Effects of Intake Geometry on the Occurrence

of a Free-Surface Vortex
Kerem Taştan 1 and Nevzat Yıldırım 2

Abstract: Both the profile and orientation of an intake entrance basically determine the characteristics of the flow (especially the velocity
distribution and therefore the hydraulically developed subsurface depressions) toward the intake. The profile and orientation of the intake
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

affect (1) the process of accumulation and growth of vorticity (circulation) along the pathway axis of the swirling vortex from the free surface
to the intake; (2) position of the pathway of the swirling-vortex filament; and (3) location of the point of entry of the free-surface or subsurface
vortex at the entrance of the intake. Test results indicate that the intake-entrance profile has limited effects on the occurrence of the air-core
vortex and critical submergence. Regardless of the profile of the intake entrance, one common physical mechanism exists for development of
free vortices occurring at intake entrances of different profiles. It is proven that spherical sink surface sectors (SSSSs) are essential for the
vortex to exist. DOI: 10.1061/(ASCE)HY.1943-7900.0001439. © 2018 American Society of Civil Engineers.
Author keywords: Air entrainment; Critical submergence; Entrance profile; Intake; Vortex.

Introduction the effect of the profile of the intake entrance on the air core or free
vortex and critical submergence is limited. The main mechanism of
The vertical distance between the level of intake center and water occurrence of the air-core vortex is identical for all types of intake-
surface level is called submergence, S. If S is below a certain mini- entrance profiles.
mum, air enters the intake via an air-entraining vortex [Fig. 1(a)].
The submergence at which the tip of the air-core vortex reaches the
intake is called the critical submergence, Sc [S ¼ Sc , critical intake Physical Considerations
condition, Fig. 1(b)]. If S < Sc , air enters the intake [Fig. 1(c)],
causing hydraulic problems such as discharge reduction, vibrations, The flow visualization in the study by Kocabaş and Yıldırım (2002,
and loss of efficiency in pumps, turbines, water-conveying struc- Fig. 2) indicated that the pathlines of the particles are in the form of
tures, and pressure conduits of hydraulic power stations. Studies descending and converging unclosed swirling lines. These lines
relating to the critical submergence and air-entraining vortices form the lateral surfaces of an infinite number of coaxial swirling
occurring at circular intakes include those by Denny (1956), Anwar paraboloid revolutions (hydraulically developed imaginary Stokes’
et al. (1978), Jain et al. (1978), Gulliver and Rindels (1987), stream surfaces) of circular cross sections extending from the free
surface to the intake, which appear as a swirling vortex filament
Yıldırım and Kocabaş (1995), Kocabaş and Yıldırım (2002),
[Figs. 2(a and b)]. The intake-entrance profile has two main effects
Yıldırım et al. (2011), Taştan and Yıldırım (2010), Sarkardeh et al.
on the occurrence of the air-core vortex, as discussed in the follow-
(2010), Naderi et al. (2014), Suerich-Gulick et al. (2014), and
ing subsections.
Taştan (2017). In practice, noncircular intakes are also frequently
used. Studies relating to noncircular intake entrances include those
by Yıldırım (2004), Eroğlu and Bahadırlı (2007), Yang et al. (2014), Velocity Distribution Effect
and Taştan (2016). Those studies indicated that intake-entrance The intake-entrance profile affects the velocity distribution at the
geometry affects the occurrence of the vortex and critical submer- intake entrance and thus the velocity distribution within the ambi-
gence. However, they did not give explanations or reasons regarding ent fluid flow toward the intake entrance. The intake-entrance
why the intake-entrance geometry affects the occurrence of the profile determines the continuous streamline of the geometrical lo-
air-core vortex and critical submergence. cation of the maximum velocities and their magnitudes at all levels
In the present study, the development mechanism of the air-core within the ambient fluid flow toward the intake entrance (including
vortices occurring at circular and noncircular intakes and the rea- the intake entrance itself) all the way from the free surface to the
sons relating to why the critical submergences differ for different intake entrance. Accordingly, it also determines the characteristics
intake-entrance geometries are examined. Test results indicate that of an infinite number of coaxial subsurface depressions (SSDs) of a
point-sink character that hydraulically develop within the ambient
1
Postdoctoral Research Fellow, Dept. of Civil Engineering, Faculty of flow field. This occurs as follows.
Engineering, Gazi Univ., Maltepe, Ankara 06570, Turkey (corresponding Each particle [e.g., particles A1 , A2 , A3 , etc., in Fig. 2(c)] on the
author). E-mail: [email protected] streamline of the geometrical location of the maximum velocities
2
Professor, Dept. of Civil Engineering, Faculty of Engineering, has the largest velocity among all particles at the same level as that
Çankaya Univ., Merkez Kampüs, Etimesgut, Ankara 06790, Turkey.
particle. Therefore, as the particles with the maximum velocities on
E-mail: [email protected]
Note. This manuscript was submitted on November 1, 2016; approved this streamline move toward the intake, they drag their neighboring
on October 11, 2017; published online on January 31, 2018. Discussion particles along and thereby cause the development of coaxial elon-
period open until June 30, 2018; separate discussions must be submitted gating SSDs of circular cross sections, as explained by Yıldırım
for individual papers. This paper is part of the Journal of Hydraulic et al. (2012, Fig. 1). These SSDs behave as coaxial subsurface in-
Engineering, © ASCE, ISSN 0733-9429. takes of point-sink character for the ambient fluid. The SSDs may

