Aquacultural Engineering 6 (1987) 97-110

Design of Airlift Pumps for Water Circulation and

Aeration in Aquaculture

Nick C. Parker and Mary A n n a Suttle

US Fish and Wildlife Service, Southeastern Fish Cultural Laboratory, Marion,

Alabama 36756, USA

ABSTRACT

Water flow rates were measured in airlift pumps 3"75-30 cm in diameter

to develop performance data that might be useful to aquaculturists. Flows
were determined when submergence of airlifts was 100% and when the
center-line of the discharge was between 12"5 cm above and 5 cm below
the water surface. Air was injected at 15-cm intervals from 15 cm to
120 cm below the discharge of the airlift and airflow was varied from 28
to 1416 liters rain- 1. A n increase in the vertical lift reduced flow rates
greatly in large-diameter pipes, but only slightly affected flow rates in
small-diameter pipes. Water flow increased linearly as airflow increased
logarithmically.

INTRODUCTION

Water circulation and aeration in aquaculture ponds have increased

primary productivity, reduced stratification, increased nutrient solubili-
zation, reduced organic accumulation on the bottom, and increased fish
production. Pond aeration techniques have been investigated to increase
the growth, survival, and production of both fish (Ito et al., 1974; Sarig
and Marek, 1974; Parker, 1979, 1983) and crustaceans (Morrissy, 1979;
Apud and Camacho, 1980). Airlift pumps of various sizes and con-
figurations have been used to circulate and aerate pond water, but due to
fluctuating water level in ponds, not all systems have worked reliably and
efficiently. The design and flow predictions for airlift pumps have
typically been based on data derived from small systems suitable for
aquaria and tanks, or from performance charts showing the vertical lift
capacity of airlifts that are 40-90% submerged (Spotte, 1970; Castro
etal., 1975; Murray etal., 1981).
97
98 N. C Parker, M. A. Suttle

Ivens (1914) reported that airlift pumps were tested in a laboratory in

Germany as early as 1797 and that their first practical application in the
United States was in pumping oil in Pennsylvania in 1846. In the aquatic
sciences, airlift pumps have been used to collect samples of seawater
(Tokar et al., 1981); to provide aeration, agitation, and pumping in
sewage treatment plants (Andeen, 1974); and to provide aeration,
destratification, and pumping for aquaculture (Rappaport et al., 1976;
Wheaton, 1977; Parker et al., 1984).
Several investigators have reported the flow rates of small-diameter
airlift pumps used to lift water vertically. Spotte (1970) presented data on
the vertical lift capacity of airlift "pumps 2.5-15 cm in diameter and
40-70% submerged. Spotte (1979) revised the water flow rates from his
earlier edition and also included flow rates of air. Castro et al. (1975)
reported on the pumping rate of airlift pumps 1.27-7.62 cm in diameter,
30 cm to 3.7 m long, and 40-70% submerged. Todoroki et al. (1973)
developed equations for predicting the flow in airlift pumps 2.5-10 cm
in diameter, 4-42 m long, and 40-80% submerged.
Airlift pumps used to circulate water in ponds operate almost totally
submerged and need to move water only from the bottom of the pond to
the surface. The theory of operation and equations describing per-
formance for airlift pumps operated in this mode have previously been
reported by Nicklin (1963). Murray et al. (1981) defined the nomencla-
ture used to describe airlift pumps, discussed theory of operation, and
presented performance data on pumps of 1-78-3.65 cm in diameter
operated at 50-80% submergence. Performance data on airlift pumps
operated at 90-100% submergence have not been presented previously.
The first objective of our study was to determine the influence of three
variables -- pipe diameter, depth of air injection (commonly referred to
as submergence ratio), and volume of air injected -- on the water flow
rate of airlift pumps suitable for use in aquaculture ponds. Our further
objective was to report these data in a format reflecting actual and
expected performance under field conditions.

