American Society of Civil Engineers.

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

. (page 89 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 89 of 167)
Font size
QR-code for this ebook


was enough clayey soil in the neighborhood to cover the bottom in
an up-stream direction from the core-wall, although it is stated in
the paper that there was not enough suitable material to build a
plain earth-fill dam.

• San Francisco, Cal.



DISCUSSION : RECONSTRUCTION OF STONY RIVER DAM 1071

Of course, the author was not concerned with what could have Mr.
been done in the first place when he was confronted with the task °''^^°^^°-
of repairing and strengthening the dam.

The author makes extensive studies for spillway provision, and
arrives at a high run-oS figure which seems to be justified. The
breakable flash-board support arrangement is a good feature, consid-
ering the fact that the treated pins were found to break within
narrow limits of load, that is, within between 4 and 4.6 ft. of head
above the base of the flash-boards.

Valuable information as to the frictional resistance and shearing
value of clayey soil and shale under various conditions is given in
the paper. The anchoring wall at the heel and the work of tying it
into the dam must have been among the most difficult details of the
construction.

The stresses used, 700 and 500 lb. per sq. in., are somewhat higher
than those to which the writer is accustomed, but, with good concrete,
should leave sufficient margin for safety.

The drainage system has evidently been a problem of no small
magnitude in this case, where the foundation is of such a character
that muddy water is likely to continue flowing through the drains
where these are placed very close to the reservoir pressure. The fact
that the vertical drain pipes only fill themselves, but do not overflow,
seems to indicate that there could not be very much uplift pressure,
although the writer does not understand why these pipes just fill but
do not overflow.

It is specified under Class C concrete that boulders should consti-
tute not more than 40% of the total volume, and it would be of in-
terest — at least to the writer — to know if the quantity actually placed
in this concrete approximated this figure, as he has always felt that
20% of "plums" was about the maximum that could be "thrown in",
unless there was considerable hand work. Concrete containing many
"plums" is not likely to be as water-tight as when a medium percentage
is used, say, 20%, which is not so small a percentage at that.

Edward Wegmann,* M. Am. Soc. C. E. — The speaker has been Mr.
greatly interested in this description of the repair and reconstruction esmann.
of the Stony River Dam. The conditions were very unfavorable, and
the author is to be complimented on the very thorough manner in
which the work was done, giving the dam large factors of safety
against all possible kinds of failure which might occur.

In order to correct some erroneous statements which have been
made — some of them in print — the speaker must state his connection
with the inception of this dam. He was engaged by the West Virginia
Pulp and Paper Company, in the fall of 1911, to visit the site of a

• New York City.



1073 DISCUSSION : reconstruction of stony river dam

Mr. proposed storage reservoir in the mountains of West Virginia, which
eg:inan . ^.^^ ^^^^ been selected by an engineer employed by the Paper Company
for this purpose. Owing to the steepness of the river beds in that
region, it was very difficult to find a suitable site for a storage reser-
voir, and the one finally chosen, after considerable investigation,
appeared to be about the only one available. The speaker was requested
by the Paper Company to give his opinion about this site, and to
recommend the type of dam which he thought should be adopted.

In accordance with this request, he made one visit to the site of
the reservoir, late in the fall of 1911, in the company of Mr. R. P.
Bloss, the engineer of the Paper Company. At that time, only a few
test pits had been excavated. They showed that the material over-
lying the rock consisted of yellow clay mixed with very fine sand,
and underlain by compact blue clay. The rock surface was found
to be at a considerable depth. The surface of the ground was cov-
ered with numerous boulders, and in most of the test pits boulders
were found at a certain depth below the surface. An alarming feature
was the fact that veins of some fine black material which looked like
coal appeared in some of the test pits. Analyses, made later, proved
that this material was not coal.

Both ]\Ir. Bloss and the speaker were much impressed by the
possibility that leakage under the proposed dam might occur through
the black veins in the clay, unless a proper cut-off was provided to
rock bottom, either by a masonry wall or by sheet-piling. The speaker
requested Mr. Bloss to dig additional test pits and to make soundings
with an auger, in order to determine, as nearly as possible, the posi-
tion of rock bottom. This work was done subsequently, and, early
in 1912, Mr. Bloss si;bniitted to the speaker a cross-section of the
valley of Stony River, showing the probable line of rock bottom.

Based on this cross-section and the data obtained by the test pits
and borings, the speaker reported to the Paper Company as follows :

1. — That the site selected for the reservoir would be satisfactory,
if the valley were closed by a well-constructed dam.

