Page 302 - Applied Process Design For Chemical And Petrochemical Plants Volume II

Packed Towers 291

Lobo et al. [47] proposed the packing factor, F, and
experimentally determined that it better represented the
J
data than the calculated term. Values calculated
g 2 using surface area per cubic foot and percent free gas
V
-
space from manufacturer's tables can be as much as 40%
< off. The values are dependent upon the method of pack-
a ing the tower, i.e., dry dumped, wet dumped, or wet
I
. I dumped and shaken. The latter condition may approxi-
z
- mate the situation after a tower has been running a while
E
8 and the packing settled.
2 0.s Experience definitely indicates that the packing factor, F,
<
c increases with hours of operation for ceramic materials up
B to some limit. This is due to settling, breakage, plugging,
5 0.3 etc. For design of commercial towers, values of F should be
3
K
L? increased from 15 to 73% for ceramic materials, over val-
y 0.2
a ues read from Tables 9-26A-E. The percent increase
depends upon the tendency of the shape to disintegrate
into smaller pieces during operations-flooding, gas surg-
0.1 ing, etc. In general, circular shapes exhibit the least ten-
5 6 7 8 10 15 20 so loo dency to break up. As a reasonable value where data are
PACKING FACTOR. Fp
available, the average of the wet dumped, and wetdumped-
Figure 9-23. Flood pressure drop vs. packing factor for random and-shaken values for tower voidages is recommended.
packings. Reproduced with permission of the American Institute of Leva [40] has correlated the data of Lubin into correc-
Chemical Engineers, Kister, H. 2. and Gill, D. R., Chemical Engineer- tion factors to apply to a non-irrigated bed pressure drop
ing Progress, V. 87, No. 2 0 (1991); all rights reserved.
to end up with pressure drop for a liquid-gas system in the
loading to flooding range. In general this does not appear
any more convenient to use than Figure 9-21D.
of an existing tower, because this change affects the pack- Relations expressing the fractional voids in a ring
ing factor. Tables 26A-E present specific packing factors packed bed are useful in estimated the ''E" dues for a/$
from the manufacturers. determinations [47]. The average deviation is 2.6%.
Many of the packings of the various manufacturers are Dry packed tower:
essentially identical in shape, size, and performance fac-
tors. Some packing manufacturers suggest adjusting pack- E = 1.046 - 0.658 Q (9 - 22)
ing factors for vacuum and pressure distillations; however, Wet packed, unshaken tower:
this should only be done after consultation.
The experimentally determined packing factors are the E = 1.029 - 0.591 Q (9 - 23)
only reliable values to use for design calculations;
although estimates can be made for packing shapes when Wet packed and shaken tower:
no data are available. The packing characteristic is
expressed as: E = 1.009 - 0.626 Q (9 - 24)
where
F = a/E3 (9 - 21)
where at = specific surface of packing, ft2/ft3 1 - (di /do )' not valid if I$ < 0.20 or for
a = effective interfacial area for contacting, ft2/ft3 ' = (id: )o.o170 ' extra thick walls or solids
E = fractional voids
1 = ring height, in.
The values of determined experimentally by Lobo et do = outside diameter of ring, in.
al. are indicated [47]. These are the values in the develop di = inside diameter of ring, in.
ment of the basic relation expressed in Figure 9-21A with
correction of Q2 suggested by Leva [41]. These values The generalized correlations of Sakiadis and Johnson
were found to correlate a considerable amount of the liter- [59] are reported to satisfy a wide variety of systems.
ature data within 12%. This would mean about a 6% error Manufacturers of commercial packings provide packing
in tower diameter determined at flooding conditions. factors for their products. Many of the commonly used (not