October 20～23 2008, Hohai University Nanjing, China 1

DESIGN CONDITIONS FOR MORNING-GLORY SPILLWAYS:

APPLICATION TO POTRERILLOS DAM SPILLWAY

Claudio A. Fattor, Jorge D. Bacchiega

National Institute of Water, Casilla Correo 21 – B1802WAA-Ezeiza – Argentina, e-mail: [email protected]

Abstract. Design of morning-glory spillways has special features, because the hydraulic behavior
of this structure is strongly related to the boundary topography, crest diameter, curve or bend
downstream, spillway profile and the tunnel dimensions towards the outlet section. In the case of
Potrerillos dam, located on Mendoza River, Argentina, a morning-glory spillway was selected. This
spillway must be able to pass a design discharge of 1800 m3/s at free surface, being characterized
by a 31,76 m-crest diameter, 13.40 m-cross section diameter, a vertical 90° curve of 44.67 m-radio
and a 400 m-length horseshoe cross section tunnel of 11.80 m-diameter. A special study was
developed at National Institute of Water in order to analyze the hydraulic performance of the
spillway. The physical model studies allowed determining a final location different from that
considered in the project, optimizing the boundary conditions close to the spillways, modifying the
number of piles on the crest and improving by these means the hydraulic performance of the whole
structure, including flow characteristics inside tunnel. Final hydraulic design was compatible with
structural design, allowing besides that a strong reduction of rock excavation.

Key words Morning-glory spillways; Potrerillos dam; physical model.

1 INTRODUCTION
Ra
Morning-Glory spillways must be able of De
Dv
Hd

discharging floods related to a determined return

period under safe conditions for the dam. This kind Dg

y= f(x)

of spillways can be characterized by its spillway

profile, a vertical curve and a discharge tunnel.
R

Main dimensions, as it can be observed in Figure 1,

Dc

include radius at crest elevation (R), spillway

Dt

profile, the radius of the downstream edge of the

profile (Rt), curve radius (Rc), angle of the vertical Figure 1 – Morning-glory spillway
bend, diameter D and the longitudinal slope (i).
The discharge capacity of the structure is The criteria mostly extended for this kind of
governed by the hydraulic regime, which can spillway is that must be designed for working in
present three different phases, as it can be seen in phase 1, it means at free surface flow, with the
Figure 2. discharge as a function of the discharge coefficient
 16th IAHR-APD & 3rd IAHR-ISHS

at the inflow that could anticipate the submergence

and the drowning of the spillway. If this were to
Reservoir elevation

Phase III - Q=f(J) happen, the efficiency is lower and the water
surface elevation would be increased.
On the other hand, inflow conditions are
Phase II - Q=f(H1/2) potentially able to influence on the flow inside the
tunnel, where high flow velocities and cavitation
Phase I - Q=f(H3/2)
Change of tendency: fromspillway risks would be present. This conducts to try to get
regime to orifice regime
approximation flow conditions as symmetric as
Discharge
possible, which may reduce presence of waves and
instabilities in a supercritical flow regime.
Figure 2 – Discharge-elevation relationship
Flow rotation at the inlet section of the
Cd, the effective length L and H3/2, where H is the spillway and vortexes are reduced by means of piles
hydraulic head. Following these design criteria, located at the crest. However, the singular boundary
instabilities caused by submergence are avoided conditions involving the spillway geometry and the
(Fattor C. and Bacchiega J., 2001). topography in the immediate surroundings can be a
If maximum flow discharge were exceeded, for source of instabilities inside the spillways and along
instance due to the effect of river sedimentation in the discharge tunnel.
the routing of the flood, the hydraulic behavior As a consequence, the design of morning-
could be given by a total or partial submergence of glory spillways must make compatible the
the morning-glory spillway. These conditions optimization of the discharge capacity and the flow
should be carefully analyzed by means of a physical pattern, avoiding negative effects over its
model due to strong instabilities that could be performance and the presence of asymmetric flows
occasioned, being able to reach water surface along the tunnel.
elevations with some risks for the dam safety.
The elevation-discharge relationship can be 2. BACKGROUND ANALYSIS
expressed by (1):

Q = Cd ⋅ (2 ⋅ π ⋅ R ) ⋅ H
3
2
(1) Technical bibliography concerning morning-
glory spillways is not so abundant, but despite of
, where: Cd: discharge coefficient, R is crest- that, there are references that allow analyzing main
diameter and H is the hydraulic head. issues of hydraulic behavior. For instance, the
Cd depends on Hd/R. For Hd/R<0.50, Cd has guidelines described at Hydraulic Design Criteria
similar values to a conventional spillway, but the are based on laboratory experiments carried out by
decrease strongly for higher values of Hd/R, Wagner (1956) and other researchers.
reaching values as low as 0.50 for Hd/R≈2. Bearing in mind the purpose of making
Due to usual locations of these spillways, at compatible the free discharge of the spillway with a
short distances from hills or dams, local boundary hydraulic behavior without rotation or with a
conditions could strongly affect the flow conditions substantial reduction of circulation flows in the
and its desirable symmetry. Under these adduction region and irregular flows inside the
circumstances there would be tangential tunnel, an analysis of antecedents, which is resumed
components of the velocity, inducing flow rotation in Table 1 was carried out.
 October 20～23 2008, Hohai University Nanjing, China 1