MATERIALS AND METHODS

Water flow rates were determined in eight airlift pumps 3.75-30 cm in

diameter into which various volumes of air were injected at different
depths. Replicate measurements were made under each test condition
until three similar readings were obtained. The mean of these three
readings was accepted as the flow rate under the test conditions.
 Airlift pumps for water circulation and aeration 99

Airlift pump operation

Airlift pumps with nominal diameters of 3.75, 5.0, 7.5, 10, 15, 20, 25
and 30 cm (actual internal diameters were 4.4, 5-2, 7.6, 10"2, 15-4, 21.0,
26.4 and 31-3 cm) were constructed from polyvinyl chloride (PVC) pipe.
Each pump consisted of a vertical section of pipe fitted with a 90* elbow
at the upper end (Fig. 1). Air was injected into the side of the vertical
pipe through a series of 1.25-cm holes placed at 15-cm intervals from 15
to 120 cm below the horizontal arm of the PVC elbow. There was only
one hole per 15-cm interval in 3-75-15 cm diameter pipes, two holes per
interval in 20 cm diameter pipes and three holes per interval in 25 and
30 cm diameter pipes. Each 1.25-cm hole was tapped to receive a 1.25-
cm hose adapter. The bottom of the vertical pipe (the airlift pump intake)
was 130 cm below the horizontal arm of the elbow (the airlift pump dis-
charge).
The bottom of the discharge elbow was placed 1.25 cm above the
water level in a flume, located just above the water surface (Fig. 1) and

Ptel|ure gauge
T Air * walter
separation Flume

1.25 cm

A*¢ pump ..--.b~

1,25 ¢m Air line ....... e

• 130 cm

e~

Fig. 1. Test apparatus for measuring water flow in airlift pumps. Each pump consisted
of a 90 ° elbow and a vertical riser. The discharge was 1.25 cm above the surface of the
water and was measured volumetrically or in a flume after passing through an egg-crate
type baffle to reduce turbulence. For each test air was injected into one of eight 1.25-cm
holes at 15-cm intervals in the vertical riser, while the other seven holes were plugged.
100 N. C. Parker, M. A. Suttle

water flow was measured as air was injected into the 1-25-cm holes. Air
volume was regulated with a series of rotameter type flow gauges (Dwyer
Instruments Inc., Michigan City, Indiana, USA), and adjusted from 28 to
224 liters min- 1. Values were not corrected to standard temperature and
atmospheric pressure (STP) conditions; measurements were made when
water and air temperature was about 20-25°C. Reported air flow values
may have a 7-8% error as compared with values at STP conditions. Air
was injected at only one depth at a time while the other 1.25-cm holes
were plugged. For airlift pumps 3.75.--10 cm in diameter, water flow was
measured by volumetric displacement during a timed interval. For airlift
pumps 15-30 cm in diameter, water flow was. measured in
Palmer-Bowlus flumes (manufactured and calibrated by Plastic-Fab Inc.,

Plastic bag

.sf

~'~ Pump
discharge

Fig. 2. Arrangement of plastic bag for collection of airlift pump discharge for
volumetric measurement. The bag was quickly slipped under the lower edge of the
discharge pipe (A) and arranged to allow escape of air as discharge filled the bag (B).
 Airlift pumps for water circulation and aeration 101

Tualatin, Oregon, USA). Water depth in the flume was measured and
flow rates were taken from tables published by Grant (1979). All
measurements of water and air represent the mean of three independent
measurements made at approximately 5-min intervals after flows were
stabilized.
We analyzed all data using Hewlett-Packard software packages in a
Hewlett-Packard-9845 computer. Best-fit equations were selected from
linear, logarithmic, exponential, and polynomial regression models to
predict water flow rates when air flow varied over the range of 28-224
liters min-
Flow rates were calculated with linear, logarithmic, exponential, and
polynomial regression equations for air volumes up to 1416 liters min- ].
Projected water flow rates at air flows above 224 liters min-] (up to
1416 liters rain- ~) were then verified in 5 and 10 cm diameter pipes by
increasing the air volume from 224 to 1 138 liters min-~ and measuring

Air Flow (cubic feet per minute)

1 5 10 50

ill
160 3.75 cm A 40
B

~" 80 20
I

'E 0 P---~, . . . . . . . . . . . . . 0
ce 25 100 250 1000 la
1 5 lO so
- L . . . . . . . . . . . . . .
240f 5 cm _A 60 ca
c .

