Flexural Ductility of

Structural Concrete Sections

This paper investigates in detail the flexural
ductility of structural concrete sections, i.e.,
reinforced concrete (RC), prestressed con-
crete (PC) and partially prestressed concrete
(PPC) sections. Flexural ductility is defined as
the ratio of the section ultimate and yield cur-
vatures. Realistic, rather than conventional,
standard estimates are attempted on the
basis of actual material behavior, consistent
description of all behavior states, and unified
M.Z.Cohn
Professor
analysis of structural concrete sections. The
Department of Civil Engineering effects of section shape, degree of prestress-
University of Waterloo ing, material laws, compression reinforce-
Ontario, Canada
ment, and other parameters are considered.
Finally, some ACI Code revisions are pro-
posed.

uctility of structural concrete sections ensures suf-

D ficient deforrnability and avoids premature failure

in either tension or compression zones (breaking of
reinforcement or crushing of concrete, respectively).
Thus, it may be considered a criterion for determining fea-
sible ranges of reinforcement percentages, between some
minimum and maximum (or balanced) values.
However, although reinforcement limitations are re-
Paolo Riva lated to flexural ductility, they are not its direct expression
Post-Doctoral Fellow or quantitative measure, but only reflect minimum ductil-
Department of Civil Engineering ity requirements implied by standard codes. Flexural duc-
University of Waterloo tility is usually defined by the ratio of ultimate to yield
Ontario, Canada
curvatures, and, hence, it is an intrinsic property of struc-
tural concrete sections.
Flexural ductility of structural concrete has been widely
investigated'-' 0 because of its implications in inelastic mo-
ment redistribution, energy absorbing and earthquake re-
sistant structural design, joint detailing of precast RC/PC
structures, and other applications. Some important results

72 PCI JOURNAL
from analytical and experimental investigations of the
flexural behavior of reinforced concrete (RC), prestressed
concrete (PC) and partially prestressed concrete (PPC)
sections in general and their ductility in particular have
been reported in a number of papers published in the PCI
JOURNAL. 5•8•9
Analytical investigations are based on either code spec-
ified or more realistic behavior models. To various de-
grees, structural codes:
• Neglect the need for consistent and continuous coverage
of structural concrete throughout all behavior states.
• Do not offer a unified treatment for RC, PC, and PPC.
• Do not consider all relevant factors. Fig. 1. Typical element for which the local moment-curvature
• Perpetuate some questionable assumptions, definitions, law is studied.
and design trends.
The first three points are self evident. The last one is view to removing some limitations of past investigations.
illustrated by a recent study8 which proposes a replace- Specifically, it is our aim to predict ductility factors, based
ment of existing upper limits of reinforcement: on consistent assumptions throughout all behavior states,
(p- p') /ph 5, 0.5 or (J)P + (dm Jdps) ( ro- ro' ) 5, 0.24B 1 valid equally for RC, PC and PPC sections, and consider-
for RC and PC in ACI 318-83, 11 and c!d < 600/(600 +/y) in ing all governing parameters. The objective is to offer a
the CSA CodeY The recommended upper limit 8 c/h coherent theoretical understanding of inelastic section be-
5, 120Ecu replaces these code maximum values ofthe steel havior, along with some reliable conclusions for structural
reinforcement, and is not a direct expression of section engineering practice.
ductility. Despite its generality and wide acceptance, 12·13 \
the superiority of the cJd or cJ h limitation over the one in p ANALYTICAL MODEL
or ro is not evident. Indeed, for a given section, c is un-
known before the calculation of the resisting moment, The moment-curvature (M-<1>) constitutive law pre-
whereas p or ro are basic design (geometry and material) sented in Refs. 3, 6 and 14 has been adopted for the study
data. Hence, c is adequate for analysis, but inconvenient
for design, when the section geometry and reinforcement
amount are not yet known.
For all its simplicity and versatility, the proposed limita-
tion clh 5, 120Ecu only reflects some specified ultimate de-
formation of concrete and no other relevant parameters.
Again, flexural ductility is an intrinsic property of struc-
tural concrete sections. As such it cannot properly be de-
fined by standard limitations, but rather by the basic gov-
erning parameters, i.e., section geometry and mechanical
properties of materials.
Structural designers may require some documented in-
formation on:
1. Flexural behavior (M-<1> curves), based on realistic
models of material behavior.
2. Behavior throughout States I, II and III (i.e., un-
cracked, cracked and post-yield, respectively).
3. Ductility ofRC, PPC and PC sections.
4. Effects of section shape, degree of prestressing, ma-
terial 0'-£ laws, in addition to material strengths,
flange width, and compression to tension steel ratios,
which have previously been investigated.
5. Lower and upper bounds of the effective flexural
ductility that can be counted upon in design, when
current practice is followed (i.e., realistic p, p',/y and
t: values, and other parameters). wife
An attempt to provide such information has been pre- (b)
sented in a comprehensive analytical study published else-
where.3 As a followup, this paper explores in detail the Fig. 2. Constitutive laws for concrete: (a) compression; 7 (b)
flexural ductility of structural concrete sections with a tension. 18

March-April 1991 73
of the ductility properties of structural concrete sections. ing relation after the onset of cracking.
This M- cp law is based on realistic material laws and takes The stress-strain laws adopted for the mild and reinforc-
into consideration the effects of cracking, steel-concrete ing steel and prestressing steel are also those proposed by
interaction and tension stiffening through the analytical Sargin.? A third linear branch has been added to the origi-
model developed in Refs. 15 and 16 for RC elements and nal <J-£ law for the prestressing steel. This linear branch
later extended to PC or PPC elements. 17 In the following, starts at the stress/;, for which the tangent to the nonlinear
only the main assumptions inherent to the moment-curva- portion of the analytical model passes through the ultimate
ture law are recalled. strength point (Fig. 3 and Ref. 3).
The models assumed to represent the bond of mild rein-
Assumptions forcing steel and prestressing steel are shown in Fig. 4. 17
1. Quasi-static (monotonic, nonrepeated, nonreversi-
ble) loading. Computational Features
2. Negligible shear effects. A general computer program, MOCURO (MOment
3. Linear strain distribution. CUrvature ROtation), has been developed to automati-
4. Known material stress-strain relationships (analyti- cally handle the governing conditions of section response
cal, experimental, and other studies). at all loading states.6 Any symmetrical concrete section
5. Uni-axial stress-strain laws valid for section analysis. with multiple layers of mild reinforcing steel and/or pre-
The moment-curvature law is determined from the stressing steel, under either negative or positive moment,
study of an element of length equal to the crack spacing lc , may be analyzed. The program accepts any experimental,
assuming the moment is constant along the element and analytical or assumed point by point material constitutive
expressing the compatibility and equilibrium conditions at law.
the cracked section B (Fig. 1). The curvature is defined as Initially, the program computes the cracking, yielding
the ratio between the relative rotation of two sections (A and ultimate limit states. Successively, any desired num-
and C in Fig. 1) and the crack spacing lc. ber of points of theM- cp curve in the first (uncracked), sec-
ond (cracked) and third (post-yielding) states can be com-
Material Laws puted by imposing the curvature and solving the
equilibrium equations by an iterative procedure.
The material laws adopted for the section analysis are
The output includes a summary of the section data, the
shown in Figs. 2 to 4, along with the relevant notation.
The stress-strain relationship adopted for concrete in
compression was proposed by Sargin,7 and accounted for
the effects of the loading rate and duration, degree of con-
finement, strain at peak, strain softening effect and other
factors.
The law for concrete in tension (Fig. 2b) is based on ex-
perimental results. 18 It assumes a parabolic stress-strain
relation up to cracking and a hyperbolic stress-crack open-
To
IDEALIZED BEHAVIOUR

