A NEW CORRELATION FOR PRESSURE DROP OF

PACKED COLUMN
Teruo TAKAHASHI
Department of Industrial Chemistry, Okayama University,
Okayama 700
Yasuharu AKAGI and Katsuki UEYAMA
Department of Applied Chemistry, OkayamaUniversity of Science,
Okayama 700

Bythe use of the experimental data of manyprevious studies, a newcorrelation for the pressure
drop of a packed column was obtained in this work. The total pressure drop is expressed as the
sum of the dry pressure drop and the wet pressure drop. The dry pressure drop originates mainly
from the friction of gas rising through the void of the packed bed, and is expressed by the Fanning
equation. The wet pressure drop originates from the increase in net gas velocity because of the
smaller void of the packed bed owing to the liquid holdup, and therefore it is expressed as a function
of liquid holdup, gas velocity, etc. The liquid holdup consists of static and dynamic holdups, and
is correlated separately.

vious work1'3"10'12"21'24'255 are shown in Table 1. The

Intr o du cti on
In previous studies, the method of dimensional
analysis has been frequently used to correlate experi-
mental values of the pressure drop of a packed column.
Moreover, correlations of the pressure drop for dry
packing (dry pressure drop) and for irrigated packing
(total pressure drop) have been carried out separately,
and no report expressing the total pressure drop con-
sisting of the drop for dry packing and the residuals
hasBecause
apparently been published to date.
of the smaller void of the packed bed in
an operation with irrigation, the net gas velocity
through the void becomes larger than that for dry
packing. It is inferred that the difference between
the total and dry pressure drops (wet pressure drop) is
influenced strongly by the liquid holdup in the packed
bed, because of the increase in the net gas velocity.
However, there is a lack of studies referring
relation of wet pressure drop and liquid holdup.
In this paper, the experimental values of the many
previous studies of the total pressure drop of a packed
columnare correlated as the sumof the dry pressure
drop and the wet pressure drop which is a function of
liquid holdup, etc.
1. Data and Characteristics of Hydraulic Performance
The kinds of packing, the range of packing size,
the superficial gas and liquid velocities used in pre-
Received August 30, 1978. Correspondence concerning this article should
be addressed to Y. Akagi, K. Ueyama is now with Fuji Bread Co., Ltd., Hira-
kata, Osaka 573.
VOL. L2 NO. 5 1979
effects of physical properties of gas and liquid phases
on pressure drop and liquid holdup have been widely
examined. In the data rearrangements of this paper,
the void fractions of various packings reported by
Leva11} and others12'13) are used. They are shown in
Table 2. Since some of the reports showed only cor-
relations, the pressure drop and liquid holdup were
reproduced from the correlations using the originally
reported data of void fraction.
Before the correlations of the experimental values
of the liquid holdup (HL) and the pressure drops (JPd
and APT) of the previous studies, the hydraulic prop-
erties are characterized. The liquid holdup is ex-
pressed as a volume fraction as is schematically shown
in Fig. 1 on the basis of comparison with that of a
plate column without downcomer. Although HL
directly affects the pressure drop in a plate column,
to the passage of rising gas in a packed bed reduces the void
fraction from e to (s-HL) by HL and thus the net gas
velocity and pressure drop increase. APd originates
mainly from the friction of the gas which rises through
the void of the packed bed.
Figure 2 shows the experimental values of the pres-
sure drop for spheres measured by Broz et al.3). The
broken points of the lines of the pressure drop for
the irrigated packing in the figure are the loading
points, and in the region of gas velocity less than
these points, HL is not influenced by gas velocity.
The data rearrangements in the present study are
carried out with the use of the values of HL and the
pressure drop in the region.
341
 Table 1 Experimental ranges of liquid holdup and pressure drop
Packing ulc ugc Literature cited
Liquid holdup 1/4-3/2 in. Raschig ring 1.66-234 7, 9, 14, 15, 17, 18, 20, 21, 24, 25
1/4-4/3 in. Berl saddle 2.10-239 7, 9, 17, 18, 20, 25
3/8-1 in. Sphere 1.10-171 3, 7, 9, 19, 21
Dry pressure drop 3/8-4/3 in. Raschig ring 7.38-5580 1, 4, 5, 8, 24
2 in. Pall ring 434.0-19400 16
1-2 in. Berl saddle 968.0-5150 5, 8
1/20-1 in. Sphere 62.3-8500 3, 5, 12
Others 113.0-4790 12, 13
Total pressure drop 3/8-4/3 in. Raschig ring 1.98-100 95.4-4310 1, 7, 8, 24
2 in. Pall ring 24.5-342 411.0-12300 16
1/4-1 in. Berl saddle 2. 14-95.8 143.0-2850 7, 8
2 in. Intalox saddle 9. 79-343 414.0-8240 6
3/8-5/8 in. Sphere 1.ll-71.9 38.0-1650 3, 7, 10, 19

