Page 389 - Dust Explosions in the Process Industries

358 Dust Explosions in the Process Industries

of dust per unit length of duct correspondingto the desired nominal dust concentration.
Ignition was sufficiently powerful to start dust entrainmentand flame propagation through
the dust cloud,but subsequentpropagation depended on whether a sufficiently strong flow
field was generated ahead of the flame for entraining the dust further downstream.This,
as expected, depended on the extent to which the tube ends were closed or open and on
the location of the ignition point. Some examples are given in Figure 4.65.

-0- 7-_--
A. [I--
-.
1 ' --.
0, = %AX. FLAKE SPEEI
'1 0- MAX. EXPLOSION
PRESSURE
\.

250 440 700
INTERNAL DIAMETER OF EXPLOSION TUBE [mmJ

Figure 4.65 Maximum explosion pressures and maximum flame speeds during wheat flour/air
explosions in one-end-open or fully closed tubes of lengths 42 m and three different diameters. The
nominal average concentration is 470 g/m3. ignition is at the closed tube end (Replot of data from
Pineau and Ronchail, 1982).

The results for the 700 mm diameter tube show that the maximum explosionpressures
were low and nearly the same; that is, 1bar(g) for the one-end-opentube and 1.5bar(g)
for the fully closed one. In the closed tube, the low pressure means that the flame speeds
and associated gas velocities were too low to cause entrainment and dispersion of the
majority of the dust. In the one-end-opentube, the speed of the blast ahead of the flame
was much higher, due to the venting at the open end. This, in turn, entrained the still
unburned part of the dust in the tube and gave rise to a sufficiently high combustionrate
in the resulting dust cloud to generate 1 bar(g) pressure in spite of generous venting. As
the tube diameter decreased,the maximum flame speed increased, even in the closed tube;
more dust was entrained and burned, and a higher maximum explosionpressureresulted.
Pineau and Ronchail (1982) also conducted a number of experiments in tubes of
smaller diameters in the range 25-100 mm connected to a vessel in which the explosion
was initiated.The tube lengths varied between 10m and 40 m and the volume of the vessel
was either 1m3or 0.1 m3.The influence of a 90" bend in the duct was also investigated.
Furthermore, the effect of venting, either of the vessel or at the bend, was studied.