American Society of Civil Engineers.

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

. (page 2 of 167)
Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 2 of 167)
Font size
QR-code for this ebook


remaining on, and passing, the SO-mesh
sieve.



Mesh.



+ 100
+ 200

— 20()

— 200



Character.



Sand

Fine sand and silt
Clay



Percentage,
by weight.



7.86
10.06



19.20

21.27

40.00

1.61



100.00



Chemical Analysis.
Air-dried sample.



Constituents.



Woody fiber

Insoluble

Silica (Si02)

Ferric Oxide (Fe208)

Alumina (AI2O3)

Lime ( CaO )

Magnesia (MgO)

Water at 110° cent

Lesson Ignition (CO2), Organic,

etc

Undetermined



Percentage,
by weight.



0.62
52.16



44.48
2.46
5.92

18.27
5.70
0.60

21.58
0.99



This combined material was then tested in the tank to determine
the hydraulic gradient. The head of water in the tank above the soil
immediiitely came to a maximum, and showed only a very slight fall



TRANS.



PLATE I.

M. 60C. CIV. ENGRS.

VOL. LXXXI, No. 1379.

HAYS ON

DESIGNING AN EARTH DAM.



0.04 0.05



Diameter of Particles
0.09 0.10



0^ 0.3O



0.60 0.70



l[







1


1




1
1




























1


















_,_-


;;4


46


34


20


— 1


0.00340


32






100


+
+


47


40


14


10


7


52


39


23





0.00374



* This column gives the seepage or percolation, in cubic feet per second per linear
foot of model dam.

t Experiments were taken with water receding in order to compare with others ;
also, no percolation readings were taken.

t Figures excessive, due to water breaking through up-stream section of model
dam.

Note. — The figures in this table may be converted to inches by pointing off one
place and reading direct, thus 17 equals 1.7 in.

The writer wishes to state at this point that where the figures were
given in feet, for convenience in stating the case relative to the actual
structure, they represent actually tenths of inches on the model, that
is, 19 ft. as stated above was 1.9 in. measured on the model. The
figures given in the preceding paragraph were for a full head (10 in.
equals 100 ft.).

The hydraulic grade line, represented by the upper row of tubes,
indicated that there was an appreciable loss of head at the down-
stream toe of the dam, and the lower row of tubes showed a complete
loss of head at Tube K.

Fig. 10 (6) shows the hydraulic grade line of the lower row of
tubes when the distance is computed from the up-stream toe of the



20 DESIGNING AN EARTH DAM

dam to each tube directly, and it will be noted how close the hydraulic
gradient is to a straight line. Tube J was evidently affected by the
sheet-piling to some extent.

Fig. 10 (c) shows the seepage loss from the body of water above
the dam. This subject will be discussed later.

The failure of the "line of creep" theory to work out on the model
as it was expected, caused the development of another theory as to
the effect and usefulness of sheet-piling as a cut-off, preventing or
reducing the underground flow of water in dams designed or built
under similar conditions.

The theory finally evolved was that the sheet-piling cut-off was
greatly similar in its effect to a partly closed valve, wherein the water
is retarded and shows a higher pressure head just above the cut-off and
a lower one just below. This theory is well known in regard to the
hydraulic gradient of a pipe line, and explains the inefficiency of the
upper row of sheet-piling as a cut-off. Hence, in the final design of
the dam, this upper or shorter row of sheet-piling was omitted, and the
second row was lengthened.

The model was then reconstructed according to this idea, making
the piling 12^ in. (125 ft.) instead of 8^ in. (85 ft.). Tests were then
made in exactly the same manner as before, and the results obtained
showed a very noticeable effect due to the increased length of sheet-
piling, which bore out the theory as to effect of sheet-piling and its
similarity to the partly closed valve.

Fig. 11 shows typical results taken from the last series of tests,
as in Table 3. Fig. 11 (a) shows a graphical record of the readings
taken on the tubes. It will be noticed that the hydraulic grade line
cuts the base of the dam well up stream from the toe. Fig.
11 (&) shows the hydraulic grade line of the lower set of tubes,
where the distances from the up-stream toe of the dam to each tube
were plotted horizontally, to show more correctly the hydraulic grade
line of the water flowing through the gravel beneath the dam. Fig.
11 (c) shows the seepage, in cubic feet per second, per linear foot of
the model.

