Professional Documents
Culture Documents
V O R T E X D R O P SHAFTS ....
By Michael C. Quick,1 Member, ASCE
INTRODUCTION
309
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit h
sign might prove to be a useful, economical alternative design, especially
for underground installations where size is a factor.
310
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit h
For various practical reasons, it was considered desirable to try to simplify
the drop structure entrance design. For a particular application, extensive
spiral guide walls would have been very expensive to build and therefore
trial designs were examined. The slot design shown in Fig. 1 was discovered
by one of those happy accidents. It was discovered, that a simple guide wall
leading into a 90° slot cut in a vertical circular pipe gave a very satisfactory
flow pattern. A generous air core existed and the flow further down the drop
shaft looked almost identical to the flow produced by the spiral guide wall.
Clearly the fluid mechanics of the two designs are different, but the head-
discharge relationships and the resulting drop-shaft flow are surprisingly sim-
ilar. Consequently, a more detailed experimental and theoretical study of the
two designs was carried out and the results are presented in the following
sections.
The free spiral vortex has been analyzed by Binnie and Hookings (1948)
who were the first to show that the flow is critical at the contracted throat
section, radius b, as shown in Fig. 2. Consequently, the discharge is max-
imum for the given head and swirl conditions. However, the solution is not
complete, because the value of the circulation, T, must be specified, where
T = 2TTUaR = 2TTC (2)
0.6
FIG. 3. Theoretical and Measured Head-Discharge Results for Spiral Guide Wall
Vortex Drop Structure
wall and drop shaft geometry. In this sense, the control is not totally at the
critical section, because the inlet geometry is asymmetrical, which creates
a torque, or swirl, in the approach flow. This swirl, represented by the vor-
tex strength, Cj limits the discharge by increasing the vortex core diameter,
b. In contrast to the more usual critical section where only the upstream
energy value needs to be specified, in the present situation both the upstream
energy and the upstream swirl, or vorticity, must be specified. A similar
method of analysis was presented by Ackers and Crump (1960), in which
the emphasis was on the vortex core diameter at the critical throat section.
The present approach emphasizes the dual nature of the control that is shared
between the critical section at the throat and the circulation produced by the
asymmetry of the approach flow. Also the present approach leads to a simple
procedure that can be used to calculate discharge for any specified value of
total energy, H, and for a specified geometry. The results of the theory are
compared with experimental measurements plotted in Fig. 3.
The critical control section occurs at the lip of the shaft, which is the
region of maximum contraction. Using the symbols defined in Fig. 2 and
setting the datum at the level of the lip, which is also the bed of the approach
channel, the energy on the free surface at the critical section is purely kinetic
and can be written at radius b
V2 ul
H =— + — (3)
2g 2g
where V = the vertical velocity of flow; and ub = the tangential vortex
velocity at the critical section where R = b. Assuming no energy loss, H is
also equal to the energy in the approach flow, which, if the approach ve-
locity is small, is essentially equal to the depth of the approach flow.
Assuming a free vortex distribution
c = R„Mm = bub (4)
312
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit
where Rm = the maximum radius at entry, and the total velocity, u„, = the
tangential vortex velocity, ue.
Combining Eqs. 3 and 4
2\ '/2
V=[2gH~- (5)
The discharge is equal to the vertical flow at the critical section, given by
the product of the vertical velocity, V, and the flow cross section,
Q = ir(a2 - b2)V (6)
Note that no contraction coefficient is specified, and checks against mea-
sured values indicate that the swirling flow appears to suppress the contrac-
tion.
Substituting for V from Eq. 5,
/ c2\i/2
Q = TT(a2 - b2)i 2gH - - 1 (7)
um = ^- (12)
Rm
and the swirl velocity, ub, at the critical section is estimated from
ub=C- (13)
b
An estimate, Qu of the discharge at the critical section is calculated from
313
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit ht
Eqs. 5 and 6. A second estimate, Q2, for the discharge at the inlet section
is also calculated from the estimated values of velocity and depth at the inlet
section using Eqs. 14-17 set out herein.
Assuming a free vortex distribution at section 0, the tangential velocity,
which is also the total velocity at radius R, is
u=- (14)
R
The depth, HR, at a general radius R at the entry section is, by energy
U2 c2
H
x=H-~ — == HH-T-i (15)
2g 2QR1
The discharge, Q2, at the entry section is then estimated from the integral
HBubR (16)
i.e.
c3 (\ 1: \ / W
Q2 = — I -7 ; I - He In I 1 - — I (17)
2 2
4g \R m (Rm - W) ) \ R,
This latest estimate of discharge, Q2, is used to find a second estimate of
the vertical velocity at the critical section from Eq. 6
Qi
TT(a — b)
314
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit
develop and test the physical dimensions.
