3-1 EM 1110-2-1602

15 Oct 80

SLUICES FOR CONCRETE GRAVITY DAMS

Section I. Basic Considerations

3-1. Location. Sluices for concrete dams are generally located along
the center line of spillway monoliths (plate C-18). When more than one
sluice per monolith is required they are spaced appropriately in each
monolith (plate C-19). A sluice should never be located close to or
straddling a monolith joint. Since it is also general practice to place
crest piers on the center line of spillway monoliths, the sluice air
vent intakes can be placed in the crest pier, eliminating any danger of
submergence during spillway flow. Air vents should not be cross-
connected below the highest possible pressure grade line. In some cases
it may be desirable to locate the sluices in the nonoverflow section of
the dam. Such a location requires either (a) a separate energy dissi-
pater or (b) a careful design for discharging into the spillway energy
dissipater.

3-2. Size, Shape, and Number. The sluices for concrete gravity dams
usually have arelatively small cross-sectional area. One of the prin-
cipal reasons for making the sluices small in cross section is adverse
structural effects of large openings in a concrete gravity section. In
addition, the use of a large number of small sluices, each controlled by
individual gates, provides a finer degree of regulation than could be
obtained from a smaller number of sluices of larger cross-sectional area.
The flood control sluices installed in Corps of Engineers' dams are
predominantly rectangular in cross section. The size of sluices usually
varies from 4 ft 0 in. by 6 ft 0 in. to 5 ft 8 in. by 10 ft 0 in., de-
pending on discharge requirements. Larger sizes may be indicated in
certain cases. All sluices should be large enough for inspection, main-
tenance, and repair purposes.

3-3. Elevation and Alignment.

a. General. The reservoir operational requirements normally play

an important part in determining the elevation of the flood control
sluices. The inlets of the sluices must be set low enoughdto drain the
reservoir to the required limits of drawdown (ER 1110-2-50 ). In a dam
for flood control only, the reservoir is normally dry and the sluice in-
let elevations are set at, or slightly above, the streambed with due con-
sideration of the sluice outlet elevation relative to stilling basin de-
sign. In a multipurpose dam with fixed reservoir storage allocations
and in which high reservoir stages may be maintained for long periods of

3-l
EM 1110-2-1602 3-3a
15 Oct 80

time, it may be desirable to have both high- and low-level sluices

(plate C-18). Low-level sluices are sometimes desirable for the passage
of sediment through a reservoir and for aiding in water quality control
if a special intake tower is not provided. If the sluice intake is
permanently or frequently submerged, the servicing and inspection neces-
sary for maintenance are more costly than for a high-level sluice. A
high-level sluice usually requires that the outlet portal be sloped to
direct the flow along the face of an agee spillway section or into a
stilling basin. The invert may slope on a straight line from the intake
to the outlet portal, or curve downward at some point downstream from
the intake. Setting the outlet portal at a lower elevation than the
intake reduces the pressure at critical locations such as the intake,
gate slots, and bends. An area reduction is usually provided in the
vicinity of the outlet portal of sluices to assure positive pressures
in these sluices when operated under full gate openings, or the sluice
is enlarged downstream of the gate to ensure open-channel flow at full
gate openings. Area reductions may be used to spread the emerging jet.

b. Bends. Flow around conduit bends results in acceleration of

flow along the inside of the bend accompanied by a local pressure reduc-
tion and the potential for cavitation (particularly for short-radius
bends). Cavitation is not likely to occur in bends where long-radius
curves are used. Pressure drop coefficients to evaluate cavitation
potential for 90-deg bends are given in plate C-20. The minimum pres-
sure occurs at 22.5 deg and 45 deg from the beginning of curvature for
circular and rectangular conduits, respectively. Since the computed
minimum pressure is an average pressure, the guidance given in para-
graph 2-16 should be adhered to.

Section II. Sluice Intakes

3-4. General. Sluice intakes are integral parts of concrete spillways,

and are usually rectangular in shape and flared in four directions. The
curved entrance is followed by the sluice passage, normally having a
height-width ratio of about 1.5:1 to 2:1. In some cases considerable
economy in stop log costs can be effected by projecting the intake
curves upstream beyond the face of the dam. This permits a reduction
in the required size of the stop log or bulkhead gate. Bulkhead slots
must extend vertically above the maximum reservoir pool or be provided
with slot covers. Open roof slots for closure bulkheads at Kinzua Dam
permitted flow through the slot and resulted in extensive cavitation
damage downstream (item 20). Plate C-21 shows typical designs for
flush and protruding sluice intakes.