'7.. 1 8 o I 4o ;.

0 0
25 100 250 1000
Air Flow (liters min -1)
Fig. 3. Water flow rates in airlift pumps with diameters of 3.75 cm (upper panel) and
5.0 cm (lower panel) when air was injected into the vertical riser at different distances
(cm) below the surface of the water: 120 (A), 105 (B), 90 (C), 75 (D), 60 (E), 45 (F), 30
/G). and 15 (H). All measurements were made with the flow-line of the discharge
1.25 cm above the surface of the water.
102 N. C Parker, M. A. Sunle

water flow. The effects of depth of submergence and vertical lift were
also evaluated. Air at a flow rate of 85 liters min- 1 was injected into air-
lift pumps with diameters of 5"0, 7-5 and 10 cm at a depth of 60 cm
below the center-line of the discharge from the pump. Under these con-
ditions, flow rates were measured when the center-line of the airlift dis-
charge was at the surface of the water and when elevated 2.5, 5"0, 7.5, 10
and 12.5 cm above the surface or submerged 2.5 or 5.0 cm below the
surface. The maximum submergence was limited to slightly less than the
diameter of the airlift pump tested; at no time was the top of the
discharge pipe completely submerged. Volumetric measurements were
made during the submergence tests by collecting airlift discharges in
large plastic bags (with tops open to allow air to escape) held under the
discharge pipes (Fig. 2).

RESULTS A N D DISCUSSION

Logarithmic regression provided the best-fit model for empirically

derived water flow rates for airlift pumps; data were plotted on a semilog

Air Flow (cubic feet per minute)

1 5 10 50
• , • , . , . . l . , • • , i A
140
7.5 cm /
50O
B
120 1
400 C IO0
D
o
t..
0 E m

" C
300 80 o
F
60
200 • G I

0
,o i
100 @

I0 "
25 100 250 1000
Air Flow (liters rain " 1 )

Fig. 4. Water flow rates in an airlift pump 7-5 cm in diameter, when air was injected
into the vertical riser at different distances (cm) below the surface of the water: 120 (A),
105 (B), 90 (C), 75 (D), 60 (E), 45 (F), 30 (G), and 15 (H). All measurements were made
with the flow-line of the discharge 1-25 cm above the surface of the water.
 Airlift pumps for water circulation and aeration 103

scale (Figs 3-9). Flow rates in 5 and 10 cm diameter pipes with air flows
of 2 8 - 2 2 4 liters min- 1 were used to project water flow beyond the range
of the initial test. The predicted flow rates were verified in additional
laboratory tests of 5 and 10 cm diameter pipes when air flow was 1 138
liters min-1. In airlifts of all sizes, flow rates increased as either air flow,
the depth of air injection, or both, increased. The no-flow conditions
described by Murray et al. (1981) were not approached in our test even
in the small-diameter pipes. An increase in air flow rate increased water
flow rate in all size pumps even though the increase was slight in small-
diameter pipes.
Slight differences in submergence of airlift pumps altered water flow
rates significantly. Water flow was highest when the center-line of the
discharge was at or slightly below the surface of the water (Figs 10-12).
Flows decreased as the discharge was elevated above the water surface.
Pickert (1932) cautioned that only flows from airlifts of identical
submergence and length could be compared with each other. Recogniz-
ing that caution, we compared the flow rates of our 3.75 cm diameter air-

Air Flow (cubic feet per minute)

1 5 10 50
600 . . . . . . . . . . . . . .

10 cm
A 140 •
soo C

_¢ 120 E

400 E ,-,
100 =
•~ c
F 0
300 . G 80 "

o
u. 60
; 200 o

(u 40 u.

100
20

0 0
25 100 250 1000

Air Flow (liters rain -1)

Fig. 5 . Water flow rates in an airlift pump 10 cm in diameter when air was injected into
the vertical riser at different distances (cm) below the surface of the water: 120 (A), 105
(Bt, 90 (C), 75 (D), 60 (E), 45 (F), and 30 (G). All measurements were made with the
flow-line of the discharge 1.25 cm above the surface of the water.
104 N. C. Parker, M. A. Suttle

Alr Flow (cubic feet per minute)

1 5 10 50

180
1000

:40 ,.'

