2017 2nd International Conference for Convergence in Technology (I2CT)

Vortexing During Draining Through Twin Drain

Ports In A Cylindrical Tank
Gopikrishnan T. H., Anandakrishnan U., Ayilavarapu Sai Kiran, Ajith Kumar R. and Anith T. A.
Department of Mechanical Engineering,
Amrita School of Engineering, Amritapuri,
Amrita Vishwa Vidyapeetham, Amrita University, India
gopikrishnan.pillai@gmail.com, onsaku@gmail.com, saikiranayilavarapu@gmail.com,
amritanjali.ajith@gmail.com,55anith55@gmail.com

Abstract—In this paper, the results of an investigation on the (1986) studied the main factors affecting the critical
formation of vortex air core commonly known as ‘vortexing’ submergence (Hc) of vortex formation. He derived an
occurring during draining of a liquid (water) column from a analytical expression for Hc in terms of Froude number,
cylindrical tank through two eccentric ports (diameters d1 and Reynolds number, Circulation number and Weber number.
d2) simultaneously, are presented and discussed; d1 is fixed Baum (1974) has shown that the onset of gas entrainment
whereas d2 is varied. Before draining, the liquid column is depends on Reynolds number, Froude number and Weber
induced with a rotation (N rpm) using a motorized stirrer where number and also proposed a correlation to predict this
rotational speed can be controlled. Drain port eccentricities are entrainment phenomenon. Baum and Cook (1975) have
equal (e). During draining, as the liquid level reaches certain
conducted experiments with different fluids to determine the
height called ‘critical height’ (Hc), a vortex air core develops and
necessary conditions for the inception of gas entrainment in a
extends down touching the drain port. Results show that as N is
increased, Hc is increased. At N=200rpm (highest speed tested), vortex with an unstable gas core. They have shown that surface
at the smallest eccentricity, lowest critical height occurs for the tension plays a significant role and hence, cannot be neglected.
largest value of d2 (equal to d1).At larger eccentricities, a These are some of the papers which deal with the fundaments
significantly different highly non-linear variation of Hc w.r.t N is of vortex-induced gas entrainment and subsequent ‘vortexing’
seen with vortexing nearly suppressed for the case with drain phenomenon.
port sizes equal. In many engineering systems, occurrence of Literature reveals that some interesting works were carried
‘vortexing’ is reported and hence, the present study carries
out intending to suppress the vortexing phenomenon. Baffles
significant practical relevance. All the studies reported so far
were with single drain port and hence, the present study is first
were used by Abramson at al.(1962) to suppress sloshing
of its kind. which also prevents vortex formation. Ramamurthi and
Tharakan (1992) have shown that stepped drain port could
Index Terms—vortexing, liquid draining, critical height, time arrest vortex formation. Gowda (1996) has shown that in tanks
of emptying. with square and rectangular cross-sections, vortexing is
arrested. One of the previous study has shown that a device
‘dish-type suppressor’ can prevent vortexing (Gowda (1996)).
I. INTRODUCTION
Studies have revealed that eccentric drain ports control vortex
Free surface vortex formation is a commonly observed formation (Sohn et al.(2008)).
phenomenon in many engineering applications such as in Base surface roughness was found to reduce the critical
hydraulic engineering, metal casting, nuclear reactors and height of vortex formation (which indicates the intensity of
rocket propellant systems. Swirling flows contained in a tank, vortexing) by about 70% (Prabhu et al (2015)). In a recent
when allowed to drain though a port, entrains air or gas from experimental study by Ajith Kumar et al (2016), it was
the environment and forms a vortex-gas core which develops proposed that drain port size has a remarkable influence on the
and extends down to the drain port partially blocking the port critical height and suggested that below a critical value of port
area. This process eventually leads to a reduction in the flow size, vortexing disappears or completely suppressed. It is to be
discharge rate. Since, it is generally causes undesirable effects, noted that, all these studies pertain to the application of a single
scientists/engineers have been trying to avoid its occurrence drain port.
through various means. But, before attempting to ‘kill’ such
vortex formation, it is essential to understand the phenomenon As mentioned earlier, two of the previous studies (Sohn et
from the fundamental point of view, its characteristics and the al (2008) and Ajith Kumar et al (2016)) have revealed two
factors influencing it. strategies for vortex suppression, viz., suitable drain port size
(below a critical value) and port eccentricity. But, when we
Lubin and Springer (1967) have studied the liquid draining reduce the drain port below the critical size to suppress
process through a circular port from a cylindrical tank vortexing, then, it would adversely affect the magnitude of
experimentally and have derived an analytical expression for discharge flow rate which would obviously result in longer
the critical height for various fluid combinations. Odgaard time of emptying of the tank. On the other hand, in the case of

