Combining CFD simulations and PIV Measurements to Optimize the Conditions for Dust Explosion Tests

Abstract

Two studies were developed in parallel on the dust dispersion in a tube during a flammability test. On the one hand, an experimental set-up composed of a Particle Image Velocimetry (PIV) system, a high speed video camera and a laser diffraction sensor was used to characterize the dust cloud, notably the mean velocity of the particles, the root-mean-square velocity and the turbulence intensity. On the other hand, a Computational Fluid Dynamics (CFD) simulation was developed by using an Euler-Lagrange approach. Good agreements were obtained between particle velocities and turbulence levels measured by Particle Image Velocimetry and those determined by simulations. The relation between the initial turbulence and the homogeneity of the dust dispersion has also been discussed. Three stages have been identified during the dust dispersion: a first phase of turbulence intensity increase due to the presence of powerful air jets, a second phase of the turbulence decrease during which the velocity vectors are less oriented and the dust cloud tends to be more uniform and a third represented by the particles settling. The relevance of the electrode positioning as well as the choice of the ignition delay tv in order to perform reproducible flammability tests have also been discussed. In the short term, these results will improve our predictive models on dusts explosions. In the medium term, this study will advocate modifications of the existing procedures/standards in order to define, ab initio, the suspension characteristics which will better correspond to actual industrial conditions or, which will lead to the worst case scenario.