This iterative procedure recognizes the dual nature of the flow control that
is defined by the critical section at the top of the shaft, together with the
vorticity induced by the approach flow geometry. The two discharge esti-
mates are calculated at these two sections and iteration proceeds until these
estimates agree.
It is worth noting that the head-discharge relationship for this spiral-walled
vortex drop structure is almost linear, so that the various complex interac-
tions of the vertical and vortex flow yields a very simple result.
The design is simply a vertical pipe which has a 90° rectangular slot cut
into its top and has a vertical guide wall that leads the approach flow into
the slot, as shown in Figs. 1 and 4. In the present tests, the approach flow
was unconstrained apart from the guide wall, but a more constrained ap-
proach flow would produce a similar result.
The vortex flow, shown in Fig. 4, it is more like a weir flow, but is
modified by its own flow, that spirals around and impacts on the side of the
incoming jet. This "feedback" of the weir crest flow reduces the total flow
through the slot by raising the pressure in the merging slot flow and by
reducing the jet velocity because of the sharing of momentum as the primary
slot jet and the "feedback" secondary jet mix together. In contrast to the
spiral guide wall vortex, the critical section is now near the inlet, essentially
like a weir flow.
Referring to Fig. 4, the momentum analysis will be applied to the control
volume defined by sections 1 and 2, firstly for a pure main flow with no
returning side jet to impinge on the flow, and secondly for the side jet im-
pinging as shown in the figure.
For the pure weir flow, the upstream condition at section 1 is difficult to
SECTION A-A
FIG. 4. Plan and Section of Vertical Slot Vortex and Definition of Control Volume
315
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit h
define because of the three-dimensional nature of the flow, but it is essen-
tially equivalent to the momentum flux at section 2, which can be defined
from standard weir theory
Momentum flux at 1 s pQ0V0 (23)
where Q„ = the normal weir discharge for a simple rectangular weir; and
Va — the mean velocity for such a flow. The assumption can be justified as
follows.
Considering the situation for a simple weir flow, with no intersecting side
jet, the momentum flux across section 1 in Fig. 4 is made up of a large
pressure term together with a small velocity flux term. For the control vol-
ume bounded by sections 1 and 2, there will be additional pressure and shear
terms on the walls cd and ab of Fig. 4 , but only the pressures on ab and
the small shears on cd will act in the direction of the required momentum
flux. The net effect of the momentum flux at section 1 and these other ex-
ternal forces is equal to the momentum flux at section 2 expressed by pQ0Va.
With the introduction of the side jet, or "feedback" jet, it is assumed that
the momentum flux across section 1 will be essentially unaltered for the
same value of approach flow depth H. There will be a slight reduction in
the already small velocity term and perhaps a small change in pressure and
shear on the walls ab and cd, but there will also be partly compensating
increases in pressure in section 1. Only the pressure changes on wall ab can
influence the required component of momentum flux and it is therefore con-
sidered that, even for the reintersecting side jet, pQ„Va is a reasonably ac-
curate assessment of the momentum flux in the jet direction across section
1, where
Now turning to the flow condition when there is a "feedback" jet im-
pinging on the main flow, the momentum equation in the direction of the
main entering weir flow can be written as
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit htt
momentum of the main and side jet flows; and (3) the force exerted by the
side jet on the main jet, in the direction of flow. The dimension, A, is the
depth to the jet intersection point, Fig. 4, and the depth of the free surface,
e, is Vl/2g.
It should be noted that because the side jet is at right angles to the main
jet, the side jet has no momentum in the main jet direction. This is a sim-
plification that occurs for this particular slot geometry. It is assumed that
the side jet mixes with the main jet and in the process the main jet mo-
mentum is shared with the side jet.
The final term, the force on the main jet, requires some justification be-
cause it is assumed that the total momentum of the side jet produces a dis-
tributed pressure within the main jet. On impact with the main jet, the side
jet velocity is reduced to zero and there is a corresponding increase in pres-
sure equal to the original side jet momentum, pqsVs. This pressure is om-
nidirectional and therefore produces a force component at section 2 of pqsVs
that opposes the main jet.
The four terms of Eq. 26 are integrated as follows:
D f D ,
pqmVdh = pCD- 2ghdh = pCfl - gA2 (28)
^ Jo *•
H *H
This final integral contains a reduction factor, a, because qs, the returning
side jet, is only part of the original incoming flow. This qs flow arises from
the upper part of the flow entering the slot, which impinges on the curved
wall of the pipe and splits into two distinct portions, an upper "splash" por-
tion, called the splash jet, and the lower main jet. This lower main jet drops
more quickly and, unless H is very large, does not re-intersect the incoming
jet. An expression for the a factor will be derived later.
