Page 656 - Dust Explosions in the Process Industries

Research and Development 623

and the maximum rate of pressure rise generated in a closed 23liter explosionbomb. It was
found that both parameters increasedlinearlywith the initialpressure. Pilao et al. also found
that the minimum explosivedust concentration increased linearly with the initial pressure.

9.3.7.3
Explosions in Systems of Interconnected Process Units: Explosion Isolation

The objective of explosion isolation is to prevent dust explosionsfrom spreading from the
primary explosion site to other process units, workrooms, and the like. An outline of the
explosion isolation concept is given in Section 1.4.4in Chapter 1.A basic understanding
of flamepropagation and pressurebuild-upin coupled process equipment(“interconnected
vessels”)is required for specificationof performancecriteria of various types of active and
passive isolation equipment. Wingerden and Alfert (1992),Wingerden et al. (19941,and
Wingerden, Pedersen, and Eckhoff (1995)reported on dust explosion experiments in a
system of two vented vessels connected by a duct. They found that the passage of flame
from the primary ignition vessel through the duct and into the second vessel could result
in substantially higher explosion pressures there than would have occurred in a single
vented vessel of the same size and vent area. The combined reasons for this were pressure
piling, jet-initiated high initial turbulence, and turbulentjet ignition in the second vessel.
Further extensive and most valuable large-scale experimental work in this area was
reported by Lunn (1992a,1996),Lunn et al. (1996), and Holbrow, Andrews, and Lunn
(1996). These investigations comprised dust explosion experiments in various systems
of interconnected vessels of volumes ranging from 2 to 20 m3. Both vented and fully
enclosed systems were studied. The experiments confirmed the findings of Wingerden
and Alfert (1992)and Wingerden et al. (1995).In fully enclosed systems, significantpres-
sure piling effects, yielding maximum explosionpressures approaching 20bar&), were
observed in some experiments.It was also found that the pressure in the primary vessel
could rise to higher values than expected for single vessels, but to a lesser degree than in
the secondary vessel. The combined effects depended on the connecting pipe diameter,
the pipe length, the vessel volumes and their ratio, in which of the two vessels the explo-
sion was initiated, the explssibility of the dust, and the vent areas of the two vessels.
Holbrow, Lunn, and Tydesley (1999)summarized the status on this problem and pre-
sented coherent quantitative guidance for design of interconnected process equipment,
focusingon the two protectiontechnologies-explosion containmentand explosionventing.
Vogl(1994)presented further results from a comprehensiveexperimentalprogram in
Germany on propagation of dust explosions in interconnected process systems (see
Section 4.4.7and Figure 4.66in Chapter). Pipe lengths up to 48m and pipe diameters
up to 200 mwere used. The influencesof a range of experimentalparameters on the flame
speed and explosion pressure were studied: initial air velocity in pipe, pipe diameter,
ignition source location, dust concentration in pipe, and the K,, value of the dust. Vogl
(1996)presented results from further experimental studies of the propagation of dust
explosions in a system consisting of a 9.4m3 vented vessel into which powder or dust
was conveyed pneumatically at up to 30m/s through a 25 m long pipeline and immedi-
ately extracted pneumatically from the vessel and conveyed further, through another 20
rn of pipe, to a cyclone collector.The test dusts were corn starch and wheat flour. With
corn starch, flame speeds of 600m/s were observed in the pipeline. Further aspects of
this research program, in particular the prediction of flame speeds in pipes and ducts, were