© ASCE 04018009-1 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

 air-core vortex that is also an SSD filled out by both the fluid and
air, these three cases have been observed and witnessed to exist
during experiments by several researchers mentioned earlier, in-
cluding Padmanabhan and Hacker (1984, Fig. 1), Denny (1956,
Fig. 22), and Jain et al. (1978, Fig. 2).
For any type of intake entrance, there exists a single line of geo-
metrical locations of particles of maximum velocities at each level
from the free surface to the intake entrance with an infinite number
of coaxial SSDs or DIs of point-sink character with their own re-
spective spherical sink surface sectors (SSSSs) that hydraulically
develop within the ambient fluid.
Fig. 1. Flow at a single pipe intake: (a) no air-core vortex; (b) air-core The intake-entrance profile determines both the location and
vortex just reaches the intake (critical intake condition); (c) air- magnitude of the maximum velocity at the intake entrance. There-
entraining vortex fore, it also determines the location and configuration of the stream-
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

line of the geometrical location of the points of maximum velocities

or coaxial elongating SSDs or DIs within the ambient fluid all the
way from the free surface to the intake.
also be called depression intakes (DIs). The boundary of the SSD
or DI is a Stokes’ stream surface [Figs. 2(a and b)], as explained
earlier. As indicated in Fig. 3, continuity relating to the ambient Magnitude of the Portion of Intake Discharge
fluid flow filling out the SSDs dictates that there can be three Participating in the Vortex Filament (Vorticity-Feeding
cases relating to the lower end (tip) of the SSD (or Stokes’ stream Intake-Discharge Effect)
surface): (1) the lower end (tip) of the SSD (or Stokes’ stream The intake-entrance profile determines the amount of ambient fluid
surface) is a fixed stagnation point because of the ambient fluid flow (portion of the intake discharge participating in the vortex
flow filling out the SSD stops [Fig. 3(a)]; (2) the lower end (tip) of filament) toward the hydraulically developed SSDs of point-sink
the SSD continuously moves downward toward the intake to in- characters. It also determines the magnitude of the total vorticity
crease its dimension as the ambient fluid flow fills it in [Fig. 3(b)]; harvested or inherited by this flow through the ambient fluid
and (3) the lower end (tip) of the SSD has to have an open end medium and entrained or fed into the flow along the pathway axis
(opening) because the ambient fluid flows into the intake entrance of the vortex all along the way from the free surface to the intake
by passing through this open end [Fig. 3(c)]. Indeed, for the entrance for the vortex. This causes an increment in the strength of

Fig. 2. (a) Subsurface depressions, or Stokes’ stream surfaces, and an SSSS; (b) velocity components V rB and V θB of the particle B; (c) geometric
locations of maximum velocities and development of elongating subsurface depressions

© ASCE 04018009-2 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

 depth of 0.806S − S is solely attributable the effect of the bottom
boundary. To the authors of the present study, the increment in cir-
culation across S is unavoidable whether or not a bottom boundary
exists (it does not solely depend on the boundary effect). Similar
statements relating to the upper vortex or the upper SSDs are also
valid for the development of the lower vortex or the coaxial lower
SSDs (within the ambient fluid below the level of the intake en-
trance) extending from the bottom boundary to the intake entrance
(Fig. 4).

Theoretical Considerations
Fig. 3. (a) Nonelongating subsurface depression; (b) elongating
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

subsurface depression; (c) open-ended subsurface depression Hypothesis

This study proposes the hypothesis that the geometrical location
of fluid particles that have identical magnitudes of radial velocity
components toward any chosen point on the axis of the vortex
(vortex pathline) extending from the free surface to the intake en-
trance is a hydraulically developed SSSS whose center is the same
as that point. The air-core vortex or free vortex in general is the
result of stages of concurrent collapse and rotation of an infinite
number of coaxial SSSSs (under the effects of Coriolis forces
and imposed or induced circulations) as they and their changing
centers travel toward the intake entrance through their respective
SSDs or Stokes’ stream surfaces.
The verification of this hypothesis is as follows. Consider a hy-
draulically developed Stokes’ stream surface (imaginary boundary
of an SSD or DI) and any point P on the axis of symmetry of the
Stokes’ stream surface (or DI or the pathway axis of the vortex) as
indicated in Figs. 2(a and b). To avoid overcrowding of these fig-
ures, the spiral pathlines of the infinite number of symmetrical par-
ticles about the axis line of the Stokes’ stream surface, e.g., N, U, Y,
H, E, and so on, other than B, are not indicated on the lateral sur-
Fig. 4. Upper and lower vortices for an inclined or horizontal intake
faces of the SSDs (Stokes’ stream surface). The velocity vector of
the particle B lying on this Stokes’ stream-surface is denoted by V B.
The radial component of V B toward the center point P is denoted by
circulation, stimulating the occurrence of the air-core vortex. This V rB . The geometrical location of whole fluid particles that have an
occurs as follows. identical magnitude of radial velocity of V rB toward any chosen
Regarding the development of the upper swirling vortex flow center point P on the axis of the vortex (vortex pathline) is an SSSS,
(Figs. 2 and 4), every free-surface or subsurface fluid particle fol- as presented in Figs. 2(a and b). These particles belong to an infinite
lows a descending and converging unclosed swirling pathline about number of coaxial Stokes’ stream surfaces (Fig. 2). In reality, the
the vertical axis of the vortex. These particles input the sum of the total velocity V B is the result of the radial velocity V rB and circu-
imposed or induced vortices, Coriolis effects (Earth’s rotation), latory velocity about the axis line of the SSD, V θB , which are attrib-
nonuniformities present within the ambient fluid, and rotations that uted to the point-sink flow character of the DI and effects of the
have been inherited along their entire paths to the swirling vortex Coriolis and vorticity on the fluid particle, respectively. In addition,
filament at the points of their arrivals on the pathway axis of the V B , V rB , and V θB are tangential to the lateral surface of the Stokes’
vortex. Afterwards, it continues on its swirling pathway to reach the stream surface. V rB is tangential to the Stokes’ stream surface in the
intake by colliding with the particles coming from the upper levels plane on which both points B and P (or pathway axis of the vortex)
(i.e., at its upstream). Therefore, it causes further disturbances, in- occur. For example, in Fig. 5, particles G, D, A, E, and F of differ-
stabilities, and rotations for those particles reaching the swirling ent Stokes’ stream surfaces (DIs) are on the same SSSS I and have
vortex filament below the level of its point of arrival in the swirling identical radial velocity toward the center M1 of SSSS I. Similarly,
vortex region, thereby causing the circulation to increase further at particles B, C, K, J, L, N, and T of different Stokes’ stream surfaces
the lower levels (i.e., at its downstream). This should be the reason (DIs) are on the same SSSS II and have an identical radial velocity
why the magnitude of the circulation at each level along the vertical toward the center M2 of SSSS II. Each SSSS has the same dis-
axis (pathway) of the air-core vortex or the swirling vortex in- charge as its own SSD, which determines its maximum base. The
creases as one goes from the free surface to the intake. radius of the SSSS for the DI with the center at the free surface is
In fact, the test data provided by Sun and Liu (2015, Fig. 7) zero. Based on the test data in the studies by Sun and Liu (2015)
indicated that as one goes along the vortex flow from the free and Daggett and Keulegan (1974), the verification relating to the
surface to the intake, the increment in the magnitudes of both the existence of the SSSS is given a little later.
circulation and radius of the vortex core from the free surface to the Because the radial velocity toward the center point P
depth of 0.806S are negligibly small. However, they increase con- [Figs. 2(a and b)] of the SSSS is tangential to the Stokes’ stream
siderably from the depth of 0.806S to the intake or S. Sun and Liu surface or lateral surface of the SSD at point B and normal to the
(2015, Fig. 7) claimed that the sharp increment in circulation for the spherical sink surface, the SSSS has to normally (perpendicularly)

