PROTOTYPE BEHAVIOR OF MORNING-GLORY SHAFT SPILLWAYS

By
J. N. Bradley
Engineer, Bureau of Reclamation
Design and Construction Division
Denver, Colorado

A paper for the

Summer Convention, ASCE,
Denver, Colorado, June 1952
 PROTOTYPE BEHAVIOR OF MORNING-GLORY SHAFT SPILLWAYS

J. N. Bradley, Assoc. M, ASCE 1/

This is the first of three talks on morning-glory shaft spillways.

The first lecture will be of a general nature, while the following two will
deal with specific problems The second talk will deal with the shaping of
the morning-glory portion of the spillway, and the third lecture will con-
sist of an account of experiences and observations at the Heart Butte and
Shadehill Dam spillways operating under flood conditions.

This talk will deal with an investigation of prototype behavior of

morning-glory shaft spillways. We will begin with a short discussion of
the general features of this type of spillway as an introduction to all three
talks,

The morning-glory shaft spillway is comprised of three essential

parts: a collecting structure or the morning-glory proper, a vertical or
inclined shaft, and a horizontal tunnel, In some cases, a stilling basin is
constructed at the downstream end of the horizontal tunnel, Usually, a
vertical shaft is used and this is connected to the horizontal tunnel through
a rather sharp 90 0 bend. In fact, in only two cases on record have inclined
shafts been used.

The morning-glory spillway may operate with free flow or, de-
signed properly, it can operate submerged. For free flow, the discharge
characteristics are similar to those for a straight overfall dam section,
where an increase of discharge is proportional to the three halves power of
the head, When operated submerged, the flow characteristics change com-
pletely, An increase in discharge is then proportional to the square root of
the head but, if part of the shaft flows full, the effective head may exceed
the head measured above the crest. A point to be remembered is that, once
submerged flow occurs, a further increase in head on the crest results in
a very limited increase in discharge. Thus, if a morning-glory spillway is
designed to operate submerged, an additional factor of safety should be ap-
plied to flood predictions to guard against the spillway capacity ever being
exceeded.

The morning-glory spillway is attractive in that it can often be con-

structed at less cost than other types; it is readily adaptable to dams in nar-
row steep canyons; and where a diversion tunnel is used to conduct the river
around the site during construction, it can do double duty by serving as a
portion of the spillway tunnel, When designed to operate submerged, this
type of spillway serves well for retention reservoirs where the flow down-
stream is to be limited,

1/ Hydraulic Engineer, Design and Construction Division, Bureau of

Reclamation, Denver, Colorado
 The principal drawback is the fact that information is quite meager
on prototype operation of these structures, Consequently, designers have
had a tendency to shy away from the morning-glory spillway if some other
type will do,

The design is sound from a hydraulic standpoint. The question-

able factors are structural, namely Will the concrete or other material
composing the curved surface of the vertical bend and horizontal tunnel
withstand the velocities and scaling effects of the fast-moving water? Will
ice or- debris present difficulties in operation? Will vibration or noise be-
come objectionable?

With .the construction of the Heart Butte Dam spillway, which con-
sists of a morning-glory designed for submerged flow with as much as 54
feet of head on the crest, and the design of the Hungry Horse Dam spillway,
which will feature a total drop of 475 feet from headwater to tail water, the
above questions take on added significance. For example, vibration of the
Heart Butte spillway, which is embedded directly in the earth dam, would
create a serious situation. Erosion of the tunnel lining in the Hungry Horse
spillway at the very high velocities contemplated would produce intolerable
difficulties,
In an attempt to par'tially clarify some of these questions in light
of prototype experience, a questionnaire was sent to persons connected
with morningeglory shaft spillways throughout the world, The text of the
questionnaire varied, but these three questions were common to all .of them:

1. Has any erosion of the concrete been experienced in

the tunnel, especially in the vicinity of the vertical elbow?

2, Has .objectionable vibration or noise been noticeab..

in. the structure during operation?

3. Has any difficulty .1:peen experienced in the pa,ssing

of debris Or ice ?

The response and cooperation received in reply to this inquiry were ex-
cellent. As to the value of the information collected, you may be the
judge.

