Page 625 - Dust Explosions in the Process Industries

592 Dust Explosions in the Process Industries

under normal gravity to resolve the specific influence of gravity, and a significant effect
of gravity was indicated by some results.
Tse et al. (1999) developed and used an ultrasonic tomographic imaging technique to
examine smolderingcombustionprocesses in a permeable medium.The method provides
information about the local permeability variations within a smoldering sample, which
can, in turn, be converted to information on the propagation of the smolderingreaction.
Results from studies of two-dimensional smoldering processes were reported.
Peters (1999) studied the combustion of coal particles in a packed bed and identified
four distinct combustion regimes, characterized by certain values of the dimensionless
Damkohler number and Thiele modus. Two of the regimes resembled a well-stirred
reactor, and the two others were characterizedby a conversion front propagating with a
characteristic velocity through the packed bed.
A question asked by many is whether combustion in dust layers and deposits can be
initiated by metal particle sparks. Hesby (2000)conducted a series of experiments in
which layers of a wide range of dusts were exposed to showers of burning steel parti-
cles generatedby forcingrods of various steel types against a grinding wheel.The exper-
imental parameters included rod/wheel contact pressure and contact time and distance
between the rodwheel contactpoint and the dust surface.The minimum numbers of par-
ticles required for ignition were estimated for various experimental conditions.A main
conclusion was that the number of steel sparks from single accidentalimpacts would be
all too low to cause ignition of the dust layers studied.

9.2.3.4
Ignition of Dust Clouds

A brief introduction to this topic is given Section 1.1.4 in Chapter 1. In dust explosion
statistics, the frequency data for the occurrence of various ignition sources sometimes
contain categories that are not entirely unambiguous. The categories “friction sparks,”
“frictionheating,”orjust “friction,” constitute one example. Sometimeshot surfaces,gen-
erated either by repeated impacts on the same spot or by sliding friction, are included in
the “friction spark” or “mechanical spark’, category. The perception of the basic mech-
anisms causing ignition in these cases is not always clear. Ignition may have been caused
by either the burning metal particles or the hot surface produced at the object being
ground, cut, or hit by repeated impacts. Furthermore, one does not always distinguish
between the comparatively scattered transient showers of burning metal particles from
single accidental impacts and the dense semi-stationaryspark showers of such particles
produced by grinding and cutting equipment.A central objective of research in the area
of dust cloud ignition is to identify the details of the various ignition mechanisms.
We1(1993) conducteda series of experimentsin which laminar dust clouds were ignited
by a shortfocused laser light pulse (100 ps or 10 ns) from a Nd-YAG laser (1064 nm
wavelength). This kind of experiment can provide basic information about dust cloud igni-
tion processes and flame propagation processes in dust clouds. We1 et al. (1994) used a
simple,modified Semenov theory for autoignition(no temperaturegradientsinside the ini-
tially heated volume) to transform the experimental laser-light-pulse ignition data to pre-
dicted minimum ignition temperatures and energies. The predicted values were in
approximateagreementwith minimum temperatures and energies for the ignition actually
measured. Proust (2002),also using a Nd-YAG CW laser, determined experimentally the