© ASCE 04018009-3 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Position of air-core vortex in conjunction with S and Sc : (a) air-

core vortex is not visible; (b) air-core vortex is visible; (c) air-core
vortex becomes a fully developed air-entraining vortex

vortex in suspension [Fig. 6(b)]. In reality, the lower end (tip)

of the air-core vortex (dimple) should have a tiny opening, but
the surface tension (capillary) forces overcome the small inertia
and gravity forces at the lower end of the air-core vortex. There-
fore, the lower end (tip) of the air-core vortex is seen as closed and
slightly curved.
Fig. 5. SSSSs and Stokes’ stream surfaces or SSDs
If the depth of the ambient fluid above the SSD is far from being
sufficient (i.e., S ≤ Sc ) to fill out the SSD with an open end, both
the outside ambient fluid and air together fill in the SSD, and as
a result, the vortex is seen as a fully developed air-entraining vortex
intersect the Stokes’ stream surface (Figs. 2 and 5). As the SSSS [Fig. 6(c)].
goes from the free surface toward the intake, it rotates and becomes
small (collapses), and concurrently, its center P moves downward
toward the intake on the axis line of the Stokes’ stream surface. Verification of Existence of SSSSs
Continuity dictates that during this motion, the velocity at the SSSS
Considering the test data in the study by Daggett and Keulegan
has to increase.
(1974, Fig. 8) relating to the velocities measured across different
Because the fluid body of the spherical sector is continuously
vertical depths at different horizontal radial distances of r=ðD=2Þ ¼
rotating and collapsing while it moves toward the intake entrance
2.64 (or r ¼ 13.4 cm), 3.54 (r ¼ 17.98 cm), and 4.44 (r ¼ 22.5 cm)
(the Stokes’ stream surface narrows toward the intake), the radial
(herein, r = horizontal radial distance and D = diameter of the cir-
velocity at the surface of the SSSS increases greatly. Therefore, the cular orifice intake), it can be shown that the geometrical location
shear deformation, vorticity, and circulation, and thus the circula- of the points (i.e., G, M, and N in Fig. 7 herein) of identical velocity
tory velocity, increase in the region at and close to the lower section (2.769 cm=s) at these different vertical and horizontal radial distan-
of the Stokes’ stream surface. At the level of the intake entrance, the ces is the same SSSS that has the center C (on the vertical central
circulation and circulatory velocity attain their maximum magni- axis of the intake) 19.487 cm from the free surface or 10.993 cm
tudes as measured and observed by Sun and Liu (2015, Fig. 7). from the bottom floor boundary and a radius of 23.5 cm, as indi-
Because of the increments in both the circulation and circulatory cated in the scaled Fig. 7. If desired, one can draw more SSSSs.
velocity in the lower section of the Stokes’ stream surface, the out- This indicates that SSSSs exist. Using a miniature propeller flow
ward centrifugal forces on the particles in the lower section of the meter, Daggett and Keulegan (1974) believed that they had mea-
Stokes’ stream surface increase. This also means that the circula- sured the horizontal radial velocity, U (horizontal radial velocity
tory velocity at the circumference of the base of the SSSS is larger at the lateral surfaces of the coaxial cylinders corresponding
than those at and close to the top of the SSSS. If the centrifugal to the vertical depths with the specified horizontal radial distances),
force overcomes the pressure and surface tension (capillary) forces in accordance with the vertical axis of the circular orifice intake.
acting on the particles, the very bottom closed end (tip) of the SSD In reality, they had measured the radial velocity, V rs , in accor-
elongating (traveling) from the free surface to the intake to reach dance with the center of C of the SSSS in the present study
the intake entrance becomes an open end through which the am- (Fig. 7). Therefore, the measured magnitudes of U in Daggett and
bient fluid filling in the SSD reaches and enters the intake entrance. Keulegan’s (1974) Fig. 8 are identical to the magnitudes of V rs in
If the depth of the ambient fluid above the SSD is large enough the present study (V rs ¼ U).
(i.e., S ≫ Sc ) to fill out the SSD with or without an open end, the As stated earlier, an air-core vortex can also be considered as a
outside ambient fluid alone comfortably fills out the SSD, and the subsurface vortex or SSD. By assuming a hydrostatic pressure dis-
vortex is seen as a swirling fluid vortex filament without an air core tribution, Sun and Liu [2015, Eq. (12)] rewrote the theoretical for-
(air-core vortex is not visible at the free surface) [Fig. 6(a)]. mula derived by Vatistas et al. (1991) for the profile of the air-core
If the depth of the ambient fluid above the SSD is barely suffi- vortex relating to a circular intake entrance as follows:
cient or slightly less than sufficient (i.e., S ≈ Sc ) to fill out the SSD
alone, the SSD is filled out by both the outside ambient fluid and air Hr − H0 2
¼ arctan R2 ð1Þ
together, and thus, the vortex is seen as a shallow or deep air-core h π

