American Society of Civil Engineers.

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

. (page 88 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 88 of 167)
Font size

they cannot be utilized in estimating the resistance of the dam against
sliding, as previously noted.

It has been stated previously that the
unit passive thrust at the free surface of
the earth was not zero. To find approxi-
mately its amount at the point, D, let us
find the thrust on a vertical plane extend-
ing 1 ft. below D, Fig. 30, by repeated use
of Equation (4). The minimum value of
E, which is the true thrust, is found to ^

be 2 700 lb. per sq. ft. It corresponds to i i

a = 61°, nearly. By consideration of the ^ig. 30.

total thrusts on the areas, FJ and FG of Fig. 28, it was estimated
that the unit thrust at G was about double that at F. The unit thrust
is not uniformly increasing, but varies about as shown by the little
arrows in Fig. 30.

The unit compression on FG, Fig. 28, can be supposed to follow
the same law, though, for simplicity, it can be regarded as uniformly
increasing from a stress of 2 000 lb. per sq. ft. at F to double this,
or the maximum allowable, 4 000 lb. per sq. ft., at G. This gives an
allowable average stress on FG of 3 000 lb. per sq. ft., as used above.

It is seen from this investigation that the author is warranted in
using a larger average safe stress on FG tlian the specified 2 000 lb.
per sq. ft.

It may be observed that, if the force tending to cause sliding of
the dam is entirely resisted along the plane. AB, then only the active

1058 DISCUSSION : reconstruction of stony river dam

Mr, pressure of the earth to the right of FC is exerted. If, then, the resist-
*'°' ance along AB is supposed to diminish, first the active thrust of the
mass, CFML, will be overcome and then more or less of the full passive
thrust of this mass will be exerted, the amount always being less than
that which would exceed the safe compressive stress of the earth, or
would cause flow of the material along CF. Consequently, it does
not appear that the passive resistance of any pile of earth placed below
a dam to help resist sliding will ever come into play unless there is
a slight movement of the dam down stream, since earth is not a
rigid body.

A year or so ago, the writer's attention was called to the case of
a high dam, where a large pile of earth was placed below and against
the dam, to prevent possible sliding. He stated then that the coeffi-
cients of friction and cohesion were both needed to investigate fully
the extra resistance to sliding supplied by the earth. He is glad
of this opportunity to offer the full solution for such cases. Whatever
the free surface contour, the graphical method of Fig. 29 offers a quick,
approximate solution, though it is well, for the reasons stated, to apply
a factor of safety to the result, or otherwise to use conservative
values of c and /.

No better illustration could be given of the need of "comprehensive
experimentation to determine the coefficients of friction and cohesion,"
than is afforded by this constantly recurring problem of estimating
the resistance to sliding of dams and retaining walls. The writer is
firmly convinced that a correct solution can only be attained by the
use of the theory of coherent earth.

In conclusion, the writer wishes to state his appreciation of the
very thorough and painstaking manner — down to the minutest details —
in which the author has done the work of reconstructing the Stony
River Dam.

Mr. Charles E. Gregory,* Assoc. M. Am. Soc. C. E. — A paper of this

regory. y^^^ -g unusually interesting, because it deals with the failure of a

structure. The speaker believes that one can always learn very much

more from the failure of a structure than from any discussion of what

one believes ought to happen in accordance with theories.

The author has given a most lucid and well-written description of
how he corrected the mistakes of the first dam, and has most ingeniously
met the difficulties of the problem while utilizing the portions of the
old dam which remained after the failure.

The reconstruction of this dam is an interesting subject, because
it deals with a number of uncertainties pertaining to both the site and
the type of dam, and uncertainties are always interesting. About the
only relatively definite elements in the whole problem are the strength

• Mt. Klsco, N. Y.

DISCUSSION : RECONSTKUCTION OF STONY RIVER DAM 1059

of concrete and steel. Nearly everything else is open to more or less Mr.
uncertainty. Notwithstanding this, the speaker believes all will admit ''^^"''y-
that any large dam should be built so that its safety cannot be ques-
tioned, and one cannot say that any structure is safe until the uncer-
tainties have been eliminated.

