Chute Spillway Aerators - McPhee Dam

Model/Prototype Comparison

K. Warren Frizell 1
Clifford A. Pugh, M. ASCE 2

Abstract

In 1983 a decision was made to include aeration in the chute spillway

at McPhee Dam in southwest Colorado to prevent possible cavitation
damage to the chute floor near the stilling basin. Model studies
were conducted to develop the design of the aerator in 1984 and field
tests were conducted to compare to model measurements in 1987. This
paper describes the measurements taken and compares model and prototype
results.

Introduction

McPhee is the main storage and regulation reservoir in the Dolores

Project in southwest Colorado. The dam is an earthfill structure
82.3 m high and 396 m wide at the crest. The chute spillway and
stilling basin are located in the right abutment of the dam. The
chute is 18.3 m wide, 303 m long and drops 90 m in elevation from
the maximum water surface to the stilling basin. The stilling basin
is a combined hydraulic jump/flip bucket energy dissipator. At a
flow of about 425 m3/s the jump will wash out of the stilling basin
and the flow will flip into a plunge pool downstream from the basin.
Figure 1 is a section through the spillway.

A model study was conducted before the dam was constructed to develop
the design of the spillway [Pugh, 1981]. This study included the
approach channel, the chute, the stilling basin, and the exit channel
in a 1:36 scale model. When the spillway was completed, the surface
tolerances required to prevent cavitation damage during high releases
were not obtained. A decision was made to include aeration to prevent
cavitation damage and to minimize future maintenance costs associated
with maintaining close surface tolerances. A prototype test was run
upon completion of the spillway modifications.

Model Study

Since the 1:36 scale model was still available, it was modified to
study the addition of aeration. Initially, two aeration devices were
located at stations 13+99 and 15+94. Figure 2 shows the model operating
with two aerators and the prototype operating at 142 m3/s.

1 Hydraulic Research Engineer, Bureau of Reclamation, Denver, CO

2 Head, Hydraulic Equipment Section, Bureau of Reclamation, D-1532,
P.O. Box 25007, Denver, CO 80227

Frizell, Pugh
Subsequently, additional analysis of the spillway indicated that one
aeration device located at station 15+29 would provide adequate aeration
to prevent cavitation damage.

The aeration device location was determined by the cavitation potential

and the geometry of the spillway. The cavitation potential is deter-
mined by analysis of the flow with a computer program [Falvey, 1989].
The program indicated that the major cavitation potential is in the
lower part of the chute where the cavitation index drops as low as
0.135. The aeration device was located where the minimum cavitation
index is 0.195, this location is above the point where damage would
be expected. The ramp angle was chosen to throw the jet onto the
last vertical curve in the spillway to minimize impact pressures and
splash. This design allowed 1.52 m of freeboard on the chute walls
for the maximum flow depth of 1.66 m, to allow for bulking. The aera-
tion consists of a ramp at a 6.4° angle with the chute floor. The
ramp, at station 15+29, is 0.91 m high and 7.8 m long. A ramp without
a slot in the chute floor was chosen since there was adequate freeboard
available and a slot would require cutting into the chute floor, thus
destroying the continuity of the reinforcing steel. Air vent openings
(0.91 m by 1.22 m) were cut into the side walls at the downstream
end of the ramp, and towers (1.22 m by 1.22 m) were placed on the
outside of the walls to provide the air. Figure 3 shows the configura-
tion of the aeration device. The exact location of the ramp was deter-
mined by the location of the structural panel in the approximate loca-
tion desired. The entire ramp, ports, and towers were contained within
one panel to facilitate construction.

Model Measurements

Measurements taken in the model included:

(1) Piezometric pressure profiles upstream and downstream from

the ramp and on the downstream side of the ramp.

(2) Cavity length under the jet at the sides and the center of
the chute.

(3) Flow profiles along the side walls.

(4) Air demand through the air vents.

(5) Pictures and video tape at various flows.