© ASCE 04018009-4 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 7. Scaled drawing relating to the test set-up of Daggett and

Keulegan (1974) and an SSSS

where Q = intake discharge; V = average intake-entrance velocity;

R = dimensionless (R ¼ r=rm ) horizontal radial distance from the
vertical centerline of the vortex or intake entrance, where rm is vor-
tex core radius (outer limit radius of the viscous region); H r = water
surface elevation at R; H 0 = elevation of the lower end or tip of the Fig. 8. Geometric variables’ definitions
air-core vortex; and h = total depth of the air core, as indicated in
Fig. 8(a). Sun and Liu [2015, Eq. (14)] introduced a multiplication
coefficient of 2 for R2 in Eq. (1) (which may be attributed to the
2
Hr − H0 a
nonhydrostatic distribution in reality, as stated at the beginning of R2 þ − ¼ R2s ð4Þ
the present study) to get it fitted very well with their test data relat- h h
ing to the profile of the air-core vortex occurring at a circular intake
entrance. They also proposed a semiempirical expression for the where Rs = dimensionless radius of a different type of SSSS, in-
profile of the air-core vortex, as follows: dicated as SSSS (in the dimensionless coordinate system of H r =h
and R) in Fig. 8(b). Both H0 and a can take positive and negative
Hr − H0 2 signs (Fig. 8). From Eq. (3)
¼ arctanð2R2 Þ ð2Þ 

h π d Hr − H0 4βR
¼ ð5Þ
Eqs. (1) and (2) indicate that the profile of the air-core vortex
dR h air-core πð1 þ β 2 R4 Þ
may change with flow conditions. For general purposes, it may be
From Eq. (4)
appropriate to write Eqs. (1) and (2) in a single form as follows:


d Hr − H0 R
Hr − H0 2 ¼ − Hr −H0 a ð6Þ
¼ arctanðβR2 Þ ð3Þ dR h circular arch −h
h π h

where β = coefficient relating to the pressure distribution in Because the SSSS and profile of the air-core vortex are normal
conjunction with the flow conditions. The magnitude of β changes to each other, the product of their slopes should be equal to −1.
with the flow conditions. Eq. (3) can also be used for SSDs Thus, considering this condition, from Eqs. (5) and (6)
provided that the magnitudes of β and their depths are known or " #
4βR −R
given. Because the SSSS is the revolution of a circular arch, con- × Hr −H0 a ¼ −1 ð7Þ
sider an SSSS whose center C on the axis of symmetry of the intake πð1 þ β 2 R4 Þ h −h
entrance or air-core vortex is above the lower end (tip) of the air-
core vortex for an amount of a or a=h, as shown in Figs. 8(a and b). Eq. (7) yields (for the intersection of the SSSS and profile of
For the circular arch, one can write the air-core vortex)

© ASCE 04018009-5 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

 " sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
1 1 2
R2 ¼ Hr −H0 a þ 
2 H r −H 0
 − 0.25
a 2
ð8Þ
π h −h π h −h β

Test data from Sun and Liu (2015, Fig. 12) indicate that the plus
sign before the square-root term on the right side in Eq. (8) is to be
considered. Thus, for the intersection of the SSSS and the profile
of the air-core vortex, Eqs. (8) and (3) result in

 ( "
Hr − H0 2 1
¼ arctan 2 Hr −H0 a
h π π h −h
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #)
1
þ 
2 H r −H 0
 − 0.25 ð9Þ
− ah 2
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

π h

Being independent of β, Eq. (9) indicates that ðH r − H0 Þ=h

for the intersection point of the SSSS is independent of the flow
condition and depends solely on a=h. In Eq. (9), the term within
the square root must be positive, for which 1=fπ2 ½ðH r − H0 Þ=
h − ða=hÞ2 g ≥ 0.25, and therefore, the upper limit for ½ðHr − H0 Þ=
h − ða=hÞ is


