Technical Note

Vortex Suppression through Drain Port Sizing

R. Ajith Kumar, Ph.D. 1; Josy Joykutty 2; Rahul Korah Shaji 3; and A. R. Srikrishnan, Ph.D. 4

Abstract: This paper reports on an experimental study motivated by the issue of vortex formation in fuel tanks of liquid propulsion
rockets. In this study, vortex funnel (vortex-air core) formation during draining of liquids from cylindrical tanks is suppressed by means
Downloaded from ascelibrary.org by University of California, San Diego on 03/08/16. Copyright ASCE. For personal use only; all rights reserved.

of a simple, yet effective method of controlling or adjusting the size of the base drain port. This is particularly relevant in spacecraft and
rocket applications where suppression of such vortex-air core is very much warranted because of the possible drain port blockage they
cause and the consequent adverse impact (of vortices) on the generation of propulsive thrust. It is found that in the range of rotational
speeds provided to the liquid column (between 120 and 200 rpm) in a cylindrical tank, for d=D ∼ 0.03 (where D is the tank diameter held
constant in the study, and d is the port diameter that is varied in the experiments), vortex funnel formation is completely eliminated for
concentric drain ports. For eccentric ports, this occurs at a slightly higher value of port size, viz., d=D ∼ 0.04. The liquid used is water at
room temperature with its free surface open to the atmosphere. DOI: 10.1061/(ASCE)AS.1943-5525.0000609. © 2016 American Society
of Civil Engineers.

Introduction Ramamurthi and Tharakan (1993) used a stepped drain port to

arrest vortex formation. Gowda (1996) has shown that vortex
During the flight of space vehicles and rockets, a gas-core vortex formation can be avoided by using tanks of square and rectangular
develops due to drainage of liquid from the propellant tank to the cross-sections. Further, Gowda et al. (1996) used a dish-type sup-
engines (Basu et al. 2013). Further, the rotational flow of liquid pressor to prevent vortexing. Sohn et al. (2008) used eccentric drain
and consequent extension of the vortex core into the outlet reduce
ports to control vortex formation.
the liquid flow rate significantly by reducing the effective area of
In the present study, unlike many other previous studies,
the drain outlet, subsequently increasing the time of draining
suppression of vortex formation is achieved without using any add-
(Abramson et al. 1962; Ramamurthi and Tharakan 1993). Pump-
ons or external devices. Instead, it is made possible merely by
fed liquid rocket engines must therefore be cut off before the pres-
controlling the port size (d) for a given cylinder diameter (D).
surant gas enters the pump to prevent overrotation and possible
In other words, by varying the relative port size ratio (d=D), keep-
critical launch failure (Basu et al. 2013). In liquid propellant rocket
ing D constant, the present study proves that vortex formation
motors, vortex formation can adversely affect the performance and
could be completely eliminated.
also could lead to underutilization of the propellant (Basu et al.
2013). The present study is motivated by the impact of air core
vortex formation on the functioning of propellant feed systems
of liquid propellant rocket engines. An important parameter that
Experimental Arrangement
characterizes the core vortex formation is the critical height (Lubin The experimental arrangement is shown in Fig. 1(a). In Fig. 1(b),
and Springer 1967), which refers to the height of the liquid column e is the eccentric distance between the center of the cylinder axis
at an instant when the core air vortex just reaches the drain port. and the center of the drain port. A cylindrical glass tank is used
Lubin and Springer (1967) have studied the liquid-draining pro- with an internal diameter (D) of 100 mm and height (H) of
cess through a circular port from a cylindrical tank experimentally 400 mm with a base plate made of acrylic. Concentric and eccen-
and have derived an analytical expression for the critical height for tric drain holes of different diameters varying from 4 to 9 mm
various fluid combinations. were drilled to perform a wide range of experiments. The glass
Several methods of preventing vortex formation have been cylinder was glued to the acrylic plates by using silicone gel.
explored in earlier studies. Baffles were used by Abramson et al. A variable-speed motor-powered stirrer was used to impart the re-
(1962) to suppress sloshing, which also prevents vortex formation. quired rotations to the liquid (water) in the cylindrical tank. After
filling the tank with water with an initial height (Hi ), the stirrer
1
Professor, Dept. of Mechanical Engineering, AMRITA Univ.,
was inserted to a depth of three-fourths the water level. Rotation
Amritapuri, Kerala 690525, India. was provided for 2 min so that a uniform angular velocity is im-
2
Graduate Student, Dept. of Mechanical Engineering, AMRITA Univ., parted to the liquid column. After 2 min, the drain port stopper is
Amritapuri, Kerala 690525, India. removed and draining is allowed; before removing the stopper,
3
Assistant Professor, Dept. of Mechanical Engineering, ASIET, Kalady, the stirrer was lifted slowly and cautiously so that it would not
Kerala 688522, India. interfere with the vortex formation. As draining starts, a dip forms,
4
Associate Professor, Dept. of Aerospace Engineering, AMRITA Univ., which extends to the drain port almost instantaneously in the
Ettimadai Campus, Kerala 641112, India (corresponding author). E-mail: form of a vortex with an air core, and the liquid column height
ar_srikrishnan@cb.amrita.edu; ARSrikrishnan@gmail.com
at which this occurs is measured as the critical height (H c ) in all
Note. This manuscript was submitted on July 2, 2015; approved on
November 25, 2015; published online on February 26, 2016. Discussion the experiments. Initial rotation (N rpm) is provided in the range
period open until July 26, 2016; separate discussions must be submitted for of 60–200 rpm. Four drain port sizes are employed in the present
individual papers. This technical note is part of the Journal of Aerospace study, viz., d ¼ 4, 5, 7, and 9 mm (d=D ¼ 0.04, 0.05, 0.07 and
Engineering, © ASCE, ISSN 0893-1321. 0.09, respectively).