2.— That a masonry dam would cost more than $500 000, owing
to the great depth to rock, which is about 45 ft. in the center
of the valley.

o. — That an earth dam was out of the question, as no earth or
gravel, for mixing with the clay at the site of the dam.
could be found within a reasonable distance from the pro-
posed reservoir.

4. — That the only type of dam which could be built at the pro-
posed site, within a reasonable sum — say, from $150 000 to
$200 000 — was a hollow dam of reinforced concrete.



discussion: reconstkuction of stony river dam 10?3

After receiving the speaker's report, the Paper Company decided Mr.
to construct a hollow dam of reinforced concrete, and requested the ^^^™*''°-
speaker to prepare a contract and specifications for this work. Three
different companies which had had experience in this kind of con-
struction were invited to submit plans and bids for the construction
of a hollow dam of reinforced concrete, based on the cross-section
of the valley prepared by Mr. Bloss. On this plan Mr. Bloss, in
consultation with the speaker, had marked the least depth to which
the foundation of the cut-off wall would probably be excavated, and
the contract provided that if the foundation should go deeper, the
additional work involved should be paid for as an extra. At both
ends of the dam, where it was thought that the cut-off wall would
probably not go down to rock, sheet-piling was shown.

Four different plans* for constructing the dam of reinforced con-
crete were received, with lump-sum bids, ranging from $143 000 to
about $200 000, for building the dam to the depth shown on the cross-
section of the valley.

Only one among the bidders, the Ambursen Hydraulic Construc-
tion Company, of Boston, Mass., had had much experience in the
construction of dams of the proposed type. This Company had built,
at that time, more than seventy dams of this kind, ranging in height
up to 150 ft. All these dams were standing, and although one of
them — that at Pittsfield, Mass.,t the cut-off wall of which had not
been carried deep enough in a foundation of gravel — had been under-
mined at the center of the valley, the dam had merely sagged, but
had not been ruptured. As the hole made under this dam by under-
mining had been 20 ft. deep, 53 ft. wide, and about 50 ft. long, both
up stream and down stream, the structure had certainly shown remark-
able strengtli. It had been jacked up, a deeper cut-off wall had been
provided, and no further trouble had been experienced.

In view of the wide experience of the Ambursen Company in con-
structing hollow dams of reinforced concrete, the speaker strongly
advised the Paper Company: (1) to engage the Ambursen Company
to design the proposed dam: and (2), in case this Company should
not get the contract for building the dam, to employ one. of its expe-
rienced engineers or superintendents to be constantly on the ground
while the dam was being built, in order to insure that the plans of
the Ambursen Company would be properly carried out.

These recommendations were adopted by the Paper Comi)any, and
the speaker's connection with the Stony River Dam then terminated.
The speaker had never seen the plans finally adopted, imtil after the
partial failure of the dam in January, 1914.

* Engineering News, September 5th. 1912.
r Engineering Neics, April 1st, 1909.



1074 DISCUSSION : reconstruction of stony river dam

_ Mr. It appears that the westerly half of the dam was constructed by

■ the Webber Construction Company, the lowest bidder, and the easterly
half was built by the Ambursen Hydraulic Construction Company.
According to the plans shown to the speaker by the Paper Company,
after the partial failure of the dam, and according to various published
accounts, the cut-off wall was only 5 ft. deep below the floor of the
dam from the west end to Buttress 15, although the depth of water
at the latter point was more than 25 ft. This part of the core-wall
was founded on what appeared to be hard-pan, a tough clay which
was very hard to pick but became soft after being under water. For
the remaining length, the cut-off wall was carried down to what was
thought to be bed-rock. As the usual rule is to make the depth of
the cut-off wall at least half the depth of the water, the inconsistency
of the manner in which the dam was constructed was apparent, and
it is difficult to understand how the engineers in charge of the work
could have been satisfied with such a shallow cut-off wall, especially
as sheet-piling had been omitted at the ends of the dam, although
they knew that there were porous seams in the clay formation.

After the partial failure occurred, the Ambursen Company issued
a bulletin about "The Facts as to the Blow-out under the Stony
River Dam at Dobbin, West Virginia." In this pamphlet the Com-
pany stated that the plans submitted to them by the speaker "showed
sheet-piling under each edge as a cut-off", but that its Mr. Ambursen,
after visiting the site of the dam, "expressed disbelief as to the
possibility of driving sheet-piling, and recommended a cut-off trench
to be carried down into sound material." The pamphlet continued as
follows: "The test pits were at that time* examined jointly by the
owner's engineer, their consulting engineer, Mr. Wegmann, and Mr.
Ambursen."