Gathering up the terms for Eq. 26 from Eqs. 2 7 - 3 0 , and canceling the
common factors, the final form is
k2(H2 - A2) + (1 + a)A2 - 2otAH - (1 - a)H2 = 0 (31)
This equation is used to calculate k when A, H, and a are specified.
317
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit ht
FRACTION OF CROSS SECTION
FIG. 5. Definition of Main and Side Jet Fractions for Calculating Factor/
Q,
'f.H*[ §qm (32)
Below the intersection of the splash jet with the main jet, Fig. 5 defines
the flow through a horizontal slice. The main jet flow, hqm, occupies a frac-
tion (1 - / ) of the total cross section, while the returning side jet, hqs,
occupies the remainder. The following equation can be written by recog-
nizing that hqs arises from the incoming flow at a distance A above the sec-
tion under consideration, and a fraction, a, of this flow forms the returning
D
§qs = CD-a V2g<7! - A) hh . (33)
i.e.
2 r)
fir = - CD - V2g {A3/2 + k(H3/2 - A3/2) - a(ff - A)3/2} (40)
a FACTOR
The entering main jet collides with the back wall of the pipe and divides
into an upwards splash jet and a downwards main jet. An estimate of this
subdivision can be made by comparing with the two dimensional problem
shown in Fig. 6.
By vertical momentum at the wall
QV sin 6 = QdV - QUV. (41)
where Q = the incident flow that splits into an up component, Qu, and a
down component, Qd, so that
Q = Qu + Qd (42)
— = - (1 - sin 8) = a (43)
2 2
where a = the flow reduction factor for the splash flow, introduced in Eq.
30.
The angle 8 is defined from the free fall of the jet, where, as defined in
Fig. 6
u, « VgHc ••• (44)
in which Hc = the critical depth.
The time, t, for the jet to reach the wall is,
-D
f = P— (45)
Uo
where (3 = the fraction of the circumference traversed by the jet.
The vertical velocity, v, of the jet at the wall will be
g(3-£>
v = gt = —== (46)
'gffc
The angle 8 is defined from Eqs. 4 4 and 4 6 by
IT
-D
v 2 3 D
tan 6 = - = (3 — = - PIT - (47)
u„ Hc 4 H
where, approximately, Hc » 2 / 3 H. It would b e possible to reevaluate Hc
iteratively, but the results indicate that this is not necessary.
For the given geometry, which consists of three quarters of a circle, the
(3 value is of the order 0 . 7 5 , allowing 0 to be calculated for any specified
H value. Knowing 8, Eq. 4 3 defines the a value.
J E T INTERSECTION DEPTH, A
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit
, 1 ., ,D2
A = - trir — , (49)
8 Hc
The right-hand side is a constant for the system, so that the product, AHC,
is constant. Writing Hc equal to 2 / 3 H,
3 1
- - pVD- (51)
2 \8
1. The (3 factor is defined from the geometry, and is 0.75 in this situation
because the flow traverses three quarters of a circle.
2. The jet angle, 6, associated with the upward jet splash is calculated from
Eq. 47.
3. The flow reduction factor, a, for a splash flow is calculated from Eq. 43.
4. Eq. 51 is used to calculate A, the depth to the jet intersection point.
5. The velocity reduction factor, k, is calculated from Eq. 31, utilizing the
a, A, and H values.
6. Finally QT is calculated from Eq. 40.
The calculated and measured values of the flow are plotted in Fig. 7,
together with the curve for pure weir flow. At low values of H, the flow
becomes simple weir flow because there is no reintersection of the flow with
0.06
0.0 5
J L__L
0.1 0.2 0.3 0.4 0.5
HEAD(m)
FIG. 7. Theoretical and Measured Head-Discharge Results for the Vertical Slot
Vortex Drop Structure
321
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit http:
the entering jet. As H increases, there is a steady divergence from the weir-
type flow, and it will be noted that the vertical slot vortex discharge is lin-
early related to the head, H. This linear relationship contrasts with the three-
halves power law for weir flow. It can also be seen that the theoretical flows
are in excellent agreement with the measured values. It should be noted that
the only empirical coefficient used is Cd, the discharge coefficient for a rect-
angular weir: all the other coefficients and parameters, such as a, (3, K, and
A, are calculated, admittedly with some approximations. Of course, it could
be argued that Cd is also theoretically derived from the well-known con-
traction coefficient value of TT/(IT + 2).
EXPERIMENTS
The model drop structure tests were carried out using 200-mm (8-in.) di-
ameter drop shafts mounted in a 900-mm-wide by 600-mm-deep channel.
