ISSN 0015-4628, Fluid Dynamics, 2021, Vol. 56, No. 5, pp. 718–731. © Pleiades Publishing, Ltd., 2021.

Liquid Draining through Multiple Ports: An Investigation on Air

Core Vortex Formation
M. Prabhua,*, R. A. Kumara,**, R. Warriera,***, B. S. N. V. Moulia,****, and S. Kirana,*****
a
Department of Mechanical Engineering, Amrita Vishwa Vidyapeetham, Amritapuri, India
*e-mail: prabhumahadev1729@gmail.com
**e-mail: amritanjali.ajith@gmail.com
***e-mail: rajwar31497@gmail.com
****e-mail: vinaymouli1997@gmail.com
*****e-mail: kiransanthosh11@gmail.com
Received July 11, 2020; revised October 6, 2020; accepted December 25, 2020

Abstract—The phenomenon of air core vortexing in liquid draining through cylindrical tanks can have
serious adverse effects on many engineering systems. Hence, suppression of such vortices is necessity
in these applications. In the present work, the influence of initial rotation on vortexing phenomenon
with different drain port configurations with constant port diameter (d = 10 mm) are investigated using
multiple ports. In all the configurations considered, there is a concentric port (circular) in common
surrounded by other ports (circular) placed equidistant from the center. In this paper, vortexing phe-
nomenon is studied by varying the number of surrounding ports and their eccentricity. Draining is
allowed through one or more surrounding ports simultaneously. Before draining, circulation is pro-
vided to the liquid column by means of rotation (40–240 rpm) using a motorized stirrer where rota-
tional speed can be controlled. Results of this study reveal that for all the drain port configurations, as
eccentricity increases, the vortex gets suppressed and hence the draining time decreases. The result
also shows that the central port is the key contributor to the vortex formation and for lower values of
eccentricity, as the number of surrounding ports increases, vortex is progressively suppressed. Litera-
ture reveals that all vortexing studies reported so far were either with single or two drain ports. On mul-
tiple drain ports (number of drain ports more than 2), no investigations have ever been reported.
Hence, the current study is the first of its kind being presented in this paper.

Key words: vortexing, Multiple ports, critical height, head loss

DOI: 10.1134/S0015462821050116

Draining of rotated liquid column from a tank through the drain port located at the tank base can result
in the formation of surface dip or “kink” on the free surface. Surface dip thus formed will soon penetrate
the liquid column vertically down and transform in to a tapering pillar like structure reaching the drain
port, called as “air core vortex” as shown in Fig. 1. The height of the liquid column at which the vortex air
core just touches the drain port is called as critical height or critical submergence (Hc). The formation of
such an air core vortex in liquid draining through cylindrical tank has wide and prime applications to the
areas of aerospace, metal casting and hydraulic engineering. The air core formed in liquid propellant tanks
of rockets blocks the drain port and reduces the designed fuel flow rate which can adversely affect the per-
formance of rocket engines. This air core is also capable of triggering cavitation in feed pumps that are used
to supply fuel to the rocket engine. Such occurrences of cavitation could be expected in hydraulic appli-
cations as well. Formation of air core during draining of molten metal can invite casting defects in casted
metal. Due to such objectionable consequences, suppression of vortex air core has received great research
attention in the aforementioned engineering fields.
The phenomenon of vortex air core formation during liquid draining had been studied and reported in
the literature. In [1], an experimental study was performed on fluid draining from a cylindrical tank where
it was found that critical height can be empirically related to fluid density ratio and fluid drain rate for
draining of two or more immiscible liquids. In [2], the use of shaped discharge ports was suggested to sup-
press vortex phenomenon. It was found that stepped drain ports are very effective in suppressing the vortex
formation compared to bell mouth and conventional circular ports. A numerical investigation was con-
ducted in [3] on withdrawal of layered fluids from an open tank through a hole centred on the bottom

718
 LIQUID DRAINING THROUGH MULTIPLE PORTS 719

Hc

Fig. 1. Photograph of a typical vortex air core.

