Page 342 - Dust Explosions in the Process Industries

Propagation of Flames in Dust Clouds 3 11

Table 4.9 Minimum explosible dust concentration (g/m3)determined by three methods in four
laboratories

Zinc I 600(0.9) I 600(0.3) I 400 (0.3) 1 400 (0.1) I 500 I 650 (0.8) I 565+65*A
*Standard deviation.
Note: Figures in parentheses( ) indicate measured maximum explosion pressure in bar(g) minus that due
to the ignitor (assumed to be 1.3 bar(g) for the 20 liter sphere). Figures in brackets [] are from more-recent
data.
Source Eckhoff, 1988.

The 20 liter Siwek test was based on a rather weak and vaguely defined pressure rise
criterion. In addition, the very strong 10kJ pyrotechnical ignitor may cause combustion
of dust even if the dust concentration is below that required for self-sustained flame prop-
agation at constant pressure. Therefore, it is not unexpected that the 20 liter minimum
explosible dust concentrations were generally lower than for the two other methods. The
work of Continillo et al. (1986) with the 20 liter Siwek sphere indicates significant pres-
sure rise for coal dudair clouds even at dust concentrations as low as 50 g/m3,in accor-
dance with the low minimum explosible concentration for coal dust for this apparatus
in Table 4.9. Furthermore, the real, local dust concentration in the region of the ignition
source was not known. The problem of generating nonhomogeneous distributions of dust
concentration in small-scale experiments has been emphasized by Eggleston and Pryor
(1967).
The 1 m3 method also involves a 10 kJ ignition source and a pressure rise criterion,
but because of the large size of the vessel, the net influence of the 10 kJ ignitor on the
pressure rise is small. However, the distribution of dust concentration is not
The Nordtest Fire 011is essentially a constant pressure method, because the top of the
15 liter vessel is covered only by a weak paper diaphragm. The explosion criterion is
independent, upward flame propagation through the experimental dust cloud to an extent
that the flame, as observed visually, is clearly detached from the ignition source. A spe-
cial feature of this method is that the actual local dust concentration in the region of the
ignition source is measured directly gravimetrically.Most ofthe Nordtest data in Table 4.9
are based on an earlier, quite restrictive criterion of explosion, requiring fairly extensive
flame propagation. More recent data, based on the present criterion of any flame prop-
agation clearly detached from the ignition source, are given in brackets.
kovachev (1976) discussed some unrealistically low values for the minimum explosi-
ble concentration of some dusts reported in USSR, and he emphasized the necessity of
observing self-sustained flame propagation through the dust cloud, beyond the influence
of the ignition source (see also Section 7.13).