In this case the type of dam and the site having been determined,
there remain a great number of very uncertain features to be con-
sidered. The author has pointed out most of them, and has told how
they were removed.

In any hollow dam, the uplift is the great doubtful feature and the
great enemy of a dam of this type. There is no way of knowing to a
certainty just what the uplift pressure is and how it is distributed. In
designing a dam of this type, certain assumptions are made as to this
pressure. It is assumed that if uplift exists, it can be controlled and
limited by the weep-holes. Therefore, numerous weep-holes are placed
in the foundation, and are supposed to relieve the upward pressure,
and limit it to a relatively small amount. The cut-off wall must be
effective, and must cut off the water from the impounded reservoir.

It is common experience that it is exceedingly difficult to build a
cut-off wall, of any great depth and length, which will be absolutely
water-tight. A concrete wall of great length will shrink and crack, in
spite of anything that can be done to prevent it. Construction joints
form weak planes, if not actual joints. All these conditions are favor-
able to leakage, even when one is successful in securing impervious
rock foundation. Some water is boimd to get through, and then it is
bound also to show itself under the bottom of the dam.

When weep-holes are provided and are effective, they will allow this
water to escape and relieve the pressure, but if they are not effective,
the upward pressure will obtain. If they are effective, however, the
water flowing through the soil under the bottom of the dam is a new
source of danger.

A safe structure should not be menaced by flotation or by under-
mining. The flowing water from the weep-holes is certain to carry
with it more or less soil, especially if it consists of fine, light grains;
and undermining of the bottom of the dam will gradually occur.
Eventually, the flow will become greater, heavier grains of soil will be
carried, and ultimately there will be sufficient undermining to cause
settlement and cracking, increasing the upward pressure so as to cause
the dam to overturn or slide.

Probably this dam started to fail by overturning due to under-
mining. As soon as the tendency to overturn developed tension at the
upper toe sufficient to open a joint and admit the water, the upward
pressure floated the whole section off its base. The failure was probably
a combination of overturning and sliding.

1060 DISCUSSION : reconstkuction of stony river dam

Mr. The speaker has not made any check computations whatever of the

regory. Qpjg.jjjjj| structure, but from the author's statement of the various forces
acting on this dam it would seem that, unless high coefficients of fric-
tion are assumed for the base, it would not have been safe from sliding.
There seems to be no good ground for assuming high coefficients for a
soil which is saturated with upward flowing water to the weep-holes.

In the speaker's opinion, the initial mistake was made in selecting
a dam of this type for such a site. Although the speaker has not
visited the site, and knows nothing about it, except what is disclosed
in the paper, there appears to be in the vicinity considerable clay,
gravel, and sandy clay which would be excellent material for an earth
dam. An earth dam could have been built on this site economically.
and would have eliminated nearly all the very uncertain features which
obtain with this hollow concrete dam ; the speaker believes an earth
dam is the proper type for this site.

Another interesting part of the paper is the discussion of the spill-
way capacity. The author certainly has increased the capacity of the
original spillway to a point beyond all precedent. The speaker believes
that any spillway should have a very safe capacity. It should be large
enough to care for the very largest storm that can possibly be con-
ceived of, but the speaker believes that, in this case, the author has
surely gone beyond the necessary limit. A flood as great as that pro-
vided for by the new spillway certainly could be produced only by a
rainfall far greater than the maximum curve of rainfall for the eastern
United States, as indicated by A. N. Talbot, M. Am. Sec. C. E. This
curve is supposed to cover the greatest rainfall in this part of the
country.

On page 930, the author cites certain storms of great intensity for
24 hours, hut docs not give the rate for the critical time for the water-
shed. Probably a rainfall of 2 or 3 hours would be much more nearly
the critical period for this water-shod. Fig. 31 and Table 9 present
data from the automatic rain gauge in Central Park, IS^ew York.
recorded continuously since 1869, and are self-explanatory. The
maximum curve for Greater New York gives rates materially outside
the Talbot curve, and is shown on Fig. 32.

On the illustration of the new spillway (Plate X) is shown a
joint between the buttress wall and the slab covering the bays. This
joint is filled with three-ply tar-paper. x\pparently, such paper was not
used in the old dam, or it would not have shown the bond that it did
along those joints.