© ASCE 06016002-1 J. Aerosp. Eng.

J. Aerosp. Eng., 06016002

 Results

Results on vortex formation and suppression are presented in this

section as Hc =H i versus N plots for H ¼ 400 mm and also for dif-
ferent drain port sizes as shown in Figs. 2 and 3. As mentioned
earlier, the port size ratio d=D is varied by changing the port diam-
eter d, keeping D constant at 100 mm.

Influence of Initial Rotation on Vortex Formation

In Fig. 2, a common aspect noticed for both the concentric and
the eccentric configurations is that the nondimensional critical
Downloaded from ascelibrary.org by University of California, San Diego on 03/08/16. Copyright ASCE. For personal use only; all rights reserved.

height (H c =Hi ) of vortex formation is higher for higher values

of port diameter (d). For a given value of N, at larger port sizes
(d=D > 0.05), the tendency for vortex formation increases. Also,
the larger the rate of rotation imparted, the larger the value of criti-
cal height of vortex formation will be. The time required for empty-
ing the tank increases with an increase in the critical height of
vortex formation, i.e., the rate of discharge reduces as the vortex
formation occurs (Fig. 3). The experiments were repeated several
times (a minimum of five times) for each case to check this feature,
and the same results were obtained every time. Referring to Fig. 2,
at higher values of d (d=D ¼ 0.07 and 0.09), Hc =H i increases
with N for eccentric ports as well. However, in general the values
of critical height are lower than those for the concentric ports.
That is, in eccentric ports, vortex formation is delayed compared
to concentric port configurations. At lower d=D ratios (= 0.04,
0.05), the comparative trends are the same. When the port size
is 3 mm (the curve corresponding to d=D ¼ 0.03 in Fig. 2), for
concentric port configurations, vortex formation does not occur
over the complete range of stirrer rpm (N) tested, whereas for ec-
centric ports, vortex formation is completely arrested at d ¼ 4 mm
(d=D ¼ 0.04). Correspondingly, Hc =H i ¼ 0 for these cases (Fig. 2).
That is, in both cases vortex formation is eliminated by way of em-
Fig. 1. Experimental setup: (a) cylindrical tank; (b) eccentric drain
ploying a critical port size. A possible physical explanation for the
port
observed suppression of vortex growth by reduction in port size is

Fig. 2. Influence of initial rotation on vortex formation

© ASCE 06016002-2 J. Aerosp. Eng.

