American Society of Civil Engineers.

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

. (page 74 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 74 of 167)
Font size
QR-code for this ebook

that the center line of the arch ring and the line of pressure will coin-
cide very closely for all depths.

It will be necessary to use a three or five-centered curve for the form
ribs, and this will entail an extra expense, as the circular form ribs will
have to be made in sections also. Furthermore, with a horizontal sec-
tion circular, the up-stream and down-stream edges will be concentric,
with a uniform radial thickness of arch ring. This will result in the
normal section of the arch ring being much thicker at the buttresses
than at the center, giving it greater stability.

The author has provided a light but very effective method of stiffen-
ing the buttress walls, and it is worthy of more general use.

Gardner S. Williams,* M. Am. Soc. C. E. (by letter).— No dis- Mr.
cussion of the multiple-arch dam would be complete without mention
of the famous Meer Alum Dam,t near Hyderabad, India, built about
ISOO, which is made up of a series of full-centered vertical brick arches
with radii varying from 70 to 147 ft. These arches are arranged in the
arc of a circle, having a crest length of about i mile and a height of
about 40 ft. This structure appears to be the dean of multiple-arch

On September 1st, 1902, the writer tabulated the bids received by
the Ithaca Water Works Company, and a few days later awarded the
contract for the construction of the so-called Six-Mile Creek Dam, near
Ithaca, N. Y. This dam embodied in its design a section of the inclined
multiple-arch type, and is believed to be the first structure of the kind
for which a contract was ever let.:{:

Prior to the design of this dam, which was made by the writer in
the early part of 1902, plans and specifications for an inclined multiple-

* Ann Arbor, Mich.

t Transactions, International Engineering Congress, 1915, Waterways and Irriga-
tion, p. 719.

t The Six-Mile Creek Dam was discussed in Transactions. Am. Soc. C. B., Vol.
LIII, and the multiple-arch section is particularly described on page 195 thereof.


Mr. arch dam were prepared by Henry Goldmark, M. Am. Soc. C. E., for
■ the Pioneer Power Company, of Ogden, Utah,* the suggestion of this
design being due, the writer is informed, to George H. Pegram, Presi-
dent, Am. Soc. C. E. The inclination of the arches was about 25° from
the vertical. Subsequently to the work of Mr. Goldmark and prior to
1902, the multiple-arch dam was investigated by James H. Fuertes,
M. Am. Soc. C. E., and by the late Emil Kuichling, M. Am. Soc. C. E.,
independently, but neither of their designs came to the constructive

Since the construction of the Six-Mile Creek Dam, dams of the
multiple-arch type have been designed and built under the supervision
of the writer in Michigan, Minnesota, and Iowa.

The principles of design laid down by the author differ somewhat
from those followed by the writer, and appear to transgress the rules
of economical design, since the structure, as described by the author,
does not follow the equilibrium curve. As is well known, the curve of
equilibrium for an arch under uniform normal pressure is a circle,
and, therefore, since the inclined arch may be conceived as made up
of an infinite number of infinitesimally thick horizontal arches,
the horizontal section of the arch should be a circle and not an ellipse,
as seems to have been contemplated by the author. A section of the
arch normal to the axis will then be elliptical, and the inclination
should be chosen so that this ellipse approaches most closely to the
hydrostatic arch for the condition of most constant load. When de-
signed in accordance with these principles, the arch is subjected to the
minimum of bending stress, and requires the least material. An
increasing thickness of the arch section to the bottom of the dam is
hardly necessary, as over the lower one-third, in ordinary cases, the
beam action is such as to relieve to a considerable degree the arch
action, so that the shell may very properly be made of uniform or even
diminishing thickness from the point of one-third the height down.

J!^one of the dams built by the writer has been on a rock foundation,
and although it is necessary to be assured of the carrying capacity of
the material under the buttresses, this does not preclude the use of
such structures on sand, gravel, or clay.

Mr. George W. HowsoN,t Assoc. M. Am. Soc. C. E. (by letter). —

HowRon. -j^^ Jorgensen has presented a very excellent paper on the design and

construction of multiple-arch dams. The method of working out the

design shows much careful study, and appears to be the last word in

the design of dams of this type.