It would seem to the speaker that the placing of tar-paper in the

joints of this dam, or in any dam, would be very poor practice. In the

speaker's opinion, tar-paper could not be classed as a permanent

■ material which would stand up under wet and dry, freezing and thaw-

DISCUSSION : KECONSTKUCTION OF STONY RIVER DAM

1061

,0 5 10

INTENSITY DIAGRAM

SHOWING MAXIMUM RAINFALL INTENSITY CURVES

WITH OBSERVED INTENSITIES

BOROUGH OF MANHATTAN

CITY OF NEW YORK.

o Indicates obserreil intensities

120
4

Mr.
Gregory.

•15-Year Curve

o Greatest Rainfall

T_ 260

^ T+18

2 22V4-Year Curve
2d Rainfall
r_ 220

■'■ r-t-18
II

15- Year Curve
, 3d Rainfall
1 7_ 185

r+17
III

.5 10 15 20

40 50 60 80

Duration, in Minutes

Fig. 31.

10- Year Curve

Interpolated

T_ 150

-* T-t-16

IV

7/i-Year Curve
6th Rainfall

■' r+16

V

5-Year Curve

9tli Rainfall

._ 130

•^ r+17

120 YI

1062

DISCUSSION : RECONSTRUCTION OF STONY RIVER DAM

Mr. TABLE 9. — Maximum Eates of Rainfall, in Inches per Hour, as

Gregory.

Eecorded by Rain Gauge in Central Park, New York City,

Date.

Duration, in Minutes.

15

40

50

60

100

120

Aug. 5,
July 3,
July 31,

'78

'92

'10

May 30,
Oct. 11,
Aug. 6,

'12

'71

'95

July 12,
Aug. 20,
July 28,

'80

•93

'03

Aug. 5,
Sep. 4,
July 31,

'78

'13

'10

July 28,
May 30,
July 28,

'13

'13

'03

Aug. 5,
Aug. 20,
July 26,

'02

'93

'75

Aug. 5,
Sep. 4,
July 28,

'78

'13

'13

Aug. 20,
May 22,
July 28,

'93

'81

'02

July 26,
Aug. 5,
May 30,

'75

'02

'13

Aug. 5,
Sep. 4,
Aug. 20,

'78

'13

'93

July 28,
Aug. 5,
July 38,

'02

'02

'13

July 26,
May 33,
Oct. 1,

'75

'08

'13

Aug. 5,
Sep. 4,
July 38,

'78

'13

'02

Aug. 20,
July 28,
Aug. 5,

'93

'13

'03

Aug. 19,
July 26,
May 23,

'04

'75

'81

Sep. 4,
Aug. 5,
July 28,

'13

'78

'03

00

DISCUSSION : RECONSTKUCTION OF STONY EIVER DAM 1063

TABLE 9.— (Continued.)

Date.

July 28, '13..

Oct. 1
Aug. 20,

Aug. 19,
Aug. 5,
Sep. 23,

Sep. 4,
July 28.
Aug. 5.

July 28,
Oct. 1,
July 6,

Aug. 23,
Aug. 2,
Aug. 19,

Sep. 4,
July 28,
July 28,

Aug. 5,
Oct. 1,
July 6,

Sep. 23,
Aug. 23,
Aug. 19,

Sep. 4,
July 28,
Oct. 1,

Sep. 23,
Aug. 19,
July 6,

May 21,
July 5,
June 29,

Oct. 1,
July 5,
Sep. 23,

Oct. 9,
Aug. 19,
Oct. 4,

Sep. 4,
June 29.
Oct. 23,

July 5,
Sep. 23,
Oct. 9,

Sep. 4,
Oct. 4,
Aug. 19,

Aug. 4,
Aug. 23,
Aug. 23,

'13.
'93.

'13
'02.

'78.

'13.
'13.
'96.

'97.
'93.
'04.

'13.
'02.
'13.

'78.
'13.

'82.
'97.

'04.

'13.
'13.
'13.

'82.
'04.
'96.

'83.
'01.
'03.

'13.
'02.

'82.

'03.
'04.

'77.

