(70b) Assessing the Hazard of Marginally Explosible Dusts

With the current focus on combustible dust
hazards in industry, an increasing number of dusts are being tested by
facilities to determine their explosibility.  For marginally explosible
dusts (MEDs) with low P_max and K_St values, it can be a difficult task to
determine if the material is truly explosible, and if
so, what level of protection is required for a specific process.  In this paper we address ways to measure the explosibility of the MED and to analyze the consequences of
a deflagration involving an MED in an enclosure.

Explosibility screening tests are often performed in a
spherical 20-L vessel using strong pyrotechnic igniters, in accordance with
ASTM E1226 Standard Test Method for
Explosibility of Dust Clouds.  The
20-L vessel was original designed by Siwek to
correlate with data in the 50 times larger 1-m³ (1,000 L) vessel,
while requiring less dust and time to perform testing.  Previous studies have shown that some MEDs
may appear to be explosible when tested in the 20-L
vessel but non-explosible when tested in the 1-m³
vessel.  This phenomenon has been attributed
to ?overdriving? by the strong chemical igniters in the 20-L vessel.  The 1-m³ vessel is seen as a more
reliable vessel to determine whether a deflagration will propagate through a
cloud of MED because of the lower chance of overdriving in the larger
vessel.  It has been hypothesized that
overdriving may be relevant to situations that can occur in industry involving
strong ignition sources, such as an initial gas or hybrid explosion which then
propagates through a combustible dust.

In this
study, we review existing literature comparing tests of MEDs in both the 20-L
and 1-m³ vessels.  We also
present new test data of MEDs in both vessels for organic and metal dusts.   We find that many of these dusts are explosible in the 20-L vessel but not the 1-m³ vessel.  However, we also find that some of these low K_St materials are explosible
in both vessels, and may produce higher P_max
and K_St values in the 1-m³
vessel.  For dusts that have tested as explosible in the 20-L vessel, but not the 1-m³
vessel, we conduct additional testing with smaller energy pyrotechnic igniters
to determine the igniter energy that matches data from the 1-m³ vessel.  Similar to previous studies, we find that the
ignition energy required to match 1-m³ date varies for different
dusts. 

Hybrid (flammable
gas and dust) explosion tests are performed on MEDs in the 20-L vessel, using a
spark igniter and lean mixtures of flammable gas to determine if the dust will contribute
to the severity of a gas explosion.  We
find that some MEDs have little effect or decrease the maximum overpressure and
rate of pressure rise of the hybrid mixture, relative to a lean gas
explosion.  In contrast, similar hybrid
explosions performed with more explosible dusts significantly
increase the maximum overpressure and rate of pressure rise of the hybrid
mixture, relative to a lean gas explosion. 
We also measure the amount of gas necessary to cause a deflagration
using a spark igniter with the MED.  These
type of hybrid explosions are compared to possible
ignition scenarios in industry.  We
propose that these methods can be used to evaluate the potential of MEDs to contribute
to gas explosions.

Finally,
vented deflagration calculations are performed on MEDs in enclosures.  The analysis shows that for some enclosures
with existing openings, a deflagration of an MED with a low P_max
and K_St value will not produce a damage
causing overpressure and no additional explosion venting is required.  However, the deflagration will still produce
fireballs that can create a thermal hazard to personnel and property.