Cornell University Library

TC 547.C91
Engineering for masonry dams,

3 1924 004 025 213

TC547

(Cornell Mniuetaitj} ffithratij

3tf?ara, Sfetn IJatk

BOUGHT WITH THE INCOME OF THE

SAGE ENDOWMENT FUND

THE GIFT OF

HENRY W. SAGE
1891

ENGINEERING LIBRARY
 Cornell University
Library

The original of this book is in

the Cornell University Library.

There are no known copyright restrictions in

the United States on the use of the text.

http://www.archive.org/details/cu31924004025213
 ENGINEERING
FOR

MASONKY DAMS
 ENGINEERING
FOR

MASONRY DAMS
BY
WILLIAM PITCHER CREAGER, C.E.
Mexbek, Ajikkicax Socasrr of Cmi. Exgixkebs

FIRST EDITIOX

>TEW YORK:

JOHN WILEY & SONS, Inc.

Losdos : CHAPMAX & HALL, Limited

1917
 Copyright, 1917
BY
WILLIAM PITCHER CREAGER

PRESS OF
BRAUNWORTM & CO.
BOOK MANUFACTURERS
BROOKLYN- N. V-
 PREFACE

In reviewing this book the reader will probably be impressed

with the fact that many of the fundamental assumptions of design
are based on very obscure data. This is particularly true regard-
ing uplift pressure, ice pressure, and the distribution of stresses
in high dams.
It is believed, however, that such assumptions are all on the
side of safety ; for each of the recorded failures of masonry dams
may be attributed to a violation of one or more of the standard
rules on which the theory is based.
It is considered that the methods of design described, and the
assumptions recommended, represent present conservative prac-
tice, and correspond to a proper degree of safety for the average

enterprise, and where considerable damage to property and loss

of human life would result if failure occurred.
Some of the designing assumptions, and particularly those for
the usual unit working stresses, may seem ultra-conservative,
as compared with those allowed for other masonry structures.
It must be remembered, however, that the theory on which the
design of masonry dams is based, is not exact, and, moreover, has
not been verified by satisfactory experiments. The failures
which have occurred have furnished valuable information regard-
ing some of the limiting conditions, but not all. For instance, no
failure due to crushing of concrete masonry has been recorded,
arid it is not known just what margin of safety the usual work-
ing stresses afford. Designs of extremely radical tendencies are
being made continually, particularly for dams in unsettled regions,
and where there are no government restrictions. There is a
possibility, therefore, that future failures will furnish much
needed information in this respect.
For valuable suggestions, criticisms and other help, the author
desires toacknowledge his indebtedness to Messrs. H. L. Coburn,
A, A, Conger, A. S, Crane, A. D. Flinn, N. C. Grover, R, C. Lat-
vii
viii PREFACE

imer, Daniel Moran, G. S. Thompson, and many of his office

associates.
Special thanks are due to Mr. J. W. Van Demburg for generous
and intelligent assistance in the preparation of examples of design
and in research work.
Much information was obtained from several branches of the
Federal Government, many of the State and Municipal engineer-
ing departments, and from engineering literature, as noted in the
text.
Acknowledgment is also made to Mr. T. J. McMinn and Mrs.
W. P. Creager for invaluable assistance in the editorial work of
publication.
 CONTENTS

CHAPTER 1

Investigations and Surveys

PAGE
1. The Choice of Location 1
2. The Nature of Investigations 3
3. Preliminary Investigations 4
4. Final Investigations 6

CHAPTER II

The Choice of Type op Dam

5. General Considerations 11
6. Solid Gravity Masonry Dams 12
7. Hollow Gravity Masonry Dams 13
8. Arched Masonry Dams 14
9. Embankments 14
10. Timber Dams 15
11. Other Types 15

CHAPTER III

Forces Acting on Dams

12. Nomenclature 16
13. General Considerations 19
14. External Water Pressure 19
15. Internal Water Pressure. Uplift 25
16. Earth Pressure 33
17. Atmospheric Pressure 35
18. Ice Pressure 37
19. Wave Pressure 40
20. The Weight of the Dam ."
40
21. The Weight of the Foundation 42
22. The Reaction of the Foundation 42
ix
 .

x CONTENTS

CHAPTER IV

Requirements fob Stability op Gravity Dams

PAGE

23. Causes of Failure 48

24. Rule 1, Governing the Location of the Resultant 49
25. Rule 2, Governing the Inclination of the Resultant 50
26. Rule 3, Governing Compressive Stresses 52
27. Rule 4, Governing Tension in Vertical Planes 56
28. Rule 5, Governing the Margin of Safety 57
29. Rule 6, Governing Details of Design and Methods of Construction . 58

CHAPTER V

General Equations for Design of Gravity Dams

30. General Considerations 60

31. Equations for Rule 1 64
32. Equations for Rule 2 70
33. Equations for Rule 3 71
34. Equations for Rule 4 75
35. Equations for Rule 5 76
36. Equations for Rule 6 77

CHAPTER VI

The Design of Solid Non-Overflow Gravity Dams

37. General Considerations 78

38. Example No. 1. 200-ft. Solid Non-Overflow Dam 79
39. Example No. 2. 102-ft. Solid Non-Overflow Dam 99
40. Comparison of Non-Overflow Dams 104

CHAPTER VII

The Design of Solid Spillway Gravity Dams

41. General Considerations 105

42. The Shape of the Crest 105
43. DischargeCapacity HO
44. The Bucket '.'.'.'..'....'... '.

115
45. Example No. 3. 91-ft. Solid Spillway Dam without Ice Pressure .... 117
46. Example No. 4. 87-ft. Solid SpilwayDam with Ice Pressure 118
47. Example No. 5. 30-ft. Solid Spillway Dam .120
48. Comparison of Solid Spillway Dams 130
 CONTENTS xi

CHAPTER VIII

The Design of Hollow Dams

PAGE
49. General Considerations 132
50. Example No. 6. Hollow Non-overflow Dam 40
]

51. Example No. 7. Hollow Spillway Dam 145

CHAPTER IX
The Design op Arch Dams
52. General Considerations 148
53. Arch Stresses 149
54. Vertical Beam Stresses 157
55. Recommendations for Design 158
56. Details 160
57. Multiple Arch Dams 164
58. Allowed Stresses 165
59. Examples of Arch Dams 165

CHAPTER X
Preparation and Protection op the Foundation

60. General Considerations 172

61. Rock Foundations 172
62. Earth Foundations 181

CHAPTER XI
Flood Flows

63. General Considerations 194

64. High-water Marks 195
65. Comparison with Other Rivers 196

CHAPTER XII

Details and Accessories

66. Masonry for Dams 203

67. Water-Proofing 205
68. Contraction Joints 206
69. Drainage Systems 210
70. Architectural Treatment 211
71. The Regulation of High- Water Surface 214
ENGINEERING FOR MASONRY DAMS

CHAPTER I

INVESTIGATIONS AND SURVEYS

1. The Choice of Location. A dam isusually a unit in a

more or less extensive project involving the construction of a
number of structures of various types. The general location,
therefore, is fixed by factors varying with the purpose of the
project, and is affected by
considerations but remotely allied
to the text of this book. Bearing in mind the fact- that the
general location to be adopted is that which, at reasonable cost,
will be best suited to the purpose for which the dam is intended,
we have only to consider here those factors which affect the cost
and safety of the dam, and the choice of its exact position.
The general location having been chosen, the exact position
will be fixed after careful consideration of each of the following
factors:

a. The character of the foundation;

b. The configuration of the earth and rock surfaces at the
site and on the length of the dam, quantity
its effect

of material to be excavated, and other factors;

c. Availability and character of materials for construction;
d. The value of the necessary lands and water rights;
e. Requirements as to coffers, pumping, conduits, and other
provisions necessary for unwatering the site;
/. Transportation facilities and the accessibility of the site;
g. Availability of suitable sites for construction equipment
and camps;
h. The safety of the structure.

The foundation * is one of the most important factors in the

final location. It should be practically impervious, or capable
* See Chapter X.
2 INVESTIGATIONS AND SURVEYS [Chap. I

of being made so, and have sufficient strength to sustain the weight
of the dam and prevent sliding.
valley should have sufficient width at the crest of the
The
proposed dam for the requisite spillway,* and in some instances
for other structures,such as the power-house in a water-power
development. A width than that necessary for the
greater
spillway and other structures is, of course, undesirable, because

of the greater cost of a long dam. The configuration of the

earth and solid rock at the site affects greatly the cost of the dam,
because of its influence on many items of cost, such as excava-
tion, coffers, and masonry.
The location of sand, and stone or gravel for the masonry,
the cost of quarrying, the facilities for transportation to the
concrete mixers or stone yard, the quality of such material and
its effect on the strength, durability, and appearance of the
masonry; the hardness of the stone and its effect on the cost
of crushing for concrete or shaping for ashlar facing, all influence
the cost of the work and therefore the choice of location.
An important item in the cost of the project will be the value
of the lands to be provided for the dam, for the reservoir area,
for construction equipment and camps, and for the right of
way for a construction highway or railroad to the site. The in-
crease in the elevation of the water surface by the dam may
necessitate the relocation of existing railroads or public highways.
In some instances the abandonment of towns and villages has
been necessary.
The removal of water from the site of the dam, in order to
facilitate construction, commonly called " unwatering " the site,

often involves a large percentage of the total cost of construction.

If the depth of water at the site is considerable, and if it is prob-
able that floods will cause great increases in the stage of the river
during the period of construction, the cost of the coffer-dams may
be excessive. Many otherwise attractive sites have had to be
abandoned on account of the probable great cost of unwatering
caused by the great depth and velocity of the water and the dif-
ficulties to be overcome in constructing and sealing the coffer-
dams.
If the construction is to be accomplished at a reasonable cost,
it is important that the site be accessible for the transportation
* See Art. 43.
Art. 21 THE NATURE OF INVESTIGATIONS 3

of plant, materials,and supplies. It is often necessary to con-

struct a highway or a railroad from the nearest shipping point
to the site. The cost of constructing, maintaining, and operating
the highway or railroad is a controlling factor in the choice of
location.
A convenient site of adequate size for the construction equip-
ment and camps is essential for economical construction. The
area required for this purpose varies considerably, depending on
the size of the structure, and the desired rate of progress in con-
struction. A typical site for the construction plant is a broad,
fiat stretch of land adjoining and just below the dam site, and
above maximum high water. If the materials for the masonry
are to be delivered to the dam by cars passing over the coffer-
dams, the equipment should be nearly at the same elevation in
order to permit of moderate grades. If, however, delivery is to
be made by cableway stretched across the valley, or by chutes,
or compressed air, most of the plant will be more economical at
a higher elevation.
In cold climates, floating ice sometimes enters the reservoir
faster than the slack vater can cany it to the dam. If the quan-
tity of ice is sufficiently great, and other conditions are favorable,
an ice jam will form, and may be greater than has ever been
experienced under the natural conditions of the river. This may
cause back-water and damage to property, perhaps for miles
above. The probability of such ice jams should be given con-
sideration in fixing the site of the dam.
In general, a location below, rather than above, a community
of considerable size is to be preferred, principally because of the
necessity of providing a greater margin of safety where a failure
would result in great destruction of property or an appalling loss
of life.

Another consideration which, for storage dams, affects the

choice of general location, rather than the exact position, is the
quantity of silt carried by the
In some cases this is
stream.
enormous, and may, in the course of a few years, completely fill
the reservoir and destroy its usefulness. Sluices in the dam are
never effective in preventing silting of the reservoir, except near
the dam.
2. The Nature of Investigations. In considering a site for a
|

dam, a preliminary investigation or reconnoissance must be made,

4 INVESTIGATIONS AND SURVEYS [Chap. I

in order to determine whether or not the project is feasible. This

may include considerable surveys, if a preliminary estimate of
cost is necessary.
The final investigation is made after the project has become
active, and includes all necessary studies up to and sometimes
after the start of construction.
There are often many intermediate steps which partake of the
nature of both of these divisions of the investigations, so that a
clear distinction cannot be made. The extent to which the pre-
liminary investigations should be carried will depend on the
nature of the project and the information necessary for a complete
report.
3. Preliminary Investigations. The preliminary examina-
tion of a dam site should be made with the following objects in
view:

a. Advising as to the general adaptability of the site to the

purpose of the project, and as to the desirability of pro-
ceeding with further investigations if other features are
favorable;
b. Locating one or more possible sites for the dam;
c. Determining the character and extent of future investi-
gation to be made;
d. Obtaining information on which to form a basis for a pre-
liminary and approximate estimate of cost.

If results of instrumental surveys are not available, as is gen-

erally the case, the preliminary investigation should be as com-
plete asit is possible to make it, without going into detailed sur-

veys, as these will constitute a part of the final investigations

which will follow if the preliminary report is favorable. The
engineer should see every part of the proposed site or sites, make
full notes,and take photographs or everything that may be of
interest. Unless he is experienced in such work, he is likely to
regret, when it is too late, that he did not record certain
features
which at the time of the investigation appeared to be of insuf-
ficient importance to warrant attention.
At each prospective site a study should be made of the mate-
rials of the foundation as far as it is possible
to make such study
from surface indications. If a foundation of rock is
expected, the
elevation of rock surface, the dip, direction, and
character of the
Art. 3] PRELIMINARY INVESTIGATION'S 5

strata and the probable quantity of overburden should be esti-

mated. It is seldom that the exact character of the foundation
can be ascertained from surface appearances, but sometimes a
reasonable estimate can be made, especially if the foundation is

to be of earth, or if rock is exposed throughout the width of the

valley and the stream is small.
In case of a proposed important and expensive structure,
where an accurate estimate of cost must be made, it will ulti-
mately be necessary to complete the examination of the founda-
tion by digging tost pits and making borings, in order to ascertain
the elevation and character of the rock. In his preliminary inves-
tigation, the engineer should decide, therefore, on the number
and location and borings to be made later.
of test pits
Possible locations for construction plant and camps should be
sought, and their positions, estimated areas, and elevations
recorded. The problem of unwatering the t-ite should be care-
fully considered on the ground, and notes made as to the method
to be adopted, the approximate depth of water, the estimated
height cf coffer-dams, and the effect of the velocity of the cur-
rent and the nature of the foundation on their construction and
the work of making them tight.
Specimens of materials available for construction and to be
removed from the foundation should be obtained and preserved.
Rough estimates should also be made of the area and value of
the land to be used, of the length and cost of railroads and public
highways to be relocated, and of the area and cost of clearing and
grubbing the reservoir, if required.
Search should be made for suitable stone, sand, gravel, timber.
and other materials for construction purposes. The suitability
of materials to be excavated from the foundations for use in con-
struction should be considered. The quantity of timber required
for coffer-dams and forms for concrete is generally so large that it

is necessary to obtain this material from local sources of supply.

if it is do so. In cases where the work is of great mag-
possible to
nitude, and very remote from the nearest shipping point, it has
been found feasible to manufacture locally the cement for the
masonry.
Search should be made for high-water marks, as an aid in
estimating the magnitude of the maximum flood * and the prob-
* See Chapter XI.
6 INVESTIGATIONS AND SURVEYS [Chap. I

able highest elevation of water below the dam, in order that its
effect on the design and on the temporary and permanent works
below the dam may be properly evaluated.
The accessibility of the site should be carefully investigated.
The condition of existing highways, railroads, bridges, and navi-
gable waters should be examined, having in mind the transporta-
tion of materials and supplies. If it is probable that the exist-
ing means of transportation are inadequate, or incapable of being
made satisfactory, reconnoissance for one or more possible loca-
tions for a highway or railroad should be made.
The site should be studied for the purpose of arriving at a
tentative conclusion as to the best type of dam to be built.
Limited surveys, including cross-sections of the valley, showing
water surfaces, rock outcrops, and other information will be nec-
essary if a preliminary rough estimate of cost is required.
4. Final Investigations. The final investigations are usually
carriedon under the supervision of the engineer who has conducted
the preliminary examinations, or at least in accordance with his
recommendations. The principal items are,

a. To determine the relative merits of two or more sites for

the dam in question, so that a final location can be
adopted;
b. To settle beyond a doubt the nature of the foundation, as
affecting the safety and cost of the dam;
c. To fix the limits of the lands to be controlled for flowage,
for the sites of structures, and for other necessary pur-
poses;
d. To determine the length and character of relocation of
railroads and public highways necessary on account of
raising the water surface;
e. To ascertain the character of the Government regulations
to be observed;
/. To obtain sufficient information for an accurate estimate
of cost;

g. To fix the final location of the dam, construction equip-

ment, camps, coffer-dams, construction highways, and
railroads, as well as the probable source of materials of
construction, and all other information needful to
the constructing engineer;
Art. 4) FINAL INVESTIGATIONS 7

h. To obtain all necessary information affecting the design

of the dam.

Usually, there are not many the final exam-

sites to consider in
ination. Of all an intelligent preliminary
that are available,
examination usually reduces the problem to a consideration of a
very few, and, in many cases, where the problem is simple, it is
sufficient for a final choice of site. Thus, a rigorous investigation
of two or more sites is not always necessary in order to make a
selection; in fact, such a requirement is the exception, rather than
the rule. At any rate, it is never necessary to conduct the final
investigation throughout its whole scope for more than one loca-
tion, as the final choice of site can be made before the investiga-
tions have proceeded very far.
The foundation is one of the most important features to be
investigated. If it is to be earth, a series of test pits or borings
should be made, in order to determine the nature and extent of
treatment which will be necessary to produce a stable and prac-
tically impervious footing. The investigations should include tests
to determine the probable bearing power of the earth, or the number
and length of bearing piles which will be necessary, and the nature
and length of sheet-piling to be driven to prevent excessive leakage.
Except for low dams, the foundation should be rock. For
rock the investigation should fix, not only the depth of the ledge
below the surface, but its nature or suitability for a foundation.
This is usually accomplished by test pits or wash-borings to rock
surface and core drilling into the rock. With the latter it will be
possible to determine, from the samples produced, not only the
character of the rock, but the location of pervious or soft pockets,
seams, and other faults, by the action of the drills during the boring
operations. Water forced into the drill holes under pressure will
indicate the perviousness of the foundation.*
Too much stress cannot be placed on the necessity of making
extensive investigations of the foundations before construction
begins, if an accurate indication of the construction difficulties,
the cost of the work, and the tightness of the foundations is
desired. A
very good description of modern methods of conduct-
ing wash and core borings is to be found in Chester W. Smith's
" Construction of Masonry Dams."t
* See Art. 61. t McGraw-Hill Book Co., Inc., 1915.
8 INVESTIGATIONS AND SURVEYS [Chap. I

The problem of determining the proper location and depth of

borings and the interpretation of the results obtained is very dif-
ficult. It is impossible to give a general rule for the spacing and
depth of borings, as each case presents a different problem. All
probable zones of weakness, such as the plane of contact between
intrusive igneous rock and rock of sedimentary origin should be
thoroughly investigated.
It should be remembered that a valley, in the making, will
usually be started at the weakest part of the geological formation.
One is, therefore, likely to find at every dam site a geological
reason for the particular location which the stream has adopted,
and such indications will assist materially in the determination of
the zones requiring more thorough investigation.
The spacings and depth of borings will be governed, not only
by the geological formation, but also by the height of the dam, and
the extent to which the importance of the work will justify the

expense of the investigation. A careful study of the data obtained

from the borings as they are drilled will influence the depth and
location of additional borings.
In important work, particularly for high dams, the services of
an expert geologist should be obtained, and his recommendations
should be given full consideration. The danger of excessive leak-
age is not confined to the vicinity of the dam; such leakage may
extend for considerable distances on each side of the dam site,
and even through the ridge to an adjacent valley lying at a lower
elevation.
Surveys should be made of the lands necessary for the site of
the reservoir and of the dam, and for various other purposes. The
on maps of adequate
results of these surveys should be indicated
scale, depending on the value of the lands to be obtained and,
therefore the accuracy with which such information must be
shown. These maps should indicate:
a. The original edge of the stream at low, ordinary, and
high water;
b. The edge of the proposed reservoir at low, ordinary, and
high water, usually with a sufficient number of inter-
mediate contours to enable an estimate of the storage
contents to be made;
c. The location of the dam, the sites for construction equip-
ment, camps, and other proposed structures;
Art. 4] FINAL INVESTIGATIONS 9

d. The property lines of the lands required. For the res-

ervoir,a distance from the original shore line to the
edge of the survey equal to two or three times the dis-
tance from the original to the new shore line will usually
be However, where part of the land required
sufficient.
includes a large percentage of an individual estate, it may
be necessary to purchase the entire tract. In this
event the surveys should be extended to cover the total
area.
e. The location of all railroads, public highways, bridges,
dams, important structures, estates, enterprises, and
other items of importance on and in the neighborhood
of the proposed reservoir and dam site;
/. The proposed new location of railroads and public high-
ways which must be relocated .on account of raising the
water surface, showing the boundaries of the lands
required for such purpose;
g. A brief description of each parcel of land required should
be placed on the maps, but a complete description should
be given in a supplementary report, including the ap-
proximate value of lands, water rights, structures, etc.,
to be taken over.

Government requirements will include the observance of ripar-

ian rights, if the water is to be retained or diverted; the necessity
of providing log chutes, fish ways, or navigation locks; regulations
covering the design and construction of dams, and allied matters.
The Federal Government has jurisdiction over all navigable
streams; also in most States it is required that the plans for the
dam must receive the approval of the State Engineer or other
official before construction starts.
A map of the dam site should be prepared showing,

a. Contours of the natural surface within and near the area

to be occupied by the dam, construction equipment,
camps, and other proposed structures. The required
contour interval will be determined generally by the
height of the dam. Usually, an interval of about 5
per cent of the height will be found to be sufficiently
close.
10 INVESTIGATIONS AND SURVEYS [Chap. I

6. The location of the river, existing and proposed structures,

coffer-dams, test pits, borings, rock outcrops, high-
water marks, proposed quarries, sand banks, and all
other items of use to the designing and constructing
engineers.

This map should be accompanied by cross-sections of the site

at frequent intervals, showing the depth of test pits and borings,
and the estimated elevations of all . underground materials and
their nature.
For estimating purposes, contour maps and profiles of the rail-
roads and public highways to be relocated, as well as the railroad
or highway for providing access to the site, may be necessary.
The results of the final investigations should be incorporated
in a complete report, and this should be carefully preserved, as
years may elapse between the time of making such investigations
and the beginning of construction operations.
 CHAPTER II

THE CHOICE OF TYPE OF DAM

5. General Considerations. It is not within the scope of

this volume to describe in detail all the advantages and disad-
vantages of the different types of dams which have been built,
but merely to outline briefly the adaptability of each type, in
order that the reader may be able to comprehend the relative
limitations of masonry dams.
The usual types of dams may be summarized as follows:

Solid gravity masonry dams,

Hollow gravity masonry dams,
Arched masonry dams,
Earth and rock embankments,
Timber dams, and
Other types.

Many combinations of these types have been constructed.

Thechoice of the type best suited to a particular location or
use is a matter on which experienced engineers will often differ
considerably, and is quite often purely a matter of judgment and
experience. However, an intelligent study of the existing condi-
tions and requirements will assist materially in the choice.
Safety, of course, is the first consideration. It is impossible to
build with safety some types of dams under certain conditions of
foundations and other characteristics of the site. Consideration
of this question will often decrease considerably the number of
possible types from which to choose.
The first cost of the structure, as affected by the availability
and price of construction materials and other characteristics of
the site, is, perhaps, of next importance.
The choice of type is often limited by the funds available for
construction of the dam and other requirements of the project.
It will sometimes be found that the difference in cost between an
expensive, permanent dam and an inexpensive structure of short
H
12 THE CHOICE OF TYPE OF DAM [Chap. II

lifeand high maintenance charges, if set aside at compound inter-

est, willbe more than sufficient to provide funds for the higher
maintenance cost and a sinking fund to cover the rapid deprecia-
tion of the less expensive type. It may be said, however, that, in
general, the most permanent dam will be found to be the most
economical, and it is usually adopted for ordinary sites, unless the
structure is for temporary use, or if sufficient funds are not avail-
able.
A
comparison of the several types of dams follows:
Solid Gravity Masonry Dams.*
6. There is no type of dam
more permanent than one of solid masonry, nor does any other
type require less for maintenance. It is adaptable to all localities
except where a sufficiently impervious cut-off at and below the
surface is impractical of attainment, where there is danger of con-
siderable uplif t, or where the low bearing strength of the foundation
prohibits its use. It is imperative that high masonry dams be
built on rock foundations. Low dams of this type are some-
times built on earth or piles, but such support, for dams more than
about 30 ft. in height, should be adopted with caution.
The solid gravity masonry type, being the most common of
all masonry dams, is the safest, according to the popular idea;
and, in this respect, has an advantage when the enterprise is
affected, to a large extent, by public opinion, as in municipal or
other public works.
The between solid and hollow gravity
difference in first cost
masonry dams the subject of considerable debate. The solid
is

dam requires less cement per cubic yard of concrete, less form
work, less expense in placing concrete, and has no steel reinforce-
ment. On the other hand, the hollow dam requires considerably
lessconcrete per linear foot of dam.f It is the author's opinion,
based on a number of comparative estimates, that for a remote
location,where materials of construction are expensive, the hollow
type will usually cost less to build than the solid type; but, in
an ordinary location, comparatively near a railroad, where there
is a good quarry, and a sand bank is convenient, the
reverse is
true.
The solid gravity masonry dam will usually cost more than a
timber dam. However, this may not be the case if a first-class,
* See Chapters VI and VII.
t Usually from 35 to 40 per cent of the concrete required for a solid
dam.
Art. 7] HOLLOW GRAVITY MASONRY DAMS 13

rock-filled, timber crib dam

adopted at a site where a coffer-
is

dam is required for its and if timber is expensive.

construction,
An earth or rock embankment will almost always cost con-
siderably less than any form of gravity masonry dam, if materials
for the former are found convenient to the site. Therefore, if
conditions admit of an embankment, that type of dam is usually
to be preferred. The limitations of embankments will be men-
tioned later.
There is considerably less material in an arch dam than in any
other masonry type, and consequently it will cost much less to
construct. However, as will be pointed out later, a site suitable
for an arched dam is the exception, rather than the rule.
7. Hollow Gravity Masonry Dams. Most hollow dams have
been constructed of reinforced concrete of the types described in
Chapter VIII.
Compared with most methods of construction, reinforced
concrete is in its infancy. Although its durability has not been
tested for as long a period as plain concrete, it has, thus far, shown
itself to be as permanent as can be desired.

When the element of time is a governing consideration, the

hollow dam possesses some advantage over the solid dam, because,
there being less concrete to deposit, it can be constructed in a
somewhat shorter period.
Turbines and other apparatus have often been placed within
hollow dams, thereby making a saving in the necessary housing for
such appliances.
A hollow dam is often adopted in preference to a solid gravity
masonry dam in localities where considerable uplift * under the
latter type would be expected. The hollow dam has a distinct
advantage, owing to the fact that the narrow walls and buttresses
are subjected to a practically negligible uplift pressure, the water
under the dam having a direct exit.
Another advantage claimed for the hollow dam having an up-
stream face with considerable batter, is that it is impossible for
it to overturn; as the resultant of all forces, for any depth of water,
falls well within the base. However, this advantage is more
fanciful than practical, as either type, if properly designed, should
give no cause for worry in that respect.
Hollow dams, being lighter per square foot of area covered,
* See Art. 15. fi ^"U
14 THE CHOICE OF TYPE OF DAM [Chap. II

can, by having spread footings, be made to exert less unit pres-

sure on the foundation than solid dams. For this reason the former
type is sometimes adopted where the requisite support for a solid
dam is lacking.
8. Arched Masonry Dams.* This type is adaptable when
the length is small in proportion to the height, and when the sides
of the valley are composed of good rock which can resist the end
thrust. It is the ideal permanent dam, containing much less
material than other masonry types, and being equally permanent,
it is always adopted where conditions permit. Unfortunately,
however, sites suitable for this type are seldom found.
The weight of the arched dam is not counted on to assist
materially in the resistance of external loads. For this reason
there is always sufficient weight of masonry to resist any possible,
uplift on the base.
Combinations of arch and solid gravity masonry dams are
common for sites where the length is thought to be insufficient
to permit of the adoption of the pure arched type. Such dams are
designed to resist the loading by gravity, but are curved in plan
and designed with a smaller margin of safety than if straight
and with no arching possible.
9. Embankments. When plenty of materials are convenient
to the side, embankments can usually be built for considerably
less cost than any form of masonry gravity dam. The use of
this type, however, is often limited by the necessity of providing
a more suitable spillway for the passage of floods. It is not safe
to allow water to spill directly over the embankment, even if it

is well paved, unless the volume of the flood per linear foot of
crest is small. Therefore a spillway of more suitable character is
a necessary adjunct. In some instances such a spillway would
require most, if not all, of the available length of the dam; in
which case an embankment would be out of the question.
The quantity of seepage through pervious material is propor-
tional to the distance the water is required to travel. An earthen
embankment, having the longest base in proportion to the height,
is particularly adaptable to sites having pervious foundations.
With proper maintenance, the embankment dam should be as
permanent as the best. The necessary maintenance charges are
comparatively high during the first year or two, but become rapidly
* See Chapter IX.
Abt. 11] OTHER TYPES 15

the structure settles into its final position and becomes well
less as
compacted, tight, and overgrown with proper vegetation to with-
stand wash from rains.
Earthen dams possess a distinct advantage in landscape work
where it is desired to change as little as possible the appearance
which Nature has given to the site.
10. Timber Dams. A timber dam is the ideal temporary
type; although when well designed, constructed and maintained,
it may last fifty years or more. Maintenance charges, however,
are very high, compared with other types.
Timber dams are seldom very tight. In fact, a small leakage
is necessary for the proper preservation of the timber. Such leak-
age, however, is of importance only when the value of the stored
water is exceptionally high.
This type is often used on soft foundations where masonry
dams are out of the question, as a slight settlement, which, in the
former would be permissible, would, in the latter, be an element
of considerable danger.
Owing to a scarcity of funds, a timber dam is sometimes
adopted with the intention of utilizing it later as a part of the
necessary coffer-dam for the construction of a more permanent
structure.
11. Other Types.Various other types of dams have been
designed and built. These include structural steel dams, pecu-
liarly shaped masonry dams, the many forms of movable dams,
and others. These, however, may be considered as either struc-
tures of unique character, suitable for special conditions not ad-
mitting of comparison in the general sense, or types which were
the creation of fanciful engineers of radical tendencies.
 CHAPTER III

FORCES ACTING ON DAMS

12. Nomenclature. The following nomenclature will apply,

in general, to all parts of the text. Special nomenclature, applic-
able to arch dams, is given in Chapter IX. Unless definitely-
mentioned, all forces are stated in pounds and all dimensions in
feet.

W =A vertical force; positive when directed downward;

P =A horizontal force; positive when directed toward
the left;

P t =Ice pressure per linear foot of dam;

R = A resultant of forces;
2(W) =The algebraic summation of the vertical components
of all forces acting on the dam above a given joint,
including uplift, but excluding the reaction at the
joint; positive when directed downward;
2(P) = The algebraic summation of the horizontal components
of all forces acting on the dam above a given joint,
excluding the reaction at the joint; positive when
directed toward the left;
2(Wa;)=The algebraic summation of the moments, about a
given point, of the forces contained in the summa-
tion, 2(TF); positive when counter-clockwise;
2 (Pa;) =The same for the summation 2(P);
A =An area, in square feet, or the area of a water-shed,
in square miles;
a =The distance from the top of the dam to the water
surface;
C =A constant;
c =The percentage of area of joints or base subjected
to uplift;
e =The eccentricity of a loading (see " Irregular Bases,"
Art. 22, and Fig. 12) ;

e =A subscript used to represent the condition of empty

reservoir;
16
 ;

Art. 12] NOMENCLATURE 17

r =a subscript used to represent the condition of full

reservoir;
/' =The actual coefficient of static friction at a given joint
or the base;
/ =The coefficient of static friction of the same materials
as indicated by well-dressed test specimens;
g = The acceleration of gravity = approximately 32.2
h =A vertical distance, a height of masonry, a head of
water, etc.;
he =The measured head on a spillway crest;
h, =The total head on a spillway crest;
K =The head corresponding to a given velocity;
H =The total height of a dam above a given elevation;
I =The moment of inertia of a figure;
k =The percentage of voids in earth or silt, expressed as
a decimal;
L =The top width of dam;
= The known length of a horizontal joint;
=The unknown length of the horizontal joint next below;
= The total length of a spillway crest;
= The net, or effective, length of a spillway crest, Art. 43;
= The average width of the channel of approach to a
dam;
m =A distance to the right or left of the center of gravity
of a figure; see definition given in " Irregular
Bases," Art. 22;
N =A flood coefficient, Art. 65;
n =The number of complete end contractions on a spill-

way crest;

p =Unit pressure or compressive stress, in pounds per

square foot;
p' =The same at the down-stream extremity of the base;
p" =The same at the up-stream extremity of the base;
(The foregoing system of primes and double primes
applies also the following special values of p.)

pr =The unit vertical reaction * of the foundation at a joint

or the base, exclusive of uplift pressure;

p e =The unit vertical compressive stress * at a joint or the
base, inclusive of uplift pressure;
* See foot-note, p. 43.
 ;

FORCES ACTING ON DAMS [Chap. HI

18

p, =The unit maximum inclined compressive stress in the

masonry or the foundation;
p„ =The unit effective uplift * on a joint or the base;
=The unit normal pressure of water or earth on the face
pn
of the dam;
total quantity of water passing over the
spill-
Q =The
way crest, in cubic feet per second;
=The maximum flood from a given drainage area, likely
Qm
to be exceeded only once in T years;
q =The quantity of water passing over each linear foot
of effective spillway crest;
r =The radius of a circle;
S = A factor of safety, Art. 25;
t =A period of time, in seconds;
T =A period of time, in years;
u =The horizontal distance from the down-stream ex-
tremity of a joint to the point of intersection of
the resultant, R, with that joint;
v =A velocity; in feet per second;
w =Unit weight;
w\ =Unit weight of masonry;
W2 =Unit weight of water;
W3 =Unit weight of earth;
x =In Also used as the lever
general, a vertical distance.
arm both vertical and horizontal forces;
of

y =The horizontal distance from an origin of moments to

the up-stream extremity of a joint or the base;
z =The horizontal distance from an origin of moments
to the point of intersection of the resultant, R, with
a joint or the base;
a = The angle of repose of earth;
8 = The angle of inclination with the vertical, f of the result-
ant, R, of the forces, 2(W) and 2(P)
<j> =The angle of inclination, with the vertical,* of the face
of the dam;
<f>'
= The same for the down-stream face at a given elevation;
<f>"
= The same for the up-stream face at a given elevation.
* See Art. 15 and foot-note, p. 45.

t This is the common

definition, as, in general, the joints and bases are
horizontal. For inclined joints or bases, the angles 8 and should be measured
<f>

from a normal to the joint or base.

Art. 14] EXTERNAL WATER PRESSURE 19

13. General Considerations. The first consideration in

designing a dam is the determination of the nature of the forces
acting on the structure. These forces may be considered as con-
sisting of the following:

a. Water pressure,
b. Earth pressure,
c. Atmospheric pressure,
d. Ice pressure,
e. Wind pressure,
f. Wave pressure,
g. Weight of the dam,
h. Weight of the foundation,
i. Reaction of the foundation.

The nature of most of these forces, unfortunately, is such that

they do not admit of exact determination, and their amounts,
direction and location must be adopted by the designer after a
thorough consideration of all obtainable facts bearing on the case,
and the exercise of his best judgment, based on his experience and
that of others who have had to deal with similar problems.
must always be borne in mind that conditions in no two dams
It
are alike,, and that a general theory must never be applied to a
particular case without thought as to the possible need of modifica-
tion to suit the conditions peculiar thereto.
14. External Water Pressure. The weight of a cubic foot of
fresh water has been determined to be

62.42 lb. per cu. ft. at 32° F.

62.26 lb. per cu. ft. at 75° F.
62.00 lb. per cu. ft. at 100° F.

The weight usually adopted in the design of dams is 62.5 lb. per
cu. ft.

The total pressure, P, of quiet water on any vertical submerged

plane of area A is

P=w2Ah3 , (1)

where W2 is the weight of 1 cu. ft. of water and hz is the distance

from the center of gravity of the plane to the surface of the water.
The force, P, will be horizontal.
20 FORCES ACTING ON DAMS [Chap. Ill

rectangular
In Fig. 1, let 1-2 represent a submerged vertical
perpendicular to the paper, and
plane of unity width, measured
having its top edge parallel to and a distance, hi, from the surface
plane unity, the length, 1-2
of the water. As the width of the is

= h, will be a measure of its area, A*

 '

Art. 14] EXTERNAL WATER PRESSURE 21

the total pressure, P. The force, P, will pass through the center
of gravity of the area, 1-2-3-1, which is a vertical distance
above point 2 equal to
_3hh+h2
x=
6/ii+3/i

In the design of dams it is found convenient to deal with hori-

zontal and vertical forces only. If the plane, 1-2, is inclined, as
in Fig. 2, the resultant total pressure, R, on the plane may be
resolved into horizontal and vertical components, P and W. The
horizontal component, P, will be equal to the pressure on the pro-
jection, 2-5, of the plane, 1-2, and its amount and location can be
calculated from Eqs. (2) and (3). The vertical component, W,
will be equal to the weight of water above the plane, 1-2, namely,
within the boundaries, 1-2-6-7. The force, W, will pass through
the center of gravity of the figure, 1-2-6-7.
Water Surface

g Water Surface

Fig. 2. Fig. 3.