without swirl motion and the results are in agreement with the experimental results [1]. In [4], it was
attempted to derive an expression for velocity of air core vortices at hydraulic intakes. Equations for the
tangential, radial, and axial velocities at hydraulic intakes were derived in the work. The effect of base
inclination of tank on the critical submergence during vortex formation was studied in [5] and it was
revealed that vortex gets completely suppressed when the base inclination is above a limiting value which
depends on the size of the drain port. Rankine vortex formation during draining of liquids from tanks of
square and rectangular cross sections was analysed in [6]. It was found that vortexing can be suppressed by
making use of square and rectangular tanks instead of conventional circular tanks. It was also postulated
that the presence of sharp corners on rectangle and square cross sections results in vortex suppression.
In [7] the use of eccentric drain port was suggested to prevent the vortexing phenomenon when a liquid is
drained from cylindrical tanks. For complete suppression, it was found that ratio of port eccentricity to
tank radius should be at least 0.4. An experimental study on vortex air core formation was conducted in
[8]. It was found that when ratio of port diameter to tank diameter is less than 0.3, vortexing can be fully
suppressed for concentric ports. In [9], it was suggested to use the eccentric drain port to prevent vortexing
phenomenon when a liquid is drained from cylindrical tanks. It was found that the ratio of the port to tank
diameter required to suppress the vortex formation is a strong non-linear function of port eccentricity.
Influence of polygonal ports on air core vortex formation was investigated in [10] and it was found that
octagonal geometry is the best in effectively suppressing the vortex compared to other polygonal ports.
Recently, a study [11] on the vortexing phenomenon occurring when a liquid drains through two ports
located equidistant from the tank centre was conducted. It was found that vortexing can be efficiently con-
trolled and fully suppressed when the diameters of both drain ports are kept the same.
All the previous vortexing studies, except for the study [11], employed only a single port to drain the
liquid. Experimental study in [11] revealed that two drain ports located at equal eccentricities from the
tank centre could effectively control vortex formation. It was concluded that better vortex suppression
could be achieved by employing two drain ports compared to an equivalent single port under identical
conditions. The aforementioned study [11] left behind an interesting question as to how would the vortex
formation be affected if the draining is admitted through more than two ports. Aiming to address this
question in the present study, multiple drain ports (more than two) are introduced for the draining of a
rotated liquid column with the following objectives (a) devise alternative drain port configurations feasible
for effective vortex suppression and (b) assess whether the critical height can alone represent and quantify
the vortexing phenomenon. Accordingly, we have introduced additional drain ports surrounding a main
concentric drain port.

1. EXPERIMENTAL SETUP AND METHODOLOGY

The experimental setup and experimental methodology adopted are the same as those in [9]. The setup
consists of a cylindrical tank of internal diameter D = 96 mm and height of 410 mm as shown in Fig. 2a.
The diameter of both concentric as well as surrounding drain ports was chosen to be 10 mm and the ports
were drilled in an acrylic base plate as shown in the Fig. 3. Acrylic plate was attached to a cylindrical glass
tank using silica gel. Water was then filled in the cylindrical tank to an initial height (Hi) of 300 mm.

FLUID DYNAMICS Vol. 56 No. 5 2021

720 PRABHU et al.

(a) (b)

RPM indicator 1
RPM control knob
2

Stirrer
1. Rubber cork
Glass tank
2. Strands made
of metal

Handle
Acryllic
base with drain ports

Fig. 2. Experimental setup with port opening mechanism.

n=1 n=2
е = 17 mm

е = 29.5 mm
n=3 n=4 n=0 е = 42 mm
(a) (b)

Fig. 3. Multiple drain port configurations studied; red dots indicate drain ports.

A variable speed motorised stirrer was used to impart rotation to the liquid column. In the present study,
initial rotation (N) was varied from 40–240 rpm for all the cases considered. Rotation of two minutes was
imparted to the liquid column for providing uniform angular velocity to all the liquid particles. After
2 min, rotation was stopped and the stirrer was taken out cautiously from the liquid column. The drain
port stoppers from all the drain ports were simultaneously pulled vertically down under the tank to allow
draining. Figure 2b shows the mechanism used to open two multiple ports simultaneously at a particular
value of eccentricity (e/D = 0.438). Two rubber corks were attached to the strands of the metal frame as
shown in Fig. 2b and were inserted in to the drain ports before imparting rotation. For other eccentricity
values, separate but similar drain port openers were made. After two minutes of liquid rotation, a vertical
force was applied to the handle to open the drain port. For each eccentricity value, to simultaneously open
more number of ports (more than two), similar port openers but with more number of strands were made.
In the initial instances of draining, the dip formed on the surface of the liquid penetrated into the liquid
column to form a fully grown air core vortex and the free surface height of the liquid when air core just
touches the drain port, i.e., the critical height (Hc) was noted. The critical height was measured using a ver-
tical graduated scale attached to the outer surface of the cylinder. As mentioned earlier, Fig. 1 shows the
photograph of a typical vortex air core that formed during draining in a cylindrical tank where Hc marked
as the critical height. Studies by previous investigators [8, 9, 11] mentioned that strong vortexing is indi-
cated by the higher critical height. In previous studies, the critical height was considered as the prime
parameter to investigate vortexing phenomenon. Time of emptying the tank was also measured using a
stop watch.

