American Society of Civil Engineers.

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

. (page 75 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 75 of 167)
Font size
QR-code for this ebook

to be adequate, was made by test pits and auger borings. After this
examination it was concluded that the depth to bed-rock was so great
as to prohibit the construction of a solid masonry dam, and that the
surface soil in the neighborhood of the dam was not of such a nature
as to warrant the construction of an earth fill dam. It had been the
intention of the Company to build a dam of one of these types, but,
under the circumstances, it was decided to construct a hollow rein-
forced concrete structure with a height of about 50 ft. above the original
water surface and a length of about 1 075 ft. The resulting capacity
,,,,:,. •Transactions, Ajn. Soc. C. E., Vol. LXXVII, p. 1069.


was approximately 1 533 000 000 gal. below the elevation of the spill-
way crest.

Certain companies were invited to submit plans and bids for the
construction of such a dam. On the basis of the bids received,* a con-
tract was entered into with Mr. Frederick G. Webber, President of the
Webber Construction Company, for the construction of a dam accord-
ing to plans submitted by the Ambursen Hydraulic Construction Com-
pany, licensor.

Work was begun by the contractor about Jime 1st, 1912, and carried
on until February, 1913. At that time the Company took the work
from the original contractor and engaged the Ambursen Hydraulic
Construction Company to complete it. The latter company took
over the work in March, 1913, and completed it in July, 1913.

The details of the original construction have been fully described
by G. H. Bayles, M. Am. Soc. C. E., Resident Engineer for the owner
during the construction. f Variations of actual conditions from those
assumed in the original design and construction will be discussed later
in this paper. Fig. 1 shows the dam as originally built.

The storage of water in the reservoir was begun in May, 1913, that
is, prior to the completion of the dam. The reservoir, however, was not
completely filled until the late autumn, and the water level had been
at the spillway crest for about 65 days prior to the faihire of the dam.
There had been but little flow over the spillway.

The failure occurred on January 15th, 1914, the cause, in the
opinion of the writer, being the washing out of the foundation soil by
leakage from the reservoir under an up-stream cut-off wall of inade-
quate depth. It is probable that there were conclusive signs of impend-
ing trouble for several days earlier (perhaps as early as January 10th),
in the way of water flowing through weep-holes in the floor or footing
of the dam at or near the section where failure finally occurred. With
proper attention on the part of the operating attendant at the dam,
the reservoir could probably have been drained in time to avoid the
failure which occurred.

It is not the purpose in this paper to discuss the failure in detail,
inasmuch as it has been described adequately in articles in engineering
periodicals.:}: For the sake, however, of making clear. the .description

* Engineering News, September 5th, 1912.
■f Engineering News, January 22d, 1914.

t Enaineering News, January 22d, 1914: Enaiveerina Record. January 24th.


of the design, and the work involved in the reconstruction, Figs. 2, 3,
and 4 are submitted, showing the portion of the dam which failed.

As a matter of personal interest, the writer visited the site as soon
as practicable after the failure, and spent portions of January 18th
and 19th, 1914, there. At the time of his inspection the main outlet-
gate in the dam had been opened, and the reservoir had already
been completely drained, so that Stony River was flowing in its normal
channel at the center of the valley, and not through the break, which
occurred at a higher level, and on the west bank of the valley. Based
on the inspection made at that time, the writer prepared and furnished
to the Public Service Commission of West Virginia, at its request, his
conclusions as to the failure of the dam. These conclusions, dated
January 22d, 1914, were as follows :

1. — Complete failure occurred between Buttresses 11 and 16, in-
volving a length of approximately 75 ft. out of a total crest
length of approximately 1 075 ft. At least three additional
bays, viz., as far eastward as Buttress 19, were damaged so as
to require practically complete rebuilding. Similarly, the
faulty cut-off wall construction and certain damage done west-
wardly from Buttress 11 must be remedied.

2. — Failure was caused by the undermining of the over-burden or
soil under the up-stream cut-off wall. The over-burden is in
general clayey, but non-homogeneous. Undermining was
initiated by leakage of water from the reservoir under approxi-
mately 25 ft. head through permeable over-burden not pen-
etrated by the cut-off wall.