3-5. Trash Protection. The intake may be equipped with struts or

3-2
3-5 EM 1110-2-1602
15 Oct 80
trashracks, depending upon the need for protection against clogging and
debris damage to gates and turbines.

a. Trash Struts. A simple trash str~t usually of reinforced con-

crete with clear horizontal and vertical openings not more than two-
thirds the gate or other constricted section width and height, respec-
tively, should be adequate for highly submerged flood control outlet
conduits. The purpose of such struts is to catch trees and other large
debris which may reach the entrance but would not pass through the gate
passage, thereby possibly preventing closure of the gates. Trash struts
should be located to effect local net-area velocities not greater than
15 fps. A flow net or model test should be used to determine local
velocities through this area (items 99, 101, and 135). The struts
should be circular cylinders or have rounded noses and square tails, ..
depending upon the structural design requirements and economy. Teardrop
designs are not required if the local velocity guidance is maintained.
Trash strut losses are usually included in the overall intake loss.
If necessary to consider separately, use of equation 2-12 is recommended
with a loss coefficient K value of 0.02. V in this equation is the
flow velocity in the uniform conduit section just inside the intake.
Trash struts should be provided with a working platform located above
conservation pool elevation to facilitate removal of debris. Additional
information on the design of trash struts is given in EM lll0-2-2400.j

b. Trashracks. Trashracks are provided where debris protection

for downstream devices such as valves or turbines is required (item 22).
Such racks are designed to retain debris of such size and type of mate-
rial that could result in damage to these devices. Because of danger
of overstressing from clogging, trashracks should be located in lower
velocity areas than trash struts and must be provided with raking or
cleaning facilities. They should be designed for safe operation with
50 percent clogging. Such devices can be fabricated from circular bars
and pipe. Trashracks should not be located in velocities exceeding
3 to 4 fps. Where additional strength is required, elongated sections
with rounded noses and tails can be used. Trashrack head losses depend
on the flow velocity and area construction (items 22, 39, 100, 108, and
135). The design of vibration-free trashracks is necessary to prevent
failure from material fatigue. It is especially important where reverse
flow can occur (items 21, 37, 53, 63, and 110).

3-6. Entrance Curves.

a. General. The curved converging section, which begins at the

upstream face of the dam or intake structure and terminates in tangency
to parallel walls, is commonly referred to as the entrance section.

3-3
EM lll0-2-1602 3-6a
15 Oct 80

The curves that determine the rate of convergence are designated as en-
trance curves. It is the function of the entrance section to guide the
flow with minimum disturbance until it is contracted to the dimensions
of the gate passage or to the upstream transition of an ungated intake.
If the entrance curve is too sharp or too short, negative pressure
areas may develop in the entrance section where the jet is inadequately
supported or improperly guided. On the other hand, a long and gradual
entrance curve may require an unnecessary amount of expensive forming.
The objective is to design an entrance of minimum length in which posi-
tive pressures can be maintained at all flows.

b. Circular Inlets. A bell-mouthed entrance, which conforms to

or encroaches very slightly into the free jet profile of a circular
orifice, eliminates occurrence of negative pressure in localized areas
at the entrance to a circular conduit (seep 414 of item 101). An
elliptical entrance curve for a circular conduit will satisfy the re-
quired streamlining and jet contact requirements if the curve is ex-
pressed by the following equation:

(3-1)

where X and Y are coordinates measured parallel to and perpendicular

to the conduit center line, respectively, and D is the diameter in
feet.

c. Noncircular Inlets. The sluices of a concrete dam are commonly

rectangular in cross section. WES (item 128) has tested entrance curves
of various shapes. A laboratory-tested elliptical curve is shown in ·
figure a, plate C-22, with the pressure drop coefficients. This simple
ellipse is normally satisfactory. For designs of high-head dams and
when the conduit has insufficient length to produce substantial back
pressure, the compound elliptical curve (fig. b, plate C-22) should be
used. HDC 211-l/2n shows the effect of upstream face slope of the dam
on the entrance curve pressures.