(a)
PRESTRESS lNG STEEL

Tp~------------------

Sp
(b)

Fig. 3. Constitutive laws for reinforcing steel Fig. 4. Bond-slip constitutive relations: 17 (a) reinforcing steel;
(nonprestressed) and prestressing steel. 7 (b) prestressing steel.

74 PCIJOURNAL
elastic stiffness, the cracking, yielding, and ultimate mo- values for each computed point of theM-cj> law.
ments and curvatures, the ductility factor cl>u /cj>Y and M, M I A complete description of the program characteristics
(J:bd 2), cj>, cj>lc, the tangent stiffness, reinforcing steel and and of the input-output procedures are given in the
prestressing steel stresses, as well as the crack opening MOCURO-User's Manual. 19

Table 1 Numerical constants for material laws

CONCRETE
COMPRESSION ~ ·~E == 40 MPo
29930 MPa
;
lo = 0.00264
A iii 2.5 i
0:.0.362
k3•0.8
c

~
CONCRETE c2 • 2
fct = 4.5 MPa ;
TENSION c1 I lc • 12000

MILD
STEEL
""" t=_ f1 y = 400 MPa
E1 = 200000 MPo'
f10 • 600 MPo
lsu • 7%
l Sh = 1%
E1 h= 6500MPa

MILD STEEL
BOND t= t'o = 3 MPo ;
T10 • 10 MPo

Su • 0.5 mm

~
fpop • I 300 MPo fp I •1580 MPo, ~, = 1% f. • 17 40MPa
PRESTRESSING p
STEEL EP •190000 MPo fpu •1860 MPo, ~pu • 3.5°/o fp • 1.9%

PRESTRESSING
STEEL
BONO c T'p • 4 MPa

Table 2 Definitions of reinforcement index yielding and ultimate conditions

'
AU'l110R Ill
+y +u

- 0.35%

[:: - J
£c
A,f"+A,f,
CEB- ttC 78 - 'r or £ - 1.00%
bdf.' [
PY •
or lit p - £ p -[ pe - 1.00%

ACI
(318 - 83)
A,f,.+A,f,
btl f.' ('· .,- /cr
[p p py >
) £c - 0.3%

A,/,.+A,f. £. • £ • y # 1
COHN,
BARTLETT b df.' £
p
y
• E(fp,0.05)' y • 1 -3M
3+
- 0

NAAMAN,
HARAJLI,
WIGHT
A,f,_+A,/,
b df.' Mn
A 8 C Dl' -3+3M
3M
- 0

COHN,
_ A,f,. + A.f_,
q- b.. df.'< 1:. - £y - 0.27., y # 1 -3+ - 0
RIVA or t 81 • lltp :. 3%
ll£
p - £ p-£ pe- 0.2%, y - 1 - £ pu -£ pe
f . , • a(£ 8 ~ 3%) or £ p • Epu
1 = 3.5%

) • implied definition

March-Apri11991 75
 PARAMETERS GOVERNING the section shape, reinforcement index (ro), and mixed re-
FLEXURAL DUCTILITY inforcement index (y). Various definitions of ro and of the
yielding and ultimate curvatures (q>y and<!>.), can be found
A comprehensive study was conducted to ascertain the in the literature. A summary of some available definitions
influence of the basic material and geometric parameters is given in Table 2.
that govern the behavior of flexural elements on both the The yielding curvature <I>Y is defined as the curvature at
M-<!>constitutive law and the ductility factor<!>. /<I>Y. which the transition from the second (cracked) to the third
The material parameters are: cr-E law for concrete in (post-yielding) state of the moment-curvature law takes
compression; cr-E law for concrete in tension; cr-E law for place. In reinforced concrete sections, this is a well de-
mild steel; cr-E law for prestressing steel; bond-slip consti- fined point that coincides with yielding of the reinforcing
tutive law for mild reinforcing steel and prestressing steel. steel. In partially and fully prestressed sections, it is not
The geometric parameters are: shape of the section; ten- possible to identify a unique transition point: in these
sion reinforcement index; compression reinforcement cases the curvature <I>Y has only a conventional meaning,
index; prestressing to total reinforcing steel ratio (mixed which is especially useful for the definition ofthe ductility
reinforcement index); degree of prestressing; and stirrup factor <l>J<l>Y. Various defmitions of <l>Y have been proposed
percentage. in the literature (Table 2); in this study the yielding curva-
The loading parameters (axial loading, loading repeti- ture <l>Y of reinforced and partially prestressed sections ("( <
tion, duration and reversal) have not been investigated in 1) is the curvature at which reinforcing steel reaches its
this study. yield strain Ey = 0.2 percent.
Unless otherwise specified, the material parameters The yielq curvature of fully prestressed concrete sec-
have been kept constant. The values assumed for the pa- tions (y= 1) is the curvature at which the strain increment
rameters involved in the cr-E laws (Figs. 2 through 4) are of prestressing steel from its effective value equals the
summarized in Table 1. yield strain of reinforcing steel (i.e., ~EP = EP- epe = Ey = 0.2
The most important geometric parameters appear to be percent, where EP, is the effective prestressing steel strain,

Table 3. Specimen data for parametric study, effects of section shape, q, y, t;.
SECTION A SECTION B SECTION C SECTION D SECTION E SECTION F
K •I