Hmáx Dg Dc Rc Cd
Dam Country Year Dv (m) Qd (m3/s) Piers P (m)
(m) (m) (m) (m) (m1/2/s)
Akongtien Taiwan 1951 18.29 3.51 92.0 2.99 2.99 4 0.244
Ben Metir Tunisia 10.67 14.93 623.0 6 0.322
Burnhope England 15.24 0.82 73.6 3.66 3.66 4.88 - 1.22 2.071
Campilhas Portugal 7.70 2.9 210.0 1.758
D'Aït Aadel Morocco 24.90 1500.0
Davis Bridge USA 1926 24.38 2.44 764.6 6.86 6.55 16.76 16 2.619
Eagle Mountain USA 1934 17.07 623.0 12
Ebenezer South Africa 1957 19.20 3.2 612.0 4 1.771
Gibson USA 1930 28.04 6.1 1415.9 8.99 10.97 17.98 8 1.83 1.067
Heart Butte USA 1949 8.22 16.46 154.3 4.27 4.27 6.4 6 0.089
Hungry Horse USA 19.51 5.15 1500.8 1 2.095
Jubilee Hong Kong 22.56 2.86 481.4 4.57 4.57 9.14 - 1.404
La Regadera Colombia 1935 16.76 2.74 424.8 11.12 6.19 8 1.779
Lady Bower England 1944 24.38 1.98 283.2 4.57 4.57 10.67 12 1.327
Manuherika Falls New Zeland 1935 31.08 1.83 424.8 5.18 5.18 6 1.757
Maranhao Portugal 3.9 1600.0
Owyhee USA 1932 18.29 3.66 849.3 3.89 3.89 15.24 1 2.111
Pegodo do Altar Portugal 3.5 1000.0 4
Pontial Ketchil Singapur 1931 15.24 0.82 76.5 15 0.76 2.151
Potrerillos Argentina 2000 31.76 4.4 1800.0 13.4 11.8 51.37 4 5.00 1.955
Procana Portugal 1952 19.62 1650.0 9.5 20 6
Robert Lee USA 20.30 12
Shade Hill USA 1950 9.97 12.19 141.6 4.72 3.81 4.27 6 0.106
Silent Valley Irland 24.38 0.7 73.6 4.88 4.88 5.49 4 1.52 1.641
Siya Rhodesia 15.00 350
South Holston USA 39.01 3.78 1699.0 6 1.886
Taf Fechan Gales 1927 20.11 0.85 85.0 4.88 3.96 4.57 4 2.00 1.716
Watanga USA 39.01 3.78 1699.0 6 1.886

Table 1 – List of some morning-glory spillways

Cd values informed in previous table were Besides, height and upstream slope of the
calculated with expression (1) without considering spillway adds another parameter. Figure 5 presents
the effect of piles. Table 1 show some structures as the variation of Hs/Hd as a function of Hd/R for
those of Akongtein, Ben Metir, Heart Butte and
Shade Hill, with very low values of discharge
coefficients, being supposed that hydraulic control
is not performed like a spillway, with incidence of
orifice control or frictional control. This situation
was avoided in the design of Potrerillos spillway.
In relation to the characteristics of the area
close to the spillway structure, Figures 3 and 4
show configuration for those particular dams. Figure 3 – Dillon dam spillway
 16th IAHR-APD & 3rd IAHR-ISHS

different relations P/R. Discharge coefficients are The morning-glory spillway, whose design
not so sensitive to small changes of P/R for values discharge is 1800 m3/s, has 31.76 m-crest diameter,
of Hd/R that are almost constant (USBR,1970). located in a sector of the hill that was specially
excavated close to the right abutment of the dam.
The vertical parameter of the spillway was
projected as a 5 m-high wall. Downstream of the
crest, the spillway develops a curve profile until
reaching 13.40 m-diameter. Beyond this section, a
44.67 m-radius and 90° vertical curve is extended
until the section where a transition to an 11.80 m-
diameter horseshoe section begins. The discharge
tunnel keeps this last cross-section in a length of
400 m, reaching a sky-jump at the outlet. Figures 6
and 7 present main dimensions of the spillway.
Figure 4 – Owyhee dam spillway

Figure 6 – Potrerillos dam spillway: layout

Figure 5 – Hs/Hd vs. Hd/R relationship

Finally, the location of piles as anti-vortexes

elements has been extensively applied, though the
number of these piles can vary depending on the
spillway diameter and boundary conditions.