'78.
'03.
'12.

'01.
'82.
'03.

'78.
'77.
'04.

Mr.
Gregory.

Duration, in Minutes.

10 15 20 30 40 50 60

100 120

1.88
1.44
1.36

1.29
1.23

1.20

1.20
1.20
1.13

1.32
1.31
1.21

1.10
1.08
1.07

1.02
0.96
0.96

1064

DISCUSSION : RECONSTRUCTION OF STONY RIVER DAM

Mr.
Gregory.

ing, as should have been provided in a permanent structure. These
joints, though they should have been made free to move, should have
had some more permanent methods of stopping leakage. Even a plain,
smooth joint, without any filling, would have been far better than the
tar-paper. Some form of metal tongue across the joint would have
been still better.

When designing the Ashokan and Kensico Dams, of the Catskill
Water Supply System, the speaker had occasion to divide these gravity
masonry dams into sections for the purpose of taking up the contrac-
tion due to temperature changes. In the Ashokan Dam the joint was
formed with concrete blocks, as shown on Fig. 33, and the leakage
through the opening of the joint was prevented by constructing a 6-in.
offset, with a face very accurately finished in a plane parallel to the
longitudinal axis of the dam, so that when the joint opened the con-
crete surfaces would slide on each other without opening.

INTENSITY DIAGRAM

MAXIMUM RAINFALL INTENSITY CURVE

FOR GREATER NEW YORK

January, 191i

o Indicates observation taken at the Central Park
Rain Gauge-Manhattan, Aug. 6, 1878.

A Indicates observation taken at the New Brighton
Kain Gauge -Richmond, Oct. 1, 1913

Maximum for Greater
New York

Talbot-Maximum for

^^^" Eastern U.S.

6 10 15 20

40 50 60

Duration, in Minutes

Fig. 32.

Just down stream from this offset, as shown, is the drainage well
to collect and carry to a lower gallery whatever leakage might come
through the joint. Experience has shown the leakage through the
fifteen joints of this dam, aggregating about 1 500 lin. ft. of exposed
joint, to have been about 450 000 gal. per day as a maximum, and to
have been reduced to about 25 000 gal. per day on November 2d, 1916.

In the Kensico Dam, similar joints were provided with a copper
strip arranged diagonally across the sliding joint, so that, when the

DISCUSSION : RECONSTRUCTION OP STONY RIVER DAM

1065

1066

DISCUSSION : KECONSTRUCTION OF STONY RIVER DAM

Mr. joint opens, the copper strip (as shown on Fig. 34) will crumple
Gregory. j,^^]^gp than tend to pull out. These structures have been very effective
in all joints but two where there was considerable leakage at first, due
to a poor batch of concrete near the bottom of the well in one case,
and a contraction crack below the bottom of the well and copper
strips in two cases which admitted more water than the entire twenty-
two joints in the dam. After these cracks were grouted the leakage
was reduced to from 13 000 to 18 000 gal. per day for the 1900
lin. ft. of exposed joint.

COPPER STRIP AT
EXPANSION-JOINT

1 2 3 i 5ft.

I I I ( I M

Mr.
Grant.

Fig. 34.

Kenneth C. Grant,* M. Am. Soc. C. E. (by letter). — This paper
is very interesting to the writer, who made a careful examination of
the dam on January 16th, 1914, and agrees with the author's conclu-
sions as to its failure. The writer is impressed by the thoroughness
of the studies and designs for the repair of the dam, and cannot but
feel that it is unfortunate that the same care was not taken in the
original design and construction, when equal safety could have been
obtained at much less cost.

All readers of this paper must have been impressed by the large
spillway capacity provided in the reconstructed dam. The writer
was particularly pleased to see the retarding effect of the storage above

* Washington, D. C.