3. — The pressure of ice on the reservoir against the deck of the
dam did not contribute to the failure.

4. — The dam was of the Ambursen, hollow, reinforced concrete
type. The type of dam was not in any way a cause of failure.

5. — The quality of concrete and character of construction were

6. — Failure occurred where the up-stream cut-off wall extended
only a short depth (5 to 7 ft.) into the over-burden. Where
the up-stream cut-off wall extended to a comparatively greater
depth (said to be to rock), failure was checked, and the cut-off
wall is still intact.

^^HB^v^B^^^^ '

1^^ Pmk._

STOney RfvtH UAM -*«a^ij!»- -

....,, 7 9^m


Fig. 1. — Stony River Dam, as Originally Constructed.

Fig. 2. — Break and Wash-Out of Stony River Dam, January 18th, 1914.

■p^Q 3 — Eastern Edge of Break in Stony River Dam. January 18th, 1914.

Fig. 4.- — ExTTTRESS 11 OF Stony River Dam. January 18th, 1914.






7. — The preliminary exploration and investigation of the dam site
were not sufficiently comprehensive, and did not develop essen-
tial and all-important facts.
8. — The technical advisers consulted in connection with the founda-
tion of the dam were not familiar with local conditions. The
geological factors at the dam site were not svifficiently con-

9. — The conditions obtaining at the Stony River dam site are such
that a safe dam could be built there. Whether the portion of
the dam still intact, and with its core-wall reported to extend
down to rock, has sufficient factors of safety, depends on
facts of which we do not have accurate knowledge. Further
investigation should be made as to this question.

10. — The mechanism for operating the outlet-gates and valves is
not sufficiently protected against ice formed by otherwise harm-
less leakage through the deck of the dam; nor is such mecha-
nism readily accessible under all conditions.

11. — Although the dam should, and could, have been designed and
constructed so as to be absolutely safe, regardless of the gate
mechanism and the character of the attendance, yet, even in
this instance, had the attendant been observant and resource-
ful, it would have been possible to open the sluice-gates in
ample time to drain the reservoir sufficiently to have obviated
failure of the dam.

12. — Published newspaper reports have grossly exaggerated the
details of the failure and the extent of the damage to property.

In general, the foregoing summarized conclusions on the part of
the writer still hold good. It may be well to point out, however, that
the writer distinguishes between the "character of construction work"
and the design of the structure.

The location of the break in the dam, with reference to the entire
profile, is shown on Plate IX. Complete failure, as previously stated,
involved a length of about 75 ft., viz., between Buttresses 11 and 16,
but certain additional damaged work was removed, so that the gap in
the dam to be closed during the reconstruction work comprised a total
length of 135 ft., viz., between Buttresses 10 and 19.

In February, 1914, the writer was engaged as Consulting Engineer
by the West Virginia Pulp and Paper Company to take charge of the


proposed reconstruction. He desires to emphasize the point that he
was not requested to place the responsibility for the failure, nor did he
attempt to do so. It is fair to note, however, that in this case, unlike
most other failures, no criticism has been heard directed against the
owner, that is, against the executive officers of the West Virginia Pulp
and Paper Company, on the score that economy was preferred to safety.
Moreover, in connection with the proposed reconstruction, the execu-
tive officers of the Company proceeded on the principle that either the
dam should be reconstructed so as to be reliably safe, or no attempt
should be made to utilize the damaged structure. The same attitude
was manifested throughout the entire period of reconstruction.

Prior to the final decision to proceed with the actual work of recon-
struction, the entire structure and foundation conditions were inves-
tigated carefully, in order to discover any features of design or con-
struction involving insufficient margins of safety. Not until about
August 1st, 1914, did the owner finally decide to reconstruct. The work
was finished in May, 1915, though the storage of water in the reservoir
had been resumed several months earlier.

Geological and Foundation Conditions.

The available information regarding foundation conditions was
found to be so inadequate that the investigation which followed had
to be made practically as complete as though there had been no dam on
the site. In fact, the existence of a dam at the site necessitated addi-
tional exploration to determine the adequacy of the cut-off.