978-1-5090-4307-1/17/$31.00 ©2017 IEEE 325

eccentric ports, port eccentricity itself act as a vortexsuppressor 10mm were drilled on the base plate (Fig. 1(b)) to perform a
(Sohn et al (2008)) thereby enabling a larger critical port wide range of experiments to meet the objectives of the present
diameter (compared to a concentric drain port case). But, in study. The drain port eccentricities are equal, viz. ‘e’ as marked
this case (port eccentricity), eccentric location of drain port in Fig. 1(b); e=11mm, 21mm and 31mm are employed. The
might induce a de-stabilizing couple in aerospace engineering glass cylinder was glued to the acrylic plates by using silicone
applications such as in liquid propellant tanks (Agarwal et al gel. The experimental procedure is the same as that detailed in
(2014)). It is thought that, this problem could be defeated by Ajith Kumar et al (2016). As the draining starts, at first a dip
employing two eccentric ports for the draining of liquids. In forms at the free surface which extends to the drain port almost
this way, the total discharge flow rate also could be increased instantaneously in the form of a vortex with an air core and the
avoiding the delay in time of emptying. This forms the liquid column height at which this occurs is measured as the
motivation for the present study with two drain ports critical height (Hc) in all the experiments. Critical height was
eccentrically located at the base of a liquid (water) filled tank. measured using transparent scale attached vertically on
As already mentioned, to the author’s knowledge, all the cylinder. Initial rotation imparted to the liquid column (N rpm)
previous studies were conducted with only one drain port. is provided in the range from 60rpm to 200rpm. Draining is
Hence, the present study is thought to be first of its kind being allowed through two drain ports with diameters d1 and d2
reported in this paper. The main objective of this study is to where d1 is fixed at 10mm varying d2 assuming values 4mm,
investigate the characteristics of vortexing phenomenon when 6mm, 8mm and 10mm. Also, the study was conducted at three
liquid draining takesplace through two eccentric drain ports values of initial height of the liquid column (Hi), viz., 300mm,
simultaneously. 350mm and 400mm. The liquid used is water at room
temperature.
II. EXPERIMENTAL SET UP DETAILS
III. RESULTS AND DISCUSSION