800 E
.-" ,,oo ;
0 O.

'- m
S ,-
e 600 60 o
• 01
• ,* ot
_ v

20
400 o
O
Ik

; ~o ;
e
• 200 •
~0

O
25 100 250 1000

Air Flow (liters rain "1)

Fig. 6. Water flow rates in an airlift pump 15 cm in diameter when air was injected into
the vertical riser at different distances (cm) below the surface of the water: 120 (A), 105
(B), 90 (C), 75 (D), 60 (E), 45 (F), and 30 (G). All measurements were made with the
flow-line of the discharge 1.25 cm above the surface of the water.

lift p u m p with air injected at 73 liters m i n - l, 60 and 120 cm below the

surface (Fig. 3) with flow rates reported by Castro and Zielinski (1980).
O u r flow rates were 51 and 77 liters m i n - l, respectively, for the 60 and
120 cm d e p t h of air injection, c o m p a r e d with the rates of 73 and 98
liters min-~ reported by Castro and Zielinski. T h e s e differences were
probably due to differences in submergence. O u r m e a s u r e m e n t s were
m a d e with the b o t t o m or flow-line of the discharge pipe 1.25 cm above
the surface of the water. Castro and Zielinski measured flow at 100%
submergence, presumably with the center-line of the discharge pipe at
the surface of the water.
T h e flow rates presented in Figs 3 - 9 are about 1 0 - 4 0 % less than the
m a x i m u m flows obtainable from airlifts of similar size with similar rates
of air injection. For example, in a 5 cm diameter airlift operated with 85
 Airlift pumps for water circulation and aeration 105

Air Flow (cubic feet per minute)

1 5 10 50

80

1000

40

800 E
- oo ;
C

m 600 60 _o

~ 20

400 o
0
D

o ;

20o
o

C
25 100 250 100o

Air Flow (liters min "1)

Fig. 7. Water flow rates in an airlift pump 20 cm in diameter when air was injectexi into
the vertical riser at different distances (cm) below the surface of the water: 120 (A), 105
(B), 90 (C), 75 (D), 60 (E), 45 (F), and 30 (G). All measurements were made with the
flow-line of the discharge 1.25 cm above the surface of the water.

liters min- ~of air injected at a depth of 60 cm below the center-line of the
discharge, the flow rate was about 12% greater when the vertical lift was
zero than when it was 1.25 cm (Fig. 10). The flow in a 7-5. cm diameter
airlift increased about 16% when the vertical lift was reduced from 1.25
cm to zero (Fig. 11 ). Changing the vertical lift in a 10 cm diameter airlift
from 1.25 to zero increased flows about 38% (Fig. 12). The depth of
submergence or, conversely, the vertical lift, affected flow rates in
proportion to the diameter of the airlift. The effect of change in depth of
submergence on flow was only slight i n small-diameter airlifts, but was
very substantial in large-diameter pipes.
In our test, flow rates in 7.5 cm airlift pumps were increased about
5 - 1 0 % when air was injected through a series of 1.6-mm holes around
the pipe rather than through a single 1.25-cm hole. However, the
106 N. C. Parker, M. A. Suttle

Air Flow (cubic feet per minute)

1 5 10 50

120i 120

~ao
.-" 1 001 .c
, E
C
--
S ~
" 40 ;
IIL
el
,- 801
• !00 ¢
" 0
v

]= 60~ 60 ~=a
o
u. I
o
20 --
u.
"" 401

20t
|o

L .)
25 100 250 1000

Air Flow (liters rain "1)

Fig. 8. Waterflowrates in an airliftpump 25 cm in diameterwhenair was injectedinto

the vertical riser at different distances (cm) below the surface of the water: 120 (A), 105
(B), 90 (C), 75 (D), 60 (E), 45 (F), and 30 (G). All measurements were made with the
flow-line of the discharge 1.25 cm above the surface of the water.

