H

ILLINOI

UNIVERSITY OF ILLINOIS AT URBANA-CHAMPAIGN

PRODUCTION NOTE
University of Illinois at
Urbana-Champaign Library
Large-scale Digitization Project, 2007.

ABSTRACT
A GENERAL METHOD OF ANALYSIS OF
PRESTRESSED CONCRETE BEAMS AT ULTIMATE
IS

PRESENTED.

EQUATIONS FOR THE

DETER-

MINATION OF ULTIMATE MOMENT ARE DERIVED.

SIMPLIFIED EXPRESSIONS ARE PRESENTED FOR
ULTIMATE MOMENT FOR USE IN DESIGN.
METHOD

IS PRESENTED BY WHICH PRESTRESSED

CONCRETE BEAMS CAN BE DESIGNED ON THE

ACKNOWLEDGMENT

BASIS OF STRENGTH AND DUCTILITY.

PLES OF DESIGN ARE CONSIDERED,

The
for the

authors

extend their thanks

comments and

suggestions of

ONLY

EXAM-

IN WHICH

THE REQUIRED DUCTILITY VARIES.

COMPARISON OF THE

THREE EXAMPLES

SHOWS

Professor K. Preiss, University of

THE

Illinois,

ON THE DIMENSIONS OF THE BEAM, AND THE

Carl E. Thunman, Jr.,

Illinois Division of Highways,

and

INFLUENCE OF THE REQUIRED DUCTILITY

INFLUENCE OF COMPRESSION STEEL ON DUC-

Professor J. L. Merritt, University

TILITY.

of

AT SERVICE CONDITIONS ARE CALCULATED

Illinois, who reviewed the manu-

script.

THE STRESSES AT TRANSFER AND

AND TABULATED.

CONTENTS
I.
II.

III.
IV.

V.

VI.
VII.

CALCULATION OF ULTIMATE MOMENT

INTRODUCTION

A.

Assumptions

B.

Flexural Strength of Sections in which

the Neutral Axis at Ultimate Falls in
the Flange . . . . . . . . . . . . . .

C.

Flexural Strength of Section in which

the Neutral Axis at Ultimate Falls
Below the Flange . . . . . . . . . . .

D.

Summary of General Equations

E.

Calculation of Ultimate Moment by the

Properties of Stress Block . . . . . .

F.

Provisions of the American Concrete

Institute Building Code
. . . . . .

PROVISIONS FOR SAFETY AND DUCTILITY IN

ULTIMATE DESIGN
. . . . . . . . . . .

12

ULTIMATE DESIGN OF SECTIONS WITHOUT COMPRESSION REINFORCEMENT . . . . . . . . .

l1

A.

Method of Analysis Used

l4

B.

Determination of Area of the Beam

C.

Illustrative Example 1

15

15

ULTIMATE DESIGN OF SECTIONS WITH NONPRESTRESSED COMPRESSION STEEL

. . . .

A.

Method of Analysis Used

B.

Illustrative

C.

Comparison of the Three Solutions

Example 2

SUMMARY AND CONCLUSIONS

REFERENCES

19

19

20

22

23

2h

FIGURES
1.

Idealized I-Section

2.

Flanged Section, Neutral Axis in the

3.

Flanged

Section, Neutral Axis

below the Flange

4.

Flanged

Section, Neutral Axis

in the

5.

Flanged Section, Neutral Axis

6.

Relationship
the Section

7.

Stress-Strain Diagram for Steel

8.

Section of a beam with low Ductility

Weight of Beam = 300 lb/ft

9.

between

Flange

Flange

below the

Flange

p and Geometric Parameters of

= 0.01

su

Section of a Beam with very High Ductility

Weight of Beam = 377 lb/ft

= 0.03) -

(e
su

10.

Section of a Beam in Which the High Required Ductility

(E
= 0.03) is Obtained by the use of Compression
Reinforcement - Weight of Beam = 300 lb/ft

11.

End

12.

Elevation of Beam of

Section of the Beam with Compression Reinforcement

Example 3

NOTATION

distance from the neutral axis to

the top fiber

F(e

gross cross-sectional area of the

beam

A's

=
=

sr

=
=

f , equation of the stress-strain

diagram of prestressed steel

f(e)

stress in the concrete, equation of

the stress-strain diagram of concrete

area of prestressed steel

area of non-prestressed compression

steel

cylinder strength of concrete at 28

days

stress in prestressed steel at

failure

ultimate strength of prestressing

steel

area of prestressed steel that dein

velops the rectangle (b - b')t
the compression zone

ft
su

stress in non-prestressed compression steel at failure

yield point of non-prestressed

compression steel

f - 0.85 f'

G(e'su )

==

fIsu , equation of the stress-strain

diagram of non-prestressed compression steel

over-all depth of the beam

ratio of the width of bottom flange

to that of top flange

kI

ratio of the average to maximum

stress in the compression zone

- A
s
sf
width of compression zone or top
flange

b'

web thickness

ratio of average to maximum stress

for the flange

c2

ratio of distance between the point

of action of flange force and the
top fiber, to the flange thickness

distance from the center of gravity

of prestressed steel to the top fiber

d'

distance from the center of gravity

of the non-prestressed compression
steel to the top fiber

a compatibility factor

f
y

ratio defining the position of the

center of gravity of the compression
force contributed by concrete; ak
is the distance of center of gravity
from the top fiber

the ratio

p fsu/ft

the ratio

p' f*/fI

su

flange thickness
one half the theoretically required
length of non-prestressed compression bars

ratio of strength of concrete in

beam to that of cylinder

span length of a simply supported

beam

unit weight of concrete

moment due to the weight of the beam

strain in concrete at the level of

steel due to effective prestress
strain in prestressed steel due to
effective prestress

strain

moment due to the live load

moment due to the superimposed dead

load

required flexural
beam

load factor for the dead load

strain in the non-prestressed compression steel at ultimate

load factor for the live load

the ultimate strain of concrete in

flexural compression

percentage of prestressed steel,

As/bd

strain at yield of non-prestressed

steel

percentage of non-prestressed compression steel, A'/bd

a capacity reduction factor

Mu/bd2 f.
u

strength of the

limiting strain in prestressed steel

strain in the prestressed steel at
ultimate

curvature of the section

a dimensionless shape factor,

A/bh

I. INTRODUCTION

Design of prestressed concrete beams is

visions for ultimate design in our present

based upon two distinct concepts which lead to

specifications are more suitable for calculat-

two design methods known as service load de-

ing the flexural strength of a given section.

sign or working stress design, and ultimate

design.

It should be pointed out that there is a

relationship between working stress design and

In service load design the stresses in

ultimate design.

Although they have different

the beam are calculated on the basis of the

bases, in fulfilling the objective of one, the

assumption that concrete is an elastic material,

objective of the other is satisfied to a

(1)*
certain extent.

These calculated stresses are to be less than,

or equal

to, certain limiting stresses known as

allowable stresses.

The allowable stresses are

It can be shown that the provisions of

ultimate design can be used to proportion a

chosen so that the structure will perform its

section with a more rigorous control of duc-

intended service satisfactorily under service

tility.

conditions while providing indirectly for the

sign can then be used to check working stresses

safety of the beam.

in the section so designed.

In ultimate design the flexural strength

The provisions of working stress de-

Furthermore,

rational design of a section is considerably

or ultimate moment of the section is calculated

simpler by ultimate design than by service

based on the knowledge of behavior of the beam.

load design.

The calculated ultimate moment

is to be equal

The purpose of this study is to develop a

to, or greater than, the sum of moments of all

method by which a prestressed concrete beam can

forces each multiplied by a factor.

be proportioned by the provisions of ultimate

These are

known as load factors and are chosen so that

design.

the structure will be sufficiently safe under

of ductility in its influence upon the di-

the service conditions.

mensions of the beam.

Moreover, ultimate

It is intended to show the importance

In addition it is intend-

design also requires certain ductility in the

ed to study the influence of compression steel

beam, so that prior to failure the beam will

on ductility and the proportions of a section.

deform sufficiently.