As far as it is possible to determine, there are approximately 25

true morning-glory shaft spillways either in existence or under construction
today. The first morning-glory spillway was built in 1896, designed by
James Mansergh, for the Blackton Reservoir, England. The .second morning-
glory spillway on record was built in 1916, at the Taf Fechan Dam in Wales.
The remaining 23 spillways have been constructed since 1925. This would .
indicate that the morning -glory •spillway is a rather recent innovation; or
it may indicate that the designer has played safe by avoiding it ft is in-
teresting. to note that model.. studies have been made to aid in the design of
practically all of the morning-glory spillways, dating back to 1.96. Un-
fortunately,
• model studies do not tell all that is desired,

2
 Why the anxiety over morning-glory spillways? The following
example should illustrate the point: The Arizona spillway at Hoover
Darn, which has a fall of approximately 570 feet, went into operation for
the first time in August 1941. A week later, the gates were raised and
a hurried inspection was made of the tunnel, No erosion was apparent,
and operation was continued until December of the same year, The average
flow during the 4-month period was 13, 500 cfs. The maximum flow was
38, 000 cfs, which lasted for only a few hours. At the end of this period,
a hole was discovered 115 feet long, 30 feet wide, and 45 feet below in-
vert grade, immediately downstream from the vertical bend. The veloc-
ity was approximately 150 feet per second, The spillway was designed
for 200, 000 cfs, so the average flow was about 7 percent of capacity. You
have probably heard this story before, but we have not forgotten it.

The Hoover Darn spillways are not the morning-glory type, but
the tunnel arrangement is similar to some morning-glory spillways. There
are a number of theories as to what actually happened at Hoover Dam, but
nothing conclusive, In short, there was a misalignment of the invert of the
tunnel, there was evidence of cavitation, and the large hole is indicative of
impact and scaling. This we do not want to happen again in any structure.
But a question presents itself: If this can happen to an inclined shaft, what
can be expected in the case of a vertical shaft terminating in a sharp 90 0
bend?

On the other side of the ledger, we can cite experiences relative

to the spillway at Fontana Darn of the TVA which was completed in 1945.
The Fontana Darn was in the process of design at the time of the incident
at Hoover Dam, so special care was exercised in the tunnel design, This
spillway also is not the morning-glory type, but it has an inclined tunnel
with a fall of 410 feet, 160 feet less than at Hoover Dam. It has operated
every year since its completion and has carried discharges of 20, 000 cfs,
or approximately one-ninth of capacity. Nothing but roughened surfaces
have been evident to date.

Now for the results of the investigation of morning-glory shaft

spillways, Starting with the Taf Fechan morning-glory spillway in Wales,
which is the earliest one on which there is any concrete information, you
may get a picture of the evolution of this type of spillway as well as infor-
mation on prototype operation. The Taf Fechan (Figure 1) was designed in
1916 to handle a maximum discharge of 3, 000 cfs. The drop from head-
water to tail water is about 100 feet, with a maximum head of 2.8 feet of
head over the crest. The design is sound--in fact greatly overdesigned by
present standards. The long sweeping curves of the early designs are very
suggestive of the morning-glory flower. Correspondence with the late
W. J. E. Binnie, who acted as consultant on the Taf Fechan job, indicates
that no trouble has been experienced with this spillway and no unusual re-
pairs have been required. It is interesting to note the presence of the anti-
vortex device on this early spillway.

There is no record of a morning-glory spillway being built for the

next 10 years, but there were 10 constructed between 1925 and 1931. The

3
 Davis Bridge spillway on the Deerfield River near Whitingham. Vermont,
was completed .about 1925, This spillway (Figure 2) was designed for a
maximum discharge of 27,000 cfs„ a .drop of 180 feet, and a head on the
crest of 8 feet,. Reports indicate that this structure has discharged a cum-
bet' .of. times for short period,s, The maximum discharge occurred in the .
.September hurricane flood of 1938 when the water reached an elevation of
6 feet on.. the crest„ The computed discharge at that time was about 19,000. .
second feet„ or 72 percent of 1:„:apacity. The spiaduringthis flood exceeded
10, 000 second. feet, for about 15 hours" duration. Inspection after the flood
showed n.o appreciable erosion of the concrete in the horizontal part of the
tunnel, Several smal). erosion pockets were apparent at the vertical bend,
but it was not certain whether this was caused by erosion, due to velocity .
or 'spalling due to freezing and thawing of the concrete surfaces, In any
event, the repairs were of a minor nature, The report goes on to .say that
.all ice and debris which passed over the crest between the piers has readily
gone through the spillway„ Logs and trees 50 to 60 feet long have passed
through the spillway at times, The spillway makes some noise when it is
in operation, but this can be heard only a short distance from the dam, and .
vibration appears to be no problem,