FLUID DYNAMICS Vol. 56 No. 5 2021

 LIQUID DRAINING THROUGH MULTIPLE PORTS 721

Hc/Hi
0.6 n=1
0.5
0.4
0.3
0.2
e/D = 0.177
0.1 e/D = 0.307
e/D = 0.438
0
0.1
0 50 100 150 200 250
N, rpm

Fig. 4. Hc/Hi versus N at n = 1.

All experimental cases were repeated 10 times to ensure the repeatability of measurements. From such
experimental trails, the standard deviation values of critical height and time of emptying were estimated.
Correspondingly, the uncertainty in the measurements of these parameters, viz., critical height and time
of emptying, was estimated to be 11% and 7%, respectively. All the multiple port cases tested consist of a
concentric drain port surrounded by one or more other drain ports. The number of surrounding drain
ports n introduced around the concentric port varies from 0 to 4, and they are located at specific geomet-
rical positions on the base plate of a cylindrical tank (diameter D = 96 mm) as shown in Fig. 3. The
red spots in Fig. 3 show drain port positions. Figure 3b shows the eccentric positioning of the surrounding
drain ports. The position of the surrounding drain ports is fixed at eccentricity values, viz., e = 17, 29.5, and
42 mm and the corresponding non-dimensional eccentricity values are e/D = 0.177, 0.307,
0.438, respectively. The cases were grouped under five drain port configurations (n) namely, n = 0, 1, 2,
3, and 4 as shown in Fig. 3. The configuration n = 0 corresponds to the single concentric drain port case.
The experiment was carried out with water at the room temperature (28°C).

2. RESULTS AND DISCUSSION

The experimental study on vortexing behaviour during draining through multiple ports at equal eccen-
tricity (e/D = 0.177, 0.307, 0.438, respectively) and port diameter d = 10 mm was done on the configura-
tions shown in Fig. 3. Results are shown in Figs. 4–14. As mentioned earlier, the critical height has been
considered as a prime parameter by previous investigators to characterise the vortex formation.
Figures 4–7 show the variation of dimensionless critical height (Hc/Hi) with the initial fluid rotation
rate (N rpm). In general, it could be seen that, Hc/Hi increases with initial rotation whereas, it decreases
with non-dimensional eccentricity. The non-dimensional draining time (Td/To) of the liquid is plotted
against stirrer rotation in Figs. 8–11; Td is the time of emptying the tank with rotation and To is the emp-
tying time without rotation. It could be observed that, the higher the rate of rotation imparted, the higher
is the value of Hc which indicates longer time of emptying of the tank (larger Td/To value). At larger Td/To
values, the average rate of discharge flow reduces due to vortex formation. Furthermore, at a given value
of N, Hc/Hi reduces with e/D which is particularly evident for n = 4 at higher values of N (Fig. 7). This
effect of eccentricity qualitatively agrees with the results of previous investigators [7, 9]. Correspondingly,
Td/To also reduces as seen in Figs. 8–11. Effect of e/D is more distinctly seen in Fig. 7 for the configuration
with n = 4, where at e/D = 0.438. For this particular configuration, it is observed that the critical height is
zero for all values of initial rotation. Standard deviations of the measured quantities, viz., critical height
and time of draining are typically shown as error bands in Figs. 7 and 11, respectively.
Among all the configurations tested, n = 4 gave the highest percentage reduction (100%) in the non-
dimensional critical height (Fig. 7) followed by n = 3 (20%, see Fig. 6), n = 2 (16.13%, see Fig. 5) and n = 1
(7.6%, see Fig. 4) when e/D is varied from 0.177 to 0.438 at 240 RPM. From these results, it could be
inferred that the lower values of critical height is achieved not only by increasing the eccentricity but also
by increasing the number of surrounding drain ports and this is evident from Fig. 12. Figure 12 depicts the

FLUID DYNAMICS Vol. 56 No. 5 2021

722 PRABHU et al.

Hc/Hi
0.6
n=2
0.5
0.4
0.3
0.2
e/D = 0.177
0.1 e/D = 0.307
0 e/D = 0.438

0.1
0 50 100 150 200 250
N, rpm

Fig. 5. Hc/Hi versus N at n = 2.