The speaker cannot understand how this statement got into the
pamphlet, as it is absolutely incorrect. He never met Mr. Ambursen
and Mr. Bloss at the site of the dam, and, at the time mentioned, had
no connection with the work.

A few days after the dam had been ruptured, Mr. A. G. Hillberg,
representing the Engineering Record, visited the site. According to
the published account of his observations,! the core-wall had only
been carried down 5 ft. below the flooring of the dam, at the point
where the break occurred, although the depth of the water in the
reservoir at this point was more than 25 ft. Mr. Hillberg fotmd, at
the point of rupture, a "pervious seam of coal and sand with some
clay as a binder, about 8 in. below the level at which the cut-off wall
had been stopped." The seam was from :| to 6 in. thick, and about
4 ft. wide. This discovery gave a clear proof of the cause of the

• April, 1912.

t Engineering Record, January 24th, 1914.



DISCUSSION : RECONSTRUCTION OF STONY RIVER DAM 1075

failure of the dam, and as Mr. Scheidenhelm, also, has stated in his Mr.
report to the Public Service Commission of West Virginia that ^^™*°°-
"failure was caused by the undermining of the over-burden or soil
under the up-stream cut-off wall", the speaker thinks that this question
might be considered as definitely settled.

Mr. Scheidenhelm has not only repaired the breach in the dam,
but has made a number of important changes in its design and con-
struction which the speaker will discuss.

Increasing the Spillway. — As originally built, the spillway was
only 150 ft. long, and 3 ft. deep below the crest of the main dam.
In case of a severe freshet, the whole dam could have acted safely
as a spillway, if a suitable apron had been constructed on the down-
stream side.

Assuming the probable maximum freshet at 1 386 sec-ft. per sq.
mile — the figure given by the author for Cane Creek — the bulkhead
section of the dam, as originally built, would have had to pass a sheet
of water about 2.5 ft. deep. In all probability, the dam would have
been able to pass this water, but, as no apron had been provided for the
bulkhead section, the latter would have been gradually undermined.

In reconstructing the dam, Mr. Scheidenhelm increased the spill-
way capacity to about 1 840 sec-ft. per sq. mile. Although this may
seem unusually large, it made the dam very safe.

Increasing the Storage Capacity of the Reservoir. — By placing
flash-boards on the two spillways provided in the reconstructed dam,
Mr. Scheidenhelm raised the water level 3.5 ft., and thus increased
the reliable storage capacity of the reservoir by 25 per cent.

Resistance to Sliding. — The author made some experiments on
the frictional resistance of clay moving on clay, shale on shale, and
concrete on shale. In all these experiments, the coefficient of friction
appears to have been determined for materials in motion. The force
required to start the motion is known to be much greater than that
needed to keep a body moving. As sufficient data on these points are not
yet available, the author acted wisely in adopting conservative figures.

The original dam had withstood successfully for about 65 days
the pressure due to a full reservoir, before failure occurred on January
15th, 1915, and there was no indication that the dam had not sufficient
stability against sliding. As the weepers in the floor of the dam had,
doubtless, been closed by ice during this period, and as the reservoir
had not, at that time, been made water-tight by silting up, the dam,
in all probability, had been subjected to upward pressure.

Resistance to Overturning. — One of the advantages of a hollow
dam of reinforced concrete, having its deck on an angle of 45°, is
its stability against overturning. As the water rose in the reservoir
and finally over-topped the crest of the dam, the line of pressure was



107G DISCUSSION : reconstruction of stony river dam

Mr. first drawn up stream and then down stream, and intersected the base
egmann. ^£ ^^^ ^^^ near its center when the water reached the highest flood
level. The speaker thinks the original dam would not have failed b.v
overturning. The provisions made during the reconstruction to in-
crease the stability against sliding, at the same time, gave the dam a
still larger factor of safety against overturning.

Extension of the Cut-Off Wall. — Mr. Scheidenhelm has extended
the shallow cut-off wall near the west end of the dam to rock. His
explorations by test pits and drill-holes showed that, in some places,
w^ere the cut-off was thought to have been founded on rock, it was
really built on boulders, and permitted leakage under its base. In
such cases, Mr. Scheidenhelm carried the cut-off deeper, and, in other
places, he made, by a small V-shaped trench filled with concrete, a
water-tight seal between the up-stream side of the cut-off wall and
the bed-rock. The width of this wall — only 2^ ft. — did not seem to
be suificient for the maximum pressure it had to sustain, and, there-
fore, it is not surprising that Mr. Scheidenhelm fomid that in some
places water leaked through the cut-off wall.

