American Society of Civil Engineers.

Transactions of the American Society of Civil Engineers (Volume 81) online

. (page 82 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 82 of 167)
Font size
QR-code for this ebook

the reservoir water immediately up stream from Bay 51 with potassium
permanganate, it was finally proved that the leakage came through
the cut-off. Probably it was due to inadequate bond along a horizontal
construction joint in the new "heel", within a foot or two in elevation
of the old construction joint at the top of the cut-off wall. The leakage
was practically stopped by the simple exiDedient of dumping clay into
the reservoir over the new "heel" from a float extending out from
the east shore.

It is probable that in time the leak at Bay 51 would have sealed
itself, as has since become true of the leak at Bay 36. In the latter
instance the same remedy was applied as at Bay 51, and fine cinders,
as well as clay, were dumped from boats into the reservoir water over
the "heel". The cure, however, was not effected in this manner; but
it was noted that the leakage was gradually decreasing, even though
the water level of the reservoir grew higher, until by November 1st,
1915, there was hardly any leakage at this point.

If the leakage at these two points was due to carelessly treated
construction joints in the new "heel", the stopping of the leakage
may in turn have been due — in part, at least — to the taking of some
of the horizontal thrust by the new "heel", with consequent tightening
of the construction joint by reason of the compression caused in the
up-stream portion of the "heel".

Other leaks, of a minor nature, which appeared in the original
portions of the structure have stopped or are stopping themselves in
the same manner. Thus on May 25th, 1915, there were eighteen
leaks or traces of leakage from weep-holes or drain pipes in the
foundation soil, whereas, by November 1st, 1915, there were only
three, and these the attendant designated as "traces". In no case


has a pressure test indicated a head of more than 2 ft. above the tops
of the weep-holes or drain pipes.

At times certain of the drainage openings yielded muddy water,
instead of the normally clear water. In the opinion of the writer,
these phenomena, usually temporary, were due to adjustments in the
foundation soil caused by increases in loading as the reservoir refilled.
In a few instances such openings were grouted shut, rather than take
any chances of harmful erosion under the footings.

It may not be amiss to refer here to another phenomenon, namely,
that in a number of pipes which are sealed into weep-holes and com-
municate through the footings into the foundation soil, the water
level remains practically at the top of the pipes, which are about 2 in.
in diameter and extend about 6 in. above the top of the concrete
footings. These pipes do not actually weep, except in the sense that
when water is dipped out of the pipes, they refill in a day or two, but
again do not overflow. The possibility is indicated that capillary
tension over the surface of the ground- water rising in the pipes is
sufficient to prevent overflow, but the writer has been unable to advance
any explanation which is satisfactory to himself.

2. — Drainage of BocTc Foundation at New Spillway. — In the case
of the shale forming the foundation for the buttress footings of the
new spillway, there is, of course, no danger that deep drains will
facilitate erosion of the foundation. This is especially true because
of the fact that the shale contains practically no lime which might be
taken out in chemical solution. Moreover, as described in detail
under the heading "Geological and Foundation Conditions", the rock
contains water-bearing seams, and it was possible that ground-water
communication might exist between the reservoir and certain springs
which appeared in the area excavated for the new spillway footings.
Hence it was concluded to provide more drains at the new spillway
than in the foundation soil under the intact portions of the original

The system of drainage provided at the new spillway consists in
general of three deep drains per bay, placed as shown in the typical
sections of Plate X. The drains were sunk with the Calyx, shot-
drilling apparatus. In the bed-rock they remain unlined. Above the
shale, however, 4-in. pipes were sealed as well as possible into the
rock and carried up at least to Elevation 96, in order that leakage


might be detected above the level of the water which ordinarily stands
in the new spillway bays as high as the elevation of the horizontal
drain through the foot of the spillway apron (Plate X). It was
practically impossible to prevent the continuously flooded condition
of the space between the new spillway footings; at the same time, this
condition has the merit of preventing disintegration of any shale by
exposure to the atmosphere.