Inasmuch as it is possible for a dam to be entirely submerged,

with water pressure on every square foot of surface, we will now
investigate the forces due to water pressure in the following order:

a. On the up-stream face,

b. On the top,
c. On the down-stream face,
d. On the bottom.

In Fig. 3, the horizontal component, P, of the total water pres-

sure on the up-stream face of the dam is equal to the total pressure
on the plane, 3-8, as indicated by Eq. (2). In this case, hi being
equal to zero, and fe being equal to h,

P = ^# (2a)
22 FORCES ACTING ON DAMS [Chap. Ill

The distance, x, from point, 3, to the force P, may be found

from Eq. (3). For this case there results

h (3a)

The vertical component,equal to the weight of water

W, is

within the area, 2-3-8-7, and passes through the center of gravity
of that area.
In Fig. 4, the height, h, of the dam is not as great as the depth
of water, h 2 so that the water is constantly spilling over the crest.
,

In this case the dam will not be subjected to horizontal water

"Water Snrface

'Fig. 4.

pressure on the plane, 1-8, above the crest. On account of the

velocity of thewater passing the plane, 8-9, which approaches
very nearly spouting velocity in well-designed dams, the vertical
component of the pressure on the plane, 9-10, and on the top of
the dam is always neglected. To be on the safe side, however,
the horizontal component of the pressure on the plane, 9-10, is

included.
Considering, again, vertical and horizontal forces only, we
have a vertical force, W, equal to the weight of water within the
limits, 7-8-2-3, passing through the center of gravity of that area,
and a total horizontal water pressure, P, equal to the pressure on
the plane, 11-3, found by Eq. (2), and located by Eq. (3).
The heads, hi and fa, should correspond to the measured head
Aut. 14] EXTERNAL WATER PRESSURE 23

on the dam; that is, to the water surface in the channel of ap-

proach sufficiently remote from the dam to be beyond reach of the

surface curve.*
An additional head, equal to twice. the velocity head of the
water in the channel of approach, should be added to allow for
impact on the face of the dam.f

If <'i=t^ is the velocity of approach, the average unit impact

pressure is approximately

p-211^,-2 — =_ (4)

and the total impact pressure on the dam is

2
lC-2l'\ h
P' = 2w2 hji = -

where g is the acceleration of gravity (about 32.2), q is the discharge

per lineai- foot of crest, and icj is the unit weight of water.
Fig. 5 represents a dam containing a barrier, such as flash-
boards, projecting above the crest, or closed gates held between
piers." In such cases, the weight of the water within the bound-
aries, l-2-3~i, should be included in the calculations as well as

the horizontal pressure on the plane, 1-2.

The stream of water indicated in Fig. 4, flowing down the face
of the dam, is, on account of its velocity, considered as transmit-
ting no pressure to the dam.
Pressure on the dam from water standing in the lower pool, as
indicated in Fig. 5, may be treated in the same manner as described
for water pressure on the up-stream face.
In the case of a spillway, tail-water does not always exert a
pressure on the dam, as its depth adjacent to the dam is often
practically eliminated by the impact of the water spilling over the
crest. J This condition is indicated in Fig. 4.

Air. A. H. Gibson, § in Ins discussion of standing waves in non-

* See Art. 43.

t See Gibson's "Hydraulics," Art. 96; D. Van Nostrand Co., 1908.
J See Karl R. Kennison's paper "The Hydraulic Jump in Open Channel
Flow at High Velocity," Transactions, Am. Soc. C. E., Vol. LXXX, p. 338.

§ Gibson's "Hydraulics and its Applications," Art. 86. J. Wiley & Sons,
1908.
24 FORCES ACTING ON DAMS [Chap. Ill

uniform channels, states that the conditions under which this

phenomenon will occur are

v
2
=>gh3 ,

and
2
j2h 3 v 2
3 3 ,h h
hs
-V~r + T-2--
For spillway dams, v 2 is always greater than gh 3 hence we , may
expect a reduction in depth of tail-water, provided hs is less than
the value given in the second equation,

Water Surface

Head-water

Fig. 5.

A more convenient form may be obtained by writing the

second equation as follows:

, /2feV ,h3 2 h3

But h3 2v 2 = q2 where
, q is the discharge, in cubic feet per second,
'
per linear foot of crest. Making this substitution, there results

/*5
9
2
| h2 h
-VI6.M3 4 2' (5)
Abt. 15] INTERNAL WATER PRESSURE 25

The any given

thickness, hs, of the sheet of falling water for
head, hi,and height of dam, h, may be found from Figs. 25 and 26:
A method of determining the discharge, q, corresponding to the
head, hi, is given in Art. 43.
If As is greater than the value given in Eq. (5), the water flow-
ing over the spillway will expend its energy in creating eddies,

and will cause no reduction in tail-water level.

If hs isequal to this value, a standing wave, 14-15, (Fig. 4),
or " hydraulic jump," will occur at the toe of the dam.
If hs is less than this value, the standing wave will occur at a
distance down-stream, where the depth, hs, has increased an
amount sufficient to fulfill the conditions of Eq. (5).
Therefore, if, in any example, it is found that hs is less than
the value given in Eq. (5), the pressure of the water on the down-
stream face of the dam may, theoretically, be neglected.
The foregoing has been based on theory alone. Although
Gibson's equation has been fairly well substantiated by experi-
ments on a small scale, the actual height, hs, of the standing wave
for large structures may differ considerably from that calculated.
It isrecommended, therefore, that when the depth of tail-water is

within 20 per cent of the value given in Eq. (5), the pressure of
tail-water should be considered doubtful and the dam tested for
stability with and without it.

For a practical application of this theory see Art. 47.

Obstructions or " baffles," have often been placed in the tail-

race, adjacent to the dam to prevent the standing wave from occur-
ring below the apron and the swiftly moving water from having
access to the unprotected portion of the foundation. Such an
arrangement is indicated in Figs. 6 and 7. This feature will be
further discussed in Art. 61.
Water Pressure. Uplift. If the foundations
15. Internal
are pervious, an upward pressure, or " uplift " will occur on the
base of the dam, as indicated by the force, W, Fig. 5. The
amount and location of this force depends on the relative pervious-
ness of the foundation at various points and the details of the
dam.*
In order to comprehend more clearly the principles of uplift
under solid dams, consider a solid dam to be slightly raised from
* Such details refer to the difference between solid and hollow dams, and
the use of drainage systems.
26 FORCES ACTING ON DAMS [Chap. Ill

si
«

I
Pi

Ia
02

CO

O
Abt. 15] INTERNAL WATER PRESSURE 27

|
"3.

a
 .

28 FORCES ACTING ON DAMS [Chap. Ill

its foundations, so that water flows through from the upper to

the lower pool. The laws governing the pressure over the base
are the same as those for the flow through pipes. In this case,
neglecting the inappreciable loss at entrance, the pressure would
diminish uniformly from W2I12 (Fig. 5), at point 6 to W2I15 at point
5. The total.uplift would be

W = W2l —~—
The location of the force, W, can be obtained (for this as well
as for other cases to be considered later), by locating the center of
gravity of a force diagram constructed in the manner described
for Fig. 1.

Assume, however, that although the dam is slightly raised, a

portion of each square foot of the base has absolute contact with
the foundation. The force, W, would then be reduced an amount
in proportion to the percentage, c, of such area not in contact,
Then,

w=CW2 ib+h (6)

If there is an impervious barrier at point 5, the pressure over

the portion of the base not in contact would be uniform and equal
to u>2h,2, and the total uplift on the base would be

w = CW2II12
If the impervious barrier is at point 6, the uplift would be uni-
form and equal to
W = CW2lhs

If the barrier is at any point between 5 and 6, the uniform unit

pressure above the barrier would average cw-^,2, and below the
barrier CW2I15.
In practice, where the joint between the dam and the founda-
tion, or other horizontal joints in the dam, are conducive to the
passage of water, uplift will occur, and the foregoing laws will
apply. The conditions are also the same for a horizontal seam in
the rock below the base, with the exception that the points, 5 and
6, should be considered located at the entrance and exit of the seam,

respectively.
Art. 15] INTERNAL WATER PRESSURE 29

Most engineers recognize the likelihood of uplift. The ques-

tion of amount and distribution is, of course, impossible to deter-
mine exactly, and, of late years, this question has been the source
of much debate. In order to bring the subject before the engineer-
ing profession for discussion, the late C. L. Harrison, on Decem-
ber 20, 1911, presented a paper before the American Society of
Civil Engineers, in which he made some brief suggestions regarding
the methods of estimating the amount and distribution of uplift
on the foundations and horizontal joints of dams. The paper was
quite freely discussed by some of the most prominent engineers.
It is quite surprising to note the difference of opinion on the sub-
ject, varying practically from one extreme to the other.
The following are extracts from Mr. Harrison's paper * and
his analysis of the discussion thereon:

For convenience in discussing this subject, reference is made particularly

to masonry dams on rock foundations. The principles involved will apply
equally to other foundations and to dams built of other materials. The
upward pressure may be due to water getting into the foundation of the
dam or into the dam itself.

Foundations vary so much in character that it is necessary to study

each particular site before deciding to what extent water may get into them.
(1) In the case of a foundation of hard, sound rock, without either
horizontal or vertical seams, there is no reason to expect that water will
get into it and produce an upward pressure, and, in the design, no allow-
ance should be made for it. In such cases the junction between the masonry
and the foundation can easily be made water-tight.
In the case where the foundation is stratified with well-defined
(2)
horizontal seams, and the dam is located near a fall or rapids in the stream,
so that the water may flow from the seams at the toe of the dam as freely
as it enters them from the reservoir, the upward pressure will be approx-
imately equal to the static head at the heel and gradually decrease to zero
at the toe of the dam.
(3) Take a foundation similar to the foregoing in every respect except
that the water in the seams of the rock cannot escape freely near the toe
of the dam, but must flow some distance down-stream through rock or
other materials before it reaches the surface of the ground, or must rise ver-
tically to the surface. Then the upward pressure at, the heel will be equal
to the static head, and that at the toe will be equal to the head required
to overcome the resistance to the water escaping at that point.
While these three cases present well-defined conditions, it is probable
that at most sites the conditions will he between those presented in Case
1 and Cases 2 and 3, that is, the water will not be in the foundation through-

* Tramactiotis, Am. Soc. C- E., Vol. LXXV, pp. 142-225.

30 FORCES ACTING ON DAMS [Chap. Ill

out its entire area, but will cover only a part of this area. This makes it
necessary to study the foundation carefully at each site in order to deter-
mine to what extent water may get into it. When this upward pressure
exists,weight must be added to the dam by additional masonry to counter-
balance it. Generally, it will be found cheaper to make large expenditures
to provide a cut-off in the foundation, which will not only reduce the uplift,
but will also save the water. Such a cut-off should be located at the heel
of the dam. If it is located under the middle of the dam, there would be
an upward pressure under the up-stream half of the dam, due to the full
head of the water in the reservoir.
* * * ,* * * *

In order to determine what allowance to make for pressures due to

water which gets into the dam itself, one must first decide on the character
of the construction. With suitable stone, sand, and cement, it is possible
to build a masonry dam which will have no horizontal cracks or seams, and
it is also possible to provide against vertical cracks, to » large extent, by

expansion joints. Water in vertical cracks, however, does not produce an

upward pressure. In such structures very little, if any, allowance should
be made for the upward pressure due to water getting into the masonry.
If the materials for building water-tight masonry are not to be had at
the site of the dam, and it is very expensive to import them, it is generally
advisable to adopt a, different class of masonry, which will probably be more
pervious and also more difficult to construct without horizontal cracks or
seams, thus allowing the water to enter the dam, and resulting in upward
pressures. The extent of such pressures will depend on the character of the
masonry and the care with which it is built, all of which must be known
before an estimate can be made of the extent to which the water will get
into the dam. The effect of this upward pressure, however, must be counter-
acted, either by increasing the section of the dam or by increasing its height
above the water level in the reservoir, or by both. In many cases it may
be advisable to provide drainage wells near the up-stream face to intercept
the water and carry it off through pipes at the toe of the dam, thus reducing
or eliminating its effects in the main body of the dam. After determining
the type of masonry to be constructed, it is still a question of judgment,

the dam

An
1.
********
based on observation,
will be.
tests, and experience, as to what the upward pressure

analysis of the discussion indicates the following conclusions:

For any stable dam the uplift in the foundation cannot act over the
in

entire area of any horizontal seam, and in the masonry it cannot act over the
entire area of any horizontal joint.
2. The dam can never be more, and
intensity of uplift at the heel of the
is generally than that due to the static head. Also, this uplift decreases
less,

in intensity from the heel to the toe of the dam, where it will be zero if the
water escapes freely, and will be that due to the static head if the water is
trapped.
3. The uplift in the foundation should be minimised by a cut-off wall,
 .

Art. 15] INTERNAL WATER PRESSURE 31

under-drainage,* and grouting when applicable; and in the dam itself by

using good materials and workmanship, and by drainage when advisable.
4. The design should be based on the conditions found to exist at each
site aftera thorough investigation by borings, test pits, and otherwise, and
modified if found necessary after bed-rock is uncovered.

Mr. Harrison's paper should be read by all who are interested

in the design of dams.
In practically all the designs of solid gravity dams embodying
upliftwhich have come to the author's attention, the assumption
has been a pressure varying uniformly from wohz (Fig. 5), at the
heel to N'i/is at the toe, Eq. (6), being directly applicable. The
only difference has been the percentage, c, of the base assumed to
be subjected to the pressure. In Table I is given a list of some
recently constructed high dams in the design of which extreme
values of c were adopted.

TABLE I

Percentage of Area of Base Subjected to Uplift, Adopted

for Various Dams

Dam. Value of c. Per Cent.

Wachusett 66
Cross River. . . . 66
Elephant Butte 33 (approximately)
Olive Bridge . . 66
Kensico 66
Lock Raven . . . 66

All the dams mentioned in Table I contain drainage systems,

cut-offs, and other features designed to reduce, as far as possible,
uplift on the base. It must be remembered, however, that these

dams are among the largest and most important structures built
in this country, and failure would undoubtedly result in immense
loss, dam but to other structures below it, and, in
not only at the
most an appalling loss of life. For these reasons an unusu-
cas?s,
ally large margin of safety has been provided, not only in the
assumptions of uplift, but in other features. Many dams, includ-
* Author's Note Except for very high dams, comparatively few have
:

been provided with under-drainage.

32 FORCES ACTING ON DAMS [Chap. Ill

ing the largeNew Croton Dam of the New York City Water Sup-
ply system, contain no provision for uplift in the design. On the
other hand, the failure of dams on very poor foundations, in a
number of cases, may be attributed to insufficient provision for
this force.
The author is of the opinion that a value of c = f is the extreme
limit for dams on fair rock foundations, and for the most impor-
tant structures; and that ordinarily a much lower value may
be adopted.
It is impossible to recommend definite values of c for the many
classes of foundations which exist, no two of which are alike. The
final choice must be made in accordance with the judgment and
experience of the engineer after a thorough investigation of the
site after the foundations have been exposed.
For earth foundations, a value of c = 1.0 is necessary, but the
distance which the water must travel may be considered as being
the length of the base plus twice the depth of the impervious
cut-off.*
Little, if any, uplift can exist in hollow dams of the types de-
scribed in Chapter VIII. The pressure of water finding a passage
through the deck or cut-off will be almost immediately relieved
through the sides of the buttresses before it can penetrate any
appreciable distance. The author knows of no case in which an
assumption of uplift was used in the design of such structures.
Mr. Arthur P. Davis f gives the following general rules for the
relative perviousness of several classes of foundations:

The determination of the perviousness of natural formations is one of

the most difficult things in Nature. Any examination of such formations
which disturbs them changes the conditions which it is desired to know.
For this reason, it is necessary to allow a large factor of safety in any estimates
which involve this factor.
In general, it may be said that water will more readily follow seams
or bedding planes than devious paths through the material of the rock. It
follows that it will pass more readily and in larger volume in the direction of
stratification than in a direction normal thereto. Similarly, stratified rock
will permit percolation more easily and in greater volume than good, massive
rock, such as granite.
Granular rock, such as sandstone, is. likely to transmit more water through
the rock itself than one of denser or finer grain, such as limestone or shale,

* See Art. 62.

f Transactions, Am. Soc. C. E., Vol. LXXV, page 208.
Art. 16] EARTH PRESSURE 33

but no exact rule of this nature can be laid down, because there are many
varieties of each kind of rock, with various percolating capacities. In gen-
eral, however, the following rules may be taken as a rough guide:
1. Massive or crystalline rocks, such as granite, gneiss and schists, will

transmit water less freely than those of sedimentary origin.

2. Stratified rocks will transmit water much more readily in the direc-

tion of stratification than transverse thereto.

3. In the direction normal to stratification, sandstone will generally

transmit water more readily than limestone, and the latter more readily
than shale.
4. Stratification on a plane approximately horizontal is the worst pos-

sible condition for introducing upward pressures beneath a dam. Con-

versely, the most favorable position in this respect for stratified rock is in
vertical beds.

It is rational to assume that, in some cases, there may be less

proportionate uplift on horizontal joints above the base than on
the base. However,
customary to adopt an uplift on all hor-
it is

izontal joints equal, proportionately, to thatassumed for the base.

This wijl lead to no appreciable error, and will greatly simplify
the calculations, because, if the assumption of uplift for the joints
were less than that used for the base, the upper part of the section
designed would not suit those parts of the structure where the
foundation is at a higher elevation.
The effect of silt deposits in reducing uplift pressure is dis-

cussed in Art. 16.

16. Earth Pressure. Practically all streams transport silt

and gravel, particularly during floods, when, in some cases, the

quantities are enormous. The construction of a dam across a
stream results in a stretch of slack water which causes the material
to deposit. It fills the upper end of the reservoir first, and grad-
ually advances until it reaches the dam.
Quite often, sluices are constructed in the lower part of the
dam, and, if periodically flushed in the proper manner, this limits
the depth of such deposits adjacent to the dam.
An earth fill is sometimes deposited against the down-stream
face of the dam, and may or may not be submerged.
The horizontal component of the silt or earth pressure, P"
(Fig. 5), exerted on the dam may be derived from Rankine's
well-known equation:

'
,_ w3 h
2
/ l-sma \
' K
, 2 \l+sin«/ '
34 FORCES ACTING ON DAMS [Chap. Ill

Where w3 = the weight, in pounds per cubic foot, of the silt or

earth, in air or submerged, as the case may be;
A=its depth, in feet;
and a = its angle of repose.

2h
P" is located a distance, —o , from the surface of the earth, and,

when it is the result of submerged earth, it is in addition to the

water pressure. The uncertain element of static friction between

the earth and the dam is usually neglected, as it is on the side of
safety.
When the face of the dam on which the earth or silt pressure
acts not vertical, as in Fig. 5, the weight of the material ver-
is

tically above the plane (the area, 6-9-10-11), should be included

in the forces acting on the dam, in the same manner as described
for water pressure. As a matter of fact, the determination of the
vertical and horizontal components of earth and silt pressures
parallel exactly that described for water pressures, except,that the
portion of Eq. (7) in parentheses modifies the pressure in accord-
ance with the internal resistance to movement, as measured by
a, the angle of repose.
If wz is the weight of earth per cubic foot in air, and k is the
percentage of voids, then in 1 cu. ft. of the fill there will be (1 — A;)
cubic feet of solids, weighing wz pounds. The weight of water
displaced when this cubic foot of submerged will be w>2(l — k)
fill is

pounds. Therefore, the weight, wz, of the submerged fill will be

wz=wz'— 102(1 — k) (8)

For example, if the material weighs 110 lb. per cu. ft. in air and
has 30 per cent of voids, there will be (100-30 per cent) or 70 per
cent of solids in each cubic foot. Therefore, 1 cu. ft. of the fill will
displace 0.7 cu. ft. of water weighing 0.7X62.5 = 43.75 lb., and the
weight per cubic foot of submerged fill will be 110—43.75 = 66.25
lb. =wz-
The weight, voids, and angle of repose vary, of course, with the
nature of the material, and should be made the subject of investi-
gation for each case.
Safe values often used for sand and gravel are 103'= 110,
A; = 30
per cent, and a = 30°. This corresponds to a value for w 3
of 66.2. For material capable of acting as liquid mud the usual
Art. 17] ATMOSPHERIC PRESSURE 35

values are ws = 125, k = 0,

'
and a = 0. This corresponds to a value
for w3 of 62.5.
Thus it is seen that the usual unit weight of submerged earth
isnot far from that of water, but the angle of repose for use in
Eq. (7) varies from zero for liquid mud to about 30° for sand and
gravel.
It is silt pressure on the up-stream face and
not often that
uplift pressure from head-water are assumed to act simultaneously
against the dam, unless the silt is of an unusually pervious nature.
Ordinarily, the silt deposited by streams is an impalpable clay
mixed with very fine sand or similar material, particularly im-
pervious to the passage of water. It is the usual practice, where
silt of an impervious nature is expected, to use in the calculations

for the design that one of these two forces which has the greatest
influence on the shape of the section.
It has been claimed that silt deposits relieve the water pressure
on the up-stream face of the dam, but this is an assumption incon-
sistent with conservative design.
17. Atmospheric Pressure. Except as described later, atmos-
pheric pressure is exerted on every square foot of the surface of

Water Surface

Fig. 8.

the dam. Fig. 8 represents a section of a spillway where the over-

flowing sheet of water is not in contact with the down-stream face
of the dam.
In such cases the friction of the moving water surface, 1-2-3,
entrains the air contained within the boundaries, 1-2-3-4, and
carries it away. If the chamber, 1-2-3—1, has free access to the

atmosphere, air will be easily supplied as fast as it is removed.

It is not always feasible to provide such access, particularly if
36 FORCES ACTING ON BAMS [Chap. Ill

the dam is very long and the head, h, on the crest is relatively
great.
If the chamber, 1-2-3-4, does not have free access to the
atmosphere, a partial vacuum or reduction of atmospheric pres-
sure will occur therein. An
adjustment of conditions, as indicated
in Fig. 9, will then occur. The
difference in. atmospheric pressure
on the two sides of the sheet will cause it to move toward the dam,
the difference in pressure being balanced by the force overcome in
changing the direction of the sheet.
The reduction of atmospheric pressure on the area, 3—4, will
water surface between the sheet and the dam
result in raising the
an amount, hi, sufficient to balance such reduction. The reduc-
tion of atmospheric pressure on the face, l-4rAa, of the dam must
be balanced by the stability of the structure. The resulting force,

Fig. 9.

R, on the dam is equal to the trapezoid, l-5-6-4a, in which the

distances, 1-5 and 4-6, represent the reduction in atmospheric
pressure, hi, over the area, 1-4.
Quite often, in such cases, the reduction of atmospheric pres-
sure becomes intermittent; that is, a partial vacuum within the
chamber, 1-2-3-4, accumulates up to a certain amount, then a
break occurs in the sheet admitting air, with a sudden return to
normal pressure. The operation is repeated periodically. Such
periods sometimes become of very short duration, causing a strong
vibration, with a consequent tendency to loosen the dam from the
foundation and thus decrease its stability considerably beyond
that corresponding to the steady force, R.
As an example of the possible magnitude of such vibrations,
the late J. P. Frizell * stated that the vibrations set up in this
way
* "Water Power," by Joseph P. Frizell, 3d ed. J. Wiley & Sons, 1905.
Aet. 18] ICE PRESSURE 37

by the water flowing over the old crib dam at Holyoke, Mass.,
rattled the windows in Springfield, six miles away.
On account of the impossibility of determining accurately the
amount and effect of the force, R, it is customary to fit the down-
stream face of the dam to the lower nappe of the jet corresponding
to the maximum head, h, which may occur on the crest of the dam.
A method of determining the shape of the jet is given in Art. 42.
18. Ice Pressure. In common with other materials, ice
expands and contracts with changes of temperature. In a reser-
voir completely covered with ice, a contraction due to a decided
reduction in the temperature of the air will take place, opening up
large cracks in the ice in which, subsequently, the water freezes
solid. When the next rise of temperature occurs the ice expands,
and if it is not free to slide up the banks of the reservoir, it will
exert considerable pressure on the dam. This pressure usually
causes the sheet of ice to buckle or crush. Should the conditions
be favorable, the ice may exert an overturning force on the dam.
The thrust of ice is impossible of exact determination. It is,

of course, limited to the crushing strength of ice which, however, is

variously reported between 100 and 1000 lb. per sq. in. The
latter value corresponds to the enormous amount of 144,000 lb.
per sq. ft., and, where ice attains a thickness of 4 ft., amounts to
the absurd value of more than a half a million pounds per linear
dam.
foot of
That ice under usual conditions, can never approach
thrust,
the latter value proved by the fact that a great many dams are
is

standing to-day which otherwise would certainly have failed.

This, in part, may be explained by the fact that the full thrust of
ice cannot be exerted on a dam:

1. If it is in a narrow gorge where the full effect of expansion

cannot be felt;
2. If the opposite banks are sloping so as to allow the ice to
up on them and thus relieve the pressure;
ride
3. If, as has been claimed, the ice next to the dam cannot
attain its full crushing strength owing to the greater
temperature of the masonry.

Mr. C. L. Harrison's paper * on this subject brought out a

* Transactions, Am. Soc. C. E., Vol. LXXV, p. 142.
3$ FORCES ACTING ON DAMS [Chap. Ill

very free discussion. The following is quoted from his closing

remarks in reply thereto:

It is generally agreed that this force [ice pressure] should be considered,

and allowance made in some but not in others. The limitations of
cases,
these cases are not definitely stated by those who have discussed the subject.
In order to give the matter more definite shape, it may be suggested that,
under the following conditions, it is not necessary to provide for ice pressure:
1. For the ordinary storage reservoir with sloping banks, in
climates
—
where the maximum thickness of ice is 6 in. or less for dams with southern
exposure this limit may be placed as high as 1 ft. None of the discussions
fixes this limit, but it is what the writer has in mind as a reasonable provision.
2. For reservoirs which are filled during the flood season and from which

all the stored water is drawn off each year during the low-water season.
This would include even the Jarge reservoirs on the head-waters of the Mis-
sissippi River,where the ice has a thickness of more than 4 ft., and the
atmospheric temperatures reach 50° below zero.
3. For storage reservoirs where the water will be drawn off each year

during the winter to a level where the dam is strong enough to resist the
ice pressure.
4.

level is
*******
For reservoirs where the contour
such that the expansive force of the

The dams cited in Table II,

pressure in the design, are in the vicinity of
of the
ice will
ground at the high-water
not reach the dam.

where a value has been given to the ice

New York or Boston, where the
maximum thickness of ice may be taken as about 2 ft. No allowance is
made for ice pressure in the New Croton Dam, which is in the same climate.

TABLE II

Ice Pressure Adopted for Various Dams

Dam.
Art. 18] ICE PRESSURE 39

are that local conditions will prevent the full ice thrust from reaching the
dams, and also that they are located up stream from the New Croton. If
either of these dams should fail, no valuable property would be damaged,
and the waters would flow into the New Croton Reservoir. Those respon-
sible for the design of the New Croton Dam believed that no allowance
should be made at this dam for ice thrust. At first glance, this looks like
a wide range in judgment, but it must be remembered that the foregoing
statement gives only a part of the facts, and to this must be added the local
conditions and the service the dam is to render in each case before judg-
ment is passed on the wisdom of the design.
Reservoirs for domestic supplies are generally drawn down during the
ice period, and the greatest expansion of the ice occurs at the end of this
period, thus applying the pressure at a point below high-water level, where
the dam is strong enough to resist it. If, however, such reservoirs are to
be at high-water level during, and especially at the end of, the ice period,
at any time during their service, then the proper allowance should be made
for ice pressure in the design. The daily fluctuations in the water level
in the forebay at power dams will usually prevent the ice from freezing to
the dam, which, therefore, will not be subjected to thrust caused by the
expansion of the ice in the pool above the dam. In such cases, the proper
course seems to be, not to reduce the allowance, but to omit it altogether.
If, however, a storage reservoir is to be filled to the high-water level during

the full ice period, at any time during the life of its service, then not a partial,
but the full ice pressure should be allowed for in the design of the dam.
It is contemplated that the Kensico Reservoir is to be kept at or near the
high-water level at all times, and therefore will be subject to full ice pres-
sure at high-water level; also, the Olive Bridge and Wachusett Dams may
at intervals be subject to this pressure at high-water level. It is entirely
possible that it would be proper to allow for ice pressure on a dam in a given
locality and also proper to make no allowance for such pressure on another
dam in the same locality, depending on the service each is to render.
The fact that so many dams have been designed and built without making
a specific or separate allowance for ice thrust, and have for years stood the
test of actual service without failing,, is an indication that ice pressures may
not be as great as sometimes thought, or that the factors of safety allowed
for other purposes are sufficient to take this pressure. On the other hand
in the cases mentioned in this discussion, there seems to be good and suf-
ficient reason for allowing for ice pressure in the designs.

Ice floes are capable of exerting comparatively little pressure

against a dam. If the velocity of the approaching water is high
and the crest is not clear, the most that can be expected is a local
thrust, which need not be very large, as the ice in such cases is

always soft.

Possible ice thrust on the down-stream side of the dam is

usually neglected in the calculations.

40 FORCES ACTING ON DAMS [Chap. Ill

19. Wave Pressure.seldom necessary to consider the

It is
effect of waves on the dam. Stevenson's formula
stability of the
for the determination of the height of waves is:

h= lWF+(2.5-\/F), (9)

where h is the height, in feet, and F is the fetch, or straight length

of clear water, in miles.
Unless the waves break, due to the relative shallowness of the
water near the dam, very little impact results. Except for dams
which are unusually small compared with the size of the reservoir
the usual dimensions obtained from other considerations are ample
to resist the effect of wave action.
20. The Weight of the Dam.
The unit weight of masonry
varies considerably, depending on the ingredients of which it is
composed. Table III, giving weights of masonry, is taken from
the " American Civil Engineer's Pocket Book." *

TABLE III

Weight of Masonry in Pounds per Cubic Foot

Ashlar:
Granite, syenite, gneiss.
Limestone, marble
Sandstone
Mortar Rubble:,
Granite, syenite, gneiss
Limestone, marble
Sandstone

Table IV gives the approximate average weight of plain Port-

land cement concrete. Although the weight of concrete is affected
greatly by the materials of which it is composed, it is influenced
very little by changing the proportions of the ingredients. A wet
mixture will ordinarily result in a somewhat lighter concrete than
if the ingredients are mixed dry and well compacted.

* By Mansfield Merriman, John Wiley & Sons.

Art. 20] THE WEIGHT OF THE DAM 41

TABLE IV

Approximate Weight of Plain Concrete and Coarse Ingredients,

in Pounds per Cubic Foot

Coarse Aggregate.
42 FORCES ACTING ON DAMS [Chap. Ill

One cubic foot of cyclopean concrete requires:

0.8 cu. ft. of plain concrete at 148.4 lb . . . . 118 7

. lb.

0.2 cu. ft. of plums at 175 lb 35.0 lb.

Weight of 1 cu. ft. of cyclopean concrete. . . 153 7

. lb.

This method of computing the weight of concrete is quite ap-

proximate, and adopted values equal to 95 per cent of computed
values are not too conservative. In important work, an exact
analysis of the materials or tests of the actual weight of concrete
blocks, should be made.
21.The Weight of the Foundation. Dams have sometimes
been tied down to the rock foundation in order to increase their
resistance to overturning, sliding, or both. The arrangement
has consisted of steel bars grouted into holes bored in the rock and
extended into the dam near the up-stream face. This practice,
however, has been severely criticised lately, on the ground that
a satisfactory anchorage of the bars in the foundation is seldom
possible.
22. The Reaction of the Foundation. A. Rectangular Bases.
Let 2(TF)* (Fig. 10) be the resultant of all vertical forces
acting on the dam above the
foundation, including uplift,
and 2(P) the resultant of all

horizontal forces. The result-

ant, R, of 2(1*0 and 2(P) will
represent the resultant of all
forces.
For the dam to be in static
equilibrium,the resultant, R,
must be balanced by an equal
and opposite reaction, it!, of the
foundation, consisting of the
total vertical reaction, 2 (IF),
Fig. 10.
and the total horizontal shear
or friction, 2(P).
The masonry of a dam and all classes of foundations, in com-
mon with other materials, are known to be elastic. The effect of
*2(TF) and S(F) are here used to represent a general condition, and
may be applied to either full or empty reservoir.
Art. 22] THE REACTION OF THE FOUNDATION 43

such elastic properties makes the absolute determination of the

distribution of compressive stresses very difficult, if not impossible.
Because of the lack of knowledge on the subject, the distri-
bution is assumed to follow a law of uniform variation, as in the

design of beams, although it is known that, on account of such

elastic properties, the unit pressures at the extremities of the base
will be somewhat less than this method of distribution would
indicate.
In Fig. 10, p/ represents the unit vertical reaction* at point 1,
and pT " that at point 2, due to the load, ~2(W). An ordinate at
any point from the base to the straight line, 3—4, will represent
the unit vertical reaction at that point. The area, 1-2-3—4, will
then represent the total vertical reaction, S(W).
To derive an equation for the value of p/, take moments about
point 2. Then
•n "I 2 I
2

whence,
"
,_ 3Z(TT) 32(W)u pT
Vt ~ I P 2
But,

or,

Pr
22W ,
Pr.
l

Substituting this value of p r" in the second equation, there

results

~
22W
Vt
I
(*-?) .• • <"*>

By taking moments about point 1, there may be derived,

in like manner,

"v-^fr-O <">

Dividing the base, 1-2, into three equal parts, as in Fig. 11,
and calling the middle part the " middle third," it is seen from<Eq.
* The reader should bear in mind the distinction between p r the unit ,

vertical reaction due to load 2(PP); and p v the unit vertical compressive stress.
,

It will be shown later that pt=pr+p«i as in Fig. 10.

 .

44 FORCES ACTING ON DAMS [Chap. Ill

(10) and (11) that when the resultant, R, intersects the base at the
exact down-stream extremity of the middle third,

w = ~, p r =0, and pr ' = —— -,

(12)

When the resultant, R, intersects the base at the exact up-

stream extremity of the middle third,

21
u = -g-, P>
, n
=0, and
,

pr =
„ —
2Z(TF)
j—*-. (13)

When the resultant, R, intersects the base at the middle

point,

u=- and pr " = pr'= , .

When the resultant falls outside the middle third, tension

will exist at the opposite end of the base. This is indicated

by substituting a value of u less than = in Eq. (11), whereupon

pr" becomes negative. Also, by substituting a value of u greater

than — in Eq.
21
(10), p/ becomes
o
negative. The former case is

indicated in Fig. 11.

The unreliability of tension in
masonry has led to the common
practice of neglecting it as a
possible factor in the stability
of dams. Moreover, for reasons
which will be mentioned later,
the resultant for well-designed
dams always made to inter-
is

Fig. 11. sect the base within the middle

third.
The values, p/ and p T ", derived from Eq. (10) to (13), inclu-
sive, represent the unit vertical reaction of the
foundations on the
dam, corresponding to the load, S(Tf). As S(TT) includes the
force of uplift and as the water pressure causing uplift is exerted
both upward and downward, the corresponding unit vertical pres-
sures, p„' and p v ", on both the dam and the foundation, can be
 Art. 22] THE REACTION OF THE FOUNDATION 45

obtained by adding to the calculated values of p/ and pT " what-

ever effective * unit uplift has been assumed to exist, as indicated
in Fig. 10.
To sum up:
For the resultant within the middle third:

p ,> = *Jfl(2-f) +P; (10a)

r
22(TT )/3w
p. -
t
(^-l)+pu " (11a)

For the resultant at the down-stream extremity of the middle

third:

Pw
> = Wp.+ Pu '
i
(i2a)

p," = 0+ Pu " (12&)

For the resultant at the up-stream extremity of the middle

third:

Pv
j
tPu , (Ida)

p,' = 0+pj (136)

The foregoing equations are apphcable also to the vertical

reactions and pressures on any horizontal plane above the foun-
dation. They are simply modifications of the theory of flexure
and direct stress as applied to horizontal rectangular sections.
B. Irregular Bases. If the base is not rectangular, as in the
case of some hollow dams, the foregoing equations should be
modified.
As indicated in Fig. 12, the total pressure at any point in the
base will be the sum of the flexural stress at that point due to the
eccentricity of the loading, a uniform direct compressive stress
corresponding to the loading, and the uplift pressure.
From the principles of mechanics, the flexural stress at any
point in the base is,

Flexural stress =± j
,

* p„ is termed "effective unit uplift" as it is the absolute uplift multiplied

by the percentage, c, of the area over which it acts, as explained in Art. 15.
 —

46 FORCES ACTING ON DAMS [Chap, m

where ra = any distance to the right or left of the center of gravity,
in feet; positive on the side of the resultant, 2(TT),
and negative on the other side;
e= the eccentricity of the loading, in feet;
7= the moment of inertia of the base about an axis through
its center of gravity perpendicular to the dam sec-

tion, in feet units.