Miscellaneous Construction. — Mr. Scheidenhelm has strengthened
the original footings of the buttresses, remedying faulty conditions
in places by pressure grouting. He has also provided a proper drain-
age system for taking care of the leakage through the foundation soil,
and has housed in the higher bulkhead portions of the original dam
by curtain-walls and roofs, in order to prevent serious freezing in
the drainage system, etc.

In conclusion, the speaker compliments the author on the very
thorough manner in which he has repaired and improved the dam,
and on the detailed account of this work which he has given in
his valuable paper.

IiJviNo P. Church,* Assoc. A>r. See. C. E. (by letter). — In con-
nection with Fig. 12, of his valuable and exhaustive paper, the author
refers to a tentative process by which such a curve as a-h-g, or "Dis-
charge over Spillways", is determined; that is, a curve showing the
rate of discharge, Q, over the spillway as a function of the time; so
that with Q known as a function of //, the depth on the spillway, it
becomes possible to compute H for any epoch ; this being a case where
the rate of influx into the reservoir, or "flood discharge", is given as
a function of time, in such a curve as a-c-d, or graph of flood dis-
charge, in Fig. 12.

As a matter of this kind is very rarely treated in books on hydraulics,
it may be of interest to consider the strict mathematical nature of
such a problem, that is, where efflux takes place through an orifice
or over a spillway from a very wide vessel or reservoir, simultaneously

* Ithaca, N. Y.



Mr.
Church.



discussion: reconstructiox of stoxy river dam 1077

with an influx into the reservoir; and where it is desired to determine Mr.
the value of the head, H (head on orifice or over crest of spillway), as *^^"''^^-
a function of the time, t.

At any instant of time, suppose the surface of the water in the
reservoir to be rising, and assume the following notation :

A = the area of that surface, at any instant ;
i'^ = area of orifice, if one is used;
IX = coefficient of discharge of orifice ;

and, if a spillway is used, let

6 = length of crest ; and
c = coefficient of discharge.
Also, let

Q = rate of efflux, in cubic feet per second, at any instant ;
and Q' = rate of influx, in cubic feet per second, at any instant.

During any short interval of time, d t (time-increment), H increases
by dH, and the gain of the volume of water in the reservoir must be
equal to the excess of influx over efflux during this tim^e; that is,

Q' . d t — Q . d t = A . d II (1)

(If the surface were sinking, we would have — A . d U instead of
+ A . dH.)

Case I. — Let Q' be constant and the reservoir have vertical sides,

with efflux taking place through an orifice; then A is constant. Let H^

be the initial value of H, and let the constant, Q\ be written in the

form, Q' ^ jii F \^ 2 g . K^ , where K is an ideal constant head, easily

computed. Equation (1) now becomes:

A dH
d t = = . (2)

1.1 F sf '2.g Ki — Ili

This is readily integrable by a temporary change of variable, K^ — Hs
being denoted by Z ; so that d H = — 2 H^ . d Z, = 2 (Z — Xi) . d Z.

Now, H = i^Q for t = zero and, finally, after integration, the form
is obtained,



t






giving the time, t, as a function of H. If t is given and H sought,
resort must be had to solution by trial. (By ''log.^" is meant J^aperian
logarithm.)

Case II. — Let the reservoir have a rectangular spillway on which
initially (that is, for t =^ zero), the head is H^ though at any later
instant it is H, the rate of eiflux being then Q = c .h s/ 2g . H^. Let the



1078 DISCUSSION : KECONSTEUCTION' of stony RIVEK DAM

Mr. rate of influx, Q', be constant, and let an expression be written for it

Church. / 3

in the form, Q' = c . 6 V Sgr . K^i, where K is an ideal constant head.

The sides of the reservoir are vertical; hence A \% o. constant.

Assuming that Q' is greater than the initial value of Q, we have
from Equation (1) :

A dH

dt = ^; . — (4)

c . b V 2 g' K's — H^

The integration* of this equation is quite roundabout, but finallyf
leads to the result

— V 12 tan.- ^ ^ — ^ ^^ — ^ . (5)

\3 7v + (2 V i^o + V ir) (2 V iT + V k)I _\

in which, as before, log.^ signifies " Naperian logarithm of " ; and
tan. - -^ denotes "anti-tangent of", or "arc whose tangent is"; for
example, tan. - ^ 0.488 = 0.454 (since tan. 26° = 0.488, and 26°
expressed in arc (" radians ") is 0.454).

