See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/262376367

Flexural Deformability of Reinforced Concrete Beams

Article  in  Journal of Structural Engineering · September 1998

DOI: 10.1061/(ASCE)0733-9445(1998)124:9(1041)

CITATIONS READS

73 399

4 authors, including:

Alessandro P. Fantilli Daniele Ferretti

Politecnico di Torino Università di Parma
122 PUBLICATIONS   1,360 CITATIONS    70 PUBLICATIONS   1,199 CITATIONS   

SEE PROFILE SEE PROFILE

Paolo Vallini
Politecnico di Torino
15 PUBLICATIONS   267 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Sustainable concrete masonry: AAC and LWAC View project

Numerical simulation of RC elements reinforced with FRCM View project

All content following this page was uploaded by Alessandro P. Fantilli on 20 May 2015.

The user has requested enhancement of the downloaded file.

 FLEXURAL DEFORMABILITY OF REINFORCED CONCRETE BEAMS

By Alessandro P. Fantilli,t Daniele Ferretti,z Ivo Iori/ and Paolo Vallini4

ABSTRACT: In reinforced concrete beams, the sectional definition of the moment-curvature relationship is not
straightforward owing to the presence of cracks. In general, this problem has .been solved by con~idering a
representative portion of the beam and by defining the curvature 1/r as the ratio between the rotatIOn of the
portion considered and its length. Large portions of the beam with many cracks lead to "average" moment-
curvature relationship, while small portions delimited by two consecutive flexural cracks lead to "local" mo-
ment-curvature relationship. The difference between these two definitions is studied in this paper introducing a
"general model." With this model the entire beam is modeled through a succession of blocks divided by flexural
cracks taking into account the bond between steel and concrete. Due to the practical impossibility of univocal
definition of the crack pattern evolution, it is proposed to use a "range model." This "range," delimited by the
curves of maximum and minimum deformability, includes the moment-curvature relationship (locals and average)
obtained from all possible crack patterns.

INTRODUCTION ter in portions of the beam characterized by steep gradients of

The problem of reinforced concrete (RC) beams deforma- the bending moment (Giuriani and Sforza 1981). Due to the
bility has been addressed by most authors on the basis of the characteristics of the testing equipment employed (displace-
moment-curvature relationship. As is well known, this ap- ment transducers and extensometers), in the past wide beam
proach, which is the favorite tool of the "engineering beam portions were generally considered, giving rise to average-type
theory," establishes a correlation within a generic cross sec- M-<I>IL diagrams. The data supplied by experimental investi-
tion of the member between a single mechanical action (Le., gations were used to calibrate most of the early sectional mo-
the bending moment M) and the corresponding deformation ment-curvature models described in the literature.
(i.e., the curvature llr). Although referring to a single section, these models are able
In RC beams, the sectional definition of this relationship is to simulate, with various degrees of approximation, the be-
not straightforward owing to the presence of cracks. How can havior of the entire beam portion considered (Branson 1966;
a single representative cross section be identified, if the open- Beeby 1968; Rao and Subrabmanyam 1973; Gilbert and War-
ing of cracks gives rise to a complex configuration (Fig. 1) ner 1978; Espion and Halleux 1988; "Building" 1989; Alwis
involving cracked sections (Sections ~), uncracked sections 1990; Prakhya and Morley 1990; CEB 1993). Among such
with no slip between steel and concrete (Sections (1), as well models, it is worth mentioning both the classical trilinear re-
as sections that are still uncracked but exhibit slip (Sections lationship [Fig. 3(a)] and some of the latest codes [Le., Comite
(1)-~)? Euro-International du Beton (CEB) (1993)]. This allows 1Ir to
In general, this problem has been solved by considering a be computed as a suitable linear combination, through a 'Y
representative portion of the beam and by defining the cur- coefficient, of curvatures 1Irl and 1Ir2' relating to Stage I and
vature Ilr as the ratio between the rotation <I> of the portion Stage II, respectively (Fig. 3(b)]
considered and its length L. 1 I 1
In early studies, the reference portions of the beam were -r = 'Y'-
rl
+ (1 - 'Y)'-
r2
(1)
wide and contained many cracks. Their effects on the moment-
curvature diagrams, M-<I>IL (of a type that might be defined Recently, various researchers moved their attention from
as "average"), were therefore markedly smeared and gave lit- wide portions of the beam to the block delimited by two con-
tle evidence for the formation of each crack (Fig. 2). secutive flexural cracks. By means of modern experimental
To account for these consequences, some authors [for in- techniques that observe the whole surface of the beam (i.e.,
stance, Gelfi and Giuriani (1982)] have introduced a "local" the geometric moire technique used by Giuriani and Ronca
definition of the curvature, that is, referring to a suitable por- 1979) researchers developed mechanical models of the entire
tion around each crack. The moment-curvature diagrams ob- block. These models (known as "block models") introduce a
tained from these two definitions may differ to a considerable suitable bond-slip relationship and adopt simplified strain pro-
extent, as clearly shown in Fig. 2. files in the cross sections. By assuming both the stress-strain
In general, it should be noted that, in beams characterized relationships of steel and concrete to be known, the models
by a virtually constant bending moment, average-type rela- impose the requirements of strain compatibility and equilib-
tionships may provide an accurate representation of physical rium of forces in all sections of the block obtaining a system
behavior. On the other hand, local-type relationships work bet- of nonlinear differential equations. Through suitable boundary
conditions, the solution of this system makes it possible to
'Doctoral Student, Dept. of Struct. Engrg., Politecnico di Torino, Corso