J. Aerosp. Eng., 06016002

 significant difference in the values of critical height because the
vortexing phenomenon is observed to be nonlinear characterized
by intermittency behavior (Fig. 4).
Similar results were obtained at different initial heights of 300
and 350 mm (not presented here).

Influence of Initial Rotation on Time of Emptying

In Fig. 3, the time of emptying with rotation ½T empðrÞ  is normalized
with the time of emptying without rotation ½T empðnrÞ  for each case.
It could be seen that the T empðrÞ =T empðnrÞ ratio increases as N in-
creases for a given port size. Compared to concentric drain port
Downloaded from ascelibrary.org by University of California, San Diego on 03/08/16. Copyright ASCE. For personal use only; all rights reserved.

configurations, the T empðrÞ =T empðnrÞ ratio is smaller in eccentric port

cases, complying with the fact that Hc =H i is also smaller in eccen-
tric cases (Fig. 2). In majority of the cases, T empðrÞ =T empðnrÞ follows
a slightly nonlinear trend with respect to N. From Fig. 3 it can be
noted that the time of emptying increases with port size for a given
rotational speed (N). This is because the larger port size entails an
earlier formation of the air core vortex (as shown by the increase in
critical height in Fig. 2), which in turn partially blocks the discharge
flow area.
Fig. 3. Influence of initial rotation on time of emptying with
eccentricity Vortex-Air Core Shape and Intermittency
The term vortex intermittency refers to the intermittent nature of
vortex formation during draining in the context of the present study.
It includes the instability in the vortex dip length, the phenomenon
given in a later section. In the eccentric drain port case, the critical of appearing and disappearing of a vortex and the phenomenon of
value of d=D is higher because eccentricity itself has a suppressing breaking of the vortex tip. In the present set of studies, the observed
effect on vortex formation (Sohn et al. 2008). Sohn et al. (2008) vortex intermittency is shown in Fig. 4 where the temporal evolu-
have explained the fluid dynamics of vortex suppression achieved tion of vortex-air core at different time instances is shown for a
by eccentricity of the drain port, based on the consideration of the typical case. As could be seen, the vortex dip length undergoes non-
relevant forces (their experiments pertain to a rotational speed of linear variations with respect to time. This intermittency phenome-
120 rpm). They conjecture that when the port is eccentrically non has been observed for all cases with d=D > 0.06 for all H i
placed, its location will be at a higher static pressure compared to values tested (= 300, 350, and 400 mm) and at all values of N em-
that of a concentric port, and this offers additional resistance to the ployed. Another interesting feature is the curvilinear shape of the
propagation of the core vortex downwards. Results of the present vortex-air core profile as against straight vortex core formed in the
study (with eccentricity) are in agreement with these observations. case of concentric ports (Fig. 4). In the images, the dotted white line
Quantitative differences between the results of the present study shows the cylinder axis and the long black vertical line (extending
and those of Sohn et al. (2008) could be attributed to the difference to the base of the cylinder) shows the drain port axis. The shorter
in the experimental conditions, viz., in the magnitude of Reynolds vertical line is along the axis of the vortex at the fluid surface. It
number (R ¼ ω · D2 =υ) and Froude number [F ¼ ωðD=gÞ1=2 ]; could be seen that the position of the vortex surface dip sways to
Reynolds number and Froude number differ by about 20% and either side during draining. More detailed investigations are re-
4.5% respectively (between the two studies), which could bring quired to find out the exact reasons for this fluid motion behavior.

Fig. 4. Vortex intermittency with curvilinear profile: (a–d) observed at d=D ¼ 0.09 at 200 rpm; (e) straight vortex air core in concentric
port

© ASCE 06016002-3 J. Aerosp. Eng.