3-7. Intake Energy Losses. Intake head losses are considered to in-
clude all the energy losses between the reservoir and the sluice proper.
The head loss includes the form losses generated by the entrance curves,
bulkhead or stop log slots, gate passage and gate slot, air vents, and
the transition between the intake and the sluice proper. They also in-
clude the friction losses occurring in the intake structure. Intake
losses are experimentally determined (model and prototype) by assuming

3-4
3-7 EM 1110-2-1602
15 Oct 80
that the fUlly developed turbulent friction gradient exists between the
conduit exit portal and the intake as shown in plate C-2. On the basis
of limited model and prototype intake loss data for sluices, an intake
loss coefficient value of 0.16 is recommended for capacity design and a
value of 0.10 when high velocity is critical. When gate slot losses
are not included in the intake loss, a value of 0.01 for each gate may
be considered. If trashracks are provided this value should be in-
creased in accordance with data referenced in paragraph 3-5b.

Section III. Gate Passage, Gates, and Valves

3-8. General. The gate passage may be defined as the passageway in

which the gate leaves operate. The hydraulic design problems of the
gate passage are often closely associated with the structural and
mechanical problems in the design of the gate, gate frames, and gate
hoist. One of the most important problems in design of gates and
appurtenant features is to eliminate cavitation. A basic condition is
whether the gate will be required to operate partially open or will
only be operated fully open. When high-head gates are operated under
partial opening, they may be subject to severe cavitation and vibration
and have a high air demand. When valves are used for regulation they
are commonly.placed at or near the downstream end of the outlet conduits.
This location permits the valves to discharge freely into the atmosphere
and eliminates most of the cavitation potential. In some cases, however,
the spray so produced may be troublesome to power plants and switch-
yards. Gate passages of circular cross section are designed when nec-
essary to accommodate circular gates or valves, such as knife or ring-
follower gates or butterfly, fixed cone, or needle valves. Rectangular
gate passages are used for ordinary slide, tainter, and tractor or
wheel-type gates.

3-9. Gate Types.

a. Vertical Lift. Vertical-lift gates for outlet works are de-

fined according to their method of movement. Due to the friction be-
tween the gate and the vertical guides, slide gates are generally
operated by hydraulic cylinders. Tractor and fixed-wheel gates are used
where closure of large openings is required. Tractor gates move on an
endless chain of rollers on each side of the gate. Fixed-wheel gates
have a series of wheels down each side of the gate which bear on verti-
cal guides in the gate slots. Vertical-lift gates are operated either
by cables or a rigid stem connection to the hoist mechanism. Cable-
suspended gates operate in open wet wells which fill to the reservoir
pool elevation when the gate is closed; therefore, the hoist mechanism
is located at an elevation above the maximum pool level. This type of

3-5
EM 1110-2-1602 3-9a
15 Oct 80

operation is not usually used for gates which operate partly open for
long periods of time because of possible vibration. See paragraphs
4-18 and 4-19 for design problems concerning cable-suspended gates.
Hydraulically operated gates are preferred for high heads and for long
periods of operation at partial openings. These gates have rigid
riser stems that recess into bonnets or extend to a higher floor level
where the hydraulic hoist mechanism is located. The hydraulically
operated slide gate is used preponderantly in designs for service gate
installations in sluices of concrete dams. The rectangular slide gate
generally has a height greater than the width to minimize both the
flexure on the horizontal members and the unit loads on the vertical
guides, and to reduce the possibility of binding in the slot. The
cross-sectional shape of the gate passage in the sluice is usually the
same as the shape of the gate. The upstream face of vertical-lift
type gates must be flat rather than "bellied," as some gates were in
the past, and the 45-deg lip should terminate in a l-in. vertical
extension (see plate C-23). Rating curve computations are discussed
in paragraph 4-16 and in Appendix D.

b. Tainter Gates. Tainter gates have been used in the Pacific

Northwest as service gates in sluices operating under extremely high
heads (>250 ft). The characteristics of the tainter gate are favorable
to its use for accurate reservoir regulation in both concrete and
embankment dams. Advantages of the tainter gate over the vertical-lift
type gate include: gate slots are not required in the walls of the
gate passage, which is favorable in partly open gate operation; a rela-
tively small hoist capacity is required because the direction of the
resultant water load is through the trunnions; and the friction between
the gate seals and the gate passage walls is low. A disadvantage of
the tainter gate is that the entire gate cannot be easily lifted out
of the well for maintenance. Tainter gates are placed in an enlarged
section of the sluice and some have eccentric trunnions to facilitate
movement and sealing under a very high head. The enlarged gate section
may include an invert step-down as well as side and roof offsets to
provide for complete sealing and ~or aeration of the jet which most fre-
quently discharges as open-channel flow downstream at full gate opening.
Under this condition, back ' pressure in the intake section is essentially
nonexistent and the boundary layer is not fully developed. A model
study is usually required to resolve pressure and vibration problems
in pressure flow conduit designs. Discharge coefficients of a partially
opened tainter gate in a rectangular conduit are shown in plate C-24.
In general, the discharge coefficient can be considered the same as the
contraction coefficient based on a study of the jet profile (HDC 320-3n).