increase in flow rate was too small to justify the additional expense or
effort required to provide this radial air injection. Ward (1924), who
tested different types of air injectors, concluded that there should be no
central air injection nozzle or other obstruction to the flow of water. He
found that special devices were not required to mix air with water in
airlift pumps. He also reported that small fine bubbles provided no
advantage because small bubbles quickly coalesced into larger bubbles
as they traveled up through the water column. Threaded holes fitted with
90* hose adaptors have worked well as injection ports in our tests and in
subsequent field applications at this laboratory and in field tests in US
Fish and Wildlife Service National Fish Hatcheries in San Marcos,
Texas; Natchitoches, Louisiana; Edenton, North Carolina; and the
Marion State Fish Hatchery, Marion, Alabama.
 Airlift pumps for water circulation and aeration 107

Air Flow (cubic feet per minute)

1 5 10 50

80

1000

40
A ¢

' E
¢ 800
oo ;

• C
600 60 o

0
- 20
400 o

~ ~0

200
o

o j
25 lO0 250 100o

Air Flow (liters rain " 1 )

Fig. 9. Water flow rates in an airlift pump 30 cm in diameter when air was injected into
the vertical riser at different distances (cm) below the surface of the water: 120 (A), 105
(B), 90 (C), 75 (D), 60 (E), and 45 (F). All measurements were made with the flow-line of
the discharge 1.25 cm above the surface of the water.

100"
,~ C m
90"

~'-' 80.

[,~ ~ ?0-

~: "" 5o,
i
40t

30
-'a ~ ~ i ~ ~ io ,'2 1,
VERTICAL LIFT (era)
Fig. 10. Water flow rates in an airlift pump 5-0 cm in diameter, into which 85 liters
min- t of air were injected 60 cm below the center-line of the discharge, while the center-
line of the airlift discharge was below, at, or above the surface of the water.
108 N.C. Parker, M. A. Suttle

2oo T

120~-

40 i i 12 I ~ p
-~ ~ ~ 1'o 1'4
VERTICAL LIFT ( c m )

Fig. I I. Water flow rates in an airliftp u m p 7.5 c m in diameter, into which 85 liters
rain- ~ of air were injected 60 c m below the ccnter-line of the discharge, while the centcr-
line of the airliftdischarge was below, at, or above the surface of the water.

2eo T
T i0 cm
220+

I~ 140.

~ tOO-

eo+

201 J
-6 -2 2 6 10 14

VERTICAL LIFT ( c m )

Fig. 12. Water flow rates in an airlift pump 10 cm in diameter, into which 85 liters
min- ~of air were injected 60 cm below the center-line of the discharge, while the center-
line of the airlift discharge was below, at, or above the surface of the water.

O n the basis of the cost of materials, installation, and operation, we

found 7.5 and 10 c m diameter airlift p u m p s to be m o r e appropriate than
either larger or smaller ones for destratifying 0.02- to 2.0-ha ponds. O n e
airlift p u m p per 0"02- or 0.04-ha p o n d prevented stratification when 85
liters m i n - t of air was injected into the vertical riser at a point 60 cm
below the surface of the water. In larger p o n d s 7.5 and 10 cm diameter
p u m p s have been installed at the rate of 25 per ha at this laboratory and
in other US Fish and Wildlife Service facilities.
U n d e r these conditions, a regenerative blower (see Parker, 1983, for
description of types of blowers and compressors) with a nameplate
rating of 1 kW (1.34 hp) operated 27 airlift p u m p s with water flow rates
 Airlift pumps for water circulation and aeration 109

of 150-230 liters min -1 each; the total combined flow rates from 27
airlifts ranged from 4050 to 6210 liters min -1 for 7.5 and 10 cm
diameter pumps, respectively. Vertical lift was essentially zero as these
airlifts were adjusted for maximum flow to produce circulation and
vertical mixing of water in ponds.

ACKNOWLEDGEMENTS

We thank C. Suppes, and I. Quinata, Missouri Department of Conserva-

tion, and G. Chamberlain, Texas A & M University, for their critical
review of the manuscript.
Reference to trade names in this paper does not constitute endorsement
by the US Government.

