Page 401 - Dust Explosions in the Process Industries

370 Dust Explosions in the Process Industries

4.5.2.2
Unconfined Dust Cloud Detonations
As discussed by Lee (1987), Borisov et al. (1984) canie to the conclusion that uncon-
fined dust clouds may be considered as practically impossible to detonate. This was
because the ignition delay times of clouds in air of wheat flour and similar materials are
at least one order of magnitude greater than for methane/air, which requires at least 22 kg
of high explosives to detonate in the unconfined state. By assuming that the minimum
detonation charge for dust clouds is proportional to the cube of the induction time, as
for premixed gases, one finds that a minimum of 20 tons of high explosive would be
required for direct initiation of detonation in an unconfined cloud of wheat flour in air.
However, the induction time for dust clouds decreases with decreasing particle size or
increasing specific surface area, and therefore unconfined detonations in clouds of very
fine dusts become less unlikely than in wheat flour/air.
The only direct experimental observation of a self-sustained detonation wave in an
unconfined dust cloud in air that has been traced was made by Tulis and Selman (1984)
and Tulis (1984). They used aluminum dusts of various fineness and found that detona-
tion could be initiated only with a very fine aluminum flake powder of specific surface
area 34 m2/g,correspondingto sphericalparticles of diameter smaller than 1pm. In the
first phase of this work, Tulis and Selman (1984) worked with an unconfined dust cloud
of approximately cylindrical shape, 6 m in diameter and 1 m in height, containing 4.5 kg
of the fine aluminum powder, corresponding to an average nominal dust concentration
of 160g/m3.The centrally located initiatorcharge was 2.3 kg of high explosive.Although
indications of self-sustaineddetonationswere demonstratedin these experiments,the size
of the dust cloud was too small to eliminate the influence of the initiation charge on the
detonation wave propagation. Therefore, as reported by Tulis (1984), a further experi-
ment was conducted,using three simultaneouslygenerated aluminum flake clouds of the
same size and concentration as just described, forming one large elongated flat uncon-
fined cloud of length 10 m.
When this cloud was initiated at one end with a 2.3 kg high explosive charge, stable,
self-sustained detonationwas achieved. The average velocity of the detonation wave was
1750 m/s, and the peak pressures in excess of 28 bar. The corresponding calculated C-J
values were 1850 m/s and 26 bar. This close agreement between experiment and theory
seems to confirm that the phenomenon observed was in fact a proper, unconfined dust
cloud detonation.

4.5.3
THEORIES OF DETONATION

4.5.3.1
The Chapman-JouguetTheory
As pointed out by Lee (1987), there is no a prim4 reason for assuming that the classical
Chapman-Jouguettheory for detonation,which has been successfullyapplied to premixed
gases and solid and liquid explosives,does not apply even to explosibledust clouds.This
theory, developed independentlyby Chapman (1899) and Jouguet (1905,1906), predicts