An F-l Davis Calyx core-drill was placed on the work, and by it ten
holes were sunk. The deepest hole was sunk to approximately
100 ft. below ground surface; several other holes were carried
to depths of approximately 80 ft. Usually, the strata penetrated were
sufficiently tight to retain in the hole the water necessary for the shot-
drilling process. In other cases, however, especially on the east bank
of the valley, seams were encountered in the rock, and these were of
such capacity that it was difficult to keep sufficient water in the holes
to allow the drilling to proceed.

iSTumerous test pits were sunk, the majority being immediately
adjacent to, ;nid up stream from, the original cut-off wall. Such test
pits not only exposed the cliaracter of the over-burden, but also per-
mitted an examination of the original cut-off and the efficacy of its
seal into the bed-rock. Several of the test pits extended to a depth of


approximately 45 ft., and were sunk at considerable expense, because
of the inflow of water and the necessity for sheeting the pits for prac-
tically their entire depth.

On Plate IX are sho\vii the locations of the core-drill holes, the logs
thereof, and the "lay" of the rock strata as deduced from the logs and
other records. This plate shows, also, the profile of the original ground
surface of the valley. It is apparent that the axes of the original
valley and of the present valley do not coincide. By "original" valley
is meant that valley which existed at the time the stream was still
eroding the bed-rock and was in immediate contact with that rock.
At that time the axis of the valley was approximately at the west end
of the old spillway, or at about Station 5 -j- 50. As time went on, the
stream ceased its cutting or eroding action and began to build up the
floor of the valley by depositing sediment and gravel. At the sides of
the valley this process was apparently aided materially by "slips" and
talus from the hillsides.

During this process Stony River meandered back and forth across
the valley until at present the axis of the valley is approximately at
Station 4 -|- 40. If, in the profile shown on Plate IX, the original and
the present locations of the stream-bed were to be connected by a line
passing through all the intermediate locations of the stream-bed, such
connecting line would have an extremely irregular or zigzag course.
Water-worn boulders and gravel were found strewn through various
portions of the over-burden as exposed by the exploratory work, tes-
tifying to the fact that they had been deposited by the stream in
ages past.

The Stony River Valley lies in the geological formations of the
Carboniferous period. The lower portion of the valley, at the dam site,
lies in what is known locally as the Savage formation, one of the prin-
cipal features of which is the Davis or Kittanning coal, which is of
commercial importance throughout this region.

The general characteristics and slope of the formations are indi-
cated roughly in the Piedmont folio (West Virginia) of the United
States Geological Survey. The detailed geological correlations at the
dam site, however, were far from simple. Variability, rather than uni-
formity or persistency, characterizes the Savage formation, and no-
where is this exemplified better than at the site in question. Prom the
core-drilling data it is apparent that lentils abound, that relatively soft


strata are interbedded with hard strata, and, what is of even greater
importance, that the geological horizons, which at one location offer
impervious strata, show decidedly pervious strata at other locations.

The characteristic rocks are shales of a laminated or seamy char-
acter, as shovsTi in the typical section. Fig, 6. Here and there lentils
of coal and black slate are interbedded in them. At approximately the
location of the break, viz., in the neighborhood of Station 2 + 00, the
shales were found to be so crumpled and crushed by folding near the
surface as to require their excavation in the process of reconstruction.

It is not believed that this condition had any relation to the failure
of the dam, but it is probably true that, once the dam had failed and
the reservoir was being emptied through the break, the erosion of the
bed-rock immediately down stream from the place of failure was
facilitated considerably by this local surface crumpling of the shales.

At the east side of the valley (the left side of the profile in Plate
IX) heavy sandstones were encountered. Evidently, at the time the
valley was formed by the process of stream erosion, the sandstones
caused the formation of overhanging cliffs, under which the softer
strata were eroded. Frost and other natural forces, however, had their
effect, even on these sandstone cliffs, with the result that they became
fissured, though the boulders which were thus cracked off probably did
not move far from their original location. In fact, excavations made
under the original cut-off showed that in the original construction
certain of these boulders or false cliffs had been mistaken for bed-rock,
and that under them there was actually pervious material, such as the
disintegrated coal shown in the profile immediately to the west of
Station 9 + 00.

