American Society of Civil Engineers.

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

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

from the trench excavation for the "heel" on the up-stream side of
the cut-off confirmed these findings. It was definitely established that
there was easy communication for water between the up-stream and
down-stream sides of the cut-off wall at Bay 28. Probably some of
the water which flowed back and forth, depending on the hydrostatic
pressure conditions, passed through the cut-off wall itself; but, at this
point, undoubtedly the greater part of the communication was under-
neath, through openings between the cut-off wall and the bed-rock.

At Bay 30, also, it was evident that the seal of the cut-off into
bed-rock was imperfect. In fact, at one point a layer of sandy mud,
averaging \ in. thick, was plainly visible between the concrete and the
shale, indicating that, in the original construction, the trench had not
been properly cleaned preparatory to the placing of concrete. Moreover,
the cut-off wall itself was porous to the extent of allowing a constant
trickle of water to pass through from the down-stream side. That
there was any such leakage at all was the more noteworthy because


of the fact that the hydrostatic pressure on the down-stream side was
due solely to the ground-water in the clayey foundation soil — though
it should be remembered, of course, that the foundation soil under
this maximum height portion of the dam is of poorer quality than
that at any other part of the structure.

In view of the foregoing findings, it was determined to underpin
the original cut-off wall for approximately 200 ft., viz., from about
Buttress 32 westward to the new spillway. A notch, about 24 in. deep
and 18 in. wide, was cut into the bed-rock, with its center approximately
at the up-stream edge of the original cut-off wall. The notch was
carefully cleaned and kept free from water while it was being filled
with a rich concrete which was carried to a height of at least 12 in.
above the bottom of the old cut-off.

This work proved to be the most expensive part of the entire recon-
struction, partly because of the constant fight with the water which
flowed into the trenches through the pervious soil under the river-bed,
and partly because of the fact that the character of the soil required
practically continuous sheeting and bracing. So difficult, in fact,
was it to carry on the work and to hold the up-stream side of the
excavation for the "heel" that at Bays 30 and 31 it was found advisable
first to pour the "heel" concrete and then, after that concrete had
set, to tunnel under it in order to investigate the efficacy of the cut-off
seal into the rock. Such tunneling was done from the adjacent bays,
and, to insure solid work under the "heel" at Bays 30 and 31, grout
pipes were inserted while the concrete for the cut-off underpinning
was being placed, the ends of the pipes being exposed under these
portions of the "heel". Later, the voids at the ends of these pipes
were filled with grout, by using the Ransome-Caniff grout mixer, under
a pressure of about 75 lb. per sq. in.

At certain places, as shown on Plate IX, the cut-oft' underpinning
concrete was extended on the up-stream side of the cut-off up to the
elevation of the bottom of the "heel". In such cases a slip joint was
made between the "heel" and the underpinning concrete. This joint
consisted essentially of a longitudinal groove, about 12 in. wide and
6 in. deep, lined with tar-paper. The groove was centered on the
bottom of the "heel". When the "heel" concrete was poured, therefore,
a tongue of concrete, likewise 12 in. wide and 6 in. deep, was formed
in the groove integral with the "heel" concrete. On all horizontal


surfaces, and on the up-stream vertical face of the groove, single sheets
of two-ply tar-i^aper were used to separate the two bodies of concrete.
On the down-stream vertical face of the groove, however, three sheets
of two-ply tar-paper were used.

The purpose of the slip joint is to allow the "heel" to move down
stream a fraction of an inch under the horizontal thrust acting
against the dam and hence transmitted by the tie-steel into the "heel".
Without such "play" an unfavorable loading might be caused at the
lower end of the long cantilever member of the "heel", because of
the resistance of the relatively inelastic underpinning concrete. The
latter, of course, is practically immovable because of its bond into
the bed-rock. The utility of the slip joint is probably to be justified
on theoretical rather than practical groiinds, but, in view of the small
expense involved, there seemed to be no warrant for omitting this

At about the middle of Bay 26 it was found that there was a
pervious spot in the bed-rock strata just below the practicable limit
of underpinning. Therefore a churn-drill hole was sunk into the
strata in question, and grout pipe connections were made, which, after
the completion of the underpinning and the new "heel", were utilized
in an attempt to force grout into the stratum. However, a pressure
of 75 lb. per sq. in. did not suffice to force down an appreciable quantity
of grout.