There are several phases of the paper, however, which hardly seem
to be in accord with the writer's experience, and there is a general

* Transactions, Am. Soc. C. E., Vol. XXXVIII, pp. 290 to 297.

t San Francisco, Cal. "*


tendency to compare the multiple-arch with the rock-fill dam. These Mr.
are two entirely different types, each having its own advantages. Howson

In general, it may be stated that a rock-fill dam may be built on
any foundation suitable for a multiple-arch structure; but the reverse
is not true. As an example, a rock-fill dam is best adapted to a site
where the underlying bed-rock across the floor of the canyon is covered
by, say, from 20 to 30 ft. of over-burden. Here it may be necessary to
carry only the thin cut-off wall down through the overlying material
into solid rock; and the great bulk of the dam, the loose drop fill,
may be built on the over-burden. It must be known, of course, that
the foundation under the rock fill is suitable, though it may not be a
foundation on which the buttresses of a multiple-arch dam should be
built. Mr. Flinn has brought out this point in comparison with a
solid masonry dam, but liis comparison of economy relative to the
foundation is even more applicable in the case of a rock-fill structure.

It is then seen that the problem of choice between rock-fill and
nmltiple-arch types presents itself only at a site where a good founda-
tion for the buttresses is exposed, or is found within a limited depth
of excavation. Here the final economic balance may be in favor of
either type, depending on the particular cross-section of the canyon,
the transportation problem, and the general topography of the site.
The writer, however, believes that, in a majority of cases where it is
possible to consider a choice of types, the relative cost of a multiple-
arch dam and a rock-fill dam will be in favor of the former.

Mr. Jorgensen is correct when he states that the water-tightness of
a dam depends to a much larger extent on the quality of the concrete
in the arches than on the thickness of the latter. The Strawberry
Eock Fill Dam, built recently by the Sierra and San Francisco Power
Company, may be said to be bottle-tight, although it depends on a
reinforced concrete apron only 18 in. thick, imder a 130-ft. head, which
thickness tapers to 9 in. at the crest. A multiple-arch dam should be
built only under the most competent engineering supervision. A few
failures of dams of this class due to poor construction would condemn
all such structures, in the public opinion.

L. E. Jorgensen,* M. Am. Soc. C. E. (by letter). — Throughout Mr.
the discussion there has been a tendency to compare the cost of
multiple-arch dams with that of gravity dams. As a general rule, it
can be said that the multiple-arch dam is the cheaper, the more ex-
pensive the building material. For dams of moderate height, it is
probably always cheaper, but not for heights of more than 130 ft.
The only sure way of finding out is to estimate the cost of each type
for any given location.

The system of struts has not been designed to help to prevent set-
tlement or sliding of the buttresses; it cannot do that, and it is not

*San Francisco, Cal.


Mr. necessary, for, in these particular cases, the rock foundation is pos-
Jorgensen. gjj^iy twice as strong in shear and crushing as the concrete in the
buttresses. The struts are to stiffen the buttresses, and to transmit
any vuiequal arch thrust which normally might exist at any place,
due, for instance, to the adjacent arches being built, during periods
of different temperatures, into the side-hill where the struts are an-
chored. The two struts close to the arch (Fig. 4) will also transmit
to the side-hill the abnormal unbalanced thrust which might occur
should one arch accidentally fail, thereby preventing a complete col-
lapse of the dam. Should the struts fail at the same time as the
corresponding arch — which is probable — the unbalanced thrust would
have to be taken up in tension by the steel in the line of struts on
each side of the break. The eight 1-in. square rods should be able
to hold 400 000 lb., as there are no laps between the knee-braces belong-
ing to the different buttresses. The triangular girder mentioned on
page 867, simply stiffens the up-stream ends of the buttresses so that
they will stand up between points of application of the struts. The
struts then transmit the unbalanced thrust to the side-hill, either by
tension or compression, or by both.