3-6

I:'
v
3-10 EM 1110-2-1602
15 Oct 80

3-10. Control Valves.

a. Valve Hydraulics. Knife gate, needle-type, fixed-cone, and

various commercial valves have been used for flow control. Discharge
rating curves for a valve discharging freely into air or into an en-
larged, well-vented conduit can be developed from the equation

Q = CAI2gH (3-2)

where

Q = discharge in cfs

C = discharge coefficient
2
A = nominal conduit or valve flow area in rt

H = energyhead immediately upstream and generally measured from

the center line of the conduit in feet of water
2
g = acceleration due to gravity in ft/sec

Discharge coefficients for freely discharging valves of many types

have been determined empirically and will be presented in subsequent
discussions on specific valve types. Head loss across in-line valves
in pressure conduits can be computed by equation 2-12 using the dimen-
sionless valve-loss coefficient K determined experimentally for the
particular valve and valve opening.

b. Butterfly Valves. Butterfly valves have been used extensively

for cutoff valves but are not recommended for flow regulation. There
is evidence that the butterfly valves in the 11-ft-diam flood control
conduits at Summersville Dam may have contributed to the failure of
the 9-ft-diam fixed-cone valves immediately downstream (item 80).

c. Needle-Type Valves. The needle valve opens and closes by the

horizontal movement of a needle; the valve is closed when the needle
is advanced to its extreme downstream position. The water flows in an
annular passageway first diverging and then converging past the needle.
Discharge from needle valves can be computed using equation 3-2, where
A and H are the area and energy head, respectively, at the inlet end,
and C is a discharge coefficient. Kohler and Ball (in Davis and
Sorensen, item 24) show the full open coefficient to be about 0.60 when

3-7
EM 1110-2-1602 3-lOc
15 Oct 80

the ratio of outlet diameter to inlet diameter is 0.95. Thomas (item

120) gives discharge coefficients for partly open 86-in. needle valves.
The hollow-jet valve is a modification of the needle and the needle
moves upstream to close the outer casing of the valve. Model tests
of the hollow-jet valve for Anderson Ranch Dam showed fully open dis-
charge coefficients of approximately 0.70. Thomas also presents dis-
charge coefficients for partly open valves in item 120. Nag presents
a good summary of the characteristics, the uses, and the limitations
of free discharge regulating valves in item 78.

d. Fixed-Cone Valves. The fixed-cone valve is similar in princi-

ple to the hollow-jet valve except that the cone pointing upstream on
the downstream end is stationary and a sleeve of the outer casing moves
downstream to close the valve. The shape of the issuing jet is a
hollow cone. The discharge coefficient curves for fixed-cone valves
are shown in plate C-25. The coefficients for the six-vane valve are
based on tests by TVA (item 29). A comparable coefficient curve for a
four-vane valve reproduced from HDC 332-ln is also shown in this plate.
Model-prototype confirmation of the hydraulic characteristics of these
valves has been studied by Lancaster (item 58). The shell of a six-
vane valve has been found to be less likely to vibrate than that of a
four-vane valve. In a number of cases, flow-induced vibration of
fixed-cone valves has resulted in serious and costly damage (items 71
and 80). Hoods can be designed for these valves to control the spray
of the jet (items 31 and 81).

e. Commercial Valves. Many types of commercially available valves

are available for small conduits and water-supply systems. Some of
those most commonly used are the knife gate and other gate valves. Head
loss coefficients for lenticular- and crescent-shaped opening, in-line
gate valves are given in HDC 330-l.n KnLie gate valves are recommended
for free discharge installations.

3-11. Metering Devices. Where accurate monitoring of outflow is re-

quired the inclusion of a metering device in the system should be con-
sidered. Many schemes can be considered, varying from venturi and
elbow meters to acoustic and electronic systems. The installation of
such devices eliminates the need for extensive calibration of gates and
valves under varying operating conditions and generally results in flow
measurements with errors from about ~5 percent to +1 percent. It is nec-
essary that all flow measuring devices of these types be installed
according to standard specifications for proper, cavitation-free opera-
tion. If the provision of metering equipment is contemplated, WES
should be consulted relative to available types and to their installa-
tion and operation requirements and limitations.