3. POTRERILLOS SPILLWAY

Potrerillos dam is located in Mendoza River,

30 km upstream of the homonymous city, in the
western region of Argentina. It is a 116 m-high and
395 m-long CFRD dam, whose purposes are flood Figure 7 – Potrerillos dam spillway: cross section

control, irrigation and hydroelectricity.

 October 20～23 2008, Hohai University Nanjing, China 3

4. POTRERILLOS SPILLWAY: These rotation was induced for the asymmetric

PHYSICAL MODELING location of the spillway respect to the surrounding
terrain, inducing flow velocities on the right side
4.1. PURPOSES that were almost three times the flow velocities on
the right side, the closest to the dam, with the
The design discharge is 1800 m3/s, being
superficial flow pattern observed in Figure 8.
necessary to verify free discharge conditions until
2250 m3/s, being this flow related to peak discharge
4.3. ALTERNATIVES
for a reservoir having lost 25% of the volume
because reservoir sedimentation at a estimated Some alternatives were proposed to improve
period of 50 years. The purposes of the study by the hydraulic behavior of the spillway. These
means of a physical model are the analysis of water measures include its relocation in order to correct
surface elevation-discharge relationship, the study the asymmetry of velocities, the modification of the
of flow pattern upstream of the spillway, the flow approximation area (excavated in the hill) to
functioning of the vertical curve and the discharge contribute to improve flow velocity distributions on
tunnel, as well as the analysis of local scour the perimeter of the spillway, the increase of the
downstream of the flip bucket, among other issues. height of the vertical upstream face of the spillway
Taking into account these reasons, a 1:33.71 length and the redistribution of anti-vortexes piles.
scale physical model was built. This model One of the most important changes
reproduced a part of the adjacent reservoir, the introduced to the project was related to the layout of
excavated volume where the spillway was located, the surrounding area to the spillway, as it can be
the geometry of the spillway, the vertical curve and seen in Figures 9 and 10 (Bacchiega et al, 2000).
around 200 m-long of the discharge tunnel.

4.2. ORIGINAL DESIGN: RESULTS

Although free discharge was observed until
flows exceeding 2250 m3/s, the approximation flow
to the spillway was characterized by strong
circulations that had influence on the hydraulic
regime in the curve and the tunnel (Figure 8).

Figure 9 – Layout of the original project

The changes of the position and the shape were

followed by increasing the height of the vertical
upstream face to go from 5 m-high to 6.6 m-high at
the toe of the spillway and 10.0 m-high respect to
the excavated channel around the spillway. Besides,
only the pile located close to the hill and aligned to
Figure 8 – Flow pattern for Q=1800 m3/s the longitudinal axis of the tunnel was maintained.
 16th IAHR-APD & 3rd IAHR-ISHS

5. CONCLUSIONS

A general analysis of different morning-glory

spillways was made, with emphasis in conditions of
discharge with the section control at the crest,
without influence of orifice or frictional control.
Potrerillos dam morning-glory spillway was
analyzed by means of a 1:33.71 length scale
physical model, where some problems to be solved
were found for the original design.
Figure 10 – Layout of the selected alternative The main measures implemented to allow a
substantial improvement of the hydraulic behavior
Figure 11 allows observing the water
were the displacement of the spillway location, the
elevation-discharge relation of the selected and
modification of the approximation physical
original project, being able to see the best
boundaries, changes in the number and location of
performance of the final alternative and a spillway
anti-vortexes piles and more excavation to deepen
control to flow discharges higher than 2250 m3/s.
the area close to the upstream face of the spillway.
Finally, Figure 12 shows an image of the prototype.
The changes allowed increasing discharge
Water surface elevation vs. Flow discharge capacity, reducing circulation currents immediately
upstream of the inlet section and ensuring the
1384.3
Original design
1383.3

discharge of 2250 m3/s without influence by

Final design
Water Surface Elevation (m)

1382.3

1381.3 submergence or drowning.

1380.3 The criteria followed for the optimization of
3
Q = 2250 m /s
1379.3
3
Q = 1800 m /s
Potrerillos spillway can be taken like basic
1378.3
guidelines, being necessary to consider singular
1377.3
0.0 500.0 1000.0 1500.0
Flow discharge (m3/s)
2000.0 2500.0
conditions for any singular project.

Figure 11 – Water elevation – discharge relationship REFERENCES

Bacchiega J., Fattor C., Lopardo M. and Barchilón M.

(2000). Flow pattern conditions upstream of Morning-
Glory spillways (in spanish). XIX Congreso
Latinoamericano de Hidráulica, Córdoba, Argentina.

Fattor C. and Bacchiega J. (2001). Analysis of instabilities in

the change of regime in morning-glory spillways.
XXVIII IAHR Congress, Beijing, China.

United States Department of the Interior, Bureau of

Reclamation (1970). Design of Small Dams (in
spanish). Editorial Dossat, Madrid, España.

Wagner W. (1956) Determination of pressure-controlled

Figure 12 – Morning-glory spillway: prototype profiles, American Society of Civil Engineers, Paper
No.2802, pp.345-384.