Page 455 - Dust Explosions in the Process Industries

422 Dust Explosions in the Process Industries

Table 5.7 ignition of dust clouds by metal sparks and hot spots generated by forcing a metal rod
against a rotatinggrinding wheel (roddiameter 6.3 mm; contactforce 13.2 N;estimatedcontact time
between rod and wheel before ignition <I s)

*NI = No ignition up to a peripheral grinding wheel speed of 20 m/s.
Source: Dahn and Reyes, 1987.

impacts againstrusty steel (thermiteflash ignition).Apositive correlationbetween the fre-
quency of ignition and the minimum electric spark ignition energy is indicated.
Singleimpacts with steel as the spark-producingmaterial generated a very low number
of sparks as compared to the number produced by titanium under the same impact con-
ditions.The temperaturesof individual steel sparks,however, could reach the samelevel
as those of titanium sparks (-2500°C).
Impacts of standard quality aluminum against rusty steel generated no sparks nor any
other luminous reaction at all, only a thin smear of aluminum atop the rust (see Section
1.1.4.5in Chapter 1).Impacts with hard aluminum-containingalloys were not investigated.
In most cases, ignition by titanium sparks (e.g., from titanium against concrete) was
observed very close to the point of impact. However, occasionally ignition was also
observed 10-30 cm downstream of the impact point. Ignitionby a singlemetal spark was
never observed. A fairly dense cluster of sparks seemed necessary to ignite the clouds
of corn starch.
Any moving object in the dust cloud reduces the ignition sensitivity of the cloud in the
vicinity of the object by inducing turbulence.The experiments showed that, at a given net
impact energy, the ignition frequency dropped when the impact velocity increased.
Therefore, at a given net impact energy,objectsgeneratinglow turbulencerepresenta greater
ignition hazard than objects generating high turbulence. This was illustrated by an exper-
iment in whch the impactingobject on the spring-loadedarm(Figure7.40) was withdrawn
slightly, allowingit, once the arm was released, to passjust above the anvil withouttouch-
ing it. Instead, an electric spark was discharged at the point where the impact would nor-
mally have occurred and the frequency of ignition measured as a function of electric spark
energy for various tangential arm-tip velocities. The results are shown in Figure 5.24.
The microscopic nature of the anvil surfaceis decisive for the spark formation process.
For example, impacts against worn concrete surfaces exposing naked stone and gravel
faces produced considerably more sparks than impacts against a fresh concrete surface
covered with cement.
The overall practical conclusion of the investigation by Pedersen and Eckhoff (1987)
is that, up to net impact energies of 20 J, tangential accidental single impacts between