CONTINUOUS FRAMES

OF REINFORCED CONCRETE

CHAPTER I
INTRODUCTION

General Considerations on Continuity. Continuity affects very much

the moments in a structure and affects appreciably the shears and re
actions. It does not require any great amount of mathematics to see
that this is true. It is a matter of common observation that, where a
beam is rigidly connected to another beam or to a column, bending in
one member will produce bending in the others. In structures of rein
forced concrete, reinforcing steel must be provided for the forces pro
duced, otherwise cracking will lead to unsightly appearance and pos
sibly to failure of the structure.
In many cases we can sketch the shape taken by the deflected struc
ture, see where cracking might occur, add reinforcing steel to prevent
If,

it and be satisfied. however, the structure important and we

is

wish to know accurately how much steel or concrete needed, we will

is

make more thorough investigation. But important to recognize

is
it
a

clearly that we do not want to carry this investigation beyond the point
where will influence the design; we must always remember that what
it

we want a structure,
not merely an analysis.
is

It almost necessary to study our subject in the order: methods of

is

analysis; a consideration of physical constants; consideration of the

a

application of these values to design. But the engineer must often

approach the subject in the reverse order; certainly its phases assume
importance to him in reverse order. First, what can he do with these
analyses; then what value shall he choose for the physical constants;
finally how will he analyze it? difficulty in the past has been that
A

the student approached the matter in the succession of analysis, con

stants, design. The practical designer was forced to approach the
subject in the order of design, constants, analysis. The student became
so involved in methods of analysis that he never got far enough to con
sider the uncertainty of physical constants or the elements of judgment
involved in design. The busy designer, knowing well enough what
1
2 INTRODUCTION

values he wanted for his design, recognizing quite well the uncertainty
of the physical constants, never had time and frequently had little in
clination to devote much effort to analysis.
One of the greatest problems in this field, as it has been in all fields of
structural work, is to simplify the principles at both ends so as to make
it possible for a man to correlate his knowledge. The analysis of a
structure for continuity should be less complicated than the determina
tion of anchorage and stirrup spacing under some specifications. Many
writers on the subject of reinforced concrete seem to have held a strange
philosophy which attempts to substitute exact rules in every detail for
the judgment of the designer and builder. Some rules must check judg
ment, but there is not and will never be any substitute for the sound
judgment of the designer. All that we can hope to do is to furnish simple
and usable tools of analysis as aids to that judgment. Analysis must i
be thought of as a guide to judgment, not as a substitute for it. •

In the design of an important indeterminate structure, two analyses

of the forces and moments are usually needed. The first analysis is
necessarily more or less approximate, and it is essential that it be made
quickly. The structure designed from this analysis should usually be
so near the final structure that the designer can proceed to detailing
and to estimates of quantities. A final analysis is needed to check the
stresses in the structure as designed.
Of these two analyses the preliminary one requires greater skill.
It need not be very precise, but it should not overlook any important
factors. The effect of minor variations in proportions and of the less
important elements in analysis may be determined in the final analysis.
Analysis of the shears and bending moments in a continuous structure
implies knowledge from three sources. First and foremost are the laws
of statics; anyone who does not clearly understand these laws and their
application to structural analysis is wasting time in attempting to study
continuous structures at all. Second are the facts of geometry; these
appear in various forms in the literature of indeterminate structures,
but always the fundamental geometrical facts are very simple and very
familiar. Finally, the properties of the materials enter to some extent
into the problem; Hooke's Law is usually assumed, and the moduli
of elasticity of the component parts of a structure are involved.
In most problems of determining the moments, shears and reactions,
the properties of the material are not very important; the principles
involved are nearly all from statics and geometry, matters about which
there can be no question. In the design of the structure to resist these
forces we are, of course, concerned very largely with the properties of
the materials. If we can accurately picture the deformation of the
 PHYSICAL CONSTANTS INVOLVED IN ANALYSIS 3

structure under load, we can analyze it by statics alone. The methods

of analysis serve to make this picture accurate; but unless, after we
have analyzed a structure, we can draw its deflected shape, we do not
know what we are talking about. Moreover, the exact analysis is no
more scientific — though it is more exact — than is the elementary
process of drawing a consistent picture of the deformed structure by
sketching.
No indeterminate analysis — no structural analysis of any kind —
is complete until the computer has satisfied himself:

(1) That the forces balance, at least within the accuracy of compu
tation used.
(2) That he has not overlooked any forces.