~~ TT~~ ]Ij ~ ~

S!Tj
bw• !500

tiD
r-IOOOjj
IT

'~ ~50 H:W'

•10 (i) 300
STIRRUPS
~0
_j
n=
~o
IT
fa' fp SEE TABLE I
2~ E
•
11,•250 Q~ __t
~8
"'Jo oo '
'• • 0.5 lli111 ALL CASES
U: _t iL- u____: il___:
t0 -!500~ -w.. --l!lool- ~ ...!----!.
on .. •Z!IO ••• 250
., 0 O.n ~ .nO
., o.n ono ono ono one ono onO
NO') o- oneN ~~ 0 - !!~ ~~ 0 - !! ~ N.., 0 - - N ~~
lc
""""
fc TaiMPal Tp
IMP•I Te T• IMPel ~ 0
0 00
NN ~8
00 00 00 0 00 00
-
00 00 00 oo 00 00 oo 00 oo co oo
100 40 5 10 4
0.0 • • • • • •
1.0 • • • • • •
50 :s 10 4
0.0 •••••• •• ••• •
1.0 •• •• •• ••• •••
2 7 4
0.0 • • •
o.e • • •
5
0.5 • • •
1.0 • • •
0.0 •••••• • ••• • •••• ••• •• •••• ••• ••••• ••••
0.2 •••••• • •••• •••• ••• • • •••• ••• • • •• • •• • •
zoo 40 :s 10 4
0.4 •••• • • • •••• • ••• • •• • • •••• • •• • • •• • •• ••
0.6 •••••• • •• • • •••• ••• •• •••• ••• • ••• • • • ••
0.8 •••••• • •••• •••• ••• •• •••• ••• • • • • • • • • •
1.0 •• •••• • •••• •••• • •• •• • ••• ••• • ••• ••• ••
5
0.5 • • •
I .0 • • •
4 13 4
0.0 • • •
0.5 • • •
50 3 10 4
00 • • • • • •
1.0 • • • • • •
60 3 10 4
0.0 • • • • • •
1.0 • • • • • •
300 40 3 10 4
0.0 • • • • • •
1.0 • • • • • •
76 PCIJOURNAL
·and f y is the reinforcing steel yield strain). This definition higher than the selected limiting strain (since the ultimate
has the advantage of simplicity and ensures a consistent strain for mild steel is much higher than 3 percent, and can
yielding behavior description for RC, PPC and PC sec- be as high as 15 + 20 percent), the adopted definition has
tions. the advantage of ensuring a consistent description of the
Various definitions of the ultimate curvature <Pu are behavior of RC, PPC and PC sections at the ultimate limit
available in the literature (Table 2). Typically, the defini- state.
tion of the ultimate limit state in building codes is based on The reinforcement index essentially reflects the ulti-
the assumption of a limit strain for concrete (ACI 318- mate flexural behavior of concrete sections/ and its defi-
8311) or for both concrete and steel (CEB-MC 78 13). On the nition should depend on that assumed for the ultimate
other hand, a frequently used definition is that the ultimate limit state of the section. As one of the main parameters
limit state corresponds to the maximum moment capac- that characterizes the section behavior, ro is especially
ity.2,s useful in the design process. Because the final design and
In this study the latter definition has been adopted for actual response are not known, the definition of ro should
sections in which the ultimate condition is characterized not be dependeQt on the steel stresses at the ultimate limit
by concrete crushing. For all other sections, the ultimate state and should only reflect the material characteristics
limit state is identified by the failure of prestressing steel and section geometry.
for PC or PPC sections (y :t:. 0) or by the reinforcing steel To avoid confusion between the definitions of the rein-
reaching a given limiting strain for RC sections (y = 0). forcement index commonly adopted in the literature and
This limiting strain is assumed approximately equal to the the definition adopted in this study, the former (listed in
strain increment of prestressing steel from its effective to Table 2) are indicated by the symbol ro, and the latter will
its ultimate value in an equivalent PC or PPC section, i.e., be referred to as q, where:
£ s1 =~£P = f pu - f pe ~ 3 percent.
Although in some cases the actual ultimate limit state of (1)
RC sections could occur for reinforcing steel strains

Table 4. Specimen data for parametric study, effects of analytical model, compression flange width and
compression reinforcement.

SECTION A SECTION C SECTION F

,. . ~
jiOO~_j
0
on
bw • ~00

rr~~
il
,., .
rr[E•..
40 MPo

nco
c
400 MPo oo t oo f
2~L ono
,,,"" . II I 6
f • 1860 MPo zso o2
MPo ( K • 1) ~ _t IL.: 1L
w~
'
~~ o-
~
100 10 0
....+--+.-
••• zso
~~
10
A1 1Ap

~~
ltZ
100 100 0 10 0
blbw 0- - N - N 0 -: -: 1\j
MODEL
""• 00 00 oo 00 0 0 00 d 00 0 00
1. 0 00 • • ••• •
1. 5 0 .0 • • ••••
2 .0 0 .0 • •• • • •
2 .5 00 • •• • • •
0.0 30 o.o ••• • • •
3.5 0.0 •• • • • •
COHN 40 0 .0 •• • • • •
RIVA 4 .5 0 .0 ••• • • •
50 0 .0 ••• • • •
0 .00 - 0 .0 • • • • • •
0 .25 - 0 .0 • • • • • •
0 .50 - 0 .0 • • • •• •
0 .75 - 0 .0 • • • • • •
1.00 - 0 .0 • • • • ••
COHN
0.0 - 0 .0 ••• ••• • • • •• •
RIVA 1.0 • •
• ••• • • • •• •
ACI 0.0 - 0 .0 ••• ••• • • • •• •
1.0 ••• ••• • • • •• •
CEI 0 .0 - 0.0 ••• ••• • • • •• •
1.0 ••• ••• • • • • ••