Ductility is measured

by the deformation of the beam at failure.

In our present practice prestressed con-

We will consider a simply supported bonded beam and assume that the strength of the
beam is measured by flexure.

We will assume

crete beams are in most cases designed and

that the only loads acting, in addition to

proportioned by working stress design.

the prestressing force, are the weight of the

The

provisions of ultimate design are used to check

beam, the superimposed dead load, and live load.

the flexural strength of a section that has

*The numbers in parentheses refer to the

already been designed.

Furthermore, the pro-

entries in Chapter VII,

References.

II. CALCULATION OF ULTIMATE MOMENT

A.

ASSUMPTIONS

In addition to the above assumptions, the

For the sake of simplicity the determination of ultimate moment is discussed here

tension contributed by concrete is usually neglected since it is small at ultimate.

for prestressed concrete beams which have an

idealized section as shown in Figure 1.

The

section considered is flanged; the prestressed

is assumed to be bonded to concrete, and

steel

The neutral axis at failure may be either

in the flange or below the flange depending
upon the dimensions of the beam, the amount of
steel and the properties of both steel and con-

in addition to prestressed steel the section

crete.

is assumed to have non-prestressed compression

which the neutral axis falls in the flange.

steel.

B.

Practical sections sometimes contain nonprestressed tensile reinforcement which in-

FLEXURAL STRENGTH OF SECTIONS IN

WHICH THE NEUTRAL AXIS AT ULTIMATE
FALLS IN THE FLANGE

crease the flexural strength of the section

and reduces its ductility.

Non-prestressed

tensile reinforcement is not considered here

since it does not contribute to our primary

We will first consider the case in

In this case the width of the compression

flange is constant and is equal to

b,

as

shown in Figure 2.
We will take

f = f(e)

as the equation

purpose which is to develop a method of design

for the stress-strain diagram for concrete and

which leads to the lightest section for a

assume that this equation is the same for all

given strength and ductility.

the fibers in the compression zone of the beam.

The calculation of the ultimate moment is

usually based upon the following assumptions:
1.

The strain distribution in concrete

Since the width of the compression flange is

constant and the strain distribution in the
compression zone

is assumed to be linear with

varies linearly with depth in the compression

depth, the ultimate moment can be written as

zone of the beam.

follows:

2.

The stress-strain diagrams for the

prestressed as well as non-prestressed rein-

Failure occurs when the strain in con+ A'f'

(a - d')
s su

crete at the top fiber reaches a limiting value.

4.

ef(e) de + Afsu(d - a)

forcement are known.

3.

Mu= ab

The strain in non-prestressed com-

pression steel

is equal to the strain in con-

where

= distance from neutral axis to the

= the limiting strain

top fiber

crete at the level of compression steel.

5.

The average strain in steel

is not

greatly different from the maximum strain.

u
fsu = stress in prestressed steel

adding the term

(A' f(-6
(a - d'))
5
a
right side of the equation.

fl = stress in non-prestressed comsu

pression steel at failure
b

= width of compression zone or top

= distance from the center of gravity

flange

to the

The strain in prestressed steel at failure can be expressed as the sum of the following quantities:

of prestressed steel to the top

fiber

d' = distance from the center of gravity

It may be included by

is neglected.

steel

at failure

1.

The strain in prestressed steel due

to effective prestress designated as e

se
Effective prestress is the magnitude of pre-

of the non-prestressed compression

stress after losses, or at the time of de-

steel to the top fiber

termination of ultimate moment.

= area of prestressed steel

2.

A' = area of non-prestressed compression

The additional strain in steel that

is induced by sufficient load to make the

strain in concrete at the level of prestressed

steel
The above equation shows the sum of moment

steel equal to zero.

It can be shown that in

of all forces about the neutral axis which is

bonded beams this addition to the strain in

the same as the bending moment at the section

steel

due to the external

at the level of the prestressed steel due to

loads.

The first term on

the right side of the above equation is the

compression force contributed by concrete and

is equal to

ece,

the strain in concrete

effective prestress.
Hence the total strain in steel at the

does not take into account the area of concrete

load corresponding to zero strain in concrete

replaced by compression steel.

at the level of steel

This effect is

usually small, but it can be included by subtracting the term

As f(-a

3.

It can be seen from Equation 1 that for a

Mu

depends on three quantities

that are not known, namely

the neutral axis;
stressed steel; and

fsu,

a,

the depth to

f' ,

the stress in the

Additional strain in steel from the

load corresponding to zero strain in concrete

at the level of steel to ultimate.

This ad-

dition can be expressed as follows:

- a) F.
_u
a (d

f =

the stress-strain diagram for concrete

were known.

+ e.
ce

the stress in the pre-

non-prestressed compression steel -- if

f(e)

se

(a - d')) (a - de )

from the right side of the equation.

given section

is

Therefore, it is necessary to

The quantity

in the above expression is

a compatibility factor.

If concrete were not

obtain other relations in order to be able to

cracked and were bonded with steel its magni-

compute these unknown quantities.

tude would be one.

From the equilibrium of horizontal forces

in the section we have the following equation:

ab

e j
u 0

Since at ultimate, con-

crete at the level of prestressed steel is

cracked, the value of

in a particular

section depends on the position of crack and

f(e) de + A'f'

s Su

= A f

S SU

(2)

As before, in the above equation the effect

of the area of concrete replaced by compression

condition of bond.

Its magnitude is usually

less than one at ultimate, however it is

difficult to predict its value for a particular section in a given beam.

A detailed

discussion of its magnitude is available.(2)

Hence

esu,

the strain in prestressed

e'
and f'
as well as
su
su
Equations 4 and 6, vanish and the problem is
no compression steel,

steel at failure, can be expressed as the sum

reduced to the solution of four equations for

of the quantities listed above

four unknowns.

In this case Equations 1, 2,

3, and 5 are the relations, and

e

su

= e

se

e ce + -a

(d - a)F.

C.
Since

e
ce

e
+ se
a

is small

e
su

and

M , a, f

FLEXURAL STRENGTH OF SECTION IN

WHICH THE NEUTRAL AXIS AT ULTIMATE
FALLS BELOW THE FLANGE

in comparison with

When the neutral axis at ultimate falls

(d - a)F

below the flange the expression for

will be taken as 1.0 for bonded beams.

In the following discussions the value of

u ef(e) de
0

+ (b - b i)a 2

ef(e)

steel at failure can be written as follows:

u
su

= e

e
+ ece +-u
a
se

(d - a).

where

concrete at the level of non-prestressed comis equal to the strain in com-

pression steel, we have

+ A'f'
s su (a - dn)

b'

= the web thickness

= the flange thickness.

The first term on the right side of the

above equation is moment due to the compression
force developed by a rectangle of width

e'
= ~
su
a

de

e (a-t)/a

+ A f
(d - a)
s su

Since we have assumed that the strain in

pression steel

is

Hence

the expression for the strain in prestressed

the following:

it is often neglected.

are the unknowns.

(a - d')

(4)

and depth

about the neutral axis.

b'
The

second term is the moment of the compression

e'
= the strain in the non-prestressed
su
compression steel at ultimate

where

We have also assumed that the stress-

force contributed by a rectangle of width

(b - b')

and depth

about the neutral axis.

Figure 3 shows all the forces of a section in

strain diagrams for prestressed steel and non-

which the neutral axis at ultimate is below

prestressed compression steel are known.

the flange.

The

equations for these diagrams are designated as

F(esu )

and

G(e' )

As before, the area of concrete

replaced by compression steel

The expression

respectively.

f(e),

is neglected.

the stress-strain

fsu = F(e su)

su
su

(5)

diagram in concrete, is assumed known and

su= G(e' )
su
su

(6)

applicable to each compressed fiber of the

In order to obtain

it is necessary to

solve Equations 1,2,3,4,5, and 6 simultaneously

for the six unknowns
fo .

M , a, e

su

su

, e'

su'

and

su
In a special case in which the section has

beam.