DISCUSSION : RECONSTRUCTION OF STONY RIVER DAM 1067

spillway level worked out by actually routing the assumed maximum Mr.
flood through the reservoir. The usual method of expressing spillway ''^^^^■
capacity, in terms of run-off per square mile of drainage area, does
not tell the whole story. If the reservoir has a large surface area at
spillway level, a rise of a foot in the water surface represents a very
considerable storage capacity; and the maximum flood that can safely
be taken care of by spillway and storage combined may be much
greater than in the case of a reservoir with the same spillway capacity,
but a small area of water surface. In the assumed maximum case
shown on Fig. 15, the maximum spillway discharge amounted to
about 1 600 sec-ft. per sq. mile, and the assumed flood causing this
outflow reached a maximum of nearly 2 600 sec-ft. per sq. mile. Thus,
the spillway capacity of 1 840 sec-ft. per sq. mile given by the author
really means that it would take a flood about two-thirds larger than
this to overtop the dam.

This reducing action, which every full reservoir exerts on a flood
wave, is the same in principle as the control effected by retarding
basins such as those which are to be built for the protection of the
Miami Yalley in "Western Ohio, except that, in the latter case, the
reduction is much greater, inasmuch as the entire capacity of the
basins is available for flood storage. The writer has noted that the
capacity of some of the flood-control reservoirs in Europe, in which
the conduits can be closed by gates, is correctly considered to be
the total capacity up to the elevation at which the spillway discharge
reaches the maximum outflow that can be safely delivered to the
channel below.

It may be of interest to explain briefly the method found to be
the simplest for routing floods through the Miami Valley retarding
basins, by applying it to the flood wave assumed by the author in
Fig. 12. In the method here to be described, the time interval is fixed,
and the rate of outflow is obtained by trial. The rate of inflow is
fixed by the adopted time interval, and the trial rate of storage and
elevation of reservoir surface are fixed as soon as the rate of outflow
is assumed. As compared with the method of flxing the elevation
of reservoir surface and finding the time interval by trial, this method
has the advantage of enabling one to pick the breaks in the inflow
curve, especially the peak, and the points where inflow equals outflow.

Fig. 35 shows a capacity curve of the Stony River Reservoir above
spillway level, derived from Fig. 10; also a curve showing spillway
discharges plotted against reservoir capacities, derived from Fig. 11.
Fig. 36 shows the inflow and outflow curves, and Table 10 illustrates
the method of recording the routing operation.

Conditions of starting similar to those given by the author, were
assumed, namely, that, at 7 a. m., the outflow is equal to the inflow.

1068

DISCUSSTOX : RECONSTRUCTION OF STONY RIVER BAM

Mr.
Grant

or 100 sec-ft. From Fig. 35, this corresponds to a capacity of
208 400 000 cu. ft., or an elevation of 136.2.

Take a time interval of 1 hour. At this time the rate of inflow
is 500 sec-ft. By trial, a rate of outflow at this time can quickly be

18

CURVES SHOV\[ING CAPACITY

1

OF Sf

'ILLWAY A

ND RESERVOIR

— /

Capacity

referred t

capa

3f spillway
reservoir-N.
:ities

4-

u
|l42

U

o

J

/

/

/

'^

3

s

^140

y

o
a
o

Capac
referi

ity of reserve
ed to elevati(

ir y

138

//

^

137

/

200 220 240 260 280 300

Capacity of Reservoir, in Millions of Cubic Feet

Fig. 35.

320

14 IS

log

8"^

CO

5

existing gives a total storage corresponding to the assumed outflow.
Thus, by assuming a rate of outflow of 150 sec-ft. at the end of 1 hour,
the rate of storage is 350 sec-ft., and the mean rate of storage is

DISCUSSION : RECOXSTEUCTION OF STONY RIVER DAM

1069

175 sec-ft. Multiplying this by 3 600 gives the storage increment, in Mr.
cubic feet, or 630 000 cu. ft. Adding this to 208 400 000, the storage *^''^°*-
at Elevation 136.2, gives 209 030 000 cu. ft., the total storage. Entering

14

12

olU

eg 8

CANE CREEK FLOOD OF MAY, 1901,

ROUTED THROUGH STONY RIVER RESERVOIR.

SHOWING REDUCTION DUE TO STORAGE ABOVE SPILLWAY.

Assumed Stony River tUscharg'e derived from Cane Creek
discharges by direct ratio of drainage areas. No flash-boards
on spillways. Reservoir full at beginning of flood.

I

I

iflow.^

1

/

iMa

-JMa

Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 88 of 167)