March-April1991 77
 Consistent with the proposed definition of the ultimate tion of prestressing steel to the total steel in a section:
limit state, the reinforcing steel stress..f. at this limit state is
assumed equal to the steel stress at the proposed limit "(= Ap/pu (2)
strain,..f.1 =a, (e,1 ), and, therefore, it is lower than the ulti- Ap/pu + As..f.I
mate stress of the material, ..f..= a, (e,.).In the above def-
inition, q is based on the specified ultimate prestressing This definition implies that, for sections with a given q
steel and mild steel stresses,,[p. and ..fs1 , respectively. value and whose ultimate limit state corresponds to the
The other ro definitions in Table 2 are based on either steel reaching its limiting strain, the ultimate moment
the prestressing steel and reinforcing steel stresses at the value is independent of y.
ultimate limit state of the section,,[p, and..f., respectively, 2•5 Finally, the degree of prestressing K is defined2 as:
or the steel yielding stresses,,[py and [y .11 •13 The proposed q
is referred to the web width of the concrete section, bw, "K=~ (3)
/pa
rather than to the width of the compression zone, b. This
more explicitly illustrates the effect of the flange width on where,[p, and,{pa are the effective and admissible (i.e., ser-
the ductility of structural concrete sections. vice stress,,[pa =0.6 ,{p.) prestressing steel stresses, respec-
The mixed reinforcement index2 y expresses the propor- tively.

c
M ..,c
PC
r =1.0
~
N ""
N .. ····0 --B£.. r =0.0
N N
-~~~
~u

...'-
"CN
- N
-~~~
~u

'-
"CN
- 0.3
.................. o

0.2
.J:l- .J:l-
......
I:
ID
... ...... ···················
~

q: 0.1 c: ......
I:
ID
··········· ...............
q =0.1 c.
c c
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
e>d (%) e>d 1%1
~ ~

-
~
N N

N N .....o
-=-
"t_u
-=
"'(._U

u
"'oN '\JN
.J:l- .J:l-
...... . ........ ······· ...... -
I:
ID q: 0.1 c: %:
ID
·········· . ...............

c c
o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
c 4ld ( I.) 4ld (%)
M lij

~ ~
N N

N N ..... o
-!!!
t_u
..."CN
······
-m
(._U
N
- ........... 0

.J:J-
"ON
.J:J- .. 0.2
··············· . ............
......
%:
ID q: 0.1 c: ......
%:
ID
. ......................

D
c c
o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 o.s 1.0 1.5 2.0 2.5 3.0 J.S
4ld (I.) e>d t%1

Fig. 5. Effect of section shape on the moment-curvature relationship.

78 PCIJOURNAL
 PARAMETRIC STUDY OF
were examined. 3
't., 'tP
FLEXURAL DUCTILITY
Table 4 identifies the specimens investigated for com-
A set of 468 computer simulated tests was performed to paring the proposed analytical model with those implied
study the influence of geometric parameters, material by the ACI Standard 11 and the CEB Model Code, 13 as well
characteristics and various analytical models on the mo- as for emphasizing the major effect of compression flange
ment-curvature law and ductility factor. The program of width and compression reinforcement on the ductility fac-
the parametric study is summarized in Tables 3 and 4, tor.
where black dots indicate the combinations of parameters For all tests, the stirrup area and spacing, and degree of
investigated. prestressing K ( K =1, withfpe =/pa =0.6 /pu ), are constant.
Table 3 assembles the data of specimens selected For all cases studied the yielding point is assumed to cor-
mainly for studying the effect of section shape (A ... F), respond to Es =Ey =0.2 percent if y < 1, or~ Ep =EP- Epe =
degree of prestressing y (0 ... 1), reinforcement index q 0.2 percent if y = 1. The ultimate limit strains for mild re-
t:
(0.05 ... 0.30) and concrete grade, (30 ... 60 MPa or inforcing steel and prestressing steel are assumed to corre-
4.35 ... 8.70 ksi). In addition, the influence of crack spac- spond to Es1 = 3.0 percent and Epu = 3.5 percent, respec-
ing, lc [100, 200, 300 mm (4, 8, 12 in.)] and bond stresses tively (Table 2).

0 ,.,0
""
~
N "'
N
;-..: ;-..:

.."''N-
-CI)
-~
.. u
~ u~~).'K' ."''N
'- u'.~).'K' ~~)
..c- ..c-
......
......
I: I:
LD CD
lJ). 0) q -o.1o lF ).0)-
0 0
o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
~d (%) ~d (%1
,.,0 ,.,
0

•
N "'
N
if'.~).df' -
;-..: ;-..:
-CI)

,.c_u - .
~
-~

"''N "''N
..c-
......
..c-
......
I: I:
LD 1-o.oo CD 1!-].00
q -o.2o q-0.20
Q 0
o.o 0.5 1.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 J.S
~d I I. I
Cl ,.,
0

""
.... ~ •
ir~~).'K'
N

N ;-..:

- ...c--
-CI) '-'II)

""u
~ '-
']~ lF) .0) "''N
...... ......
I: 1-o.oo I: '8-l.OO
LD LD
q-0.30 q-0.30
Cl 0
o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 o.s 1.0 1.5 2.0 2.5 3.0 J.S
'd (%) 'd (I. I

( 0) {b)
Fig. 6. Effect of section shape on the moment-curvature relationship: (a) reinforced concrete sections; (b) fully
prestressed concrete sections.

March-April1991 79
Influence of Section Shape, q and y • For Sections A, B and C, <1>. /<I>Y is scarcely affected by
A thorough discussion on the influence of these param- q. For these sections, assuming a constant value of the
eters on theM- <1> law and a complete set of plots is given ductility factor approximately equal to 14 appears to be
elsewhere.3•6 The present discussion is focused on the ef- conservative for any amount of reinforcement.
fects of these parameters on the ductility factor <1>. /<I>Y only. • For rectangular and T -sections under negative mo-
Figs. 5 and 6 show the influence of the section shape, ment (Sections F and E), the ductility factor shows a hy-
reinforcement index q, and mixed reinforcement index y perbolic variation with q. In these cases <1>. /<I>Y varies be-
on the moment-curvature law. They demonstrate that PC tween approximately 14 (for q = 0.05) and 2.5 (for q ~
sections have a flatter inelastic response than RC sections, 0.25).
which display a more pronounced strain hardening. • The different trends of the <1>. /<I>Y curves may be ex-
The influence of the section shape on <1>. /<I>Y is illustrated plained by the varying types of failure that characterize the
by Fig. 7. The combined effects of the section shape, q, ultimate limit state of various sections. For wide flanged
andy, shown in Fig. 8, may be summarized as follows: sections (A, B and C), an almost constant ultimate curva-
• The ductility is generally improved by the introduc- ture is obtained because failure is always governed by the
tion of prestressing, as already shown. 2•5 For design prac- steel reaching its limit strain. For the T -section under neg-
tice, it is conservative to consider the ductility factor of an ative moment (E) and rectangular section (F), where fail-
RC section for any PPC or PC section with the same q. ure is governed by concrete crushing, curvatures are re-

c c
"' "'
1/) Ill

~ ~
>- >o
&c eo
~-
&
}-
L/)
~xA 1/)

c.1f D
c c
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
q q
F;l F;l

L/) ~I Ill
k±ae 1 a
.... Y=O >o
&c eo
)- ~-
Y=o
L/)

c
~Jrs
Ln

0
(.LJE
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
q q
c c
"' "'

,..
~
~Y=O
se ~~
~

>o
&0 ec
~-
&
}-
Ill

0
o.o
~LJ
0.1
c
0.2 0.3
Ill

0
cO
o.o
F

0.1 0.2 0.3

q q

Fig. 7. Effect of mixed reinforcement index on the ductility factorvs. mechanical steel percentage relationship.

