Effect of scale on freely propagating flames in aluminum dust clouds. (July 2015)

• Record Type: Journal Article
• Authors: Julien, Philippe; Vickery, James; Whiteley, Sam; Wright, Alex; Goroshin, Sam; Bergthorson, Jeffrey M.; Frost, David L.
• Is Part Of: Journal of loss prevention in the process industries. Volume 36(2015:Jul.)
• Page Start: 230
• Page End: 236
• Publication Date: 2015-07
• Subjects: Aluminium dust cloud -- Flame speed -- Scale -- Radiation heat transfer -- Turbulence; Chemical industries -- Safety measures -- Periodicals; 660.2804
• DOI: 10.1016/j.jlp.2014.12.022
• Languages: English
• ISSNs: 0950-4230

Abstract

Abstract: The majority of experimental tests done on combustible dusts are performed in constant volume vessels that have limited or no optical access. Over the years, McGill University has been developing alternative experimental techniques based on direct observation of dust flames, yielding reliable fundamental parameters such as flame burning velocity, temperature and structure. The present work describes two new experimental set-ups allowing direct observation of isobaric and freely propagating dust flames at two sufficiently different scales to test the influence of scale on dust flame phenomena. In the laboratory-scale experiments, a few grams of aluminum powder are dispersed in transparent, 30 cm diameter latex balloons that allow for full visualization of the spherical flame propagation. In the field experiments, about 1 kg of aluminum powder is dispersed by a short pulse of air, forming a conical dust cloud with a total volume of about 5 m 3 . High-speed digital imaging is used to record the particle dispersal and flame propagation in both configurations. In the small-scale laboratory tests, the measured flame speed is found to be about 2.0 ± 0.2 m/s in fuel-rich aluminium clouds. The burning velocity, calculated by dividing the measured flame speed by the expansion factor deduced from thermodynamic equilibrium calculations, correlates well with the previously measured burning velocity of about 22–24 cm/s from Bunsen dust flames. Flame speeds observed in field experiments with large-scale clouds, however, are found to be much higher, in the range of 12 ± 2 m/s. Estimations are presented that show that the presumably greater role of radiative heat transfer in larger-scale aluminium flames is insufficient to explain the six-fold increase in flame speed. The role of residual large-eddy turbulence, as well as the frozen-turbulence effect leading to large-scale dust concentration fluctuations that cause flame folding, are discussed as two possible sources for the greater flame speed. Highlights: Two new apparatus are presented to study dust combustion. Small-Scale experiments are performed in 14L latex balloons. Large-scale experiments are performed in the field on clouds of about 5 m 3 in volume. Burning velocities derived from small-scale flames correlate well with literature. The greater flame speed in large-scale clouds is believed to be due to turbulence.