Previous studies [Pinto, 1982] have shown that a model scale of 1:10
to 1:15 is needed to overcome surface tension and viscous effects
which limit air entrainment in a scale model. Pinto found that model
air demand was considerably less than prototype measurements. There-
fore, additional turbulence was induced (as suggested by Pinto) in
the 1:36 scale McPhee model by placing 16-gauge wire mesh screen (about
3 mm per square) - 150-mm long - on the ramp. The screen ended 50 mm
upstream from the end of the ramp. This screen was large enough to
increase the turbulence on the bottom of the flow nappe, yet not large
enough to significantly increase the flow depth. This increased the

2 Frizell, Pugh
air demand in the model by 40 to 100 percent over the air demand without
induced turbulence (fig. 4). The jet trajectory length was also influ-
enced by the induced turbulence. With the increased turbulence, the
cavity length downstream from the ramp was reduced by 30 to 40 percent
in the model (fig. 5). Figure 6 shows the aerator operating at 940 m3/s
(note the turbulent flow entraining air on both the upper and lower
surfaces of the jet). The model results are compared to prototype
measurements in a subsequent section of this paper.

Prototype Tests

Field tests of the aeration ramp on the McPhee Dam spillway were per-
formed in May 1987. The tests, which were designed to verify operation
of the spillway and to evaluate the effectiveness of the aeration
ramp, were limited to a maximum discharge of 142 m3/s due to downstream
channel capacity. The observations and data collected were important
in the verification of hydraulic and numerical model results, even
though the maximum discharge tested was only about 15 percent of the
maximum design capacity.

The tests consisted of measurements and observations at three spillway

discharges, 28 m3/s, 71 m3/s, and 142 m3/s (fig. 7). Quantities which
were measured included air demand through the vents and pressure dis-
tribution on the downstream side of the ramp beneath the jet. Observa-
tions of the jet trajectory length and details about the free surface
aeration on the chute were recorded for each flow.

Measurements were accomplished with electronic instrumentation and

data acquisition equipment. Due to the remote location, power was
provided with a portable generator.

Air demand through each vent was measured by placing an orifice plate
over the duct entrance and measuring the pressure drop across the
plate with a differential pressure transducer. The orifice and duct
configuration were modeled just prior to the test at a scale of 1:7.5
at the Bureau of Reclamation's hydraulic laboratory in an air test
facility. With given tap locations, the coefficient of discharge,
CD for the orifice and duct combination was determined. A scaled
ladder was included in the model and its effect is shown in figure 8.

The pressure distribution beneath the jet was measured with five static
pressure transducers, mounted flush on an aluminum plate which was
secured in the chute, figure 9. The flow observations were aided
by the installation of a staff gauge mounted on the chute wall at
the downstream end of the ramp, and also by lines painted on the chute
walls every 3.05 m downstream from the ramp.

Each of the three test discharges were maintained for about 1 hour.
The readings from the electronic instruments were taken with a computer
controlled A to D scanner and a magnetic tape recorder. Flow observa-
tions were noted and video and still photographs were used to further
document the test.

3 Frizell, Pugh
Model-Prototype Comparison

Common measurements between the model and the prototype included:

(1)Air demand

(2)Jet trajectory length

(3)Air pressure downstream from the ramp

A comparison between model and prototype air demand is shown on

figure 4. The model air demand was increased considerably by increasing
the turbulence; however, the prototype air demand was still higher.
If the prototype measurements are projected to the design flow, the
maximum air demand would be f3 . 0.20, resulting in a maximum air
velocity in the vents of 63 m/s.

The jet trajectory length was close to the prototype observations

before turbulence was added to the model. After turbulence was induced
on the bottom of the nappe, the trajectory length was reduced in the
model due to additional losses at the air water interface and reduced
cavity pressures (fig. 5).

The pressure distribution under the nappe at the ramp is the driving
force for pulling air in through the vents. Thus, it is important
in any type of model (physical or mathematical) to reproduce the proper
pressure distribution.