80 PCIJOURNAL
 Influence of Flange Width and
Compression Reinforcement
A set of numerical tests was performed on a T -section
and on a doubly reinforced rectangular section (Sections C
and F, Table 4) to assess and compare the influence of
A; !As on the ductility factor$" !$y. Only the casey= 0 has
been studied, as previous results (Fig. 7) have indicated
that yhas little influence on the ductility and that the duc-
tility factor of RC sections gives a conservative estimate
of G>u !$y for any value of y. Figs. 9a and 9b summarize the
effects of blbw and A; !As on G>u !$y.
o~--~----~~~~~~--~~-7~--~ Fig. 9a shows that increasing the size of the compres-
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
q sion flange increases the ductility of the section. For blbw =
1.0 the ductility factor varies hyperbolically with q. In-
Fig. 8. Effect of section shape on the ductility factor. creasing blbw to 4 results in the ductility factor becoming
almost constant and independent of q CG>u IG>Y"" 14). Values
of blbw higher than 4 do not affect the G>u IG>Y ratios.
duced by increasing q values. Section D shows a behavior The progressive change of behavior with the increase of
intermediate between the two preceding cases. More pre- blbw is related to the type of failure that characterizes the
cisely, for q ~ 0.15 the ultimate limit state is characterized ultimate limit state. For any considered q value, the ulti-
by steel failure and the ductility factor has a trend similar mate limit state corresponds to the concrete crushing for
to Sections A, B and C. For q > 0.15 failure is governed by blbw = 1 and to the steel reaching its limit strain for blbw
concrete crushing, resulting in a behavior analogous to ~ 4. For blbw values between 1 and 4, the ultimate limit
Sections E and F. state corresponds to the steel reaching its limit strain for
low q values and to the concrete crushing for high q val-
ues.
Comparison of Figs. 9a and 9b demonstrates that the ef-
0
N fect of A; !As on the ductility is similar to that of blbw. As
R.C. Y•O expected, the ductility factor increases A; !As values, and
becomes almost constant for A; !As~ 0.75. From Fig. 9a
-
.,., b/ b. =
5,4.5
4.0
we note that for A; !As = 0.25 the ductility factor G>u !G>y is
always higher than 5, even at rather high q (p) values (q =
-!{o 3.5
3.0
0.25 or p = 2 percent).
~- b

~i
2.5
.,., 2.0 Influence of the Analytical Model
....w.... 1.5 A set of tests with different analytical models was per-
b. 1.0
0
formed to compare results based on the ACI 318-83 11 and
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 CEB-MC 78 13 assumptions and definitions with those in
q
this investigation.
(a) The ACI Code allows the use of any realistic material
law for the nonlinear analysis of concrete sections. The
0
N
code only prescribes a maximum allowable concrete
R.C. Y•O strain Ecu = 0.3 percent and suggests an unlimited elastic
perfectly plastic behavior for reinforcing steel. Accord-
~
A'1 1As• ingly, a comparison with the ACI standard model was per-
1.00 formed using for concrete under compression and pre-
~:: 0.75 stressing steel the same constitutive laws adopted
~
throughout this study, neglecting reinforcing steel strain

c~
0.50 hardening, and limiting the concrete ultimate strain to 0.3
"' 0.25 percent.
0.00 According to CEB-MC 78, a parabola-rectangle and an
0
0.25
elastic-plastic stress-strain relationship for concrete and
0.00 0.05 0.10 0.15 0.20 0.30 0.35
q mild steel, respectively, were adopted. For the prestress-
(b)
ing steel the material law given in the CEB-MC 78 was
adopted. In both the CEB and ACI models the tension
Fig. 9. Ductility factor vs. reinforcement index variation for stiffening effect is neglected.
RC and PC sections: (a) effect of flange to web width ratio; Only the I and rectangular sections (A and Fin Table 4)
(b) effect of compression reinforcement. were analyzed for the cases in which prestressing is either

March-Apri11991 81
 ~ r------------------------------,
..
N
..
N
-COHN-RIVA
- · • ·- AC I
~ ---CEB

.. -
-m
.. u
~
'-
"l:)N
.D-
'
~
____ . , . . . . . . . . . . c:
0.1
N
"tJN
..c-
'
~ ~~-~--=-·=-=·-=·-~oo_.=_~-~---·~;~-o-.o-o--~·[]
I ~-o.oo
I
0 0 ~--~--~--~----~--~--~--~
o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5
CDd 1%) ¢d 1%1
g ~----------------------------~
0
M

..
N
.
N

;-..:
-m

....c-
"i....U
"l:)N
----.0 ....... ·0 0.1
'
~ ,.
"' 11-1.00
0 L---~--~--~----L---~--~--~
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5
CDd ( %) CDd 1%1

(a) ( b)

Fig. 10. Moment-curvature relationships according to models by Cohn-Riva, ACI and CEB codes:
(a) 1-section; (b) rectangular section.

full or absent (y= 1 ory= 0, respectively). To obtain com- Some differences are noted for prestressed sections where
parable moment and curvature results as functions of q (or the ultimate moment according to ACI is slightly larger
co) and y, the reinforcement area for each section was de- because no limitations are imposed on the steel strain.
termined by defining co for the ACI and CEB models with • Comparing the ultimate moments according to the
respect to the web width bw . The moment-curvature plots ACI and CEB codes and those resulting from the proposed
in Fig. 10 and the ductility factor plots in Fig. 11 suggest analytical model, two different cases can be identified. For
the following remarks: sections whose ultimate limit state corresponds to the steel
• The ultimate moments according to ACI 318-83 and reaching its limit strain, as defined in this study, the ulti-
CEB-MC 78 models are generally in good agreement. mate moments are in good agreement with the code values

~ ---COHN- RIVA R.C. y =0 P.c. r .. ,

------ ACI
----CEB
.
ft.. 0 .&' 0
....
~- ~-

···.. lJ]
•=-===:t:==!'~~-!=.:=.~~~- .•~ •===- =•=;~=':·::.t=.:.:.:.~.~-............ ----
........... ·····•
0~--~----~--~--~----~--~--~ 0~--~----~--~----~--~~~~--~
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
q q
(o)
(b)

Fig. 11. Ductility factor vs. reinforcement index relationships according to models by Cohn-Riva, ACI and CEB codes:
(a) RC sections; (b) PC sections.