Physical Constants Involved in Analysis. An important field of

study is that of the sensitiveness of the structure to variations in the
properties of the materials or to other underlying assumptions. All
analyses are based on some assumptions which are not quite in accord
ance with the facts. From this, however, it does not follow that the
conclusions of the analysis are not very close to the facts.
Three sets of elementary physical constants are involved in the
geometric relations used in the analysis of continuous structures.
These are: the change of length, per unit of force per unit of length, in
members subject to axial forces; the rotation, per unit of moment per
unit of length, in members subject to bending moments; the rotation,
per unit of torque per unit of length, in members subject to torsion.
Probably these constants should have some special name, although
none has ever been assigned to them. Perhaps
" constants of deforma
tion " is as good a term as any.
Understand clearly that these are actual deformations which occur
in the structure under the conditions which we are discussing. It will
be found in reinforced concrete that they are not invariable for a given
member. The rotation, in a given length of concrete beam subject to a
given bending moment, per unit of length per unit of moment, will vary
with the dimensions of the section, but it will also vary with the magni
tude of the moment and with the duration of the moment. This makes
it a rather complicated quantity to deal with. These variations are
not small. In a given beam of reinforced concrete we may reasonably
expect that if the rotation per unit of length per unit of moment is a
for a small moment, it may be 2a for a large moment. These values
may be doubled if the moment persists for one year, and will increase
for any duration of loading.
Fortunately, it is usually the relative values of these deformation
4 INTRODUCTION

constants, not their absolute values, which are needed. If, however,
we wish to find the moments for a definite displacement produced by
shrinkage, by change of temperature, or by movement of abutments,
we need to know the absolute values.
What is the value of E to be used? What is the value of / to be used?
The very definite answer to these questions is that we do not know
enough to assign very definite values. The exact values of /, the exact
values of E are subject to great variation and are to no small extent a
matter of chance. We can only investigate the probable effect on the
design of such variations as may reasonably be expected and choose
such values as will give conservative design. There is nothing unique
in this situation. It occurs in practically all engineering design.
Reference to the elementary geometry presented in Chapter IIIwill
show that what we are dealing with in structures of reinforced concrete,
is not E or /at all; the quantity with which we are dealing is the ro
tation per unit of moment per unit of length which exists in the struc
ture in question under the loading conditions in question and for the
duration of load being considered. Iis simply a myth here.
The whole subject of the values to be used for E and especially for I
is at present highly controversial. Space does not permit a review of
current investigations along these lines, and the discussion here is re
stricted to very brief statements of the problems involved and of what
seem to the authors the most reasonable conclusions.
What is E? At one time it was thought that the modulus of elas
ticity of concrete at rupture varied approximately with the ultimate
compressive strength, and present codes and specifications are based
on this view. There no doubt, some relation between the two val
is,

ues, but by no means consistent. The value of varies in what at

E
is
it

present seems to be an erratic way over range from 2,000,000 lb. per
a

sq. in. to 4,000,000 lb. per sq. in. Occasionally values as low as 1,500,000
and as high as 5,000,000 may be found, but in general we may expect the
variations to be limited to ±33 1/3% from mean value of 3,000,000.*
a

This the value at rupture. But the ratio of stress intensity to

is

strain varies both with the intensity of stress and also with the duration
of the stress. The authors see no reason to think that either variation
affects seriously the moments produced by loads on the structure, but
on this point also there some difference of opinion.
is

See "Reinforced Concrete Columns in Flexure," Hardy Cross, Proc. A.C.I.,

*

Vol. XXVI (1930); also discussion by N. H. Roy and F. E. Richart, and that by
W. M. Dunagan.
See "Flow of Concrete Under the Action of Sustained Loads," Hardy Cross,
t

Jour. A.C.I. December, 1931.