Fig. 7 shows the tank for testing the model, under a full head. The
water was colored red in order to show clearly in the photograph.

The down-stream section of the dam was constructed of the same
material as the foundation, in order to provide adequate drainage and



DESIGNING AN EARTH DAM



31




Cubic Feet per Second per
Linear Foot.



s



22



DESIGNING AN EARTH DAM



for economy. It will be noted, from the plotting of the readings of
the upper row of tubes, that there was no water in this section of the
dam, except possibly that due to capillary action. This contributes a
great deal to the stability of the dam as a whole.

It will be noted from Figs. 10 and 11, that the water traveling at
great depths is not materially affected by the sheet-piling, at any
particular point; but, throughout the whole distance under the dam,
the general effect is obtained. This may be attributed to the ex-
tremely slow velocity of the water.

TABLE 3.— Second Series (a).



No.


Date,
1915.


Time.


Head.


A


B


C


D


E


F


H


I


J


K


S*


1


7-19


10 a.m.


50


16


13


8











14


9


4


—4


0.00142


2


1. 1 .1


75


27


22


16











23


17


9}i


—3


0.00202


3


" •'


90


33


25


21


4








30


23


12


—2


0.00237


4


"




100


40


30


25


6


2





35


26


15


-IJi


0.00304


5


7-20




50


13


12


6











10


7


3


—4


0.00121


6


"




75


21


19


13











18


14


7


-3


0.00174


7


"




90


30


25


18^


2








27


19


W'A


-2^


0.00203


8


"




100


34^


27


21


4








30


22^


12


-2


0.00217


9


7-21




55


15


8


6











11


7


%


—4


0.00127


10


"




75


23


15


12»^


1








19


13


7


—3


0.00170


11


"




91


31


22


17


1








26


19


10


—3


0.00203


13


'•




100


44


26


20


3








30


21


12


—3


0.00229


13


9-16


2 p.m.


30
50


10^
23




7
















^3


4


—4

-1


-6}^
— 5


0.000418
0.000857


14










2


1


15


"




75


40


22


7











8


5


3


-W2


0.00142


16


"




90


t


30


13











14


9


5


-3


0.00182


17






100


+


36


16


6








20


13


8


—3


0.00222



* Percolation, in cubic feet per second per linear foot of model.

t Readings for Tube A were excessive, due to breaking through the thin section of
model dam over that tube.

Minus readings show that the pressure, as indicated by the tubes, did not reach
the base of the dam.

Figs. 10 (c) and 11 (c) show the seepage water coming through
under the dam. The results show that seepage water will follow the
straight-line rule, after the soil is once filled with water and a suffi-
cient head has been reached to overcome the friction. The rate
of increase of flow, for each additional unit of head applied, varies
for sands and gravels under different conditions, such as porosity
of sand or gravel, character of voids, etc. As it would take a very
extensive set of experiments to make any determinations in regard to
the laws of seepage, the writer merely wishes to present the results so
that others may use them.

In any dam, the correct place for a cut-off wall, under conditions
such as these, namely, a pervious foundation, is at, or near, the up-



DESIGNINa AN EAETH DAM M

stream toe. However, in an earth dam, it must of necessity be placed
nearer to the center, or axis, of the dam. For construction reasons,
the sheet-piling in this dam was placed a short distance up stream from
the center line, but, at the same time, it was stifficiently far back to
prevent any water from seeping through the dam to the down-stream
side of the piling in any appreciable quantities.

Conclusions.

Earth dams fail from four main causes, and the writer desires to
show how structures of this type may be made safe.

One cause is from overtopping, due to high waves or a flood. This
is provided against by designing an adequate spillway and providing
sufficient free-board. This point has not been discussed in this paper,
as it is governed by local and not general conditions. The concrete
core-wall was designed to extend to the top of the embankment, which
would prevent waves from cutting through.

A second cause of failure is by water seeping along a conduit placed
in the built-up portion of the dam. In the present case, the outlet
consists of a tunnel through the rock cliffs at one end of the dam,
thus eliminating this cause of failure.

A third cause, and an unseen one, is burrowing animals. Where
the down-stream section is loose rock this cannot happen, and it is
much the same in loose gravel.