The up-stream drain of each trio extends down to Elevation 70,
well below the bottom of the combination cut-off and anchoring wall.
The middle and down-stream drains extend down only to Elevation 75,
which elevation also is below that of the bottom of the cut-off. The
number of drains was increased at Bays 11 and 12, being the two most
westwardly bays of the new spillway ; and, inasmuch as the strata slope
upward toward the west bank, it is probable that the rock-water, fol-
lowing the dip of the strata, will be intercepted at these bays.

Ever since the deep drains of the new spillway were sxmk there
has been a flow of rock-water in Bays 12, 13, 16, and 19. However,
both the quality of the water and its relatively constant yield prove
that there is little, if any, direct communication between such rock-
water and the reservoir. The reservoir pressure doubtless has some
influence on the rock-water in question, for there was a minor increase
in the yield as the reservoir gradually refilled. Yet it is also true that
during "dry" periods the rock-water yield has decreased somewhat, even
though during the same periods there was but little fluctuation in
reservoir level. The maximum rock-water pressure, measured at the
tops of the new spillway drains, was equivalent to a head of about
2 ft. The drains overflow freely into the surrounding pools of water.

In view of the drainage system provided throughout the higher
portions of the structure, it appears reasonable to assume that, if the
drains are not frozen shut at times when uplift pressure may be
accumulating under the footings, no serious uplift pressure is to be
feared. The assumptions of uplift pressure under the "most severe
conditions within the limits of reason", as set forth in Table 3, are
based on the drainage provision above described.

Curtain-Wall and Roof. — It was primarily to prevent any consider-
able freezing at the top of the weep-holes and deep drains throughout
the higher bulkhead sections of the dam that a curtain-wall and roof
were constructed, as illustrated in the typical section B-B of Plate XII.


Manifestly, an ice plug, say, 6 in. deep, at the top of a weep-hole or
drain pipe, would for practical purposes serve to confine water under
the footings as effectively as though the openings had been filled
with concrete and made a part of the footings.

It is believed that, as the result of housing in all bays having
deep drains, namely. Bays 11 to 47, inclusive (Figs. 16 and 17), the
temperature of the corresponding footings will be prevented from
falling to more than a few degrees below the freezing point. During
cold weather the radiation of heat outward through the curtain-wall
and roof will be largely counterbalanced by the radiation upward of
heat from the unfrozen foundation soil and by the further radiation
of heat from the reservoir water through the deck of the dam into
the enclosed space. Water is at its greatest density at about 39° Fahr.,
with the density gradually decreasing as the temperature decreases
to the freezing point. Consequently, water in the reservoir under an
ice covering is at least several degrees above freezing temperature;
the difference may be as great as 7 degrees. The assumption made as
to the effect of the curtain-wall and roof is supported by observations
said to have been made at various hollow dams which are provided
with both deck and apron, thus being entirely enclosed. It is reported
that at some such structures freezing does not occur in the enclosed

As soon as practicable in the late winter of 1914 the writer had
maximum and minimum thermometers placed in the old spillway
section of the Stony River Dam, and had the larger openings into
that section closed so as to allow a series of records to be made,
covering exterior and interior temperatures. Conditions were not
favorable for conclusive results, because the reservoir, of course.
had been emptied at the time of failure, thus exposing both apron and
deck to the cold, with the result that the only source of heat was the
foundation soil under the old spillway. Nevertheless, even these obser-
vations showed clearly that variations of temperature inside the enclosed
spillway section were far less than the corresponding variations in
outside temperature; also, that periods of relatively higher and lower
temperature inside the structure lagged considerably behind the corre-
sponding periods of relatively higher and lower temperature outside.

During the winter of 1915-16 the lowest temperature within the
enclosed portions of the dam was + 26° Fahr., and the lowest exterior


temperature was — 12° Fahr. Under these conditions ice formation
has taken place, but not of a serious nature. At the very least, the
enclosing of the bays enables the attendant to inspect carefully the
drainage system, regardless of weather conditions, and also protects the
operating mechanism of the outlet gates. When ice forms within the
enclosed portions, the attendant can readily keep it broken or cut
away so as to prevent any harmful results.