In Equation 7 there are quantities

a,

which are unknown and should be

, f
su, su
determined by other relations.
f

From the equilibrium of horizontal forces

we have:

bla
u

uf(,) de + (b
0

f(e) de

-b

Mu

e (a-t)/a

+ A'f'
= A f .
s su
s su

uef(e) de
0

+ (b - b')a

(8)

eu

ef(e) de

e (a-t)/a

As before the area of concrete replaced by the

compression steel has been neglected.

+ Af

(d - a)

(a - d')

+ Asf'

(7)

Equations 3, 4, 5,and 6 are equally applicable in this case.

M

Hence in order to find

3, 4, 5, and 6 simultaneously for the six

uf(e) de +(b -e b)a

eu f
0

f(e) de
f

e (a-t)/a

M , a, f , e , f' , and e .
u
su
su
su
su

unknowns

D.

it is necessary to solve Equations 7, 8,

SUMMARY OF GENERAL EQUATIONS

+ A'f'
= A f
s su
s su

The ultimate moment in a section in which

(8)

the neutral axis at failure is in the flange,

or when

a < t,

is obtained by the simul-

taneous solutions of the following equations:

su

e
M
u

+ A'f' (a - dl)
s su
ab
u

f(e) de + Af'
lul

= A c f

e
+ e
+ -- (d - a)
se
ce
a

(3)

= -u (a - d")
a

(4)

= F(e

(5)

= G(e" ) .

su

ef(e) de + A f
(d - a)
s su

2 j
u 0

= e

su

(6)

su

It should be remembered that in Equations

e

su

= e

+ e ce + a u (d - a)
se

(3)

I and 2 as well as 7 and 8 the area of concrete replaced by compression steel is not

e,
su

= -

(a

- d')

(4)

taken into account.

E.
fu = F(e u )

CALCULATION OF ULTIMATE MOMENT BY

THE PROPERTIES OF STRESS BLOCK
The procedure for the calculation of

f'

su

= G(e' ).

(6)

su

ultimate moment discussed in the preceding

sections is the most general formulation for

The ultimate moment for a section in

the problem of computing ultimate moment and is

which the neutral axis at failure is below the

consistent with the assumptions made.

flange, i.e., when

assumptions are reasonable and are in agree-

t < a,

is obtained by the

simultaneous solution of the following equations:

ment with observations.

These

Practically, however, there are diffi-

the stress-strain diagram for concrete.

culties in solving the problem by this procedure.

Equations I and 7 for ultimate moment,

as well as Equations 2 and 8 for equilibrium

f(e)

of horizontal forces, depend upon

stress-strain diagram for concrete.
stress-strain diagram for concrete

the

The
is non-

This can be achieved if the average stress

in the section and the point of action of the
compression force contributed by concrete were
known.

the stress block were known, the ultimate

moment could be determined.

linear and may change shape as concrete

strength changes.

In addition the stress-

That is, if the area and centroid of

Let us first consider the case in which

the width of compression zone is constant,

strain diagram obtained from a concrete cylin-

i.e.,

der may not truly represent this relationship

the flange.

for all

the fibers of the beam in the com-

pression zone, and obtaining a stress-strain

the neutral

reasonable expression for

f(e)

In this case let us express the

force and moment contributed by concrete as

follows:

diagram for the concrete in the beam by testing a beam is a lengthy procedure.

axis at ultimate falls in

eu

Even if a

were es-

ab

f(e) de = k

fc ab

(9)

tablished from tests, the non-linear nature of

the expressions for the compression force and

2b
a2b

moment contributed by concrete, make the so-

lution of the non-linear simultaneous equations

u
f ef(e)

de = k,

k3 fc ab (a-ak 2 )(10)

tedious.
Since strain varies linearly with depth

where

kI = ratio of the average to maximum

in the compression zone, by adopting a stress-

stress

k3 = ratio of strength of concrete in

strain diagram for concrete, the stress distribution with depth will be defined.

in the compression zone

beam to that of cylinder

From the

k2 = a ratio defining the position of

shape of stress-distribution in the compression

zone both the force and the moment contributed

the center of gravity of the com-

by concrete can be calculated for any section

pression force contributed by con-

even if the width of the section is variable.

crete;

The shape of the stress distribution in the

center of gravity from the top

compressed zone of concrete is called the

fiber.

is the distance of

Equation 9 gives the compression force

stress block.
To simplify the problem for a practical
solution a somewhat different approach is introduced.

ak2

In the equations of equilibrium of

contributed by concrete, and Equation 10 gives

the moment of this force about the neutral
Figure 4 shows the forces in the section.

axis.

kl,

k3 , and k2

can be de-

horizontal forces and moments, it is only

If the ratios

necessary to know the compression force and

termined, this procedure is very convenient.

moment contributed by concrete.

If we can

Substituting Equation 9 for the com-

somehow find a way of estimating the force and

pression force contributed by concrete in

moment contributed by concrete without knowing

Equation 2 we have

the actual distribution of stress in the compressed zone, there would be no need in having

k3 f

ab + Af'su = As su

(2a)

or

k =
S3

Equation 11

A f
- A'f'
s su
5s SU
f ab

()

and the compression force developed by the

rectangle

indicates that

k I k3

(b - b')t

can be

measured from tests, since all the quantities

(b - b')a

at the right side are either known or can be

measured.

The quantity

kI k3

as

f(e) de = clfi

for the ma-

terials commonly used is in the neighborhood

where

of 0.7.

stress for the flange.

In a similar fashion

k2

can be

evaluated since the moment in the section can

be measured.

The quantity

(b-b')t,

(a-t)/a

k2

cI

is the ratio of average to maximum

Since the stress

distribution is undefined, the quantity

varies around

is undefined.

0.42 and has little influence on the ultimate

0.7 and 1.00.

moment.(3)

c1

It can vary widely between

The moments of these compression forces

This approach is convenient in cases where

about the neutral axis are

the width of compression zone is constant.

Equation I can now be written as follows:

b a2

Cu

f(e)e de
0

u
Mu = kl k3 f' ab (a - ak)

+ Asfsu (d - a)

+ A'f'
(a - d').
S su

= k1 k3 f' ab" (a - ak2 )

This equation represents the sum of

moment of all forces in the section about the
neutral axis.

and

(b

) a2

In this case it can also be

ef() ) de

C (a-t)/a

written in the following convenient form by

taking moments about the centroid of the com-

= c I f' (b -

b')

t (a - c2t)

pression stress block:

Mu = As su(d - ak2 ) + Asf u(ak 2 - d').(la)

where c2 is the ratio of the distance between

the point of action of flange force and the

Equation 2 for the equilibrium of hori-

top fiber, to the flange thickness.

zontal forces in the section becomes the same

quantity

as Equation 2a.

not influence the moment appreciably.

This method is inadequate when the width

of the compression flange becomes non-uniform.

is undefined; however, it does

in the neighborhood of 0.5.

Equation 7 can now be written:

For the

section shown in Figure 5 let us express the

= k

k3 f

abl

(a - ak2

compression force developed by the rectangle

b'a

as
-a
u
C

+ cl f
uf(e) de =k
0

a b'

(b -

+ As su ssu
(d -

bi)

It is

Figure 5 shows

the forces in the section.

This condition occurs when the neutral axis at

ultimate falls below the flange.

c2

The

t (a - c2 t)

a) + A'fs
S SU (a - d').

The equation gives the sum of moments of

all forces
axis.

ce

se

su

in the section about the neutral

e

Often it is more convenient to take

= _- (a - d')
a

'

su

moments about the center of gravity of prestressed steel.

Mu = k

ab' (d - k2 a)

k3

+ c

In this case we have

f' (b - b')

su == F(e
F(su )
su = G(e su ).

fl

t (d - c2 t)

When the neutral axis is below the flange,

or when

+ A'f'

(d - d').

(7a)

t < a,

the ultimate moment may be

calculated by a simultaneous solution of the

following six equations.

It is also possible to take momentsabout

the compression force contributed by the web.