Case III. — Rate of influx, Q', not constant, but proportional to
the time. Efflux over a spillway. If Q' is a linear function of t, that
is, is proportional to t, when t is reckoned from a special origin, we have
a case that is suggested by the graph of the Cane Creek flood of the
author's Fig. 12; since this curve, a-c-d, etc., may be considered to be
made up of a number of consecutive straight lines. Among the more
prominent of these straight lines is the portion, c-d; and this portion
has been treated by the writer in the attempt to discover the law con-
necting the variables, H and t, holding good for any instant of time
between about 9.30 and 10.30 a. m.

If a straight-edge be applied to the diagram in Eig. 12, it will be
found that the prolongation of the straight line, d-c, cuts the time-
axis very closely at 9 a. M. ; and that the point, d, corresponds to a value
of Q' = 5 000 cu. ft. per sec. at 10.30 a. m., that is, 5 400 sec, along the
time-axis from 9 a. m. Hence, if we reckon t from 9 a. m., we have the
proportion, Q' : 5 000 : : i : 5 400, or Q' = 0.926i (for the foot and second
as units).

Since the elevation of the surface of the water at this stage of the
flow is less than 140 ft., only two of the three spillways will be in
action, namely, the "old" and the "new"; and these have the same

* Detail will be found on p. 200 of Frizell's "Water Power", First Edition,
t On pp. 362 and 430 of Engineering News, November and December, 1901, will be
found interesting matter relative to this, by the Messrs. Gould.



discussion: reconsteuction of stony river dam 1079

crest elevation of 136.0 ft. and hence are equivalent to a single spillway Mr.
for which the rate of discharge, or "capacity" (see Fig. 11), is Church.

Q = 3.65 (150 + 133) . Jfl;
that is,

Q = 1033 . H^ cu. ft. per sec. (with Jf in feet).

Again, in Fig. 10, it is to be noted that the curve of reservoir
capacity is nearly straight between Elevations 134.0 and 138.0, the
gain of volume in that interval of 4 ft. being 500 000 000 gal. or a
fairly constant rise of 125 000 000 gal. per ft. Division by 7.58, and
by 1 ft., gives 16 710 000 sq. ft.* as a fair estimate of the area. A, in
the region of Elevation 136.0. These expressions and values having
been inserted in Equation (1), there is obtained, after reduction and
division (for foot-second imits),

10 000 000-—- = 0.554 t —618.4 if I (6)

This differential equation not admitting of a strict mathematical
solution,t resort was had to plotting a curve with R as ordinate and t as
abscissa (^ is reckoned from 9 a. m. and H is the head on the spillway),
a special method being used for the purpose, giving a very close
approximation.

It is a property of a circular arc that the tangent lines drawn at
its extremities intersect at a point equi-distant from the points of
tangency, and the method in question consists virtually in drawing,
"tandem", a number of very flat circular arcs, the radius of each being
the average radius of curvature for the small extent of curve involved.
The arcs themselves are not drawn, being sufficiently defined by the
extremities and their tangents.

For each of the five values of E: 0.6, 0.7, 0.8, 0.9, and 1.0 ft., the
cl 77
value of the derivative, — - — -, was computed from Equation (6) for

each of some six or eight values of t, chosen so that, by previous inspec-
tion, the true value of t for the assumed H would probably lie within
the range taken; and these thirty or forty values were tabulated.

In Fig. 37, for the starting point, D (corresponding as regards time
to points, h and c of the author's Fig. 12), we have H = 0.50 ft., and
t will be taken as 2 034 sec. (that is, from 9 a. m. ; see Fig. 12). Also,

cl J-f

the value of the " slope " of the tangent line at Z), namely, , is

found from Equation (6) to be — — . A straight line, D F, is then

* Prom the curve "area in acres" of Fig. 10, we might also have taken A = 386
acres X 43 560 = 16 820 000 sq. ft.

t Information on this point was kindly furnished by James McMahon, Professor of
Mathematics, Cornell University.



1080 discussion: reconstructiox of stony river dam

Mr. drawn through D at such an angle that a portion having 1 000 sec.

Church. ^^ horizontal projection has 0.0910 ft. as vertical projection. Use



Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 89 of 167)