A portion of the curtain was constructed in the form of a roof,
instead of continuing the wall vertically to intersect the deck, in
order to avoid enclosing what would have been a triangular section
at the top, exposed to cold air on the down-stream side and frequently
to the ice covering of the reservoir on the up-stream side. For the
bulkhead section between the old and new spillways, the curtain-wall
was made 12 in. and the roof 7 in. thick, and, for the eleven bays
immediately east of the old spillway, the thickness of the curtain-wall
was reduced to 6 in., the roof remaining 7 in. in thickness. At the base
of each curtain-wall there is an opening, about 12 by 18 in., fitted
with a wooden flap door, as shown on Plate XII, so as to allow the
outward flow of any leakage water, but at the same time to keep air
currents and snow out of the enclosed space. Wooden ladders bolted
to the buttresses are provided between the walkway and the floor at
each enclosed bay.

In addition to this primary purpose, the curtain-wall and roof
furnish effective lateral stiffening for the buttresses. The writer does
not suggest that the bracing of the buttresses in the original structure
was insufficient, yet he confesses that he feels better satisfied with
the buttresses braced by the curtain-wall, especially as regards the
lower portions of those of maximum height, where originally an addi-
tional lower set of horizontal brace-beams might have been desirable.

Furthermore, the curtain-wall tends to equalize deflection in the
foundation soil under the footings, and, consequently, to shift the
load from weaker to firmer soil. Finally, the curtain-wall was utilized
in a minor way in the design of the strengthening of the intermediate
flooring, allowing, as it did, the omission of transverse reinforcement
for a reasonable distance on each side of it.

Excepting for these secondary ftmctions of the curtain, it would
have been well to make the wall and roof hollow, for instance, by the


use of cement plaster on metal lath. A hollow construction could have
been obtained with concrete, of course, but at a sacrifice of economy.
Hollow construction would have better conserved the heat available
for warming the enclosed portions of the bays. Relative to this it
is interesting to note that more freezing has taken place in the eleven
bays immediately east of the old spillway, where the curtain-wall is
only 6 in. thick, than in the remainder of the enclosed bays, where the
wall is at least double that thickness.

Leakage in Superstructure. — At the time of failure there were pro-
nounced leaks at certain contraction joints in the deck. In the type
of construction used the deck slabs are not in contact with each other
at the buttresses. Instead, a tongue, integral with the buttresses and
varying from 18 to 24 in. in width, separates the slabs. Contraction
joints exist, therefore, between these tongues and the slabs, over all
buttress haunches. It appears that, in the original construction,
insufficient care was taken in forming these joints, with the result
that in most cases they did not fulfill their intended functions, and
contraction cracks occurred only at intervals averaging about 75 ft.
The tenacity of the bond in certain of the intended joints is well illus-
trated in Fig. 5, which shows the deck of Bay 18 overhanging west-
wardly for a distance of about 13 ft., the adjacent Buttresses 17 and
16 with their intermediate deck having fallen away due to the under-
mining of the buttress footings. Naturally, such joints as did act
opened excessively and allowed considerable leakage to take place.

The worst leakage of this kind occurred on the east side of Buttress
29, where, during the severely cold weather obtaining at the time of
the failure, ice formed to such an extent as to block the stairway
leading from the walkway down to the outlet gate-operating mechanism.
It was considered impracticable to attempt to reconstruct all the
contraction joints so as to make them perform properly their intended
function. Instead, all joints which were acting too freely were made
as nearly water-tight as possible by cleaning them out thoroughly,
packing them with tar-paper placed on edge, and then pouring in
hot asphalt. The asphalt has a tendency to flow from the joints during
extremely hot weather, but it is believed that such asphalt as remains
in the joints, together with sediment deposited by leakage water, will
prevent any serious leakage. As a matter of fact, the joint at Buttress
29 still leaks somewhat, especially in cold weather. However, by


reason of the bay being enclosed by the curtain-wall and roof, the
attendant can now with little difficulty clear the stairway of the trifling
quantity of ice that forms during severely cold weather.