The over-burden or foundation soil at this site is mainly of a clayey
nature. It is far from being homogeneous, as is evident from Fig. 7,
which shows typical soil conditions at Buttress 10. In general, the soil
near the surface is loamy and more pervious than that immediately
overlying the bed-rock. Near the middle of the valley, however, the
conditions are reversed, and there deposits of pervious materials, such
as gravels, were found at considerable depths.

A catalogue of the constituent materials of the over-burden shows :

Yellow clay impervious, of fair bearing value.

Blue clay impervious, of fair bearing value.

1 ^ /^/^


p ..:§: '-fl

^ i^... "^^^^^HRr^*" t^^^

Fig. 5. — Buttresses 16 to 19, Stony River Dam,
March 16th, 1914.

Fig. 6.

-Typical Section of Sha-le at Site of Stony River Dam,
Near Point of Failure.



liJHIBIU «• *i^>'-^


^H i?i|^^^^KB

HB 1





Fig. 7. — Typical Foundation Soil Conditions at Buttress 10, Stony River Dam.

'f% ■ lir

Fig. 8. — Test of Frictional Resistance.


Black clay or gumbo, .impervious, of fair bearing value when dry
or only moist, but of poor bearing value
when saturated with water.

Sand and gravel pervious, but of good bearing value.

Loamy sand in minor quantities, pervious, of poor bear-
ing value.

Sandy clay fairly water-tight, of fair bearing value.

Shale fragments not in place, and in various stages of dis-
integration, mostly permeable, of fair bear-
ing value.

Boulders in considerable numbers, and, of course, of

excellent bearing value.

The bearing values, as referred to in the foregoing, are relative;
comparison is made with the bearing values of other materials present
in the over-burden, and not with those of granite or even of the shale
rock on which the new buttresses, 11 to 18, inclusive, are founded.
The materials occurred in no regular order.

Such foundation conditions required the most serious consideration
and care in construction. In general, the designs involved in the work
of reconstruction were based on the assumption that the worst condi-
tions might exist at any bay. In view of the findings of the exploration
work, it appeared logical, however, to attribute higher bearing values
and greater resistance to percolation to the foundation soil of the east
hillside, that is, roughly, east of Station 5 + 00, for in that portion of
the valley there is more clay, and it is of a better character.

Simultaneously with the work of core-drilling and sinking test
pits, the wrecked portions of the dam were removed, and the under-
lying foundation soil was stripped from the bed-rock, thus permit-
ting a first-hand study of the strata on which were to be founded
the buttresses which were to replace those taken out. Such excava-
tion was actually a part of the reconstruction, inasmuch as it would
have been necessary later, if it had not been made as part of the
exploratory work. The investigation was completed in June, 1914,
and designs and recommendations for reconstruction based on the
findings were prepared immediately.

An explanation of the treatment of the more important problems
of the reconstruction follows.


Ti' Spillway Provision.

Original Spillway Capacity. — A problem affecting most of the
other features of design — because it determined the probable maxi-
mum head-water level, and hence affected the probable maximum
stresses — ^was that of the spillway capacity to be provided. The orig-
inal spillway provision was apparently inadequate, being only about
2 800 sec-ft. with water level at the elevation of the original crest
of the bulkhead sections of the dam, viz., Elevation 139 (Plate IX).
In fact, it is likely that the actual safe flood discharging capacity
would have been somewhat less, for the original spillway crest had
not been given the most advantageous shape.

Assuming the drainage area to be 11.4 sq. miles, as shown by the
"Piedmont" topographical quadrangle issued by the U. S. Geological
Survey, the original spillway capacity was less than 250 sec-ft. per
sq. mile of drainage area. ,

Pertinent Records of Flood Flow. — In general, in a given region,
the smaller the drainage area the larger the maximum unit run-ofi
and flood flow. So-called cloudbursts are limited as to area affected,
but certainly the Stony River drainage area is sufficiently small to
fall within the territorial limits of a cloudburst. It happens that
there are two important and applicable records* of flood discharge,
namely, those of the floods on:

