' .

Explore the Potential

of Air-Lift Pumps
t'.'

, ..::

and Multiphase
pplications of so-called "air-lift This is in accord with analogous experi­

A
pumps" in fields other than ence in the injection of flashing feeds into
petroleum (I) have included risers in fluid catalytic cracking units,
the handling of hazardous flu­ an alo gously carrying upwardly bulk
ids (2), the design of bioreactors (3,4), solids (particles) as opposed to bulk liq­
the recovery of archeological artifacts uids (molecules), and also accounts for
(5). recycle aeration in sludge digestors some amount of scatter in published
(6), deep sea mining (7), and the recovery experimental data on air-lift capacity.
of manganese nodules (8,9,10) from Since injected gas bubbles are dis­
ocean floors. Interest among a host of placed upwardly by the downward "slip"
domestic (//-17,18) and foreign (/9) flow of liquid (that is, dense phase (23))
organizations dates back several decades; at the walls, the cross section through an
Use this having passed EPA hurdles (20), recovery upward moving gas bubble represents a
new capacity of ocean resources has been further sanc­ countercurrent flow as depicted in Figure
tioned by U.S. legislation (/7). With the le despite a net upward transport of slugs
correlation for exception of "bioreactors" the practical of liquid. This local countercurrent flow
pump design. design and operation of an air-lift pump is analogous to the empty pipe flooding
lies in the dense-phase slug-flow regime phenomena illustrated in Figure Id for
of cocurrent gas-liquid upflow (21). which correlations already exist (24). In
view of the identical flow phenomena in
Figures le and Id, accounting for the sys­
The transition to slug flow tem variables should also be identical
The entrainment of spray or droplets whether the net flows are cocurrent or
f. A.Zenz. as occurs from a fractionator tray consti­ countercurrent
AIMS tutes a form of dilute phase cocurrent
gas-liquid upflow. Since bubble caps
have been used as foot pieces in gas lift Dense phase liquid
pumps, a pipe set over a cap on a tray, as transport by gas-lift pumps
illustrated in Figure 1 a, constitutes con­ In Figure I b, gas compressed to a
c e p t u a lly a low efficiency bubbling pressure level equivalent to the depth of
upflow gas lift operating at very shallow submergence is introduced to the lift pipe
submergence. In practice, a gas lift oper­ to displace liquid upwardly. As the liquid
ates with deeper submergence, and more flows back down into the bubble-form­
efficiently in slug flow, as depicted in voids, more gas is continually introduced
Figure I b where the foot piece is simply to establish a steady state of refluxing of
a gas injection nozzle. upwardly displaced liquid. The net rate of
In practice, multiport injectors have conveyance or refluxing decreases with
been found to be more efficient (22) in height of lift. If the pipe is cut off at some
terms of increased capacity, particularly height, below the maximum commensu­
in the bubbling and slug-flow regimes. rate with the given gas rate, or of the dif-

CHEMICAL ENGINEERING PROGRESS • AUGUST 1993 • 51

 FLUIDS/SOLIDS HANDLING........,

(bl 104 l>j.-<

fa) ·.

Vo �
t ri-VL
' .

11)3 .·� 11J3

-.,.o · > .
.
·. ·· · ·
:'��' A= Pii>u cross·seelion, sq. h.
D PiJ!e Ld� inches ·

l .. Lilt. h.
"
..

S .. Submerguncu, h.
_·
.

p6 Gas density
. .

. ,-'' ' � ' ' �- '1·:·} "

PL = Fluid density
;' • • • .... :
. �. " f

101 101

' . \· , :,_•'

. .

.
'· f: .'�t,� · .. - '..' .

Vo'· VL."
-
, , •

--.:....

•Figure 1. (above) Dilute-dense phase manifestations. 10 . , ... 10

10 100 1000

• Figure 2. (right) "Con-elation" ofair-waler lift data.

