Page 473 - Dust Explosions in the Process Industries

440 Dust Explosions in the Process Industries

Figure 6.8 Vented corn starch explosions
in 236 m3 steel silo in Norway. For a much
clearer picture, see Color Plate 7.

most probably due to the different ways in which the dust clouds were generated in the silos
and the difference between the silo heights (10 m and 22 m), as discussed by Eckhoff (1989).
In the 236 m3 silo, the dust cloud was essentially quiescent at the moment of ignition.
There is little doubt that the marked increase of the maximum explosion pressure as the
ignition point was shifted downward in the silo was caused by flame acceleration due to
expansion-induced flow. This is in complete accordance with what has been found in the
past in numerous dust and gas explosion experiments in one-end-open tubes, ducts, and
galleries.
In the 20 m3 silo, the strong, turbulent dust jet, extending from the top of the silo and
several meters downward, was maintained during the ignition and explosion process. This
most probably caused very rapid propagation of any flame initiated at the top of the silo
to the central parts. This would explain why the top ignition and central ignition gave
almost the same explosion pressures. In bottom ignition, the initial flame propagation
was probably comparatively slow. But, as soon as the flame front reached the turbulent
dust zone in the central parts of the silo, a much more rapid flame propagation pattern,
similar to that generated by central ignition, probably developed.