The stress will be positive (compression) on the side of the resultant

Fig. 12.

and negative (tension) on the other side, as indicated by the line,

4-5-3, in Fig. 12.

The direct stress is equal to \

A ,
'

where A is the area of the base, in square feet. This stress is rep-
resented by ordinates between the lines, 4-5-3 and 6-7. Ordi-
nates from the base, 1-2, to the line, 6-7, will represent the total
flexure and direct stress, or the reaction of the foundation.
The unit effective uplift, if existing, is laid off below the line

6-7; the line, 8-9, finally represents the total unit pressure in the
dam and the foundation.
If wl and m" represent distances to the extremities of the base,
as indicated, then

+ J
1

J
— r-Vu, .... (14)
Art. 22] THE REACTION OF THE FOUNDATION 47

and

p
,,
= _2^+ ^) +pu
„
(15)

C. General. It is well to note here that the vertical unit pres-

sures, p/ and p„", are not the maximum compressive stresses in
the foundation or the dam, but only the vertical components of *
such stresses. The determination of the maximum stresses will
be taken up in Art. 33.
The distribution of the total horizontal shear, 2(P) (Fig. 10),
isprobably not possible of exact determination. Many attempts
have been made to determine such distribution, with varied
results.* Fortunately, it is not necessary to utilize the unit
shearing stresses in order to determine the safety of the structure
or to design the section. The amount of the total shear, S(P),
is sufficient.

*See Chapter VIII, "Masonry Dam Design," by Morrison and Brodie,

2d ed. John Wiley & Sons, Inc. 1916.
 CHAPTER IV
REQUIREMENTS FOR STABILITY OF GRAVITY DAMS

23. Causes of Failure. There are two direct ways in which

a dam will fail:
1. By on a horizontal joint above the foundation,
sliding (a)
(b) on the foundation, or (c),on a horizontal or nearly
horizontal seam in the foundation.
2. By overturning (a) about the down-stream edge of a hori-
zontal joint, (b) the base, or (c) a plane below the base.

The direct cause of sliding is the presence of horizontal forces

greater than the combined shearing resistance of the joint or base
and the static friction induced by the vertical forces.
The direct cause of overturning is the presence of horizontal
forces, greatenough, in comparison with the vertical forces, to
cause the resultant, R (Fig. 10), of all forces acting on the dam
above any horizontal plane, to intersect that plane outside of the
limits of the dam.*
A dam may start to overturn, but finally fail by sliding. This
iscaused by the fact that, when overturning starts, not only is
the shearing resistance at the plane of rupture materially reduced,
but the admission of head-water pressure to the fissure at once
reduces the effective weight of the dam, so that the frictional
resistance due to the pressure of vertical forces is also diminished,
and sliding ensues.
A dam with the resultant well inside the joint may overturn if

the toe of the dam fails by crushing or other causes so as to reduce

the effective length of the joint or base sufficiently to cause the
resultant, R, to pass outside.
Theoretically, the dam extends indefinitely below the base,
but, the area of the vertical section of the foundation being unlim-
below the base is usually assured. Failures have
ited, stability
been caused, however, by an erosion of the foundation caused by
* Tension in the dam at the plane disregarded.
48
Art. 24] LOCATION OF THE RESULTANT 49

water spilling over the crest. In this event, one of two things
may happen: the dam may become undermined so as to cause
overturning, or the erosion may expose a horizontal seam filled
with clay or other practically frictionless material, in which event,
the rock ledge down-stream from the dam having been removed,
it offers no resistance to sliding and the dam fails in that manner.
The gradual disintegration of the dam by weathering and other
causes will, of course, finally result in its failure. Modern masonry
dams, particularly if of concrete, are practically indestructible
if well built and composed of proper ingredients. However, if

proper precautions are not observed in the choice of materials

and methods of construction, there may be rapid deterioration of
the structure.
24. Rule 1. Governing the Location of the Resultant It
is obvious that, if the resultant of all forces acting on a dam above
any horizontal joint, including uplift, passes outside that joint, the
dam will overturn unless the joint is capable of resisting tensile
stresses. As the tensile strength is always disregarded as inde-
terminate and unreliable, we have as the first consideration
that the resultant must intersect the joint. Further restrictions,
however, are necessary.
It was shown in Art. 22, that, when the resultant falls outside
the middle third, tensile stresses are set up at the other end of the
joint. If the joint is incapable of resisting these tensile stresses,
the elasticity of the masonry will cause a slight opening of the joint.
Such an opening is particularly objectionable at the up-stream
side when the pond is full, as it will admit full head-water pressure
over the entire area not in compression, a condition considerably
more severe than usually adopted for uplift. This additional
uplift would result in a movement of the resultant toward the toe
of the joint, with a further opening, of the joint in tension, and a
further increase in uplift. The progression may be sufficient to
cause failure.
It customary in most government requirements, and to a
is

lesser degree in practice, to prohibit tension at the down-stream

edge of the joints when the pond is empty. The logic of this
requirement is open to debate, as one cannot imagine any dam of
the usual type overturning up-stream before the water is let into
the pond. Parker * states that if tension, reservoir empty, occurs
* "The Control of Water." D. Van Nostrand Co. 1913.
50 STABILITY OF GRAVITY DAMS [Chap. IV

over 0.04 to 0.10 of the joint, it does not necessarily mean a bad
design.
However, out of deference to precedent, and the usual govern-
ment requirements, the author has adopted this condition.
We can now write the first designing rule.
RULE 1. GOVERNING THE LOCATION OF THE RESULTANT
Tension shall not exist in any joint of the dam, under any
condition of loading. For dams with rectangular joints, the re-
sultant of all forces acting on the dam above any horizontal
joint (including uplift) shall, for full or empty reservoir, intersect
the joint within the middle third.*

25. Rule 2. Governing the Inclination of the Resultant.

The resultant, 2(P), (Fig. 10), of all the horizontal forces acting
on the dam above any horizontal joint has a tendency to slide
that part of the dam over the lower part. The shearing and
frictional resistance of the joint must be sufficient to withstand
this tendency.
The planes weakness are the necessary horizontal joints
of
(including the base), between two days' work. The shearing
value of such joints, though sometimes considerable, f is unreliable.
It is customary, therefore, to neglect any possible assistance from
shear, and rely solely on the frictional resistance due to the ver-
tical forces, 2(T7), acting on the dam above the joint.
If /' represents the coefficient of static friction of the materials
above and below the joint, then/'2(W) will be the frictional resist-
ance to sliding.
For equilibrium, f'2(W) must be equal to or greater than
2(P). This may be expressed algebraically by:

/'2(F) = >2(P),
or,

|£}=tan *=</'.
* For joints which are not rectangular, as in a hollow dam, the resultant,
in consideration of other requirements, usually falls so close to the center
of gravity of the base as to obviate the possibility of tension. Where doubt
the joint should be tested for tension, as explained in Art. 22.
exists,

t W. A. Mitchell, in
" Professional Memoirs," Corps of Engrs., TJ. S. A.,
Jan. and Feb., 1915, reports shearing strength on planes between new and old
concrete amounting to nearly 50 lb. per square inch, as determined by tests.
Art. 25] INCLINATION OF THE RESULTANT 51

where 8 is the angle of inclination, with the vertical, of the resultant

R (Fig. 10).

RULE 2. GOVERNING THE INCLINATION OF THE RESULTANT

The tangent of 0, the angle of inclination with the vertical,
of the resultant of all forces (including uplift), acting on the dam
above any horizontal plane, shall be less than the coefficient of
friction at that plane.

The coefficient, /', in carefully constructed dams on rock foun-

dations, with particular attention paid to obtaining very rough
surfaces at the base and between two days' work, is usually con-
sidered to be at least twice as great as indicated by experiments
on well-dressed specimens of the same materials. Therefore, if

tan is made equal to or less than the coefficient of friction, as

indicated by such
tests, a factor of safety in this respect of at
least two be provided; and the neglect of the probable con-
will
siderable adhesion or shearing resistance at the joints and founda-
tion will serve to increase further the factor of safety. The last
equation can then be written:

For horizontal joints and rock foundations:

S(P)
= tan0=</, (16)
2(F)
where / is the coefficient of friction of the materials on each side
of the joint or at the base, as indicated by tests on well-dressed
specimens of the same material.
Values of/ for masonry on masonry and masonry on good rock
foundations have been assumed variously between 0.6 and 0.75.
In general, and for careful work, a value of 0.75 is not excessive.
Proper allowance, however, should always be made where the
rock foundation is poor, or where it contains nearly horizontal
seams close to the finished surface of the foundation. Such seams
are particularly. dangerous if they contain clay or other unstable
material. The allowance to be made will depend on the charac-
ter of the seam and its contents, its inclination, and the ability
of the rock above the upper seam to resist movement.
On earth foundations, a large factor of safety should be pro-
vided, as the dam is as likely to slide on planes below the surface
52 STABILITY OF GRAVITY DAMS [Chap. IV

as at the junction between the earth and the masonry. An arti-

ficial bond and increase in the coefficient of friction obviously

cannot be obtained, as in the case of rock foundations. Eq. (16),
therefore, should be rewritten for earth foundations as follows:

For earth foundations:

!$-«-•-£ (16o)

where S is the factor of safety desired.

For masonry dams on gravel, sand, and wet clay, approximate
values of / are 0.50, 0.40, and 0.30, respectively. In conservative
designs the dam is usually anchored to deep cut-off walls or piles,
or a factor of safety, S, of 3 or more is adopted. The weight of the
apron, which is an adjunct of every spillway dam on earth, will
assist materially in reducing the tendency to slide.

26. Rule Governing Compressive Stresses. Equations

3.

for the derivation of the vertical compressive stresses on the base

and horizontal joints have been derived in Art. 22. It will be
shown later that the maximum vertical compressive stresses are
not the maximum stresses which occur in the structure. The
maximum stresses occur on inclined planes, the vertical stresses
being only their vertical components.
Many theories have been advanced for the determination of
the maximum inclined stresses in dams; but, so far, no general
agreement has been reached. The problem is not a determinate
one, and, whatever theory is used, the solution is necessarily ap-
proximate.
For these reasons, it has been common practice to design dams
with reference to maximum allowed vertical stresses, such stresses
being chosen low enough, in the designers' judgment, to allow for
the difference between them and the larger inclined stresses.
This method has resulted in the present type of solid high
masonry dams having the down-stream face concave outward,
and maximum inclined stresses which, in the lower part, are not
constant at all elevations. To design a section in conformity
with allowed maximum most
inclined stresses, in accordance with
of the theories advanced, would result in a section
having a partly
convex down-stream face, maximum vertical stresses which are
not constant at all elevations, and different also in other minor
respects.
Abt. 26] GOVERNING COMPRESSIVE STRESSES 53

Until the matter has received much needed attention by our

engineering societies, and a general agreement is reached, it seems
best to adhere to the present practice for shaping the section.
However, a rational method of determining the maximum inclined
stresses, no matter how approximate the assumptions on which
it is based, is much more accurate than a pure guess. It is rec-
ommended, therefore, that, after the dam has been designed to
conform to the allowed vertical stresses, the inclined stresses
should be investigated in accordance with whatever theory is
thought to be the best,* and changes made in the section if the
inclined stresses are found to be too high. This course is all the
more justifiable because the present method affects only the
economy of the structure and not its safety, provided the inclined
stresses are within allowed values, f

RULE 3. GOVERNING COMPRESSIVE STRESSES

The unit vertical and inclined compressive, stresses in the
dam and the foundation shall not exceed certain prescribed
values.

The strength of masonry, as indicated by well-constructed test

specimens, not a satisfactory measure of its strength in large
is

masses, owing to the less favorable conditions under which the

latter is usually built. For this reason, as well as on account of
the uncertainty in the determination of the maximum stresses, it is
necessary to adopt working stresses for masonry dams correspond-
ing to an unusually large factor of safety. It is, indeed, advisable
to limit the stresses to values closely approximating that which
precedent has shown to be safe.

*A combination of the theories of Enger and Bouvier, as described in

Art. 33, is thought to give results more nearly correct than that of any of
the others.
f This method is quite unsatisfactory, in that the stresses in the dam
are limited; by
first, by usual practice according
vertical stresses fixed
to the present method of design, and second, by inclined stresses which, it
is thought, should receive equal attention, although seldom considered by

designers. However, though the author does not deem it advisable to pro-
pose a rational change in the present method of design, without the con-
sideration and discussion of the Engineering Profession, he cannot con-
scientiously countenance the acceptance of any design without due regard
for the more important actual maximum or inclined stresses in the structure.
54 STABILITY OF GRAVITY DAMS [Chap. IV

Table V indicates the maximum vertical and inclined stresses

in several large existingdams, computed, in accordance with the
theory indicated in Art. 33, from scaled values of the angle, </>,

and from published records of computed maximum vertical stresses.

These may be considered to be the greatest stresses which have
thus far been adopted for conservative designs. The table also
indicates the maximum stresses corresponding to examples of
design of two authorities, and values recommended by others.

TABLE V

Various Recommended Working Stresses and Existing Stresses

in Masonry Dams

Dam.
Art. 26] GOVERNING COMPRESSIVE STRESSES 55

TABLE VI
Relative Ultimate Strength of Portland Cement Concrete,
in Pounds per Square Foot

Proportion.
56 STABILITY OF GRAVITY DAMS [Chap. IV

values which are within the limits of those commonly used for
masonry dams, and which, moreover, will result in a section in
which the maximum inclined stresses are not excessive. The fol-
lowing values will ordinarily fulfill both these conditions:

At the toe of solid dams ^t of the ultimate strength.

At the heel of solid dams ^
of the ultimate strength.
At the toe and heel of hollow dams x§- 0I the ultimate strength.

For the foundations, it is universally agreed that bed-rock

suitable for a high dam will in every case be stronger than the
masonry placed upon it. For dams on earth foundations, the
requirements of Rule 2 necessitate a very small angle of the result-
ant with the vertical. It will be shown later that the difference
between the maximum inclined and the maximum vertical stresses
in the foundation is a function of this angle and, therefore, so small
as to be negligible. Common values for allowed stresses on earth
foundations are as follows:

Clay 8,0001b. per sq. in.

Coarse sand. 4,000 to 8,000 lb. per sq. in.
Fine silt 2,000 to 4,000 lb. per sq. in.

It must be borne in mind that the qualities of such materials

vary greatly, and that adopted values, in all cases, must be based
on complete investigation at the site.
27. Rule 4. Governing Tension in Vertical Planes. Mr.
L. W. Atcherly, * of London University, claims to have proved that,
although the resultant, in dams of the usual type, may fall well
inside the middle third of the base, there may still be considerable
tension in vertical planes near the toe of the dam. Messrs. Mor-
rison and Brodie | give an outline of Atcherly's theory, as well as
the opinions of a number who have entered into a discussion of
the paper. The paper attracted a great deal of attention at the
time, and even influenced the design of several important struc-
tures. Later studies, however, have led to the general conclusion
that such danger is not ordinarily to be feared. Atcherly's theory
* In a paper "On Some Disregarded Points in the Stability of Dams " '
See foot-note, p. 71.
" Masonry Dam Design."
s
t
Ast. 28] MARGIN OP SAFETY 57

isbased on an assumed distribution of horizontal shearing stresses,

which remains to be demonstrated,. A great number of existing
dams, if investigated according to Atcherly's theory, would be
found to be quite unsafe, if not incapable of sustaining their loads.
Considering the hundreds of dams which have been conserva-
tively designed and built without a
thought of tension in vertical planes,
not one of which is known to have
failed through this source of weak-
ness, it must be assumed that there
is some error in Atcherly's methods,

and that they impose too severe a

condition on structures of the usual
type.
However, one must not lose sight
of the fact that tension will certainly
FiQ 12o
exist in vertical planes if the inch-
nation of the down-stream face of the dam is too great to transmit
the loads to the foundation, as in Fig. 12a. In such cases there
is danger of the toe cracking off on the line, 1-2, even though

resistance to shear on that plane is ample.

The influence of these contingencies on the design of the dam
is taken up in Art. 34.

RULE 4. GOVERNING TENSION IN VERTICAL PLANES

The inclination, with the vertical, of the down-stream faces
of the dam shall be limited to prevent or safely withstand all
possible tensile stresses in vertical planes.

28. Rule 5. Governing the Margin of Safety. The usual

meaning of the expression " factor of safety," in the design of
structures, is the ratio of the loading which will be just sufficient
to cause damage or failure to the loading which has been adopted
in the design. Factors of safety may be applied to the balance
or resistance of a structure to overturning or to unit stresses. In a
cantilever bridge it is usual to design the shore span so that it will

have a resisting moment about the pier equal to twice the moment
of maximum combined dead and live loads on the river end, thus
providing a factor of safety of two against overturning. A steel
tie-rod designed to be stressed to one-half its elastic limit will have
58 STABILITY OP GRAVITY DAMS [Chap. IV

a factor of safety of two against damage (permanent elongation),

and a factor of about four against failure.
In dealing with unit stresses, the term " factor of safety " is

misleading, as it is usually applied to the ultimate strength, whereas

in most cases, it is the elastic limit which is the limiting stress.
With regard to the balance of a structure, or the location of
the resultant of forces, a factor of safety of one is, theoretically, all
that is necessary, provided the loading is certain. However, on'
account of the usual uncertainty as to the loading, a factor of
safety greater than one is necessary. In a dam, the head- and tail-
water pressures for any given depth may be accurately calculated.
A careful assumption of the weight of the dam should not vary
more than 4 per cent from its correct value. Then we must adopt
pressures for ice, silt, and uplift, which we are certain will not be
exceeded, in order to justify a location of the resultant at the
extremity of the middle third. If this is done the dam, if cor-
rectly designed, may
be considered perfectly safe, provided, also,
that the calculated induced unit compressive stresses amount to a
safe proportion of the ultimate strength of the materials. The
dam have an additional element of safety on rock foundations,
will
because the adhesion of the concrete to the foundation and in
horizontal planes above the foundation, though certainly consid-
erable, is always neglected. In fact, to this feature, alone, can
be attributed the existence of a number of poorly designed dams.

RULE 5. GOVERNING THE MARGIN OF SAFETY

All assumptions of forces acting on the dam shall be unques-
tionably on the safe side; and all unit stresses adopted in the
design shall provide an ample margin against rupture.

The margin of safety varies considerably in recent dams, and

depends to a great extent on the use to which the structure is to
be put, its magnitude, and the probable damage and loss of life
which would result from failure. One of the most conservative
designs of modern dams is that of the Olive Bridge Dam, of the
New York City Water Supply System (Fig. 23). Probably few
private enterprises could stand the burden of such conservatism.
29. Rule 6. Governing Details of Design and Methods of
Construction. The shape of the section of the dam having been
determined in accordance with the foregoing rules, careful atten-
Art. 29] METHODS OF CONSTRUCTION 59

tion must be paid to the details of the design and the methods of
construction, in order that the structure may be satisfactory in
every respect.
The location and extent of vertical building joints, passage-
ways, and other planes of weakness must be well within proper
limits, in order that the stresses used in the design will not be seri-
ously increased. Such features as drains and cut-offs, on which
'theassumption of uplift has been based, must be carefully worked
up, and other matters of much importance attended to. The
engineer should be in a position to insure that the masonry in the
structure will be of a quality to withstand safely the working
stresses adopted in the design, practically water-tight, and dur-
able. The final rule may be written as follows:

RULE 6. GOVERNING DETAILS OF DESIGN AND METHODS OF

CONSTRUCTION

All details shall be adopted so as not to interfere with the

assumptions used in the design; the masonry shall be of a quality
suited to the working stresses adopted, and shall be practically
water-tight and durable.
 CHAPTER V
GENERAL EQUATIONS FOR DESIGN OF GRAVITY DAMS

30. General Considerations. It is impossible to derive a

general equation for the direct determination of the shape of the
section. The only possible solution is to design the dam, joint
by joint, beginning at the top, making -each joint conform to the
designing rules given heretofore. Assuming the dam to have been
designed from the top to a certain horizontal joint, 3-4, Fig. 13,

^ *--Point of reference,
or orgin of momenta

Fig. 13.

of length, lo, equations can be written for the determination of the

location, y, and
length, I, of the next joint a distance, h, below, to
conform to that one of the foregoing rules which happens to govern
at the particular stage of the design.
The length of the joint will be mathematically exact, but the
width of the section between the joints will be approximate only.
By taking the vertical distance, h, small enough, say 15 per cent of
the height of the dam above the joint considered, the resulting
error will be negligible. The section will have polygonal faces,
which, of course, may be smoothed up later, for appearance, with
no appreciable change in the stability of the section.
Each step in the design is based on one of the rules which is

60
Art. 30] GENERAL CONSIDERATIONS 61

thought to govern. After each joint is calculated, and conforms

to that rule, it is necessary to investigate the joint to determine if

the other rules are also complied with. It is essential, therefore,

for each designing rule, to derive two sets of equations, as follows:

1. " Equations of Determination " for each rule, with which

to determine the length and location of successive joints
in conformity with that rule;
2. " Equations of Investigation " for each rule, with which
to investigate any pre-designed portion of the dam for
conformity with that rule.

Such equations will be derived for each of the designing rules,

it being found convenient to derive the latter set first in each case.

Fia. 14.

Before proceeding further, however, it seems desirable to indi-

cate the influence of each designing rule on the general shape of
the section of the dam. In order to do this, reference is made to
Figs. 14 and 15, which indicate typical sections of a solid non-
overflow and spillway dam, respectively.
 DESIGN OF GRAVITY DAMS [Chap. V
62

The section of the dam may be divided into a number of zones,

as indicated, it being necessary to design

each zone in accordance
with a different rule or combination of rules.*
Water Snrface

Non-overflow Dams
When ice pressure occurs, the quantity of masonry in Zone I,
above the water surface of non-overflow dams, is fixed by Rule 2,
as sufficient weight must be provided to prevent the portion,
l-2-3^i, from sliding.
In Zone II the resultants, reservoir full and empty, lie well
within the middle third, due to the fact that the width of the top
is always greater than necessary to conform to Rule 1. Both up-
and down-stream faces, therefore, will remain vertical f until,
* As far as the author is aware, Wegmann was the first to use similar
divisions of the section of the dam to explain the methods of design.

f It is generally recognized that there is no economy in battering the

faces of the dam unless that is necessary in order to conform to one of the
designing rules.
Art. 30] GENERAL CONSIDERATIONS 63

at the plane, 5-6, the resultant, reservoir full, intersects the joint
at the exact extremity of the middle third.
At the top of Zone III, the down-stream face must begin to
batter in order to accord with Rule 1, reservoir full; and the
resultant, reservoir empty, still being within the middle third,
the up-stream face remains vertical until at the plane, 7-8, the
resultant, reservoir empty, also intersects at the extremity of the
middle third.
Therefore, at the plane, 7-8, the up-stream face must begin to
batter, in order to accord with Rule 1, reservoir empty. In Zone
IV the resultants, reservoir full and empty, intersect the joint at
the extremities of the middle third.
The upper Zone V is fixed by the condition of limiting
limit of
vertical pressures, Rule 3. Usually, the maximum allowed unit
pressure is reached at the down-stream face first. Below the
plane, 9-10, the length of the joints must be determined by Rule 3,
for full reservoir, and by Rule 1, for empty reservoir. This will
result, for Zone V, in the resultant, reservoir full, intersecting well
within the middle third, and, for reservoir empty, the resultant
continues to intersect at the extremity of the middle third.
The vertical pressures at the up-stream face, however, will
gradually increase, and, at the upper extremity of Zone VI will
just reach the allowed working value. In Zone VI, therefore, the
length of the joints will be determined entirely by Rule 3, the
resultants, reservoir full and empty, intersecting well within the
middle third.
As the section increases in height, the batters of both up-stream
and down-stream faces increase. The down-stream face, in solid
gravity dams, always has a flatter slope than the up-stream face.
Consequently, at some elevation, such as 13-14, the inclination of
the down-stream face from the vertical may reach the maximum
allowed value. Zone VII, therefore, represents a portion of the
dam in which the inclination of the down-stream face is apt to be
greater than the limit fixed by Rule 4. It is unfortunate for the
peace of mind of the designer if this happens, for, in that event,
as will be indicated later, be necessary for him to start the
it will

design all over again, and provide a different arrangement in the

upper part which will result in a steeper down-stream face.
 "

64 DESIGN OF GRAVITY DAMS [Chap.

Spillway Dams

Owing to a deficiency in the weight of the top of all spillway

dams, the resultant, reservoir full, within Zone I, will necessarily
fall outside the limits of the middle third, in direct violation of
Rule 1. This condition must be met by special construction, as
explained later. At the extreme top of the dam, where the weight
of the masonry is negligible in comparison with the water pressure,
the resultant will intersect at a distance infinitely remote, as indi-
cated by the trend of the line of resultants, reservoir full, shown in
Fig. 15.
For the same reason, above the bottom of Zone la, the inclina-
tion of the resultant, reservoir full, will make an angle with the
vertical greater than the allowed value, in direct violation of Rule 2.
At the extreme top of the dam, where the weight is negligible in
comparison with the water pressure, there will be no frictional
resistance and sliding of the upper part of the dam must be pre-
vented by shearing resistance.
It is obvious that, above the bottom of Zone la, these two rules
must be violated, as it is not feasible to supply enough masonry
without obstructing the water spilling over the crest. Fortu-
nately, however, the conditions are not severe and can be easily
met by special treatment.
The conditions fixing the limits of Zones II to VII, inclusive,
for spillway dams are exactly as previously described for non-
overflow sections.

It should be borne in mind that the arrangement of zones indi-

cated in Figs. 14 and 15 represent the usual conditions met in the
design of solid dams. In hollow dams the arrangement will be
different, but the general theory will apply. The equations for
design given in the following articles will be derived with respect
to solid dams; but they are equally applicable, in general, to hol-
low dams.
31. Equations for Rule 1. Equations of Investigation. To
derive a general equation for the location of the intersection of
the resultant, R (Fig. 13), with the joint, 1-2:
Let S(TF) represents the algebraic summation of the vertical
components of all forces acting on the dam above
the joint, 1-2, including uplift; 2(JF), also rep-
Ast, 31] EQUATIONS FOR RULE 1 65

resents the equal and opposite vertical reaction of

the foundation;
2(Wa;) represent the algebraic summation of the moments
about any convenient joint, 17,* of the forces
above the joint, contained in the summation,
r
2(W ); positive when counter-clockwise;
2(P) represent the algebraic summation of the horizontal
components of all forces acting on the dam above
the joint, 1-2; 2(P) also represents the equal
and opposite horizontal reaction of the founda-
tion;
2 (Pre) represent the algebraic summation of the moments
about the point, 17, of the forces above the joint,
contained in the summation 2(P); positive
when counter-clockwise;
Subscript e represents the condition of empty reservoir;
Subscript f represents the condition of full reservoir;

Other symbols as indicated in Fig. 13.

The moment of all forces, acting on the dam, above the joint,
about the point 17, is 2(TFa;) + 2(Pa;). The moment of the reac-
tions about the same point is 2(TF)z, the moment of the force
2(P) being zero. For equilibrium, these moments must be equal;
therefore, for the general case of either full or empty reservoir, we
have:
2(Wx)+2(Px) = I,(W)z (17)

Solving for z, there results:

Z
_ 2(PTx) + 2(Px) (18)
2(TF)

21
According to Rule 1, the distance, z-y, must be less than — for
o

full reservoir and greater than - for empty reservoir.

The usual forces acting under the condition of " full reser-
voir " are indicated in Fig. 16. Equations for the determination

* Point 17 should be on any convenient line of reference, and at the

elevation of the joint considered.
66 DESIGN OF GRAVITY DAMS [Chap. V

of the amounts and locations of these forces have been given in

Chapter III.
For " empty " reservoir, the force due to the weight of masonry-
is usually considered to be the only one acting on the dam. This is
consistent with the loading after completion, but before the coffer-
dam enclosure is allowed to fill, and before earth or silt is deposited
against the dam. It usually gives the severest condition, as far
as Rules 1 and 3 are concerned.

Hoto:-
Reaclions on the joint,
1-2, not indicated.

Fig. 16.

Equations of Determination. First. The general case, where

the location of the resultant for both full and empty reservoir
governs the design of the section, as in Zone IV (Fig. 14).
In Fig. 13, suppose the dam to have been designed from the top
down to the joint, 3-4, and that it is desired to determine the
length, I, of the next joint, 1-2, and its location,
y, relative to the
convenient point of reference, 17.
Art. 31] EQUATIONS FOR RULE 1 67

For full reservoir we may substitute (y+%1) for z in Eq. (18).

Adding the subscript, F there results: ,

v+V ^'tmf*' - <19)

For empty reservoir, by substituting y+%1 for z in Eq. (18)

and using the subscript, E there results: ,

_ S(ITs) g +S(Ps) E
y+ 3~
I

Thus we have two equations with two unknowns, y and I.

^ m
,__*

The factors in the second terms of these equations, however, are

functions of y and I. By substituting in the second terms the
values of the various forces and moments, in terms of y and I,
exact equations may be derived from which the values of y and I
can be solved directly. Unless, however, the vertical components
of water and silt pressure on both faces of the dam are neglected,
such equations will be too intricate for practical application.*
A sufficiently accurate tentative method of solving for y and I may
be used, as follows:
From the general trend of the slopes of the up- and down-
stream faces, tentative approximate values of y and I may be used
for the determination of the factors in the second terms of the
equations. If the height, h (Fig. 13), is not more than 15 per cent
of the height of the dam above the joint considered, the resulting
calculated values of y and I will generally he within 0.5 per cent
of the truth, so that usually, a second trial substitution will be
unnecessary.

* Messrs. Morrison and Brodie, in their "Masonry Dam Design," by neg-

lecting the items mentioned, have derived such equations applicable to
Rules 1 and 3 and combinations thereof. The author has found, however,
by actual comparison, that his tentative method, hereinafter described, is
sufficiently accurate, involves no more labor, and has the advantage of not
only including all forces acting on the dam, but requires a tabulation of forces,
moments, and other features in such a manner as to allow comparatively
little opportunity for omissions or misapplications.
Parker, in his "Control of Water," has also derived similar equations
but states: "I am not aware that these equations have ever been applied
in practice, and as the result of actual experiment, I am inclined to believe

that a return to first principles and trial "and error is probably more rapid."
68 DESIGN OF GRAVITY DAMS [Chap. V

Second. The where the location of the resultant,

special case
reservoir full, is the only governing condition, as in Zone III.
For
this case the up-stream face is vertical and, therefore y, in Eq. (19),
is known. For this reason, and also because the resultant, res-
ervoir empty, is well within the middle third, Eq. (20), is not
necessary, and the lengths, I, of successive joints in this zone may
be solved from Eq. (19) alone.
Third. The special case where it is required to determine the
elevation of the joint where the resultant, reservoir full, first inter-

sects the down-stream extremity of the middle third, as at the

bottom of Zone II. For most types of dams it will be found con-
venient to determine this elevation by trial, using Eq. (18). For

Water Surface

Fig. 17.

the non-overflow dam, however, it is possible to derive an exact

equation applicable to the usual conditions of loading, namely,

the external forces acting on the dam consisting of head-water
pressure, ice pressure, and uplift.
In Fig. 17 1-2-3-4 represent the top of such a dam. Let it
let
be required to determine the distance, h, below the water surface,
at which the resultant, R, reservoir full, lies just at the extremity
of the middle third.
The forces assumed to act on
this portion of the dam are indi-
cated in Fig. 17, and may
be expressed as shown below. The
origin of moments in this case is taken at the down-stream extrem-
ity of the middle third. Remember that counter-clockwise
moments are considered positive.
Art. 31] EQUATIONS FOR RULE 1

Item. Force. Moment.

Vertical forces:

Weight of masonry, -\-wiLH

wiLW
6
CW2I1L cw2hL2
Uplift,* ~~2~ "
l

v ™ TV cw 2 hL
2 '
v
.

"
.

J
it>iL

6
2
H t

'
cw 2 hL2
6

Horizontal forces:
Ice pressure, +P { +PJi
Water pressure, -\— %r- -\ —|—
2(P)=P t +^ S(Px)=P^+^

Eq. (18) may now be written:

w\L2H ,
cw2hL2 . _ . ,

—W2h 3
'-P t h-\ -T-
-E(Wx) + 2(Px)
= _ =— -
6 ^ 6
2
2(W) TTT cw 2 hL
wiLH y~
Substituting h+a for H, there results:

(cw 2 L 2 +6P t -w L 2 )h+W2h3 = wiL2a,

1 . . . (21)

from which the value of h may be derived by proper substitutions.

Fourth. The special case in which it is required to determine
the elevation of the joint, where the resultant, reservoir empty,
first intersects the up-stream extremity of the middle third, as at
the bottom of Zone III. No equation can be written for this case.
It must be solved by trial, using Eq. (18).
Fifth. The special case where the location of the resultant,
reservoir empty, and limiting pressures at the down-stream ex-
tremity of the joint (Rule 3), reservoir full, governs the design, as
in Zone V. For this case, the resultant, reservoir full, being well
within the middle third, Eq. (19), should be disregarded and Eq.
(20) used in conjunction with another equation embodying the

* The uplift is assumed to vary uniformly from a value, cw 2 h, at the up-

stream end of the joint to zero at the down-stream end.

70 DESIGN OF GRAVITY DAMS [Chap. V

condition of limiting toe pressures. This combination will be

taken up in the derivation of equations for compressive stresses.
32. Equations for Rule 2. Equations of Investigation. The
tangent of the angle, 6, of the inclination, with the vertical, of the
resultant, R, Fig. 12, is evidently

^ e
=0) c°J

According to Rule 2, tan 6, for rock foundations, must be equal

to or less than the coefficient of friction, /; and for earth founda-
/
tions, equal to or less than ".

Equations of Determination. As mentioned heretofore, it is nec-

essary, in non-overflow dams, to provide sufficient masonry above
the water surface to counteract the sliding force of ice. It is not
practicable to write an equation for this condition, for, if the
width of the top and the superelevation adopted on account of
other considerations is not sufficient for this purpose, there is
nothing to indicate whether an increase in the top width or an
increase in the superelevation would provide more economy for the
dam as a whole. A slight increase in either has little influence on
the quantity of masonry in the dam, because an addition of mate-
rial at this place will effect a greater or less reduction in the lower
parts. The choice must be made in accordance with the judg-
ment of the designer.
In spillway dams it not practicable to provide sufficient
is

masonry at the top to conform to Rule 2, to provide for ice thrust.

In this case it is necessary to violate the requirements of Rule 2
and compensate by some special construction, such as keys, in-
clined joints, or monolithic concrete.* Violations of this kind
are also sometimes found in the tops of concrete non-overflow
dams, where it is possible to provide monolithic concrete within
the danger zone.
Except as mentioned previously, it seldom happens that Rule 2
is a factor in the design of solid dams on rock foundations. Tan 9
should always be determined, however, and compared with the
allowed value, after each step in the design has been completed.
If, at any stage of the design, it is found
that tan 6 is greater than
the value allowed by Rule 2, then several courses of procedure
are
* See Example No. 4 of Art. 46.
 Art. 33] EQUATIONS FOR RULE 3 71

open to the designer. It is evident, in such a contingency, that

the fault lies in the lack of weight, or, in other words, lack of
resultant vertical forces, 2("FF). Additional weight may be ob-
tained by adding to the top width, increasing the surcharge, in-
creasing the batter of the up-stream or down-stream face, or a
combination of these expedients. Each
case must be decided on
its own characteristics. be found most economical
It will usually
to increase the batter of the up-stream face, as this will include an
additional vertical component of water pressure.
One of the principal reasons for the flat slope of the up-stream
face of a hollow dam is the need for the large vertical component
of water pressure to fulfill the conditions of Rule 2. A hollow dam
with a vertical up-stream face would contain, on account of this
rule, nearly as much masonry as a solid dam. The batter of' the
up-stream face of a hollow dam, if fixed by Rule 2, is found by
successive trials, using Eq. (22).
33. Equations for Rule 3. Equations of Investigation.* In Art.
22, the following general equations were derived for the deter-
mination of the maximum vertical compressive stress in the hori-
zontal joints and base:*
At the toe,
p =
, 2^Wl^
22(F) /_ 3«^
+pu> (1Qa)

At the heel,
2S
V W(f-l) +Pu "
I
(Ha)

The symbols are explained in Art. 12 and indicated in Fig. 10.