Hc/Hi
0.6
n=3
0.5
0.4
0.3
0.2
e/D = 0.177
0.1 e/D = 0.307
0 e/D = 0.438

0.1
0 50 100 150 200 250
N, rpm

Fig. 6. Hc/Hi versus N at n = 3.

Hc/Hi n=4
0.5

0.4 e/D = 0.177

e/D = 0.307
e/D = 0.438
0.3

0.2

0.1

0.1
0 50 100 150 200 250
N, rpm

Fig. 7. Hc/Hi versus N at n = 4.

FLUID DYNAMICS Vol. 56 No. 5 2021

 LIQUID DRAINING THROUGH MULTIPLE PORTS 723

Td/To
1.5 n=1

1.4

1.3

1.2

1.1 e/D = 0.177

e/D = 0.307
1.0 e/D = 0.438

0.9
0 50 100 150 200 250
N, rpm

Fig. 8. Td/To versus N at n = 1.

Td/To
1.5 n=2

1.4 e/D = 0.177

e/D = 0.307
1.3 e/D = 0.438

1.2

1.1

1.0

0.9
0 50 100 150 200 250
N, rpm

Fig. 9. Td/To versus N at n = 2.

Td/To
1.5 n=3
1.4 e/D = 0.177
e/D = 0.307
1.3 e/D = 0.438

1.2

1.1

1.0

0.9
0 50 100 150 200 250
N, rpm

Fig. 10. Td/To versus N at n = 3.

FLUID DYNAMICS Vol. 56 No. 5 2021

724 PRABHU et al.

Td/To n=4
1.35
1.30 e/D = 0.177
e/D = 0.307
1.25 e/D = 0.438
1.20
1.15
1.10
1.05
1.00
0.95
0 50 100 150 200 250
N, rpm

Fig. 11. Td/To versus N at n = 4.

Hc/Hi
e/D = 0.177
0.8 e/D = 0.307
0.7 e/D = 0.438
0.6
0.5
0.4
0.3
0.2
0.1
0
0.1
0 2 4 n

Fig. 12. Variation in Hc/Hi versus number of drain ports (n).

variation in Hc/Hi with n for various e/D values at Hi = 300 mm and N = 240 RPM. At e/D = 0.177, when
the number of drain ports is increased from n = 0 to n = 4, approximately 30% reduction in the critical
height could be obtained. When the eccentricity increases to 0.307 and 0.438, percentage reduction in the
critical height increases to 61% and 100%, respectively. Figures 13a–13b show the photo shots of the air
core vortex just at the instance when it attains the critical height at n = 1 and n = 4, respectively at N = 200
RPM and e/D = 0.177. As evident in Fig. 13, the critical height decreases with increase in the number of
surrounding ports.
Again, from Fig. 12, at n = 0 (there is only concentric port), it could be seen that, Hc/Hi is the highest
and up on increasing the value of n, critical height decreases for all e/D values. Furthermore, on extrapo-
lation, the trends of critical height depicted in Fig. 12 indicate that there is a minimum number of sur-
rounding ports (n) required to reduce the value of dimensionless critical height for a given e/D ratio. To
make this point clear, at e/D = 0.177, for n > 2, Hc/Hi ratio remains more or less constant (~0.55) indicat-
ing that critical height cannot be brought to zero by a finite increase in the number of surrounding ports.
In other words, in this case, for a possible reduction in critical height, n tends to infinity which could also
means that the resulting drain port geometry asymptotically approaches an “annular ring.” Whereas, at
e/D = 0.307, graphical trend indicates that, critical height could be brought to zero possibly at n = 5. At
e/D = 0.438, as Fig. 12 clearly shows, for n = 4, there is no occurrence of air core vortex formation. The
limiting values of 1/n to achieve zero critical height for each e/D ratio is shown in Fig. 14. As this figure
shows, limiting 1/n increases as e/D increases.

FLUID DYNAMICS Vol. 56 No. 5 2021

 LIQUID DRAINING THROUGH MULTIPLE PORTS 725

(a) (b)

D = 96 mm

Hс

n=1 n=4

Fig. 13. Critical height formation for the n = 1 and n = 4 configurations.