Hr − H0 a 2
− ≤ ð¼ 0.63694Þ ð10Þ
h h π

From Eqs. (10) and (8), the upper limit for R is

1
R ≥ pffiffiffi ð11Þ
β

Eq. (9) indicates that for every point on the vertical axis of
the vortex, there corresponds an SSSS (or SSSS). Eq. (9) can be
solved for ðH r − H0 Þ=h by a trial and error procedure for a chosen Fig. 9. (a) SSSS for the cases of a=h ¼ 0 and a=h ¼ 0.3; (b) param-
magnitude of a=h. Eqs. (4)–(8) indicate that for each magnitude of eters for the critical intake condition; (c) CSSS
β, there is a corresponding vortex profile or SSD and an SSSS
normally intersecting the vortex profile (intersection point of the
vortex profile or SSD and the SSSS ) for a given a=h. To prove
the existence of the SSSS , the procedure is as follows: This computational procedure is also valid for the free vortex or
1. Use the same scale for both the vertical ðH r − H0 Þ=h and air-core vortex and other SSDs provided that the magnitudes of β
horizontal R axes of the coordinate system; and their depths h are known or given.
2. Choose a=h and mark the center C of the SSSS on the verti- To present the existence of the SSSS , two different examples
cal axis; are given next.
3. Choose a magnitude for β;
4. By means of Eq. (3), draw the profile of the vortex (or SSD); Example 1
5. Draw a tangent line (line of the velocity vector toward C) from Consider β ¼ 2 [same case as in the study by Sun and Liu (2015),
the center C of the SSSS to the vortex profile. Measure or read Fig. 12] and a=h ¼ 0 for which the center C1 of the SSSS is the
the magnitudes of R, ðHr − H0 Þ=h, and Rs corresponding to same as the lower end (tip) Point T of the vortex profile (Points C
the tangent point on the vortex profile; and T coincide in Fig. 8). By means of Eq. (3), the profile of the
6. Draw the SSSS that has a center of C and radius of Rs ; vortex is drawn in Fig. 9(a), which has the same scale for both the
7. Measure the slopes and slope angles of the tangents of both horizontal and vertical axes. For β ¼ 2, Eq. (2) or (3) almost
the vortex profile (SSD) and SSSS , the radial velocity vector perfectly match the test data relating to the vortex profile provided
toward C, and the angle between the tangent lines of the SSSS by Sun and Liu (2015, Fig. 12). Therefore, their test data are
and vortex profile; not indicated in Fig. 9(a). In this figure, from the center C1,
8. From Eqs. (4)–(9), compute R, ðHr − H0 Þ=h, Rs , the slopes draw a tangent line to the vortex profile and mark the tangent
and slope angles of the tangents of both the SSSS and vortex Point B1 on the vortex profile. For the Point B1 , in Fig. 9(a),
profile, the radial velocity vector toward C, and the angle the measured or read magnitudes of the parameters are as follows:
between the tangents of the SSSS and vortex profile; and ðH r − H0 Þ=h ¼ 0.603, R ¼ 0.834, Rs ¼ C1 B1 ¼ 1.0295, slope
9. Compare the computed and measured magnitudes of R, angles of the tangent line of the vortex profile = 35.88°, and radial
ðHr − H0 Þ=h, and Rs , the slopes and slope angles of the tan- velocity vector toward C1 ¼ 215.88°. The SSSS of radius Rs ¼
gents of both the SSSS and vortex profile, the slope and slope 1.0295 is drawn in Fig. 9(a). The slope angle of the tangent line
angle of the radial velocity vector, and the angle between the of the SSSS ¼ 125.88°, and the angle between the tangent lines
tangents of the SSSS and vortex profile. If they are the same of the SSSS and vortex profile is measured as 90°. For point B1 , by
(or acceptably close), it means that one or more SSSS exist. means of Eqs. (4)–(9), the computed magnitudes of the parameters

© ASCE 04018009-6 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

 are as follows: ðH r −H 0 Þ=h ¼ 0.6033; R ¼ 0.8343; Rs ¼ 1.0296; 2πS2c ¼ constant ð16Þ
and the slope angles of the tangent line of the vortex profile, radial
velocity vector, and SSSS are 35.87°, 215.87°, and 125.87°, re- This is the surface area of an SSSS in a critical intake condition
spectively. The angle between the tangent lines of the SSSS that is the same as half of the so-called critical spherical sink
and vortex profile is ð125.87–35.87Þ ¼ ð215.88–125.87Þ ≅ 90°. surface (CSSS) in the related available literature (Yıldırım and
These results indicate that the measured and computed magnitudes Kocabaş 1995; Yıldırım et al. 2000). Eq. (16) is for the flow of
of the corresponding parameters are acceptably close. Therefore, the upper vortex region above the intake level. Considering the am-
the SSSS exists. bient fluid region below the level of the center of the intake
entrance, the lower vortex region should also have half of the
Example 2 CSSS of 2πS2c . Hence, for an intake that has flow from the regions
Choose β ¼ 2 and a=h ¼ 0.3. By following the same procedure of both the upper and lower vortices, Eq. (15) or (16) can be
as in the previous example [center of SSSS and the tangent rewritten
point are denoted by C2 and B2 , respectively, in Fig. 9(a)], the
measured or read magnitudes of the parameters are as follows: 4πS2c ¼ constant ð17Þ
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

