Page 257 - Dust Explosions in the Process Industries

Generation of Explosible Dust Clouds 229

where pp is the particle density, v, is the actual “drag velocity” of the air (see the para-
graph following equation (3.22)), and v,,~~is the minimum “drag velocity” for entrain-
ment of particles (Figure 3.20). The v,,~~values for the four powders were 20, 23, 24,
and 27 ds, in order of increasing particle density.
Akiyama and Tanijiri (1989) used a wind tunnel of 3.6 m length and a rectangular cross
section of 30 mm width and 150 rnm height in their study of reentrainment of dust par-
ticles from a powder bed having its plane surface flush with the wind tunnel floor. The
particles studied included glass beads, talc, alumina, and fly ash of volume-surface diam-
eters ranging from 15 to 80 pm, solid densities in the range from 2.3 to 4.0 g/cm3,and
bulk porosities in the uncompressed state from 0.47 to 0.77.
The bed of particles to be tested in the wind tunnel was conditioned in a humidistat
of relative humidity H for more than 6 hours before being exposed to the reentrainment
experiment. The humidity of the air in the wind tunnel was not controlled, but it was
assumed that the short test period of about 60 s did not significantly influence the
humidity inside the bed, To obtain H= 0, the particle bed was kept at 177°Cfor more than
10 hours.
With a powder bed of 220 mm length and 30 mm width and an average air velocity
of 15 ds in the wind tunnel, the entrained particle mass per unit time was independent
of relative humidity up to 65%. For higher humidities, there was a drop of the entrain-
ment rate with increasing humidity, increasing with decreasing particle size. However,
at the given conditions, some of the particle systemstested could not be entrained at all,
even at low air humidity.It should be pointed out that the particles investigated were non-
hygroscopic, in the sense that moisture did not penetrate into the bulk of the individual
particles but accumulated only on the particle surface. For some natural organic mate-
rials, the influence of the relative humidity may therefore be more complex.
Akiyama and Tanijiri then investigatedthe relationshipsbetween the entrainmentrate
and the four powder mechanical properties: angle of repose, angle of spatula, com-
pressibility, and cohesivenessor cohesion.All these parameters are somewhat arbitrary
and not easy to relate to the more fundamentalpowder mechanical properties described
in Section 3.4.They are determinedin a set of somewhatarbitrary tests, specified in terms
of apparatuses and procedures. An overall dimensionless flowability coefficient F was
defined as a function of the four measured parameters, and the rates of reentrainment
measured in the wind tunnel were correlatedwith F for the various powders. Reasonable
monotonic correlations comprising all seven powders were obtained for the three over-
all wind tunnel velocities 8, 12, and 15 m/s investigated.
Urd (1989a, 1989b)postulatedthat the dispersibilityof dusts can be characterizedby
two parameters: the minimum aerodynamic shear stress required for dust entrainment
from a horizontal surface and the settling velocity distribution of a dust cloud. This is
an interesting approach, which will be discussed in greater detail in Chapter 7, treating
various test methods related to the dust explosion hazard.
It should finally be noted that Bagnold (1960) briefly mentioned the reentrainment
of a powder layer by a sudden blast of gas rather than a steady flow. This clearly is an
important case in the context of dust explosions. Even if the Mach number is consid-
erably smaller than unity and the static pressure gradient in the direction of air move-
ment is negligible, the dynamic pressure gradient (gas velocity gradient) can be
considerable. Section 9.2.2.4 in Chapter 9 gives further references to works on gener-
ation of dust clouds by dispersion of dust layers and deposits.