Limiting (1/n)
0.3

0.2

0.1

0
0.2 0.3 0.4 0.5
e/D

Fig. 14. Variation in limiting (1/n) versus e/D.

For discussing the flow physics of this problem, analysis of hydraulic head loss incurred during liquid
draining is carried out for all values of initial rotation, eccentricity and number of ports. The exit velocity
of the liquid (efflux velocity) at the drain port just at the beginning of draining can be expressed by
Eqs. (2.1) and (2.2):

Vo = 2 g ( H i − h lo ) = εo 2 gH i , (2.1)

Vd = 2 g ( H i − h ld ) = εd 2 gH i , (2.2)
where hl is the head loss and ε, the loss coefficient defined here with subscripts o and d indicating no-
rotation and rotation cases of liquid, respectively. The head loss occurring during the draining process var-
ies with the liquid height and this includes (a) the head loss at entry and exit of the drain port (b) the head
loss due to friction in the drain port (c) other hydraulic losses possibly occurring in swirling flow (fluid
rotation). The coefficients εo and εd takes care of the head losses (a), (b), and (c) mentioned above. To
specifically mention, just in the start of draining, loss is limited to (a) and (b) only. As time progresses,

FLUID DYNAMICS Vol. 56 No. 5 2021

726 PRABHU et al.

hld/hlo
2.2

2.0

1.8 n=1
n=2
1.6 n=3
n=4
1.4

1.2

1.0
0 50 100 150 200 250
N

Fig. 15. Variation in hld/hlo with N for various values of n.

other hydraulic losses come in to picture which possibly varies during draining. Hence, coefficients εo and
εd vary with respect to time (during draining). Therefore, the values assigned to these loss coefficients are
the time averaged values in Eqs. (2.3) and (2.4):
 2H i   D 2 
To = 1   , (2.3)
εo  g   (n + 1)d 2 
 2H i   D 2 
Td = 1   . (2.4)
εd  g   (n + 1)d 2 
To find the loss coefficients, these sets of equations were used. The equations used here can be derived
from the continuity and Bernoulli equations, which are the same as those used in [11].
The total time required to drain the liquid fully from the tank, To (no-rotation case) and Td (rotation
cases) being measured from the experiments, the associated loss coefficients (εo and εd ) can be found
from Eqs. (2.3) and (2.4). As previously mentioned, the efflux velocity of liquid from the drain port under
no-rotation and rotational cases is given by Eqs. (2.1) and (2.2), respectively. From the value of loss coef-
ficients, head losses (hlo and hld) can be found from Eqs. (2.1) and (2.2).
The following equation
hld = hlo + hls (2.5)
gives how hlo and hld are related to each other and subsequently, head loss due to swirl (hls) can be found
from Eq. 2.5. All losses are non-dimensionalized with respect to hlo. Figures 15 and 16 are the plots
between N and dimensionless parameters, viz., hld and hls respectively and Fig. 17 depicts the relationship
between N and non-dimensional efflux velocity (Vd/Vo) for various values of n. All these plots corresponds
to e/D = 0.177. For all other values of eccentricities, similar trends are not observed. For the higher values
of eccentricity, time of emptying increases with the number of surrounding ports though the critical height
reduces. This is because, at the higher values of e/D (e/D = 0.307 and 0.438) hydrodynamic losses are
higher when the number of surrounding ports increases.
From Fig. 15 it is seen that hld/hlo increases with N and decreases with the number of ports; at N =
240 rpm, for a single surrounding port (n = 1), hld/hlo ~ 2.0 which indicates that the swirl loss (hls ) is almost
equal to hlo. Following Eq. (2.5), this indicates that, swirl loss increases with N and decreases with the
number of ports as depicted in Fig. 16. To highlight further, once fluid is rotated, net loss (hld) is highly
contributed by swirl loss. Figure 17 shows the variation of non-dimensional efflux velocity (Vd/Vo) with N
which reveals that, reduction in losses will enhance the efflux velocity. For a specific value of n, efflux
velocity is highest for lowest value of initial rotation. As the value of initial rotation increases, the strength
of vortexing increases which leads to blockage of drain port. In the light of the current results, it could be
conjectured that swirl losses which manifest only due to initial rotation, is a measure of blockage of drain
port by the vortex air core.