Cane Creek, at Bakersville, N. C, May 19th-20th, 1901; drain-
age area, 22 sq. miles ; estimated maximum discharge, 1 386
sec-ft. per sq. mile; elevation above sea level at point of
measurement, approximately 2 450 ft. (U. S. G. S. datum) ;
elevation of highest point in water-shed, approximately
5 330 ft; and
Elkhom Creek, at Keystone, W. Va., June 22d, 1901; drainage
area, 44 sq. miles ; estimated maximum discharge, 1 363 sec-ft.
per sq. mile ; elevation at point of measurement, approxi-
mately 1 600 ft. ; elevation of highest point in water-shed,
approximately 3 365 ft.

The fact that both of these extreme floods occurred in the South-
ern Appalachian mountain system, in which the Stony River dam is
situated, and the further fact that the several drainage areas in ques-

• Engineering News, August 7th, 1902.u:0waij'XJi;XiO';


fibiii including that of Stony River (with the dam at approximately
Elevation 3 350 and the highest point in the water-shed at approxi-
mately Elevation 4 200), are at not widely differing altitudes, cause
these two records to be exceptionally valuable for purposes of com-
parison. These drainage areas are apparently quite similar in their
physical characteristics.

The records referred to were reported by Mr. E. W. Myers, then
Engineer of the J^orth Carolina Geological Survey, and Assistant
Hydrographer of the U. S. Geological Survey. It is to be admitted
that the method of measurement was probably not such as to warrant
expressing the discharges to the nearest second-foot; on the other
hand, an examination of the data does not warrant one in disregard-
ing the records or assuming that the maximum flood discharges were
materially less than those stated.

Although the Cane Creek flood was estimated to have been actu-
ally of slightly greater intensity than that of Elkhorn Creek, yet
relatively, and for the purpose of predicting the maximum probable
flood discharge from the drainage area above the Stony River dam,
the Elkhorn Creek flood is of considerably greater importance, because
of the fact that the drainage area at the point of measurement is
twice as great as in the case of Cane Creek. For detailed data, how-
ever, one must work from the records of the Cane Creek flood, as in
that case, fortunately, Mr. Myers made an estimate of the average
discharge during each hour of the flood, as well as an estimate of
the maximum discharge during the flood. In the graph of this flood,
Fig. 12, the run-off of Cane Creek has been reduced in proportion
to the drainage area, so as to be applicable to the drainage area above
the Stony River Dam. Granting that the Cane Creek data are only
approximate, they nevertheless appear to be well worth while adopt-
ing as a basis of study.

It is noteworthy that all available records indicate that, in the
region under consideration, maximum rains, of suiEcient duration to
tax the equalizing capacity of the reservoir, do not occur in the early
spring, and hence do not coincide with the melting of the snow. For
this reason it was assumed that the greatest flood to be anticipated
at the Stony River Dam would be due solely to rainfall.

Comparison with Records of Maximum Rainfall. — For practical
purposes, it may be assumed that the Elkhorn and Cane Creek floods


were of equal intensity when reduced to a unit basis. The following
remarks, based on the Cane Creek record, are consequently applicable
also to the even more important Elkhorn Creek record.

Both these floods were undoubtedly caused by local cloudbursts
which occurred after previous rains had saturated the respective drain-
age areas. For the purpose of comparison with rainfall records,
therefore, it may be assumed that the maximum run-off per hour
was approximately equal to the maximum rainfall per hour. In the
case of the Cane Creek flood it is reported that, during the hour of
maximum run-off, the discharge was at an average rate of 1 202 sec-ft.
per sq. mile, corresponding to a rainfall at the rate of approximately
1.86 in. per hour.

Referring to Fig. 9, showing the duration of maximum intensi-
ties of rainfall in the United States, it will be noted that the maxi-
mum rate of rainfall previously derived for the Cane Creek flood
falls well within the limits of the highest records shown.

Bulletin D of the U. S. Weather Bureau, the basis of the princi-
pal data of Fig. 9, includes only records obtained after the establish-
ment of self-registering gauges in 1893, but prior to 1897. Because
of the limited period which it covers, and, further, because of the
limited number of precipitation stations involved, it is certain that the

Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 75 of 167)