,*
Duca degli Abruzzi 24, 10129 Torino, Italy.
2Asst. Prof., Dept. of Civ. Engrg., Univ. of Parma, Viale delle Scienze,
40300 Parma, Italy.
'Prof., Dept. of Civ. Engrg., Univ. of Parma, Viale della Scienze, 40300
Parma, Italy.
'Prof., Dept. of Struct. Engrg., Politecnico di Torino, Corso Duca degli
Abruzzi 24, 10129 Torino, Italy.
Note. Associate Editor: John B. Mander. Discussion open until Feb-
ruary I, 1999. To extend the closing date one month, a written request
must be filed with the ASCE Manager of Journals. The manuscript for
M;:eMcr
(
-----------~------------

<D.@
\
I <D.@
I /
1

---------~----{-----
cb
I
M;:eMcr
)
;W :[
Ac
As
I
I

I
this paper was submitted for review and possible publication on Decem-
ber I, 1997. This paper is part of the Journal of Structul'lll Engineering,
Vol. 124, No.9, September, 1998. ©ASCE, ISSN 0733-9445/98/0009-
1041- 1049/$8.00 + $.50 per page. Paper No. 17075. Flexural Cracking in RC Beams

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/1041

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 ~ L ~ 'average' curvature plied (Maldague 1965; Priestley et al. 1971; Giuriani and Pliz-

M M(~&&i)M
zari 1989). The strain profile between the cracked section [Fig.
4(c)] and the uncracked one [Fig. 4(a)] is no longer plane but
! \. C%I follows the bilinear pattern depicted in Fig. 4(b).
\ ..,. Although aiming to a simplified representation of physical
behavior, the assumptions find confirmation both in laboratory
stag~ I
..... L tests (Giuriani and Ronca 1979; Jones et al. 1980) and in nu-
merical analyses (Ngo and Scordelis 1967; Plauk and Hees
.... I----of 1981).

stage 1.1 •• ••
M~M For a generic section of the block, together with the com-
patibility assumptions shown in Fig. 4, the equilibrium equa-
'it tions are considered, i.e.
'local' curvature

(2a)
1/r =C%l1L
FIG. 2. Definition of Local and Average Moment-Curvature Re-
lationship (Gelfi and Glurianl1982)

a) b)
L A<
<Te'Ye dA e + <T,·y,·A. + <T; 'Y; 'A; = M(z) (2b)