However, as explained in Art. 26, the vertical compressive
stresses are not the maximum stresses in the structure, the latter
occurring on planes which are not horizontal. The question of
the maximum, or " inclined," stresses in dams has been the sub-
ject of considerable discussion, f more particularly abroad than in

* Eqs. (12a) to (136) inclusive, being for special cases, have been omitted
for brevity, and may be easily included by the reader.

t L. W. Atcherly, Dept. of Applied Mechanics, University College, Uni-

versity of London, Draper's Co. Research Memoirs, Technical Series II;
Sir Benjamin Baker, Minutes of Proceedings, Inst. C. E., Vol. 162, p. 120;
L. W. Atcherly, Engineering, Vol. 79, p. 414; W. C. Unwin, Engineering,
Vol. 79, pp. 513, 593, and 825; Ottley and Brightmore, Minutes of Pro-
ceedings, Inst. C. E., Vol. 172, p. 89; Wilson and Gore, Minutes of Proceedings,
Inst. C.E., Vol. 172, p. 107; E. P. Hill, Minnies of Proceedings, Inst. C. E.,
Vol. 172, p. 134; William Cain, Transactions, Am. Soc. C. E., Vol. 64, p. 208.
72 DESIGN OF GRAVITY DAMS [Chap. V

America. It is unfortunate, however, that no general agreement

has been reached on this subject.
The most commonly accepted theory is that of Enger, as ex-
panded by Cain, Hill and others. The result of Enger's theory is
an inclined stress at the extremity of the joint equal to

Pi=Pv sec 2 <j).

where pv is the vertical stress and 4> the angle of inclination of

the face of the dam with the vertical. If a normal pressure, p n ,

of water or silt acting on the face of the dam, is included, this

equation will reduce to

2 2
Pi=(pv sec <t>—p n tan <t>) or pn .

In the last equation, the two values of pt correspond to the two

principal planes of stress; the governing condition being, of course,
the greater of the two.
Enger's theory, however, is not generally applicable, because

the whole argument based solely on a triangular dam, with

is

vertical up-stream face, loaded with water pressure level with the
top. The author has found that the theory will not apply to
any other shape of dam or loading; and that, for cases where the
angle, 6, of the resultant, R, is greater than pt, from the fore- <t>,

going equation, will give results too low. It is, indeed, rational
to assume that for all values of greater than the theory of <j>,

Bouvier * will apply more closely. This theory gives as the value
of the inclined stress,

Pt=pv sec2 0.

This equation probably the most accurate of any for the actual
is

maximum stress in the

foundation for all conditions.
Combining the two theories, the author proposes the following
general equations for the true maximum, or maximum inclined
compressive stresses in the dam and the foundation.
For maximum stress in the dam:
At the toe

p/ = (p/ sec V-p n

'
tan V) or p„' or p/ sec 2 0. . (23a)

* "Calculs de Resistance des Grands Barrages en Maconnerie." Annaks

des Ponts et Chausstes, Aug., 1875.
Art. 33] EQUATIONS FOR RULE 3 73

At the heel,

Pi" = (p„" sec V-P»" tan V) or p." or p„" sec 2

0. (236)

For maximum stress in the foundation:

At the toe,

pt' = p/
2
sec (24a)
At the heel,
p," = p„" sec 2
(246)

In these equations the general expression (p„ sec2 <j>—p n tan2 <j>),
gives the approximate flexural stress on planes, A, Fig. 12a, lying
close to the point, 3, and normal to the face, 2-3. The value,
p„, is, of course, the normal pressure of external forces on the plane,
2-3, close to the point, 3. The general expression, p„ sec2 0, gives
the stress between the dam and the foundation, on planes, B,
lying normal to the resultant, R.
Eq. (236) will not apply to the heel of hollow dams, if the
horizontal area of the deck is included in the area of the base (as

explained in Art. 49), the equation applying only to solid sections.

Fortunately, however, it is known that, in all ordinary types of
hollow dams having sloping up-stream faces, the maximum stress
at the heel corresponds to the bearing of the deck on the buttresses.
For practical application to hollow dams, see Art. 50.
Although the foregoing equations and the theories on which
they are based are not exact, and, for certain conditions, probably
not closely approximate, nevertheless the equations certainly indi-
cate a value for the maximum stress considerably nearer the truth
than the vertical pressure, p„, generally used.
The author does not propose a change in the present practice,
in which the shape of the section is by certain allowed maxi-
fixed
mum vertical pressures, p„, as indicated by Eqs. (10a and 11a).
It is recommended, however, that the completed section should
always be tested for maximum stresses in accordance with his
proposed Eqs. (23) and (24), and altered if found necessary.
Equations of Determination. It has been pointed out that it

is usual to consider only vertical pressures in the determination of

the shape of the section, the section afterward being investigated
for maximum stresses and changed if found necessary.
When the condition of limiting vertical toe
First. pressures,
reservoir full, and the condition requiring the resultant, reservoir
74 DESIGN OF GRAVITY DAMS [Chap. V

empty, to intersect at the up-stream extremity of the middle third

governs the design, as in Zone V. This is a combination of Rules
1 and 3.

The location of the resultant is given by Eq. (18), from which,

by using the subscript, F to represent full reservoir, there results,
,

z(Wx) f +?;(Px)f
Zf ~
X(W) r

The distance, uF , Fig. 13, is:

uF =l+y-zF =l+y ~^'

xCWh

The maximum vertical compression at the joint for full reservoir

is at the toe of the joint, and is given by Eq. (10a). Transposing
that equation, using the subscript, F , and substituting, for uF the
value just given, there results:

42(W) Fl-p,'l2 +Pu'l2 _

' ~ t+V
7 ,
2(Wx) f +2(Px)p
~
6Z(TP), 2(TF),

Solving for I, there results:

Pl^plp _ X(W)Ll = x(W)i^_ 2(wx)f+ 2 {Px)f] {

. . (25)

For empty reservoir, Eq. (20) applies:

y+ Z~ 2(WS (20)

Thus we have two equations, (25) and (20), with two unknowns,
y and I, from which the location and length of the joint may be
determined. The values, 2(1*0, 2 (Pa;), and 2(T7x), for both full
and empty reservoir, are also functions of y and I. Here, as in the
equations of determination (first case), applicable to Rule 1,
these values, in terms of y and Z, may be substituted in Eqs. (25)
and (20) and exact equations derived. However, as in the former
case, certain omissions must be made in order to provide equations
of practical application.* The author, therefore, proposes the

* See foot-note on page 67.

Aet. 34] EQUATIONS FOR RULE 4 75

same method of trial substitution of approximate values of 2(W),

(2P#), and 2(Wx),' derived by anticipating, from the general
trend of the faces of the dam, the location, y, and the length, I.
If such substitutions are made in Eqs. (25) and (20) the resulting

calculated values of y and I will be, after the first or second trial,
sufficiently accurate for all practical purposes.
Second. When the conditions of limiting pressures, reservoir
and empty, govern the design, as in Zone VI.
full

The location of the resultant is given by Eq. (18), from which,

by using the subscript, E to represent empty reservoir, there
,

results:

?(Wx) B +'2(Px) B
Ze ~
2(TP)*
The distance, uE , is:

2(Wx) E +2(Px) B
uE =l+y—zE =l+y—'-
i(ff)i

The maximum vertical compression at the joint, for empty reser-

voir, occurs at the heel, and is given by Eq. (11a). Transposing
this equation, using the subscript E and substituting for uE the
, ,

value just given, there results:

pv"l
2
+2Z(W) E l-p«"l2 _. Z(Wx) E +2(Px) E
~™
62(TT) B 2(W) E
Solving for I, there results

P"-P*'\ 2_2^^ l= 2 {why _ {

2 iWxh+ z (px)El (26)

For full reservoir, Eq. (25) applies:

-^^l -^^l = ^(W) Fy-{2(Wx) F +i:(Px) F

P 2
}. . (25)

Here we have two equations with two unknowns, y and I, from

which the location and length of the joint can be determined by
the author's method of trial substitution, as in the preceding case,
the discussion of which also applies here.
34. Equations for Rule 4. It was explained
in Art. 27 that
tensile stresses in vertical planes may
be caused by an excessive
inclination of the down-stream face with the vertical. It is seldom
76 DESIGN OF GRAVITY DAMS [Chap. V

that the shape of the section of the dam is fixed in any way by
Rule 4. However, where this rule is a governing feature, there is
no way of proportioning the shape of the section except by the cut-
and-try method.
It is obvious that the direct cause of such tension at the toe is a
relatively greater movement of the triangle, 1-2-3 (Fig. 12a),
toward the left than that of the rest of the dam. This condition
may be brought about by failure of the frictional and shearing
resistances of the plane, 1-3, or by unequal inclined settlement of
the foundation. The allowed inclination, of the face, there-
</>',

fore,must be a function of the coefficient of friction, /' (shear on

1-3 being neglected, as usual), and the settlement coefficient of the
foundation. The author knows of no satisfactory theoretical
analysis of this feature. His empirical equations, based on safe
approximate assumptions, are as follows -

For earth or pile foundations,

tan0'=<^, (27)

where H
is the height of the dam.

For rock foundations,

tan<£'=<|/
or whichever allows ,

"
/io '*" '

the greater value '

tan<2>'=< .

A/fi

where / is the allowed coefficient of friction.

The value (tan <t>'= <$•/) in Eq. (28) embodies the condition
of no tension in vertical planes, but Eq. (27) and the second part
of Eq. (28) embodies the condition of tensile stresses so small as
to be negligible.
Eq. (28), for good rock foundations, does not impose a severe
condition. Substituting a usual value of /=0.75, there results
<£' = 45°,
which is about the limit reached in high existing dams.
According to the second part of Eq. (28), a value of <j>' greater than
45° would be allowed in dams less than 10 ft. high.
35. Equations for Rule 5. Rule 5, governing the margin of
safety, is applicable only to the determination of constant assump-
Abt. 36] EQUATIONS FOR RULE 6 77

tions used throughout the design, and does not require equations
for its correct application.
36. Equations for Rule 6. It is evident that no universal
rules or equations can be written to provide against defective
details and methods of construction. Details vary considerably
with each dam, and depend mostly on the judgment of the designer
to guard against infringement of the designing rules.
 CHAPTER VI

THE DESIGN OF SOLID NON-OVERFLOW GRAVITY DAMS

37. General Considerations. As it always be found

will

convenient to start design at the top of the dam, the shape of the
top of the section is the first consideration.

A superelevation of the top above high-water surface is some-

times desirable in order to get beyond the reach of waves,* for
appearance, and for other incidental purposes. When ice pressure
is assumed to act at the level of high water, some superelevation

will be found necessary in order to fulfill the conditions of Rule 2

at the point where the ice pressure is applied. In any case it is

probable that a superelevation of about 5 per cent of the height

of solid non-overflow dams is, in general, productive of economy,
rather than an expenditure of material, although sufficient proof
has not been presented to verify this for all cases. However,
thereis no evidence to show that the adoption of a slight super-

elevation is uneconomical.
The most .economical width of top of a solid non-overflow
gravity dam is a direct function of the height of the dam, and is

dependent on the assumptions used in the design. For dams of

fairly uniform height, and designed in accordance with the usual
designing assumptions, the most economical top may be assumed
at about 14 per cent of the maximum height, f but for dams of
considerable variation in height this figure should be somewhat
reduced. It is usually considered thata value of 10 per cent is
about the minimum advisable. The width
of top. of low dams is
usually somewhat greater than that dictated by economy, as a
roadway or passageway is often necessary, as well as sufficient
width to withstand the shock of floating bodies. The details of
the top of the dam having been determined, the up-stream and
down-stream faces must be designed to conform strictly to the
designing rules hereinbefore given, it being remembered that there
*For height of waves see Art. 19.
tSee "The Economical Top Width of Non-overflow Dams," by the
Author, Transactions, Am. Sop. C. E., Vol. LXXX, p. 723.

78
 : : ;

Art. 38] EXAMPLE NO. 1 79

is no economy in battering either face unless necessary to conform

to such rules. The determination of the shape of the section is
simply a practical application of the general equations for design
previously derived. In order to explain the application of such
equations to solid non-overflow dams, the following examples are
given
38. Example No. 1. 200 ft. Solid Non-overflow Dam (Fig.
21) . Assumptions
w\ = weight of masonry =145 lb. per cu. ft.;
w>2 = weight of water = 62.5 lb. per cu. ft.;
p,' = maximum allowed vertical compressive stress at the toe
= 18,000 lb. per sq. ft.;
" = maximum allowed vertical compressive stress at the heel
p v

= 25,000 lb. per sq.ft.;

=
p maximum allowed inclined compressive stress, in the dam
t

and foundation, at the toe or heel = 42,000 lb. per

sq. ft. ;

H = maximum height of dam = 200 ft.;

c = the area of joints and base subjected to uplift = 50 per cent.

The uplift is assumed to vary uniformly from head-
water pressure at the heel to zero at the toe (there being,
for this example, no tail-water)
/=the working value of the coefficient of friction of the joints
and base = 0.75;
L = the width of top = 12 per cent of the height = 24 ft.;
o = the distance from the top of the dam to the level of highest
water surface = 10 ft.;
a' = the distance from the top of the dam to the level of the
spillway crest = 20 ft.;
P« = ice pressure =40,000 lb. per lin. ft. of dam, assumed to act
*
at the level of the spillway crest, or level 20.

As the foundation is assumed to be rock, the allowed inclina-

tion of the down-stream face of the dam may be taken from

Eq. (28), or, maximum allowed tan (j>' = -^f= 4X0

4 —^ — 75
= 1.0. There-
fore, maximum allowed angle #'=45°.
* If a rise of water surface above the spillway crest can be caused only
by a freshet in the stream, it is reasonable to assume that, during the period

of high water, the ice in the pond will be incapable of exerting pressure on
the dam, owing to the larger area of water surface.
 ;

80 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

It is seen that there are two conditions of loading, for full reser-
voir, for both of which the designing rules must be fulfilled.

First Condition. —Low water and pressure. ice

Second Condition. —High water and no pressure. ice

It is necessary, therefore, to determine, at every stage of the design,

by inspection or trial, which of these loadings will impose the more
severe conditions. be found that, for this example, the
It will
first condition of loading will govern from the top down to Level
126.0 below which level the second condition must be used. It
will be sufficient for the purpose of this example to consider in the
explanation only that loading which, for any particular joint, is

the determining condition.

The width of top and superelevation having been adopted, the
first consideration is to ascertain whether there is sufficient ma-

sonry above the level of the application of ice pressure to fulfill

the conditions of Rule 2. Referring to Art 32, the equation of
investigation for Rule 2 is

n 2(P)

AJbove Level 20:

S(P) = Ice pressure = 40,000 lb.

S (W) = weight masonry = 24 X 20 X 145 = 69,600
of lb.
. 40,000 ._„,
tan * = = 574
.

-
'
69lo00
which is seen to be well within the allowed value of 0.75.

The next consideration is to determine the level of the bottom

of Zone at what level the down-stream face must
II, Fig. 14, or

begin to batter. This evidently comes under the third case given
in Art. 31, and Eq. (21) will apply.

(cw 2 L 2 +6Pi -w L2 )h+w2 h3 = w

1 x l?a,

(0.5X62.5X24X24+6X40,000-145X24X24)/i-r-62.5/i3
= 145X24X24X20.
This equation must be solved by trial. The final value will be
found to be h = 9.27, and the bottom of Zone II will be 9.27 ft.
below the water surface, or at Level 20+9.27 = 29.27.
Art. 38] EXAMPLE NO. 1 81

It is evident that the resultant, reservoir empty, intersects the

center of the joint as long as both faces are vertical. With the
resultant, reservoir empty, at the center of the joint and the
resultant, reservoir full, at its down-stream extremity, all condi-
Rule 1 are met.
tions of
For Rule 2 we have, from Eq. (22)

2(P);
tan0 = ^
2(117
2(P) =
Ice pressure 40,000 lb.

__ 9.27X9.27X62.5
.

Water pressure ~ = ORfir -

^,ooo 1Klb.

42,685 lb.

2(TT) =
Masonry 24 X 29.27 X 145 = 101,800 lb.
9-2-X62.5X24X0.5 = _ 3>4751b
Uplift .

98,325 lb.

which is well within the allowed value of 0.75.

The requirement of limiting compressive stresses (Rule 3), is

never a governing condition for solid dams less than 100 ft. high,
on good rock foundations.
As the down-stream face is vertical, Rule 4 is not a governing
condition.
Therefore, . it is proved that the dam above Level 29.27 is

stable, and we are now ready to proceed to the design of lower

joints. The height of the dam should now be divided into a num-
ber of parts by imaginary horizontal joints, each joint being a dis-
tance from the one next above not greater than about 15 per cent
of the distance of the former below the top of the dam. Such joints
are indicated in Fig. 21, at Levels 29.27, 34.00, 39.0, etc. It now
remains to determine the length and location of such joints.
The joint at Level 34.0 is evidently in Zone III, and Case 2, of

Art- 31, applies. Choosing the point of reference on the vertical

82 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

up-stream face of the dam; y, in Eq. (19), is zero, and there

results,

21 2(Wx) F +-2(Px)r nq ,
3~ Z<TP),
k '

In order to solve this equation, we must assume, tentatively,

from a general knowledge of the trend of the down-stream faces of
dams of this type, a length of joint with which to derive factors for
substitution in the second term. Assume, for the first trial, a
length of 26.0 ft., as indicated in Fig. 18. A convenient arrange-
ment of the necessary calculations is indicated in Table VII.

lLu.o'

»•«:

Point of reference
T
\
; =26.0 (First trial) ,
|

J =26.10 '(Second trial) *J

Fig. 18.

Making the necessary substitutions, from Table VII, in Eq.

(19a), there results,
21 _ 1,975,400 = "
17.44;
3 113,256

Z = 26.160.

It is seen thatour first assumption of 1 = 26.0 ft. was a little

too small. Using the value, Z = 26.16, as a new tentative value of
length of joint, as indicated in Fig. 18, we now construct Table
VIII.
Making the proper substitutions in Eq. (19a), there results,

21 _ 1,976,210 _
°'
3 113,279

Z=26.17;

this value of I is seen to agree almost exactly with the value

assumed, and will be calledfinal. Forces and moments, for use in
Art. 38] EXAMPLE NO. 1
83

OOO
OOOI oo oo OOO
oo oo
OO-H HO
cocoas o*o oo
oo

£ I 0,

+
ft.

OOiO (DO
OtDcO OQxC oo
OtN 0<N
od»o* ON Oh
3 I
H
CO

o
o
* * fc,

M
l-l
TO
> J g g > g g ft.

•3 +
P-l-
« B +
o |3x .So
S xx x «
,|x X
o gx
§ 13
o §X5 X
U |x s
S
g s
o
g O
4) O
a

JO V e>J3

&£> EH <3
a a
o o
3a
I.SW

0) O 'Co 0) O 'CO
w
84 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

Table
subsequent calculations should, of course, be taken from
VIII. The two sets of calculations afford a valuable check on one
another.
To determine the location of the resultant for any predesigned
section, we may use Eq. (18). Placing the subscript, E to repre- ,

sent an investigation for empty reservoir, we have:

-2
,{Wx)e+^{Px)e
*""'
Z(TP),
From Table VIII we get
*
1,437,510+0 _ 10/vy
z= 12 07
- '

119,004

The distance from the point of reference to the up-stream extrem-

ity of the middle third being 26.17-f-3=j8.72 ft., it is seen that the
resultant, reservoir empty, lies well within the middle third. Since
the resultant, reservoir full, has been made to intersect at the
down-stream extremity of the middle third, all conditions of Rule 1

have been observed.

To test for Rule 2, we have,

w= W, = nt2T9 =0
, „ 2(P) P 46,120 _.„._
- 407 '

which is well within the allowed value of 0.75.

As explained heretofore, the requirement of limiting compres-
sive stresses (Rule 3) is not a governing condition for this height of
dam.
For Rule 4, we have,

tan <^=|^ = 0.459,

which is well within the allowed value of 1.0.

Thus we have designed and tested the dam above Level 34.0.
The design is continued in the same manner to the bottom of
Zone III. The results of all calculations should be tabulated, as
* In every case, that condition of loading which will give the severest

requirements for the up-stream portion of the joint should be taken for
"empty reservoir." Usually, the assumption that the reservoir is empty
the heel, and the nearest location of
will result in the greatest pressures at
the resultant to the up-stream extremity of the middle third.
Aet. 38] EXAMPLE NO. 1 85

PS
H
PS

o
o
-J

a
g u
pa
H

13
O

I
Ph
S
o
O
86 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

in Table XIII, and the location of the resultants, tan 6, and other
characteristics plotted for both conditions of loading, as in Fig. 21,
to an extent sufficient at least to foresee when it will be necessary
to change from one zone to the other, or from one condition of
loading to another, in the determination of the shape of the section.
Ordinarily, it is not necessary to calculate the compressive
stresses occurring in the upper part of the dam. This has been
done for Fig. 21, however, for the reader's information as to the
variation of such stresses.
It is seen, by referring to Fig. 21, that, when Level 87.0 is

reached, if the up-stream face is continued vertical, the resultant,

reservoir empty, will intersect outside the middle third at the next
joint to be designed. Level 87.0, therefore, marks the location of
the last joint in Zone III, and the joint at Level 102.0 must be
designed in accordance with the principle governing Zone IV.
The first case of " Equations of Determination " of Art. 31 will

apply. It will now be found convenient to locate the point of

reference well outside the base, say 10 ft., as in Fig. 21. Eqs.
(19) and (20) will apply to this case.

2l_ 2(Wx) F +2(Px) F

V± d 2(F) P '
Uy;

l__ 2(Wx) B +2(Px) E

m
V {M)
^Z~ 2(W) B
It is now necessary to adopt tentative values of y and I (Fig.
13), for the derivation of the factors for substitution in the second
term of each of these equations. From the general trend of the
down-stream face of the dam, and a knowledge that the up-stream
face must begin to batter slightly, values of y=9.7 and Z=59.5
will be adopted, as indicated in Fig. 19.
Making the proper substitutions from Table IX in Eqs. (19)
and (20), there results,

y
8, 22,460,100 y
4
^S 453,982 ''

l_ 15,675,730+0
y ^3 y,b ''
528,826
Solving these two equations, there results:
1 = 59.40,
y =9.87.
Art. 38] EXAMPLE NO. 1 87

These values of I and y are fairly close to the assumed values,

namely 3=59.5 and y=9.7, but closer values, as well as a check on
the foregoing calculations, will be obtained by adopting them for
new tentative values of I and y for a second set of calculations.

I = 69.40 (Second trial)

Fig. 19.

Substituting valves from Table X in Eqs. (19) and (20), there

results,
21 22,452,670
y+ 3 ~
453,553
=49.50,

I 15,679,410+0
y+ 3 = 528,721 = 29 68
-
'

Combining these two equations, and solving for I and y, there

results:
1 = 59.46,
y = 9.86.

These values of I and y agree very closely with the tentative values,
namely, £ = 59.4 and ?/ = 9.87, and will be called final. Thus, the
joint has been designed in conformity with Rule 1.
To test for Rule 2, we have, for the first condition of loading,

Tan for the second condition of loading, determined in the same

way, will be found to give the greater value, as indicated by Fig. 21.
Tan 6, however, for both conditions of loading is seen to be well
within the allowed value of 0.75.
88 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

<!

§
O
o
a
w
o
o

M g
a

o
H
P
0.
a
o
O
Art. 38] EXAMPLE NO. 1 89

It will be evident, from a little study, that the maximum com-

pressive stresses, both inclined and vertical, will occur at the toe
of the joint for full reservoir, and at the heel for empty reservoir.
For full reservoir we may substitute in Eq. (10a)* of Art. 33,

2(TF) = 2 (lF)r =453,550 from Table X,

1 = 59.46,

W = L= 19.82,

p„' = 0, as there is no tail-water.

The maximum vertical compressive stress for full reservoir

is then found to be:

2X453 ,550/ 3X19.82X ,„_,,„,„

Q~ [2r-59M~) + °
,
15 25 °-
'
59.46

For empty reservoir, Eq. (11a) * applies,

2(^0=2(1^ = 528,720,
Z
m= = 39.64,
|
P*" = 0,

as there is, for empty reservoir, no head-water pressure.

The maximum vertical compressive stress for empty reservoir

is then found to be:

2X528,720 / 3X39.64 \
Vv ^+U-l/,77U.
59.46 \, 59.46

For the maximum inclined compressive stresses in the dam,

Eqs. (23a) and (23&) apply. Eqs. (24a) and (246) have no prac-

* As this is a special case, with the resultant, reservoir full and empty,
at the exact extremity of the middle third, special Eqs. (12a) and (13a) of
Art. 22, would be more simple of application. In order to reduce the number
of equations to be dealt with, these equations have not been used in the
examples, but may be readily applied by the reader, simply by omitting
the parts of Eqs. (10a) and (11a) in parenthesis, which, for this special case,
are equal to unity.
90 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

tical use unless the strength of the foundation is less than that of

the masonry.
8 67
At the toe of the dam, tan <*>' =^-=0.577; sec 2 <f' = 1-330.

At the heel of the dam, tan ^" = ^=0.0080; sec2 <J>" = 1.00.
For full reservoir, tan =0.552; sec 2 =1.305.
For empty reservoir, tan =0; sec 2 =1.00.
zero at the toe for full reservoir and at the heel for empty
p n is
reservoir.

Using these values in Eq. (23a), the maximum inclined com-

pressive stress for full reservoiris found to be:

P '= (15,250X1.330-0) or 0, or (15.250X1.305),

t

P = 20,300, the greatest of these values.

f
'

In the same way, from Eq. (236), the maximum inclined com-
pressive stress for empty reservoir is found to be:

P " = (17,770X1-0), orO, or (17,770X1),

4

P" = 17,770, the greatest of these values.

All vertical and inclined pressures at this joint are seen to be
well within the allowed values, and, therefore, Rule 3 is followed.
To test for Rule 4, we have only to observe that the value of
tan <t>', as just derived, is well within the allowed value of 1.0.
Thus we have designed and tested the section above level
102.0.
The design is continued in the same manner to the bottom of
Zone IV. After the joint at Level 120.0 has been reached, it will

be seen, by reference to Fig. 21, that the lines of resultants for the
two conditions of loading will cross before the next joint is reached.
It will be obvious, therefore, that the second condition of loading
will govern from here on; namely, water surface at Level 10 and
no ice pressure.
When the joint at Level 140.00 has been designed, the diagram
on Fig. 21 will indicate that, unless proper precaution is taken, the
vertical pressures at the toe for full reservoir will exceed the allowed
limit of 18,000. Level 140.0, therefore, marks the elevation of the
last joint in Zone IV, and the joint at Level 160.0 must be designed
Art. 38] EXAMPLE NO. 1 91

in accordance with the principles governing Zone V, namely, Rule

3 for full reservoir and Rule 1 for empty reservoir. Eqs. (25) and
(20), of Art. 33, apply to this case.

6
V— 2(If%l = 2,(W) Fy-{Z,(Wx) F +2(Px) F }. . (25)

l_ V(Wx) E +-2(Px) E
y+ 3 ?j(W)e
(20)

As before, it is necessary to adopt tentative values of I and y.

These will be assumed at 107.0 and 6.0, respectively.

Zo= 87.01

I .= 107.0 (First trial)

I =100.6 (Second trial)

Fig. 20.

Making the proper substitutions from Table XI in Eq. (25),

there results,

0zJM22p_9?if9i=991,900 2/ -74,513,440,

which reduces to
2
Z +110.21Z=-330.63?/+24,838. . . (25a)

Making corresponding substitutions in Eq. (20), there results:

I 49,580,000+0
y+ 3 =
,

1,212,780
An 87
40 C7 -
/nn .
(20a)

Combining Eqs. (20a) and (25a), and solving for I and y, there
Vesults,
1= 106.5,
y=5.37,
92 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

a
a
o
s

d
O
J

I
P
§
O
o
Abt. 38] EXAMPLE NO. 1 93

3
I
Q
S3
O
U
H
02

O
O
to

H
M

K
O
ft

la
o

S
o
O
94 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

These values of I and y are somewhat different from the values,

1= 107.0 and j/ = 6.0 adopted tentatively. '
Closer values, as well
as a check on the calculations, will be obtained by using them for

new tentative values for a second set of calculations.

Making the proper substitutions from Table XII in Eq. (25),

there results,
0-18,000
;2 _ 997 470 = QQ?
;
_ 74^430,
6 o

which reduces to
2 +110.83Z=-332.492/+24,875. . . (25a')
Z

Making corresponding substitutions in Eq. (20), there results,

"f^SSr— ™
Combining Eqs. (20a') and (25a'), and solving for I and y, there
results,
= 106.4,
1

y = 5.30;

which is sufficiently close to the tentative values of Z = 106.5 and

y = 5.37 and will be considered as final.
The necessary investigations to determine whether or not the
other designing rules are complied with are, in every respect,
similar to those described for the upper joints, and need not be
repeated. The results are indicated in Fig. 21.
The joint at Level 180 is determined in the same manner, and,
after the necessary calculations have been completed and plotted
in Fig. 21, it will be noticed that, unless proper precautions are
taken, the maximum allowed vertical pressures at the heel, for
empty reservoir, be exceeded. The joint at Level 200,
will
therefore, will he within Zone VI, and must be designed by the
application of Eqs. (26) and (25), of Art. 33, in the same manner
as Eqs. (25) and (20) were applied in the determination of the
length and location of the joint at Level 160.
The angles in the faces of the designed section may be smoothed
up, if desired, as indicated by the broken line for the angle at Level
29.27 . In low dams the down-stream face is usually made straight,
as indicated in Fig. 35, in order to save the expense of curved forms,
for the concrete.
To face 'page 95
Art. 38] EXAMPLE NO. 1 95

be noticed that the design has not been influenced by

It will
the maximum allowed values of tan
8, tan </>', or maximum inclined

pressures. Although this is usually the case for solid dams not
exceeding 200 ft. in height, it will be well to indicate here the proper
method of procedure, if, at any stage of the design, it is found that
such values are determining conditions.
Referring to Fig. 21, it is seen that tan 6 reaches a maximum
value of 0.712 at Level 180.0. If it is considered that this value is

too high, be necessary to provide an additional vertical,

it will
downward force, namely; an increased S(W), as it will be seen from
Eq. (22) that tan 8 decreases as 2 (IF) increases. This may be
accomplished simply by adding more masonry, but, usually, it
will be found best to increase the batter of the up-stream face,
as by doing this, a larger vertical component of head-water pres-
sure will be included.
A reduction in the value of tan <f>'may be made in a number of
ways. If the section is redesigned with a greater superelevation
or width of top, the resulting value of tan <j>' will be less; but, in
very high dams, the change will be relatively small. The desired
reduction may also be obtained by arbitrarily thickening the mid-
dle and upper part of the section and redesigning the lower part.
Such increase in thickness can be obtained by increasing the batter
of the up- or down-stream faces. It is probable that the best
remedial expedient, for most cases, to increase the batter of the is

up-stream face from about mid-height to the base, and to redesign

the lower part of the dam with the new up-stream face as a fixed
condition in the calculations.
The inclined pressures, for a dam of this type, will always be
greater at the toe than at the heel, notwithstanding the fact that
the opposite is true of the allowed vertical pressures. It is seen
from Eq. (23a) and (24a), that the maximum inclined stress in
the masonry at the toe of the dam is proportional to the vertical
stress, p,', and to sec </>' or sec 8, whichever is the greater; and
that the maximum inclined stress in the foundation is proportional
to p/ and sec 9. It will be found that a reduction in the inclined
stresses in the dam can be obtained to best advantage by reducing
sec <j>', if 4>' is greater than 8, or by reducing pv ', if 8 is greater than
(j>'. To reduce the inclined stress in the foundation, a reduction
in p/ will usually prove best.
96 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

TABLE XIII

Results of Calculations fob Example No. 1 of Art. 38

Art. 38] EXAMPLE NO. 1 97

TABLE XIII—Continued
 Tangent 9
o.o 0.4 as 0.3 0.1

Loading 1-Wat*r Surfaoe and lee Pressure at Level ^5

I Tail-water At Level 78
Loading 3 -Water Surfaoe at Levelli

140 100 80 60 20
Cubic Yards per Linear Foot

To face page 99
 ; ; ;

Akt. 39] EXAMPLE NO. 2 99

39. Example No. 2. 102-ft. Solid Non-Overflow Dam

(Fig. 22).In this example the assumptions are the same as in
Example No. 1, Art. 33, with the following exceptions:
H = maximum height of dam = 102 ft.;

L = width of top = 14 per cent of height = 14 ft.

a = the distance from the top of the dam to the level of high-
water surface = 5 ft.
a' = the distance from the top of the dam to the level of the
spillway crest = 15 ft.

a" = the distance from the top of the dam to tail-water surface
= 72 ft.;

c = the area of joints and base subjected to uplift = 50 per cent.

The uplift is assumed to vary uniformly from head-
water pressure at the heel to tail-water pressure at the toe.

Except that the negative force and moment, of tail-water is

present, the method of design to be followed is exactly the same
as that indicated for the upper part of Example No. 1, and will
not be repeated.
The design is governed by low water, with ice pressure down
to Level 65.0, and high water with no ice below that elevation.
The results of computations are indicated in Table XIV, and are
plotted in Fig. 22.
It will be noticed that the dam does not extend below Zone IV,
the pressures at the base with resultants at the middle third
extremities not exceeding the allowed values, as indicated by the
calculations which follow.
It will be evident, from a little study, that the maximum com-
pressive stresses, both inclined and vertical, will occur at the toe
of the dam for full reservoir, and at the heel for empty reservoir.
For full reservoir we may substitute in Eq. (10a) *, of Art. 33,
the following known values from Table XIV.

2(F) = Z(T7F) =395,700,

1=68 A,
68.4
w = -g— _
= 22.8,
'= 0.5X30X62.5 =937.5.
pu
* See footnote, p. 89.
100 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

TABLE XIV
Results of Calculations fob Example No. 2 of Akt. 39
Abt. 39] EXAMPLE NO. 2 101

TABLE XIV—Continued

Moments of Vertical Forces.

Level
of
Joint.
102 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

TABLE XIV—Continued
Art. 39] EXAMPLE NO. 2 103

The maximum vertical compressive stresses for full reservoir

is found to be:

For empty reservoir, Eq. (11a) * applies,

S(T7) = S(Tf) B = 497,300,
2
w=
-><f^=45.6,
'

Pu"=0.
The maximum vertical compressive stress for empty reservoir
is then found to be:

„ 2X497,300/3X45.6 A,. ,_„

1+0=14,650.
.

P» =-^Ta
68.4 \ "«Ql
68.4

For the maximum inclined compressive stresses in the dam,

Eqs. (23a) and (236), of Art. 33, apply. Eqs. (24a) and (246) have
no practical use unless the strength of the foundation is less than
that of the masonry.

At the toe of the dam,

19 (VI
tan d> '=——— = 0.75
tan0'=^p O.752; tan2 <t>'
=0.566; sec2 <f' =1.568.
lo

At the heel of the dam,

0.66
,
tan<^"
.„
tan 4," = ^g- °^ = 0.0412; tan2 0" = 0.0017; sec2 (/." = 1.0016.

For full reservoir,

tan = 0.672; sec 2 = 1.455.

For empty reservoir,

tan = 0; sec 2 0=1.00.

Using these values in Eq. (23a), the maximum inclined com-

is found to be:
pressive stress for full reservoir

p/ = (12,500 X 1.568 - 30 X 62.5 X 0.566)

or 30X62.5 or 12,500X1.455,
p/ = 18,540 or 1,875 or 18,200,
pt = 18,540, the greatest of these values.
* See foot-note, p. 89.
 8

104 SOLID NON-OVERFLOW GRAVITY DAMS [Chap. VI

In the same way, from Eq. (23b), the maximum inclined com-
pressive stress for empty reservoir is found to be:

Vt" = (14,550-0) or or 14,550X1.0016,

p/' = 14,550, the greatest of these values.

40. Comparison of Non-overflow Dams. A comparison of

high, solid, non-overflow dams is given in Fig. 23. With the
exception of the Olive Bridge final section, the designs of these dams
were all made in accordance with the same general theory, the
differences in areas and shapes being affected solely by the assump-
tions, as indicated in Table XV.
The theoretical section, of the Olive Bridge Dam was arbi-
trarily increased to its final section, on account of the importance
of the structure.
TABLE XV
Comparison of Solid, Non-overflow Dams, Showing Assumptions
Used in Design
(See. Fig. 23)

Maximum
Percentage Allowed Verti-
Unit Wt. Total Ice
of Area of cal Pressures,
of Mason- Pressure,
Dam. Base Sub- in Lbs. per Sq.ft.
ry, in Lbs. in Lbs.
jected to
per Cu.ft. per Lin.ft.
Uplift.
Toe. Heel.