82 PCIJOURNAL
 ,..,c 0
1"1

...
N
..
N f~ =60,50, 40,30
~ ;-..!

-=- -= q: 0.3
..
-u
(,_

'"0 '"
0.2 ..
-u
(,_

"QN
0.2
~- ~-

'
I:
"' 0.1
-a-o.oo
c: '
I:
ID 0.1
-a-o.oo
c 0
o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
~d (%) ~d (%)
~ ..,
0

- .
;-..!
N
fc=Go 50
40
30
;-..!
...
N
I
fc =60,50, 40,30
-= -=
..
~
0.2
.-u
lx
(,_

"ON "QN
~- ~-

'
I: 0.1 '
E:
Ul
"' 1-1.00 1-1.00
0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 J.S
~d [%) Gld 1%)

(a) (b)
Fig. 12. Effect of concrete grade on the moment-curvature relationship: (a) 1-section; (b) rectangular section.

(Sections A and F with q or ro =0.10). However, for sec- codes and the definition of q adopted herein (Table 2) is
tions whose ultimate limit state is governed by concrete that, while both codes define ro with respect to the yield
crushing (Section F with q or ro > 0.10), the ultimate mo- stress of the steel, we define q with respect to a higher
ments resulting from analysis with the proposed model are stress value, equal to the stress at the proposed limiting
lower than those given by analyses according to the ACI strain. Hence, for an RC section with a given steel area we
and CEB models. can write:
These differences are mainly due to the adopted mate-
rial laws and the definition of q. The main difference be- Asfy h
ro = b d+' = --;:- q z 0.8 q (4)
tween the definitions of ro given by the ACI and CEB w ':lc Js/

~r--------------------------------, 0
N
_,!"_~-)'=I

~
c-O ..B.Ly •0

.... -
1lfo

"'
30
60
0'-----'----'-----...____.........____.____.......____. 0
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
q q

(o) (b)

Fig. 13. Effect of concrete grade on concrete and reinforcing steel strains at the ultimate limit state: (a) t; = 30 MPa (4.35
t;
ksi); (b) = 60 MPa (8.70 ksi).

March-April1991 83
 f~: 30
0 I 0
0 fc = 30 0
0 0

.••.. --- 6-
:I :I
··-·------. Y= I
~u ~ wu ~

lo
0 0 Y=O
min
c
8
0
)':0,1
0"""'" cJr c
8
0
0.0 0.1 0.2 0.3 o.o 0.1 0.2 0.3
...
~
q ....
0
q
c 0
Y= I A······•·······•······A·· .. •..... ·•
~ &.
f cI -- 30
....,
0
l"l
c - ...
:I
0
•
r=o :I
0
•

c. c.
"'... I'll
q
0
"',.. . I'll
0
0

cO
:I :I

"'
en-
~
0

8
0
f c1 : 30 cJ:C .,-
w ~
0

8
0
6
-- ·----- .• Y:
Y=o
I

0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3

q q
(a}

0 0

~ f I : 60 0
c c 0

~
"'ULn
~ ~
--· ---- ., .
- -
-- ~------•-·-·--•Y=t
0

§ §
c D =
min ec.. 0.45%
a ay. 0
-

~------~--------~--------~--~
0
0
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
q q
~ ~
0 r--------------------------------, 0 r--------------------------------,
A· .•..• A· •.• --·&- •. ···A·----· ... ___ Y= I A_ f cI : 60
8 8•
:::10 'A
:I
0
c. c.
Y=o
Col N
c
Col N
Cl
Cl
.
,.,· ~ 0
w• ~
Cl
····-·-· "A······· vI : I
Cl
8 0

0 ~------~--------~--------~--~ Cl ~------_.--------~--------._--~
0.0 0.1 0.2 0.3 0.0 0.1 0.2 0.3
q q
(b)

Fig. 14. Effect of concrete grade on the ductility factor: (a) !-section; (b) rectangular section.

84 PCI JOURNAL
 Thus, if two sections are such that q for the former is calculations, but are excessively conservative for defor-
equal to ro of the latter, the former will have less reinforc- mation analyses. This is particularly true for !-sections or
ing steel, and, if Esu < Est , its ultimate moment will be lightly reinforced rectangular sections, when the actual
lower. failure takes place for steel strains well beyond the CEB
• My values according to the CEB and ACI formula- limit of Es = 1.0 percent.
tions are in good agreement. The My value determined The greater ductility present in highly reinforced sec-
with the proposed analytical model is always lower be- tions is due to the adopted concrete law, which considers a
cause of the lower amount of steel provided to the section plastic (rather than a strain softening) behavior of concrete
if q =ro. in compression. This assumption may be considered ac-
• The yielding curvature is not much affected by the ceptable, since it influences only highly reinforced sec-
adoption of different analytical models or definitions of tions, which have a limited application in engineering
the reinforcement index. practice.
• Some major differences between the ultimate curva-
tures (Fig. 10) and ductility factors (Fig. 11), obtained by
adopting the ACI or CEB recommendations or the pro- Influence of the Concrete Compression Strength
posed analytical models, are noted. The influence of/; on theM- <1> law and the section duc-
The material models and limiting strain values (Figs. 10 tility <1>. /<I>Y has been studied on the I and rectangular sec-
and 11) of the ACI318-83 Codell result in ductility factors tions (A and F in Table 3). The results are plotted in Figs.
inferior to the theoretical <1>. /<I>Y values for all cases, except 12-14 and suggest the following remarks:
for RC !-sections and q < 0.22. In this case, the assumed • The effect of concrete grade on theM- <1> curves is rel-
infinite plastic behavior of the reinforcing steel results in atively minor, although higher grades tend to reduce the
considerable overestimates of <1>. /<I>Y (Fig. 11). ultimate concrete strains, particularly at low steel percent-
The CEB-MC 78 model 13 leads to gross underestimates ages (Figs. 12, 13a and 13b). The concrete grade does not
of ductility factors for rectangular and I-sections of both significantly affect the ductility ofl -sections, but has some
RC and PC sections with (practical) reinforcement indices impact on that of rectangular sections (Fig. 14). It is inter-
q ~ 0.20 (Fig. 11) because of the limiting concrete strain t:
esting to note that for increasing values, the ductility
Ecu = 0.35 percent and steel strain Est = 1.0 percent. Code factor <1>. /<I>Y decreases under a constant q value, but in-
recommended limiting strains are realistic for strength creases under a constant p value, as independent analytical

~r---~~------------------~~-----------r----~
range of feasible designs · · ·
."

