Page 507 - Book Hosokawa Nanoparticle Technology Handbook
P. 507

15 CERAMIC FILTER FOR TRAPPING DIESEL PARTICLES                               APPLICATIONS

                     100000
                                                                   100000
                             [Porosity 40%]                                  [Porosity 60%]

                      80000
                                                                    80000

                    Number of soot  60000                         Number of soot  60000



                      40000
                                                                    40000


                      20000                                         20000



                          0                                             0
                          10              100             1000           10             100            1000
                                  Particle diameter (nm)                         Particle diameter (nm)
                                       Forward of the filter
                                       Rearward of the filter: After 0 g/L of soot is trapped
                                       Rearward of the filter: After 0.1 g/L of soot is trapped
                                       Rearward of the filter: After 0.25 g/L of soot is trapped
                                       Rearward of the filter: After 0.5 g/L of soot is trapped
                  Figure 15.5
                  Characteristics of trapping efficiency by DPF, and their change as time goes by.



                  As this figure shows, the trapping efficiency is worst  3. Pressure loss
                  in the area near 100 nm where the largest amount of
                  PM particles is discharged. In general, three forms of  When pressure loss is considered, it is necessary to
                  trapping act on filters [3–5]: (1) diffusion, (2) inter-  study pressure loss by dividing it as shown below:
                  ception, and (3) impaction.                      (1) Initial condition.
                    (1) shown above is Brownian diffusion, a phenom-
                  enon which grows in strength as the particle diame-  (2) When PM is trapped.
                  ters get smaller. (2) and (3), on the other hand,  (3) After a filter is regenerated (after the ash
                  become stronger as the particle diameters get larger.   piles up).
                  It can be also seen from above that the trapping effi-  (4) After a catalytic coating is applied (for the
                  ciency is the worst for medium-sized particles with a  catalyzed DPF).
                  diameter of about 100 nm. This is consistent with the
                  earlier result shown in Fig. 15.5, in which the same  The method to reduce the pressure loss differs from
                  tendency was conspicuous in a high-porosity filter.  case to case. It is easily speculated that the pressure
                  However, the trapping efficiency for particles with  loss can be effectively reduced by increasing the pore
                  diameters of several tens nanometers or less is high  diameter, making the porosity higher, and thinning
                  from the initial stage one. This seems to be a charac-  the filter wall. But if these methods are implemented
                  teristic of DPF. Further, in the case of a 40% porosity,  at the same time, the trapping efficiency will notice-
                  the trapping efficiency after the trapping of 0.1 g/l of  ably decline. Therefore, it is necessary to select the
                  PM (time of use: a few minutes or less) reaches  most effective method for a given filter to ensure that
                  almost 100%. The PM initially trapped is thought to  the trapping efficiency is never impaired. In the initial
                  improve the trapping efficiency by forming another  condition, for example, a reduced wall thickness will
                  filter layer.                                  provide the highest effect. Later, when the PM is

                                                                                                        479
   502   503   504   505   506   507   508   509   510   511   512