Olive Bridge, theoretical section 145.8 47,000 40,000 40,000

Olive Bridge, final section 145. 661 47,000 12,200 23,000
New Croton 156.2 33,400 30,800
Elephant Butte 140.0 33J 22,000 28,000
Wegmann's Practical Profile No. 3. . . 145.8 16,800 20,600
Morrison and Brodie's example of
design 146.0 28,000 36,000
The author's Example No. 1 145.0 50 40,000 18,000 25,000

The maximum vertical pressures indicated for the Olive Bridge

probably do not exist, as the maximum section is in a
final section
narrow gorge confined on both sides by good rock, between which
the dam is wedged, without the possibility of movement.
In the upper part of the New Croton Dam the vertical pres-
sures were limited to 16,400 and 20,600 lb. and heel,
at the toe
respectively, and, in the lower part, 33,400 and 30,800. This
difference in allowed pressures, in the top and bottom of this dam,
has been severely criticised.
To face page 104
 CHAPTER VII

THE DESIGN OF SOLID SPILLWAY GRAVITY DAMS

41. General Considerations. The general method of design

of solid spillway dams differs in no way from that previously
described for solid non-overflow dams, except at the crest which, as
mentioned in Art. 17, should be proportioned to fit the lower
nappe of the sheet of water spilling over the dam during maximum
flood.
The application of the general equations of design has been
described in Example 1, Art. 38, and will not be repeated for the
following examples.
42. The Shape of the Crest. It was pointed out in Art. 17
that, if the sheet of water spilling over the crest leaves the face of
the masonry, there is danger of the existence of a considerable

indeterminate overturning force due to the formation of a partial

vacuum under the sheet. This, of course, will not occur if the

continuous length of crest is unusually short and free access to

the atmosphere is At any rate, it is very
provided at the ends.
desirable to avoid impact and vibration, and the best practice
dictates that the crest and down-stream face should always com-
pletely fill the space under the lower nappe of the sheet corre-
sponding to the maximum flood to be expected.
Experiments have been made to determine the shape of the
sheet of water flowing over aerated sharp-crested weirs. The
general form of the sheet is indicated in Fig. 24. If the area below
the lower nappe is filled with masonry, the shape of the sheet and
the discharge will not be changed appreciably. The resulting
section will not only fit the sheet of water, but will provide the
maximum possible discharge, as the water will pass the masonry
crest with no disturbance or unnecessary contraction.
This shape of crest has become standard in modern designs
and, with few exceptions, differs for individual cases only in the
methods used in determining the shape of the sheet. For refer-
ence, therefore, it will be designated the " standard dam crest."
105
106 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

A comparison of the shape of the crests of several modern dams is

made in Fig. 38.

Bazin's experiments for the shape of the sheet of water flowing
over a sharp-crested weir have been translated by Messrs. Arthur
Marichal and John C. Trautwine, Jr.* The experiments for weirs
with vertical water faces and those with water faces inclined 45°
apply directly to the determination of the shape of the crest of
the ordinary types of solid and hollow dams, respectively. The
curves indicated in Fig. 25 and 26 to about x = +0.12 for the upper
nappe and x = +0.65 nappe are plotted directly from
for the lower
the experiments. The extension
of the experimental data was
made by the author in the following manner:
The average velocity in any normal section of the sheet of
water was found by Bazin to he very close to one-third of the dis-
tance from the lower nappe, as at point 4, Fig. 24. The curve,
2-3-4, was drawn through three scaled points of average or result-
ant velocities. The curve has the form of a jet spouting with an
initial horizontal velocity, v h . Its equation, therefore, is that of a
parabola, and was derived as follows:
In the time, t, a particle on the curve will have fallen, from
rest, a vertical distance from point 1, equal to

x'-*V-
X ~ 2'
therefore,

e-K
In this same time, t, the particle will have moved horizontally,
from point 1, a distance of
y' = vj,t;

therefore,

-(9"
Equating these two values of t 2 , there results:

„ 2v h2
V
~T
,

'
••••••• (29a)

* Proceedings Engineers' Club of Philadelphia, Apr., 1893.

Art. 42] THE SHAPE OF THE CREST 107

which is the equation of the curve, 1-2-3-4, referred to the origin

at point 1.

Water Surface
108 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

Fig. 25.
Art. i'2] THE SHAPE OF THE CREST 109
110 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

where A is the thickness, 6-7. Values of q for the two types of

crests, calculated in this manner, are given in Table XVIII. It is

surprising to note that the calculated discharge agrees almost

exactly with the discharge found by independent experiments
for the same types of weirs.
The thickness of the sheet corresponding to any value of x'
may now be derived from Eq. (30), the values of v h and q being
taken from Table XVIII. The area of the sheet lies one-third
below and two-thirds above the curve.
In this manner the path of the falling sheet was determined
and plotted in Figs. 25 and 26. The paths are only approximate,
as they are extended from experimental points relatively close to
the top of the dam, and, moreover, theymay be somewhat affected
by masonry crest. Consequently, it is advis-
irregularities in the
able to provide a margin of safety by extending the masonry line
well into the theoretical sheet as indicated. This line is usually
formed by a series of arcs of circles, as indicated in Fig. 36.
The experiments were made with a negligible velocity of ap-
proach. The effect of velocity of approach on the shape of the
sheet not being known, the best that can be done is to let h, Fig.
head corresponding to the velocity of approach and
24, include the
to consider that, when the velocity of approach is large, the
results so obtained are correspondingly uncertain.

TABLE XVIII
Factors in the Determination op the Shape of the Crest, (h =1.0 ft.)

Factor.
Akt. 43] DISCHARGE CAPACITY 111

discharge of water over dams may be expressed by the following

equation:
Q = ql„ = CL{(h c +hf 2
-h 3/2 },
e .... (31)

where, Q=the total discharge, in cubic feet per second;

g = the discharge per linear foot of effective crest;
Z»=the net or effective length of crest, i.e., the total length
of crest corrected for end contractions due to piers
and sharp-cornered abutments;
7k = the actual or measured head on the crest, taken at a
point sufficiently remote from the dam to avoid the
surface curve;
ft«=the head corresponding to the velocity of approach;
and
C=a which depends on the shape
coefficient of the crest
and the head on the crest.

In determining the head on the crest corresponding to the maxi-

mum flood to be expected, the following approximate equation
may be used; provided the head on the crest is not greater than
the depth of the channel of approach, the error for that head
being much smaller than the error to be expected in determining
the maximum flood:
'

Q = ql„ = Cln (h +Kf*

c (31a)

Francis' equation for the necessary correction due to complete

end contractions is,

ln=lt — 0.lnh e ,

where, It = the total or gross length of crest between abutment

and piers (Fig. 27), and
n = the number of complete end contractions.

If the crest is obstructed by piers having considerable widths

and sharp corners, as indicated in Fig. 27, n represents the number
of corners which serve to deflect the water, there being six com-
plete contractions in this instance. Usually, however, the piers
are relatively thin and are provided with sharp up-stream ends, as
indicated in Fig. 28. In such cases the contractions for the piers
are not complete, and Francis' equation would give values of h,
112 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

too small. The number of experiments has not been sufficient

to determine closely the effect of piers on the effective length of

crest.
Unless the piers are unusually thin, relative to the head on the
crest, or very considerably pointed up-stream, they may be con-

A N

! 1
1 ili\v

Fig. 27. —Complete Contractions Due to Abutments and Large Piers.

sidered to offer a partial contraction, probably amounting to not

more than 0.04 hc for each contraction, in case the piers are pointed
as indicated in Fig. 29, and varying between this limit and 0.1 hc
for thick, blunt piers, depending on the degree of sharpness and
the relative thickness.

Fig. 28. —Partial Contractions Due to Sharp Piers.

Francis' contraction equation may then be written,

t = &— A<,(C n„-r-C»n» . . . Cnn„), . . . (32)

where C„, d etc., represent the contraction coefficients applicable

to the several different contractions which may be
expected, and
na nt
,
etc., the number of contractions having contraction
coefficients, Ca Cb
, , etc., respectively.
 .

Art. 43] DISCHARGE CAPACITY 113

Thus, if the piers in Fig. 28 are shaped as indicated in Fig. 29, the
effective length of crest would be,

ln = l,-h (0.lX2+0MX-i) =
c lt -0Mh c.

Obviously, the effective length of crest between any two piers

of whatever shape or, in the absence of piers, between the abut-
ments, can never be less than 0.788 of the clear length of crest, as
this reduction corresponds to
the contraction of the width of
a jet through an orifice in a thin
plate.
The head corresponding to the
mean velocitv, *>i, in the chan-
2
vx
nel of approach is -^

The velocity in the channel

of approach, however, is not uni-
form, the filaments above the
elevation of the crest of the dam
sometimes having a velocity con- Fig. 29. Typical Sharp Nose Pier. —
siderably greater than the mean,
depending on the depth and width of the channel, and its surface
conditions. The energy of the filaments above the elevation of
the crest have a proportionately greater effect in increasing the
discharge. The true value, h may be represented by ,

hr==Cf
Yg'

where C depends on the condition of the channel of approach

described above. Values of 0„ usually adopted in experiment
work vary from 1.0 to 1.5..
However, in view of the fact that the velocity of approach
above the elevation of the crest may be materially affected by
wind, ice, and other conditions, it would seem advisable to use a
value of C
= 1.0, particularly as such assumption is on the safe
side, when determining the capacity of the dam to pass the max-
imum flood.
114 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

Values of C, for Eqs. (31) and (31a), applicable to standard

dam crests, may be taken from Fig. 30 which, in the absence of
authentic experiments on this type of crest, was derived by the
author from theoretical considerations and a comparison of a
number of experiments on similar shapes of crests.
The values of the coefficients, for the head ratio of unity, are
probably within 3 per cent of the truth; but, at other points on
4.0

3.9

8.8

&3.7

au
.§3.6

8
3.3

6.8

'3.1
Art. 44] THE BUCKET 115

If the crest of the dam is submerged, as in Fig. 31, the dis-

charge coefficient, for use in Eq. (31) and (31a), should be modified
according to the degree of submergence, as indicated in Table
XIX.* C is the coefficient tor free discharge over a
In this table,
similar crestunder the same head, and C" is the modified coeffi-
cient due to the submergence. The heads are hs and hc as in ,

Fig. 31.

It will be noted that, for values of ^ less

hc
than 0.30, the reduc-

in discharge is less than 3 per cent.

Fig. 31. —Submerged Crest.

TABLE XIX
Relative Coefficients, Submekged Crest and Free Crest

h,
he
116 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

Although a parabola has been used in some instances for the

curvature of the bucket, an arc of a circle has been adopted almost
invariably, the former being an unnecessary refinement to establish
a gradual transition. The proper radius of the bucket depends on
the thickness of the sheet of falling water at the toe of the dam, the
height of the dam, the character of the foundation, the quantity
of debris and ice to be passed, the frequency of floods, and the
depth of tail-water. For usual conditions, without tail-water,
radii corresponding to those indicated in Fig. 32 are recommended.

I 3 -TLL
 ;

Akt. 45] EXAMPLE NO. 3 117

construction may be deferred until such time as its need becomes

apparent; providing that the toe of the dam is not submerged,
rendering it accessible only by coffer-dams. For a submerged
toe, the tendency is to neglect inspection, and the omission of a
bucket, under such conditions, may affect the future safety of the
structure.
The bucket, when used, is an addition to the predesigned sec-
tion,and is not considered in the design.
45. Example No. 3. 91-ft. Solid Spillway Dam, Without Ice
Pressure (Fig. 33). Assumptions:
H = maximum height of dam = 91 ft.;

c = the area of joints and base subjected to uplift = 50 per cent.

The uplift is assumed to vary uniformly from head-water
pressure at the heel to zero at the toe (there being for
*
this example no tail-water) ;

Pi = ice pressure = none

h c = maximum head on the crest = 10 ft.

Other assumptions as given for Example 1, Art. 38.

The first step in the design is to fix the shape of the crest and the
upper part of the down-stream face, as described in Art. 42, to
prevent the sheet of flowing water from leaving the face of the
masonry. The origin of co-ordinates being taken on the pro-
longation of the up-stream face with Level 15. The co-ordinates
of the masonry fine and the theoretical sheet may be obtained
by multiplying all quantities in Table XVI by 10, the maximum
head on the crest, in feet.
As the and la have no
considerations applicable to Zones I
influence on the shape of the section, they will be treated later.
The lower limit of Zone II, where the resultant, reservoir full,
first intersects the down-stream extremity of the middle third

can be found only by trial calculation, using Eq. (18), Art. 31.
In Zone III the slope of the down-stream face lies outside of
the theoretical sheet of water in order to keep the resultant, res-
ervoir full, within the middle third.
In Zone IV the resultants, reservoir full and empty, intersect
the joints at the exact extremities of the middle third.
For the use of the general equations of design the reader is
referred to Example 1, Art. 38. Results of calculations for Ex-
* See foot-note, p. 126.
118 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

ample 3 are given in Table XX. The maximum compressive

stresses at the base are, for this example, well within allowed
limits, and have no influence on the design.
Above the bottom of Zone la, at the top of all spillway dams, the
tangent of the angle of inclination of the resultant with the ver-
tical is greater than the allowed value, being equal to the allowed
value at the bottom of Zone la and increasing to infinity at the
crest of the dam, as indicated in Fig. 33. It is evidently impos-
sible to provide sufficient weight of masonry to prevent sliding
by friction alone and, above the bottom of Zone la, monolithic
concrete is generally used, although in the lower part of this zone,
building joints inclined at an angle of 90° — with the resultant
would be acceptable. Key-ways, having a shearing value suffi-

cient to resist the horizontal forces, may be used when it is desir-

able to have horizontal construction joints. It is permissible in
many cases to increase considerably the allowed value of 6, within
Zone la, when a failure by sliding of the top of the dam would
cause comparatively damage, particularly if a considerable
little

percentage of plums is may be sometimes found to

used, which
possess sufficient shearing and bearing value to resist the hori-
zontal forces.
It will be noticed, in Fig: 33, that near the top of the dam
(within Zone I), the resultant for full reservoir falls outside the
middle third. This is also a condition common to all spillway
dams, and can be met only by monolithic construction (if, as is
usual, the tension is inconsiderable), or the use of steel rein-
forcement, as indicated in Art. 46.
46. Example No. 4. 87-ft. Solid Spillway Dam, with Ice
Pressure (Fig. 34). This example is given in order to indicate
the effect on the shape of the section, of the assumption of con-
siderable ice pressure. For this example, ice pressure of 20,000 lb.
per lin. ft. of dam is considered as acting at the elevation of the
crest. Aside from ice pressure and tail-water, the assumptions
for this example are the same as those for Example 3, Art. 45. It
will be noted that there are two conditions of loading, as in Exam-
ple 1, Art. 38.

First. —Low water, and pressure at the

ice crest of the dam;
Second. —High water, and no pressure. ice

It is seen that, even with the greatly increased thickness which

Tangent 9
Art. 46] EXAMPLE NO. 4 119

TABLE XX
Results op Calculations fob Example No. 3 of Art. 45
 ; ; J

120 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

has been adopted for the top of the dam, tension exists above
Level 45.0 for the first condition of loading. The height of the
portion of the dam in which tension exists may be reduced by-
increasing the thickness of the top of the dam. The proportions
can be settled only by the judgment of the designer.
As it would be impracticable to provide monolithic concrete
between the crest and Level 45, steel reinforcement must be used
to resist the tension. The reinforcement should be computed in
accordance with the theory of "flexure and direct stress,"* and
a uniform uplift equal to head-water pressure should be assumed
to act over the entire area up-stream from the neutral axis of any
horizontal plane.
Fig. 35 indicates sections of the dam for water storage on the
Upper St. Maurice ftiver,Province of Quebec. The temperature
of the site of the dam drops to —60 or — 70° F., and the range is
about 160°. Ice pressure of 50,000 lb. per lin. ft. was assumed to
act at crest level. The lines of resultants in the figure indicate
clearly the need of reinforcement at the up-stream face.
47. Example No. 5. 30-ft. Solid Spillway Dam (Fig. 36).
Assumptions:

H = maximum height of dam = 30 ft. above. good rock; '

c = the area of joints and base subjected to uplift = 30 per cent.

The uplift is assumed to vary uniformly from head-
water pressure at the heel to tail-water pressure at the
toe;f
w\ = weight of masonry = 145 lb. per cu: ft.
w2 = weight of water = 62.5 lb. per cu. ft.;
w>3 = weight of silt in air = 125 lb. per cu. ft.;

a = angle of repose of silt in water = 0°; %

K = voids in silt = per cent;{
/=safe value of the coefficient of friction of the joints and
base = 0.75;
Pi = ice pressure = none;
Qm = maximum flood to be expected = 32,400 cu. ft. per sec;
= total length of crest = 200 ft.
It

*See "Principles of Reinforced Concrete Construction," by Turneaure

and Maurer. 2d Edition. John Wiley & Sons. 1910.
f See foot-note, p. 126.
J Assumed as liquid mud. See Art. 16,
 Tangent 9
0.4 0.3

120 100 80 60 40 20
Cubic Yard3 per Linear Foot, not Including bucket or cut-oa

Fig. 34. To face page ISO

Art. 47] EXAMPLE NO. 5 121

Table XXI

Results op Calculations for Example No. 4 op Art. 46

122 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

/„ = effective length of crest = /, minus the effect of two com-

plete end contractions;
lc — width of the channel of approach = 210 ft.

Maximum Head on the Crest

The velocity of approach, from Fig. 36, is

Qn 32,400 154
»i = ~~
4.(*.+8) ~210(M-8) Ac+8"
ELSS3.0

: :S^i> ::A v.

NON-OVERFLOW SECTION
Fig. 35. —Upper St. Maurice River Dam, Showing Reinforcement to Resist
Ice Pressure. (Eng. Record, Vol. LXX, p. 394.)

The head corresponding to the velocity of approach is

v£ / 154 \ 2 368
_J = *

2flf 2X32.2^+8/ (Ac+8) 2

As there are two complete end contractions, corresponding to
two sharp-cornered abutments having crest coefficients of = 0.1, C
Eq. (32) may be written,*
h = lt-hc (0.1X2),
l„ = 200 -0.2fc r.

* The effect of end contractions, for this example, is negligible, but is

included in order to indicate the application of Eq. (32).
Art. 47] EXAMPLE NO. 5 123

The shape of the crest is to be proportioned to fit the sheet of

water corresponding to the maximum flood to be expected; there-
fore the coefficient of discharge, from Fig. 30, will be C = 3.94,
corresponding to a head ratio of unity.
Substituting these values in Eq. (31a), there results,

3'*
368
32,400 = 3.94(200 - 0.2k)
1
* c+ 2
(k+8)

This equation, being solved by successive trial substitutions, there

results,
K= 11.0 ft.,

whence
t'i = 8.1 ft. per sec,

A„=1.0ft.,

g = -^=164 cu. ft. per sec.

These values are indicated in Fig. (36). The head, hc , is the actual
pressure head on the crest, and should fix the water surface to be
used in detennining the static head-water pressures on the dam.
The head, K+hr, is the total head on the crest as affecting the
discharge, and should be used, as hereinafter described, in fixing
the shape of the crest. The curve of the upper nappe may be
drawn approximately, as indicated, from the plotted curve to the
actual water surface.

Impact of the Approaching Water

From Art. 14 it is seen that the unit pressure from the impact
of the approaching water corresponds to approximately twice
velocity head, or

Unit pressure = 2u>2k=2X62.5Xl.0 = 125 lb. per sq. ft.

The pressure is assumed to be distributed uniformly over the area

above the silt.

Tail-water

The depth of tail-water at maximum flood is assumed to be

20 ft. From Art. 14, it is seen that the pressure of tail-water may
or may not act upon the dam, depending on its depth relative to
 '

124 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

the height of the dam, and the discharge. The determination of

the depth, h 5 of tail-water which is apt to be swept away from the
,

toe of the dam, as indicated in Fig. 4, is obtained from Eq. (5),

<? h 32 h3
h= 4.I6.U3 4 2

where, g = the discharge per linear foot of crest, as previously

determined;
and fa = the thickness of the sheet of water taken from Fig. 36,
the shape of the sheet having been determined as
hereinafter described. The thickness, hz, should
always be measured at the elevation of the toe of the
bucket.

Making the proper substitutions, there results,

.
/l5 = /164X164.+_
3.2X3.2 3.2 = 01 _
4 ^- 2L3
,,
ft '
Vl6.1X3.2

This value, hs, represents the minimum depth of tail-water

which will resist the force of the falling water and remain at its
full depth adjacent to the dam. Theoretically, a depth less than
/is would be swept away, and that greater than hs would remain in

contact with and press against the dam. It was explained in Art.
14 that, owing to the inexactness of Eq. (5), the dam should be
tested for stability with and without tail-water, if as in this exam-
,

ple, the computed value, h 5 is within 20 per cent of the actual

,

depth.
It will be seen that, in this example, the pressure of tail-water
has a negligible effect on the location of the resultant. This con-
dition, however, is not common to all cases.

Silt Pressure

From Eq. (8), Art. 16, the effective unit weight of submerged
silt is

W3 = W3' — u)2 (l — k),

where, = its unit weight in air;

103'

w% = the unit weight of water and ;

= the percentage of voids.

fc
Art. 47] EXAMPLE NO. 5 125

Making the proper substitutions, there results,

w 3 = 125 -62.5(1-0) =62.5 lb. per cu. ft.

The total pressure, P, above any horizontal joint may be found

from Eq. (7),

p _ wzh /l — sin a\
2

~2~\l+sma)'

where h is the depth of silt above the joint and a its angle of repose.
Therefore, since a is zero,

p = 62|^ = 3125A2

As explained in Art. 16, it is unusual to consider silt pressure of

this kind and uplift to act same time. It is customary to
at the
test the dam for stability with uplift and before sedimentation, and
also for no uplift and complete sedimentation. The two forces,
however, are here assumed to act simultaneously in order to sim-
plify the calculations.

Shape of the Sections

As previously indicated, the total head, hc +K = 12.0 ft.,

should be used in fixing the shape of the crest and down-stream

face. The proper shape of the crest and down-stream face to fit
the sheet of spilling water may be obtained by multiplying all

quantities in Table XVI by 12.0.

The masonry line having been laid out, the bucket may be
added, as indicated in Fig. 36. The proper radius of the bucket
for ordinary rock foundations may be taken from Fig. 32. A
height of dam, above the bucket, of 28 ft. and a head on the crest
of 12 ft. corresponds to an 18-ft, radius of bucket.

Test foe Stability

The shape of the section having been adopted, it should be

tested for stability, and found necessary.
altered if

It is evident that the total length of bucket is probably in-

sufficient to transmit the stresses to the foundations. The effective
length of base may be determined by first calculating, from Eq.
(28) of Art. 34, the maximum allowable inclination, <£, of the
 ^

126 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

down-stream face (Rule 4), it being usual to neglect that portion

of the section lying outside of the line of maximum inclination, as
indicated by the shaded area in Fig. 36:

tan0 = g/, or ^-^.

Making the proper substitutions, there results,

tan <j> —
= 4X0.75 . or
(10
-\/™ = l-0
,.
or
_ ___
0.577,

tan = 1.0, which is the greater of the two.

Therefore = 45 degrees.
The necessary calculations, including tail-water, for the deter-
mination of the stability of the dam above Level 42.0 willnow be
given. Calculations excluding tail-water may be made in the
same manner.
The location of the resultant (Rule 1) may be determined
from Eq. (18) of Art. 31. Using the subscripts, E and F to repre- ,

sent empty and full reservoir, respectively, referring to Fig. 13,

and taking the point of reference on the up-stream face of the dam,
we have,
_ 2(Wx) E +2(Px) E

*
_ 2(Wx) F +2(Px)r
2(W) F
The necessary calculations are indicated in Table XXII.
Substituting in the foregoing equations,

1,359,000+0*
'
102,600

Z* = 1,994,000 = 2L4
01 .

-937I90-

Plotting these values on Fig. 36, it is seen that the resultants

he well within the middle third.

* The pressure of silt is neglected for empty reservoir, representing the

condition before the water first fills the reservoir.
Art. 47] EXAMPLE NO. 5 127

1—1

g J
EH g
128 SOLID SPILLWAY GRAVITY DAMS [Chap. VII
Art. 47] EXAMPLE NO. 5 129

03
a
o
a
o
6
CO

3
aB
O
PQ

ffl
O

a
130 SOLID SPILLWAY GRAVITY DAMS [Chap. VII

investigate for safety against sliding (Rule 2), we may use

To
Eq. (22), of Art. 32. Evidently, Rule 2 is fulfilled for empty res-
ervoir. For full reservoir, we have,

tan5 - 534 '

-2(l0;-937l90-

which is well within the allowed value of 0.75.

Referring to Fig. 36, it is seen that the more severe condition,

as affecting Rule 2, is without tail-water, although tan 8 is still

well within the allowed limit.
Unit pressures in the masonry and the foundation (Rule 3),
are obviously safe for this height of dam, but may be determined,
if desired, as explained in Example 2 of Art. 39.

Results of calculations, for all elevations, and with and with-

out tail-water, are indicated in Fig. 36. Above Level 18.8,
(within Zone la, Art. 30), tan 8 is greater than the allowed value;
and above Level 14.3 (within Zone I), the resultant, reservoir full,

lies outside the middle third. This condition necessitates careful

treatment of the upper part of the dam, as described in Example 3
of Art. 45.
It will be noted that the shape of the section is governed solely
by the shape of the sheet of falling water. In other words, the
whole dam lies above the elevation of the bottom of Zone II.
Should the calculations indicate a location of the resultant outside
the middle third, the dam should be widened, as indicated in
Example 3.

48. Comparison of Solid Spillway Dams. The fundamental

theory of design, for solid spillway and non-overflow dams, differs
only in the upper part, which, in the former, is proportioned to
conform to the shape of the sheet of water spilling over the top.
A comparison of this feature, therefore, would be of interest, and
is given in Fig. 38.
In order to make a
comparison with Example No. 5,
direct
the dimensions of each dam
were multiplied by a constant, thus
reducing each head on the crest to a value of 10 ft. This can
properly be done, as it was shown in Art. 42 that the co-ordinates
of the issuing sheet, and hence the curve of the top of the dam, are
direct functions of the head on the crest. As a matter of fact, the
Art. 48] SOLID SPILLWAY DAMS 131

stability ofany dam, subjected to head-water pressure only, will

not be changed if all dimensions, including the head on the crest,

are increased in the same proportion. This, of course, provided

the compressive stresses remain within the allowed working values.

Water Surface

COMPARISON OF SECTIONS
OF SOLID SPILLWAY DAMS
All reduced to a common head-on crest of 10 feet

Feet
- Author's Example No.5 of Art-17
- Keokuk Dam
of the Mississippi River Power Co.
- Dam No.2 of the Appalachian Power Co.

Parr Shoala Dam of the Parr Shoals Power Co. {Fig. 37)

— —— Ocoee No.l Dam of the Tennessee Power Co.

McCaU.Feny Dam of the Pennsylvania Water and Power Co.

Fig. 38.

The differences in the tops of the dams indicated in Fig. 38

may be attributed chiefly to.differences in the methods used in
determining the shape of the sheet of spilling water and the dis-
tance the masonry line was extended into the sheet to provide a
margin of safety.
 CHAPTER VIII

THE DESIGN OF HOLLOW DAMS

49. General Considerations. Hollow dams may be made in

an almost unlimited variety of forms; but, thus far, they have been
largely constructed entirely of reinforced concrete, of the but-
tressed type, having substantially triangular buttresses set par-
allel to one another at fixed intervals across the bed of a stream.
Two types of hollow dams have become common. Fig. 47
indicates a typical " Ambursen " hollow dam. The distinguishing
feature of this type is the flat reinforced concrete decks. Some-
times these decks are arched, as in Fig. 46. Such a small amount
arching, however, will not be effective while the reinforcement is

intact, as a shrinkage of the concrete due to setting, or a fall in

temperature, will result in a slight opening of the joints and destroy
all arch action until the reinforcement has failed. Therefore this
feature should be considered as limiting complete failure, and not a
partial failure, such as a cracking of the concrete and a rupture of
the steel.

Arch " hollow dam in

Fig. 43 indicates a typical " Multiple
which the decks consist of series of arches spanning between the
buttresses. The arches are often reinforced. For design of the
arches see Art. 57.
The up-stream edges of the buttresses have a slope of any
desired degree; but, in by far the larger number of cases, this
slope is about 45°. The slope of the down-stream edges of the
buttresses varies usually from zero to about 15° with the vertical
for non-overflow dams. For spillway dams, the shape of the top
and down-stream edge is usually fixed by the necessity of providing
a crest and apron designed to fit the shape of the sheet of water
spilling over the dam.*
The width and superelevation of the top is fixed by practical
considerations, mentioned in Art. 37; but, in the case of hollow
dams, a considerable superelevation or top width is not wholly or
* See Art. 42.
132
Art. 49] GENERAL CONSIDERATIONS 133

>

.4

03

-a
§
3
O

J3
*3

3
134 THE DESIGN OF HOLLOW DAMS [Chap. VIII

partly compensated by a reduction in masonry in the lower parts,

as in the case of solid non-overflow dams.
The apron of the spillway type is a reinforced concrete slab,
%'-'Cot. Bar^, lapping
Bare laced with No
6,{'° Cor. Bars,
(batted at Butt
C.L.
/-El. 1635

1:3:6
Struts and Deck 1:2:1
Buttress reinforcement 1 °Cor. Bars
two in each course, lapping 30." Three
inches between steel and face of buttress.
Column reinforcement (l°Cor. Bars)
made continuous by 24" lap at lift.

•% ° Cor. Bars, HI C. t

SECTION B-B

Fig. 40. — Details of Mathis Dike Dam. {Eng. News, Vol. LXXIV, p. 592.)

but, the loading being indeterminate, its thickness and reinforcing

are matters of judgment. It is essential to provide ample concrete
at the discharge lip, as indicated in Fig. 47, to withstand shocks
Art. 49] GENERAL CONSIDERATIONS 135

from floating ice and logs, and the pressure of the jet in changing
its direction of flow.
On soft foundations the buttresses may be provided with plain
or reinforced concrete spread footings, in order to keep the com-
pressive stresses within reasonable amounts. In extreme cases,
these footings may
have a width equal to the spacing of the but-
tresses, so that the dam virtually rests on a concrete mattress, as
indicated in Fig. 41. Such footings should be provided with large
weep-holes at close intervals, in order to preclude the possibility
of uplift from head-water.

Fig. 41. —Foundation Mattress, Mathis Dike Dam.

The buttresses should be braced with horizontal struts at inter-

vals, as indicated in the illustrations. The spacing of the struts
should not exceed twelve times the thickness of the buttresses
when the latter are stressed in compression to their full working
value. Additional stiffness is obtained by horizontal steel rein-
forcement between struts, as indicated.
The reinforcement in the struts is usually continuous through
at least three bays, but in some cases it has been carried contin-
uously throughout the structure, with no deleterious effects from
contraction. The last method was used for the Mathis Dike
Dam, Fig. 39, etc. The struts should abut solidly against the
136 THE DESIGN OF HOLLOW DAMS [Chap. VIII

should
buttresses. The horizontal building joints in the buttresses
be at the elevation of the center line of the struts, if possible.
The open holes through the buttresses, indicated in the illus-
trations, are found convenient for the passage of
men and materials
provided,
during construction. An inspection gallery is usually
low dams, access
unless, in may be had from the down-stream side
at ground level.
Provision for draining the interior of hollow spillway dams is
usually made, as shown in Fig. 47, where, for this purpose, an
opening of considerable size is indicated under the bucket. In

Fig. 42. — Down-stream View of Mathis Dike Dam.

such cases the high velocity of the water at the end of the bucket
will entrain the air and cause a partial vacuum to form within the
dam, unless sufficient air inlets are provided. If a partial vacuum
is allowed to occur, the loads on the deck and apron will be ma-
teriallyincreased, as explained in Art. 17. Large openings
through the buttresses and an adequate open shaft at each end
of the dam are usually provided. The mistake has sometimes
been made of providing solid doors at the entranc.es to the interior
of the dam.
In hollow dams of the usual type, the resultants for full or
empty reservoir always intersect the joints relatively close to the
Art. 49] GENERAL CONSIDERATIONS 137

f -j ^ F ~
 138 THE DESIGN OF HOLLOW DAMS [Chap. VIII

center of gravity. Rule 1, therefore, is seldom a governing con-

sideration in determining the shape of the structure.
prevent
As the weight of masonry alone is never sufficient to

shding (Rule 2), it is necessary to batter the up-stream face to

include the vertical component of considerable water pressure.
Limiting the compressive stresses to safe values (Rule 3)
necessitates an adjustment of the length and thickness of the
buttresses.
Art. 49] GENERAL CONSIDERATIONS 139

I
I
f
a
6
i
I

0!

Pi
o

I
1

I
a
"3

Ol
e
 ; ;

140 THE DESIGN OF HOLLOW DAMS [Chap. VIH

In hollow dams there is some question as to the effective area

of base to distribute the pressures on the foundation. It is com-
mon practice to neglect the horizontal area of the deck in computing
the compressive stresses. This procedure, however, is- not always
on the side of safety. The addition of the area of the deck to the
area of the base will always reduce the direct stress (Fig. 12), but
the eccentricity, e, may, in some cases, be increased a sufficient
amount to increase the fiexural stress to a greater extent, resulting
in an increased total stress at the extremity of the base.
The complete design of hollow dams cannot be made, step by
step, as in the case of solid dams. The general outlines of the
dam are usually chosen tentatively, in accordance with the judg-
ment of the designer, tested for conformity with the designing
rules, and adjusted if found necessary. The spacing of buttresses
is limited by the economical span of the deck. In general it will
be found that a considerably greater spacing can economically be
adopted for Multiple Arch Dams than for the Ambursen type.
For large dams, a spacing of from 30 to 40 ft. for the former type
and from 15 to 25 for the latter, are common.
In order to indicate more fully the several features entering
into the design, reference ismade to the examples which follow.
50. Example No. 6. Hollow Non-Overflow Dam (Fig. 46).
Assumptions:

w, = weight of masonry = 150 lb. per cu. ft.

w% = weight of water = 62.5 lb. per cu. ft.
p,' and p," = maximum allowed vertical compressive stress =
31,000 1b. per sq. ft.;

Pt and p" = maximum allowed inclined compressive stress =

50,0001b. per sq. ft.;
/= working value of the coefficient of friction of the
joints and base = 0.60.

The section having been laid out, as indicated in Fig. 46, in

accordance with the judgment of the designer, it remains to inves-
tigate it for conformity with the designing rules,
and make the
requisite alterations found necessary.
if The methods will be
described for the joint at Level 73.0 only. A length of dam of 18
ft. will be considered,
Abt. 50] EXAMPLE NO. 6 141

Rule 1. Location of the Resultant

The may be found from Eq. (18) of

location of the resultant
Art. 31, moments being taken about the point of reference, A,
at the up-stream extremity of the joint.

For empty reservoir,

2(Wx) E +2(Px) B
Zb
Z(W) e
For full reservoir,

Zf ~
_ X(Wx) F +2(Px) F
2(W) r

The necessary calculations are indicated in Table XXIII.

Making the proper substitutions in the foregoing equations,

there results:
i

For empty reservoir,

**~
_^'- ift
_ 91,550,000 _4 1U -
1,943,000
For full reservoir,
_ 235,450,000
F
4,938,000

These values are indicated in Fig. 46.

Rule 2. The Inclination of the Resultant

From Eq. (22) of Art. 32,

taneeF ZtfV,_ 2,916,000

tan 591,
-sa^"4,938,000-°-

This value is close enough to the maximum allowed value of tan

6 = 0.60, and;no adjustment is necessary. The resultant, reservoir
empty, is, of course, vertical. !

A change in tan 6F could best be obtained, if found necessary,

by altering the inclination of the up-stream face of the dam,
thereby changing most effectually the total force, 2(W) F on
,

account of (the additional water pressure. It is essential, for

economy, to have as steep an up-stream face as the allowed max-
j

imum value jpf tan 6F will permit, i

142 THE DESIGN OF HOLLOW DAMS [Chap. VIII
Art. 50] EXAMPLE NO. 6 143

In this connection, it should be noted that a much steeper up-

stream face is possible in the upper part of hollow dams; tan 6,
for this example, being very small at the upper levels. A few
recent hollow dams have been shaped in this manner.

Rule 3. Compressive Stresses

a. Reservoir Full. As the base of the dam is not rectangular,

the vertical compressive stresses must be found by the method
indicated in Art. 22 for irregular bases, and Eq. (14) and (15)
will apply, p u and pu " being zero
'
for hollow dams.