M>M cr M>M cr For a portion of the beam of infinitesimal length dz (Fig. 5)

M
CL -J) I

L
the equilibrium equation for the reinforcing bar alone can be
written as
I·
-d<T, =-.
p,
T[S(Z)] (3)
M>M er {J, M>M cr dz A.

crm:UJ L(l·'Y)
-01--+-
where A, and P. = area and perimeter, respectively, of the re-
inforcement in tension, while T[S(Z)] = bond stress.
Referring to the same portion of the beam, it is also possible
FIG. 3. Moment-Curvature Relationship: (a) Ti'iIInear; (b) Com- to define the slip s(z), corresponding to the difference in the
bination of Curvatures of Stages I and II displacements between two initially overlapping points be-
longing to steel and concrete
determine, for the acting moment M, the states of stress, strain, S(z) = s.(z) = S;,(z) (4)
slip, and curvature in each cross section (Westergaard 1933;
Maldague 1965; Priestley et al. 1971; Gelfi and Giuriani 1982; Hence, by derivation with respect to z
Cohn and Riva 1987; Cosenza et al. 1991; Eligehausen and
Li 1992; Sato et al. 1992; Creazza and Di Marco 1993; Vas- ds ,
iliev and Belov 1993).
dz = E.(Z) - Ee,(Z) (5)
In most cases, the block models assume a suitable length of
where E, = strain of the reinforcing bar; and £;, = tensile strain
the block rather than obtaining it as a solution of the problem.
in the surrounding concrete (Fig. 5).
Moreover, the steel-concrete bond is evaluated by neglecting
To solve the problem, the equilibrium and compatibility
a possible unloading of bond stresses. The unloading phenom-
equations must be associated with the constitutive laws of the
enon may occur, even with monotonically increasing moment,
materials and with the bond-slip relationship T-S.
when the boundary conditions change due to the formation of
In particular, since the study concerns the service load
a new crack (Ferretti et al. 1995). The aforementioned prob-
lems suggest the need for a block model that is able to take
into account these effects in a more generalized way. Stage I Stage II

DEFINITION OF "GENERAL MODEL"

The model proposed in this paper occupies an intermediate
dI lEe lb) Ee / )
Ee

position between two-dimensional finite-element models and Es Es Es

the cross-sectional ones. C
By means of appropriate boundary conditions, the entire Ect Ect
beam is modeled through a succession of blocks divided by
flexural cracks. In the cracked sections, the blocks interact via FIG. 4. Different Strain Profiles Assumed for Cross Sections:
the concrete and the reinforcement. Hence, the problem ap- (a) Stage I; (b) Generic Cross Sections; (c) Stage II
pears to be statically indeterminate, and the solution would
require a three-dimensional modeling of the entire beam (Vas-
iliev and Belov 1993). Nevertheless, the tests performed on
statically determinate elements showed that the formation of a
new crack does not modify significantly the stage of stress and
strain in the other preexisting blocks (Giuriani and Ronca
1979). Accordingly, in the proposed model, the blocks are as-
sumed to be mutually independent and delimited by cracks
that start forming when the cracking moment Mer is attained.
In the cracked sections, the crack tip is assumed to be located
on the neutral axis, and hence, Stage II assumptions are ap- FIG. 5. Free-Body Diagrams

1042/ JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 t [N/mm2 J

a)
I
b) 1',
I ,
/~Ec I ,

fc ----~- : \
I
/
"I'
I \
I'
I
I
't-----t I
I
/
"
:
I
).
I
I
I
I
,,~EC1 I'
I"
I
Ec Es s[mm)
" 1 :
Ec1 ECU ----,,, Sz200N/mm 3
,,
,,
~

