of flame properties in flames containing particulate matter present additional problems, due to the impaction of particles on any physical probes introduced into the flame. A series of studies has been carried out at Sheffield University¹⁷’^18,19 »²⁰»^2,» on spray flames burning liquid fuel particles injected into the spray by liquid atomizers. Much of this work has concentrated on the measurement of particie sizes of the liquid droplets in the flames and the particie dynamics. A double-spark high-speed photographic system was developcd in which two sparks are fired consecutively at electronically controlled times. A narrow depth of field of < 1 mm thickness is photographed. The droplet images are magnified 100 times and particie sizes are determined by direct measurement of the diameter of drop images. Measurement of the distance between the two images, coupled with the measurement of the time interval between the two sparks, gives the droplet velocity, and the angle of flight of the droplets is also determined from analysis of the photographs. The optical system, photographic equipment and the electronic control system, together with the method of analysis of photographic plates, is described morę fully in Reference 17. Changes in mass of liquid droplets, as measured in a liquid spray burning in the wake of a stabilizer disc,¹⁸ are shown in Figurę 8. The mass of droplets < 50 /im as a proportion of a total mass of droplets is plotted as a function of axial distance along the spray. The effects of variation of the annular air veIocity from 8-10 m/s are shown in Figurę 8 and the differences in drop size are compared between the unignited cold conditions and the hot burning conditions. Measurements of drop velocities are shown in Figurę 9 as a function of axial distance along the spray for drop size ranges < 50 /im and > 100 /im . When the velocities of drops are compared with measurements of air velocity it can be seen that there are very significant deviations between the particles and the air. There are also significant differences between velocities of droplets in the hot burning spray as compared to those in the cold unignited spray.

For the prediction of flames with particulate matter, such as liquid sprays, the spatial distribution of drops and their size distribution needs to be known. The rates of evaporation and burning of particles are a function of local temperaturę, vapour pressure, heat transfer and mass transfer conditions. The spray studies at Sheffield have shown that for both hollow-cone pressure jets and for twin fluid atomized jet spray flames, temperaturę conditions are too Iow and mixtures are too rich to allow burning to take place within the spray cloud. Instead of the classical concept of envelope flames surrounding individual droplets the main reaction zones were found at the outer periphery of the sprays. These flames are mixing controlled, as in diffusion flames, and not controlled by the ratę of vaporization of the droplets. Theoretical predictions madę on the assumption of envelope flames surrounding individual droplets assume that rates of heat transfer are governed by molecular conduction and rates of mass transfer are governed by molecular diffusion between the flame front and the droplet surface. For the cases of the spray flames studies at Sheffield, with flame fronts at the outer periphery of the spray cloud, rates of evapor-ation of droplets become dependent upon the turbulent transfer of heat and mass between the droplet surface and a flame relatively far removed from the drop surface. These studies have shown the important role that management has and must continue to play in providing the appropriate physical model and boundary conditions, which are a necessary prerequisite for the correct formulation of the theoretical problem.