At the toe,

p/~S(W) r \\+
A '
I
At the heel,
1 em"
p„" = 2(WV \

A I

The following values for Level 73.0 are easily found:

A=313.1;
7 = 291,000;

e = 14.20;
m' = 62.71;

ra" = 33.29.

Using the value of 2(W)p from Table XXIII, and making the
proper substitutions in the foregoing equations, there results:

I4 ° 2 71
P/^4,938,000(^+ |9 i y o ) =30,870,

^"=4,938,000(3^-^^9)^,740.

The maximum vertical stress is seen to be at the toe of the dam,

where it equals, approximately, the maximum allowed value
adopted.
It will be found, upon investigation, that by properly reducing
the thickness and increasing the length of the buttresses, and at
 ;

144 THE DESIGN OF HOLLOW DAMS [Chap. VIII

the same time slightly reducing the area of the joint, the vertical
stress at the toe willremain the same and that at the heel will be
increased. is possible, owing to the resultant
This adjustment
decrease in the eccentricity, e, of the loading, with a corresponding
reduction in the flexural stress. On account of the decrease in
the area of the joint without an increase in the maximum vertical
stress, the resulting arrangement would be more economical.
However, the thickness of the buttresses is limited by practical
considerations, and, as it is considered that, in this example, they
are as thin as it is advisable to have them, a more economical
arrangement, in this respect is not possible:
For the maximum inclined compressive stresses in the dam,
Eq. (23a), of Art. 33, applies. Eq. (24a) has no practical use
unless the strength of the foundation is less than that of the
masonry.

At the toe,

tan<£' = 0.333; tanV = 0.111; sec V= 1-153.

For full reservoir,

tan 9 = 0.591; sec 2 6= 1.352.

pn is zero at the toe.

Using these values in Eq. (23a), the maximum inclined com-
pressive stress for full reservoir is found to be:

At the toe
r
jH = (30,870X1.153-0), orO, or 30,870X1.352,

p/ = 35,600 or 41,750,

pi = 41,750, the greatest of these values.

It is noted in Art. 33 that the greatest stress at the heel, for

this type of dam, is the reaction of the deck on the buttresses.
This will amount to,

p/
,
= ^5X72X18 =37300

The maximum inclined stress is seen to be within the adopted

limiting value of 50,000 lb. per sq. ft.
 Stress In Founds per Square Foot.
Note. —Stresses have been calculated only at the levels indicated by circles.

Unless indicated in the diagram, the horizontal area of the deck has been included in the
area of the joints.

Fig. 46.
To face page 145
Art. 50] EXAMPLE NO. 6 145

b. Reservoir Empty. As, for empty

reservoir, both the eccen-
tricity, e, and the weight, 2(W) E , than for full reservoir, it is
is less

evident that the stress, for this condition, will not govern.

The results of all these calculations are plotted in the diagram,

Fig. 46, together with corresponding results for the upper joints.
It will be noticed that the pressures grow rapidly less in the upper
joints. To be consistent with theory, they should be constant
at each joint. However, the thickness of the buttresses near the
top of the dam is usually fixed by practical considerations. At
the top of the dam a minimum thickness of 12 in. was considered
advisable. At Level 49.0 a minimum thickness of 19 in. was
required in order to keep within the allowed maximum vertical
pressure of 31,000 lb. per sq. ft. Between these elevations the
thickness of the buttresses was increased uniformly.
In order to indicate the difference in calculated stresses, with
and without the horizontal area of the deck included in the area
of the base, the vertical stresses, corresponding to the latter con-
dition, have also been calculated and plotted in the diagram. It is
seen that, at Level 73, to include the area of the deck, results in
an increase in the maximum vertical stress from 23,600 to 30,870
lb. per sq. ft., and a corresponding increase in the maximum
inclined stress. As mentioned in the last part of Art. 49, it is
probable that the area of the deck is not wholly effective in dis-
tributing its portion of the loads, on account of not being directly
bonded to the buttresses, and consequently the stresses may not
be as high as calculated. It is certain, however, that they exceed
those calculated without including the area of the deck, and it

seems best to adopt the more conservative method.

Rule 4. Inclination of Down-stream Face

The maximum allowed inclination of the down-stream face is

seldom a governing feature in the design of hollow dams. In the
first place, it is never necessary to provide a very flat face, and in

the second place, the usual horizontal reinforcement in the but-

tresses is in a position to resist possible tensile stresses in vertical

planes.
51. Example 7. Hollow Spillway Dam. Fig. 47 represents a
typical hollow spillway dam of the Ambursen type. In adopting
146 THE DESIGN OP HOLLOW DAMS [Chap. VIII

the general outlines, consideration should first be given to the

shape of the crest and apron, which can be established at once, as
described in Art. 42, Fig. 26 being used for the approximately
45° slope of deck. The radius of the bucket can be taken from
Fig. 32.
The Example 5 apply directly to
designing methods used in
this case, and not be repeated. The area of the horizontal
will
joints should not include that of the apron. This is a necessary
assumption; for, as indicated in Section B-B there is not a sub-
bond between the apron and the buttresses.
stantial The assump-
tion on the side of safety.
is

Assumptions:

Wi = weight of masonry = 150 lb. per cu. ft. ;

u)2 = weight of water = 62.5 lb. per cu. ft.;

pj and pv" = maximum allowed vertical compressive stress =

29,0001b. per cu. ft.;

Pi and p " = maximum allowed

t inclined compressive stress =
40,0001b. per cu. ft.;

/= working value of the coefficient of friction = 0.60.

The necessary calculations for the determination of the loca-

tion of the resultants, at Level 64.0, are indicated in Table XXIIIa.
The results of all calculations are plotted in the diagram of
Fig. 47. be noted that the stresses vary considerably at
It will
the different levels, and approach the maximum allowed values
only at Level 64.0. This could be remedied only by decreasing
the thickness of the buttresses. above Level 64.0. However, they
are as thin as it is desirable to have them, so that no adjustment,
in this respect, is possible.
Some economy would result if the slope of the up-stream face
were steepened slightly, as it is seen that the maximum value of
tan is less than the allowed value of 0.60.
 EXAMPLE NO. 7 OF ART. 51
TYPICAL
REINFORCED CONCRETE
HOLLOW SPILLWAY DAM
Tangent © .
:
;° :

i'>
;
::-;.<>;

Stress in Pounds per Square Foot

Note. — Unless indicated in the diagram, the horizontal area of the deck has been included in the
area of the joints.
Stresses have been calculated only at the levels indicated by circles.

Fig. 47.
 Htgh Water LeVel 0.0

EXAMPLE NO. 7 OF ART. 51

TYPICAL
REINFORCED CONCRETE
HOLLOW SPILLWAY DAM
Levell2.0 I

* .*. » '
* - » . ; » » I* ,"» =
-"

k^Wi/^=-i ;
:^illkv:V:;W-;x«:;v:V.'.'o;::i;'v'.;: •:.?;;.•.••<(•::•..•:•.•

Level 16.0

Fig. 47. To /ace jwj/e 1JB

Art. 51] EXAMPLE NO. 7 147

.a
3
P.
-^>
1-3 3
O
03

a
O
 CHAPTER IX
ARCH DAMS

52. General Considerations. Arched dams may be divided

into two general classes:

a. Curved gravity dams, or those designed as gravity dams, but

curved in plan so as to make use of arch action as an
additional margin of safety. Such dams may be de-
signed in accordance with the theory hereinbefore given,
and need no further makes
discussion, as the arch action
for a reduction in all stresses, and provides a more
favorable location of the resultant, reservoir full.

b. Arch dams, or those for which the arch action provides

most or practically all of the resistance to the forces
tending to move the structure. This chapter will be
limited to a discussion of this type.

In the discussion of arch stresses, the assumption will first be

made that all the load is taken by arch action, the deflection of the
dam being unrestrained by the shearing and frictional resistance
at the foundation. The modification of the arch stresses due to
restraint at the base will then be discussed.
The following nomenclature will be used for arch dams, all

distance being in feet and unit pressures in pounds per square

foot:

m = the deflection of the unrestrained arch at any elevation;

mi = the deflection of the restrained arch at any elevation;
Mi = the deflection at the top of the restrained arch;
= the thickness of the arch at any elevation;
i

/i = the height of the dam above any elevation;

H = the total height of the dam;

r„ = the radius of the arch at any elevation measured to the up-
stream face;
rm = the mean radius of the arch at any elevation;
148
Abt. 52] GENERAL CONSIDERATIONS 149

g=unit load taken by the unrestrained arch at any elevation:

gi = the unit load taken by the restrained arch at any elevation;
p=the stress in the unrestrained arch at any elevation;
pi =the stress in the restrained arch at any elevation;
tC2=the unit weight of water;
f>
C=—— =all constants;
C2
k, k', k " = constants;
A = horizontal area of an arch dam, between abutments, at any
elevation;
5= central angle between abutments (Fig. 52);
s=span of arch at any elevation;
x= 3.1416.
53. Arch Stresses.
In order to indicate the approximate
an arch dam, we will first consider the
distribution of stresses in
simple case of a dam of constant radius, confined between vertical
abutments, and having a horizontal base. It will be assumed
that:

o. The arch radius is very long, compared with the thickness

of the dam;
6. The dam is homogeneous throughout;

c The thickness at any elevation is proportional to the depth

of water, resulting in a triangular section with its crest
at water surface;
d. The arch stresses are not influenced by temperature
changes and other indeterminate considerations to be
discussed later.

The shaded area, 1—2-3, in Fig. 49, represents a section of the

unloaded dam at the crown of the arch. The unit water loads are
represented by abscissae to the line, 1-7.

Considering the deflection of the loaded dam to be unrestrained

at the base, the ordinary equation for the arch stress at any eleva-
tion is

P=^- (33)

This equation is not correct when the arch radius is small,

 ARCH DAMS [Chap. IX
150

close enough for all

compared with the arch thickness; but it is

practical purposes.

Fig. 48. —Arch Dam of the Shoshone Project, Wyoming.

As j and r„ are assumed to be constant for this case, p will also
be constant, and will be represented by abscissae to the line, 8-9.
 —

Abt. 53] ARCH STRESSES 151

The R. Shirreffs has shown * that, for constant radius and

late
span, the crown deflection of an unrestrained arch may be rep-
resented by,
Cjq
m=- '
(34)
t
_=2ra j
LOADING DIAGRAM ARCH STRESS DIAGRAM
DEFLECTION
DIAGRAM

l
/Arch
in restrained
dam with,
elastic beam
of apprecj
able weight

Arch ptress in
restrained dam
with inelastic
vertical beam
_ofi.no weigj

V
Arch etr
Ti- unre trained

ll

0.2 0.4 0.6 0.8

Unit Arch Loading, in
Terms otWiB Arch Stress, In Terms of p
w2 H
Fig. 49.

where G is a constant depending on the radius, span, and other

characteristics of the arch-

By assumption,
t = C2 h, (35)
and,
q=W2h,
* In his discussion of a paper on Lake Cheesman Dam by Harrison and
Woodard, Transactions, Am. Soc. C. E., Vol. LIII, p. 155.
152 ARCH DAMS [Chap. IX

therefore,
C^V2h =
m=
t
W\ i— Cw2 (36)
C2h
Therefore m at any elevation is constant, and the dam under load
will assume the position, 4—5-6.
A restraint at the base, however, on account of vertical beam
action, will result in a redistribution of the loading taken by the
arch. we assume the vertical beam to be incapable of deforma-
If
tion, ofno weight, and free to rotate about the base, the load car-
ried by the arch may be represented (as proved later), by the
curve, 1-10-11-3, and the arch stress by the line, 13-14. The
deflection of the dam will be as indicated by the lines, 15-2-3,

the section still being a true triangle on account of the inelasticity

of the vertical beam. The area between the vertical, 1-3, and the
curve, 1-10-11-3, represents the total load now being carried by
arch action, and is less than the total water load, 1—7-3-1, the

remainder being taken directly to the foundation by shearing and

frictional resistance at the base. All of this can be proved, as
follows:
Shirreffs' equation applied to the restrained arch is,

m!=^i (34a)

Substituting the value of t from Eq. (35), there results,

m _Ciqi_Cqi

For static equilibrium, the moment, about the base, of the

loading carried by restrained arch action must be equal to the
corresponding moment of the total water load.
Therefore,
B
C qi (H-h)dh=^ (38)'

From Eq. (37),

^ = -q- (39)

And since, by assumption, the up-stream face of the dam is still

straight,

m *= (40)
-jf
Art. 53] ARCH STRESSES 153

Substituting these values of qi and mi in Eq. (38), there results,

" Mj(H~h) h 2
h~
_ W2IP
I EC 6 •

Integrating between the limits and H, and solving for M 1 we

have, as the top deflection of the restrained dam,

M 1 =2Cw 2 (41)

Comparing the top deflection of the restrained arch, as given

by Eq. (41), with that of the unrestrained arch, from Eq. (36),
it is seen that the former is just twice the latter under the assump-
tion of no deformation of the vertical beam.
To derive the equation of the loading curve, 1-10-11-3, we
have, from Eq. (39), (40) and (41).

mih

MiiH-h)L K
mi= -jj ,

M 1 =2Cw2 ,

Therefore,

qi = 2MH-W > (42)

which is the equation of the curve, 1-10-11-3.

The equation of the stress line, 13-14, may be derived as fol-

lows: At the base of the dam we have, from Eqs. (33), (35) and
(42), for the restrained arch,

Vl=
qiru
T= qira
C2h —Wc —
= 2(H-h)w2 ru
2
.

(43)

Also, at the base of the unrestrained dam, Eq. (33) gives,

_ W2Hru _ W2Hru
P t CW
or,
„ w2 ru
V
 154 ARCH DAMS [Chap. IX

Substituting this value of C2 in Eq. (43), there results,

Pi-^5^. W
which is the equation of the stress line, 13-14.
be noted that, for the restrained dam, the arch stress, pi,
It will
is equal to the constant unrestrained arch stress, p, when h is

one-half of H. In the upper half of the dam the arch stress for
the restrained dam is greater than that for the unrestrained dam,
reaching a maximum of twice the latter value at the extreme top.
In the lower half, the restrained arch stresses are less, reaching a
minimum value at the base, where the arch deflection is zero.

Thus far we have been limited by_ assumptions, some of which

are impossible and others of which are unusual. In order to
arrive at a closer indication of the arch stresses, we must investi-
gate the effect of:

a. Elasticity of the vertical beam;

b. The weight of the dam;
c. Uplift water pressure;
d. Varying span;
e. Varying radius;
/. Possible variation of the modulus of elasticity at different
elevations;
g. Expansion and contraction due to temperature changes;
h. Expansion and contraction due to changes in moisture con-
tent;
i. Contraction of the concrete due to setting of the cement.

On account of the indeterminate effect of most of these items,

it is the opinion of the author that
it is impossible to write exact

equations indicating the arch and vertical beam stresses in arch

dams, although many attempts have been made.*
Fortunately, a number of dams have been constructed which,
being virtually experiments on a large scale, serve to indicate
* Transactions, Am. Soc. C. E., Vol. LIII,
p. 155, paper by Harrison and
Woodard, and discussion; Transactions, Tech. Soc. Pacific Coast, Vol. VI,
p. 75, paper by Vischer and Wagoner; Transactions, Am. Soc. C. E., Vol.
LXXVIII, p. 685, paper by L. R. Jorgensen; Engineering News, June
9,
1910, article by J. S. Eastwood; "Masonry Dam Design," by Morrison and
Brodie. 2d edition. John Wiley & Sons. 1916.
 TABLE XXIV—TABULATION OF ARCH DAMS
Arranged in Order of Height

a
 TABLE XXIV—TABULATION OF ARCH DAMS
Arranged in Order of Height

co 'S3

he
Art. 53] ARCH STRESSES 155

what may be expected from structures of this type. As far as

the author aware, there are no records of the failure of an arch
is

dam. This, though fortunate from an economic point of view, is

lamentable, as there has been furnished no direct indication of the
limitations of such structures. Of the existing dams, a few have
been designed in accordance with intricate although in the —
—
author's opinion questionable theory; but by far the greater
number have been treated simply as sections of rigid cylinders
subjected to external water pressure, i.e., thickness at any eleva-
tion determined from Eq. (33), without consideration being given
to possible restraint at the base.
Table XXIV gives the general characteristics of a number of
representative arch dams. The arch stresses indicated in Column 9
were calculated from Eq. (33). In order to indicate the relative
stability of future designs, in comparison with existing structures,
there is given the following brief discussion of the general effect,

on stability, of the items listed above.

Under the original assumptions, we have seen that, for an
unrestrained arch dam, the arch stress may be represented by the
line 8-9 (Fig. 49), and that, when the dam is restrained and the
vertical beam is inelastic and weightless, the arch stress changes
to 13-12-14.
The effect of the elasticity of the vertical beam is to decrease
the deflection at the top of the dam and increase the deflection at
lower elevations. The effect of the weight of the structure and a
resistance to rotation at the base is to increase the proportion of
the load carried directly to the foundation and to reduce the arch
stresses at all elevations. The arch stress curve under these
combined conditions will have the general shape indicated by the
line,18-16-14, the exact amount of stress at the different eleva-
tions being dependent on the relative influence of the imposed
conditions.
The approximate loading diagram and the deflection of the dam
corresponding to the arch stress curve, 18-16-14, are indicated by
the lines, 1-17-3 and 19-2-3, respectively.
The effect of the uplift pressure due to head-water on hori-
zontal joints tends only to lessen the effective weight of the dam
and consequently the influence of the weight on the distribution
of arch stresses. Uplift may be neglected in structures of this
type, as it has practically no influence on stability.
156 ARCH DAMS [Chap. IX

Other things being equal, the smaller the span or the radius,
the greater will be the load taken by an arch for a given deflection.
Most sites are V-shaped, resulting in relatively shorter spans at
the lower elevations; moreover, as will be shown later, the great-
est economy of material results for V-shaped sites when the arch
radius gradually decreases toward the base. The effect of these
two conditions on the distribution of stresses is similar, although
more pronounced in the latter.
A shortening of either the span or radius in the lower part of
the dam results in a less deflection of the vertical beam and a
greater arch stress. Consequently, the stress line, 16-14, in the
lower part of the dam will approach the line, 12-9. For the usual
shape of site, the reduction of deflection in the lower part of the
dam causes also reduced deflections in the upper part, which part,
still having the original radius, suffers a decrease in arch stress, or
a movement of the stress line, 18-16, toward the line, 8-12. The
net result, therefore, of a shortening of either the radius or span,
in the lower elevations, is a reduction in the effect of vertical beam
action at all elevations.

Thus the conditions considered permit of approximate

far,

mathematical determination of the arch stresses, distribution of

arch and vertical beam loading, and the deflection of the dam.
Such calculations, however, are long and intricate; but, as
mentioned heretofore, they have been made, and the results
used in the design of arch dams.
A theoretical dam, designed as a pure arch, using Eq. (33), a
constant arch stress, 8-9, and neglecting the effect of vertical
beam action, will, if checked by such rigorous investigations, be
found to have reduced arch stresses in the lower part and increased
upper part, as indicated roughly by the curve, 18-
stresses in the
16-14. However, the increased stresses in the upper part will
approach, but not exceed, double the stress assumed, and are
amply compensated for by the factor of safety embodied in the
usual working stresses. It is probable, indeed, that, for thin arch
dams, such as would result from the use of the customary working
stresses in Eq. (33), the vertical beam
action will not greatly
affect the stresses near the top.
Moreover, a practical design
embodies an appreciable thickness at the water surface, resulting
in a considerable reduction in the stresses in the upper part of the
dam.
Abt. 54] VERTICAL BEAM STRESSES 157

Rigorous, intricate investigations, for the determination of the

arch stresses due to a combination of arch and vertical beam
action, therefore, are hardly justified, particularly in consideration
of the sometimes compensating and sometimes cumulative effect
of stresses due to temperature changes', changes in moisture con-
tent, and other indeterminate factors which are sufficient to nullify
practically the basic assumptions in such investigations. A
glance at the deflection curves of the Barren Jack Creek Dam
(Fig. 60), will indicate at once the utter futihty of endeavoring to
arrive at an exact determination of the effect of vertical beam
action.
Another condition which affects materially the results to be
obtained from rigorous investigations is the probability of varia-
tion of the modulus of elasticity of the masonry at different ele-
vations. It is hardly reasonable to assume that, with the vary-
ing materials which must be used in different parts of a concrete
structure, the modulus would have a great degree of uniformity,
particularly when plums are used only in the lower and thicker
part of the dam.
A considerable reduction in temperature results in an opening
of vertical building joints which must be closed by deflection under
water pressure before arch action takes place. This has the
effect of increasing the load carried by the vertical beam and
reducing the load carried by the arch. An increase in tempera-
ture will have the reverse effect.
Swelling of the concrete, due to saturation as the water rises
in the reservoir, will have the same effect on the distribution of
the loading as a rise in temperature. It is probable that such
swelling is nearly compensated by the previous shrinkage of ihe
concrete due to the setting of the cement.
54. Vertical Beam Stresses. As pointed out heretofore, a
close determination of the stresses in the vertical beam is diffi-

cult if not impossible if all influencing conditions are considered.

Fortunately, such stresses are not important, except in very high
dams.
Near the base, the effect of the water load is to increase the
vertical compressive stresses near the down-stream face. Con-
crete, in common with other materials, expands laterally when
subjected to vertical compression. Being confined between the
walls of the canyon, the direct result is an initial negative arch
158 ARCH DAMS [Chap. IX

deflection in an up-stream direction which must be eliminated by-

water pressure before the vertical beam can deflect down stream
and the vertical compression be increased at the toe. Principally
to this is attributed the fact that in no existing arch dams has the
deflection of the vertical beam resulted in a rupture of the masonry
due to excessive vertical compressive stresses, although several
arch dams more than 200 ft. in height have been constructed.
The question of horizontal shear in the vertical beam, though
equally indeterminate, has no influence on the ultimate strength
of the structure, as a failure in this respect, though objectionable
as affecting leakage, would result in a better distribution of the
arch loading.
55. Recommendations for Design. It has been mentioned
that the influence of vertical beam action is to increase the arch
stresses in the top of the dam, but that the ratio of the resultant
stress to the arch stress, computed from Eq. (33), cannot theo-
retically exceed two. In view of the fact that the upper part of
the dam, which theoretically can reach zero thickness at the
water surface, is for practical reasons considerably in excess of the
computed thickness, such increase in the calculated stress is inap-
preciable in comparison with the factor of safety embodied in the
usual working compressive arch stresses.
recommended, therefore, that the influence of beam action
It is
be neglected, and that the thickness of the dam be determined
from Eq. (33). The effect of temperature changes in combination
with other influences is indeterminate, and, therefore, must be
also compensated for by the factor of safety in the unit stresses.
A discussion of working stresses is given in Art. 58. Fig. 51,
plotted from Table XXIV, indicates the relation between maxi-
mum arch stresses and the height of existing structures. The
arch stresses were calculated from Eq. (33). It is significant that
the arch stresses are much lower for the higher dams. This may
be attributed to the adoption of a larger margin of safety in the
higher structures, and also to the desire for an excess of thickness,
near the base, on account of the greater vertical pressures.
On account of the uncertainty as to the actual stresses in arch
dams, the design should be carefully compared with similar exist-
ing structures, and due allowance made for any variation in shape
and local conditions which may make for increased stresses. Until
our knowledge of the subject has considerably increased, arch
Akt. 55] RECOMMENDATIONS FOR DESIGN 159

dams differing materially from those already built must be con-

sidered in the nature of experiments.
Although dams of the arch type have proved their reliability,
on account of the great number now in successful use, without a

300
160 ARCH DAMS [Chap. IX

bring increased confidence in this perfectly safe and most econom-

ical type of masonry dams.
The outlines of most of the existing arch dams
56. Details.
have not been proportioned according to any standard practice,
thus resulting in many types. Most of them have a vertical up-
stream face and a constant radius. Mr. Lars R. Jorgensen *
has recently shown that, theoretically, the greatest economy of
material in an arch is obtained when it is subtended by a central
angle, 6, between abutments, of 133° 34' (see Fig. 52.)

Fig. 52.

This can be proved as follows:

At any elevation, the area, A, of any arch slice between abut-
ments is,

— q«o°
'"«'"}

From Eq. (33),

t =
V V
From which,
qrm
t =
v-

* Transactions, Am. Soc. C. E., Vol. LXXVIII, p. 685.

Akt. 56] DETAIIS 161

As p is the constant working stress, and q is a constant at any

given elevation,
t = k"ra .

Also,

rm = -, and s is a constant.
2sm
2

Substituting these values of r m and t in the first equation, there

results,

sin-^

Differentiating this expression with respect to 5, and equating

the differential coefficient to zero, there results for a minimum
value of A,
5 = 133° 34'.

If the cost ofform work and similar items were included in the
derivation of the most economical central angle, its value would

probably not exceed 120 degrees. As a matter of fact, the pecu-

liar configuaration of the site usually limits the value of the maxi-
mum central angle which can be adopted economically, it often
being necessary to follow certain well-defined ridges in order to
increase the average elevation of the foundations. In some in-
stances, where the canyon narrows rapidly toward the bottom,
a strict application of a constant angle would result in less thick-
ness near the bottom of the dam than at higher elevations, which is
impracticable. Consequently, in most cases, the economical
proportions can be found only by trial, adopting successively
different values of central angles at various elevations; always
being guided by the fact that, unless affected by peculiar condi-
tions, a value of about 120° will be most economical. This
method will prove tedious, but will be justified in view of the saving
in cost which may be obtained.
It will be noted that, for a constant central angle in a V-
shaped canyon, the arch radius must decrease toward the bottom
of the dam.
It has been shown already that, for spillway dams having large
flood discharges per linear foot of crest, it is not good practice to
162 ARCH DAMS [Chap. IX

allow the falling sheet of water to leave the face of the masonry.
For such cases the constant central angle theory will not apply,
as it usually results in a practically vertical down-stream face
(Fig. 56). For large discharges the down-stream face of an arch
dam may be shaped to fit the sheet of falling water for its entire
height; then, providing a vertical up-stream face, a varying radius
may be adopted to keep the arch stress within the limits desired.
In order to reduce the throw of the
sheet of water relative to the up-stream
face of the dam, the lip of the crest may
be made to overhang, as indicated in
Fig. 53.
The thickness of the top of the dam,
of course, should be proportioned to resist
ice pressure, if assumed to exist. The
ice pressure may
be considered as taken
Fig. 53. by a portion of the arch equal in height to
at least twice the thickness of the arch at
the elevation of the ice thrust plus the thickness of the ice, and
usually considerably more, depending on existing conditions, the
quantity of vertical reinforcement, if used, and the location of
the nearest horizontal building joint. The theory is identical
with that of the distribution of floor-slab concentrations. It
should be noted that, in very few of the dams listed in Table
XXIV, was it necessary to provide for ice thrust.
Steel reinforcement has been used to some extent in arch dams,
but in most cases has been insufficient in quantity to affect materi-
ally the stiffness or the distribution of stresses. The use of con-
tinuous horizontal reinforcement is not to be recommended, as it

is impractical to anchor the rods thoroughly into the rock of the

abutments, and there would be a tendency to draw the dam away
from the abutments due to temperature changes when the reser-
voir is empty. In this respect, the conditions in an arch dam are
somewhat different from those in a reinforced arch bridge, as the
latter is never without load. Horizontal reinforcement, if used,
should not extend through the vertical building joints of the
structure, such joints being placed at intervals in radial planes to
localize cracks and prevent their formation at too great an incli-
nation to the line of arch thrust.
Vertical reinforcement may be useful to distribute concen-
Art. 56] DETAILS 163

trated loads, such as ice thrust, and will serve also to stiffen the
relatively thin upper part of the dam if the length is considerable
in comparison with the thickness.
An arch dam is in reality a long column receiving lateral sup-
port only through its connection with the base. It will be noticed
that the ratio of curved length to thickness (" ratio of slenderness")
at the top of most of the existing dams, is greater than usually
allowed in long concrete columns. This is justified on account
of the lateral support which the top of the dam receives from the
relatively thicker lower portions, and the fact that the arch

Fig. 54.

stress in the upper part of the dam is much less than that adopted
for other parts of the structure.
It is good practice to provide a ratio of slenderness, at mid-
height, not greater than 25 and, at the top of the dam, a ratio not
greater than 75. The ratio at the top may be somewhat in-
creased if the thickness increases rapidly toward the lower ele-
vations, and the ratio at mid-height is proportionally reduced.
This is particularly true if considerable vertical reinforcement is

used. In the Salmon Creek Dam (Fig. 56), the ratio at the top
is more than 100, but this fact is compensated for by ample
164 ARCH DAMS [Chap. IX

thickness at lower elevations, which, at mid-height is less

than 10.
In order to prevent sliding, the rock surface at the abutments
should be excavated to the proper inclination to the line of thrust.
In Fig. 54, let W
represent the total weight of masonry and other
vertical forces between the vertical, radial planes, 5-1 and 4r-3.
Let P represent the total thrust of the arch slice between the two
horizontal planes passing through the points, 1 and 3. The rock
should be excavated to such lines that the resultant, R, of the
forces, Wand P, will have a maximum inclination, 0, with a
normal to the finished rock surface not greater than the angle of
repose of masonry on rock. The angles, 0' and 0", represent
the vertical and horizontal projections, respectively, of the angle,
6. The principle is identical with that of Art. 25 providing for
the resistance of gravity dams to sliding.
The number of steps and the depth
of excavation to be pro-
vided will depend on the character of the rock, and particularly
on the inclination of the stratifications. Probable shearing of the
rock on a direct line between points, 2 and 6, should be guarded
against.
57. Multiple Arch Dams. The theory of arch dams is applica-
ble only in a general way to the arched decks of hollow multiple

Fig. 55.

T
arch dams, described in Art. 49. n this type the loading on any
arch slice is not uniform, and Eq. does not apply. In Fig. 55,
(, 53)
it is seen that, for the arch slice, 1 -2-3-1, the unit water pressure

at the haunch is greater than at the crown. The percentage in

variation of the loading gradually decreases, of course, toward
the base of the dam; but, in the upper part of the structure, the
non-uniformity of the loading is pronounced.
 168-FT.DAM'
FOR
SALMON CREEK, ALASKA
ALASKA GASTINEAU MINING CO.
JUNEAU, ALASKA

Transactions Am. Soe. C.E., Vol. LXXVIII, Paper No. 1322. To face page 165
Abt. 59] EXAMPLES OF ARCH DAMS 165

Therefore the arches in the upper part of the dam should be

circular in horizontal planes, or well reinforced. In other words,
the decks of such dams should be designed as arches subjected
to non-uniform loads, the theory of which may be found in
many treatises on the subject, such as " Principles of Reinforced
Concrete Construction " by Turneaure and Maurer.
58. Allowed Stresses. For the larger and thicker arch dams,
1:3:6 cyclopean concrete has been commonly used. In the
thinner structures the mixture has been 1 2| 5, and even : :

1 2| 4§, and in the very thin arch decks of hollow gravity dams
: :

1:2:4 concrete, without plums, is customary.

The stresses in a number of existing arch dams have been cal-
culated from Eq. (33), and are indicated in Table XXIV and Fig.
51. It is seen that, in four cases, a stress of 56,000 lb. per sq. ft.
is exceeded, and in two cases the stresses amount to about 120,000
lb. per sq. ft.

These excessive stresses indicate, in a limited way, the factor

of safetyembodied in the more conservative designs; but they are
not in any sense indicative of the stress which should be adopted,
for, though the dams have not failed, there is no indication of how
close to their ultimate strength they are stressed.
Data on the ultimate strength of masonry have been given in
Art. 26. Although the total water load to be carried can be de-
termined quite accurately, there is much uncertainty in the
determination of the induced stresses in the dam, due to such
loading in combination with temperature changes and other inde-
terminate factors hereinbefore described. For this reason the
stress,calculated from Eq. (33), should not exceed from one-
eighth to one-twelfth of the ultimate strength of the masonry,
depending on the height and importance of the structure, and the
probability of destruction of other property and human life in case
of failure.
In very high dams the arch stresses near the base are reduced
considerably in order to provide additional width to distribute the
weight of the structure.
59. Examples of Arch Dams. Fig. 56 shows details of the
arch dam of the Alaska Gastineau Mining Co., on Salmon Creek,
Alaska. This dam indicates in a general way the application of
the constant central angle theory discussed in Art. 56, the radius
gradually diminishing toward the base. To have kept the central
 166 ARCH DAMS [Chap. IX

angle constant at all elevations would have resulted in the struc-

ture overhanging in places. The 10-ft. triangular piece at the
toe was added, in order to distribute more effectually the vertical
loads due to the great height of the structure.
Fig. 57 indicates the conservative design of the North Crow
Dam for the water-supply system of Cheyenne, Wyo. The
computed arch stress is only 23,000 lb. per sq. ft., and the ratio of
slenderness only 11 and 35, at mid-height and top, respectively.
is

A 1 : 2j : 4^ concrete mixture was used throughout the dam,

and steel reinforcement was placed in the upper 40 ft., as indi-

CROSS-SECTION
OF DAM CROSS-SECTION
AS CONSTRUCTED OF DAM
AS COMPUTED

Fig. 57.— North Crow Arch Dam. (Eng. Record, Vol. LXVII, p. 149.)

cated. A spillway was provided through a channel excavated in

the rock at one end of the dam.
Fig. 59 is a typicalexample of thirteen arch dams built by the
Public Works Department, New South Wales. These dams are
all described in Table XXIV, but they do not include
the Barren
Jack Creek Dam. In all these dams the concrete was mixed in
the following proportions:

4J parts of Portland cement,

11§ parts of sand,
10 parts of " shivers " of f-in. to |-in. gauge,
13 parts of " metal " of l|-in. gauge.
Art. 59] EXAMPLES OF ARCH DAMS 167

O
168 ARCH DAMS [Chap. IX

Except in Lithgow No. 2 Dam, the concrete was cyclopean, con-

taining a maximum number of two-man plums. Many of these

PLAN

f
T.W.L. 8.6

SECTION

Fig. 59. —Wollongong Arch Dam, New South Wales. (Eng. News, Vol.
LXIII, pp. 588-589.)

dams are noted for their slenderness, which greatly exceeds that of
usual American practice.
Art. 59] EXAMPLES OF ARCH DAMS 169

In Fig. 60 indicated the result of measurements of the

is

Barren Jack Creek Dam under variations in pond

deflection of the
level and temperature of the air. It is probable that the temper-
ature of the water has more influence on deflection than that of

Movement,!!! Inches.
lard rf CraUO*.05*j'.06*.10'.16!2o:2&*.80:85:4ol45*.5or55!oO*

Lines Show Movements of Dam.

PLAN

Fig. 60. —Barren Jack Creek Dam, Australia. (Eng. Bee, Vol. LXI, p. 664.)

the air, and it is unfortunate that it was not also recorded. The
deflection is seen to be much less than would be expected for a
structure of this type.
Fig. 61 shows a plan and section of the arch dam of the Agua
Pura Co. of Los Vegas, N. Mex. The dam is quite rigid, the ratio
170 ARCH DAMS [Chap. IX

1
*f
EU'JO
7— -
Concrete

Bubble ,

Sections '

ate'f Cut-ofl
, to Sound
^Impervious
Rock
ELEVATION

Fig. 61. —Arch Dam at Las Vegas, New Mexico. (Eng. News, Vol. LXIV,
p. 446.)
Art. 59] EXAMPLES OF ARCH DAMS 171

of slenderness at mid-height and top being only 17 and 52, respec-

tively. In this dam the radius might have been reduced consid-
erably with a reduction in quantity of masonry for the same arch
stress, the central angle being, for the proposed extension, only
88° at the top of the dam and much less at lower elevations. For
the existing structure, the central angle at the top is only 48°
 CHAPTER X
PREPARATION AND PROTECTION OF THE FOUNDATION

60. General Considerations. A good foundation is of ample

strength to withstand the weight of the structure, sufficiently
rough to provide ample margin against a sliding failure, tight
enough to prevent excessive leakage and uplift, and as clean as
possible, if rock, to insure the maximum effectiveness of the bond *
between it and the dam.
Absolute tightness is difficult, if not impossible, to obtain. It
is possible, with first-class rock, to provide against any appreciable
leakage or uplift; but, in the design of dams, the engineer has to
contend with foundations varying from solid rock, through all

grades of stratified and ruptured rock, to alluvial deposits of a very

porous nature.
Considerable preparation is always necessary in order to pro-
vide the requisites of a good foundation. It is probable that more
than 90 per cent of all failures of masonry dams has been caused by
faulty foundations. It is of the utmost importance, therefore,
that this feature of the design should receive the proper amount of
attention. It is unfortunate that the designer is not always the
builder, as many of the assumptions used in the design will depend
on the extent and character of the treatment which the founda-
tion receives. In all cases the designer should prepare the speci-
fications for the construction of the dam, and preferably have
supervision over the work.
61. Rock Foundations. It is not within the province of
this book to enter into a discussion of construction problems in
connection with the proper preparation of rock foundations. For
this, the reader is referred to C. W. Smith's " Construction of

Masonry Dams "f and works of like character. There is pointed

out, however, in addition to the necessary designing features, the

* Bond between rock foundations and the dam, although

neglected on
account of its unreliability, always exists to a certain extent, and undoubtedly
adds considerably to the stability of the structure.
t McGraw-Hill Book Co., Inc., 1915.