Mr. Blaekwell's criticism as to spillway capacity has been answered
by Mr. Duryea. There occurs no precipitation of such quantity as
to cause flood when the reservoir is full. During the season of maxi-
mum ice pressure — that is, in late winter or early spring — Gem Lake
Reservoir will be nearly empty. Agnew Lake Reservoir is to be used
as a forebay, and the daily fluctuations will break up the ice; there-
fore, it has not been deemed necessary to design these dams for ice

In comparing the three forms of multiple-arch dams mentioned by
Mr. Flinn, the writer regards the one with the sloping up-stream face
as the best and only one to consider. Li this construction, the water
pressure acting on the sloping face is the largest single factor con-
tributing to stability, and this factor would be absent in a dam having
a vertical face.

As to the leakage of different types of dams, this will depend much
more on the care with which they are constructed than on their type
and design.

The writer agrees with Mr. Scheidenhelm, that, after all, in the
case of multiple-arch dams, one must, for perfect safety, insist on an
unyielding foundation.

As an example of the influence of the time factor, let it be as-
sumed that an arch is subjected suddenly, or during an interval of
a few hours, to a temperature drop, penetrating through the arch
material, of, say, 80°, sufficient to cause a maximum tension in the
concrete of, say, 300 lb. per sq. in. If this same temperature drop
of 80° was brought about during an interval of 20 or more days, the


tension in the concrete would only amount to about 100 lb. per sq. in., Mr.
and, naturally, the tendency to form cracks would be much less than -^oi'sensen.
in the first case. Daily temperature changes will hardly penetrate
the arch sufficiently to have any practical effect, but seasonal tem-
perature changes will, and here is where the action of the time factor
tends to lessen the tendency to form cracks by making the concrete
stretch more for the same stress, because the change has been slow.
The publication by Mr. F. E.. McMillan, referred to on page 859,
gives detailed results of numerous tests of this nature, and is very

The multiple- arch dam at Ogden, Utah, referred to by Mr. Weg-
mann and Mr. Williams, is of xmusually conservative design, and it is
no wonder that the bid price was only from 12 to 15% lower than
the bid price on an ordinary gravity dam for the same site, as Mr.
Goldmark's design surely has a very much larger factor of safety than
the ordinary gravity dam.

Mr. Douglas says, "The unit price of masonry for the multiple-
arch dam is out of all proportion to that of the gravity dam, as
illustrated by the price of $22 per cu. yd. given by the author." This
high price, however, is due to the high cost of the material at this
particular place (Gem and Agnew Lakes), and not to the type. A
multiple-arch dam in an accessible location, and where conditions are
ordinarily favorable, can be constructed for $13 per cu. yd., and this
price may possibly include the steel reinforcement.

Mr. Douglas is correct in stating that frictional resistance can
only act after the section has failed under shear. One should rely
on shear or friction. The maximum unit shear of approximately
106 lb. per sq. in. existing at Elevation 80 below the crest can hardly
be considered excessive. In this case the area of the arch ring has
been added to the area of the buttress proper, as both must act to
resist shear at the same time. The fact that there is a very large
vertical load should be of some comfort, as it improves the ability of
the buttresses to withstand actual shear.

The stress on the buttresses is given between two limits on pages

865 and 866. The following formula (which can be found in some

handbooks) gives the maximum stress on any horizontal plane more

definitely :

m — 1

Ideal maximum stress = X unit vertical compression

2 wi

H V(unit vertical compression)^ + 4 (unit horizontal shear)^

2 m

(m = Poisson's ratio ^= 5 for concrete). Then, at Elevation 80,

Ideal maximum stress = 0.4 X 192 +0.6 \/l922 + 4 X 106^
= 76.8 + 171.6 = 248.4 lb. per sq. in.


It is true, as pointed out by Mr. Nishkian, that the stress produced
by the normal comiwnent of the arch thrust on the buttress is greater
than the stress (317 lb.) produced by the resultant. This, however, is
the case only toward lower elevations, where the "column" is very short ;
the normal component of the arch thrust diminishes rapidly toward
higher elevations. The ideal maximum stress of 248.4 lb. per sq. in.
diminishes less toward the upper system of struts; and, although the
column ratio is about 20, there might be, and very likely is, some
column action due to the ideal maximum stress.