,
 EFFECT OF UNCERTAINTIES AS TO PHYSICAL CONSTANTS 5

We have as elements in the problem the ordinary, long-recognized

variation of the modulus of elasticity, uncertainties as to the distribu
tion of stresses on plane sections in T-beams and in other cases, uncer
tainties as to the location and effect of local cracking, and uncertainties
introduced by the effect of time yield. We have elements of uncertainty,
then, as to the exact mechanics of internal stress, we have elements of
chance, and we have variation in the properties of the material with
time and with stress intensity. The problem of analysis of a continuous
structure now seems somewhat hopeless. The statement frequently
made that structures of reinforced concrete are not truly elastic in the
sense that they do not have complete elastic recovery is perfectly true
and entirely immaterial. But the fact that structures of reinforced
concrete are not truly elastic in the sense that we cannot predict with
accuracy and certainty the exact angle change per unit of moment per
unit of length from one section to another of a beam is also true and is
very pertinent.
What is /? Various values have been suggested for I. For rec
tangular sections some use the moment of inertia of the full transformed
section, some neglect the reinforcement and use 1/12-Acd2, where Ac is
the total area of concrete and d is the full depth. Some have suggested

It will that the exact value is not very important, and

be seen later
hence many approximations may legitimately be used. The authors
recommend the moment of inertia of the full transformed section as the
most satisfactory standard.
The greatest uncertainty arises in determining the value of /
to use
for beam and slab construction. Some use the moment of inertia of
the full transformed section consisting of the beam and the connecting
slab for a width extending halfway to the next parallel beams. This
seems the most satisfactory treatment and is recommended. Many
variations have been suggested, however.
Effect of Uncertainties as to Physical Constants. The effect of the
girders on the columns is much greater than the effect of the columns
on the girders. Moments in the girders are the sum of fixed-end mo
ments and distributed moments as explained later. Fixed-end moments,
though subject to some uncertainty, are quite definitely known. The
column simply affects the distribution of the unbalanced moment and
this unbalanced moment is only a part, and often a small part, of the
total moment in the member. But in general, the column has no trans
verse load and no fixed-end moment and the only moment in it is the
distributed moment which comes to it as an effect of loading the girder.
There is good reason to believe that the moments produced by loads on
6 INTRODUCTION

the girders can be determined with reasonable accuracy, but there

seems equally good reason to' doubt whether the moments in the col
umns can be determined with much accuracy. Nevertheless an effort
should be made to include the column effects in the analysis.*
Problems in Continuity of Structures of Reinforced Concrete. The
most common problem in the design of continuous structures of rein
forced concrete is to find the curves of maximum moments and some
times of maximum shears on a series of continuous girders, usually
rigidly connected to columns, for dead load and for five load uniformly
distributed and applied to the girders either directly or through joists
or floor beams equally spaced.
Commonly the girders and columns are of uniform section for their
full length except as this is necessarily modified by the existence of the
joint. Sometimes the girder sections are varied by end haunches, which
may be either straight or curved.
For bridges it is usually necessary to make analyses for moving sys
tems of concentrated loads.
Wind stresses in buildings are usually determined
by approximate
" exact "
methods, but there is at present an increasing interest in more
analyses.
The design of arches of reinforced concrete is a problem of a different
order from those just outlined. It is not difficult, but it requires a good
deal of sound judgment. The analysis of continuous arches on slender
piers is a special problem.
Continuous slabs are so distinct a field of study that they have not
been discussed in this volume.
Methods of Analysis. The chief methods of thought used in the
book are the methods of the deflected structure and of pressure lines in
Chapter II, the method of moment distribution in Chapter IV and later
of distribution of shears in Chapter VII. The column analogy is used
in Chapters III, V, VII and IX. Influence lines are treated in Chapter
VIII. The relative importance of these methods is probably in the
order given.
These tools are in themselves not very productive; their usefulness
depends on the skill and resourcefulness of the man who uses them.
Too many students of indeterminate structures hope to progress by
acquiring an endless variety of tools and are so busy doing this that they
never learn how to use them.
The ability of a designer of continuous structures is measured chiefly
by his ability to visualize the deformation of the structure under load.
* See "Continuity as a Factor in Reinforced Concrete Design," Hardy Cross,
Proc. A.C.I., Vol. XXV.
 RECOMMENDED ORDER OF STUDY 7

