CFD Study of the Dust Dispersion in the 20L Explosion Sphere: Influence of the Nozzle Design

Abstract

The dispersion conditions of the powders in a 20 L sphere are standardized to obtain reproducible results during explosion tests. However, experimental and numerical investigations have demonstrated that the dust cloud is often heterogeneous and that the nominal dust concentration can be significantly different from that reached at the sphere centre when ignition occurs. In order to improve the dust cloud homogeneity, and consequently, to enhance the repeatability of the method, six alternative dispersion nozzles were designed and tested. The time evolution of the turbulence and of the particle size distribution of the dust cloud were determined respectively by particle image velocimetry and in situ laser diffraction sensor. Euler-Lagrange simulations (Star-CCM+) and DEM simulations were performed to study the dust dispersion dynamics and to assess the importance of the dust fragmentation/agglomeration. Explosion tests were also run for some selected nozzles. Results showed that the turbulence level in the sphere was always lower with the 6 symmetrical nozzles in comparison to the standard rebound nozzle. A lowest turbulent kinetic energy during ignition generally leads to a more homogeneous dust dispersion and thus to a concentration profile at the sphere centre which is closer to the nominal dust concentration. Nevertheless, this decay also implies lower explosion severities and sometimes results in dust accumulation within the nozzles. The particle size distribution of the dust cloud varies from that of the original powder, but does not significantly changes after the first milliseconds of its dispersion within the sphere. Moreover, particle fragmentation was nearly identical for all the nozzles. Numerical and experimental results then demonstrated that the fragmentation is both caused by the injection through the injection valve and by the nozzle geometry. As a consequence, a specific attention must be paid to the injection procedure and to the ignition delay time, which is directly related to the turbulence of the dust cloud.