Table 2 e of various packings11~13)

\ , [in] 1/20 1/8 1/4 3/8 1/2 5/8 3/4 1 4/3 3/2 2 3
Packing \ aP[m] 0.0013 0.0031 0.0062 0.010 0.013 0.016 0.019 0.026 0.035 0.039 0.052 0.078

Raschig ring
(Porcelain,Glass) 0.73 0.68 0.64 0.64 0.73 0.73 0.73 0.68 0.74 0.74
(Carbon) 0.55 0.74 0.67 0.74 0.67 0.74 0.78
(Ceramic,Clay) 0.69 0.73 0.78 0.85 0.90 0.92 0.95
Pall ring 0.88 0.90 0.91 0.91

Berl saddle
(Porcelain, Ceramic) 0. 60 0.63 0. 66 0.69 0.75 0.72

Intalox saddle
(Porcelain) 0.75 0.78 0.77 0.78 0.81 0.79
Sphere (Glass, Clay) 0.40 0.40 0.40 0.40 0.40 0.40 0.40 0.40

Cylinder 0. 38 0. 46
Coke 0.50 0.50
Granule 0. 51

2. Liquid Holdup
The total liquid holdup (HL), denned as the total
liquid volume in the packed bed under operating con-
ditions, is expressed as cubic meters of liquid per
cubic meter of the packed bed. The static holdup
(Hs), defined as the liquid volume which does not
drain from the packed bed when supplies of liquid and
Fig. 1 Comparison of flow patterns of gas and gas to the column are discontinued, is expressed in
liquid flow between plate and packed columns the same manner as HL. The dynamic (operating)
holdup (Hd), denned as the difference between the
total and static holdups, represents the liquid volume
which drains from the packed bed. These three
kinds of liquid holdup are correlated separately in the
following sections.
2. 1 Static liquid holdup
The experimental values of Hs for the various pack-
ings reported by Otake etalU) and Others8'15'17'18'20'21),
are plotted in Fig. 3 against packing size (dp). As Hs
is mainly influenced by dp, it is expressed as follows.
Hs=1.53 x l0-*d-U2° (1)
2. 2 Dynamic liquid holdup
The examples of the experimental values of Hd for
Berl saddles7'9'17'18>25) are shown in Fig. 4. It is ob-
served that Hd increases with the Reynolds numberof
Fig. 2 Examples of measured pressure drop8: liquid based on dp (ReL), and with the decrease in dp,
342 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN
 Fig. 3 Relation between Hs and dv

Fig. 5 Correlation ofHd

Fig. 4 Relations between dynamic liquid holdup

and liquid Reynolds number

for the narrower void in the packed bed. Moreover,

Hd also increases with liquid viscosity (juL) because
of the difficulty of drainage through the packed bed
at high liquid viscosity.
Figure 5 shows the correlation of Hd. The experi-
mental values of Hd are shown by the solid line in
this figure, and the line represents the following equa- Fig. 6 Comparison between measured and cal-
tion.
culated Hd
Hd=2.90 x lO-'eRe^i^J^y^d-1-20 (2)
The values of Hd calculated from Eq. (2) are plotted 4£t -J£± , 41*
z - z + z
in Fig. 6 against the measured values. Although the (4)
measured values for Berl saddles tend to be slightly 3. 1 Dry pressure drop
larger than the calculated values and those for spheres It has been said that APd originates mainly from the
tend to be slightly smaller, it can be seen that most of friction of the gas which rises through the void of the
the measured values of Hd are correlated with Eq. (2) packed bed and is expressed by the following Fanning
equation2} :
within ±30 % in accuracy.
2. 3 Total liquid holdup
From Eqs. (1) and (2), HL is expressed as follows.
HL = Hs -\-Hd
where / is the friction factor. The values of/ cal-
={l.53 x l0-i+2$0x l0-5eRe^%VLlVw)0-n}d-1-20
culated from Eq. (5) by the use of the experimental
(3) values are plotted in Fig. 7 in the form of/versus the
Reynolds number of gas based on dp (ReG). Although
3. Pressure Drop the friction factor (/) decreases and approaches a
The total pressure drop (pressure drop with irri- constant value with increasing ReG, an approximating
correlation by the solid lines in the figure will be
gation) per unit height of the packed bed (JPT/z) are made. We obtain the following equations; for
correlated with the following equation as the sum of
ReG <200,
(APd/z) and (JPw/z) by use of the experimental values
of the manyprevious studies. f= U4Rea0'742 (6)
VOL 12 NO.5 1979 343
 Fig. 7 Correlation of dry pressure drop