FIG. 6" Constitutive Laws Adopted (CEB 1993): (a) Stress-Strain Relationship for Concrete In Compression; (b) Stress-Straln Rela-
tionship for Reinforcing Steel; (c) Bond-Slip Relationship with Unloading Branches

range, the ascending branch [Fig. 6(a)] of the relationship pro- of stress in concrete a c, act and steel as> the slip s, the bond
posed by CEB (1993) is considered as a reference for concrete stress T, and the curvature of each section JJ.. As a result the
in compression average curvature 1Ir, or the local curvature 1Ir/o<:, can be de-
termined as
ac =1Ec/Ect"EJEct - (£c/Ect)2
+ (EjEct - 2)' EJEct 'fc
(6)
E,-Ec
-1 or-=-'
1 1 1 JJ.dz=-·
11 ---dz (9)
r r/oc L L L L d
The concrete in tension is assumed to have a linear elastic
stress-strain relationship with modulus E c • where L = length of the considered reference portion (Fig. 2).
The stress-strain relationship for reinforcing steel in tension
or compression [Fig. 6(b)] is the following: BEAM SUBJECTED TO CONSTANT BENDING
(7a) MOMENT: COMPARISON WITH EXPERIMENTAL DATA

a,=asy + Ea"(E,-Esy) if E,>Esy (7b) The proposed model was applied to the analysis of a con-
stant moment beam (Fig. 7) tested by Giuriani and Sforza
For the T-S relationship, again in accordance with CEB (1981) by means of a geometric moire method. For this beam
(1993), the following law is assumed: both the average and local moment-curvature experimental di-
agrams are known. Starting from the incipient first cracking
T = T max " (~r if 0 s S < St (8a) (M = M cr = 1,350 N 'm), the numerical analysis examined the
beam portion comprised between experimental cracks @ and
T =T if SI S S < S2 (8b) ® [Fig. 7(a)] and offered the distribution of all the statical and
max
kinematical unknowns previously defined. In particular, the
(s - S2) l'f tensile stress a et in concrete at the lower edge of the beam
T = T max - (Tmax - Tf.) '( ) (8c) reaches the strength fet over an extensive portion in which a
S3 - S2
crack may theoretically originate. On the other hand, from the
(8d) tests it has been determined that, in the portion under study,
the cracks do not start simultaneously but, due to the inevi-
with the proposed CEB (1993) reduction of bond stress near table, albeit small, scatter in the concrete tensile strength fct
cracks. The unloading effects, which are normally produced they open in the chronological order shown in Fig. 7(a).
by the gradually increasing number of cracks, are taken into Moreover the cracks are not evenly spaced, and their dis-
account through the simplified cyclic relationship illustrated in
Fig.6(c).
589 mm 1000 mm 589 mm
For the solution of the problem, in addition to the foregoing I' I 'I' 'I
equations, suitable boundary conditions are also adopted, that O@@(D ~ @ @®®®@@ C1.~
is, Stage II conditions for the cracked sections and Stage I
conditions for the sections with perfect bond. From a mathe- a) r--f~lf+j ~ltxJ
matical viewpoint, (2)-(8) (which generally cannot be un-
coupled), together with the boundary conditions defined for Tl1Rn
M
(i irnl Iii
[-1- t -t--f--,
ir2l il3l i i)
M
TJIRI ~
the edges of the block, constitute a differential problem called
"two-point boundary value problem." b) I, portion examined I
In principle, this problem can be solved with the aid of
various numerical techniques, such as relaxation through finite
o
Experimental crack
differences, finite-element methods, and shooting techniques
tmax=2Vfc B=100mm o Numerical crack
tf =0.15 tmax H=150mm
(Press et al. 1992). For the numerical solution, the "multiple 81=~=0.6mm c=30mm
shooting method" was chosen by introducing, in the integra- 83=1.0 mm fc =40.42 Nlmm2
tion domain, two supporting sections where zero slip occurs. fet =3.2 Nlmm2
a =0.4 2
The position of such sections is shifted by trial and error until E c =28250 N1mm
S=2ooN/mm 3 2
all of the boundary conditions are satisfied in the integration Es =200000 N1mm
(Ferretti 1996). FIG" 7. Beam Tested by Glurlanl and Sforza (1981): (a) Experi-
The solution to the problem yields the distribution along the mental Crack Patternj (b) Analyzed Portion and Numerical
block of all of the unknowns statical and kinematical: the state Crack Pattern