ference in hydraulic head, then the despite the fact that air lift represent then the data in Figure 2 yield a fam­
net rate of liquid at that height will overall a net cocurrent , rather than ily of parallel curves spaced accord­
be continuously expelled from the countercurrent, flow. ing to the square root of the lift
pipe and hence the action referred to In Figure 2 gas volume and gas height. Multiplying the ordinate and
as a lift "pump." density are based on discharge con­ dividing the abscissa by the square
ditions; within a long lift pipe these root of L results in the correlation of
could vary significantly from inlet to Figure 2.
Experimental air-lift data discharge. However, the significance
Figure 2 displays a variety of pub­ of "corrections" to the data in Figure
Ii shed a i r-lift data c overing lift 2 would be difficult to evaluate justi­ Effect of fluid density
heights ranging from 5 in. to 65 ft fiably when even the best data are Since all the data in Figure 2 are
and pipe diameters from � in. to 15 subject to experimental error and to based on lifting water, the question
in. The curves drawn through each the effects of the air injection arrang­ arises as to how well this correlation
investi gator's results f all into a ments (22). would satisfy other fluids. The ordi­
numerical sequence with height of The submergence term, S, in the nate should be modified by consid­
lift. The basis for the choice of coor­ denomin at or of the ordinate of ering that increased liquid density
dinates in Figure 2 lays principally in Figure 2 is based on analogy to cor­ would logically result in reducing
their ability to correlate maximum relating data on entrainment from the effective volumetric yield, or
countercurrent dilute phase-dense distillation trays (25). It also reflects conversely giving the same wieght
phase flow through empty vertical the work expanded in compressing yield at equal air rates. Chamberlain
tubes and packed towers (23,24). the air in an air lift. If S is replaced (2) obtained air-lift data for both
T hat these coordinates yield a plot as by log [(S + 34)/34 ] which is pro­ , water and a 93.5 lb/ft3 caustic solu­
organized in Figure 2 lends credibili­ portional to the work of compression tion in the same pipe; his data are
ty to this approach to correlation, and still reflects the submergence, plotted in Figure 3 and show excel-

52 • AUGUST 1993 • CHEMICAL ENGINEERING PROGRESS

lent correlation. Figure 3, which rep­ \.6X4 X 144 1
Abscissa 0.0765 48
resents a smoothed ban d drawn x x- =O 0916 �
through the data in Figure 2 may ofFigure4 3.14(6)2 62.4 6 30 . 'G

therefore be considered a more gen­

•EquationA
eralized correlation.

The Ingersoll-Rand equation

\f; L
'¥- 144x14 .7 Exln [(34+S)/34]
The theoretical efficiency of an
air lift is simply the work required
•Equat ions
to lift the liquid to the point of dis­
c h arge divided b y t h e work of
isothermal compression of the air. (SCFM)Air 0.8 L
When analyzed in these terms, the (GPM)Water - 469 Exlog [ (34+S)/34]
air-to-liquid ratio can be expressed
as a function of lift height, submer­ •Equalion 6
gence, and efficiency. This relation­
ship is generally recognized as the (SCFM/fr2)(48/D)112 0.8 L
Ingersoll-Rand equation (26,27) =

and is usually accompanied with (GPMJJt2)( 48/D)112 Cxlog [(34+S)/34]

t a b ul ated v al ues o f a constant •Equation 7
derived from experimental data at
or near the point of maximum effi­
(VdA)(48/D)112 (pdp!)1121L112 0.8 (pdpJ
ciency of operation as a function of ------ =

the lift height. (\f.IA)(48/D)1f2 L lf2/log ([ S +34}':34] c

The work required to lift the
dense phase (liquid) to the point of •Equalion8
discharge is: alone, it is impossible to predict the flow regime) or greater (in the annu­
variation in a given lift pipe's per­ lar or mist flow regime), than near
W0= W,.L (1)
formance as air rates are changed. the point of optimum slug-flow
The work (isothermal compression) The equation gives the air-to-water operation (28).
expended by the dilute phase (air) in ratio only at the point of peak effi­ Multiplying numerators and
lifting the liquid is: ciency and would imply that this denominators by equal terms, the
ratio is constant at all air rates. It is Ingersoll-Rand equation can be
W1 =Pa VG In (PIP) (2) evident from Figures 2 and 3 that the rearranged for graphical comparison
and the fractional efficiency, E, is water rate falls off sharply when the with Figures 2 and 3. From Eq. 6 see
therefore: air rate is either less (in the bubbling Eq.7 or Eq. 8 . For standard air and

E= WJW1 (3 )
Since Pa= 34 ft of water or 14.7 psia, 106
P may be expressed as 34 + S, where
S is the submergence in ft, so that �
-
Data of Chamberlain
[) = 1.6"
L= 10ft.
E= W,. UPa VG In (PIP0) (4)
(See Equations A, 5 and 6.) The pub­ Band of data
lished version of the Ingersoll-Rand - in figure 3. •Figure
equations is simply Eq. 6 with the 3.Effecto/
term (469 E) replaced by a constant fluid
o S = 10ft. at PL= 62.4 density.
which in effect amounts to assigning '
• Self= 1 Sft. at PL = 93.5
a value to E. Operationaily deter­
. .

mined values of E are generally of

1� .____;_..;,_
the order of 40 to 50% at the point of
____ _
L-________...__________J

1 to 100 1,000
maximum operating efficiency, as
given in Table 1 (26). Vo�
From the Ingersoll-Rand equation A VDLPL

CHEMICAL ENGINEERING PROGRESS • AUGUST 1993 • 53

 'r\\,:;.\
FLUI OS/SOLIDS HANDLING "'-"
���;.:.:;.;.:;.:.;.;.;.;;-.-
----------
-
-
-
-
-......
�.��:

1oe----�---�,-----.

