Page 487 - Dust Explosions in the Process Industries
P. 487

454  Dust Explosions in the Process Industries


             select experimental conditions that are as close as possible to the conditions prevailing
             in the actual industrial enclosure (see Figure 6.11). Depending on the way in which the
             dust cloud is generated in practice, vent area reduction factors, with reference to VDI
             3673 (1979), may vary between 1.O and 0.1,


             6.3.5
             INTERMEDIATE (10-25  m3)ENCLOSURES OF SMALL LID

             The experimental basis is that of the VDI 3673 guideline (1979 edition) with highly
             homogeneous, well-dispersed, and turbulent dust clouds and the more recent results
             for much less homogeneous, less well-dispersed clouds (Figure 6.12). The vent area
             requirements identified by these two sets of experiments differ by a factor of up to 5.
             Adequate vent sizing therefore requires that the conditions of  turbulence, dust dis-
             persion, and level and homogeneity of dust concentration for the actual enclosure be
             evaluated in each case.


             6.3.6
             CYCLONES

             Two realistic investigationshave been traced (Tonkin and Berlemont, 1972,and Hayashi
             and Matsuda, 1988),and both suggest a significant vent area reductionin relation to VDI
             3673 (1979 edition). The early investigation by Tonkin and Berlemont using a cyclone
             of  1.2 m3, indicates an area reduction factor of 0.2. The more recent investigation by
             Hayashi and Matsuda, using a smaller cyclone of 0.32 m3,indicates a factor of about 0.5.
             Hence, for organic St 1 dusts (KstI 200 bar ds) there seems to be room for vent area
             reductions with reference to the VDI 3673 (1979 edition) by factorsin the range 0.5-0.2.
             However, for metal dusts, such as silicon, although there is no direct evidence from
             cyclone explosions with such dusts, the VDI 3673 (1979 edition) requirements should
             probably be followed, as in the case of filters (see Section 6.3.7).



             6.3.7
             BAG  FILTERS

             The experimentalbasis is the evidencein Figures 6.16 and 6.19-6.21,  produced by Lunn
             and Cairns (1985) and Eckhoff et al. (1989). If ignition inside the filter itself is the most
             probable scenario (no strong flamejet entering the filter nor any significantpressurepiling
             prior to ignition), the vent area requirements of VDI 3673 (1979 edition) for St 1dusts
             can be reduced by at least a factor of 0.5. If the dust concentrationin the feeding duct to
             the filter is lower than the minimum explosive concentration, the vent area may be
             reduced even more.
               However, in the case of some metal dusts, such as silicon,primary ignition in the filter
             itself may be less probable and ignition is accomplishedby a flamejet entering the filter
             from elsewhere. In this case, it is recommended that the vent area requirements of VDI
             3673 (1979 edition) be followed.
   482   483   484   485   486   487   488   489   490   491   492