Both Mr. Nishkian and Mr. Williams point to the elliptical arch as
being the most correct form. Thus far, the writer has contended that
the additional work of making a three-centered arch would outweigh
any possible gains; however, he will soon see how the elliptical arch
works out on a multiple-arch dam structure with the construction of
which he is connected.

The writer cannot agree with Mr. Williams that the arch can be
of uniform, or diminishing, thickness downward from the point of
one-third the height of dam. Even though the stresses may not be
excessive, the penetrating power of water increases with the depth,
and, therefore, the thickness of the wall should increase logically. It
must be kept in mind that even a sweat on the down-stream face will
produce icicles in cold weather, and it is not yet known to what extent,
if any, this will damage good concrete. It is known that it damages
poor concrete. Good concrete is fairly water-tight, and gunnite is
remarkably so, but, nevertheless, it requires great care to construct a
water-tight structure. Some of the arches on the Gem Lake Dam are
absolutely tight, some of them sweat in places, and a few drip in
places. A few small springs have formed behind the dam, and a
trickle of water comes under one arch; but, all told, the total leakage
is very small.

1^'"' Some tests were carried out in the Laboratory of the University
of' California on the water-tightness of plaster "shot" on the dam face
with a cement gun. Several plaster slabs, from § to 1^ in. thick, made
at Gem Lake, were tested with water pressures ranging from 700 to
1 600 ft., for several hours, with no moisture coming through the
slab. One 1-in. slab held a head of 1 610 ft. for 2^ hours without
showing moisture, then the water pressure was raised gradually to
3 400 ft., and the specimen broke in bending, having leaked a little
just before breaking.

Mr. Howson states the case correctly when he says that a rock-fill
dam is best adapted to a site where the underlying bed-rock across the
floor of the canyon is covered by, say, from 20 to 30 ft. of over-burden,
as it is only necessary to carry a thin cut-off wall down through the
overlying material.

The author thanks those participating in the discussion for the
interest shown.




This Society is not responsible for auy statement made or oijiuion expressed
in its publications.

Paper No. 1397



By F. W. Scheidenhelm, M. Am. Soc. C. E.

With Discussion by Messrs. J. W. Ledoux, J. K. Finch, P. P.
RuTENBERG, Fred F. Moore, W. S. Downs, H. L. Coburn, H. F.
Dunham, Orrin L. Brodie, William Cain, Charles E. Gregory,
Kenneth C. Grant, L. E. Jorgensen, Edward Wegmann, Irving
P. Church, M. M. O'Shaughnessy, Joel D. Justin, Ross M.



This paper describes the reconstruction and strengthening of the
Stony River Dam, which failed on January 15th, 1914. After the
failure, on investigation of the fomidation conditions and the parts
of the structure which remained intact, it was found that those por-
tions required strengthening and revision in a number of important
features. The lessons indicated by the results of the investigation are
significant. The dam, even as reconstructed and strengthened, is
founded largely on clayey soil, thus presenting most difficult founda-
tion conditions. It is about 50 ft. high above stream level and 85 ft.
between the lowest part of the cut-oif wall and the top of the parapet.

A brief resume of the previous history of the dam, including the
causes of its failure, is first given, together with a statement of the con-
trolling geological and foundation conditions. Then follows an exposi-
tion of the treatment of the more important problems involved in the

. r,-tt : ,; r, ■ • ^ ^

* Presented at the meeting of March 21st, 1917.


The spillway capacity was increased approximately to 1 840 cu. ft.
per sec. per sq. mile. The reasons for providing such imusually large
capacity are given, and also a method of determining the absorption
effect of a reservoir in smoothing off the peak of a flood.

The most important problem to be solved was that of giving the
original structure a sufficient margin of safety against sliding. The
results of tests of frictional resistance and shearing value of various
soils, principally clays, are stated, and also the various schemes con-
sidered for increasing the resistance to sliding. The method adopted
is believed to be new, and consists in the use of anchoring walls extend-
ing to a considerable depth into the underlying foundation soil, and,
in effect, utilizing the weight of that underlying soil as well as the
resistance (to horizontal movement) of the soil immediately down
stream from the structure. The details of construction are unique.
The same principles were applied, with important differences in detail,
at the new spillway which replaced that portion of the structure which
had failed or was wrecked.