Results are presented as non-dimensional critical height

(Hc/Hi) Versus N (rpm) and Td/To Versus N plots as shown in
Figs 2-7; Td is the time of emptying the tank with liquid
column rotation and To, the time of emptying without fluid
rotation. To reiterate, draining is allowed through two drain
ports simultaneously with diameters d1 and d2; d1 is fixed
(=10mm) whereas d2 is varied and takes values 4 mm (case1),
6mm (case 2), 8mm (case 3) and 10mm (case 4). It is to be
specifically noted that, To is different for different d1-d2
combinations.
Fig.2 shows the variation of Hc/Hi w.r.t N at e=11mm for
various values of Hi. As the figure shows, Hc/Hi increases with
increasing N for all the cases 1-4. Quite interestingly, it could
be seen that, as d2 increases, the critical height drops down for
all the cases at all Hi values tested at a given value of N. This
indicates that the intensity of vortexing reduces as d2 increases
which is an unexpected result from the perspective of single
Fig.1(a) Schematic of test cylindrical tank drain port results. For single drain ports, Hc/Hi increases as
drain port size increases (Ajith Kumar et al 2016). Overall
trends of Hc/Hi Vs N are nearly the same at all the Hi values as
seen in Figs 2(a) to (c). But, results indicate that, there are
minor variations in the Hc/Hi trends at different Hi values with
some differences in the magnitude of critical height values at
different speeds (60rpm to 200rpm) at this eccentricity
(=11mm).
Fig.3 shows the non-dimensionalised time of emptying of
the tank corresponding to the critical height depicted in Fig.2
for all the cases (e=11mm) at all Hi values. In general, Td/To
increases with N as Hc/Hi does. Further, it could be seen that,
at all Hi values, Td/To is the highest for d2= 6mm and the least
for d2 = 4mm. There is no definite trend observed for the
Fig.1(b) Tank bottom base showing two eccentric ports variation w.r.t d2. On the contrary, for a single drain port
A cylindrical glass tank is used having an internal diameter whether it is concentric or eccentric, as port diameter increases,
(D) of 96mm and height (H) of 410mm with base plate made non-dimensional time of emptying also increases because
out of acrylic plate as shown in Fig. 1(a). Two eccentric drain critical height increases (Ajith Kumar et al 2016). As critical
holes of different diameters (d1 and d2) varying from 4mm to height increases, vortex air core formed blocks the drain port

326
early reducing the discharge flow rate. As the flow rate is
reduced, it takes more time to empty the tank or in other words,
time of emptying with rotation (Td) increases for a single drain
port.

Fig.3 Variation of non-dimensionalised Time of Emptying

w.r.t N (rpm) at e = 11mm for various values of Hi.
At the same time, at much lower N values (40rpm-60 rpm),
Td/To drops down (see Fig. 3(a)) indicating that draining is
Fig.2 Variation of non-dimensional critical height w.r.t N faster than the case without rotation. This aspect also indicates
(rpm) at e = 11mm for various values of Hi that possibly a different flow physics is operating at very low
liquid rotation speeds.When the port eccentricity (e) is
Furthermore, the time of emptying without rotation (To) increased to 21mm, significant changes occur in the trends of
reduces when port diameter increases for single drain port case Hc/Hi w.r.t N as shown in Figs.4 (a)–(c) for all values of Hi.
at a given initial liquid column height (Hi) since, in that case, Except at d2=4mm, Hc/Hi follows an increasing-decreasing
flow discharge increases with larger port diameter. Therefore, trend with N ending up in negligibly small values of critical
Td/To increase is multifold in that case. Whereas, as pointed height at N=200 rpm. For d2=4mm, Hc/Hi steeply increases at
out earlier, this systematic trend w.r.t drain port size is not first at N=60 rpm and then, exhibits a gradually increasing
observed in the case of twin drain port probably because of the trend for N > 60 rpm. This shows that, vortexing is nearly
fact that the non-dimensional time of emptying with and suppressed for d2=6mm, 8mm and 10mm at the highest speed
without rotation (Td& To) undergo non-linear variations due to tested (200rpm). This is again very much contrary to the
the ‘interference’ between the flow fields around the drain expectations because, as pointed out earlier, for single drain
ports during flow discharges through them. This non-uniform ports, bigger the port diameter, higher will be the critical height
trend in Td/To w.r.t d2 is observed at all Hi values as seen in indicting more intense vortexing phenomenon.It is to be
Figs 3(a)-(c).

327
pointed out that, in this case, one drain port appears to act as a draining process with rotation (compared to that without
‘vortex suppressor’ for the other. It could also be seen that, the rotation). When the port eccentricity is further increased to
maximum critical height occurs more or less in the speed range e=31mm, the critical height trends have notably changed as
100 rpm < N < 150 rpm at all Hi values (Fig.4). It is to be shown in Fig.6. Hc/Hi increases w.r.t N reaches a maximum
highlighted that, at d2=10mm, this gives a combination of equal value and then, reduces in general. At Hi=400mm (Fig. 6(a)),
drain port diameters (d1=10mm). In this case, vortexing is d2 =4mm gives rise to the maximum critical height but still
greatly suppressed almost throughout the rotational speed much lesser than that at the lower values of eccentricity
range except around N=120rpm. The overall maximum critical (=11mm, 21mm); there is about 78% reduction in the
height observed in all these cases (e-21mm) is considerably maximum value compared to the case at e=11mm.
lesser (by about 50% at N=200 rpm) than that at e=11mm.