Ingersoll-Rand
equation -,..___,.,...,,
'.·� 10Wgt % Sand Fluid
::· . soWgt % Water Rate
E=50% ·

PG=0.0765
Air Air
Rate Rate

Pr= 62.4

101"'========='====�
' 10 100 1,000

VG j4iP;
A v 0LPr
•Figure 4. Significance ofthe Ingersoll­
Rand equation. •Figure S. A ir-lift examples.
10&
water (that is, PL = 62.4; Pa =

Nomenclature 0.0765), Eq. 8 is then

0.028
a cross-sectionul area of annular (9)
..---- 1-Bubbling

�I�
now, sq. ft. c
105
A = pipe cross-sectional area. sq. ft. Equation 9 is shown in
D pipe inside diameter, in. F i g u r e 4 as s up e r i m- - I,
2 '" �? ,
=

fractional efficiency, WJW1 ,, L

E posed on Figure 3. The c:::n
I;
..5! 1 .. �- lt.-A
L lift height, ft. Ingersoll-Rand equation -- Cl ..
�:
p = pressure at point of gas entry. plots as a straight line t?
-- /· Slugging I LA...
lb/sq. ft. with slope correspond- ::I. �
co
104 ... �
d: Ill
P. atmospheric pressure, lb/sq. ft. ing to the ai r-to-water ::I
b
'
c
s submergence, ft. ratio at peak theoretical ,...
N
.i �
=

�
VG
= gas now at discharge conditions, efficiency. T h e c u r v e
I rT
ft'lmin b ased on experimental .
VL
= dense phase now. gal/min data exhibits this same 103
w, pounds of dense ph11SC lifted per slope at only one point, >'I< Mist �
minute but then curves away to
w, worlc input ft-lb/min
yield lower water-to-air
w.. work output ft.lb/min
ratios at values of the abscissa
=

(air rate ) higher or lower than 102

Greek Letters 1 10 100 1,000
that corresponding to the ratio at
this point of maximum delivery Vo �
£ = gas void fraction in aerated
dense phase
efficiency. The Ingersoll-Rand A v OLP;'
equation is therefore useful in •Figure 6. Multiphase cocurrent
viscosity of lifted fluid.
µ, estimating the yield from an air
=

entipoise
vertical upjlow regimes.
lift only under conditions of peak
�
= surface tension of lifted fluid,
theoretical effici ency, b u t n o t Examples ;'.
dynes/cm
over the entire range of possible Consider the operable gas-liquid­
P1 = density of lifted fluid, lb/ft'
operating conditions . Figures 3 solids ratios for two situations illus­
Po gas density. lb/ft'
a n d 4 i n c o r p o r a t e as w e l l the trated in Figure 5:
Pt liquid density. lb/ft'
effect of fluids of densities other 1. Air l ifting of water 30 ft
than water. through a 6 in. pipe submerged 50 ft.

54 • AUGUST 1993 • CHEMICAL ENGINEERING PROGRESS

'
-----------· - - ____________ f
 Ordinate
=
\fX4 x144 /f -
8
x
(30)
112
=201 v,
o/Figure4 3.14(6>2 6 /og{(34+50}'34] 'L
•EquationB

Abscissa '6x4 x144 0.0765 48 -1-

= x x = 0.0844 (ACFM)
ofFigure4 3.14(6)2 66.5 6 33.1
•Equation C

Ordinate =(SlurryGPM)x4x 144 {48 66.5

x
(33.1)112
=235(SlurryGPM)
ofFigure4 3.14(6)2 V 6 62.4 log{(34+ 46.9}'34]

•Equation D

2. Air lifting of a sedimented

sand, as a slurry containing 10 wt.%
fine sand, a distance of 30 ft through
a 6 in. pipe submerged 50 ft.
Case/.
See Eqs. A, B and Table 2.
Case 2.
Slurry density= l/((0.1/165)
+ (0.9/62.4))
= 66.5 lb/ft3
Effective submergence=
50 (62.4/66.5)= 46.9 ft.
Effective Lift = 30 + 50- 46.9= 33.1 ft.
See Eqs. C, D and Table 3.