If he cannot form a rough picture of these deformations when he begins

the analysis he will probably analyze the structure in some very awkward
and difficult way; if he cannot picture these deformations after he has
made the analysis, he doesn't know what he is talking about. The
more or less gentle reader may find the constant repetition of this theme
monotonous, but it is the deliberate conclusion of the authors that the
most important aspect of the subject is the simple picture of structural
deformation.
Recommended Order of Study. The following order of study is
suggested to independent students of the subject. Although it reverses
some of the logical steps in demonstration, it gives results rapidly. An
experienced teacher will vary the procedure because he will be able to
explain the bearing of theory on design and of design on the relative
importance of various steps in analysis.
It is assumed that the student is quite familiar with statics and
reasonably familiar with the routine procedure of reinforced-concrete
design. Even such students may find Chapter II
on statics of interest.
(1) Learn the elementary process of moment distribution, Chapter
IV, accepting its theorems temporarily as proved. Practice simple
moment distribution restricted to frames made up of members of con
stant section and apply it. Then learn to balance the shears for side-
sway by methods explained in Chapter VII.
(2) Draw many curves of maximum moments for uniform loads,
using different span arrangements, different relative stiffnesses of col
umns and girders, different ratios of live to dead load. (See Chapter
VI.) Use merely the elementary procedure of constructing the curves
of moment for individual spans loaded and combining as indicated for
maxima.
(3) Make similar studies for maximum end shears.
(4) Run through the approximate design of several of these structures
with a view to observing the effect of continuity on the design.
(5) Haunch the beams in some of these problems, using the approxi
mate methods of treating haunches and the curves given in Chapter V,
and observe the effect of the haunches on the design.
(6) Proceed to a careful study of the geometry on which the whole
procedure is based, the geometry of flexure. Chapter III.
(7) Consider the uncertainties of physical data involved in the
analysis as indicated in this chapter and elsewhere.
(8) Proceed to exact studies of sidesway and of stresses in general
produced by transverse loads. Chapter VII.
(9) Study arches and other special problems. Chapter IX and else
where.
 CHAPTER II
STATICS OF DEFLECTED STRUCTURES

In this and in the next chapter the tools of analysis are exhibited and
their use is explained. First place is given to the procedure of visua
lizing the shape taken by the deflected structure under load. This
cannot be overemphasized.
Next we take up the statics of indeterminate structures. Students
familiar with statics but not with its application to indeterminate
structures overlook many of its important implications in this field,
some of which amount almost to independent theorems.
Importance of Visualizing Deformations. One of the first objects
of the analysis of indeterminate structures is to see how the structure
deforms. If the
analysis fails to give the designer a clear mental picture
of the action of the structure under load, it is dangerous to proceed
with the design. This is especially true of reinforced-concrete struc
tures where bars are often placed on only one side of a section; flexure
which produces tension on the other face may cause failure.
It is possible by training, by simple crude models, and by computa
tion, to develop imagination in visualizing the function of each part of
a structure in resisting the load. It
is then possible to sketch directly
the shape of the deformed structure. If this can be accurately done,
the forces, moments and shears can all be computed by statics. This
'
procedure is scientific in principle but, of course, lacks precision. It
is always the first thing to be tried in designing a new type of structure,
and it is the last thing that should be done in all cases after an analysis
has been made, in order to give meaning to the computations. Many
times the results obtained from sketching the deformed structure,
though somewhat lacking in precision, will be sufficiently accurate to
satisfy the needs of the designer.
What is eruditely called the theory of elasticity deals with the simple
fact that, if a structure does not break, it holds together. The word
" theory " here is used in the sense of an assembly of systematically
arranged facts and not in the sense of scientific hypothesis. Practically
all of what constitutes the theory of elasticity — it might better be
called the theory of continuity — is simply a statement of certain geo
metrical relations which must exist in order to have continuity preserved.
The relations between the forces and the deformations are based on
8
 EXAMPLES OF STUDY SKETCHING 9