Mu = kI k3 f' ab' (d - k2a)

In this case the ultimate moment can be expressed:

+ cl f' (b - b')

Mu = Asfsu (d - k2a)
u
+ l f' (b - b)

t (k2a - c2t)

bl)

t (k2 a - c2 t)

c (b

+ A'f'
su

+ A'f'su (d - d')
s

k,

k3 f'

(ka
2 - d').
(k

kl k
1

ab' + c,

f'

(b - b')

+ A'f
= A f
s su
s su

Equation 8 for the equilibrium of horizontal forces

t (d - c2 t)

su

= e

se

+ e

ce

+ _u (d - a)
a

in the section can be written:

f' a b' + c f' (b - b')

c
1 c

Ce
su

+ A'f'
= A f
S su
s su

su

= F(e

f' = G(e'

In summary, when the neutral axis at

failure is in the flange, or when

u- (a - d')
a

t > a,

su

the

ultimate moment in the section may be obtained

As before, the area of concrete replaced

by a simultaneous solution of the following

by compression steel has been neglected in

equations:

Equations la, 2a, 7a,and 8a.

M u = Asfsu (d - ak2)

stressed concrete is based upon the above

The present thinking and practice in pre-

equations.
+ A'f'
S SU (ak2 - d')

(la)

strain diagram the properties of the stress

block

kI k3 f
ab + A'f'
= A f
1 3 c
s su
ssu

Instead of defining the stress-

k,,

k3 , k2 , c,,

and c2

are specified.

It should be pointed out that while

k 1 k3 and

non-prestressed
compression steel-

A,

|
*__^______^____

"II t

prestressed steel -I

kb
FIGURE

FIGURE 2.

FLANGED SECTION, NEUTRAL AXIS

1.

IDEALIZED

IN THE FLANGE

I-SECTION

du

FIGURE

3.

FLANGED

SECTION, NEUTRAL AXIS BELOW

Eu(a-t)/a

THE FLANGE

f;u
k i k 3 f c ob
A's

Asfsu =pbdfsu

FIGURE 4.

FIGURE

5.

FLANGED SECTION, NEUTRAL AXIS

IN THE FLANGE

FLANGED SECTION, NEUTRAL AXIS BELOW THE

FLANGE

t/h
FIGURE

6.

RELATIONSHIP

BETWEEN ,* AND GEOMETRIC

PARAMETERS OF

THE

SECTION

II

228

214 ,
/

150

10c

0.01

0.02

0.03

0.04

0.05

Strain
FIGURE

7.

STRESS-STRAIN DIAGRAM FOR STEEL

0.06

0.07

18s

r 15"

FIGURE

8.

SECTION OF A

LOW DUCTILITY
BEAM =

= 0.01

300

LB/FT

BEAM WITH

4"

- WEIGHT OF
36

17

^
*

16

16"

_32"

I_

4
I

56"

16

1/2" Strands

-- --- t_.m

. S.
..

jj

16"
FIGURE

9.

SECTION OF A BEAM WITH VERY HIGH DUCTILITY

(eu =

0.03) -

WEIGHT OF

BEAM = 377

LB/FT

1/2" Strands

I___.
5

if

No. 9

2F

FIGURE

10.
HIGH

36"

IS

SECTION OF
REQUIRED

OBTAINED BY

REINFORCEMENT

BEAM

DUCTILITY
THE

- WEIGHT

USE
OF

IN
(E

WHICH
=

THE

0.03)

OF COMPRESSION

BEAM

300

LB/FT

78
7A-

o.5

Z 6A
16

FIGURE

11.

END SECTION OF THE BEAM WITH

COMPRESSION REINFORCEMENT

LU
-j

LU
L-

LU
LL
0

z
0

IULJ
-i
LU
0

LU

U-

z
to

0
z

~0

C,

IV

TABLE I

SUMMARY OF SECTION PROPERTIES AND STRESSES

FOR SECTIONS OBTAINED BY ULTIMATE DESIGN

(All section properties are based upon the transformed section assuming n = 7)
(Negative stresses are tensile)

A
SA
Section

Illustrative
Example la
e

su

2
in

in

in

4
in

2
in

A'
As
.-2
in

A
s

Stress Before
Losses (transfer)
ksi

Stress
After Losses
ksi

Weight
lb/ft

top
(tens)

bottom
(comp)

top
(comp)

bottom
(tens)

302

18.14

17.86

52,190

2.44

300

-0.14

2.56

2.38

-0.29

376

15.90

20.10

65,360

2.29

377

-0.09

2.20

1.68

-0.36

330

16.62

19.38

59,150

2.15

5.00

300

-0.14

2.29

1.90

-0.41

-0.19

2.40

2.25

-0.42

= 0.01

Illustrative
Example lb

E su = 0.03

Illustrative
Example 2
S

su

= 0.03

Allowable Stresses
American Concrete Code (318-63)
(f' = 5 ksi; f'.

= 4 ksi)

A = transformed area, yt = distance from centroidal axis to top fiber; Yb = distance from
centroidal axis to bottom fiber; I = moment of inertia; As = area of prestressed steel;
A' = area of non-prestressed compressive steel; n = modular ratio for both types of steel;
fc = strength of concrete; f'. = strength of concrete at transfer; prestress at transfer
150 ksi; effective prestress after losses 128 ksi.

f
k2

can be estimated from tests with a suf-

ficient degree of precision, there is no way

to estimate

c1 and c2

generally, so that they

will be applicable to all sections.

k I k3 e
1 3 u

su

-e

su

se

-e

ce

+ e

fsu = F(e su).

'su'

su

In spite

of this weakness our specifications at present

assign specific values to these coefficients.

The American Concrete Institute Code

k2
(318-63) gives k k
= 0.59 and requires
that the expression Tor the ultimate moment

F.

PROVISIONS OF THE AMERICAN CONCRETE

INSTITUTE BUILDING CODE
The provisions of the American Concrete

be multiplied by a capacity reduction factor

as follows:

Institute Code (318-63) do not consider the

effect of compression steel on ultimate moment
or strain in steel,

f
[Afsu d (1 - 0.59 p-Sf)
c

Mu

and instead of giving the

stress-strain diagram for concrete, the coef-

where

is a capacity reduction factor,

taken as 0.9.
ficients of the stress-block are specified. 4)
The Code implies that when the stress
In the following paragraphs the expressions
strain diagram for steel

is available,

su
solution of Equations14 and 5; if not,

su
can be obtained by a simultaneous

given in the Code will be derived and discussed.

and e
When the neutral axis is in the flange and
there is no non-prestressed compression steel

su
in Equation 15 may be calculated by the follow-

from the preceding section, the ultimate moment

ing expression:
may be obtained by a simultaneous solution of
the following equations.
f

su

= f'

(1 - 0.5 p f'/f),

Mu = Af su (d - ak 2 )
where
(2b )

k f' ab = A f
1 3 c
s su

e
= e
+ e
+-u
a
se
ce
su

f'

is the ultimate strength of pre-

stressing steel.
The Code controls the ductility of the

(d-

a)

(3)

beam by requiring that the ratio

be less than or equal to 0.3.

fsu = F(e su)

(5)

which

p fsu /f

p fsu/f'

In cases in

is greater than 0.3, the

ultimate moment may be calculated by Equation

Elimination of

between Equationslb

and 2b and between Equations 2b and 3 permits

15 provided that

p f /f' is taken as 0.3.

This
resuts
in
the
foowing equation:

This results in the following equation:

the reduction of the above four equations to

M , f

three equations in which

are the unknowns.
s

-- ,

We will

su

, and e

introduce

su

lb and 2b we have:

By limiting the value of

p fsu/f

the

Code in effect requires that the strain in

k2

mA =
d (
s su

(17)

p =

From Equations

as percentage of steel.

= t (0.25 f' bd2).

k 1uk 3

from Equations 2b and 3:

steel at failure be not less than a limiting

fsu
p

)
c

(13)

value, since from Equation 14 we have

Pf

su
c

su

3 - e
se

< 0.3.