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/1043

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 a) M=1350Nm

Mer===----------l) M
I, 892mm ,I

! O.l
0.0

~~ t _1-_-__A~__+_1- ! 0'1~
0.0

~~
_ , -_ _ A~-'I-- ! 0 ' 1 ~! 0 ' 1 ~
0.0
~.1
0.0
~.1

-
E -
gb -
..E
12
9
~
1/r
_. _ lIrloc
_11
E
..E
g~
12
9
E
..E
g~ .
12~
9
1/r
_. _ 1/r
__
e 129
.e
g~
-.p~~~Alr

~ 0 , I I ~ ~ 0 :::l. 0 ~ 0 -!-o-:-I~'_+:-4--+-'--~--
o 200 400 800 800 1000 0 200 400 800 800 1000 0 200 400 600 800 1000 0 200 400 600 800 1000
z [mm) z [mm) z [nm) z [nm)

FIG. 8. Numerical Outcomes: (a) Incipient Cracking; (b) Formation of First Crack; (c) Formation of Second Crack; (d) Final Crack Pat-
tern

tance varies randomly. For this reason the onset of the first the neighboring cracked blocks. It can be observed that, for
[Fig. 8(b)] and second crack [Fig. 8(c)] is triggered in keeping the same bending moment, local curvature values differ from
with the test results. The two cracks reduce a el and therefore one block to another, owing to the different distances between
the opening of a third crack can be attained only by increasing cracks. In Fig. 9, the values obtained for the average and local
the bending moment. In this way the tensile stress a et reaches curvatures from the numerical procedure and the laboratory
fer only in one cross section, so that the position of the third tests are compared. Their substantial agreement proves the va-
crack [Fig. 8(c)] is univocally determined and virtually coin- lidity of the proposed method. Furthermore, the small differ-
cident with the experimental one [Fig. 7(a)]. The bending mo- ence between local and average results confirms that, when
ment was increased until four cracks formed [Fig. 8(d)]. dealing with members subjected to a virtually constant mo-
For the portion examined it now becomes possible to de- ment, the use of an average instead of a local relationship does
termine, from (9), the local 1/rloe and the average 1/r curva- not yield significant differences.
tures by taking a suitable value of L (Fig. 9).
Local curvatures have been determined on the basis of the BEAM SUBJECTED TO VARIABLE BENDING
definition proposed by Giuriani and Sforza (1981), by consid- MOMENT: COMPARISONS WITH
ering L equal to the distance between the midway sections of EXPERIMENTAL DATA
The proposed model was also applied to the analysis of a
a) M[Nm] b) M [Nm] beam with variable bending moment tested by Gelfi and Giu-
3500 r-r-r-r-r....,.-..,...,...........,--.-....,-, 3500 r-r-r...............................-.-,.-,.-...-r"! riani (1982) by means of the geometric moire method. The
main geometric and mechanical properties of the beam ex-
3000 3000 amined by Gelfi and Giuriani are shown in Fig. 10.
.'
The numerical investigation, performed by increasing the
2500 2500
load Rio showed that when R 1 = 2,150 N the tensile strength
2000 2000 of the concrete is reached on the support section [Fig. II(a)],
where crack ill starts, in perfect agreement with the test results
1500 [Fig. II(b)].
The crack formation produces an instantaneous reduction of
concrete tensile stress aero as can be evinced from a compari-
•• ' - Experimental ." - Experimental son between Fig. II(a), relating to the incipient cracking con-
. ..' -"General" model : .... -"General" model
dition, and Fig. 11 (b) showing the first crack formation. The
oo~~':-'-~':-'-"""""'::-'--'.....-...,J
4 8 12 18 o0~"""""4~"""""'8""""''''''''1,.L:2~..J16 same figure shows that the new stress distribution does not
reach !ct, so that it becomes necessary to increase the load to
1/r [10-8/mm] 1/rloc [10-8/mm] reach incipient cracking conditions and the formation of crack
FIG. 9. Comparison between General Model and Experi- rID [Fig. II(c)]. By increasing the load gradually, the five cracks
mental Data (Glurlanl and Sforza 1981): (a) Average Moment- illustrated in Fig. lO(b) and in Fig. l1(d) are obtained.
Curvature Curves; (b) Local Moment-Curvature Curves Thus it is possible to compare the numerical crack pattern
1044/ JOURNAL OF STRUCTURAL ENGINEERING I SEPTEMBER 1998

