Abstract
The performances of four dispersion nozzles were studied: the standardized rebound nozzle and the mushroom-cup, and two additional nozzles called mushroom nozzles having different surface section area for air injection. Their influences on the in-situ particle size distribution (PSD) and their impact on the explosion severity were assessed. Discrepancies depend on the powder nature and initial PSD, but also on the injection dynamics. Fragmentation effects were highlighted with the rebound nozzle and a mushroom nozzle (type A), which improves the explosion severity. On the contrary, dust lifting is sometimes not sufficient with the mushroom-cup, which is perceptible focusing on the overpressure evolution. If the rebound nozzle is conservative in most cases, its use should not be systematic and alternative designs, as mushroom nozzles, can sometimes be preferred.