+
ce

--

The above requirement sets an acceptable

range for

esu,

roughly in the neighborhood

as that part of the force contributed by the

prestressed steel that develops the rectangle

of 0.01, below which it cannot go as far as

a b'

calculation of

section.

is concerned.

in the compression zone of a flanged

The quantity

For sections in which the neutral axis at

ultimate falls below the flange and there is

sf

su

= c

f' (b - b')
c

no non-prestressed compression steel, from the

preceding section the ultimate moment may be

is designated as that part of steel that

obtained by a simultaneous solution of the

develops the rectangle

following equations:

compression zone.
Since

= A

+ A

sr

in the

we have

sf

f' ab' (d - k a)

= k

(b - b')

Af
- c f' (b - b') t = A
f .
s su
1 c
sr su
+ cl f' (b - b')

ab'

t (d - c2 t)
f' (b-b')

+ c,

t = Asf

ssu

(8b)

Equation 18 we have

e
= e

su

ece

se

By substituting the above quantity in

(d - a)

M
f

= F(e

su

su

= A

d (1
f
sr su

).

Elimination of

Scf

between Equations 7b

k
2

kI k3
1 3

(b - b')

sr
b'd

t (d

f
su
f
C

c2 t).

and 8b and between Equations 8b and 3 will

result in three equations in which
and e

The Code gives

k2 /k I k3 = 0.59, cl =

0.85, and c2 = 0.50

and requires that the

fsu,

We have

are the unknowns.

su

Mu

expression for ultimate moment be multiplied

M = d[A fsu - C

s su

f'

b')

t]

(b - b')t
Af
- c f'
s su
c
k k fP bid

[l-

fj (b - b')

+ c,
f

by a capacity reduction factor as follows:

k2 1

13

(b-

1 c

k I k3

+ 0.85 f' (b - b')

(d - c,t)

1 3 eu

fr
c

su

+ c

(l

- e

se

-_

b
(su
e
+ e (1
u
ce

f
A
d (1 - 0. 5 9 -- u)
[Asr f
b'd f
sr su

t (d - 0.5t)].

The arbitrary assumption of

c2 = 0.5
(19)

c I = 0.85 and

by the Code may result in some in-

accuracies when the neutral axis

is very

slightly below the bottom fiber of the flange.

(5)

fsu = F(e su)

For sections of small

c I = 0.85

The American Concrete Institute Code (31863) designates the quantity

for

b'/b

the assumption of

may even lead to a negative quantity

sr
The Code implies that when the stress-

strain diagram for the prestressed steel is

A

sr

su

= k

f'

ab'

available

fsu and esu

can be obtained by a

simultaneous solution of Equations 19 and 5.

requiring a minimum value for

If the stress-strain diagram is not available

before, is in the neighborhood of 0.01.

the Code permits calculation of

su

by

esu,

which as

It should be pointed out that the limi-

Equation 16.

tations on the ductility of the beam as given

As before, the Code limits the quantity

Asr su
If this ratio is
to 0.3 or less.
bdb-d frc
more than 0.3, the ultimate moment may be

by the American Concrete Institute Code do not

calculated by Equation 20 provided that

It will be shown in Chapter V, Section C that

this effect is not negligible.

/f'd f'

is taken as 0.3.

This results

in the following equation:

include the effect of any compression steel

that may be present at the top of the beam.

The neutral axis may either fall in the

flange or below the flange.

M = < [0.25 f' b'

u
c

between the two cases

The distinction

is made according to the

following inequalities:
+ 0.85 f

(b

- b')

t (d

- 0.5t)]

(21)

If

t> a = p f

su

d/kl k

axis falls in the flange.

since we have

sr

f su/b
su

d f

su

-e

se

kl k3 = 0.7

this condition can be restated as follows:

fsud
When t > 1.4 p , the neutral axis

c
1

the neutral

Since the American

Concrete Institute Code gives

A

f'

I 3 c

3 u
< 0.3.
-e
+e
ce
u

Limitation of this quantity is equivalent to

falls in the flange. f

When

t < 1.4 p

d
su

falls below the flange.c

the neutral axis

III. PROVISIONS FOR SAFETY AND DUCTILITY IN ULTIMATE DESIGN

In ultimate design a section

of the beam should be large enough so that the

is

proportioned in such a way that the ultimate

moment

is greater than the moment developed

under service loads by a prescribed quantity,

strain in steel at ultimate will be at least

equal
as

esA

and that it deforms a certain amount before it

fails.

From the discussions

in the preceding

section we know that the prediction of failure

These concepts may be stated in the form

of the following requirements:

to a given limiting value designated

> N

(M

+ M ) + N

on the basis of moment depends upon the limiting strain e , i.e., a value for e
is
U
U
required for an unambiguous definition of

(22)

failure.
There are many ways that ductility of the

and

section can be measured.

su > e

(23)

In Chapter II,

Section F, it was shown that the American

Concrete Institute Code defines ductility by

where

= the required flexural strength

Nd

= the load factor for the dead load

the following quantity

of the beam

M
g

kl k3 eu

= moment due to the weight of the

su

-e

se

- e

ce

+ e

beam
M

= moment due to the superimposed

= the load factor for the live load

M.

= moment due to the live load

= strain in steel at ultimate

Ductility may also be measured by the

curvature at ultimate which may be defined as

dead load

follows:
e

u
a

su
es9 = limiting strain in steel

Expression 22 states that the required

where

to

Nd(M

+ Ms) + NMA,

which is

su

se
d

ce

+ e

is the curvature of the section.

Both of these measures of ductility are

flexural strength of the beam should be at

least equal

cp

based upon the magnitude of

which is the

a requirement for the strength of the beam.

strain in prestressed steel at failure.

For the type of loads considered here, the

esu

American Concrete Institute Code (318-63)

Nd = 1.5 and N

gives

= 1.8.

Expression 23 states that the ductility

Hence

may be used as a measure of ductility.

This method is based upon the assumptions

which defines
u
flexural failure, and we have a minimum limitthat we have a value for

quantity for

esu,

designated as

es5.

In

addition we have the stress-strain diagram for

prestressed as well as non-prestressed compression steel.

IV. ULTIMATE DESIGN OF SECTIONS WITHOUT COMPRESSION REINFORCEMENT

A.

METHOD OF ANALYSIS USED

The design procedure developed here does

not depend on the method of analysis.

f
P fc

su

0.7 e
e - e
+ e (b)
se
ce
u

+ 0.85 (1 - -)

The

(19a)

b~-

following are adopted.

When the neutral axis is in the flange,

fsu

(5)

F(e su).

14, and 5 will be adopted for the

Equations 15,

calculation of ultimate moment.

Equation 15 is

In addition to the above equations we

the expression for the ultimate moment given by

also know that

the American Concrete Institute Code (318-63).

eliminates the need of considering Equations

We will further assume

k I k3 = 0.70.

c
(15)

f
su
P f
c

(14a)

fsu

su

0.7 e
u
- e
+e
se
ce
u

(5)

When the neutral axis

is below the flange,

Equations 20, 19, and 5 will be adopted for

calculation of the ultimate moment.

Introducing

Equation

20 is the expression of the ultimate moment

When

When

M = $

[A

f
t < 1.4 p

fs

sr *su

d (I

M
2----- = q (1 - 0.59q)
f cbd

When

we have:
f
fsu)

b'd f

Se - e
- e
+
su
se
ce

t (d - 0.5 t)]

(20)

(14a)
u

q > 0.7 t/d:

I +

*bd2f

.085t

(t
d k

) - 0.59 -

q]

)
(20a)

q =

0.7 e (b'/b)
u-+
su
se
ce

+ 0.85 (1 - bl/b) t/d.

(15a)

V.,

where

where

A
- 0.59 b

+ 0.85 f' (b - b')

d
--c

the above

q < 0.7 t/d:

k I k3 = 0.70, c1 = 0.85, and

c2 = 0.5.

/f',

expressions may be written in dimensionless

given by the American Concrete Institute Code.

We will assume

q = pf

F(e su).

This condition

form as follows:

f
M =
[Afsu d (1 - 0.59 p -)]
fo
su
us
- e

> e6

17 and 21.