It is not improbable that the failure to force more grout into this
stratum was due to the fact that there was no vent for such rock-
water as might have been displaced by grout had a means of escape
for such water been provided. It happened that, for the grouting of
the underpinning of Bays 30 and 31, as previously mentioned, two
grout pipe connections had been left, and thereafter on the work,
profiting by the experience at Bay 26, at least two grout pipes were
placed in each void to be filled. Usually, the second pipe was con-
nected to the pressure grouting apparatus only after the first pipe
refused to accept any more grout. In such cases it was commonly
impossible to force grout into the second pipe, and apparently the
main function of that pipe was to serve as a vent. In the writer's
opinion, it is not unlikely that some of the reported failures of pres-
sure grouting operations at other dam sites have been due to the
fact that no such vent existed, or had been provided, for the escape


of air or ground-water contained in the voids sought to be filled by

Referring again to the faults in the original cut-off construction,
which required remedying by the means just described, the reconstruc-
tion afforded a rare opportunity for observing the results of this
construction work, which had probably been done with average care.
These observations show that great care is necessary even in foundation
work, where it is too often assumed that any kind of work will

■j'{. Extension of Cut-Off. — Inasmuch as the instifficient depth of
cnt-off wall in the neighborhood of the break in the dam was un-
doubtedly the direct cause of the failure, it was advisable to pay
especially careful attention to any further deficiency in depth of
cut-oif. At the sections of maximum height, precautions were taken
in the way of underpinning the original cut-oS wall, as discussed
previously. Such underpinning extended westward to the new spill-
way which in turn covered completely the area affected by the failure.
(Plate XI.) Here the combination cut-off and anchoring wall at the
heel extends into the shale rock to an average depth of about 10 ft.,
in the manner illustrated on Plate X. The depth is greatest in
the neighborhood of Buttresses 14 and 15, because there, as mentioned
previously, the rock was found to be somewhat shattered.

In the same region the rock yielded a considerable quantity of water,
especially after the drain holes had been drilled. Consequently, after
the combination cut-off and anchoring wall had been completed, a
hole was drilled, with the Calyx machine, through a pipe which had
been embedded in the center of the wall at about the middle of Bay 13.
The hole penetrated the underlying shale to about Elevation 70. It
was intended to grout this hole for the purpose of cutting off any
communication with the resei-voir which might exist in this neighbor-
hood through passages in the rock under the cut-off. It was assumed
that the existence of any such connection would be shown, when pres-
sure grouting was applied, through the manifestation of bubbling or
other symptoms at a drain hole extending to about the same elevation
and approximately 7 ft. down stream in Bay 13. However, a surprise
was afforded, in that the hole would accept no grout, and it was apparent
that the rock-water passages, which had been tapped a short distance
down stream in Bay 13, either came from a greater depth, or more


probably communicated to the westward — the slope of the bed-rock
(Plate IX) being upward and westward.

West of the new spillway, beginning at Buttress 10, the depth of
the original cut-off was as inadequate as at the point of failure. At
some places, especially near steps in the bottom, the depth of the
cut-off was less than 5 ft. below the bottom of the footings (Plate IX).
It is probable that everywhere in this region, except possibly at one
or two of such steps, the original cut-off extended to a depth of at
least 6 ft. below the original ground surface. That the wall was not
deep enough to cut off a pervious stratum of dark color in the over-
burden will be seen by referring to Fig. 7, in the upper right corner
of which the bottom of the original cut-off wall is shown supported
by a prop.