Fig. 10 Exponents ofugt

and for ReG>200,
f= 6.S5Rea 0-216 (7)
Figure 8 compares experimental values with calcu-
lated values of (JPd/z). It is obvious that the ex-
perimental values of (JPdfz) are correlated within
dz30 %accuracy for the most part.
3. 2 Total pressure drop
Figure 9 shows examples of the values of (JPw/z)
for spheres defined as the difference between the
Fig. 8 Comparison between measured and cal-
culated APd values of (JPT/z) measured by Kolar et al.m and the
values of (JPd/z) calculated by the use of Eqs. (5) and
(6) or (7). In this figure, the abscissa is the net gas
velocity based on (e-HL) and the broken points are
the loading points. The slope (C) of lines for the
various packings is plotted in Fig. 10 against uu. For
the region of uu where C is constant, (JPw/z) is ex-
pressed as follows.
(APwlz) *z {uJ{e -HL)Y (8)
Figure ll shows the relation between (JPw/z), ugc and
HL for Raschig rings. A similar relation is found for
other packings and {APwjz) is expressed by the fol-
lowing equation.
A
PW rri
^kJ w
For the various packings, the values of the coefficient
(k) and the applicable ranges of (uul(s-HL)) are
shown in Table 3.
From
lows.
Eqs. (5) and (9), (JPT/z) is expressed as fol-
APT 4// ugc VfPa-
Fig. 9 Examples ofAPW dp
feK^W^-)2 <io>
344 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN
 Table 3 Values of k and usable ranges of Eq. (9)
Packing ulcl(e - HL) [m/hr] k
Raschig ring <1.00x l02 9.30x l05
Pall ring <1.80x l02 5.22x lO4
Berl saddle <1.00x l02 1.65x lO4
Intalox saddle <4.00x 102 3.98 x 104
Sphere <1.00x l02 1.34x lO6