These equations are rewritten here for

fsu d
convenient reference when t > 1.4 p fsu

(19a)

B.

DETERMINATION OF AREA OF THE

M /$bd2 f' = Q.

Let us write

hence they should be made as small as possible

BEAM

Expression

without causing the dimensions of the beam to

22 can then be written as an equation in the

become unreasonably thin.

following form:

Qebd2ft

From the expression for

= Nd

(M

+ Ms) + NMM.

Q'

and Equation

25,

it can be seen that

k.

However, since the bottom flange of the

Q'

decreases with

beam should be large enough to permit the

Substituting

A/hi for b

where

is the

gross cross-sectional area of the beam,

the overall depth, and

placing of steel,

is

is a dimensionless

definitely.

The quantity
2

A
*

==,

= yAL
8

fc c N

Nd

M - M

greater than 0.9.

M

Ms +-

increases with
q

it is de-

as large as possible;

however, Expression 23 sets the upper limit

h*N d
y

Q'

sirable to make

d2 f'Q

where

Since

Nd

A=d

should be made as

that in most practical cases it can be made

Ne
or

d/h

large as possible; however, it is doubtful

N
N1

cannot be reduced in-

possible.

shape factor, we have:

M
m

It should be made as small as

for

is the unit weight of concrete.

q.
Since Expression 23 sets the required

For the idealized I-section shown in

minimum ductility of the beam as a strain in

Figure 1 the following general expression may

steel equal to

be used for

consistent with the required ductility can be

4:

es, the required maximum

computed from Equations 14a and 19a as follows:

S= A- = -t (1 + k) + - (l - 2 ).
h
b
Sbh
h
The quantity

(25)
qmax

max

in the above equation is

the ratio of the width of bottom flange to that

of the top flange.

Equation 25 is plotted in

Figure 6 for a few typical cases.

or

A study of

only.
Q

depends upon

It can be seen that

and increases with

t.

and

(14b)

0.7 e
- e
- e
e+ e
sse
ce
u

+ 0.85 (1 -

Equation 24 shows that for a given depth and

type of concrete

si

0.7 e
u
- e
se - ece + e u

) (-)
bs()

(19b)

Q
whichever applies.

decreases with

In order to decrease

It should be pointed out that Equations

the area of the beam it is necessary to in-

14b and 19b contain the additional parameter

crease

Cse,

and decrease

Since both

and

t/h, b'/b, and d/h,

variation with

*.
*

are functions of

prestress.

in order to study their

it will

be more convenient

The quantities

limit for

decrease with

se

it should be taken as
The practical upper

for the materials used in pre-

tensioned construction is about 0.005.

as large as possible.

t/d and b'/b

qmax

large as practicable.

Q' = - Q with A.
h *
In order to obtain the least area it is
Q'

It can be seen that since

increases with

to study the variation of

necessary to make

the strain in steel due to effective

C.
Q'

ILLUSTRATIVE

EXAMPLE

I.

The following example is presented to

illustrate the procedure for the ultimate

Equations 19b and 20a since

design of a prestressed concrete beam and to

> 0.7 t/d

show the influence of the required ductility

below the flange.

in this case

and the neutral axis at ultimate is

From Equation 19b we have:

on the dimensions of the beam so designed.

It

is necessary to design a simply sup-

max

0.7 x 0.004 x 1/4

0.010 - 0.0044 - 0.0006 + 0.004

ported beam of 54 foot span subjected to a

+ 0.85 (1 - I) 0.185 = 0.196

superimposed dead load of 1.0 kips per linear

foot (klf) and a live load of 0.6 klf.
load factors are given as
= 1.8,

N d = 1.5

The

and

and from Equation 20a:

and the capacity reduction factor is

= 0.9.

Design the section for:

ductility corresponding to

1. a minimum

= e

= 0.01,

M
--- u<bd fl
c

and 2. a minimum ductility corresponding to

es

su

= e

s2

0.196 [1 + 0.185 (4-1)

- 0.59 x 4 x 0.196] - 085 (0.185)2(4-1)

= 0.03.
= 0.170.

The effective prestress may be taken as

the prestress after losses which in this
problem is given as 128 kips per square inch
(ksi).

This value corresponds to a prestress

of 150 ksi at transfer if the effectiveness

se

of the section can be

obtained using Equation 24 with the following

values:

The strain due to effective

is taken as 0.85.
prestress is

The area

= 0.0044,

and

may be

e
ce

approximated as 0.0006.

M =x 54
s
8o

x 1.0 x 12 = 4370 k-in

M = I x 54
1
-9
4

x 0.6 x 12 = 2620 k-in

This approximation

may be verified after the section is designed.

Also for the purposes of this problem assume

fl = 5 ksi, e = 0.004, y = 0.15 kips per cubic
c
u
foot (kcf), and h = 36 in. The stress strain

d2f' Q
_
h*N d

diagram for steel is shown in Figure 7.

1.

-- = 1.2
Nd

Section with Minimum Required Ductility

= 0.01
Corresponding to e

crease with

A,

t/h, b'/b,

x 5 x 0.170 x 0.90
36 x 0.467 x 1.5

in-

15 x 54

= 4.6 k-in.

hence they should be taken

as small as possible.
as

and

= 31.8 k-in

In Chapter IV, Section B it was shown

that the quantities

(32
(32.4)

_c

Here they will be taken

Therefore

t/h = 1/6, b'/b = 1/4, and k = 0.8.

The shape factor of the section

is

276 in
4370 + 1.2 x 2620
- 31.8 - 4.6
=276 in

obtained using Equation 25 as follows:

=-1- (1 + 0.8) +1

) = 0.467.

d = 32.4 inches and t/d = 0.185.

values of

and Q
q
max

b = A =
276
= 16.4 in
b-h
0.467 x 36
kb = 0.8 x 16.4 = 13.1

d/h = 0.9, for h = 36 inches, we

Assuming
obtain

(1 - 2

The

can be computed from

b'

=-

x 16.4 = 4.1

in.

in

The stress in the steel at ultimate can be

the bottom flange is not overdesigned.

found from the stress-strain diagram for steel

value of

shown in Figure 7.

24 and 25 it can be seen that the area of the

k = 0.5

is selected.

From Equations

section decreases with the web thickness.

f

su

However, the web thickness cannot be reduced

= 214 ksi.

indefinitely, since the cover requirement for

The amount of prestressing can be found from

the draped reinforcement and the shearing

the definition of

strength determine the minimum thickness.

to be

this case let

0.196 x 5
2142 - 0.00458
p-

In

b'/b = 1/6.

In view of the large ductility required

the neutral axis is bound to be closer to the

= 0.00458 x 16.4 x 32.4 = 2.44 sq.

in.

top fiber than in the preceding example.

De-

termination of this position affects the se(Use seventeen 1/2-inch strands)

lection of the thickness of the top flange.

A total of seventeen 1/2-inch strands are

The value of

needed.

3 as follows:

Each 1/2-inch strand has an area of

0.1438 square inch.

can be obtained from Equation

The final dimensions of

the section in this solution are shown in

Figure 8.

0.030 = 0.0044 + 0.0006 + 0a004(32.4-a)

a

The bottom flange has been widened

to accommodate the reinforcement.

Both top

from which

a = 4.47 inches.

Use of

t>

and bottom flanges are tapered to facilitate

4.47 inches

construction.

top flange, the bottom fibers of which would

The properties of the trans-

would result in an overdesigned

formed section as well as the stresses at the

not be subjected to compressive stresses at

top and bottom fibers before and after losses

ultimate.

are listed in Table 1.

selected as a practical dimension.

2.

= 0.125

Section with Minimum Required Ductility

= 0.03
Corresponding to e
The ultimate strain

in the steel

t = 4.5 inches

is

Then

t/h

and Equation 25 yields the value of

as follows

required

for this example is very large and is not used

frequently in actual practice.