certain assumptions, which may be open to debate, as to properties of

materials. The geometrical relations themselves are open to no debate.
The best method of analyzing most indeterminate structures is to
find the forces which would exist under a simple condition of restraint
and then to determine the effect of removal of the restraint. This is
true in picturing the deformation of the structure, and it is usually true
in devising a mathematical procedure for exact analysis.
Examples of Study Sketching. Consider a fixed-ended beam with a
moment load applied at its center as shown in Fig. 1. It is obvious

Fig. 1. Deformation of a Fixed-ended Beam — Moment Load at Center.

that the axis of the beam will in the direction of

be turned at the center
the moment. Moreover, there will be no deflection at the center be
cause of symmetry, for if the forces on the right caused a downward
deflection of the center, the forces on the left would cause an equal, but
contrary, upward deflection. Since there is no deflection at the center
and since the axis must remain horizontal at the ends, the curve of

Points \of Inflection if

Points of Inflection Fixed at Ends
as a Continuous
Beam

Fig. 2. Deformation of a Fixed-ended Beam and of a Continuous Beam.

deflection shown in the figure may be drawn, having one point of in

flection on each side of the load.
This method of studying the problem is powerful and leads directly
to a sufficiently accurate solution of the problem by statics.
In the case of a continuous beam, there is at the ends of a loaded span
an obvious tendency to rotate the axis since the beam " tries " to
straighten out at that point. We can sketch the deflected fixed-ended
beam and, by modifying this, visualize the deformation of a continuous
beam. This is done in Fig. 2.
10 STATICS OF DEFLECTED STRUCTURES

After a little practice it is not difficult to sketch the simple cases of

deformed structures.
In Fig. 3 (a) is shown a symmetrical rectangular bent, subject to
transverse load. If the girder is very stiff it does not bend appreciably.
It is easy to see that the columns are deflected sidewise, remain vertical
at the top and at the bottom, and hence are symmetrically bent, and
have their points of inflection at the middle. From this a statical
solution follows as indicated in (b).
If the girder is not extremely stiff, it will bend at each end as shown
in (c). Considerations of symmetry indicate that the point of inflection
of the girder is at its midpoint, and a study of the sketch will show that
the points of inflection in the columns have moved up above the center.

_ [*— L—
—H
P | Girder Rigid y j

rr

**■
(a) (b)
Fig. 3. Deformation of a Symmetrical Bent.

Obviously, if the girder were absolutely flexible, this would

be equivalent
to hinge connections at the tops of the columns, and the points of in
flection would rise to the top. Similarly, if the foundation rotates
owing to the reactions, the point of inflection in the column will move
down, and as a limit will be at the bottom of the column when perfect
freedom of rotation exists.
This sort of study may seem extremely elementary, but it will pro
duce quite as satisfactory results in many cases as will elaborate analysis.
Most rectangular bents do not justify much mathematical analysis
because the degree of fixation of the foundation is almost invariably
uncertain. We can only strike a balance between the influences tend
ing to raise and to lower the point of inflection in the column.
As another simple illustration, consider a beam fixed at the ends and