J. Aerosp. Eng., 06016002

 On Vortex Suppression by Port Sizing of the drain port location during draining, also exhibiting an
intermittency character.
One of the key findings of the present study is that for a given tank
geometry, for the range of values of drain port diameters considered
here, the growth rate of the core air vortex decreases with decreas-
ing size of the drain port. For a sufficiently small value of the port Notation
diameter the core vortex is prevented from reaching the port. A
The following symbols are used in this paper:
physical explanation for the observation can be given by a simpli-
D = cylinder diameter (mm);
fied consideration of the dominant forces acting on the fluid
d = drain port diameter (mm);
element in the vertical direction. For a concentric port, the fluid
d=D = relative port ratio;
column above the port (a cylindrical column with diameter equal
e = eccentric distance (mm);
to the port diameter d) can be considered to be primarily acted up
g = acceleration due to gravity;
Downloaded from ascelibrary.org by University of California, San Diego on 03/08/16. Copyright ASCE. For personal use only; all rights reserved.

on by forces due to gravity, viscosity, and forces in the radial di-

Hi = initial height of water (mm);
rection (including pressure forces). The dominance of gravitational
Hc = critical height of water at which vortex formation
force on fluid motion in a similar configuration has been discussed
occurs (mm);
by Singh (2004). Considering impending vertical fall of the central
N = rotational speed (rpm);
water column (at a specific height), the downward movement can
T empðrÞ = time of emptying with rotation (s);
be theorized as being assisted by the gravitational forces acting on
T empðnrÞ = time of emptying without rotation (s);
the volume and opposed by the viscous forces acting on the vertical
ν = kinematic viscosity (mm2 =s); and
surface of the fluid column. Because the volume (πd2 × H) varies
ω = rotational velocity (rad=s).
as d2 and the (lateral) surface area (πd × H) varies as d, an increase
in the value of d (the port diameter) is favorable for a faster drain
(and hence an increase in the growth rate of the core vortex); by the
same argument, a reduction in d would lead to a reduced growth
rate of the vortex core. As far as the qualitative impact of drain port References
diameter on vortex formation is concerned, this explanation can be
Abramson, H. N., Chu, W. H., Garza, L. R., and Ransleben, G. E. (1962).
extended to the eccentric port drain as well, though the quantifica- “Some studies of liquid rotation and vortexing in rocket propellant
tion of forces in the eccentric port case would require a detailed tanks.” NASA D-1212, NASA, Washington, DC.
theoretical formulation that would account for the eccentricity. Basu, P., Agarwal, D., Tharakan, T. J., and Salih, A. (2013). “Numerical
The diameter of the tank (D) was not varied in the present set of studies on air-core vortex formation during draining of liquids from
experiments; more studies are required to quantify the impact of D tanks.” Int. J. Fluid Mech. Res., 40(1), 27–41.
on critical height. Gowda, B. H. L. (1996). “Draining of liquids from tanks of square or rec-
tangular cross sections.” J. Spacecraft Rockets, 33(2), 311–312.
Gowda, B. H. L., Joshy, P. J., and Swarnamani, S. (1996). “Device to sup-
Conclusions press vortexing during draining from cylindrical tanks.” J. Spacecraft
Rockets, 33(4), 598–600.
The results of the present study show that by adopting a suitable Lubin, B. T., and Springer, G. S. (1967). “The formation of a dip on
drain port size, the vortexing phenomenon could be completely the surface of a liquid draining from a tank.” J. Fluid Mech., 29(2),
eliminated. In the series of experiments reported here, conducted 385–390.
with a tank diameter of 100 mm and various port diameters, a Ramamurthi, K., and Tharakan, J. T. (1993). “Shaped discharge ports for
draining liquids.” J. Spacecraft Rockets, 30(6), 786–788.
port size ratio of 0.03 (or below) suppressed the vortex formation
Singh, P. K. (2004). “Scale model experiments and numerical study on a
for concentric ports. For eccentric ports, vortex suppression takes steel teeming process.” Master’s thesis, Univ. of Kentucky, Lexington,
place at a higher value of port size ratio (d=D ¼ 0.04). During KY.
eccentric port draining, the vortex-air core profile is curvilinear Sohn, C. H., Gowda, B. H. L., and Ju, M. G. (2008). “Eccentric drain port
in contrast to the linear (straight) profile for concentric port to prevent vortexing during draining from cylindrical tanks.” J. Space-
configurations. The curvilinear vortex-air core shifts to either side craft Rockets, 45(3), 638–640.

© ASCE 06016002-4 J. Aerosp. Eng.

J. Aerosp. Eng., 06016002