The value of

= (0.125)(1 + 0.5) + (1/6) [1-(2)(0.125)]

It has been

selected to show that direct design for the

= 0.312.

largest levels of ductility is possible and to

study how it affects the shape of the section.
All the quantities are the same as in
part 1
case

The values of

and

can be obtained

from Equations 14a and 15a as follows:

of this example except that in this

= 0.03,

and

k,

the ratio of the

=0 0965
= (0.7)(0.004)
0
q = 0.03 - 0.0044 - 0.0006 + 0.004

width of bottom flange to width of top flange,

is different.

The bottom flange needs only be

Q = 0.0965 [1 - (0.59)(0.0965)] = 0.091

large enough to accommodate the reinforcement.

Due to the fact that the higher the ductility

The above equations are applicable because

the wider the top flange has to be to provide

q < 0.7 t/d = 0.0973,

the required area under compression, it is

in the flange of the resulting section.

necessary to select

small enough so that

and the neutral axis is

The area of the section can be computed

from Equation 24 using the known values of

case 1
of

and 25.5 k/in as the modified value

d2f'QO/htNd

beam.

The dimension of the bottom flange is

the minimum required to accommodate the pre-

coincides with the calculated value of

Therefore,

thereby requiring no adjustments.

If

kb
kb

360 in
in

(2620)
4370 + (1.2)
25.5 - 4.6

It

stressing steel at the required depth.

360

turns out to be larger than necessary, only the

minimum required should be used, as the bottom

b =

A
h

360

-32 in

= (0.32)(36)
= 32 in
(0.312) (36)-

flange contributes nothing to ductility and

strength.

If the adjustment of the dimensions

is large, recalculation may be necessary to

kb = (0.5)(32) = 16 in

improve the shape of the section.

The proper-

ties of the transformed section and the

b i =--

= 5.34 in.

stresses before and after losses for this part

are also given in Table 1.
A comparison of Figures 8 and 9 indicates

From Figure 7, the stress strain diagram for

steel,

may be obtained

that a large ductility results in a heavy

section.

su

In this particular example increas-

ing the required

= 228 ksi

from 0.01 to 0.03

causes the weight of the beam to increase by

26 per cent.
p =

.96

= 0.00212

There is a 6 per cent saving in

the amount of prestressing steel as the more

ductile section requires one 1/2-inch strand
less.

= 2.20 in .

This is because the larger stress in

the steel at ultimate not only compensates

the additional weight of the heavier section

(Use sixteen 1/2-inch strands)

but also results in less required area of

Figure 9 shows the final section of the

steel.

V. ULTIMATE DESIGN OF SECTIONS WITH NON-PRESTRESSED COMPRESSION

A.

METHOD OF ANALYSIS USED

Determination of flexural strength of

STEEL

In Equation 26 the quantities

f

we need Equations 8a, 3, and 5.

compression steel was discussed in Chapter II,

8a we will take

and E.

and

are unknowns, and for their determination

prestressed concrete beams with non-prestressed

Sections D

k I k3 = 0.7

In Equation

and

cI = 0.85.

It was shown that the

ultimate moment of a given section in which

0.7 f' a b' + 0.85 f'

c (b c

b')t

the neutral axis falls below the flange can

be calculated by a simultaneous solution of

+ A' f* = A f
s su
s y

Equations 7a, 8a, 3, 4, 5, and 6.

e
= e
+ e
+ su
se
ce
a

For design purposes the ultimate moment

(d - a)

will be computed by Equation 7b assuming that

the stress

in compression steel has reached the

yield point, and taking

and k2 = 0.42.

In addition the expression for

the ultimate moment will be multiplied by the

capacity reduction factor

Mu =

fsu = F(e su)

cI = 0.85, c2 = 0.5,

[A f

Elimination of

between Equations 8b

and 3 will result in the following:

t.

(d - 0.42a)

su

0.70 f' e (b'/b)

c u
p [

su

ese -

+ e

ce

fl
+ 0.85 f' (b - b')
c

+ 0.85

t (0.42a - 0.5t)

+ A' f* (0.42a - d')]

s y

A'

(1 - b-) b

s.

(27)

yA

The condition that the compression steel

has yielded is satisfied by the following

where

f* = f - 0.85 f',
and f
is the
y
y
c
y
yield point of non-prestressed compression

inequality:

steel.

e
= u
su
a

(a - d') > e .
- y

Here it is assumed that non-prestressed

compression steel

is American Society for Test-

We will substitute for

from Equation

ing and Materials Billet Steel A-15 with a flat

8b in the above inequality and rearrange it to

stress-strain diagram beyond the yield point.

arrive at:

The stress

in the area of concrete replaced by

compression steel is taken into account by the

term 0.85 f'
c

d'

< (1 - -1)
e
u

which is an approximation.
- 1.21

F. Asf
s su
0.7 fS b'
c
- 1)t .

A' f*

s Y

0.7 f' b1
c

Hence for the solution of unknowns

f su, and esu
5,

Mu,

we have available Equations 26,

and 27.

The preceding equations were developed

for the case of a T flanged section in which
the neutral axis falls in the web.

Equation 26 may conveniently be expressed

This

condition can be stated

in the dimensionless form:

t > 1.4 (q - qi)
S=q

1 + (1 - -)

(---

1)

When the above inequality is not satisfied,

c
C
-

the neutral axis falls in the flange and the

0.59 q (1 -

-)

flanged section becomes a rectangular section.

For this case,

2
0.85 () 2b

-q d' 1

where

d.

2d

f
= p -

(26a)

1)

-*bd 2f

, 2

= Q = q [I - 0.59q

(1 -

-)
q

, = bs f

bd fl
c

0.7 e
e

(b'/b)

- ese

su

+ 0.85 (1 -

ce

-L)

(26b)

0.7 e
C
u
e
- e
- e
+ e
su
se
ce
u
0.7 e
q max = q' +
e -u
s
se
ce
q = q

and

and Equations 26a,

27a, and 28a yield

M

c
A' f*

b' = b

(27c)

(27d)
u

d' < 1.4d (1 - -Y) (q - q').

(27a)

(28b)

Expression 28 can similarly be presented

in a dimensionless form as follows:
e
d'

< d (1 -

)[(q-

q')

1.21 -

1)].

it is possible to increase

b
0.7 b'
(28a )

consistent with the required

ductility corresponding to

quired ductility is high.

= es

is

section, the following example is presented.

It is required to design the section in

+ o.85 (1 -

se

ce

.)

"Illustrative Example 1" in such a way that

(-

+-

-6e

-6

si

In order to show that the non-prestressed

ductility without increasing the area of the

0.7 e
e

ILLUSTRATIVE EXAMPLE 2

compressive reinforcement increases the

given by the following:

max

for a ductility corresponding to

(27b )

when

the beam has no non-prestressed

compressive reinforcement, and Equations 26a,

27a, and 27b become identical with Equations
20a, 19a, and 19b respectively.

= 0.03,

the area of the section will be the same as

that for a ductility corresponding to

q' = 0,

q'.

This relationship is very useful when the re(

The expression for the required maximum

q

q,

crease the area of the beam, by increasing

B.
value of

>
su
ss
hence to de-

required ductility corresponding to

Equation 27a implies that for a large

su

= 0.01.
The yield point stress of the compressive
reinforcement may be assumed as f = 50 ksi.
Thedesigned
section
in part 1 of
The section designed in part 1 of

"Illustrative Example 1" has a ductility corre-

ultimate.

sponding to

Expression 28b:

= 0.01.

The problem is to

This condition can be checked using

determine how much compressive steel of the

d'

type given should be placed, so that the

ductility of the section will
sponding to

= 2 in. < (1.4) (32.4) (1 -

0.00167

reach that corre(0.181 - 0.084) = 2.6 in.

= 0.03.

The neutral axis was determined in Example

1, part 2 for the same required

as being

at a distance from the top fiber given by

4.47 in.

In this case

a < t = 6 in.

and the

T section behaves as a rectangular beam.

