Page 308 - Dust Explosions in the Process Industries

Propagation of Flames in Dust Clouds 277

Bradley, Habik, and Swithenbank (1986) simulated the combustion of rapidly
devolatilizingcoal dusts by generating stabilizedlaminar flames of mixtures of 40 pm
diameter graphite dust and methane in air. The laminar burning velocities measured
agreed well with the theory of coal dust flame propagation, assuming rapid devolatiliza-
tion and subsequent gas phase mixing and no heat sink influence by the graphite parti-
cles. Apart from radiative losses from the particles, which were also accounted for in
theory, the flames were in fact close to adiabatic.The theoretical prediction also agreed
well with experimentalburning velocities for coal dusts as long as the particle diameter
did not exceed 10 pm and the volatile content of the coal was greater than about 25%.
In a subsequentstudy Bradley, Dixon-Lewis,and Habik (1989) investigatedthe burn-
ing velocities of CH,/air/graphite dust flames near the minimum explosible concentra-
tion at subatmosphericpressure of 0.14 bar(abs). On the basis of anindicatedexperimental
peak flame temperature of 1550 K at the limit concentration for flame propagation, a
theory was developed that enabled computation of chemical species concentrationpro-
files, gas temperatures, and heat release rates for flames at atmosphericpressure. As an
example, it was found that the laminar burning velocity for a fuel concentration corre-
sponding to anequivalence ratio of 0.72 decreased from 0.18 m/s for methane as the only
fuel to 0.06 m/s for a fuel mass ratio of CH$graphite of 0.2. The relevance of assuming
that CH,/graphite mixtures can be used for simulating coal dust mass was investigated
theoretically.
The lower experimentallydetermined limit of volatile content of the coal for a cloud
of coal dust to be able to propagate a self-sustainedflame at normal atmosphericcondi-
tions is about 13% according to Cybulski (1975) and Ballal(l983) and 8-10% accord-
ing to Scholl(l981).
It should be mentioned that Helwig (1965), who used a 43 liter closed bomb, found
that the rate of explosions of coal dust containing 10-50% volatiles, did not increase
monaltonicallywith decreasing particle size.Instead, the explosion rate for the finest frac-
tion, of 0-10 pm particle diameter, was systematicallylower than for the most explosi-
ble size range 20-30 pm. It is not clear whether incomplete dispersion of the finest
particle fraction contributed to this effect.
Jarlosinskiet al. (1987) measured the quenching distance for flames in air of a <74 pm
bituminous coal dust of 32% volatile matter and the same dust ground to <5 pm parti-
cle diameter. The quenching distances were 190mm for the <74 pm dust and 25 mm for
the 4pm one. The reason for these unexpectedly high values is not clear.

4.2.3.3
Organic Materials

Laminar 20 mm diameter flames of lycopodium/air and polyvinyl alcohollairwere stud-
ied by Kaesche-Knscher and Zehr (1958) and Kaesche-Krischer (1959). The burning
velocity, defined as the ratio of airflow to flame cone area, was determined photograph-
ically from the height of the flame cone. Some results are given in Figure 4.13.
Lycopodiudair flames of dust concentrationslower than 180 g/m3and higher than 500
g/m3 were difficult to stabilize (stoichiometric concentration = 125 g/m3).The appear-
ance of a stabilizedlycopodium/airflame was very similar to that of a rich hydrocarbonlair
flame, that is, a blue flame front followed by a more or less luminous soot edge.
Approximatethermocouplemeasurementsof flame temperatures showed about 1800 K