In the new spillway section precautions were taken to insure
positive action of the contraction joints by covering the entire sur-
face of such joints with two thicknesses of three-ply tar-paper, or
the equivalent thereof. The leakage here has been inconsequential.

Miscellaneous Problems.

Strength of Decks and Buttresses. — In the case of the greatest
flood reasonably to be anticipated, the deck of the original structure
would be subjected to greater loading than that for which it was
designed. The resulting increase in stresses would be greatest at the
bottom of the upper lift of the deck, namely, at Elevation 126. Here
the maximum stresses with head-water at Elevation 142.25 would be
approximately 22 000 lb. per sq. in. in the reinforcing steel and 800 lb.
per sq. in. in the concrete. The original steel is reported to have been
purchased under a guaranty of 55 000 lb. per sq. in., minimum elastic
limit, and the concrete was of a character warranting the assumption
of an ultimate strength of at least 2 000 lb. per sq. in. in 90 days.

Under these conditions, it did not appear necessary or advisable
to strengthen the decks of the original structure. Likewise, it was
not considered necessary to strengthen the buttresses, other than to
give them the benefit of the incidental stiffening provided by the
curtain-wall previously described.

At the new spillway section, on the other hand, it was considered
proper to design the deck for a loading due to head- water at Elevation
142.25, notwithstanding the fact that the new spillway deck thus has
greater strength than the deck of the original portions of the structure.
Also, the thickness of the buttresses was increased, as compared with
the original structure. Above Elevation 112 the increase in thickness
was a minimum of 4 in., and below that elevation a minimum of 6 in.
Furthermore, the new buttresses are more heavily braced laterally
than in the case of the original construction. The walkway slab is
thickened so as to serve as a brace-beam, with its principal dimension
horizontal, and two other brace-beams, with their principal dimensions
vertical, are provided lower in the structure. Detailed comparisons
may be made by examining the data on Plates X and XII.


Increase in Storage Capacity. — In view of the much larger spillway
capacity provided in the reconstruction, it appeared logical to make
temporary use of some of the depth over the spillway by providing
increased storage capacity. It was concluded to be feasible to utilize
in this way approximately 3 ft. of the 6 ft. 6 in. available depth over
the main spillways. Such use for storage purposes, however, is tem-
porary only in the sense that, during extraordinary floods, the entire
depth over the main spillways must be available for flood discharging
purposes. The resulting increase in the available storage capacity
was approximately 380 000 000 gal., or about 25% (Fig. 10), and it
afforded a certain consolation to the owner for the heavy expenditures
entailed by the failure and reconstruction of the dam. It may reason-
ably be said that, even in the case of the original structure, it would
have been possible similarly to utilize some of the 3 ft. difference in
elevation between the crest of the spillway and that of the bulkhead
sections; yet it is equally true that an appreciable quantity of addi-
tional storage could have been obtained only at the risk of overtopping
the structure during floods.

Flash-Board Supports. — The problem involved in increasing the
available storage capacity by about 25% was to provide a temporary
dam, in other words, flash-boards, which would fail within compara-
tively narrow, predetermined limits of loading. Thus, it was impera-
tive that 3-ft. flash-boards should not be overtopped by more than
3 ft. 6 in. at the very maximum. The effect of the use of such 3-ft. flash-
boards on the flood discharging capacity of the spillways was carefully
considered, and a number of hypothetical cases were worked out. One
of these cases, previously discussed in this paper, is shown in Fig. 14,
wherein it was assumed that, with the greatest flood reasonably to be
provided for, the 3-ft. flash-boards might not fail until the head-water
had reached an elevation 5 ft. above the base of the flash-boards.
The result of such studies was the conclusion that the use of flash-
boards would not affect seriously the maximum flood discharging
capacity of the dam as reconstructed.

It was impracticable to provide at small cost reliable means which
would cause the flash-boards to fail with head-water 3 ft. over the
boards, and yet not fail at lower elevations. On the other hand, it
is highly undesirable that the flash-boards should fail under any except
the more severe freshets ; otherwise the inconvenience of renewing


the flash-boards and of the loss of the 3-ft. depth of stored water would
occur too frequently.

