A methodology for quantifying combustion efficiencies and species emission rates of flares subjected to crosswind

Abstract

Despite strong indications that flares subjected to crosswinds can undergo fuel stripping mechanisms that reduce efficiency and lead to emission of unburned fuel, most published flare experiments have not considered the impact of a crosswind. Knowledge of flare performance in turbulent crosswinds representative of atmospheric wind is especially lacking. Accurate experiments to fill this gap can only be achieved in an environment where the wind conditions can be controlled. This work presents a methodology to determine the carbon conversion efficiency and species emission rates of a flare burning in a closed-loop wind tunnel. The developed methodology is based on solving an unsteady mass balance to relate measured accumulation rates in a closed-loop wind tunnel to emission rates, while extending the applicability and correcting errors of an earlier method. The developed methodology considers complicating factors such as infiltration and exfiltration of gases into and out of the wind tunnel, potential for reburning of products within the wind tunnel, and presence of inert species in the fuel stream. The methodology is assessed with a comprehensive Monte Carlo uncertainty and sensitivity analysis. Results suggest that carbon conversion efficiency can be measured with a systematic bias uncertainty of ≲ ±0.5%, ranging from 0.04% for a 99% efficient flare to 0.55% for a 70% efficient flare at 95% confidence, and methane emission rates can be measured within ±3.25% at 95% confidence. Demonstration experiments in a large closed-loop wind tunnel show that precision (repeatability) uncertainties will be larger but that overall (combined bias and precision) uncertainties of <10% are readily achievable.