172
Art. 61] ROCK FOUNDATIONS 173

results which should be obtained from such preparation, in order

to justify the adoption of the designing assumptions herein rec-
ommended.
Surface rock is usually badly weathered, and unsuitable for
the support of a dam. It is sometimes necessary to excavate to
considerable depths before rock of an acceptable nature is uncov-

ered.
In the excavation of rock foundations, it is always necessary
to take particular care in order that good rock directly beneath
the blasting charges is not unnecessarily shattered. It is often
specified that the last foot or two of the excavation shall be barred
and wedged loose. The proper method will suggest itself to the
experienced builder when it is borne in mind that no part of the
final foundation should be disturbed from its original position and
that no stratifications should be jarred loose.
It has been pointed out heretofore that there should be as
much resistance to sliding below the surface of the foundation as
at other planes. If, therefore, the foundation contains loose
horizontal or nearly horizontal stratifications on which there is

danger of sliding, the excavation should be deep enough to obtain

a " toe hold " on the rock below the dam, in order that some, if
not all, of the horizontal loading may be carried to the rock by
direct compression through the vertical plane at the toe between
the dam and the rock.
In order that the masonry of the dam shall have the maximum
possible adhesion to the foundation, it is necessary that the final
rock surface should be absolutely clean, and the presence of flowing
water should be rigidly guarded against. There is nothing better
for cleaning rock surfaces than jets of clear water, under consid-
erable pressure, and the thorough use of stiff wire brooms. The
adaptable to cleaning out vertical seams and
jet is particularly
pot-holes of considerable depth. Such seams should be well
plugged with mortar. For concrete dams it is often specified that
the finished foundation shall be covered with a thin coat of rich
mortar immediately before the concrete is poured. Springs in the
foundation are not usually plugged before the masonry is started,
the practice being to allow them to discharge through pipes until
a sufficient mass of masonry has been placed to balance any pos-
sible uplift from them. The springs are then grouted under
pressure.
174 PROTECTION OF THE FOUNDATION [Chap. X

It is essential that the plane of contact between the founda-

tion and the masonry should be as rough as possible, in order to
assist in the resistance of the dam
In the case of
to sliding.
always possible
clearly defined horizontal stratifications, it is 'not
to obtain a rough surface, and a toe hold on the rock below the
dam must be provided, for reasons explained above.
Some leakage or seepage through the foundation is to be ex-
pected. Aside from the sometimes serious waste of water, leakage
is objectionable because it provides a possible means of uplift

and, in certain classes of foundations, scour. Foundations, in

general, have a tendency to tighten gradually, particularly if silt
is present in the water. The exception is in certain classes of
limestone foundations, where the water has a solvent action on
the rock, and where there is not sufficient silt in the water to plug
the leaks thus formed. It must not be supposed that, because the
stream in its original state carried large quantities of silt past the
dam site, this condition will obtain after the dam is constructed.
In large storage reservoirs not reach the dam, even during
silt will

the greatest floods, until perhaps after a great number of years,

when the whole volume of the reservoir has become silted up and
its usefulness destroyed.
In order to confine the leakage to a reasonable quantity, it is

necessary, with poor foundations, to provide a cut-off or artificial

impervious barrier under the heel of the dam. For rock founda-
tions there aretwo general types of cut-offs; first, a trench filled
with concrete, and second, holes drilled at frequent intervals and
grouted under pressure.
The first type is much to be preferred if it can be constructed
at reasonable cost. Before the use of grouted cut-offs was com-
mon, the first type was sometimes carried, in exceptional cases,
to depths of 50 ft. or more. Its advantages are twofold, in that
it not only provides perfect inspection of the vertical area to be
improved, but is a more tangible and sure method of obtaining the

desired end. For the excavation of the trench, even more care
must be taken not to shatter the surrounding rock, particularly
at the bottom. For the cut-off trenches of a number of important
dams, where the excavation has been difficult, holes have been
drilled a short distance apart to form planes of weakness to which
the sides of the trench break without serious disturbance outside
of the limits desired. The excavation has sometimes been made
Art. 61] ROCK FOUNDATIONS 175

with a channeling machine, but this method, of course, is very

expensive and hardly justified for any case, unless the rock is
easily cut. The concrete for the cut-off, in rock foundations, is
usually built under the same specifications as for the rest of the
dam, and provided with the same system of joints.
Cut-offs of the second type have been more common of late
years. Typical examples may be found at the Estacada Dam, in
Oregon; * the Lahontan Dam, of the Truckee-Carson project,
U. S. Reclamation Service; f the Horseshoe Falls Dam on Bow
River; J and the Arrowrock Dam, of the U. S. Reclamation
Service.*
A method which seems to have given the best
of grouting
primary series of holes in a row under,
results consists in drilling a
the heel of the dam, from 10 to 15 ft. apart on centers, and of a
depth depending on the nature of the foundation and the head of
water to be sustained. The holes are usually about 3 in. in
diameter.
The primary holes are first subjected to water pressure, prefer-
ably from a tank placed at the same height as the crest of the dam,
or a little above it, and the rate of leakage from each hole recorded.
The flow of water also serves to clean out the earth seams in the
vicinity of the holes in preparation for the grout which is to follow.
After the primary series of holes is grouted, a second series
consisting of an equal number of intermediate holes in the same
row is then drilled, tested, and grouted. If necessary, a third
series is drilled and treated, thus reducing the spacing to a quarter
of that for the first series. The result of a test of any hole is con-
sidered an indication of the relative tightness of the foundation
between the two adjacent holes previously grouted. The process,
therefore, is continued until the tests indicate that the leakage has
been reduced to a satisfactory extent.
The upper ends of all holes to be grouted should be provided
with threaded pipes with which to make connection to the grout-
ing machine. These pipes must be anchored or weighted to pre-
vent a blow-out during the process of grouting. This is sometimes

* For description see paper by Harold A. Rands, Transactions, Am.

Soc. C. E., Vol. LXXVIII, p. 447.

f For description see article by D. W. Cole, Engineering News, Apr. 3, 1913.

t For description see article by H. S. Johnson, Engineering Record, Dec.
12, 1914.
176 PROTECTION OF THE FOUNDATION [Chap. X

done by cementing the connecting pipes into the drilled holes, or

placing them in a concrete cut-off carried far enough into the rock
to provide ample grip. The drilling, in the latter case, is usually-
done through the pipes. When it is desired that the drilling shall
not interfere with the erection of masonry, the pipes may be car-
ried up along with the masonry and the operations of drilling,
testing, and grouting conducted from whatever elevation the
masonry has reached.
If it is apparent from the borings that there is an open seam

close to the surface, care shouldbe taken that there is sufficient

weight above the seam to balance the grouting pressure. Part of
the dam may be previously built to assist in this respect.
The initial grouting pressure, for each hole, should be that
which is necessary to force the grout slowly into the hole. The
pressure is raised gradually, as the hole tightens, so as to disturb
the natural formation as little as possible.
The grout usually consists of a mixture of neat cement and
water, of proportions varying to suit the character of the founda-
tion. For porous rock, with fine seams, a mixture as thin as 1 of
cement to 8 or 10 of water has been found satisfactory. Where
the seams are large, and other voids exist, a thicker mixture must
be used, gradually changing to a thinner mixture as the hole tight-
ens to refusal.
The grouting operations are usually started from one or two
ends of the site, the holes being treated successively. Each hole
should be capped, if necessary, as soon as it has been grouted,
in order that the grout to be forced into the next hole will not flow
back through the completed hole and be wasted, instead of passing
on to the ungrouted portion of the foundation.
If grout issues freely from an untreated hole, indicating an open
seam between it and the hole being grouted, the untreated hole
may be capped and the grouting operation considered as serving
the two holes. The object is to supply grout to each hole in
sufficient quantities and at the desired pressure, and it is immaterial
whether it is supplied at the top of the hole or by way of an open
seam from another hole. Should the grout from an untreated
hole issue sluggishly, indicating an indirect or only partly free
connection, the hole should be capped, but should receive its
share of treatment in due course.
When thick grout is being used, care must be taken that grout
Art. 61] ROCK FOUNDATIONS 177

from one hole does not partly fill an adjoining hole and set before
the latter can be treated. When thick, grout sets much more
quickly than when thin. It is sometimes advisable to provide
sufficient shifts of men to conduct the grouting operations con-
tinuously.
All holes should be gone over the second time after the grout
has set. If the grout is thin, considerable settlement in the hole
will be observed, and it is often possible to inject an additional
quantity.
No one method of grouting
will apply for more than one site;
in fact it is usuallyfound advisable, for each case, to change the
adopted method several times during a course of treatment, due
to experience constantly being gained as the work progresses.
The following is an abstract from the previously mentioned
article on the grouting of the Lahontan foundation. The dam is
an earth embankment 124 ft. high with a core-wall, but the
process of grouting would apply equally well to a masonry dam.
The grouted portion of the foundation consists principally of a
formation resembling red sandstone, varying from solid rock to a
tough red clay or even an unstable sandy clay, containing many
intermixtures and stratifications.
In the concrete cut-off trench (Fig. 62a), which was 30 ft.

deep, were previously set 5-in. well casings in two rows 2 ft. apart.
The casings were 3 ft. apart in each row, and were staggered, so
that there was a casing for every 18 in. of length of cut-off, as
indicated in Fig. 62. Core drillings, with 2f in. outside diameter
bits, were made through the well casings and extended at least
30 ft. below the bottom of the concrete cut-off. Some of the
holes were carried 70 ft., and the average was 40 ft., below the
trench, or 70 ft. below the river bottom.
The work was divided into two sections, on account of the
necessity of reserving half the width of the site for the diversion of
the stream.
In the first section, alternate holes, 6 ft. apart in the upper row,
were drilled, tested, and grouted; the intention being to complete
the upper row, and then the lower row, in successive steps. Those
of the first series were designated " primary holes." Testing was
done under a head of 127 ft. The rate of flow through each hole
was recorded.
After the holes of the first series were grouted, the interrnediate,
178 PROTECTION OP THE FOUNDATION [Chap. X

tested, and
or " secondary holes," in the upper row, were drilled,
row. Fig.
grouted, thus completing the grouting in the upper
62 indicates the results of all tests, and the number of bags
of

cement used in each hole. It will be noticed that the grouting of

- Section 1
21 29 33 37 41 45 19 63 57 61 65 69
1 9 13 17
5
oooo»ooo» o o o_o .OOOOOOOOOO ooo o > • O O O O

- Section 2 -
97 113 121 125 129 133 187 141
OOOOOOODOO»OoOO«OOOOOOoOoOOOO«
7377 85
81 89 93 101 105 109 117
#SBass ••••

a..00-
?80-
|60-
=
«j 40

a20-
«
o o-
^100-

g sol-
as'
o 40-
3 go-
to 01=
Art. 61] ROCK FOUNDATIONS 179

claimed that the quantity of cement which the holes took was not
indicative of their relative tightness, on account of caving of the
holes resulting in varying volumes to be filled, and the escape of
grout into adjoining well casings.
The results obtained for the first section of the dam led to an
alteration in the procedure for the second section. In the latter
the holes were drilled, tested and grouted in the following order:

Fia. 62a. —Grout Pipes in Cut-off Trench of Lahontan Dam. (Eng. Record,
Vol. LXVII, p. 340.)

" Primary holes " Nos. 78, 86, 94, etc., 12 ft. centers throughout;
" Secondary holes," Nos. 74, 82, 90, etc., 12 ft. centers throughout;
"1st Tertiary holes," Nos. 76, 84, 92, etc., 12-ft. centers through-
out; "2d Tertiary holes," Nos. 80, 88, 96, etc., 12-ft. centers
throughout.
The average leakage per hole for the second tertiary holes
was 6.4 gal. per min., indicating the necessity of grouting very
few holes in the second row.
All grouting was done with a duplex-cylinder, air-stirring, Caniff
 180 PROTECTION OF THE FOUNDATION [Chap. X
grouting machine; a pressure of 25 lb. being used at the begin-
ning of each operation, increasing gradually until finally the grout
was driven home at a pressure of 100 lb. per sq. in. mixtureA
of 1 part of cement to 7 or 8 parts of water was found most desir-
able, although the mixture was thickened considerably when the
flow appeared to be too free.
Uplif t pressure from head-water cannot be eliminated entirely
unless the cut-off is absolutely tight. With a deep, well-built,
masonry cut-off or a grouted cut-off, constructed under favorable
conditions, the uplift may be negligible.
Where it is thought that the effectiveness of the cut-off is not
thoroughly reliable, and in other cases where unusual precautions
are desirable, it is the practice to provide drains in the foundation
to facilitate the escape of whatever water finds its way past the
cut-off.
For this purpose holes may be drilled at intervals across the
site ina row just below the cut-off and provided with a free con-
nection to tail-water. The holes should be drilled to or a little
above the elevation of the bottom of the cut-off. For grouted
foundations the holes should be drilled after all grouting operations
have ceased.
The holes are usually 10 or 15 ft. apart; but the spacing de-
pends greatly on local conditions. Other things being the same,
the proper spacing should vary directly as the depth of the holes,
and directly as the perviousness of the foundation relative to that
of the cut-off.
The holes also serve the purpose of indicating the extent of
leakage past the cut-off, and, for this reason, should discharge a
little above tail-water, in order that the quantity of flow may be

observed.
Box drains, open-jointed pipe, and other types of drains, used
without drilled holes, have often been placed between rock foun-
dations and the dam, and may be effective in eliminating uplift
at that elevation; but they are not adaptable to rock foundations
containing nearly horizontal open seams near the surface, as they
have no direct connection with such seams, and may be sepa-
rated therefrom by a horizontal layer of very impervious rock.
For the usual spillway section, the smooth thin sheet of swiftly
moving water issuing horizontally from the bucket at the toe of
the dam, may, under certain conditions, retain a uniform depth
Art. 62] EARTH FOUNDATIONS 181

;uid velocity until, at a point, perhaps, considerably remote from

the dam, it suddenly increase in depth to tail-water level,
will
forming a standing wave, as indicated in Fig. 4. and as described
in the latter part of Art. 14. Under such conditions the river bed
below the dam will be subjected to a velocity much higher than
normal. The impact from swiftly moving water is known to be
very destructive to rough, soft, rock surfaces, particularly if there
are seams into winch the jet may In the latter case the
enter.
velocity head may be converted into a pressure head capable of
lifting large masses of rock.
In all cases, where the rock surface is known to be soft or
stratified, an apron of concrete, or other suitable material, should
be provided to afford protection to the foundation at the toe of the
dam. However, as explained above, there may be danger, under
certain conditions, of the water not returning to normal tail-water
level until after it has passed the end of the apron. The likelihood
of undermining the toe of the apron, in such cases, is evident, and
systems of baffles forming a partial obstruction to the flow, have
sometimes been built to insure the return of the sheet of water to
tail-water level before it has reached the end of the apron.
The concrete baffles used for the Gatun spillway dam are indi-
cated in Figs. (3 and 7. Dining large flows, the jet, when it reaches
the baffles, is about 6 ft. thick, and flows with a velocity of about
60 ft, per sec. The baffles are 9 ft. high. The effect of the baffles,
as indicated in Fig. 7, is to increase the depth to tail-water level
or about 20 ft. deep, and to reduce the velocity to IS ft, per sec.
The baffles piers are heavily reinforced, and are protected on their
up-stream faces with thick iron castings.
Where the profile of a spillway dam site is not horizontal, the
water passing the ends may flow along the toe of the dam toward
the middle with high velocity and destructive force. Fig. 63 is a
view of the down-stream face of the Estacada Dam. It will be
noted that training walls are provided to conduct the water to a
safe distance.
62.Earth Foundations. Masonry dams on earth * founda-
tions are numerous; but their use has been practically limited to
structures not more than 50 ft. high for good hardpan and 30 ft.
for less resisting earth. This limitation in height may be attrib-
* Earth, as here used, may be considered as all kinds of material riot

usually classed :is bed-rock.

182 PROTECTION OF THE FOUNDATION [Chap. X

g
bO

c3

H
a

(3
Art. 62] EARTH FOUNDATIONS 183

uted to the fact that the treatment of earth foundations, to prevent

erosion and excessive seepage, requires an expenditure far in excess
of that necessary for the foundations of dams on rock. In fact,
the cost of foundation treatment for dams on earth is often the
major part of the total cost of the structure. Consequently, for
moderate and high dams it will be found best to adopt another
type of structure, or change the site. There have been few
precedents for dams higher than noted above, although struc-
turally there would seem to be no reason for a limit to the height,
provided sufficient funds are available to meet the unusual expense.
It is essential that there be no excessive, unequal settlement of
the dam, as the tightness of the structure is dependent on the
absence of settlement cracks.
The preparation of the foundation for a dam on earth must be
made with four objects in view:

a. To prevent excessive seepage under the dam;

6. To prevent scouring by the water passing over the dam;
c. To provide ample bearing strength;
d. To prevent sliding.
Excessive seepage, or Underflow, through the foundation is

objectionable, not only on account of the waste of impounded

water, but principally because of the danger of movement of the
particles of the foundation due to its erosive power.
Where practicable, seepage should be prevented by carrying a
tight cut-off, under the heel of the dam, to an impermeable forma-
tion. The cut-off may consist of a concrete diaphragm or a series
of grouted holes, as described for rock foundations; interlocking
steel sheet-piling, or tongued and grooved wood sheet-piling.
Sheet-piling, unless driven under favorable conditions and with
extreme very apt to leak badly; and this is a condition to be
care, is
avoided, as recent experiments have indicated that very slight
leakage is sufficient to destroy its effectiveness. Wooden sheet-
piling for cut-offs should never be used where the foundation con-
tains boulders which are large enough to cause the piles to buckle
or deflect. Even steel sheet-piling has been made useless under

very heavy driving.

Figs. 39 to 42, inclusive, show details of the Mathis Dike Dam,
of the Tallulah Falls water-power development. The dam is
built on sand mixed with very little clay and decomposed mica
 184 PROTECTION OF THE FOUNDATION [Chap. X

schist. A concrete cut-off wall was provided, extending from 30

to 57 ft. to rock, except at one end where, as indicated in Fig. 39,
holes were driven to rock and effectually grouted under a pressure
of 70 to 85 lb. after the reservoir was filled. Before grouting,
these holes were opened to the lower side of the dam, and were
found to leak from 0.5 to 12.0 gal. per min. The cut-off wall is

..composed of 1:3:6: cyclopean concrete, with tongued and

grooved vertical contraction joints. The upper 15 to 20 ft. of the
wall is reinforced in order to prevent shrinkage cracks. The
sheet-piling indicated in the illustration was an experiment, but,
owing to the nature of the foundation, was abandoned after 50 ft.
had been driven.
Where it is impracticable to carry the cut-off to an impermeable
formation, undermining may be prevented by providing a path of
enforced percolation of sufficient length. The velocity of the
underflow uas been found by experiments to be directly propor-
tional to the hydrostatic head * and inversely proportional to the
length of the path of percolation. Its erosive force depends on its
velocity. The required length of path of enforced percolation,
therefore, is a direct function of the head on the dam, and depends
on the nature of the material in the foundation. W. B. Bligh f
gives the following empirical formula:

l = Chn> (44)

where Z = the length of the path of enforced percolation in feet;

which = h+l2+h+h+la+h;
see Fig. 64, in l

hn = the net head on the dam, in feet;

and C = a coefficient which depends on the character of the
material in the foundation.

For C, Mr. Bligh recommends the following values, which are

based on the dimensions of a number of existing dams on earth
foundations:
C= 18 for mud or silt;

C= 15 for fine micaceous sand;

* Experiments by D'Arcy, Hagen, Hazen, and others, indicate that the
flow of water through fine sand and gravel is similar to the flow through

capillary tubes, the velocity being directly proportional to the first power
of the head.
t"Dams, Barrages, and Weirs on Porous Foundations," Engineering
News, Dec. 29, 1910.
 ;

Art. 62] EARTH FOUNDATIONS 185

C = 12 for coarse-grained sand

C= 9 for a mixture of sand and gravel;
C=5 to 9 for clay, shale, or a mixture of sand, gravel, and
boulders.

03
w
a
o

MMafe
i ---^-

11
186 PROTECTION OF THE FOUNDATION [Chap. X

Care should be exercised in the application of these coefficients,

as impossible to convey an exact description of the character-
it is

istics of the foundations on which they are based. Therefore,

ample margin should be allowed, unless careful comparison has
been made between the materials in the foundation and those in
existing structures of like nature.
It has been proved experimentally that the length of the path
of percolation, as affecting uplift pressure and erosive force, is not
the shortest distance between head- and tail-water, but the length
of the actual plane of contact between the structure and the earth,
including provided the cut-offs are not closer together
all cut-offs,

than twice their depth. In other words, the effective length of

the path of percolation, as indicated in Fig. 64, is the summation
of the distances, h, to h, inclusive.
The proper length of the path of enforced percolation may be
obtained by providing a down-stream apron, an up-stream apron,
one or more cut-offs, or a combination of these parts, as indicated
in the accompanying illustrations. Many types and combina-
tions have been proposed and constructed. There seem to be no
standards in this respect. The choice will depend, to a large
extent, on local conditions. A single masonry cut-off at the heel is
commonly provided for good earth foundations; but it is often
found advisable to substitute an up-stream or down-stream
masonry apron, which is theoretically as efficient as a cut-off of
half its length.
The Jamrao and Barra type of weirs, indicated in Figs. 65 and
66, were designed for localities where the foundations are com-
posed of very pervious material, requiring a long path of enforced
percolation.
Fig. 67 is a section of a portion of the Granite Reef Dam of
the Salt River Project. The dam is founded on a formation of
gravel and boulders. The coefficient, C, from Eq. (44), was only
4.2, the length of the path of percolation being measured on the
under surface of the concrete from the river bed at the heel to the
first drain hole in the apron. When the dam was first used, con-
siderable water passed under the cut-off and out through the
drains in the apron. Fortunately, this flow was soon stopped by
the large quantity of silt which was carried by the river and

deposited at the heel of the dam, forming an increased and sufficient

length of path of percolation.
Abt. 62] EARTH FOUNDATIONS 187

I F-t '

.a

1 5

Ph

03

O Ba l-H

a
4 -#*!

A.. ±
188 PROTECTION OF THE FOUNDATION [Chap. X

In earth foundations, the river bed below the dam

all classes of

must be protected from the wash of the water passing over the
crest, in order to prevent undermining the structure, from this
cause. The object is to provide, as far as possible, a means for
the stream to regain its normal velocity, corresponding to the

Fig. 67. —Granite Reef Dam, Salt River Project, Arizona. ("The Design
and Construction of Dams," Wegmann.)

flow, before the end of such protection is reached. It is common

practice to provide a concrete or rock-filled timber crib apron for a
short distance below the dam, the end of the apron being protected
by an extension of rip-rap, as indicated in the accompanying
illustrations. The lower end of the apron is often further pro-

Fig. 68.—Power Dam on Au Sable River. (Eng. Record, Vol. LXVI, p. 247.)

tected by a vertical diaphragm of concrete or sheet-piling, which

serves to retain the foundation, under the apron, if the rip-rap is
washed away. Similar protection has been used at the toe of the
dam to provide for a possible failure of the apron.
The usual type of spillway section, passing large flows, is not
adaptable to dams on earth foundations unless the tail-water has
Abt. 62] EARTH FOUNDATIONS 189

sufficient depth to break the force of the falling water,* or unless

baffles are provided, as described in the last part of Art. 61.
A type of apron frequently used is indicated in Fig. 68. The
water is dropped, in one or more stages (two in this instance), into

pools of sufficient depth to destroy the velocity.

In case the bed of the river is higher at the ends than at the
middle, the water passing over each end, if not taken care of, will

flow parallel to the dam toward the main channel. For steep
slopes, this flow may acquire velocities sufficiently great to scour
the foundation at the toe of the dam or the end of the apron. This
condition should be avoided by the construction of stone fill,

masonry, or other suitable training dikes at intervals extending

from the dam to a point down stream far enough from the dam to
obviate the possibility of damage to that structure. Sometimes
the training dikes are supplemented by a system of canals, parallel
to the river, in order to provide a gradual descent, f Such con-
struction is obviously expensive, particularly for large flows, and
often necessitates the limitation of the length of spillway to the
width of the level portion of the river bed.
Proper unit compressive stress on the foundation is usually
obtained by spreading the footing of the dam, although supporting
piles have been used in some cases. Many types of wood and con-
crete piles have been used for this purpose. The subject of piles
is too lengthy for proper treatment here, and the reader is referred

to one of the many books on foundations of that type, such as

" Foundations of Bridges and Buildings " by Jacoby and Davis.
J
The weight of the hollow Mathis Dike Dam was distributed
over the foundation by a mattress covering the entire base, as
The weight of solid dams may be distributed
indicated in Fig. 41.
through the aprons, which are often reinforced for that purpose.
Where sufficient length of enforced percolation is provided by a
cut-off at the heel of the dam, the base is usually drained, to pre-
vent the possibility of uplift; unless, of course, the weight of the
dam great enough to sustain the uplift pressure, as hereafter
is

discussed. The best type of drain for this purpose consists of an

inverted filter surrounding open tile pipe, although many other

* See Art. 14.

fSee "The Laguna Dam," Engineering News, Feb. 9, 1905, and Feb.
27, 1908.

J McGraw-Hill Book Co., 1914.

190 PROTECTION OF THE FOUNDATION [Chap. X

types have been used. The main drain should be placed imme-
diately below the cut-off, at the heel of the dam, as shown in Fig.
70. Drainage is often provided, even though the cut-off extends
to impervious material.
Where it is possible to reach impervious material, it is common
practice to place the only cut-off at the heel of the dam, although
lower rows of auxiliary sheet-piling and even concrete diaphragms
are sometimes provided, in order to prevent undermining from
water passing over the crest, in view of a possible failure of the
apron or apron extension. Examples of such auxiliary piling and
diaphragms are indicated in Figs. 70 and 67, respectively. In
such cases they should be well perforated or drained, in order to

=jSjTi5i 5 '
8 or over
>!< 3'5^>i< JLU
£
..

Fig. 69. —Diversion Dam of the Rio Grande Project, New Mexico. ("Irri-
gation Practice and Engineering," Vol. Ill, Etcheverry.)

hold back no leakage which may pass the upper cut-off and main
drain. Intermediate drains are also desirable, if the area of the
base of the dam or apron is large. In brief, to prevent uplift on
the base of the dam, the cut-off at the heel should be as tight as
possible, and the foundation below the cut-off as pervious as pos-
sible.

The Rio Grande Dam was drained by providing a mattress of

well-rammed broken rock under the entire base and apron. Leak-
age is discharged through the pipe over the lower row of sheet-
piling indicated in Fig. 69. The mattress of the Mathis Dike
Dam was well perforated.
Fig. 70 is a section of the spillway of the Coon Rapids power
development on the Mississippi River. This dam is built on a
Art. 62] EARTH FOUNDATIONS 191

deep bed of and is carried on a pile foundation. The

glacial drift,
sheet-piling at the heel was driven to penetrate at least 5 ft. into
impervious material, which varied from 2 to 25 ft. below the sur-
face. The filter, drain, and tunnel are intended to provide for
the escape of any flow which may pass the heel, thus relieving the
dam from upward pressure. A similar row of sheet-piling was

Operating Bridge
-$' Z
7 *i E1.84S.5

-BI-
W
vrvrfcacnaroKDrrcutnEroa?

SECTION B-B
192 PROTECTION OF THE FOUNDATION [Chap. X

pressure head of the underflow, for earth foundations, is

The
usually considered to vary uniformly, along the path of percola-
tion, from full head-water pressure at the up-stream end to full
tail-water pressure at the down-stream end. The hydraulic
gradient may be drawn, as in Fig. 64, by considering the pressure
head, K, at any point, A, on the path of percolation, to be equal
to the vertical distance, hh from that point to head-water level,
,

less the friction loss, h/t between the up-stream end of the path of
percolation, and the point, or,
s

Ku = ilh — "/

The friction loss, hf, is proportional to the relative distance

through which the water has traveled and the net head, hn, on
the dam, or,

, _, h+h+k+k+k+h
'~
"h + h + h + h + h + h'

The uplift pressure, pu , is, for earth foundations, considered as

effective over the entire area, therefore,

pu = W2h u = W2(hh —hf),

or,

The uplift pressure, pu under the apron must be balanced by

,

the weight of the apron and the weight of water on the apron.
To allow a factor of safety, the thickness of the apron is usually
made at 30 per cent greater than theoretically necessary.
least
It was shown in Art. 14 that tail-water may not always be counted
on to assist in balancing uplift, in which case the presence of tail-
water should also be neglected in Eq. (45). The uplift pressure
will vary with the changing relative elevations of head- and tail-
water, and the weight of the apron and dam should be proportioned
for the severest condition.
The thickness of the apron may be reduced, if properly anchored
to bearing piles.
When the friction of the foundation is not considered suf-
ficient to prevent sliding, piles may be used, as in Fig. 69, although
Art. 62] EARTH FOUNDATIONS 193

it isobvious that vertical piles cannot withstand lateral pressure

without deflecting to some extent. Therefore some movement of
the dam may be expected, depending on the number of piles pro-
vided. For this reason, some of the piles are often inclined in the
direction of the resultant pressure.
In the Mathis Dike Dam, Fig. 40, the longitudinal ribs, at the
toe and the middle of the base, are intended to assist in preventing
the dam from sliding.
In some cases increased resistance to slidinghas been
obtained by thoroughly anchoring the dam to a deep cut-off
wall at the heel of the structure,
 CHAPTER XI
FLOOD FLOWS

63. General Considerations. The magnitude of the prob-

able maximum flood to be expected at a dam site has a direct
bearing on the safety of the structure. This is particularly true
if, as is often the case, the masonry dam is supplemented at one

or both ends by an earthen embankment. Although, for a ma-

sonry dam, a flood somewhat greater than that for which the
structure has been designed may, because of the margin of safety,
be passed without failure or even damage, an earthen embankment
will almost invariably be breached soon after the water has begun
to flow over the top.
As the magnitude of a flood is affected by an infinite variety of
conditions, the chances of the maximum possible flood from a
given catchment area occurring within the life of the structure
is infinitely remote. It therefore remains to adopt, as the design-
ing condition, a flood which is reasonably certain of not being
exceeded; it being remembered that a flood which would be con-
sidered " reasonable " for one dam may be unreasonably large or
small for another dam having exactly the same size and character-
istics catchment area. The probability of loss of life and
of
property, of interest on the investment, of the use for which the
dam is intended, and other factors, must be considered in deciding
on the flood to be accommodated. In other words, as it is imprac-
ticable to design any, except the most important of dams, for the
maximum possible flood, the problem is resolved into a considera-
tion of how much chance it is reasonable to take. It is safe to
say that very few dams in America have been designed to accommo-
date a flood proportionally as great as that which occurred on the
Miami water-shed at Dayton, Ohio, in 1913 (see No. 72 of Fig.
71), which was one of the largest floods this country has ever
known.
The maximum flood to be accommodated from a given catch-
ment area may be determined approximately by the following
194
Art. 64] HIGH-WATER MARKS 195

methods, it seldom, if ever, being possible to make a direct dis-

charge measurement on a record flood at the time of its occurrence:

1. By a study of record high-water marks on the stream in

question;
2. By comparison with known record floods from other catch-
ment areas of about the same size and characteristics.

64. High-water Marks. Authentic Federal and State Gov-

ernment records of high-water, extending over long periods, may
be obtained for many streams. Such records are also often avail-
able from mill operators and the officials of municipalities. In the
great majority of cases, however, the determination of the ele-
vation of record high-water must be from the observations and
traditions of residents, and from physical indications on the banks
of the stream.
Observations and traditions of residents should be regarded
with caution. Individual reports of untrained observers are sub-
ject to great error and, strange to say, are often of doubtful verac-
ity, as the desire to report a high-water a little higher than that

reported by a neighbor is often, among certain classes, greater than

the love of the truth. Unfortunately, also, reports are sometimes
biased by a desire to give an impression of great or small maximum
floods, whichever, in the opinion of the observer, will better serve
his interests. However, credence may be given when a number of
observations closely agree and are referred to definite objects, such
as of doors and windows or nails driven for reference.
sills

Confirmation may be obtained from the elevation of deposi-

from floating bodies
tions of brush, logs, or alluvial matter, scars
on banks and large trees, and whatever other indications of high-
water may be discovered. High-water, in an alluvial valley which
has been formed from the sediment deposited from floods, is, of
course, always higher than the surface of the valley.
The elevation of record high-water having been fixed, there are
three methods by which an estimate of the corresponding dis-
charge can be made:

1. By determining the corresponding head on a dam which

existed at the time of such high-water, from which the
discharge over that structure can be computed from one
of the well-known weir formulae.*
* See U. S. Geol. Survey Paper No. 200 by R. E. Horton.
 § S§?S8S
000,000
600,000

1700,000

600,000

600,000
*« ,« •* » » *- oo ofe>
S 8 tf 8" S" g" S 88
Abt. 65] COMPARISON WITH OTHER RIVERS 197
%
TABLE XXV
*
Unusual Flood Discharges, United States Rivers

Refer-
198 FLOOD FLOWS [Chap. XI

TABLE XXV—Continued

Refer-
Art. 65] COMPARISON WITH OTHER RIVERS 199

TABLE XXV—Continued

Refer-
200 FLOOD FLOWS [Chap. XI

3. The capacity of the water-shed suddenly to release stored

waters.

The annual rate of rainfall on any water-shed is not always an

indication of the maximum rate or intensity of precipitation which
may be expected. A detailed study of the records of the U. S.
Weather Bureau for rain-gauging stations on the catchment area
and vicinity will prove of material assistance in determining the
maximum rate and duration of precipitation to be expected.
Storage, of whatever nature, has a tendency to reduce the size
of floods. The storage capacity of the water-shed may be divided
into the following items:

1. Storage in reservoirs, lakes, and swampy places;

2. Storage below the ground surface;
3. Storage above the ground surface.

It isseldom that storage below the normal flow line of artificial

reservoirs is effective in reducing the peak of large floods, because
at such times the reservoirs are very sure to be full due to the
excessive flow preceding the peak. Storage above the normal
flow line of all bodies of water is always available, as such bodies
of water must suffer an increase in surface elevation in order to
provide sufficient head at the outlet to accommodate the increasing
flow. The percentage of area of reservoirs, lakes, and swampy
places has considerable influence on the value of the flood coeffi-
cient.
The magnitude of moderate floods is always less on rivers
draining large sandy areas, in which the storage capacity below
the ground is considerable. Unless, however, such sandy areas
are large and extend to the higher elevations of the catchment area,
any reduction in large floods will be effected, because, at
little if

such times, the voids below the ground surface are apt to be com-
pletely filled as the result of the excessive precipitation preceding
the peak of the rainfall. Storage below the ground, therefore, is
usually considered only as increasing the interval between floods.
Storage above the ground is affected by the nature of the vege-
tation, the shape and slope of the catchment area, and the char-
acteristics of the river bed and banks. It is evident that those
characteristics which will permit of rapid run-off of the precipi-
tation to the site of the dam will result in large floods. Rocky
Art. 65] COMPARISON WITH OTHER RIVERS 201

slopes, devoid of vegetation are conducive to quick discharge.

Conversely, areas covered with dense vegetation will prove effec-
tive in holdingback the water and smoothing out the peak of the
flood. Heavy underbrush is particularly effective in this respect,
as the rivulets are held back by friction in passing around and
among the stalks of the plants and such branches as haAT e been
beaten down to the ground surface. Practically no water, at the
peak of the precipitation, is by adherence to leaves and
held back
branches above the ground surface. For this reason it is the
opinion of many engineers that it is the removal of the dense under-
brush rather than the large trees which has increased flood ten-
dencies in districts which have been deforested.
Steep slopes, of course, will produce rapid run-off. Therefore
floods from mountainous districts are relatively severe.
In rivers having tributaries extending fan-shaped from a given
point, and of approximately the same size, the peak of the flood
from each of the tributaries is apt to reach the main stream and
the dam at approximately the same time, resulting in relatively
large floods. Conversely, when the catchment area is relatively
narrow, with tributaties of different sizes discharging into the main
stream at regular intervals, the peak of the run-off from the trib-
utary areas will reach the dam at different times, resulting in
relatively small floods. A large number of tributaries is also pro-
ductive of rapid run-off.
Rivers and tributaries which have frequent restricted cross-
sectional areas,rough bottoms, and are relatively shallow in com-
parison with their widths may also be said to have a moderately
retarding effect on the rapidity of run-off.
The capacity of the catchment area to release stored water
suddenly may be indicated by:

1. The frequency and magnitude of iceand log jams, with

consequent danger of release of the impounded waters at
or near the peak of the flood.
2. The presence of other dams of questionable strength im-
pounding large volumes of water. A number of well-
designed dams have failed on account of the destruction
of defective dams above, with a resultant enormous
increase in the run-off due to the sudden release of im-
pounded waters.
202 FLOOD FLOWS [Chap. XI

due
3. Temporary partial blocking of the flow of the stream,
to lodgment of debris against submerged bridges, and
subsequent failure of the bridges, with a release of the
impounded waters at or near the peak of the flood.
4. Storage in the form of snow which may be suddenly released
by a record precipitation accompanied by a rise in

temperature.