The Gibson spillway on the North Fork of the Sun River in Montana,
a Bureau of Reclamation dam, was completed about 19260 This structure
(Figure 2) was designed for 50, 000 cfs, a drop of 160 feet, with 20 feet of
head over the crest. Later the crest portion was altered and six radial
gates were installed, The :report states that during the month of May 1938,
a total, of 435„ 00.0 acre feet of water flowed through the spillway.. The peak
of this flow amounted to 13, 000 second feet, or 26 percent of capacity. Upon
inspection. of the tunnel in :1,948, two spots about 5 feet in greatest dimension,
ovol in shape„ and about 2 inches deep, prominently displayed erosion in the
vertical bend, Water passing through the spillway in the 1948 season caused
them to show additional erosion. From the irregular and misaligned nature
of the vertical bend, it would be unfair to tie down the responsibility for this
•erosion to any one cause, The report goes on to state that no ice has passed
through the spillway, but large trees have passed. through it during flood sea-
son with no apparent damage. Noise was reported to have been considerable.,
increasing as the discharge, but this was not particularly objectionable,
Vibreion was considered normal, and in. no way detrimental, With radial
gates, vortex actioti is no problem,

Two British conservatively designe,d morning-glory spillways wq.re

completed in 1926 and .1927, the Silent Valley in Ireland and the Ponti.2-,,.,n
KetChil in :Singapore (Figure 3), The Silent Valley spillway was designed for
a (. .apac.ity of 2, 600 cfs„ a 55 -foot fall and 2,3 feet of head on the crest. The
Ponti.an Ketchil :spillway was designed for a. discharge of 2, 700 cfs, a fall of
60 feet, and .a .h.ead of 2,7 feet on the crest, Mr. Binnie informs us that the
Silent Valley- and Pontian Ketchil ,spillways have operated as expected without
appreciable erosion of any kind and without unusual repairs, Apprehension.
that•erosion might occur caused the designers of the Silent Valley spillway
to line the vertical bend with ca.st iron, The Pontian Ketchil spillway is sit-
uated in the .Malayan forest where cyc.Jonic storms OCCUr of such intensity as
to uproot large trees a„nd carry them for considerable distances, Many of

4
Max. W.S:-% ,
- - 66'Clici: -
2.9.--"
s
W . 4' Fins 9" wide
w
1
-4725

4700

4673

MAX. W.S.
E1.4724-
PLAN MAX. W.S:,
Y
-16 PIERS
61.5 DIA. -

77 31' 36 DIA.-
7E1.4623.75
22.5 DIA
154.2'
02.42 29.5' DIA.
55 R.

22.5 DIA.

SECTION SECT 10 N
SECTION
TAP FECHAN DAVIS BRIDGE B. GIBSO N

Figure 1 Figure 2
 80 Duo

SECTION
PONTIAN KETCHIL

r •

SECTION
SILENT VALLEY

Figure 3

SE ION
BURNHOPE MANUHERIKIA FALLS

Figure 4
these have passed through the spillway without damage. Antivortex devices
consist of four fins on the Silent Valley spillway and 15 piers on the Pontian
Ketchit

The capacity of another English spillway, the Burnhope (Figure 4)

in Durham County, is 2, 600 cfs, 2.7 feet of head on the crest, and approxi-
mately 100 feet of drop. The vertical bend in the Burnhope spillway was
purposely lined with cast iron to prevent erosion. One moderate flood of
1, 500 cfs, or 55-percent capacity, has been recorded with no undesirable
operating conditions or damage. In this design a curtain wall extends across
the center of the spillway parallel to the horizontal tunnel.