The spiral guide wall or the guide wall for the slot vortex were set up within
this wider channel. Leading dimensions are set out on Figs. 2 and 4. A
plexiglass drop shaft was used so that the drop shaft flow behavior could be
observed. The apparatus was set up so that flow discharged into a large basin
that was used as a volumetric tank for measuring the discharge. The dis-
charge measurements are therefore quite accurate and reliable. Measure-
ments of surface profile and velocities were also made so that various the-
oretical predictions could be checked. The primary results presented here are
the various head-discharge measurements.
The experimental results show that both designs of drop structure operate
satisfactorily and produce very similar flow conditions in the vertical drop
shaft. This statement raises the question of just what is satisfactory or un-
satisfactory behavior. Drop shafts are required to dissipate energy of the
falling water safely and with little risk of damage to the discharge tunnels.
Damage can be caused by high velocity impact of falling water. A serious
risk is the entrainment of air into the downstream flow, especially if there
is a horizontal discharge tunnel, as pointed out by Kennedy et al. (1988)
and earlier by Binnie (1938) when discussing the so-called morning glory
spillway design. Entrained air can become pressurized downstream and be-
come quite explosive when suddenly released. Sudden priming of the ver-
tical shaft from air-core to pipe-full conditions can cause serious surging of
the flow and gives rise to shock pressures. In general, a good stable air core
prevents most of these problems. With an air core present, the control is at
the top of the shaft and this limits the discharge and prevents surging.
A generous size of air core allows air to pass up or down the shaft, and
generally this large air core eliminates much of the air entrainment problems,
as reported by Jain and Kennedy (1983). However, if the shaft extends to
some considerable depth, the fast-moving water entrains considerable quan-
tities of air. This air can be carried downstream into horizontal tunnels after
the flow undergoes the transition to pipe-full flow at the bottom of the shaft.
This transition to pipe-full flow often involves the formation of an annular
hydraulic jump, as reported by the writer (Quick 1961).
It can therefore be concluded that both designs of drop structure behave
322
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit
0.06
FIG. 8. Comparison of Head-Discharge Results for Spiral Guide Wall and Ver-
tical Slot Drop Structures
well and in a very similar fashion, as can be seen from Fig. 8. The theo-
retical predictions agree well with the measured values, as shown in Figs.
3 and 7, and are therefore valid for design application. The vertical slot
design offers a useful advantage, because it takes up much less space and
is simpler to construct. This compactness of design could be a valuable at-
tribute for underground installation or where space is at a premium.
For preliminary design purposes, the discharges for geometrically similar
boundaries can be approximately scaled using
& 2i Hi
= \i Hi'
(52)
For the geometries used in this study, a combined equation for either the
log spiral or the slot is approximately
Q = 1A2DISH m3/s • (53)
Final design values should be computed by programming the equations
set out above. A copy of a program, written in Turbobasic for an IBM PC,
is available at nominal cost and user's risk. The user should carry out ad-
equate checks on the calculation accuracy.
ACKNOWLEDGMENTS
The work was started when investigating suitable designs of drop shaft
for flooding a salt mine for Cargil Corporation of New York. Further ex-
perimental investigations were made to confirm the theoretical predictions.
The first preliminary set of experiments was carried out by Yves Crits, an
IASTE summer student from Belgium, and the second, more complete set
of experiments, was carried out as a bachelors degree project by Steve Dent,
Jody Evans, Don Demque, and Ken Tomczyk. The thorough and enthu-
siastic work of all these students is warmly acknowledged.
323
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit h
APPENDIX I. REFERENCES
Ackers, P., and Crump, E. S. (1960). "The Vortex drop." Proc. of the Inst, of Civ.
Engrs., 16, Aug., 433-442.
Binnie, A. M., and Hookings, G. A. (1948). "Experiments on whirlpools." Proc.
of the Royal Society, Series A, 194, 398-413.
Binnie, G. M. (1938). "Experiments on bellmouth and siphon spillways." J. Inst,
of Civ. Engrs., 10, Nov., 65-90.
Hager, W. H. (1985). "Head-discharge relation for vortex shaft." J. Hydr. Engrg.,
ASCE, 3(6), 1015-1020.
Jain, S. C , and Kennedy, J. F. (1983). "Vortex-flow dropstructures for the Mil-
waukee metropolitan sewage district inline storage system." IIHR Report No. 264,
Iowa Inst, of Hydr. Res., Univ. of Iowa, Iowa City, Iowa.
Kennedy, J. F., Jain, S. C , and Quinones, H. R. (1988). "Helicoidal ramp drop-
shaft." J. Hydr. Engrg., ASCE, 114(3), 315-325.
Quick, M. C. (1961). "The annular hydraulic jump." Civ. Engrg. and Public Works
Review, 56(662), 1176-1179.
Vallentine, H. R. (1967), Applied hydrodynamics. Butterworths Press, London, 2nd
Ed., 117.
Downloaded 04 May 2011 to 95.131.110.102. Redistribution subject to ASCE license or copyright. Visit http