Bubbling and annular

liquid flow
At values of the abscissa less than
about 50 and to the left of the curve
in Figure 4, two-phase flow would
occur in the form of gas bubbles ris­
ing in a matrix of liquid forced to
flow upwardly by means, for exam­
ple, of a motor driven pump or sim­
ply by a hydraulic difference in head.
Under such conditions, typical in
bioreactors (3), L would be the tube
diameter and S would equal

Lp/ (1 - e}.162.4
At values of the abscissa greater
than about 50, and to the right of the
curve in Figure 4, voidage must be
so great that liquid can only be dri­
ven up the pipe in annular flow by
the surface drag of the gas core. This
region has been explored in some
detail by Dukler and others (29.30)

CHEMICAL ENGINEERING PROGRESS • AUGUST 1993 • 55

 FLUIDS/SOLIDS HANDLING

� -·: :,.'.,:-.-

. · Acknowledgment : �X Literature Cited

·�
r-�..:_
.·��
-·�: �:/'·�i.
. �\���
�
�ldr. Sylvan Cromer's kind permission to publiSh/ 19. Chem. Week, pp. 33-J4',;-::;;'
.
.

td.io_se portions of this article developed under 'a> (July 30, 1980). :-;:\;,·-
;'.;c-Ontract with Union C a rbide Corporation ls:� 20. Chem. Eng p. 20 (No'
. •

1980 ).
';1ir3t�rUlly acknowledged. . . ))' :-·

21. Padan, J, W., "The' .,,

��;./�:_.�.--:' �. ... :���}.� Mining Research � .�··
;1t��· :;:._ : ' \ . : >\.�:��t�t(� Bureau of Mines, Tit(.
iJ.=J.A.:ZENZ is president and technical dir�cw:�'ft
._·.

(May 1965).
J�
'��JMS,.Garrison, NY (914/424·3220; Fax: 914/414���. 22. Morrison, G. L., et · ·

,
' )_a non· profit industrial consortiu m:](''\' Res., 26, 387-391 (19 "
23. York, J, L., et al., ;··'
.. ,.•

' . ed his PhD in chemical engineering ..

\vt
� echnic University, New York;';� ;. �'l Column Flows Waill
1!$0r Emeritus at Manhattan ��JJ,ifllfii.'. Equation," Chent. - ·

ow"of the AIChE, the 1985 recipianf i>t � 88(10), pp. 93-98'(('
ljfrm
d in chemical engineering p����. 24. Zenz, F. A., and ·. �
�).$86 recipient of Chemical Engi11i�T,tJ8!Jl: Hydocarbon Proc .
-�. anal Achievement Award. In the p��t�PJt pp. 121-126 (1965):{
. .he has authored 90 papers, 2bookS, 19 U�Sit 25. Schweitzer, P. A;,·
:l
�� mot[' •-,,. .
_

:� : '.�nd has been 8 consultant to Separation Techni · · '

J and domestic corporations. .\'.�f�'J,
fciriiign Chemical En8ii1 �·
·:. '
' ::�:h?���-- .�· Hill, New Yorkpp
(1988). <�
26. "Chemical Etl:'
who presented experimental data relating Handbook," 4th
measured amounts of liquid, carried upward Hill pp. 6-13; Cf,.
i n annular films of measured thickness Ingersoll·Rand .ll!J!i
inside a vertical 2 in. internal diameter (ID) (July 1929) ·,7,'.�
.

�'.
•. ·

pipe, to the measured rate of the cocurrent, 27.Duckworth, C.

liquid flow inducing, upward gas stream. p. 63(March1960} ··

28. Nemet, A.G., JmJ.o

The results are shown in Figure 6 with
53(2), pp. 1SHS4,'
the ratio of film cross-sectional area to pipe
29. Zab G et- ., �·
t ..
. .

cross-sectional area as a parameter. aras,

Upward .. ,,,.,.
Cocurm..i .

Equating bubble rise velocity to displacing Annular Flow,'' Ai

liquid downflow in a slugging tube, corre­ PP· 829-843 098ii
'
_

sponds to an annulus-to-tube area ratio of 30. Taltel, Y. et � ." -

0.385, suggesting that these investigators Patterns Transit!��,.

1
were approaching slugging and that this Upward Gas-Uqui4.F.'!J
Vertical Tubes,':'Al
limit should at some time be established on
pp. 345.354 (1980).�::t..,
Figure 6 experimentally. The addition of
31. Baker,o.,0;1'&diii ,: '" ·
liquid viscosity and particularly surface ten­ 195, 53(12)(July 26$f
sion terms to the ordinate of Figure 6 has
·

32. Zenz, J. A., pm01ilif

been, somewhat arbitrarily, based on identi­ cation, PEMM Cori{�..
cal terms applicable to horizontal multi­ NY (1986). · : �:�i:
' � :-.�. --
phase flow (31) and on data obtained for
very low surface tension liquids in cocur­
rent upflow with air (32) in tubes several in. phase upflow to the extent supported
to several ft in diameter. by data published to date . Di1
Fi g u r e 6 r e p r e s e n ts n o t o n l y a
capacity correlation for gas lift pump To receive a free copy of this article, sand
design, but also suggests the correla­ in the Reader Inquiry card in this Issue
tion of all regimes of cocurrent multi- with the No. 136 circled.

56 • AUGUST 1993 • CHEMICAL ENGINEERING PROGRESS