The Manuherikia Falls spillway in New Zealand (Figure 4) is de-

signed for 15, 000 cfs, a drop of 90 feet and a head of 6 feet on the crest.
This spillway has passed several small floods since 1935, aggregating about
2 feet of head on the crest, besides a moderately large one in November 1948.
In 1948, the spillway passed a maximum flow of 5, 000 cfs, representing one-
third of capacity for a depth of 3 feet over the crest. The report states that
there are signs of wear at the circumferential construction joints in the
morning-glory proper to about 25 feet below crest level, Some of the con-
crete has been eroded to a depth of about 3 inches at these joints. Frost
action may have caused some of the erosion, as there is seepage through
the rock surrounding the spill-way. The remainder of the vertical shaft is
in good condition. The vertical curve, however, shows erosion at con-
struction joints, amounting to about 3 inches, The floor in the horizontal
section of the tunnel shows general wearing of the concrete surface of ap-
proximately 1 inch. Some of this occurred when the tunnel was used for
stream diversion. Very little debris or ice has passed through this spill-
way, and no objectionable noises have been noted. Six piers serve to
counteract vortex action.

The Jubilee spillway (Figure 5), the largest English spillway up to

this time, has a capacity of 17, 000 cfs, a drop of 245 feet, and a head of 9.4
feet on the crest. There was evidently some concern as to the advisability
of dropping the water 245 feet into a vertical shaft (100 feet greater than in
any former spillway), which is evidenced by the inclined shaft, The only in-
formation on Jubilee spillway states that it has operated as expected, Dis-
charges have been small or moderate, and no 'erosion is reported. This is
the first spillway in which special air venting has been employed. A curtain
of water tends to seal the tunnel at the entrance to the inclined shaft for the
higher discharges. An air duct was, therefore, provided downstream from
the curtain of water in the inclined portion of the tunnel, It is interesting
to note that the curtain wall on this spillway is normal to the one on :Burnhope,

The Owyhee spillway in Idaho (Figure 5); constructed in the early

thirties by the Bureau of Reclamation, was a daring design at the time, with
a capacity of 30, 000 cfs, dropping 300 feet through a vertical shaft, The
spillway was designed for a maximum head of 12 feet on the crest. A float-
ing ring gate was incorporated in the crest portion. The ring gate operates
on the same principle as the drum gate used on straight overfall
It is raised by admitting water to the chamber beneath and lowered by draining

5
water from the chamber, All reports from the Owyhee spillway (on which
we have kept special watch) are favorable, A large flow occurred through
the spillway during a. 3-month period in the spring of 1936, when approxi-
mately 300,000 ac:iee feet of water was passed. The maximum discharge
recorded was 15,000 second feet, or one-half capacity. Since then smaller
discharges have passed through the spillway frequently. The spillway op-
erated at 18,000 cfs or 60 percent of capacity in 1952. Inspection of the
vertical shaft, elbow, and horizontal tunnel downstream has disclosed no
indication of erosion of any kind, These inspections have been conducted
frequently since the spillway first operated in 1936. No evidence of vibra-
tion has been apparent in the spillway and the jet does not flutter even with
the gate in a raised position. Incidentally a: single pier serves as an anti-
vortex device and also houses the ring gate operating equipment,

Merely as a side light a few operating characteristics will be men-

tioned. For heads of 1 to 2 feet over the gate, the water falls in a solid
sheet toward the center of the shaft- apparentlyentraining air there faster
than it can be released at the outlet end of the tunnel. This causes the pres-
sure to increase until it is sufficient to break back through the sheet of 'fall-
ing water., Then air emerges with sufficient force to carry spray 50 or 60
feet above the level of the gate (Figure 6)0 This phenomenon occurs some-
times as often as once every 5 minutes, depending upon the tail-water ele-
va.tion. For heads less than 1 foot, entrained air can apparently move back
up the spillway shaft unhampered. For heads greater than 2 feet, the air
pressure is not sufficient to break back and the air is forced through the
outlet end of the tunnel causing spray to be thrown high into the canyon. The
jump caused by a rather large tail-water depth forms in the tunnel to pro-
duce this effect.

The Lady Bower (Figure 7) is the latest in English morning-glory

spillways. Constructed in the middle 1930s it was designed to operate at
a maximum of 10 feet of head on the crest with 122 feet of drop, The report
states that the spillway has operated several times with no objectionable
features, No erosion of the concrete has so far been experienced, either
in the tunnel or at the vertical bend, The internal step form of design was
arrived at by model studies, The report states that the step design is cheaper
to construct but is less efficient than the smooth morning-glory shape, The
crest is surmounted with 12 piers,

The Kingsley Dam spillway in Nebraska has operated many times.

The maximum discharge through the spillway has not exceeded 4, 500 out of
54,000 cfs, or 8 percent of capacity. There have been no difficulties ex-
perienced in operation of the spillway or the 12 vertical tractor gates. The
vertical bend and tunnel show no perceptible wear. There has been no op-
portunity to pass debris through the spillway because flow has always passed
under the tractor gates.