The entire west end of the dam, viz., between Buttresses 1 and 10,
was provided with a new "heel", the required depth of which varied
from 1 to 5 ft. below the bottom of the original cut-off. The new ''heel"
was made not less than 2 ft. 6 in. wide at the bottom, this being a
width convenient for trench excavation. A cut-off wall of this mini-
mum width was then extended downward from the bottom of the
"heel" to seal into the bed-rock. In this region the bed-rock consists
of shale of a very good quality; in fact, in some places the rock is so
hard that a notch only 6 in. deep was considered to afford a sufficiently
good seal. At other places, however, the notching was carried to
a depth of more than 3 ft. on account of penetrating a stratum of
soft shale which in part overlies the very hard stratum.

At the bottom of the required depth of "heel" a standard slip
joint was provided, in the manner previously described; and just under
the old cut-off (but above the slip joint) a shoulder of concrete, with
a minimum width of 12 in., was formed. The purpose of such a
shoulder of concrete was to make certain that any vertical load trans-
mitted into the old cut-off wall would in turn be taken up by the new
'"heel" and cut-off, thus guarding against any possible separation of
the old concrete from the new and against the placing of any excessive
shearing stress on the tie-steel connecting the new anchoring wall to
the original structure.

As will be seen from Plate IX, a core-wall had been provided in
the original construction extending 30 ft. west of Buttress 1. As
a part of the reconstruction, this core-wall was extended upward to


the elevation of the new parapet (Elevation 142.5) and downward
to seal into the shale. Also, because of the increase in assumed high
water elevation, it was extended about 15 ft. farther westward. The
precautions taken in providing a cut-off against percolation at the west
end of the dam were of a more conservative nature, for the reason
that it was near this end of the dam that the failure had occurred.

Between Buttresses 32 and 46 the depth of the original cut-off,
with a maximum of 45 ft., was concluded to be sufficient in proportion
to the maximum hydrostatic pressure. This conclusion was based on
the fairly well substantiated assumption that in this region the original
cut-off is impervious throughout its entire depth, but it did not depend
on the character of the seal of the cut-off into the bed-rock. As a
matter of fact, however, a personal inspection of the bond between
concrete and rock as exposed by the test pits at Buttresses B
and 37 convinced the writer that the seal is as effective as could be
desired. East of Buttress 46, on the other hand, certain difficulties were

It had been assumed by the builders of the original structure that
the cut-off extended to bed-rock at least as far east as the approxi-
mately vertical step in the cut-off at the middle of Bay 54 (Plate IX).
In fact, all records of the original work show the cut-off as being in
contact with bed-rock. It is not surprising, therefore, that the results
of core-drill holes Nos. 3 and 8, sunk during the investigation prior
to reconstruction, were discredited when they indicated that bed-rock
was reached only at a considerable distance below the bottom of the
original cut-off. Nevertheless, some mental reservations were made,
and subsequently a test pit was stmk at the middle of Bay 48. The
findings were by no means comforting, for, after penetrating flat-top
boulders, the upper surfaces of which offered some reason for their
previously erroneous identification as bed-rock, the pit extended through
clay deposits into strata of disintegrated coal. Rock was not reached
until at about Elevation 92. It is a shale, and quite satisfactory.

In view of the discovery of the disintegrated coal, which was
decidedly pervious, the existing cut-off was manifestly inadequate. This
was especially true of the portion east of the middle of Bay 48. It was
concluded, therefore, to underpin the original cut-off down to a line
sloping with approximate uniformity from Elevation 92 at Buttress 46
to Elevation 108 at Buttress 53, and meeting the bottom of the original


cut-off at the latter point. The proviso was made, however, that the
disintegrated coal should be cut off wherever its top came within
1 ft. below the bottom of such sloping line. The probable genesis of
the disintegrated coal and the measures taken to cut it off are shown
on Plate IX. Drill holes Nos. 9 and 10 were sunk after the discovery
of the disturbing conditions just described.

It was also concluded to underpin the original cut-off east of the
middle of Bay 54, as well as the core-wall which originally extended
50 ft. east of Buttress 56. Excavation for the underpinning in these
two localities proved to be very difficult because of the existence of
a sandstone talus in the form of huge boulders. The most economical
method was found to be the sinking of shafts from the surface of
the ground and the construction of tunnel drifts leading from such
shafts under the original cut-off. Some of the crevices discovered
between the boulders were found to be filled with clay, and others
were open. In one instance a crevice at least 2 in. wide was traced
for 6 ft., and it extended even beyond that distance. The shattered
rock was very difficult to excavate, and required frequent blasting.