Brief reference is made to the resistance of the structure — both in
its original state and as strengthened — to failure by overturning.

The bearing value of clayey soil is discussed, and the results of tests
made at the site are given. The anchoring wall added at the toe of
the original structure was utilized to decrease the pressures of the
foundation soil. The method of strengthening the original footings is
described. Pressure grouting was used to remedy other faulty footing

The means of cutting off percolation and leakage through the
foundation soil are described, also the provision made for drainage,
both of clayey soil and of shale rock, together with the controlling con-
siderations in these and similar cases.

The higher bulkhead portions of the original structure were housed
in by curtain-walls and roofs in order to prevent serious freezing in the
drainage system and to fulfill certain other functions as described.

Miscellaneous problems in the reconstruction and strengthening
involved the type of structure for the new spillway, the strength of the
original decks and buttresses, the underpinning of the portions of the
original structure immediately adjoining the new spillway, the con-
struction of a protecting and anchoring toe-wall at the old spillway, the


special method of constructing an anchoring wall at the heel at the
outlet-gate sections, and the provision of an outlet channel.

The reliable storage capacity of the reservoir was increased by 25
per cent. This increase was made permissible by the use of special
steel bars or pins acting as flash-board supports. These pins were
developed for the purpose, and automatically and reliably allow the
flash-boards to be swept off the spillway crests when the head-water
reaches a pre-determined level.

The materials, methods, and cost of reconstruction are discussed
briefly. Approximately, as much concrete was required for the recon-
struction as had been placed in the original structure.

The Stony River Dam and Reservoir have again been in service
since May, 1915, and the observed results of the reconstruction and
strengthening are stated in connection with the corresponding features
of design.

Reconstruction of Stony River Dam.

The failure and reconstruction of the Stony River Dam are of
interest, not because of the height or dimensions of the structure, but
because it is largely through failures and mistakes that engineers must
learn the limitations of previous methods of design and construction
and gather hints as to new methods to be used. Furthermore, in the
case of this structure, the extremely difficult foundation conditions
made the problem of providing a safe dam at that site one requiring
more than usual thought and care. Finally, much interest attaches to
the methods used to make safe the existing structure, that is, that
portion which remained intact after the failure — it being noted that a
comparatively small section of the dam actually failed.

It seems proper at this point to emphasize the fact that, until a
structure has actually failed, its ''safety" is relative. Seldom do any
two engineers have exactly the same conception or measure as to the
factor or margin of safety possessed by a given structure. Especially
is this true in the case of dams.

As a preface to the description of the reconstruction and the prob-
lems involved therein, the following brief review of the previous history
of the dam is pertinent. The details have been gathered from various
sources, and are based on the best information available to the writer.



In 1911 the West Virginia Pulp and Paper Company, owner of the
dam, determined to investigate the feasibility of constructing on the
head-waters of the North Branch of the Potomac River a storage
reservoir by which to increase the flow of the river at its Piedmont
pulp and paper mill at Luke, Md., near Piedmont, W. Ya. Its require-
ments of water for manufacturing purposes were such that during low-
water seasons the normal flow of the stream was at times inadequate.

After an investigation of the available sites on the drainage area
above the Piedmont mill, the owner chose the water-shed of Stony
River as offering the best conditions for the construction of a storage
reservoir. Stony River — which at the site of the dam drains only 11.4
sq. miles, and is therefore only a creek — lies entirely within Grant
County, West Virginia, flowing into the North Branch of the Potomac
River from the south a short distance above Schell, W. Va. The mouth
of Stony River is approximately 26 miles up stream from the Piedmont
mill. It is only near its upper end that the profile is flat enough to
allow the construction of a reservoir at a practicable cost per unit of
water stored.

Two sites examined in the upper reaches of Stony River were
abandoned, because of interference with a seam of commercial coal in
one case and the presence of quicksand in the over-burden in the other.
The third and final site was chosen with the approval of Edward
Wegmann, M. Am. Soc. C. E., who had been engaged by the owner as
Consulting Engineer for the purpose.* This site is approximately
3 350 ft. above sea level.

An examination of the foundation conditions, deemed at the time

Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 74 of 167)