ðHr − H0 Þ=h ¼ 0.75; R ¼ 1.10; Rs ¼ C2 B2 ¼ 1.187; and the

slope angles of the tangents of the vortex profile and SSSS of ra- As presented in Fig. 9, this is exactly the same surface area
dius Rs ¼ 1.187 and center of C2 drawn in Fig. 9(a) are 22.27 of a complete CSSS without boundary blockages that has the
(slope of the radial velocity vector ¼ 202.27°) and 112.27°, respec- same center and discharge as the intake as developed by Yıldırım
tively. The angle between the tangents of the SSSS and the and Kocabaş (1995), Yıldırım et al. (2011), and Taştan and
vortex profile is ð112.27–22.27Þ ¼ 90°. By means of Eqs. (4)–(9), Yıldırım (2014).
the computed magnitudes of the parameters are as follows: To examine the effects of the intake-entrance profiles on the oc-
ðHr − H0 Þ=h ¼ 0.75; R ¼ 1.1; Rs ¼ 1.187; and the slope angles currence of the air-entraining vortex and the critical submergence,
of the tangents of the SSSS , vortex profile, and radial velocity the following tests were conducted.
vector toward B2 are 22.23, 112.25, and 202.23°, respectively.
The angle between the tangent lines of the SSSS and vortex profile
is ð112.25–22.23Þ ¼ 90.02° ≅ 90°. The results indicate that the Experiments
measured and computed magnitudes of the corresponding param-
The test setup, presented in Fig. 10, consists of a cylindrical pool
eters are almost the same. Thus, the SSSS is present.
(still water, net working pool diameter ¼ 190 cm, net pool
For a ¼ 0 and H0 ¼ 0 (or h ¼ S ¼ Sc ), the air-core vortex is
height ¼ 120 cm, and no imposed circulation). To suppress the
just a fully developed one that corresponds to the critical intake
boundary blockage on the intake flow and thus to examine exclu-
condition for which the aforementioned results indicate that [for
sively the effects of the intake-entrance profile, the clearances of
a=h ¼ 0; Eq. (9) yields ðHr − H 0 Þ=h ¼ 0.6033, R ¼ 0.8343]
the intake entrance to the bottom and side boundaries of the pool
Hr ¼ 0.6033Sc ; r ¼ 0.8343rm ð12Þ are kept much larger than Sc (Yıldırım et al. 2000) (except for the
unavoidable blockage effect of the intake duct pipe itself, which is
Now, one can go to Fig. 9(c) (with a ¼ 0; H0 ¼ 0; Hr ¼ approximately the same order of magnitude for all intake-entrance
0.6033Sc ; and r ¼ 0.8343rm ) and write the equation for a circle profiles of identical cross-sectional area used in this study). The
(whose revolution gives an SSSS) of radius Rs as follows: pool and measurement devices relating to the intake discharge,
water depth, submergence, and so on are the same as those used
r2 þ H 2r ¼ R2s ; ð0.8343rm Þ2 þ ð0.6033Sc Þ2 ¼ R2s ¼ constant by Kocabaş and Yıldırım (2002).
ð13Þ Six different typical intake-entrance profiles with sharp edges
(circular, rectangular, square, equilateral triangle, equilateral rec-
tangle, and star, as indicated in Fig. 11) are used. These intake
S2c ¼ constant − 1.912r2m ð14Þ
duct pipes are made of mild steel sheet with a thickness of
The test results of Sun and Liu (2015, Fig. 5) indicate that rm is a 2 mm. All of the intake entrances have an identical cross-sectional
constant for the vertical distances from the free-surface level less area of A ¼ 40 cm2 , and their uniformly cross-sectioned duct pipes
than or equal to 0.7S (or Hr =S ¼ 1–0.7 ¼ 0.3). Sun and Liu (2015) are 50 cm tall. In this study, for the intake entrance, the length of the
even claimed that if the boundary effect is not present, rm is a flow-admitting perimeter (LFAP) is preferably defined and used in
constant across the entire S and is completely independent of lieu of the well-known concept of the wetted perimeter because an
the vertical depth or S of the intake. They also state that if the boun- intake entrance may or may not admit flow from all of its sides or
dary effect is present, rm is a constant up to the vertical distance of sections of its perimeter. Therefore, the LFAP may or may not
0.806S from the free surface. be the same as the well-known wetted perimeter. All types of
Eq. (12) indicates that for H0 ¼ 0 and a ¼ 0 (h ¼ S ¼ Sc , intake entrances used in this study admit flow from all their sides.
critical intake condition), the vertical distance (depth) of point B Therefore, LFAP is the same as the wetted perimeter in the present
[Fig. 9(b)] from the free surface is ðh − H r Þat B ¼ ðS − Hr Þat B or study. The hydraulic radius is Rh ¼ A=LFAP. In the present
ðS − 0.6033SÞ ¼ 0.3967S < 0.7S, indicating that for the fully study, the range of hydraulic radius of the intake entrances is
developed air-core vortex or SSSS at critical intake condition, Rh ¼ 1.143–1.785 cm.
rm in Eq. (14) can be taken as a constant. Thus, for the critical A predetermined changeable-intake duct pipe is centrally and
intake condition, Eq. (14) can be rewritten vertically located and connected to the outlet line with a treaded
coupling at the bottom of the pool (vertically flowing downward
S2c ¼ constant ð15Þ intake). The vertical clearance of the intake entrance to the bottom
boundary of the pool is 50 cm. The intake duct pipe is connected
For the upper vortex region (or the ambient fluid region above to the outlet line (suction pipe of the pump). The pool is filled
the level of the center of the circular intake entrance), one can with water (S is much larger than the Sc ). The pump is started
multiply both sides of Eq. (15) by 2π and get and the valve on the suction line is opened for a desired intake

© ASCE 04018009-7 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 10. Test setup (not to scale)