."

0.20 0.25
I I I
0.00 0.04 0.08 0.12 0.16 0.20 0.24 (J) = p /y If/
I I I I
1.2
I
1.6
I
2.0
I I

Fig. 15. Combined effects of section shape and compression reinforcement on the ductility factor
[fort;= 40 MPa (5.8 ksi) and fy= 400 MPa (58 ksi)].

March-April1991 85
studies have shownY This apparently conflicting result is CONCLUSIONS
explained by observing that, for an RC section, q =
(fs 1 If:) p. Hence, to keep a constant q value for increasing A comprehensive analytical investigation of RC, PPC
J: values, it is necessary to increase the steel percentage p and PC structural concrete members3 has attempted to
eliminate some inconsistencies and limitations of code
of the section, with a resulting overall decrease of flexural
ductility. based behavior models by: (1) adopting realistic material
• The concrete strength has a negligible effect on My laws, (2) considering a continuous description of all be-
and M", while <I>Y and <l>u decrease for increasing values havior states, (3) using a unified formulation for RC, PPC
ofJ;. For a given q value, the reduction in yielding curva- and PC, and ( 4) considering the effects of all governing
tures is negligible, but the reduction in ultimate curvatures factors. Application of the MOCURO computer pro-
is relevant for sections failing by concrete crushing [for gram19 has enabled the automatic calculation of ductility
RC rectangular sections andJ; = 60 MPa (8.70 ksi)], <l>u factors and a parametric study on a large set of concrete
can be up to 30 percent smaller than forJ; = 30 MPa (4.35 sections.
ksi). Since the ultimate curvature decreases with increas- Analysis of related results presented in this study sup-
ing J:, the ductility factor is also proportionally reduced ports the following conclusions:
[up to 40 percent of the difference between <1>/<I>Y for = J: 1. Prestressing induces a relatively flat response in the
30 MPa (4.35 ksi) andj; =60 MPa (8.70 ksi)], as shown inelastic range and, therefore, plastic analysis methods are
in Fig. 14b. applicable to continuous PC structures.
• For the rectangular section, the ultimate concrete 2. Prestressing has a positive effect on the section duc-
strain Ecu is always larger than 0.45 percent and can be as tility: <l>u I<I>Y increases with the amount of prestressing (y).
high as 1 percent (q = 0.10 andJ; = 30 MPa, Fig. 13a). Hence, ductility factors of RC members may be used as
This is due to the existence of an extended strain softening conservative estimates for the corresponding PC and PPC
branch in the crc-Ec law. members.
3. Typical prestressed beams I, boxed, and flanged sec-
tions (with b/bw ~ 4) have a ductility factor of at least 14.
Ductility Factors and Moment Redistribution
4. Moment redistribution of at least 20 percent is en-
Fig. 15 summarizes the effects of q, compression flange sured for sections of any shape and material combinations,
to web width ratio blbw, and compression to tension steel provided the reinforcement percentage, p, for RC, or rein-
J:
ratio A; /As on the ductility factor <l>u I<I>Y for = 40 MPa forcement index, ro, for PC and PPC, do not exceed 1.6
(5.8 ksi) and/y = 400 MPa (58 ksi), and determines the percent and 0.16, respectively.
range of feasible design solutions from the viewpoint of 5. Sections 8.4 and 18.10.4 of ACI 318-83 Code 11
section ductility. Although the curves refer to R C sections, could be modified as follows:
results may be considered safe lower bounds for any struc- • 8.4.1 and 18.10.4.1 ... "negative moments" ... in-
tural concrete section. creased or decreased by not more than 20 percent.
Fig. 15 shows that for sections with a reinforcement • 8.4.2 and 18.10.4.2 unchanged.
index q ~ 0.20 (p ~ 1.6 percent) and A; /As~ 0.25 (which • 8.4.3 ... p or p- p' is not greater than 1.6 percent or blbw
are common practical solutions), the ductility factor <l>u I<I>Y is at least 2 for flanged sections.
is always larger than 5. Ductility factors of at least 14 are • 18.10.4.3 ... ro + roP- ro' is not greater than 0.16 or blbw
ensured for 0.05 ~ q ~ 0.20 (or 0.4 ~ p ~ 1.6 percent). For is at least 2 for flanged sections.
RC sections still higher values may be obtained removing 6. The above conclusions ensure the possibility of mo-
the limitation imposed on the maximum allowable steel ment redistribution, even for fully prestressed concrete
strain (Es 1=3.0percent). members (y= 1). However, design solutions based on re-
This is an important fmding with regard to the amounts distributed moments are only possible if cracking under
of moment redistribution permissible in design. As service loads is permissible,23 i.e., moment redistribution
known, Sections 8.4 and 18.10 of ACI-318 11 link allow- should be restricted to reinforced or partially prestressed
able moment redistribution top, p', pb for RC and ro, ro', concrete structures only.
rob for PC.
Theoretically, the maximum amount of permissible mo-
ment redistribution Ye may be determined from earlier in- ACKNOWLEDGMENTS
vestigations21 as:
Results in this paper are part of a thesis prepared by the
Ye = (<l>u /<j>Y - 1)/(15 + <l>u /<j>Y - 1) second author under the direction of the first author in par-
i.e., as a direct function of the ductility ratio. tial fulfillment of the requirements for a PhD degree in the
It can be seen that for values <l>u I<I>Y of at least 5 (i.e., q Department of Civil Engineering, University of Waterloo,
~ 0.20), Ye ~ 0.21 follows. 21 Accordingly, the ACI 20 per- Waterloo, Ontario, Canada. The financial support of the
cent maximum redistribution is always safe for any RC or National Sciences and Research Council (NSERC) of
PC section shape with q ~ 0.20 (p ~ 1.6 percent). Canada, under Grant l-4789, is gratefully acknowledged.
A similar conclusion is obtained from an independent The authors are indebted to the reviewers of this paper for
extensive computer investigation on the effective moment their penetrating comments and valuable, constructive
redistribution of structural concrete frameworks. 22 suggestions.

