Page 337 - Dust Explosions in the Process Industries

306 Dust Explosions in the Process Industries

This facility was made available to Eckhoff and Fuhre (1975) to determine the mini-
mum explosible concentration of a wheat grain dust of 10% moisture content, taken from
a dust extractionfilter in a grain silo plant. Due to poor flow properties of the dust, a con-
stant rotation speed of the dust feeding screw did not alwaysresult in a constant dust feed.
For this reason, several dust concentration measurements had to be performed during a
test series at a given screw rotation speed, and some scatterhad to be accepted.Only flame
propagation lengths of more than about 0.5 m upward in the tube were considered sig-
nificant. Propagation lengths of about 0.5-1 .O m were classified as “marginal.”
The results of eight test series are summarizedin Figure 4.26. Each series,run at a given
set of nominal dust cloud generation conditions, comprised three to six consecutive
experiments for measurement of dust concentration or flame propagation. Figure 4.26
gives the actual average dust concentration values determined in individualexperiments
in each series and the correspondingflame propagation results.

8
I
6

0
z5
v)
Y
E 4
I. .
VI
k- VI loo 1
? 3 I
I.. .)
[ 00;
2
1 /xx/
0 50 100
AVERAGE DUST CONCENTRATION IN TUBE lg/m31
Figure 4.26 Results from determination of the minimum explosible concentration in air of a wheat
grain dust from a grain silo plant, containing 10% moisture, using the vertical large-scale dust explo-
sion tube of diameter 0.25 m and height 5.2 m developed by Palmer and Tonkin (I 971) (From Eckhoff
and Fuhre, 1975).

For dust concentrations below 50 g/m3no significantflame propagation was observed,
whereas marginal propagation was observed in the range 50-60 g/m3.From 60-80 g/m3,
flame propagated over part of the tube length, whereas full tube length propagation
required dust concentration of at least 90-100 g/m3.
This gradual increase of the extent of flame propagation with dust concentration over
a considerable range was also observed by Palmer and Tonkin (1971) and was typical
for the facility.This illustratesthat realistic dust clouds are never perfectly homogeneous
and a sharp minimum explosibleconcentration value is therefore nonexistent. However,
some numerical value may be required in practice; and in the present case, a conserva-
tive figure would be 50 g/m3.
The absence of a sharp minimum explosible concentration seems to be common also
for experiments in a smaller scale. Therefore, the specification of a given value of the