FLUID DYNAMICS Vol. 56 No. 5 2021

 LIQUID DRAINING THROUGH MULTIPLE PORTS 727

hls/hlo
1.2

1.0

0.8 n=1
n=2
0.6 n=3
n=4
0.4

0.2

0 50 100 150 200 250

N

Fig. 16. Variation in hls/hlo with N for various values of n.

Vd/Vo

1.0 n=1
n=2
n=3
0.9 n=4

0.8

0.7

0 50 100 150 200 250

N

Fig. 17. Variation in Vd/Vo with N for various values of n.

Figure 18 is the plot between dimensionless time of emptying and surrounding number of ports for var-
ious e/D ratios typically at N = 240 RPM. Except for e/D = 0.177, for all other e/D values, time of emtying
enhances after n = 2. As mentioned earlier, when the number of surrounding ports increases, vortexing is
suppressed for e/D = 0.177 as indicated by a decrease in the critical height (see Fig. 12) and a decrease in
time of emtying (see Fig. 18). The reason for the suppression can be explained as follows. It should be also
noted that, for this e/D ratio (= 0.177), the efflux velocity increases with the number of surrounding ports
(see Fig. 17). A numerical study conducted in [12] suggested that there are two kinds of fluid motion in
connection with formation and suppression of air core. The first kind of motion is related to the applied
initial rotation which tries to move the liquid away from the cylinder center because of centrifugal force.
This motion will try to enhance the formation of vortex. The second kind of motion is related to the
motion of fluid towards the center of the cylinder which tries to suppress the formation of vortex. The
efflux velocity as given in Eq. (2.2) directly depends on the radial component of velocity by two dimen-
sional continuity equation given in Eq. (2.6):
∂ρ 1 ∂ ∂ (u )
+ ( rur ) + z = 0. (2.6)
∂t r ∂r ∂z
Here, ρ is the dencity of fluid, t represents the time scale of flow, u represents the velocity vector and,
the subscripts r and z represents radial and axial directions, respectively.
From Eq. (2.6), it could be inferred that, as the axial velocity component (efflux velocity) increases,
the radial component of velocity towards the center also increases. Enhancement of the radial velocity

FLUID DYNAMICS Vol. 56 No. 5 2021

728 PRABHU et al.

Td/To

1.5 e/D = 0.177

e/D = 0.307
e/D = 0.438
1.4

1.3

1.2

1.1
0 1 2 3 4 5
n

Fig. 18. Variation in Td/To versus the number of drain ports (n) for N = 240 RPM.

hld/hlo
2.2

2.0

1.8

1.6

1.4 e/D = 0.177

e/D = 0.307
1.2 e/D = 0.438
1.0
0 50 100 150 200 250
N

Fig. 19. Variation in hld/hlo with N for various values of e/D.

component towards the center of the cylinder results in the suppression of air core as reflected in Figs. 12
and 18. For other values of e/D, i.e., e/D = 0.307 and 0.438, time of emptying enhances for values of
n > 2 as seen in Fig. 18 typically for N = 240 RPM, which means that average discharge is reducted as vor-
texing is not suppressed. Longer time of emptying indicates a stronger vortex formation which should nor-
mally correspond to the higher critical height. Contrary to this expectation, critical height reduces (indi-
cating vortex suppression) for these cases of higher eccentricity at n > 2 (Fig. 12). That is, the trends of
critical height and time of emptying indicate opposing features of vortexing in the said cases (n > 2).
Figures 19, 20, and 21 show the variation of hld, hls and Vd/Vo with respect to N for various values of e/D.
This is typically plotted for n = 1. For other values of n, similar trend is observed and hence, not plotted
here.
In [13] it was found that, with a single drain port, if the eccentricity increases, the critical height
decreases and becomes minimum at the maximum value of port eccentricity. A reason for this observation
was given as follows. Just at the commencement of draining, the axis of fluid rotation coincides with that
of the cylinder, but later, the fuid rotation axis migrates until it coincides with the port axis. At this
moment, the vorticity is strongly amplified by the gradient of the axial flow and the free fluid surface is
drawn down. Current study also notice a reduction in critical height with port eccentricity. Figures 22a–22c
show the photographs of temporal variation of vortex air core formation corresponding to the drain time
3, 5, and 9 s respectively during draining through multiple drain ports with n = 2 at N = 220 RPM. Tea
powder was introduced in to the liquid column at the start of draining to track the motion of fluid particles
and this could be clearly seen in Fig. 22a. The vortex air core seen in Fig. 22 shows that the concentration

FLUID DYNAMICS Vol. 56 No. 5 2021

 LIQUID DRAINING THROUGH MULTIPLE PORTS 729

hls/hlo
1.2

1.0

0.8

0.6

0.4 e/D = 0.177

e/D = 0.307
0.2 e/D = 0.438
0
0 50 100 150 200 250
N

Fig. 20. Variation of hls/hlo with N for various values of e/D.