86 PCIJOURNAL
 REFERENCES
1. Cohn, M. Z., and Ghosh, S. K., "Ductility of Reinforced Con- 13. CEB-FIP, Model Code for Concrete Structures, Paris, France,
crete Sections," IABSE Publications, V. 32, No.2, 1973, pp. 1978.
51-81. 14. Cohn, M. Z., and Riva, P., "Constitutive Laws of Structural
2. Cohn, M. Z., and Bartlett, M., "Nonlinear Flexural Response of Concrete for Application to Nonlinear Analysis," Proceedings
Partially Prestressed Concrete Sections," ASCE Journal of the of JABSE Colloquium on Computational Mechanics of Con-
Structural Division, V. 108, No. ST 12, December 1982, pp. crete Structures, Delft, August 1987, pp. 87-98.
2747-2765. 15. Giuriani, E., "On the Effective Axial Stiffness of a Bar in
3. Cohn, M. Z., and Riva, P., "A Comprehensive Study of the Cracked Concrete," Bond in Concrete, Bartos, Applied Science
Flexural Behavior of Structural Concrete Elements," Studi e Publishers, London, 1982, pp. 107-126.
Ricerche, Corso di Perfezionamento per le Costruzioni in 16. Giuriani, E., "Theoretical Analysis of the Early Second Stage
Cemento Armato F.lli Pesenti, Politecnico di Milano, V. 9, in R.C. Beams," Bulletin d'Information CEB, No. 153, April
1987,pp.365-414. 1982.
4. Ghosh, S. K., and Cohn, M. Z., "Ductility of Reinforced Con- 17. Giuriani, E., and Riva, P., "Effects of Cracking on the Moment-
crete Sections in Combined Bending and Axial Load," Sympo- Curvature Relationship in Partially Prestressed Concrete
sium on Inelasticity and Nonlinearity in Structural Concrete, Beams" (in Italian), Studi e Ricerche, Corso di Per-
University of Waterloo, Waterloo, Ontario, Canada, 1972, SM fezionamento per le Costruzioni in Cemento Armato F.lli
Study 8, University of Waterloo Press, 1973, pp. 147-180. Pesenti, Politecnico di Milano, V. 7, 1985, pp. 189-220.
5. Naaman, A. E., Harajli, M. H., and Wight, J. K., "Analysis of 18. Giuriani, E., and Rosati, G., "Behavior of Concrete Elements
Ductility in Partially Prestressed Concrete Flexural Members," Under Tension after Cracking" (in Italian), Studi e Ricerche,
PCIJOURNAL, V. 31, No.3, May-June 1986, pp. 64-87. Corso di Perfezionamento per le Costruzioni in Cemento
6. Riva, P., "Engineering Approaches to Nonlinear Analysis of Armato F.lli Pesenti, Politecnico di Milano, V. 8, 1986, pp. 65-
Concrete Structures," PhD Thesis, Department of Civil Engi- 82.
neering, University of Waterloo, Waterloo, Ontario, Canada, 19. Cohn, M. Z., and Riva, P., MOCURO-User's Manual, Solid
1988. Mechanics Division, University of Waterloo, Waterloo, On-
7. Sargin, M., "Stress-Strain Relationships for Concrete and the tario, Canada, 1988.
Analysis of Structural Concrete Sections," SM Study No.4, 20. ASTM Specifications, A416, A615-76a, "Standard Specifica-
Solid Mechanics Division, University of Waterloo, Waterloo, tions for Reinforced and Plain Billet-Steel Bars for Concrete
Ontario, Canada, 1971. Reinforcement."
8. Skogman, B. C., Tadros, M. K., and Grasmick, R., "Ductility of 21. Cohn, M. Z., "Rotation Compatibility in the Limit Design of
Reinforced and Prestressed Concrete Flexural Members," PCI Reinforced Concrete Continuous Beams," Proceedings, Inter-
JOURNAL, V. 33, No.6, November-December 1988, pp. 94- national Symposium on Flexural Mechanics of Reinforced
107. Concrete, ACI Special Publication, SP 12, Miami, Florida, No-
9. Thompson, K. J., and Park, R., "Ductility of Prestressed and vember 1964, pp. 359-381.
Partially Prestressed Concrete Beam Sections," PCI JOUR- 22. Cohn, M. Z., and Gattesco, N., "Computer-Simulated Tests on
NAL, V. 25, No.2, March-Aprill980, pp. 46-69. Moment Redistribution, Part 1: ULS Considerations," Studi e
10. Campbell, T. 1., Moucessian, A., "Prediction of the Load Ca- Ricerche, Corsi di Perfezionamento perle Costruzioni in
pacity of Two-Span Continuous Prestressed Concrete Beams," Cemento Armato F. Hi Pesenti, Politecnico di Milano, V. 11,
PCIJOURNAL, V. 33, No.2, March-April1988, pp. 130-151. 1989,pp.269-299.
11. ACI Committee 318, "Building Code Requirements for Rein- 23. Cohn, M. Z., and Riva, P., "Equilibrium-Serviceability of
forced Concrete (ACI 318-83)," American Concrete Institute, Hyperstatic P.C. Beams," Proceedings of the Sessions Related
Detroit, Michigan, 1983. to Design, Analysis and Testing atASCE Structures Congress,
12. CSA, Design of Concrete Structures for Buildings, CAN3- San Francisco, California, May 1-5, 1989, pp. 201-212.
A23.3-M84, Rexdale, Ontario, Canada, 1984.

APPENDIX A- NOTATION
The following symbols are used in this paper in addition ~EP prestressing strain increment
to standard ACI 318-89 notation: ~Es = reinforcing steel increment
Epu = prestressing steel ultimate strain
= reinforcing steel stress at assumed limiting strain Est = 3 percent = assumed reinforcing steel limiting
= crack spacing strain at ultimate limit state
bending moment referred to centroid of gross Esu = reinforcing steel ultimate strain
cross section K degree of prestressing
= ultimate moment crc = concrete stress
= yielding moment crP = prestressing steel stress
. d = reinforcing steel stress
q = 1 Apfvu-
b d Aptfst
~-";
. .:
= rem~orcement m ex as
w Jc 'tp = prestressing steel bond stress
= prestressing steel slip 'ts = reinforcing steel bond stress
= reinforcing steel slip cl> = curvature
ci>Jcl>y= ductility factor
'Y = ApJ;,uApJ;,u
+ Asfst
= mixed reinforcement index

March-Apri11991 87