(54bv) Influence of the Particle Size Distribution on the Dispersion Process of Combustible Dusts in the 20 L Sphere

The 20 L sphere is one of the standard apparatuses that is utilized to characterize the explosibility parameters of combustible dusts, gases and vapors. One concern about the use of this experimental setup for the analysis of combustible dust clouds is related to the variation of the particle size distribution of the powder. This physical property changes due to the agglomeration and deagglomeration of the solid aggregates occurred during the dust dispersion. These phenomena determine that a high turbulence in the cloud represents both enhancing and quenching effects for the ignition and the initial propagation of the combustion flame. Therefore, this condition might lead to a high uncertainty level during the estimation of the severity and ignitability parameters of dust explosions and induce the bad design of process equipment.

This study evaluated the dispersion and the explosion severity of dust clouds formed with wheat starch agglomerates of different Particle Size Distributions (PSD). The standard dispersion test was performed in the 20 L sphere with the standard rebound nozzle with a triplicate test at the dust minimum explosive nominal concentration (30 g/m³). The turbulence levels of the dust dispersion were analyzed with a Particle Image Velocimetry (PIV) technique. Furthermore, the PSD variations of three samples with different initial aggregation levels (d₅₀: 68.2, 270 and 420 µm) were measured before and after dispersion by using laser diffraction. Additionally, a qualitative analysis of the integrity of the particles after the dispersion was done using microscopy.

From the PIV analysis, it was found that the evolution of turbulence in dust clouds is similar for agglomerates with different PSD. On the one hand, the mean horizontal velocity remains close to 0 m/s with small variations due to the flow recirculation inside the sphere and the partial symmetry of the body. On the other hand, the mean vertical velocity is determined by the injection conditions of the standard test since it has a negative value at the beginning of the dispersion (0-10 ms) that becomes positive (10-100 ms). Lastly, at the end of the dispersion, the mean vertical velocity has a negative value for each agglomerate, but close to 0 m/s (sedimentation stage).

Subsequently, the granulometric and microscopic analyses established that an important de‑agglomeration in the three starch samples since their diameters d₅₀ were reduced until 19 µm approximately. This fact implies that the external forces that are exerted on the solid’s surface are strong enough to overcome the adhesion forces of the agglomerates. This phenomenon occurs due to the high turbulence induced by the pressure difference between the pressurized air of the dust reservoir and the air located within the explosion chamber. Moreover, this condition determines the evolution of the velocity field within the sphere due to a baroclinic effect that is an important source of turbulence and vorticity. These results confirm the relevance of the ignition delay (standard value: 60 ms) on the experimental test and allow adjusting this parameter according to the physical properties of the dust sample.