Vd/Vo

1.0
e/D = 0.177
e/D = 0.307
0.9 e/D = 0.438

0.8

0.7

0 50 100 150 200 250

N

Fig. 21. Variation of Vd/Vo with N for various values of e/D.

of vorticity takes place only along the central axis and it never shifts to either side (towards the eccentric
ports) during the entire draining process. This aspect was found to be true for all cases of multiple ports
studied. Without a shift in the air core axis, the critical height and time of emptying reduce with eccen-
tricity as shown by the present results. In other words, current results clearly demonstrate that the reason
for vortex suppression with eccentric ports is not due to vortex axis shift but it is because of a more funda-
mental reason, viz., lower hydraulic losses associated with eccentric ports during liquid draining. At given
value of initial rotation N, Figs. 19 and 20 show that non-dimensional hld and hls decreases with increase
in eccentricity which results in a hike of non-dimensional efflux velocity (Fig. 21). Following the argu-
ments made earlier based on Eq. (2.6), it can be inferred that, enhancement in efflux velocity resulted due
to increase in eccentricity have initiated second kind of fluid motion, viz., increase of radial velocity
towards the center which resulted in vortex suppression.

2.1. Can Critical Height Correctly Characterize the Vortexing Phenomenon?

As already mentioned, authors and previous investigators have made use of critical height to charac-
terize vortexing phenomenon as the prime parameter. This is because, it was found that, for liquid drain-
ing using single ports, with higher initial fluid rotation (circulation) provided, critical height increases and
subsequently, the time required to drain out the fluid from the tank (time of emptying) also increases since
the formed vortex air core blocks the drain outlet area. To highlight further, the higher critical height
(intense vortexing) always results in the longer time of emptying in cases where the liquid is drained

FLUID DYNAMICS Vol. 56 No. 5 2021

730 PRABHU et al.

(a) (b) (c)

D = 96 mm

Hc

Fig. 22. Temporal variation of air core vortex at n = 2 and N = 220 RPM.

through a single port, as reported in earlier investigations. The current results presented in Figs. 4–11 show
that both the critical height and the time of emptying increase with N for multi port configurations and
they do not contradict this argument. Furthermore, Fig. 12 reveals that critical height reduces with
increase in the number of drain ports (n) for all eccentricities. Correspondingly, like in the case of single
drain ports, liquid emptying time could also be expected to decrease. But, quite contrary to this expecta-
tion, for higher eccentricities, for n > 2, time of emptying increases as seen in Fig. 18. It can now be con-
cluded in general that, for all higher values of eccentricities (e/D > 0.177), the number of surrounding
drain ports required to suppress the vortexing should be limitted to 2.
In practical applications such as aerospace, metal casting and hydraulic engineering, adequate flow
rate condition is required to be maintained during liquid draining from containers/tanks. In other words,
liquid should be drained out without undue delays. But, air core formation blocks the flow outlet area sub-
sequently delaying the liquid discharge process and thereby increasing the time of emptying. To cite a spe-
cific example, time delay in discharging liquid propellant used in rocket engines adversely affects the
engine performance [14]. Therefore, it is not merely the critical height but the liquid emptying time also
found to be a very significant parameter which governs vortexing phenomenon. In the light of the current
study, it is proposed that while assessing the vortexing phenomenon, both these parameters, viz., critical
height and more importantly the liquid emptying time have to be considered. Critical height alone is not
sufficient to characterize vortex formation.

