Content area
Abstract
This thesis presents a detailed experimental investigation of wake-stabilized diffusion flames in crosswind, which are fundamental to the applied problems of solution gas flaring. With over 1.4 billion m3 of solution gas being flared and vented in Alberta annually and much more being flared worldwide, the relevance of this research is clear. To overcome the significant problems associated with studying emissions of open diffusion flames in a crosswind, a new measurement technique was developed. Flames were established at the exit of a burner tube that was mounted vertically in the test section of a closed-loop wind tunnel. The accumulation rates of the major carbon containing species in the products of combustion were tracked experimentally and used to calculate the carbon conversion efficiency. Efficiency measurements were conducted for a variety of different fuel types, stack diameters, and wind speeds. The results show that for any given fuel, increased crosswind speed (U∞) adversely affects the conversion efficiency, while increased jet exit velocity (Vj) makes the flame less susceptible to the effects of crosswind. Consideration of buoyancy and momentum forces as defined by a Richardson Number successfully predicted the velocity dependency of the conversion inefficiency as being U∞/[special characters omitted] and correlated data for each of the fuels. Further experiments examined the importance of burner diameter, fuel composition and energy density, and ambient turbulence. Lowering the energy density of the fuel was found to have a profound, adverse effect on the measured inefficiency and results of this work have led to a change in the regulations governing flaring in the province of Alberta. Correlations for this effect have been found. Analysis of the combustion products showed that the inefficiencies are primarily in the form of unburned hydrocarbons along with some carbon monoxide and not pyrolitic compounds. Photographic data shows a link between the flame burning in detached pockets and the measured inefficiencies. These results suggested that the observed inefficiencies could be a result of “fuel stripping” from the fuel jet before any combustion. Measurements with a single point fast flame ionization detector verified this hypothesis and showed that unburned fuel was ejected beneath the flame in highly intermittent and spatially variable bursts. Based on these results, a new fuel stripping mechanism for the inefficiencies of wake-stabilized diffusion flames in crossflow has been proposed.