3-8
3-12 EM 1110-2-1602
15 Oct 80

3-12. Gate Passageway Requirements. Normally, when reservoir outlet

flows require regulation the following are provided:

a. Two or more gate passages such that if one passage is inopera-

tive, a reasonable flow regulation as pertains to project purposes is
obtained.

b. Emergency gate provision (tandem or transferable) for each

service gate passage so that if a service gate is inoperative in any
position, closure of the gate passage can be made with the emergency
gate for any pool level.

c. Bulkhead prov1s1ons for each gate passage for inspection and

maintenance of the service and emergency gates. As a minimum, the
bulkheads must be capable of being installed at the lowest pool eleva-
tion that has a reasonable frequency and length of occurrence sufficient
for inspection and repair purposes. All judgment factors involved in
the above should be fully discussed in the design memorandum
presentation.

3-13. Gate Slots. The guide slots of rectangular gates produce a dis-
continuity in sidewalls which may cause cavitation, unless specially
designed. It has been common practice to use metal-liner plates down-
stream from the gate slots to protect the concrete from the erosive
action of cavitation. The recommended guide lines for metal liners are
given in paragraph 3-16. The gate slot in the roof of the gate chamber
and air vent slots present similar design problems. Design details for
slide gate roof, side, and air vent slot details are shown in
plate C-23. Pressure coefficients (item 123) for detailed examination
of this gate slot design for high heads (>250 ft) are given in figure a,
plate C-26. To obtain dimensional local gate slot pressure data, the
pressure coefficients given in this plate are multiplied by the flow
velocity head in the gate passage and algebraically added to the back-
pressure gradient elevation at the gate slot. Tests by Ball (item 6)
show that doubling the downstream taper length from 12 to 24 units re-
duces the severest pressure drop coefficients (C) from -0.16 to -0~.12
for comparable slot geometry. Therefore, it is recommended that for
heads >250 ft the taper downstream of the gate slot be modified to
1:24. For conservative estimates of minimum pressures at gate slots
where streamlining is not provided, the pressure coefficients in fig-
ure b, plate C-26, should be used. In detailed design studies it may be
desirable to check the gate slot design for potential incipient cavita-
tion. This can be done by solving equation 2-19 for the absolute con-
duit pressure p necessary for cavitation and comparing it with the
0
computed minimum pressure at the slots. Plate C-27 gives incipient

3-9
EM 1110-2-1602 3-13
15 Oct 80
cavitation coefficients a. for various slot geometries. These values
were obtained using relati~ely large scale (1:3) plastic models to re-
duce possible errors from scale effects. A a. value of 0.4 is recom-
mended to check cavitation potential. For con§ervative design, the
computed minimum pressure should be appreciably higher (15 ft or more)
than the incipient cavitation pressure. The head losses for gate slots
are generally included in the composite intake loss discussed in para-
graph 3-7. When gate slot losses are not included in the intake loss,
a loss coefficient K value of 0.01 is recommended for each pair of
gate slots for use in equation 2-12.

3-14. Gate Recess. Hydraulically operated control gates recess into

bonnets and cable-suspended gates into wet wells. The necessary dimen-
sional clearances for gate operation are usually based on mechanical
and structural requirements rather than hydraulic. The primary hy-
draulic consideration is the relative upstream and downstream clearance
at the roof recess when the gate passage is operated at part gate open-
ing. The upstream clearance at the roof should be appreciably larger
than the downstream clearance to assure maintenance of a hydrostatic
head in the well or bonnet for gate stability. If the downstream
clearance exceeds the upstream clearance the gate well can be sucked
dry and the gate may float or catapult or oscillate under certain
operating conditions (see para 4-18b).

3-15. Gate Seats. In general, the gate seat is flush with the floor
of the gate passage.

3~16. Steel Liners. Steel liners in concrete conduits have been used
where experience indicates cavitation is likely to occur such as down-
stream from control gates and valves where a high-velocity jet occurs.
For heads above 150 ft, a metal liner should extend 5 ft downstream
from the gate. For heads below 150 ft, no liner should be required.
If a liner is necessary, it should not terminate at a monolith joint
or in a transition.

3-17. Air Vents. The following guidance is recommended for air vent
design:

a. Control valves and gates that are located a considerable dis-

tance upstream from the exit (i.e., do not discharge into the atmosphere)
require air vents. An air vent is required for each service gate. Air
vents are not required for emergency gates when those gates are lo-
cated immediately upstream of air-vented service gates. Extreme caution
must be observed if the emergency gate is used for regulation. Air
demand will create very low pressures in the service gate recess. The

3-10