The spiliway for the Heart Butte (Figure 8), recently constructed by
the Bureau. of Reclamation, is unique in that it is designed to operate prin-
cipally submerged, Both operate in conjunction with retention reservoirs.
The Heart Butte Dam .spillway is located in North Dakota. It was designed

6
 .... ..,
CURTAIN FTAIN1NG j,
WA - 2625
„--
//
' / '
1)
1
El. 2645

2650

/3650
;MAX. W.S. PLAN
MAX. W.S. PLAN
E1.2670) ±-601x 12'
--- 1 ! R1NG GATE
--PI 9658 -x

30.75 DIA. -..' , 'AIR VENTS

-E1.2410.3
V
309'

50' R.
,
El- 9 6' D1A

SECTION
JUBILEE
•777-H.-1 OWYHEE

Figure 5

Figure 6
 32.5' DIA.-,

MAX.W.S. PLAN

--E1.2078.5
111114111 - 27' DIA.
j,-- E1.2064.5

° --El 2053 85
11.67 DIA.
AIR VENTS---
El 20 33
11D1A. 59.5'
4•::1 ,
A -21R.

DIA.

SECTION •

H. KtNS$LIY HEART BUTTE

Figure 7 Figure 8
for a capacity of 5, 600 cfs, with a drop from headwater to tail water of
106 feet, and operates submerged at a maximum head of 54 feet on the
crest. Free flow occurs for heads up to 6 feet. Submerged flow occurs
for heads above 6 feet. A large vortex forms for heads from 6 to 20 feet,
For heads from 20 to 54 feet, vortices are either small or nonexistent.
The six piers on the crest serve to break up the large vortex into a num-
ber of smaller ones reducing fluctuations in discharge and resulting in
improved flow conditions, This dam was completed in 1949, and in the
Spring of 1950 passed a major flood. The spillway operated for approxi-
mately 1 month and passed 148, 000 acre feet of water. The maximum flow
into the reservoir during the flood was 24,000 cfs, while the maximum dis-
charge released through the spillway amounted to 3, 800 cfs. An inspec-
tion after this flood showed erosion at two points in the vertical bend. In
each case the eroded areas were about 8 inches in diameter and a maxi-
mum of 1 inch deep. This spillway again went into operation in 1951 and
52.

Of the 25 morning-glory spillways in existence, only those which

have operated have been mentioned. Incidentally, there have been two new
morning-glory spillways recently completed, the Wautauga and South Holston,
constructed by the TVA in Tennessee, Each has a capacity of about 60, 000
cfs and a drop of approximately 300 feet. Others are the Hungry Horse Dam
spillway in Montana, capacity 50, 000 cfs and 475-foot fall; and several in
Greece, Spain, and Portugal, all under construction.

Summing up the above prototype information, it can be said that

at least 7 of the 25 spillways above have undergone a fair test.

The results of the investigation are by no means conclusive. The

results are encouraging, however, in that

1. No erosion of the concrete was reported at the vertical

bend at Owyhee Dam. This spillway has vperated often twice
at half capacity. It has the greatest fall of any morning-glory
spillway Which has operated, Also, in no case has erosion of
a serious nature occurred on any of the spillways investigated.
It is quite conclusively believed that the small erosion patches
which occurred at the vertical bend on at least 4 of the spill-
ways were caused by logs falling down the shaft.

2. Vibration and noise have in no case been found to be

sufficient to be considered objectionable.

3. In no case has debris or ice presented any unusual prob-

lem in operation except with respect to impact at the vertical bend.

To date, the best method of coping with erosion caused by high

velocities, regardless of the type of structure, lies in maintaining smooth
surfaces, outlawing misalignment, and eliminating construction joints in
vertical bends. This is the reasoning and procedure followed at the present

7
time by the Bureau of Reclamation on all types of tunnel spillways. It
stands to reason that once a surface has been roughened, it is more sus-
ceptible to scaling, impact forces and cavitation when exposed to high
velocity water„
In closing I would like to encourage the collection of data and the
preparation of reports on the prototype operation of hydraulic structures.
It is actually a small investment compared to the benefits which can be
derived. What better method is there of detecting defects, of effecting
economy and improvink, future designs?

8
Interior - Reclamation - Denver, Colo„