Equations

7c and 26b with

Example 1, part 1, and

Q = 0.170

d' = 2 in.

ing independent relations between

a =

Using

as in

the followq and q'

If

d'

greater than 2.6 inches had been

selected, compressive steel at ultimate would

not yield, requiring use of the actual value
of stress in compression steel which is less
than the yield stress.
The length of these non-prestressed bars
need not be the total span of the beam.

are obtained:

Theo-

retically they are not needed at a section

where the required

(0.7)(0.004)
q = q1 + q0.03
- 0.0044 - 0.0006 + 0.004

is that of the section

without the compression reinforcement.

ing the distribution of the required

0.170 = q - 0.59 (q - q')

- q'

.2

Assumq

to be

the same as the distribution of bending

moments, the theoretical section at which the

Solution of the above equations yields

q = 0.181

and

bars are no longer needed can be determined by

q1 = 0.084.

the distance

From Figure 7, the stress-strain diagram

for steel,
mately to

S= q

= 0.03

from the center line as

follows:

corresponds approxi-

fsu = 228 ksi.

su

XI

We have

(0.18)(5) = 0.00397

54
=-2-

0.084

18

,
.
18.4 ft.

Further economy can be achieved if the nonprestressed compression bars are separated in
two groups.

= (0.00397)(32.4)(16.3)

= 2.10 sq.

in.

A group of three short bars re-

presenting a

of .050 could then be cut at

a section theoretically at a distance

(Use fifteen 1/2-inch strands)

14.2 feet

Also

from the center line.

X2 =

Taking into

consideration the additional length required

P' = q' f/f

= 50-(0.85)(5)

= 0.0093

to develop bond the non-prestressed bars may

be specified as

or

2 # 9 x 40'

and

3 # 9 x 32'.

Figures 10 and 11 show the section of the

A' = (0.0093)(32.4)(16.3)

= 4.9 in.

beam at midspan and at the end respectively.

Figure 12 shows the profile of the prestressed

(Use five #9 bars of hard grade steel)

and non-prestressed steel.

The distance of these bars from the top

have been draped to prevent overstressing of

e' > e ,
if yielding of
y
s
the compressive reinforcement is to occur at

must be such that

Three web strands

the end sections of the beam.

In addition to

the 5 # 9 non-prestressed top reinforcement

an end # 5 has been added for practical construction purposes.

Stirrups have been de-

increase in ductility is accompanied with an

increase in stress in steel at ultimate.

For

signed according to American Concrete Institute

the larger ductility considered here the stress

Code (318-63).

in steel increases at ultimate from 214 ksi

As before the properties of

the transformed section as well as the stresses

to 228 ksi.

before and after losses are given in the table.

causes a decrease in the required area of

prestressing steel.

A reduction in the amount of non-prestressed compression reinforcement is possible

with a section having a wider top flange.
parameter

q'

is related to

Selection of a smaller value of

q'/q

would fix

The

by Equation

27b.

q'

or

and permit the determi-

nation of the required

This increase in steel stress

The beam of Example 1, part 2 shows that

by increasing the width of the top flange and
thereby adding concrete area to the compression
zone, high ductility can be obtained.

This,

however, increases the weight of the section

by Equation 26b.

by 26 per cent, but decreases the amount of

The area of the section and its final shape can

prestressing steel to 16 strands.

be determined as usual from Equation 24.

If

crease in stress in steel at ultimate not

the proper values of

were

only supports the additional weight of the

t/h, b'/b,

and

The in-

selected the new section will present a flange

beam, but also permits a reduction in the re-

wider than that of Example 1, part 1, but not

quired area of steel.

as large as that of Example 1, part 2.

this beam shows, however, a tendency for a

Also

Under the service loads

the compressive reinforcement required will be

large tensile stress at the bottom fiber due

This solution

to the smaller amount of prestressing force.

smaller than that of Example 2.

would show that to obtain high ductility a

The beam of Example 2 shows a different

compromise section can be obtained if some

way of obtaining high ductility.

increment of weight is tolerated with a smaller

are added to the top flange of the low

amount of non-prestressed compression steel.

ductility section of Example 1, part 1.

C.

Five #9 bars

This

COMPARISON OF THE THREE SOLUTIONS

increment in compression area raises the

It has been shown that ultimate strength

neutral axis and increases the lever arm of

design provides a convenient procedure which

leads to well proportioned sections.

The de-

sired ductility and strength were used as the

fundamental

constraints for proportioning the

the resisting couple by approximately five

per cent.

In addition the stress in the

steel at ultimate is increased from 214 ksi

to 228 ksi, approximately seven per cent.

sections, while the stresses at transfer and

These two factors combined explain the 12 per

under service loads were checked.

cent reduction in the number of prestressing

An examination of Table I shows interesting details.

The beam of Example 1, part I

strands

(from 17 to 15),

since the required

tensile force at ultimate can be obtained with

with a required ductility corresponding to

less area of steel at a higher stress and a

larger lever arm.

= 0.01

required more prestressing steel

The non-prestressed bars

(17 strands) than the beams of Example 1,

also provide additional tensile strength for

part 2 and Example 2 with a required ductility

the top part of the beam at transfer and during

corresponding to

= 0.03.

For the stress-strain diagram of prestressed steel adopted in these examples, any

handling operations.

Furthermore, they have a

tendency to reduce the inelastic deflections

due to creep.

VI. SUMMARY AND CONCLUSIONS

The work reported here presents a study

small and may be ignored.

However, neglect-

of ultimate design of prestressed concrete

ing the effect of compression steel upon the

beams.

ductility of the section is unreasonable.

It consists of a detailed discussion of

various methods for calculating ultimate moment

Compression steel contributes appreciably to

of practical sections including sections with

the ductility of the section and should be

non-prestressed compression reinforcement.

taken into account.

"Illustrative Example 2"

method is presented by which a prestressed

shows that the most expeditious way for in-

concrete beam can be proportioned by ultimate

creasing the ductility of a section is by

design.

placing non-prestressed compression reinforce-

Particular emphasis has been placed

upon the requirement of ductility and its

influence upon the dimensions of the section.

ment as near the top fiber as possible.

3.

A prestressed concrete beam can be

The design examples presented show the actual

proportioned for a given required minimum

method of proportioning as well as the in-

flexural strength and ductility.

fluence of ductility on the dimensions of the

at transfer and at service conditions may be

beam.

checked in a section thus obtained.

The following conclusions may be drawn

from the study presented in this work.

1.

Methods with varying degrees of

4.

The stresses

The dimensions of a section are in-

fluenced greatly by the required ductility.

An increase in the required ductility results

accuracy can be developed for the determination

in an increase in the required area of the

of ultimate moment in terms of the properties

section, unless compression steel

of the beam section.

For design purposes the

5.

is provided.

For a large required ductility con-

ultimate moment may be expressed conveniently

siderable saving in the area of the beam may

in a dimensionless form.

be effected by use of non-prestressed com-

2.

The expressions for the calculation

pression steel.

Compression steel has ad-

of the ultimate moment and ductility given in

ditional advantages such as its contribution

the American Concrete Institute Code (318-63)

to the crack stability of top fiber, its use

do not include the effect of non-prestressed

as spacer for the web reinforcement and its

compression steel.

function in providing more safety for the beam

pression steel

The influence of com-

on the flexural strength is

during transportation and erection.

VII. REFERENCES

1.

2.

N. Khachaturian, 1. Ali, and L. T. Thorpe,

Analytical Studies of Relations Among
Various Design Criteria for Prestressed
Concrete,(Engineering Experiment Station
Bulletin No. 463).
Urbana, Ill.: University of Illinois College of Engineering,
1962.
J. Warwaruk, M. A. Sozen, and C. P. Siess,
Strength and Behavior in Flexure of Prestressed Concrete Beams, (Engineering
Experiment Station Bulletin No. 464).
Urbana, Ill.:
University of Illinois
College of Engineering, 1962.

3.

E. Hognestad, N. W. Hanson, and D. McHenry,

"Concrete Stress Distribution in Ultimate
Strength Design," Journal of the American
Concrete Institute, December 1955, Proceedings, Vol. 52, pp. 455-479.

4.

Building Code Requirements for Reinforced

Concrete, American Concrete Institute
Code (318-63), June 1963.