The fourth cause is the occurrence of springs, or boils, on the
down-stream face of the dam, or near it, which in turn cause blow-outs
and sloughing off of the bank. A solution of this problem was the
main object of these experiments. Springs, or boils, are caused by
not consuming all the head of water above the dam. From the tests
described herein, it is seen that either an extreme length of base of
the dam must be provided, or there must be sheet-pile cut-off walls
of the proper length. Where the sand or gravel foundation is of
very great depth, a short cut-off would not be as efficient as in a
shallower foundation, but where models are constructed, results can
be obtained which enable the engineer to make correct designs better
than by any other method of study.

Finally, the writer wishes to state that he claims no new dis-
covery, but has determined a type that has solved the problem under
the conditions with which he was confronted. This type has been



M DESIGNING AN EARTH DAM

built before, but has not been advanced together with reliable infor-
mation from tests, etc. The study of the effect of the sheet-piling
will undoubtedly be of value to the Profession. The application of
the sieve analysis curves in combining the materials, for the purpose
of determining the densest mass obtainable, has not been advanced
before — as far as the writer knows — but its value as a time saver can
easily be seen. Levees, canal banks, dikes, etc., may be studied in
the same manner. Each design is a problem in itself, and the experi-
ments described herein should not be misinterpreted, or applied too
broadly.



DISCUSSION ON DESIGNING AN EARTH DAM 25

D iscxjssioisr



W. G. Bligh,* M. Am. Soc. C. E. (by letter). — This paper is of Mr.
exceptional interest, as it deals with the first actual experiment made
in many years with the object of ascertaining the value of the reduc-
tion in pressure or head caused by vertical obstructions such as
curtain-walls or diaphragms of sheet-piling projecting below the base
of a dam.

Any experiments with models, to be of authoritative value, must
be on a reasonably large scale, and, further, the conditions should
not deviate from those actually encountered in practice. In the
present case, the scale of the model is considered too small, and
might well have been 10 ft. to 1 in. However, the experiments in
question were imdertaken by private parties for a specific purpose,
and the results obtained are sufficiently reliable for the aim in view.
It is sincerely to be hoped that this good work may be continued, and
this much discussed point finally disposed of in a convincing and in-
disputable manner. With regard to the second suggested condition,
the experiment is all that could be wished, the relation between the
head of water and the length of base provided in the model being the
same as would be the case in the actual dam when constructed. This
condition has not been observed in the paper by J. B. T. Colman,
Assoc. M. Am. Soc. C. E., on "The Action of Water Under Dams",t
where the disproportion between the head of water and the length
of travel provided for the percolating undercurrent is imreal and.
in a measure, must vitiate the authoritative value of the deductions
obtained.

In the model dam, the base length was made dependent on actual
experimental data obtained by filtration. This determined the net
value of base length or travel of the percolating current requisite to
neutralize entirely the head of water. It is quite evident that the
length thus obtained must in many cases be a maximum value, whereas
the minimum safe base length is the real criterion. Supposing the
base to be shortened, the neutralization of the whole head would then
still take place, but the velocity of the percolating undercurrent would
be increased, as well as its ability to wash out and carry away par-
ticles of sand or gravel from under the dam and thus threaten piping
action, which, in time, would cause undermining. This minimum
length of travel suitable to any description of substratum could be
obtained by actual experiment to which a factor of safety would
necessarily be applied of, say, not less than 1.3. It is possible that,
in certain cases, the base length thus deduced would nearly equal the
maximum, but in nearly all instances it will be very much less. For

* Toronto, Ont., Canada.

t Transactions, Am. Soc. C. E., Vol. LXXX, p. 421.



2^6 DISCUSSION ON DESIGNING AN EAETH DAM

Mr. example, in the extreme case of a loose boulder river bed, the length
' requisite to neutralize completely a given head of water by filtration
would be altogether impracticable; on the other hand, owing to the
weight of the individual particles composing the foundation, a com-
paratively high velocity in the undercurrent would be admissible with
a correspondingly reduced length of base. As far as the writer is
aware, no comprehensive experiments have ever been undertaken to
solve this matter in a really satisfactory manner.

The so-termed "percolation factors", first introduced some years
ago by the writer, were deductions from actual failures of certain
large river works, and although they are undoubtedly reliable for
beds of pure sand, and have been accepted as such, still further definite
experimental data would be most desirable.



Online LibraryAmerican Society of Civil EngineersTransactions of the American Society of Civil Engineers (Volume 81) → online text (page 2 of 167)