Investigation of niacin and aluminum dust cloud ignition characteristics in an explosion hazard testing device using high-speed imaging

Abstract

Experimental investigations of dust explosions in standard industrial testing equipment such as the MIKE3 minimum ignition energy (MIE) testing device are promising ways of producing fundamental insights into dust cloud ignitability. However, advanced experimental methods must be developed to characterize the dust cloud ignition and combustion behavior of various types of dust. In this work, high-speed broadband and species-specific chemiluminescence imaging diagnostics are implemented to explore similarities and differences between organic and metal powder ignition kernel development. A selected set of 600-mg niacin (ϕ = 2.9) and aluminum (ϕ = 1.6) powder samples were dispersed in the air and ignited using a high-voltage spark inside a standard MIKE3 device. The resulting broadband flame emission was recorded at 4 kHz using a high-speed camera for the first 10 ms after the spark. For the niacin sample, species-specific chemiluminescence emissions from excited-state hydroxyl (OH*) and methylidyne (CH*) radicals were also recorded at 4 kHz and 1 kHz, respectively. The flame kernel in each image frame was detected using an intensity thresholding algorithm and tracked throughout the video sequence, yielding quantitative size, position, and velocity measurements of the evolving flame kernel during the first 10 ms after the ignition spark. For the niacin sample, a continuous and non-uniform reaction zone composed of burning particle clusters and excited-state radicals was observed. The niacin flame kernels grew from 5 mm to 17 mm with an initial velocity of 5 m/s. Conversely, an intensely bright reaction zone and discrete burning particles near the flame kernel boundary were observed in the aluminum sample. The aluminum flame kernels grew from 7 mm to 10 mm with an initial velocity of 3 m/s. The niacin and aluminum flame kernels traveled 7–11 mm away from the central spark gap and slowed down to 1 m/s by the end of the 10-ms period. This time-resolved imaging study, when coupled with previously reported three-dimensional particle and flow field data prior to the arrival of the ignition spark, sets the foundation for an improved multi-physics understanding of the initiation of dust cloud ignition and explosion processes.