discharge, Q. Water within the pool is recirculated by the pump. Discussion of Results
Observations are made over 1–2 h to determine whether an air-
entraining vortex occurs. If no air-entraining vortex occurs during Fig. 12 indicates that as the intake discharge increases, the change
this time period, the water level within the pool is decreased by a in Sc becomes negligible for different intake-entrance profiles
small amount (2–3 mm) by means of a valve on the drainage pipe. of identical cross-section area. The main reason for this is as fol-
When the water surface within the pool has fallen to the desired lows. As the submergence increases, characteristics of the flow,
level, the valve on the drainage pipe is completely closed to maintain especially in the upper sections of the fluid media, approach to the
a constant level. These steps are continued until an air-entraining point-sink flow character. For the small intake discharges, the
vortex occurs. Once the air-entraining vortex occurs, measurements change in Sc is approximately 0–20%, indicating that at low sub-
relating to Q, and the Sc , and so on are made and recorded. mergences, the flow characteristics toward the intakes may deviate
This test on a changeable intake-entrance profile is repeated from being point-sink flow.
for several different Q. Similar tests are conducted on the other The test results in Figs. 12–14 indicate that the effect of the pro-
intake-entrance profiles, and the results are presented in Figs. 12–14. file of the intake entrance on the air-core vortex or free vortex in
The effects of the intake-entrance Froude number F ¼ V=ðg4Rh Þ0.5, general is limited (considering the circular intake entrance as the
Reynolds number R ¼ V4Rh =ν, and Weber number W ¼ ρ4Rh V 2 =σ reference one). Therefore, the main mechanism of occurrence of
on the critical submergence are presented in Figs. 13(a–c). Herein, the air-core vortex should be identical for all types of intake-
V is average intake-entrance velocity, g is gravitational accelera- entrance profiles.
tion, ρ is fluid density, σ is surface tension, and ν is kinematic Theoretically, because its entire approach flow has the character
viscosity. In the present study, for water, values of ρ ¼ of a complete spherical sink, the entire discharge (100%) of the
1,000 kg=m3 , ν ¼ 1.02 × 10−6 m2 =s, and σ ¼ 0.072 N=m were circular intake entrance participates in the vortex filament and feed
taken (Jain et al. 1978). Because the circular intake entrance is vorticity to the vortex filament along its entire length from the free
more commonly used in practice, it is better to compare the effects surface to the intake. Conversely, because the circular intake en-
of the noncircular intake entrances with those of a circular one that trance is hydraulically the best and most effective intake, the veloc-
has an identical cross-sectional area. For this purpose, the ratio of ity distribution on its entrance cross section is close to uniform
Sc for noncircular intake entrances to that for the circular one, (although the velocity at and next to its boundary is slightly smaller
Sccir , is presented in Fig. 14. The data points in Fig. 14 were ob- than its central region). In conjunction with this, the magnitude of
tained from the real test data in Fig. 13 using the linear interpo- maximum velocities, and as a result, the depths of the SSDs are not
lation technique. expected to be excessively large. As stated earlier, under normal

© ASCE 04018009-8 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

Fig. 11. Types of intake-entrance profiles used in the tests (not to

scale); for all intakes, A ¼ 40 cm2

Fig. 12. Test data relating to Sc versus Q

conditions (no imposed or induced circulation), the vorticity-feeding

intake-discharge effect is much less than the velocity distribution
effect on the vortex.
In the case of noncircular intakes, as Rh decreases (or LFAP
increases), the magnitude of the portion of the intake discharge par-
ticipating in the vortex filament (vorticity-feeding intake-discharge
effect) decreases, but the velocity distribution on its entrance cross
section becomes far from uniform (velocity at and next to its boun-
dary becomes much smaller than that in its very central region).
Therefore, the magnitude of maximum velocities and thus the
depths of the SSDs are expected to be excessively large and narrow
for the noncircular intake entrances in comparison with that for the
circular intake entrance. This causes high shear flow and the gen-
eration of large vorticity or circulation within the SSD, which leads
to high Sc . This is the main reason why the magnitudes of Sc for
the intakes of small Rh are larger than those of large Rh , as the test
data indicate in Figs. 12–14. These figures indicate that the velocity
distribution effect and vorticity-feeding intake-discharge effect
work in opposition to each other for the development of the air-core
vortex. The greater the vorticity-feeding intake-discharge effect,
the smaller the velocity distribution effects (and vice versa).
Figs. 12–14 show that for F ≥ 3, R ≥ 5 × 104 , and W ≥ 2 × 103 ,
as Rh gets smaller (LFAP gets larger), the increment in the velocity Fig. 13. Effects of F, R, and W on Sc =ð4Rh Þ
distribution effect may well overcompensate for the loss in the

© ASCE 04018009-9 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

 In brief, for the air-core vortex to occur, it is essential that a fluid
mass of a spherical sector rotating about its vertical axis under
the effects of the Coriolis force and imposed or induced circulation
must exist or must be hydraulically developed.

Conclusions
From this study, the following conclusions may be drawn:
1. The geometrical locations of the fluid particles that have an
identical magnitude of the radial velocity component toward
any chosen point on the axis of the vortex (vortex pathline)
extending from the free surface to the intake entrance is a
hydraulically developed surface of a SSSS whose center is
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

the same as that point. The air-core vortex or free vortex

in general is the result of stages of concurrent collapse and
rotation of an infinite number of coaxial spherical SSSSs (un-
der the effects of the Coriolis force and imposed or induced
circulations) as they and their changing centers travel toward
the intake entrance through their respective SSDs or Stokes’
stream-surfaces;
2. For the air-core vortex to occur, it is essential that a coaxial
infinite number of SSDs (Stokes’ stream surfaces) of spherical
sink flow characteristics rotating about their vertical coaxes un-
der the effects of the Coriolis force and imposed or induced cir-
culation should be hydraulically developed in an ambient fluid
with an intake;
3. The effect of the profile of the intake entrance on the air-core
vortex or free vortex is in general limited. The main mechanism
of occurrence of the air-core vortex is identical for all types of
profiles of intake entrances;
4. The intake-entrance profile has two main effects on the occur-
rence of the air-core vortex:
• Velocity distribution effect; and
• Magnitude of the portion of the intake discharge participat-
ing in the vortex filament (vorticity-feeding intake-discharge
effect); and
5. Each flow and geometrical condition produces and bears its own
respective vorticity and velocity distributions within the ambient
fluid. Therefore, the magnitude of the circulation, the occur-
rence of the vortex, and the critical submergence differ for
different flow and geometrical conditions.