Fig.4 Variation of non-dimensional critical height w.r.t N Fig.5 Variation of non-dimensional Time of Emptying w.r.t
(rpm) at e = 21mm for various values of Hi. N (rpm) at e = 21mm for various values of Hi.
The corresponding non-dimensional time of emptying For higher d2 values, vortexing is only feeble with
(Td/To) is shown in Fig.5 where, in general, a monotonously substantially lower Hc/Hi values at this Hi value. At lower Hi
increasing trend is seen at all Hi values. To particularly note, values (=350mm and 300mm), overall maximum Hc/Hi value
Td/To goes below 1.0 ford2 =4mm in the speed range 40rpm further reduces where only feeble vortexing occurring with
<N < 100rpm for Hi=400mm and 350mm indicating a faster significantly lower Hc/Hi values for d2 =8mm and 10mm. That
is, extremely weak vortex air core forms when the draining is

328
about to complete. To particularly note, vortexing is then extends again further to which it once gain draws back to
completely suppressed for d2=10mm (equal port size case) at its dip configuration. This ‘on and off’ switching of air core
all Hi values. Fig.7 show the corresponding time of emptying indicates possible variation of static pressure around the drain
plots. The overall trends are notably different from the ports. In these snap shots, the cylinder center line (long white
previous cases of lower eccentricities. But, like in the previous line) and the drain port center lines (short lines) are shown so
cases, there is no definite trend for the variation of Hc/Hi w.r.t as to enable trace the vortex motion and its sway to either
N observed in this case also at all the Hi values. sides.
A common feature observed is that, for d2 =6mm, for
N>120 rpm, Hc/Hi drops down reaching lower values (Td/To<
1.0). For d2 =8mm, up to N=140rpm, Td/To is less than 1.0 at
Hi=350mm and 400mm. Both these features indicate faster
draining occurring with fluid rotation possibly indicating a
different flow physics operating near the drain ports.

Fig.7 Variation of non-dimensional Time of Emptying w.r.t

N (rpm) at e = 31mm for various values of Hi.

The observed features of vortex air core formation with

draining through twin drain ports could possibly be due to the
Fig.6 Variation of non-dimensional critical height w.r.t N changes in the flow structure around the drain ports
(rpm) at e = 31mm for various values of Hi. consequently leading to changes in the static pressures around
them as also influenced by the bottom boundary layer.
Fig.8 shows the temporal evolution of the vortex air core However, the exact reasons for typical trends observed in the
starting from the moment of critical height formation (7th sec critical height and time of emptying are not known and further
from the inception of draining) at e=11mm, d2 =4mm, investigations are needed to unfold the underlying physics
Hi=350mm & N=200rpm. It was seen that, during draining, the involved.
vortex air core sways to either sides of the cylinder axis
touching the eccentric drain ports as evidenced in the
photographs (video images) given in Fig.8.As seen in Fig.8, in
the 19th sec, vortex air core gets back to its dip configuration,

329
 drain port configuration. Also, the lower port sizes (d2) give
rise to higher critical height than others.
(d) At low port eccentricity, non-dimensional time of
emptying increases w.r.t N for all drain port size combinations.
(e) At higher port eccentricities, non-dimensional time of
emptying shows different and more non-linear trends w.r.t N.
Also, there no systematic trend observed w.r.t d2.
(f) Quite interestingly, in some lower rotational speed
range, for a certain port size combination, draining occurs more
rapidly than in the case of still water draining (no rotation)
contrasting with single drain port results.
(g) Flow visualization reveals that, during draining, vortex
air core switches from one drain port to the other with a
curvilinear surface profile. Furthermore, vortex air core also
exhibits intermittent behavior as evidenced by its temporal
evolution over a period of time.