The air pressure downstream from the ramp was higher in the model
than in the prototype. The pressure under the nappe was essentially
atmospheric before turbulence was induced. After the turbulence was
added, the model pressures near the side walls dropped to -0.11 m
of water at a flow of 142 m3/s. The pressure at the center of the
ramp was -0.03 m. The field measurements and model pressures are
shown on figure 10.

Conclusions

The use of chute spillway aerators has become increasingly popular

in the past few years. Numerous model studies have been performed
along with a few cases of prototype measurements. Although all the
details of how and why the aerators are successful in mitigating cavita-
tion damage are not yet understood, our ability to model the devices
is improving. Air demand for small models (scales less than 1:15)
can be brought closer to simulating prototype air demands by inducing
extra turbulence in the model. Turbulence measurements in both the
model and prototype are needed to determine the turbulence levels
in the model that would correctly scale prototype air demands. Jet
trajectory lengths and undernappe pressure distributions are also
closely tied to the air demand. The induced turbulence in the model
causes the undernappe pressures to more closely approach the prototype
values. However, the jet trajectory in the model is reduced by inducing
turbulence and moves further away from the prototype values.

Frizell, Pugh
 radial co/es
EL 6928 0

Crest. £1. 6097.00

alL
ow,
s .A0

Wd strudurt Chute Aeration ramp, Sta. 15+29

Avis of cres
570944 s* 11,1444%.%*
2-2970

* ,
S u•17."L _st a 19. 45
50
.
0 ;:602.00a66"°
400"Verticai carve 2„OVedico/
u v.e 4
lI0'-OVerIicoi curve

Stilt/n9 boon

Figure 1. - Section through the spillway (1 ft = m).

Figure 2. - Prototype and model operating at 142 m 3/s.

5 Frizell, Pugh
 Figure 3. - Configuration of aerator.

1. 00

0.80
PROTOTYPE DAT
mo RI n TA
„ MO EL DATA
INDJCED TURBU _ENCE
0.60

C12.

O. 40

0.20

0.00
0 200 1100 600 800 1000
SPILLWAY DISCHARGE (ciiis)

Figure 4. - Model and prototype air demand.

= Qairnwater)

6 Frizell, Pugh
Figure 5. - Cavity lengths model and prototype at 142 m 3/s.

Figure 6. - Model aerator at 940 m 3/s.

7 Frizell, Pugh
7)4
.4104"1"'im
7
 0.590
A AWITH' UT LADDER
0---DNITH LADDER

0.580

_
m :1
C) .., ,
c) 0.570 L,
...a
.L
a

,L

0.550
:00000 :25000 150000 :75000 200000 225000 ::50000
REYNOLDS NUMBER

Figure 8. - Discharge coefficient for orifice and duct configuration.

Figure 9. - Pressure transducer mounting at ramp.

9 Frizell, Pugh
 2.00
0 PRO TOTYPE DATA
n un P NTA

WAT E R)
MO pEL DATA
a IND JCED TURBU _ENCE
1.20

(m OF' 0.40

-2.00
0 3 6 9 12 15 18

SPILLWAY WIDTH (in)

Figure 10. - Model and prototype pressure distribution downstream

from ramp.

Bibliography

1. Falvey, H. T., "Cavitation in Hydraulic Structures," Engineering

Monograph No. 42, Bureau of Reclamation, 1989.

2. Pinto, N. L. de S., Neidert, S. H., "Model Prototype Conformity

in Aerated Spillway Flow," Cranfield, England, International Con-
ference on Hydraulic Modeling of Civil Engineering Structures,
BHRA, Paper E6, 1982, pp. 273-284.

3. Pugh, C. A., "Hydraulic Model Studies of McPhee Dam Spillway,"

Bureau of Reclamation, GR-81-2, Denver, Colorado, March 1981.

10 Frizell, Pugh