Notation

The following symbols are used in this paper:

A = cross-section area of the intake;
a = distance between the tip of the air-core vortex and center
of the SSSS;
D = diameter of the circular orifice intake;
Fig. 14. Comparison of Sc of noncircular intakes with that of circular F = intake Froude number;
intake Sccir g = gravitational acceleration;
Hr = water surface elevation at R;
H 0 = elevation of the lower end or tip of the air-core vortex;
vorticity-feeding intake-discharge effect (for F < 3, R < 5 × 104 , h = total depth of the air core;
and W < 2 × 103 , the increment in the velocity distribution effect Q = intake discharge;
and the loss in vorticity-feeding intake-discharge effect are of the R = dimensionless radius (R ¼ r=rm );
same order of magnitude, they balance each other). Fig. 14 indi- R = intake Reynolds number;
cates that for F > 1.5, the equilateral triangle and square intakes Rh = hydraulic radius based on LFAP;
behave in almost the same way as the circular intake of point-sink Rs = dimensionless radius of the SSSS ;
character (flows toward the equilateral triangle and square intakes r = horizontal radial distance from the vertical centerline of
may be approximated as of a point-sink flow). the vortex or intake entrance;

© ASCE 04018009-10 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009

 rm = vortex core radius (outer limit radius of the viscous Padmanabhan, M., and Hecker, G. E. (1984). “Scale effects in pump sump
region); models.” J. Hydraul. Eng., 10.1061/(ASCE)0733-9429(1984)110:
S = submergence; 11(1540), 1540–1556.
Sarkardeh, H., Zarrati, A. R., and Roshan, R. (2010). “Effect of the intake
Sc = critical submergence; head wall and trash rack on vortices.” J. Hydraul. Res., 48(1), 108–112.
Sccir = critical submergence for the circular intake; Suerich-Gulick, F., Gaskin, S., Villeneuve, M., and Parkinson, E. (2014).
U = horizontal radial velocity; “Free surface intake vortices: Theoretical model and measurements.”
V = average intake-entrance velocity; J. Hydraul. Res., 52(4), 502–512.
V rs = radial (normal) velocity at the SSSS; Sun, H., and Liu, Y. (2015). “Theoretical and experimental study on the
V θ = circulatory velocity; vortex at hydraulic intakes.” J. Hydraul. Res., 53(6), 787–796.
Taştan, K. (2016). “Critical submergence for isolated and dual rectangular
W = intake Weber number;
intakes.” Sadhana-Acad. P. Eng. S., 41(4), 425–433.
β = coefficient; Taştan, K. (2017). “Scale and flow boundary effects for air-entraining
ρ = fluid density; vortices.” Proc. Inst. Civ. Eng. Water Manage., 170(4), 198–206.
σ = surface tension; and Taştan, K., and Yıldırım, N. (2010). “Effects of dimensionless parameters
Downloaded from ascelibrary.org by University Of Florida on 02/03/18. Copyright ASCE. For personal use only; all rights reserved.

ν = kinematic viscosity. on air-entraining vortices.” J. Hydraul. Res., 48(1), 57–64.

Taştan, K., and Yıldırım, N. (2014). “Effects of Froude, Reynolds and
Weber numbers on an air-entraining vortex.” J. Hydraul. Res., 52(3),
References 421–425.
Vatistas, G. H., Kozel, V., and Mih, W. C. (1991). “A simpler model for
Anwar, H. O., Weller, J. A., and Amphlett, M. B. (1978). “Similarity of concentrated vortices.” Exp. Fluids, 11(1), 73–76.
free vortex at horizontal intake.” J. Hydraul. Res., 16(2), 95–105. Yang, J., Liu, T., Bottacin-Busolin, A., and Lin, C. (2014). “Effects of
Daggett, L. L., and Keulegan, G. H. (1974). “Similitude in free surface intake-entrance profiles on free-surface vortices.” J. Hydraul. Res.,
vortex formations.” J. Hydraul. Div., 100(HY11), 1565–1581. 52(4), 523–531.
Denny, D. F. (1956). “An experimental study of air-entraining vortices in
Yıldırım, N. (2004). “Critical submergence for a rectangular intake.” J. Eng.
pump sumps.” Proc. Inst. Mech. Eng., 170(2), 106–116.
Mech., 10.1061/(ASCE)0733-9399(2004)130:10(1195), 1195–1210.
Eroğlu, N., and Bahadırlı, T. (2007). “Prediction of critical submergence for
Yıldırım, N., Akay, H., and Taştan, K. (2011). “Critical submergence for
a rectangular intake.” J. Energy Eng., 10.1061/(ASCE)0733-9402
(2007)133:2(91), 91–103. multiple pipe intakes by the potential flow solution.” J. Hydraul. Res.,
Gulliver, S. J., and Rindels, A. J. (1987). “Weak vortices at vertical intakes.” 49(1), 117–121.
J. Hydraul. Eng., 10.1061/(ASCE)0733-9429(1987)113:9(1101), Yıldırım, N., Eroğlu, N., and Taştan, K. (2012). “Occurrence of vortices
1101–1116. at an intake of point sink character.” Teknik Dergi, 23, 5799–5812
Jain, A. K., Ranga Raju, K. G., and Garde, R. J. (1978). “Vortex formation (in Turkish).
at vertical pipe intake.” J. Hydraul. Div., 104(HY10), 1429–1448. Yıldırım, N., and Kocabaş, F. (1995). “Critical submergence for intakes
Kocabaş, F., and Yıldırım, N. (2002). “Effect of circulation on critical in open channel flow.” J. Hydraul. Eng., 10.1061/(ASCE)0733-9429
submergence of an intake pipe.” J. Hydraul. Res., 40(6), 741–752. (1995)121:12(900), 900–905.
Naderi, V., Farsadizadeh, D., Dalir, A. H., and Arvanaghi, H. (2014). Yıldırım, N., Kocabaş, F., and Gülcan, S. C. (2000). “Flow boundary ef-
“Effect of using vertical plates on vertical intake on discharge coeffi- fects of critical submergence of intake pipe.” J. Hydraul. Eng., 10.1061
cient.” Arab. J. Sci. Eng., 39(12), 8627–8633. /(ASCE)0733-9429(2000)126:4(288), 288–297.

© ASCE 04018009-11 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(4): 04018009