Before the excavation was filled with concrete, grout holes were
churn-drilled into the rock, and grout pipes were placed in crevices
so that later, when these were pressure-grouted, it is probable that
an impervious rock, concrete, and grout curtain was formed, extending
between Buttresses 54 and 56 down to at least Elevation 110. Simi-
larly, grout pipes were left in the higher portions of the tunnel drifts
so as to insure, later, the filling, by pressure grouting, of any air
pockets left in the higher sections. The same grout pipes served to
allow the escape of confined air while the main concrete was being
placed. The concrete was carried to a considerable height in the shafts
previously mentioned and, because of its slushy consistency, undoubt-
edly created considerable pressure throughout the drifts. Pressure-
grouting was not applied until the concrete had set for a day. The
east core-wall was extended approximately 24 ft. farther east. Both
the old and new portions of the core-wall were built up to Elevation
142.5, to correspond with the top of the new parapet.

Drainage. — In the original structure the intermediate flooring of
each bay had 4 by 12-in. weep-holes. A number of these holes existed
at each edge of the main buttress footings, the shorter dimension of
the holes being parallel to the footings. The weep-holes were in


XJairs, one hole at each edge of the intermediate flooring; the pairs
were about 8 ft. from center to center, though the spacing was closer
at the up-stream ends of the bays. As a matter of fact, a considerable
number of these weep-holes were found to be clogged with concrete
which had evidently flowed under the weep-hole forms at the time
the original footing or flooring concrete was being placed. Apparently,
it had not been thought necessary to provide deeper drainage for the
foundation soil.

In planning the reconstruction of the dam, however, it was con-
cluded to provide more positive means of drainage. The system
adopted for that portion of the original structure which remained
intact differed from that adopted for the new spillway section :

1. — Drainage of Foundation Soil. — In the preliminary plans for
reconstruction, the writer proceeded on the theory that the more drains
the better, and especially the more positive, the relief of uplift pressure.
Consequently, it was proposed in general to extend three drains per
bay into the foundation soil. These drains were to consist of per-
forated pipes such as used in the system of drainage finally adopted.
The up-stream drains were to be inclined, and thus were to extend to
within about 5 ft. of the down-stream face of the cut-off wall and
to a depth of about 5 ft. below the bottom of the new "heel". The
middle and down-stream drains of each bay were to be vertical, at
distances of about 20 and 30 ft., respectively, from the cut-off wall.
At the seven bays of maximum height it was proposed to place the
up-stream drains 7i ft. from center to center transversely, thus pro-
viding two up-stream drains per bay. It was also proposed to keep
all the original weep-holes open through the original concrete and the
concrete of the footing strengthening.

On the other hand, the system of drainage finally adopted, though
utilizing the same methods as proposed in the preliminary plans,
differed as to the number and location of drains. It was based on
two main considerations, namely:

(a) The necessity of relieving serious uplift pressure; but, also,

(b) The inadvisability of facilitating percolation through the
foundation soil.

It was concluded that deep drains might be harmful in earth
foundations under circumstances where they would not be harmful


in rock foundations, because, in the former, they might actually
facilitate the flow of water by providing artificial channels where pre-
viously there was no channel at all. Especially might such results
follow if the deep drains were to extend close to the bottom of the
cut-ofF wall.

A consideration of secondary importance was the effect of the low
values of the coefficient of frictional resistance assumed in the recon-
struction design. Manifestly, a given uplift pressure reduces the
margin of safety against sliding by a less amount where there is a
low coefficient of frictional resistance than it does where the coefficient
is relatively high. It was believed to be preferable, therefore, to take
such risk as might result from the existence of pockets, as it were, of
unrelieved uplift pressure near the down-stream side of the cut-off
wall rather than to invite further leakage by tapping such relatively
harmless, deep-lying pockets of water under pressure.