ACKNOWLEDGMENT
We express our heartfelt gratitude to the Department of
Mechanical Engineering, Amrita University, Amritapuri
Campus, for providing us with the apparatus, space and
encouragement for the completion of this work.

REFERENCES
[1] B.TLubinet and G.S. Springer, “The formation of a dip on the surface of
a liquid draining from a tank”, Journal of Fluid Mechanics, Vol. 29, No.
2, pp. 385-390, 1967.
[2] M R Baum andM E Cook, “Gas entrainment at the free surface of a
liquid: entrainment inception at a vortex with an unstable gas core”,
Nuclear Engineering and Design Journal. Vol. 32, No. 2, pp 239-245,
1975.
Fig. 8 Temporal evolution of the vortex air coreat e=11mm,
[3] M R Baum, “Gas entrainment at the free surface of a liquid: entrainment
d2=4mm, Hi=350mm & N=200rpm inception at a laminar vortex”, Journal of British Nuclear Energy
Society, Vol. 13, pp. 203-209, 1974.
IV. CONCLUSIONS [4] R Ajith Kumar, J Joykutty, R Shaji, and Srikrishnan A R, "Vortex
Suppression through Drain Port Sizing." Journal of Aerospace
From the results presented and discussed on the vortexing Engineering, Vl. 29, No. 4, DOI:10.1061/(ASCE)AS.1943-
phenomenon observed in a cylindrical tank with two drain 5525.0000609, 06016002, July, 2016.
ports with equal eccentricity, the following conclusions are [6] A.J. Odgaard, “Free-Surface Air Core Vortex”, Journal of Hydraulic
drawn: Engineering,Vol. 112, No. 7, July, 1986.
[7] H. N. Abramson, W. H. Chu, L. R.Garza and G. E. Ransleben, “Some
(a) The critical height of vortex formation and non- Studies of Liquid Rotation and Vortexing in Rocket Propellant Tanks,”
dimensional time of emptying for twin drain port configuration NASA D-1212, pp. 1–35, 1962.
significantly vary according to the speed of rotation of the [8] B.H.L Gowda, “Draining of Liquids from Tanks of Square or
liquid column, port sizes and port eccentricity. Critical height RectangularCross Sections,”Journal of Spacecraft and Rockets, vol. 33,
is only weakly influence by the initial height of the liquid pp. 311–312,1996.
column. [9] Ramamurthi, K., and Tharakan, J. T., “Shaped Discharge Ports for
Draining Liquids,” Journal of Spacecraft and Rockets, Vol. 30, No. 6,
(b) At very low port eccentricity, the critical height of 1992, pp. 786–788.
vortex formation monotonically increases w.r.t N but decreases [10] C.H. Sohn, B.H.L.Gowda, and M.G.Ju, “Eccentric drain port to prevent
as the size of the one of the drain ports (d2) is increased vortexing during draining from cylindrical tanks,” Journal of Spacecraft
and Rockets, Vol.45, No.3, pp. 638-640, 2008.
keeping the size of the other drain port fixed (d1).
[11] D. Agarwal., P. Basu., T. John Tharakan., A. Salih., Prediction of gas-
(c) At higher port eccentricities, critical height exhibits core vortices during draining of liquid propellants from tanks, Journal of
completely different and non-linear trends w.r.t speed of fluid Aerospace Science and Technology, Vol. 32, pp. 60-65, 2014.
rotation. Even at the highest speed of rotation of the liquid [12] J.P.Mahadev., Ragesh.R.Nair., R. Ajith Kumar., “Vortexing during
column tested, at equal drain port size combinations, vortexing draining from cylindrical tanks: Effect of base roughness”, Proceedings
of the 17th ISME Conference, ISME 17, IIT Delhi, October 3-4, 2015.
is found to be completely suppressed which is quite counter-
intuitive with regard to the previously reported results on single

330