SUMMARY
The experimental study on vortexing during draining from a cylindrical tank through multiple ports
was done for five configurations, i.e., n = 0, 1, 2, 3, and 4 for Hi = 300 mm for a wide range of initial liquid
rotation speed ranging from 40 RPM to 240 RPM and the following conclusions are drawn.
For all values of n, vortexing phenomenon intensifies (critical height and time of emptying increase)
with increase in initial rotation and decreases with increasing eccentricity. That is, air core could be sup-
pressed by increasing the eccentricity of ports. Both the critical height as well as the time of draining exhib-
its a non-linear variation with respect to the liquid rotation speed in all the cases studied. As the number
of ports surrounding the central port increases, vortex is more and more suppressed for lower values of
eccentricity (e/D = 0.177). It was also found that, at higher values of eccentricity, time of emptying gets
enhanced with increase in surrounding ports beyond n = 2. Therefore, for all higher values of eccentricity
(cases with e/D greater than 0.177) number of surrounding ports should be limited to n = 2. This study also
reveals that, along with critical height, time of emptying is also a prime parameter to describe vortexing
phenomenon particularly from practical point of view. It is found that hydraulic losses associated with

FLUID DYNAMICS Vol. 56 No. 5 2021

 LIQUID DRAINING THROUGH MULTIPLE PORTS 731

multiple ports are less compared to single concentric ports and this resulted in vortex suppression. Fur-
thermore, shift in vortex axis is not the only possible reason for vortex suppression at higher eccentricities
but vortex mitigation might also occur due to decreased swirl loss.

REFERENCES
1. Lubin, B.T. and Springer, G.S., The formation of a dip on the surface of a liquid draining from a tank, J. Fluid
Mech,1967, vol. 29, no. 2, pp. 385–390.
2. Ramamurthi, K. and Tharakan, T.J., Shaped discharge ports for draining liquids, J. Spacecr. Rockets, 1992,
vol. 30, no. 6, pp. 786–788.
3. Zhoq, Q.N. and Graebel, W.P., Axisymmetric draining of cylindrical tank with a free surface, J. Fluid Mech,
1990, vol. 221, no. 6, pp. 511–532.
4. Hite, J.E. and Mih, W.C., Velocity of air-core vortices at hydraulic intakes, J. Hydraul. Div., Am. Soc. Civ. Eng,
1994, vol. 120, no. 3, pp. 284–297.
5. Gowda, B.H.L., Akhuli, S., Anudeep, B.R., Ipe, K.R., and Kishore, K., Influence of base inclination on vortex
formation during draining from cylindrical tanks, Indian J. Eng. Mater. Sci, 2013, vol. 20, no. 5, pp. 361–366.
6. Gowda, B.H.L., Draining of liquids from tanks of square or rectangular cross sections, J. Spacecr. Rockets, 1996,
vol. 33, no. 2, pp. 311–312.
7. Sohn, C.H., Gowda, B.H.L., and Ju, M.G., Eccentric drain port to prevent vortexing during draining from cy-
lindrical tanks, J. Spacecr. Rockets, 2008, vol. 45, no. 3, pp. 638–340.
8. Kumar, R.A., Josy, J., Shaji, R.K., and Srikrishnan, A.R., Vortex suppression through drain port sizing, Journal
of Aerospace Engineering, 2016, vol. 29, no.4, p. 06016002.
9. Kumar, R.A., Nair, R.R., Prabhu, M., and Srikrishnan, A.R., Vortex formation during draining from cylindri-
cal tanks: Effect of drain Port eccentricity, Journal of Aerospace Engineering, 2017, vol. 30, no. 5, p. 06017001.
10. Prabhu, M., Kumar, R.A., Nair, R.R., and Anith, T.A., Liquid draining through polygonal ports: An investiga-
tion on gas entraining vortices, J. Spacecr. Rockets. 2019, vol. 56, no. 5, pp. 1651–1657.
11. Prabhu, M., Kumar, R.A., Gopikrishnan, T.H., Deshpande, P.J., Anandhakrishnan, U., Kiran, A.S., and Gov-
indu, R.P., Rankine Vortex formation during draining: A new twin port suppression strategy, J. Appl. Fluid
Mech, 2020, vol. 13, no. 1, pp. 147–160.
12. Nazir, K. and Sohn, C.H., Study of air core phenomenon and influence of water height during liquid draining,
Journal of Mechanical Science and Technology, 2017, vol. 31, no. 8, pp. 3831–3837.
13. Piva, M., Iglesias, M., Bissio, P., and Calvo, A., Experiments on vortex funnel formation during drainage, Phys.
A (Amsterdam, Neth.), 2003, vol. 329, no. 1, pp. 1–6.
14. Son, J.H. and Park, S., Prevention of air entrainment during liquid draining using disc-type vortex suppressor,
Journal of Mechanical science and Technology, 2018, vol. 32, no. 10, pp. 4675–4682.

FLUID DYNAMICS Vol. 56 No. 5 2021