Little or nothing has been accomplished which would indicate

a definite relation between flood coefficients and characteristics
of catchment areas. Until more information has been obtained
the matter must be left to the judgment of the engineer. It is
probable that, in general, the maximum rate and duration of rain-
fall, the steepness of the slopes, the shape of the catchment area,

and the arrangement of tributaries will, in the order given, have

the most influence on the flood tendencies of the stream. The
items mentioned as affecting the capacity of the stream to release
stored waters suddenly cannot be included in a general classifi-

cation, as their effect on floods is too uncertain.

An indication of the relative flood tendencies of two streams
may sometimes be gleaned from a study of the records of maximum
yearly floods when available. Mr. Fuller, in his paper on flood
flows, has given a unique approximate method of determining
directly the value of the flood coefficient, N, from a study of the
maximum yearly floods on the stream in question.
 CHAPTER XII

DETAILS AND ACCESSORIES

66. Masonry for Dams. During the latter part of the last
century, rubble masonry was used extensively for the construc-
tion of dams; but in recent years this type has been practically
superseded by concrete and cyclopean concrete masonry.
Cyclopean concrete masonry consists of plain concrete con-
taining a large percentage of irregular stones, or " plums." These
stones should be as large as can be economically quarried, trans-
ported, and handled; and they should comprise as large a per-
centage of the mass as possible, consistent with good work. Spalls
are rammed into the concrete between the plums. Plums and
spalls, of course, are intended to effect a saving in concrete, and,
if the work is properly done, should not reduce appreciably the
strength of the masonry.
For a small structure, the greatest dimension of any plum
should not exceed 20 per cent of the thickness of the structure,
provided the masonry is to be stressed to a point approaching a
reasonable working limit. A wet concrete is essential to the proper
setting of the plums. Fig. 72 shows the cyclopean concrete
masonry of the Olive Bridge Dam in the process of construction.
As much as 30 per cent of plums can be used advantageously,
but a larger precentage is apt to result in ineffective packing of
the concrete. The usual percentage is from 18 to 22.

has been claimed that a 1 2| 5 or even a 1 2\ 4 mix

It : : : :

for cyclopean concrete, on account of its greater fluidity, will

not only give decidedly better masonry than a 1 3 6 mix, but : :

will permit of the addition of many more plums and spalls, thus
causing an actual decrease in the quantity of cement per cubic
yard of masonry. It is also claimed that an aggregate limited to a
maximum size of 2 in. will, for the same reason, give better results
than the usual limit of 3 in. The results to be attained,
or greater.
however, will depend decidedly on local conditions and the per-

sonal equation of the superintendent, it often being the case that,

203
204 DETAILS AND ACCESSORIES [Chap. XII

Fig. 72. —The Olive Bridge Dam During Construction. Showing Plums
Projecting above Horizontal Joints.
Art. 67] WATER-PROOFING 205

through the lack of material, or by carelessness, the maximum

quantities of plums and spalls are not used, and then the richer
concrete is of no economical value.
The following proportions of concrete have been commonly
used for the various types of concrete dams:

Solid gravity dams and large

arch dams, 1:3:6 cyclopean concrete;
Thin arch dams, 1 : 2| : 4§ to 1 : 2| :
5, plain or
reinforced concrete;
Slab decks of hollow dams, 1:2:4 reinforced concrete;
Arched decks of hollow dams, 1:2:4 plain or reinforced con-
crete;
Buttresses of hollow dams, 1 : 2\ : 5 to 1:3:6 reinforced
concrete;
Struts of hollow dams, 1:2:4 reinforced concrete.

To the customary requirements for first-class concrete masonry

should be added, for masonry dams, the special provision for the
maximum possible bond at horizontal building joints. The
importance of this feature, as an additional element of safety,
has already been pointed out. The proper treatment in this
respect also serves to reduce greatly the leakage through such
joints, with a resultant decrease in uplift pressure and prevention
of unsightly discolorations of the down-stream face.
Wherever new concrete is to be laid on old, the surface of the
latter should be thoroughly cleaned by using stiff brushes and
streams of water; the dead cement and laitance should be scraped
from the old surface; and the latter should be thoroughly wet
just before the new concrete is placed. It is good practice to
spread a f-in. layer of mortar on the old surface immediately
before pouring the new concrete.
Care should be taken to have a considerable number of plums
project above the surface of horizontal building joints, in order to
increase the frictional resistance to sliding. In the absence of
plums, it is frequently the practice to groove the building joints,
as indicated in Fig. 41.
67. Water-proofing. Except as stated in Art. 69, no special
provision need be made through the masonry of solid
for seepage
gravity and thick arch dams, constructed of 1 3 6 or richer : :

concrete, other than cutting and working the concrete well against
206 DETAILS AND ACCESSORIES [Chap. XII

the forms in order to insure a skin of cement at the face. Such

dams a limited extent, but are tight enough for usual
will leak to
purposes. Leakage usually ceases after a few years, due to the
filling of cracks by silt, or by effervescence of magnesia or lime

from the cement.

For thin arch dams, where a mixture not richer than 1 2\ 4§ : :

is used, it may be necessary to provide an impervious coating on

the up-stream face.*

The 1:2:4 concrete for the decks of hollow dams should be
practically water-tight. This, however, necessitates a careful
proportioning of the aggregate, a wet mixture, complete incor-
poration of the materials, careful placing, and thorough puddling
and spading. Tight forms are absolutely necessary. For high
heads, the horizontal building joints of the decks have some-
times been coated with a layer of tar, asphalt or similar material,
from | to J in. thick.
In general, as much depends on the quality of the labor and
superintendence as on the methods specified and the materials
used. The author knows of cases where, with skillful workman-
ship, thin walls of 1 2\ 4|, and even 1 1\ 5 concrete have
: : : :

been made practically tight without the addition of any water-

proofing material or any special treatment of the finished surfaces.
68. Contraction Joints, f All monolithic masonry structures
of considerable length will crack because of restrained shrinkage
when reductions in temperature occur. Such reductions may be
due to atmospheric conditions or to the cooling of the cement
which, when setting, attains a high temperature.
Cracks are objectionable, not only on account of their unsightly
appearance, but because of their tendency to follow planes of least
resistance lying at almost any inclination to the lines of stress.
It is common practice, therefore, to construct the dam in
comparatively short sections, thus providing vertical contraction
joints in order to localize such cracks and confine them to planes
normal to the axis of the dam.

* For descriptions of water-proofing methods see:

"Report of Com-
mittee on Masonry," Bulletin of the American Railway Engineering Asso-
XV, No. 163; "Waterworks Handbook," by Flinn, Weston and
ciation, Vol.
Bogert, McGraw-Hill Book Co.; "Concrete, Plain and Reinforced," by
Taylor and Thompson, John Wiley & Sons, Inc.
t Sometimes erroneously called "expansion joints,"
Art. 6S] CONTRACTION JOINTS 207

has been found that temperature cracks will occur in small

It
walls about every 50 or 60ft. Experiments show that the inter-
nal temperature changes in any masonry structure will vary in-
versely as magnitude,* and, for this reason, contraction joints
its

are usually placed farther apart in large dams than in small ones.
For solid dams, a spacing about equal to the mean thickness of
the structure, but not less than 40 or 50 ft., seems to have become
standard practice. For this there is no logical reason, as the
height of the structure is probably also a governing condition.
In every case, however, it has apparently proved satisfactory.
Two series of contraction joints have sometimes been used in
very high dams, one series extending from the top to about mid-
height and the other completely to the foundation. Such an

Top of Dam. EL 4414

-»«M0 L

DiUnngft Gallerr

Him,.
^^i^Mk;
EL 4150

St*, i + oo a + oo s + 00 i + oo6+00 6 + 00 7 + 00 s + 00 9+00 10 + 00 11 too l^JKOO

* Put la taring lata winter ELEVATION OF ELEPHANT BUTTE DAM
showing contraction joints.
Fig. 73.

arrangement is indicated in Fig. 73. In order to minimize con-

traction, the sections marked by asterisks were specified to be
built in the coldest season.
The usual type of contraction joint is indicated in Figs. 74
and 37.
The size and shape of the vertical grooves, or "keyways,"
vary with the size of the structure, and are essentially a construc-
tion feature. It is sufficient only to provide as large a number of
corners as is practicable, in order to restrict the flow of water and
entangle as much as possible of the sediment and other matter
which may be available to tighten the joint gradually.
The contraction joint indicated in Fig. 75 is unusually elaborate,
and was designed to prevent, to as great an extent as possible,

* "Temperature Changes in Mass Concrete," by Paul and Mayhew

Transactions, Am. Soc. C.E., Vol. LXXIX, p. 1225.
208 DETAILS AND ACCESSORIES [Chap. XII

all leakage through the joint. The use of concrete blocks for one
face of the joint is not usual practice. Rectangular keyways, as
here shown, are designed with the idea of providing less area
of waterway, on the up- and down-stream sides of the keys,
should the joint open a considerable distance. It is thought,
however, that the gripping of the keys by the shrinkage of the
subsequently poured concrete on the other side of the joint, may
destroy its effectiveness and cause cracks through the base of the

SECTION B-B

SECTION A-A.

Fig. 74.—The Usual Type of Contraction Joint.

keys, if not elsewhere in the dam. Of the copper water-stop and

the inspection or drainage well, more will be said later.
In flat-deck hollow dams, a contraction joint is usually pro-
vided at each buttress. In multiple-arch dams, the arches are
usually reinforced continuously through the structure, contraction
being taken care of by arch deflection.
Contraction joints should be coated with suitable material in
order to prevent adhesion of the
masonry and to assist in checking
leakage. The author has used a' thick coating of coal-tar pitch
with complete success. The contraction joints of
the Mathis
Art. OS] CONTRACTION JOINTS 209

OOi}^ mmuisnAiOG^
210 DETAILS AND ACCESSORIES [Chap. XII

Dyke Dam, were painted with asphaltum. To insure

Fig. 40,
water-tightness, a 6-in. strip of hair felt, laid in hot asphaltum,
was placed between the deck and the haunches. In the con-
traction joints of the Farnham Dam, one face was covered with
paper between two coats of pitch, before the concrete on the other
side was poured.
87.75-t

' way Upper inspection

f? ,
-gallery
Floy line 13.355 -,

TYPICAL SECTION OF KENSICO DAM

20 10 60 ft.

Fig. 76.

69. Drainage Systems. In solid gravity and large arch dams,

the usu^rl requirements for first-class masonry and a treatment of
horizontal and contraction joints such as previously described
will result in a practically impermeable dam. Leakage through
contraction joints of the type indicated in Fig. 74
is never exces-

a few years have elapsed, and may be prac-

sive, particularly after
tically eliminated at once by using metal water-stops, such as
shown in Fig. 75. The use of these stops has been confined prin-
Abt. 70] ARCHITECTURAL TREATMENT 211

cipally todams in which a discoloration of the down-stream face,

due to leakage, would be objectionable, as affecting the appear-
ance of the structure.
Also, for thesame reason, seepage through the masonry and
horizontal joints has been prevented by special water-tight con-
struction at the up-stream face, backed by a series of vertical
drains to intercept and carry away all water which enters the dam.
The drains may also be arranged to receive any leakage which
finds itsway past the metal water-stops.
Except in very high dams, drainage systems in the body of the
dam have seldom been used solely for the purpose of preventing
uplift on horizontal planes. Unless the head is excessive, properly
treated horizontal building joints are sure to be capable of resisting,
in adhesion, the pressure of what little water finds its way into the
masonry. The conditions within the body of the dam are ideal
in comparison with those of the usual foundations, where drainage
systems are often necessary. (Art. 61.)
The elaborate system of drainage for the Kensico Dam is

indicated in Figs. 75 and 76. The up-stream face of this dam

was made relatively less permeable by the use of solid concrete
blocks. The drainage and inspection wells were formed by laying
up blocks of hollow porous concrete.
70. Architectural Treatment. Except in public works, the
architectural appearance of dams is usually neglected. In private
enterprises the need for strict economy usually will not warrant
the outlay necessary for esthetic treatment; particularly if, as is

usual, the project is in a sparsely populated district.

Perhaps the most extensive treatment for architectural effect

is that of the Kensico Dam, of the New York City Water Supply
System,* (frontispiece). In this case the dam, while serving a
useful purpose, was also intended to afford a monumental expres-
sion of the magnitude and importance of the largest municipal
water-supply system ever constructed. The entire down-stream
face of the dam is covered with a layer of granite masonry, arranged
in panels and surmounted by a continuous cut-stone cornice. A
close view of the facing is shown in Fig. 77. Extensive land-
scape work, with terraces, ramps, and artificial pools, imparts a
pleasing appearance to the structure.
"Architecture of Kensico Dam," by A. D. Flinn, Engineering News,
Vol. LXXIV, p. 433.
212 DETAILS AND ACCESSORIES [Chap. XII

To facilitate construction, the concrete heart of the dam was

completed before the setting of the face work was started. The
concrete was built in steps, arranged to receive the stone facing.
The style of ornamentation to be adopted must be in keeping
with the dignity of the structure. The most appropriate treat-
ment also depends, to a large extent, on the appearance of the
neighboring landscape.

'

Fig. 77.— Granite Facing of the Kensico Dam.

Considerable opportunity for esthetic treatment was

afforded
at the site of the City Reservoir, No. 3, at Portland,
Oregon,
Figs. 78 and 79. The type of architecture is admirably in keeping
with the beauty of the surrounding park.
On the other hand, the rugged character of the
sites of the
Salmon River and Elephant Butte Dams, Figs. 80
and 81, neces-
sitated a more simple, massive, and dignified
treatment. That of
the Salmon River Dam is thought to be the
better of the two.
 Art. 70] ARCHITECTURAL TREATMENT 213

For a solid gravity dam, the weight of copings, pilasters, etc.,

at the top, being effective in adding to the stability,
should not be

Fig. 78. —Near View of City Reservoir No. 3 Dam, Portland, Oregon.
considered an extra expense, except as to the increased cost of
forming and placing the necessary material. In a hollow dam, the
weight of water, and not masonry, offers the chief resistance to
failure, so that a greater expense is attached to such embellish-
214 DETAILS AND ACCESSORIES [Chap. XII

ments. An increased thickness, at the top of an arch dam, in the

form of copings or cornices, adds materially to the stiffness of the
arch.
A water-tight structure if a pleasing appearance is
is essential,
desired. The not only result in a discoloration
slightest leak will
of the down-stream face, on account of the darker color of damp

Fig. 79. —Distant View of City Reservoir No. 3 Dam, Portland, Oregon.
masonry, but will pave the way for a later deposit of laitance,
which is even more objectionable.*
71. The Regulation of High-water Surface. The land to be
purchased or controlled for a reservoir must include the area which
will be covered at the time of maximum flood. Obviously, the
water stored between the elevation of the crest of the dam and that
of highest water seldom serves a useful purpose, and many devices
have been used to keep the maximum rise of water surface as
small as possible.
* See Fig. 78.
Art. 71] REGULATION OP HIGH-WATER SURFACE 215
21G DETAILS AND ACCESSORIES [Chap. XII

r
Art. 71] REGULATION OP HIGH-WATER SURFACE 217

Flood regulators, in almost unlimited varieties of forms,*

have been used. Brief attention will be given here to only a few
of the types which have become common in America.
Figs. 82, 82a and 83 show typical examples of crest gates. The
gates are operated from an elevated platform, the openings being
maintained at all times at an area sufficient to provide practically
constant water surface elevation in the reservoir. The gates,
when entirely raised, should have sufficient capacity to discharge
the entire flood, and, if fragile, should be high enough to be clear
of large trees and other heavy objects which may be brought down
by floods.
The piers may be designed as separate gravity segments down to
the rock surface, or as reinforced concrete cantilevers relying for
stability on the excessive weight of the adjacent overflow portions.
Vertical grooves are usually provided in the piers, a short dis-
tance up-stream from the gates, to facilitate the placing of stop-
logs when inspection or repairs of the gates become necessary.
The rectangular type (Fig. 82), varies considerably in detail,
according to local conditions and the judgment of the designer,
not only in the nature of the materials used, but in the types of
bearings, operating mechanism, and the size of openings. Designs
and estimates may be obtained from a number of manufacturers
who make a specialty of work of this class.
Probably the most common type of crest gate, at the present
time, the sector or " Taintor " gate, Figs. 83, 70 and 68. In
is

this type the pressure of the water, passing through the pivots,
causes no resistance to opening or closing, except that of the in-
considerable friction at the pivots and the sealing strips at the
sides. Practically, the only force to be controlled by the operating
mechanism is the weight of the gate.
Small gates of this type have been made of wood, but for the
larger sizes, a steelframework with a steel or wood water face is
usually adopted.
Water-tightness at the sides of the Great Falls gates (Fig. 83),
was obtained by fastening to them strips of 8-in. five-ply rubber
belting which slide on the faces of the concrete piers. Local
smoothness of the concrete was obtained by nailing J-in. smooth
steel plates to the inside of the forms. The plates were removed
•See "Improvement of Rivers," by Thomas and Watt, 2d Edition.
John Wiley & Sons, Inc., 1913.
218 DETAILS AND ACCESSORIES [Chap. XII

-Spindle Housing

Operators Cab.
is Counterweight
Truss

Top of gate when

in raised position\

11,537

E1.525

^Normal Water

OUTLINE OF TRAVELER
KEOKUK DAM SLUICE GATES
MISSISSIPPI* RIVER POWER CO.

Fig. 82.
Art. 71] REGULATION OF HIGH-WATER SURFACE 219

Fig. 82a. —Up-stream View of Keokuk Dam Crest Gates, Showing Concrete
Counterweight and Steel Gate.
220 DETAILS AND ACCESSORIES [Chap. XII

Fig. 83.—Taintor Gates for Great Falls Power Dam, Caney Fork River,
Tennessee.
Art. 71] REGULATION OF HIGH-WATER SURFACE 221

with the forms. For the bottom seal, the lower edges of the gates
were planed to a sharp edge. In lowering, this edge cuts through
irregularities in the sills, small pieces of wood, and other debris,
and the gates come to a continuous bearing.
The pivot pins are usually made to cantilever out from the piers,
although they sometimes bear on girders spanning from pier to
pier, if danger from floating debris is not feared. The bearing
boxes are usually babbited or lined with bronze, in order to
prevent them from rusting against the steel pins.
In a plant where a large number of Taintor gates are used,
operation is usually effected by one or more traveling hoists to
which motors are attached. In cold climates steam pipes should
be provided in order to prevent the gates from freezing tight.
Objections have been raised to all form of crest gates which are
not entirely automatic in operation, particularly if subjected to
ice conditions, on the ground that constant vigilance is necessary
for their successful operation. Such antagonism, however, has
not prevented the use of non-automatic crest gates in a great many
cases, particularly where a sufficient force of men is always avail-
able for emergency operation. A few automatic crest gates have
been adopted, but their use has not become common.
Fig. 84 shows a typical sluice through the body of the dam.
Such sluices may serve a variety of purposes, and are sometimes
relied on to pass considerable of the flood flow and thus reduce,
in a measure, the rise of water surface during floods. It is seldom
found economical to provide a sufficient number of sluices to take
the maximum flood flow, as in the case of crest gates.
The Stevens Creek sluice gates are made so as to split, as indi-
cated in Fig. 84, to facilitate removal to the passageway through
the water-tight bulkhead. The gates are made accessible by low-
ering a weighted timber stop gate between the projections on the
up-stream face of the dam. If the leakage through the sluice is
too great to permit of the stop gate being lowered against the flow,
it may be lowered in a horizontal position to an elevation slightly

above the top of the sluice entrance and then allowed to swing
down over the opening. The sluice is protected from erosion
near the gate by a cast-iron lining. Should tail-water rise above
the level of the passageway floor, suction at the contracted throat
of the sluice will effectually remove all leakage through the drain-
age pipe provided for that purpose.
222 DETAILS AND ACCESSORIES [Chap. XII

fftStifc •Drainage Pipe

Vr.fi-Wj'

Fig. 84. —Arrangement of 8'X8' Sluice-gates, Stevens Creek Development

on Savannah River.
Art. 71] REGULATION OF HIGH-WATER SURFACE 223

Gates of this type should be made unusually strong in every

particular. The disturbance of the water at part gate opening
seems to impose a duty on the several parts far greater than the
capacity of the manufacturer's usual stock patterns.
Considerable trouble may be caused by logs lodging in the sluice
and preventing the gate from closing. Rack bars of 6- to 12-in.
spacing are usually placed above the gates. They should be as
far as possible from the gates, as cases have been known where
logs, although having caught on the racks, have extended through
them far enough to rest on the gate sill. For this reason some
engineers incline to the opinion that racks are a menace rather
than a protection, because, without the racks, the logs would have
been carried through the sluice. The use of racks, however, is
still quite common. There is a great deal of room for improve-
ment in this feature, as the gates are relatively inaccessible, and,
therefore, expensive to maintain and repair.
Fig. 85 shows a typical system of flash-boards with which the
normal spillway crest is lowered automatically at times of excess
flow. In this country, this is the most common of all devices
designed to control the elevation of flood-water surfaces. Although,
in certain cases, flash-boards are rather expensive to maintain,
owing to frequent renewals, they have the advantage of relatively
small first cost and the certainty of automatic operation.
The flash-boards are supported by round steel pins which are
inserted loosely into sockets set in the masonry crest of the dam.
The boards are usually fastened loosely by wire nails bent around
the pins, or by wire loops passing around the pins and fastened to
nails driven in the boards. The pins are designed with a reason-
able factor of safety, for water at full storage level, but are cal-
culated to bend over and clear the masonry crest, if the boards
are not removed, when a severe flood occurs. After the flood
passes, the pins are heated, straightened and replaced, and new
boards put in place. The boards are usually fastened loosely
enough to be removed by the flood, and are seldom recovered.
When and sufficient time
increases of flow can be anticipated
is and pins may be removed by hand before
available, the boards
the water surface rises high enough to bend them over. In order
to facilitate the handling of a barge for removing and restoring
the flash-boards, it may
sometimes be found desirable to provide
sockets at intervals into which mooring pins may be set. These
224 DETAILS AND ACCESSORIES [Chap. XII

a -a

1 I
Abt. 71] REGULATION OF HIGH-WATER SURFACE 225

are indicated in Fig. 85.The boards and pins are sometimes

manipulated from an overhead platform or bridge provided for
that purpose.
Where it is possible to remove the boards in advance of floods*
they are usually built in panels, as indicated, and provided with
handles. Handles of the type shown are not permissible if con-
siderable drift is anticipated, as this may collect and cause pre-

mature failure.
The boards are sometimes provided with planed edges in order
to reduce leakage, but more often are unplaned, and ashes or sim-
ilar caulking materials are swept along the joints.
As indicated in Fig. 85,

Let d = the diameter of solid pins, in inches;

s = the spacing of solid pins, in feet;
hi — the height of solid pins, in feet;
^2 = depth, in feet, of water above the masonry crest;
Tim = the depth, in feet, of water above the masonry crest
corresponding to the maximum flood;
/= the bending stress in the pins, in pounds per square
inch;
W2 = weight of one cubic foot of water = 62.5 lb.

The bending moment in each pin for the water surface at full
storage level is

M = ^|^
o
= 10.42/n 3
s . . . ft.-lb.

Within the elastic limit, the moment of resistance of the pin, in

foot-pounds is,

0.00818/d3 .

Equating this to the bending moment, and solving for d, there

results,
,
3
ll275hi 3 s

The use of this equation will provide for the pins having a rea-
sonable stress, say two-thirds of the elastic limit, when the pond
is at full storage level.
226 DETAILS AND ACCESSORIES [Chap. XII

Before the pins are completely bent over, they are stressed far
beyond their elastic limit, and on this account the ordinary theory
of flexure is quite inaccurate. Because of this and other uncer-
tainties, it is impossible to write an -accurate equation to indicate
at what stage of water surface the pins will bend over, and, there-
fore, we must rely on experimental data. The author has found
that, if pins of medium steel are stressed theoretically to about
two-thirds of their elastic limit when the water surface is at full

storage level, they will ordinarily not withstand a head, above h,%,

the masonry crest greater than twice their height, and that they
usually bend over for heads equal to from 1.5 to 1.75 of their
height, depending on the vacuum which forms under the overflow.*
To be amply safe, the height of the flash-boards should not or-
dinarily be more than about 40 per cent of the head, hm, corre-
sponding to the maximum flood. This will leave a margin of
about 0.2hi between maximum flood level and the highest level at
which it is thought the pins will fail.
The pins may then be expected to accommodate, without fail-
ure, a flood equal to about one-sixth to one-eleventh of the maxi-
mum expected flood, and, if not removed when necessary, will
usually bend over several times in the average year unless a part of
the moderate floods can be accommodated through sluices in the
dam, through turbines, or by other means, or unless the reservoir
is exceptionally large.
If a greater degree of accuracy is desired, it becomes necessary

to test out samples of the pins under actual operating conditions.

Experiments of this character are inexpensive, if a log chute,
wasteway has been, provided in the crest of the dam.
fish way, or
In any event, a considerable expenditure will be warranted if a
large reduction in land to be purchased can be effected. For
greater accuracy, the pins are sometimes grooved to exact diam-
eter at the elevation of the masonry crest.
Pipes have sometimes been used in place of solid pins. If the
pipe pins are short in proportion to their diameter, they may fail
by buckling at their supports, and go out sooner than solid pins
having the same initial stress.
Manytypes of automatic hinged flash-boards have been used,
in which actualfailure of the material does not occur when the
flash-boards collapse. Their use, however, has not become
com-
* See Art. 17.
Abt. 71] REGULATION OF HIGH-WATER SURFACE 227

mon, except on some of the very low river improvement dams built
by the Federal Government.
A radically different type which seems to be gain-
of regulator,
ing favor, is that of the siphon spillway.
shows the siphon
Fig. 86
spillway of the Tennessee Power Company on the Ocoee River,
Tennessee.*
In a siphon spillway, the water flows through a closed conduit,
producing a suction head that largely increases the velocity, and
consequently the discharge per unit area.
Under normal conditions, the water, at the Ocoee Siphons,
stands at El. 1089.2, or slightly trickles over the lip or crest of
the spillway. When the water surface rises, the discharge over
the crest increases and the water strikes the far side of the

Fig. 86. —Siphon Spillway, Ocoee River Development, Tennessee.

lower siphon legs. When the water surface reaches an elevation
of about 1089.55, or slightly above the top of the air inlets, the
air thus confined in the top of the siphons, as well as that in
the lower legs, is quickly entrained and ejected and the siphons
prime. The suction thus produced increases the velocity to that
corresponding to an effective head equal to the difference in ele-
vation between the water surface in the pond and the center line
of the siphon outlets, less the head expended in friction within the
siphons. If the full discharge of the siphons is greater than the
total flow, the water surface in the pond will recede. When the
upper parts of the air inlets become exposed, the air drawn in by
the suction reduces the efficiency of the siphons until the dis-
charge is automatically diminished to that required for stationary
* Engineering Record, Vol. LXXII, p. 567.
228 DETAILS AND ACCESSORIES [Chap. XII

Pig. 87.—Down-stream View of Ocoee River Siphon Spillway.

Fig. 88. —Up-stream View of Ocoee River Siphon Spillway.

Art. 71] REGULATION OF HIGH-WATER SURFACE 229

water surface in the pond. If a sufficient quantity of air is drawn

through the inlets, however, the siphonic action will be broken.
The water surface in the pond will then rise again, and the oper-
ation of priming will be repeated.
By virtue of their large capacity, relative to that of a simple
weir, of the same crest head, siphons may be used to limit the rise
of the water surface in the pond during floods, and are particularly
serviceable in cases where the fluctuations in discharge are rapid,
as at the end of a long flume supplying hydraulic turbines.
If properly proportioned, they will prime quickly enough for
all practical purposes.
The discharge of a siphon of this type may be computed from
the ordinary equation for flow through short tubes,

Q = CAV2gh,

where ft = the gross head on the siphons, in feet from the water
surface to the center line of the discharge;
A =the area at the throat, in square feet;
</= the acceleration of gravity = 32.2;
Q = the discharge, in cubic feet per second; and
C = a coefficient depending on the characteristics of the
siphon.

Tests by the author on the Ocoee Siphons indicated a value of

C of about 0.65, and this figure has been fairly well substantiated
by other tests on similar siphons.
Provision should always bemade to by-pass ice cakes and
debris which, otherwise, might clog the siphons. The throat
should be protected from erosion by a cast-iron lining,
 INDEX

A
PAGE
Ambursen Dams 132
Appalachian Pr. Co. Dam 131
Aprons 184
Arch Dams 13, 14, 148
Deflection of 169
Details of 160
Examples of 165
Ratio of slenderness 163
Recommendations for design 158
Reliability of 159
Reinforcement for 162
Spillway section of 161
Stresses in existing 159
Theory of constant angle arches 160
Architectural treatment 211
Atcherley's theory of tension in vertical planes 56
Atmospheric pressure 35, 105
Au Sable River Dam 188

B
Baffles at toe 25, 26, 181
Barra Weir 186
Barren Jack Creek arch dam 169
Bazin's experiments on weirs 106
Borings. See Grouting 5,7
Bucket at toe 115

C
Coffer Dams 1, 5, 15
Compression
Distribution of on joints 42
In various dams 54
Rule 3, equations for 71
Rule 3, governing 52
231
 232 INDEX
PAGE
Concrete
For dams „ 203
Strength of 55
Weight of 41
Construction plant location 3, 5
Contraction joints 206
Contraction of crest discharge due to piers and abutments Ill
Coon Rapids Dam 191
Core-walls. See Cut-offs.
Crest gates 217
Crest, shape of 105, 108
Crests, movable 214
Cross River Dam 31, 38
Croton Falls Dam 38
Cut-offs
General 30
In earth 183
In rock 174

D
Dams
Arch 13, 14, 148
Earth and rock 14
13,
Hollow gravity 12, 13, 132
Solidgravity 12, 148
Timber 12, 15
Discharge capacity of spillways. See also Floods 110, 114
Drainage of
Dams 210
Earth foundations 189
Rock foundations 180
Drainage systems '
30, 180, 189, 210

E
Earth dams .'.
13, 14
Earth foundations 181
Prevention of percolation through 184
Strength of 56
Earth pressure 33
Earth, weight of 34
Elephant Butte Dam 31, 104, 216
Embankments 13, 14
Equations for design 60
Rule 1, for overturning 64
Rule 2, for sliding 70
Rule 3, for compression 71
Rule 4, for tension in vertical planes 75
 INDEX 233

PAGE
Estacada Dam 54, 182
Examples of arch dams 165
Examples, general. See Contents ix

F
Factors of safety 57
Failures, causes of 48
Flashboards 223
Flood regulators 214
Floods 5, 194
Fuller's equations 196
Of record 196, 197
Forces acting on dams 19, 65
Foundation leakage 32, 174
Foundation mattresses 135
Foundations 1, 4, 7, 172
Drainage of 180
Earth 181
Leakage in 32, 174
Preparation and protection of 172
Rock 172
Strength of 56
Uplift in 25
Friction, coefficient of on joints and base 50

G
Gates
Crest 217
Sluice 221
Tain'or 217
Gatun Dam 28, 181
Government requirements 9
Granite Reef Dam 186
Gravity dams 60
Great Falls Dam 217
Grouting foundations 174
Of Lahontan Dam 177
Of Mathis Dike Dam 183
Guyabal Dam 139

H
Highways to site 3,6
Hollow dams 12, 13, 132
Ambursen type 132
Multiple arch type 164
Uplift on 32
Holyoke Dam, vibrations in 37
Hydraulic jump. See Standing wave.
234 INDEX

I
'page
Ice, floating 3,39
Ice pressure 37, 79
Reinforcement for 120
Impact from approaching water 23
Investigations 3
'.

Final 6
Preliminary 4

J
Jamrao Wier 186
Joints, contraction 206
Horizontal, treatment of 205

K
KensicoDam 31, 38, 210
Keokuk Dam crest gates 131, 218

L
Laguna Dam 189
Lahontan Dam, grouting foundations for 177
Las Vegas Arch Dam 169
Leakage in foundations 32, 174
Location of dams 1
Lock Raven Dam 31

M
Masonry for dams 203
Masonry, strength of 55
Weight of 40
Mathis Dike Dam 133, 183, 193
McCall Ferry Dam 131
Middle third theory 43
Morrison and Brodie dams 54, 104
Movable crests 214
Multiple arch dams 132, 164

N
Nappe, shape of for weirs 105
New Croton Dam 31, 38, 54, 104
New South Wales arch dams 166
Nomenclature
For arch dams 14g
General ,
,.,,.,,.,.,, 16.
 6

INDEX 235

PAGE
Non-overflow dams
Comparison of 104
Solid 62,78
Superelevation of 78
Width of top of 78
Zones in 62
North Crow Arch Dam 166

O
Ocoee Dam 131
Ocoee Siphon Spillways 227
Olive Bridge Dam 31, 38, 104
Overturning 48
Rule 1, equations for 64
Rule 1, governing 49

P
Parr Shoals Dam 129, 131
Percolation in earth foundations 184
Percolation, prevention of 184
Perviousness of foundations 32
Piers on crest Ill
Piers, contractions due to Ill
Piling
Bearing 189
For cut-offs 183
To prevent sliding 193
Pressure. See Compression.

R
Railroads to site 3,

Rating curve 196

Reservoir empty, meaning of 84
Resultants in arch dams 164
Resultants, inclination of
Rule 2, equations for. ,...., 70
Rule 2, governing 50
Resultants, location of
Rule 1, equations for 64
Rule 1, governing 49
Dam 190
Rio Grande
Rock dams 13, 14
172
Rock foundations
Strength of ,,,.,, :
. ,
, §?
 236 INDEX
PAGE
Rule 1, overturning 49
Rule 1, equations for 64
Rule 2, sliding 50
Rule 2, equations for 70
Rule 3, compression 52
Rule 3, equations for 71
Rule 4, tension in vertical planes 56
Rule 4, equations for 75
Rule 5, margin of safety 57
Rule 6, details and construction 58

S
Safety, margin of 57
Salmon Creek Dam 163, 165
Salmon River Dam 215
Seepage in earth foundations 184
Shear 47
'.

Shoshone Dam 149

Silt deposits 3
Silt pressure 33
Siltremoval 3
Silt,weight of 34
Siphon spillways 227
Site, choice of 1
Sliding 48
Rule 2, equations for 70
Rule 2, governing 50
Solid gravity dams 12
Design of 78, 105
Spillway dams 105
Comparison of 131
Discharge capacity of 110, 114
Shape of crest 105, 108
Zones in 64
Spillway, section for arch dams 161
Standard dam crest 105
Discharge capacity of 114
Standing waves 23, 124
Effect on foundations 181
Stevens Creek Dam sluice gates 221
Storage 8
Stresses
Allowed for arch dams 165
In existing arch dams. »See also Compression 159
Submerged weirs 115
Surveys 4, 6, 8
 INDEX 237

T
PAGE
Taintor gates 217
Tension
In joints and base 44, 49
Rule 1, equations for 64
Rule 1, governing 49
In vertical planes 56
Rule 4, equations for 75
Rule 4, governing 56
Test pits 5, 7
Three Miles Falls Dam 137
Timber dams 12, 15
Type of dam, choice of 11

U
Uplift 25
Affected by silt deposits 25, 35

V
Vacuum under sheet of spilling water 105
Velocity head of approaching water 23
Velocity of approach affecting weir discharge 113
Velocity of approach, impact due to 123

W
Wachusett Dam 31, 38, 54
Water pressure 19
Waterproofing dams 205
Water, weight of 19
Wave pressure 40
Wegmann's dams 54, 104
Weight of
Concrete 41
Dam 40
Foundation 42
Masonry 40
Silt 34
Water 19
Weirs
Discharge capacity 110
Form of falling sheet of water from 105
Submerged 115
Wollongong Dam 168

Z
Zones, designing 62
Warn

:''•'*'
lllMBIlMMfl
Base

WBH

HtHKKH