Figure 12 shows the comparison of experimental values

with calculated values of (APT/z). The values of
(APd/z) in Eq. (10) are calculated by the use of net
gas velocity based on (e-HL). Equation (10) is
useful in the ranges ofvariables shown in Tables 1 and
3. The mean relative deviation of the calculated
values of (APT/z) for the various packings is 33.3 %.
On the other hand, the previous correlations8'11'1^
are inaccurate, compared with Eq. (10).
4. Discussion
Fig. ll Correlation ofJPW
The correlations for liquid holdup and pressure
drop of a packed column are compared with those of
a plate columnwithout downcomerand are character-
ized in the following sections.
4. 1 Liquid holdup
It was found that HLwas a function of uu, dp, e,
pL and fiL9 and was expressed by Eq. (3) for a packed
column. On the other hand, it has been reported
that in a plate column23} HL is dependent on the
change in the gas and liquid velocities, the plate
geometry (Ah/AC9 etc.) and the physical properties of
the liquid. As e and Ah/Ae are geometrical factors
equivalent to the void in the column, it is concluded
that the effects of the factors on HL are similar in both
columns.
4. 2 Pressure drop
The total pressure drop is analogously expressed as
the sum of APd and APWin both columns. APd is
characterized by the fact that it originates mainly
Fig. 12. Comparison between measured and
from the friction of the gas rising through the void in calculated APT
a packed columnand the sudden change of the pas-
sage of gas flow in a plate column. Therefore, APd Conclusions
increases in proportion to the 1.25-1.80th power of
ugc in the former and ugc squared in the latter, respec- The pressure drop of a packed column was inves-
tively2^. The geometrical factors influencing APd are tigated by the use of the experimental data of many
e and dp in the former, Ah/Ac and the hole diameter in previous studies, and the following conclusions were
the latter22K It is concluded that the effect of the obtained.
geometrical factors is similar in both columns as well 1) The total pressure drop is correlated as the sum
asHL. of the dry and wet pressure drops.
In a plate column, APWis a function only of HL 2) The dry pressure drop, originating mainly from
influenced by ugc. However, it is a function ofHL the friction of the gas rising through the void, is ex-
and ugc in a packed column, because the experimental pressed by Eqs. (5) and (6) or (7).
values in the region where HL is not influenced by 3) The wet pressure drop is a function of HL and
gas velocity are used in the data rearrangement. ugc, and is expressed by Eq. (9) and Table 3.
4) The liquid holdup is expressed by Eq. (3) as the
sum of the static and dynamic holdups.
VOL. 12 NO. 5 1979 345
Nomenclat ur e 5) Burke, S.P. and W.B. Plummer: Ind. Eng. Chem., 20,
1196 (1928).
Ah/Ac = ratio of total hole area to columnarea 6) Eckert, J. S.: Chem. Eng., 14, 70 (1975).
7) Elgin, J.C. and F.B. Weiss: Ind. Eng. Chem., 31, 435
dp = packing size
(1939).
/ = friction factor
= dynamic liquid holdup in packed column 8) Furnas, C. C. and F. Bellinger: Trans. AIChE, 34, 251
[m3/m3] (1938).
= liquid holdup on a tray or in a packed 9) Jesser, B. W. and J. C. Elgin: ibid., 39, 277 (1943).
column [m] or [m3/m3] 10) Kolar, V., Z. Broz and J. Tichy: Collect. Czech. Chem.
= static liquid holdup in a packed column Commun., 35, 3344 (1970).
[m3/m3] ll) Leva, M.: Chem. Eng. Progr., Symp. Ser., 50(10), 51
= dry pressure drop of gas [kg/m-hr2] (1954).
= total pressure drop of gas [kg/m-hr2] 12) Morcom, A. R.: Trans. Inst. Chem. Engrs. {London), 24,
= wet pressure drop of gas [kg/mà"hr2] 30 (1946).
= Reynolds number of gas based on packing 13) Oman, O. and K. M. Watson: National Petroleum News,
size (=dpugcpG/iuGe) [-] 36, R795 (1944).
Reynolds number of liquid based on 14) Otake, T. and K. Okada: Kagaku Kogaku, 17, 176 (1953).
packing size (=dpUicpLljLtLe) [-] 15) Otake, T. and E. Kunugida: ibid., 22, 144 (1958).
gas velocity based on cross-sectional 16) Prahl, W. H.: Chem. Engrs., ll, 89 (1969).
area of column [m /hr] 17) Shulman, H. L., C. F. Ullrich and N. Wells: AIChEJ., 1,
247 (1955).
Uic
liquid velocity based on cross-sectional
area of column [m /hr][m] 18) Shulamn, H. L., C. F. Ullrich,
ibid., 1, 259 (1955).
N. Wells and A. Z. Proulx:
height of packed bed
19) Simmons, C. W. and H. B. Osborn: Ind. Eng. Chem., 26,
void fraction of packed bed [-] 529 (1934).

viscosity of gas [kg/m - hr] 20) Standish, N.: Chem. Eng. Sci., 23, 945 (1968).
viscosity of liquid [kg/m - hr] 21) Stanek, V. and V. Kolar: Chem. Eng. J., 5, 51 (1973).
viscosity of water [kg/m - hr] 22) Takahashi, T., Y. Akagi and T. Kishimoto: Kagaku
Pg density of gas [kg/m3] Kogaku Ronbunshu, 3, 569 (1977).
PL density of liquid [kg/m3] 23) Takahashi, T., S. Sudo and M. Tanaka: Kagaku Kogaku,
34, 744 (1970).
Literature Cited 24) Uchida, S. and S. Fujita: /. Soc. Chem. Ind., Japan, 39,
1) Andrieu, J.: Chem. Eng. /., 7, 259 (1974). 432, 876 (1936); 40, 238 (1937); 41, 275 (1938).
2) Bird, R. B., W. E. Stewart and E. N. Light foot: "Trans- 25) Varrier, C.B.S. and K.R. Rao: Trans. IIChE, 13, 29
(1960-61).
port Phenomena", p. 183, 196, Wiley (1960).
3) Broz, Z. and V. Kolar: Collect. Czech. Chem. Commun.,
33, 349 (1968).
(Presented at the 43rd Annual Meeting of The Soc. of Chem.
4) Buchanan, J. E.: Ind. Eng. Chem., Fundam., 8, 502 (1969). Engrs., Japan, at Nagoya, April 1978.)

346 JOURNAL OF CHEMICAL ENGINEERING OF JAPAN