Both of these considerations were subservient to a third condition,
namely, that the means of lowering the flash-boards or causing them
to fail must be absolutely automatic, inasmuch as conditions warranted
the employment of only a single attendant, and at the critical time
lie might be absent or asleep, or might for any other reason not attend
to the flash-boards when necessary. Finally, the cost of the flash-board
provision and of replacing the flash-boards was necessarily to be made
as low as possible.

For the purpose of meeting the foregoing requirements, the writer
had about seventy experiments made on steel bars or '"pins" to serve
as flash-board supports. The experiments were intended chiefly to
find a material which would fail at a predetermined stress, but would
not allow the pins to bend considerably at lower stresses. It was also
desired that, when failure occurred, it should consist in the pins
snapping oif at approximately the elevation of the spillway crest, thus
allowing the crest to be cleared absolutely of all obstruction by the
flash-boards and their supports. The experiments were performed in
the outlet channel, where, by using the sluice-gates and the previously
accumulated head in the partly refilled reservoir, it was possible to
duplicate the heads which would exist under actual operating con-
ditions, and hence use flash-boards and supports of the same size as
under actual conditions.

A controlling condition was the fact that flash-board sockets had
been provided in the original spillway 3 ft. 6 in. from center to center.
The same spacing was used for the crest of the new spillway, so as
to allow interchange of equipment. The scope of the tests covered
various conditions of head, lengths of time during which heads were
maintained, and degrees of turbulence (akin to reservoir wave action),
as well as various kinds and conditions of steel. Inasmuch as it was
desired to avoid the use of pins which would bend, and because of
previous experience with wrought-iron pins, it was not thought worth
while to spend time in testing pins of that material. However, steel
pins were tested, including soft and hard steels. The highest carbon
content was approximately 1.40 per cent.

None of the steel pins tested in their natural condition gave satis-
factory results. Those made of steel with even as high as 0.90 to


1.10% carbon (railway spring steel) persisted in bending before
breaking, whereas steels with higher carbon content, though they did
not bend unduly, were erratic as to the head or load at which they
broke. This was especially true of tool steels. Practically as a last
resort, the writer tried the process of hardening the steel; that is, the
pins were heated and then hardened by plunging into cold water.
Even under these conditions satisfactory results were obtained only
within comparatively narrow limits. The softer steels, of course, would
not harden, and the very high carbon steels became the more erratic in
their behavior by reason of such hardening. Round pins were used,
because of the uniform resistance of pins of circular cross-section,
no matter at what point the load might be applied (thus making
it unnecessary, in this regard at least, to exercise care in placing
the pins).

The most satisfactory results were obtained by the use of open-
hearth steel with a carbon content varying between 0.40 and 0.80%
and phosphorus and sulphur contents each not exceeding 0.05 per cent.
In the hardening process the pins were heated to a "cherry red" before
quenching in cold water. The heated portion extended about 8 in.
above and below that point in the pin which when in place would
be at the spillway crest. As the result of the tests on hardened
specimens with such characteristics, it was concluded that 1^-in. round
pins, vsdth carbon content of 0.62 to 0.77%, aiming at 0.72%, and
sulphur and phosphorus contents limited as above, would fail at
approximately the predetermined point, viz., when the elevation of the
head-water was about 4.3 ft. above the main spillway crests, or about
1.3 ft. above the top of 3-ft. flash-boards. It was also concluded that
such pins would not be subject to excessive temporary or permanent
deflections before breaking.

As it was impracticable to wait for further tests before making the
final selection of the supporting pins to be used, two sets were ordered
for each spillway, together with a sufficient number of extra pins to
allow for making confirmatory tests. Subsequent analysis showed that
the pins delivered actually have a carbon content of 0.68 per cent.
The flash-board pin sockets in both old and new spillways were adjusted
by grout filling to a uniform depth of 18 in. The pins were cut to

Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 82 of 167)