In consequence of these considerations, a system of drainage was
adopted which involved but a single deep drain per bay for the higher
sections of the intact portions of the original structure, resting as they
do on clayey foundation soil. The positions of the drains are indicated
in the typical sections of Plates IX and XII. From the former it is
seen that the deep drains for the foundation soil extend from Bays 20
to 47, inclusive. The drains consist of 3-in. wrought-iron pipes, perfor-
ated with four |-in. holes per linear foot, one hole in each quadrant.
Pipe drains, rather than unlined holes, were adopted for the foundation
soil, in order to make the drains relatively permanent, and also because
unlined openings would not resist erosion of the foundation soil by
leakage water.

The deep drains are about 20 ft. down stream from the cut-off wall,
except at the bays of lesser height, where they are at a minimum
distance of 12 ft. from it. The drains ordinarily extend to a
depth of 5 ft. below the bottom of the new "heel", though they were
limited to a depth equalling two-thirds of that of the cut-off,
measured from the bottom of the original footing to the bottom of
the underpinning, if any, of the original cut-off. They were sunk
in lengths of 12 ft. or less. Flush-joint piping was used, so as to avoid
couplings or other projections; the latter might leave an annular space
around the drains to serve as an outside channel for flowing water
without the protection of a metal lining.


Great difficulty was experienced in sinking the drain pipes, primarily
because of the boulders encountered in the foundation soil. Various
methods were tried, among them the process of jetting, that is, of
sinking a drain pipe, open at the lower end, by using a jet of water
under pressure playing on the soil inside the pipe near its lower end,
so as to allow the pipe to be forced down by hammer blows at the top.

The most satisfactory method, however, and that generally adopted,
involved the use of a pile-hammer. For this purpose a McKiernan-
Terry, No. 2, pile-hammer was placed on the job and operated either
by air or by steam at a pressure of about 80 lb. per sq. in. In order to
facilitate driving, the blacksmith on the work fitted the drain pipes with
solid points made from pieces of old steel shafting, approximately 3 in.
in diameter. The points were welded inside the ends of the pipes, with
results as shown in Fig. 19. This figure shows also the perforations
in the pipes. It is to be admitted that the perforations were probably
clogged with clay to a certain extent as the pipes were sunk. On the
other hand, any considerable hydrostatic pressure near the drain pipes
would soon open the perforations.

A simple wooden cap prevented injury to the pipe, and the flush-
joint pipe did not prove so weak at the joints as to buckle or otherwise
cause trouble. In a number of instances, however, boulders proved
to be impassable obstacles, and in such cases pipes were driven at
near-by locations. In each instance the drain pipes were sunk through
the weep-holes of the original structure. The process was necessarily
slow. Sometimes it required 12 hours, and even 24 hours, to drive a
pipe to a depth of 20 ft.

In general, the deep drain pipes were encased in grout or concrete
through the footings and for a depth of about 18 in. below the footings,
with the object of keeping the water percolating through the deeper
foundation soil separate from leakage water manifesting itself imme-
diately under the footings, and finding its origin perhaps through the
cut-off wall. Thus the deep drains are enabled the better to perform
their function as detectives, so to speak. The upper ends of the drain
pipes were threaded and left exposed for at least 3 in., so as to allow
the pipes to be capped or to be connected to pressure gauges. The
pipes, of course, were thoroughly cleaned out after they were driven.
Yery few of the deep drains showed leakage after the work of recon-
struction was completed, and none showed any considerable leakage.


Further, under the adopted system of drainage, the two up-stream
pairs of weep-holes in each bay were plugged with concrete in con-
nection with the pouring of the triangular section of concrete at the
junction between deck and footings. About one-half of the remaining
weep-holes — distributed as uniformly as practicable — were left open.

As the reservoir refilled during the latter part of the work of recon-
struction, leakage appeared at weep-holes in Bays 51 and 36. At first
it was thought that possibly the leakage at Bay 51 was caused by
ground-water, perhaps from the water being fed into drill hole 10
(Plate IX) which was being sunl'i at the time. However, by coloring

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