REFERENCES

Andeen, G. B. (1974). Bubble pumps. Compressed Air Mag. (January), 16-18.

Apud, E D. & Camacho, A. S. (1980). Effects of water movement and aeration
system on the survival and growth of hatchery bred sugpo (Penaeus monodon
Fabricius) in earthen nursery ponds. Aquaculture Department, Southeast
Asian Fisheries Development Center, Tigbauan, Iloilo, Philippines. Q. Res.
Report, 4 (3), 15-18.
Castro, W. E. & Zielinski, P. B. (1980). Pumping characteristics of small airlift
pumps. Proc. World Mariculture Soc., 11,163-74.
Castro, W. E., Zielinski, P. B. & Sandifer, P. A. (1975). Performance characteris-
tics of airlift pumps of short length and small diameter. Proc. World Mari-
culture Soc., 6, 451-61.
Grant, D. M. (1979). Open Channel Flow Measurement Handbook, Instru-
mentation Specialties Company, Lincoln, Nebraska.
lto, T., Toi, J. & Satomi, Y. (1974). The effect of artificial pond water circulation
and feeding on productivity in carp culture ponds. Bull. Freshwater Fisheries
Res. Lab. (Tokyo), 24(2), 65-77.
Ivens, E. M. (1914). Pumping by Compressed Air, John Wiley & Sons, New
York.
Morrissy, N. M. (1979). Experimental pond production of mairon, Cherox
tenuimanus (Smith ) (Decapoda: Parastacidae). Aquaculture, 16, 319-44.
Murray, K. R., Poxton, M. G., Linfoot, B. T. & Watret, D. W. ( 1981 ). The design
and performance of low pressure air lift pumps in a closed marine recircula-
tion system. Proc. World Symp. on Aquaculture in Heated Effluents and
Recirculation Systems, Stavanger, 28-30 May, 1980. Vol. I, Berlin, W. Ger-
many, pp. 413-28.
Nicklin, D. J. (1963). The airlift pump: Theory and optimisation. Trans. Instr.
Chem. Engrs, 41, 29-39.
110 N. C. Parker, M. A. Suttle

Parker, N. C. (1979). Striped bass culture in continuously aerated ponds. Proc.

Southeastern Assoc. Fish WildlifeAgencies, 33,353-60.
Parker, N. C. (1983). Airlift pumps and other aeration techniques. In: Water
quality in channel caO~sh ponds, ed. C. S. Tucker, Southern Cooperative
Series Bulletin 290. Mississippi Agriculture and Forestry Experiment Station,
Mississippi State University, Mississippi, pp. 24-7.
Parker, N. C., Suttle, M. A. & Fitzmayer, K. (1984). Total gas pressure and
oxygen and nitrogen saturation in warmwater ponds aerated with airlift
pumps. Aquacultural Engineering, 3, 91-102.
Pickert, I. E (1932). The theory of the airlift pump. Engineering, 134, 19-20.
Rappaport, A., Sarig, S. & Marek, M. (1976). Results of tests of various aeration
systems on the oxygen regime in the Genosar experimental ponds and growth
of fish there in 1975. Bamidgeh, 28(3 ), 35-49.
Sarig, S. & Marek, M. (1974). Results of intensive and semi-intensive fish breed-
ing techniques in Israel in 1971 - 1973. Bamidgeh, 26(2), 28-48.
Spotte, S. H. (1970). Fish and Invertebrate Culture: Water Management in Closed
Systems, Wiley-Interscience, New York.
Spotte, S. (1979). Fish and Invertebrate Culture: Water Management in Closed
Systems(2nd Edn), Wiley-Interscience, New York.
Todoroki, F., Sato, Y. & Honda, T. (1973). Performance of air-lift pump. Bulletin
of the JSMI, 16(94), 733-41.
Tokar, J. M., Harvey, G. R. & Chesal, L. A. (1981). A gas lift system for large
volume water sampling, Deep-Sea Res., 28A(11 ), 1395-9.
Ward, C. N. (1924). An experimental study of airlift pumps. Water Works,
1275-8.
Wheaton, E W. (1977). Aquacultural Engineering, John Wiley & Sons, New
York.