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 [Fig. lO(b)] with the one detected in the laboratory tests [Fig. curvature for each individual cracked block. As a matter of
lO(a)]. The good agreement between the two results confinns fact, an average evaluation does not appear to be accurate. As
the capability of the model to simulate the complex physical for the constant moment beam, the local curvature is deter-
behavior also in terms of the crack history. mined on the basis of (9) assuming L equal to the distance
Compared with constant moment conditions the presence of between the midsections of two neighboring blocks, while M
shear produces a more regular crack pattern, with nearly even is evaluated in the cracked section. Thus, the definition is the
spacing of the cracks, in accordance with the test results (Gelfi same as the one used in the testing campaign (Gelfi and Giu-
and Giuriani 1982). The presence of shear therefore seems to riani 1982). Fig. 12 shows the local moment-curvature dia-
neutralize the uncertainty into the position of cracks, which grams, for the three cracked blocks, originated by the tests and
are now univocaUy positioned. by the numerical analysis. The three blocks display virtually
Referring to Fig. 11 it may be noted that the marked vari- identical deformation characteristics and good agreement with
ation in j.L points out the need for a local evaluation of the the experimental results.

a) @G)® ® ® c ~ ENVELOPE OF POTENTIAL SOLUTIONS: DEFINITION

f-----T:rT--r-T---------------!~ l~ ~ OF "RANGE MODEL"

1)1 A I . I . . . As for the crack configuration observed in the beam sub-

I . I I I jected to a constant moment, a deeper investigation is required,
b) I!II i 1lI1 I2Ij ClI' 151' because the cracks may occur in positions other than the se-
f-- --F-::~f--FF=---=---=--=---=---=-!~ lected ones.
This represents a fundamental aspect of the problem, that
1)lA, is, the practical impossibility of univocal definition of the
I, 614mm I 1612 mm
,I crack pattern evolution. In fact, as can be seen in Fig. 8(a),
when the initial cracking stage is attained (M = Mer), (Jet
o
Experimental crack reaches the ultimate value fet in a wide portion of the beam.
max =2 ...If;;
t 8=100 mm 0 Numerical crack This justifies the opening of the first crack in any section of
tf=O.15 tmax H=150mm this portion.
S1 =112=0.6 mm c=30mm To take into account this phenomenon, the numerical anal-
s3=1.0 mm
fc =36.79 Nlmm2 ysis was repeated by assuming three different possible devel-
fet =3.2 Nlmm2 opments of the crack pattern.
a =0.4 E c =32370 Nlmm
2
S=200Nlmm 3 2
The results are illustrated in Fig. 13(a) and clearly show
E 8 =200000 Nlmm that the moment-curvature relationships are affected by the
FIG. 10. 8eam Tested by Gelfl and Glurlanl (1982): (a) Experi- variations of the initial crack pattern.
mental Crack Pattern; (b) Analyzed Portion and Aelatlve Crack This type of behavior, as pointed out by Ferretti et al.
Pattern (1995), is not a prerogative of beams subjected to constant

a) t
4
1.
l =2150 N

614
"".1. 1812mm
t
b)

4.
I
Al =2150 N

/.614
lD

I 1812mm
t
c)
I4
AI =2300N
[l1l21
J2
--I-W87mm
•
d) AI =2150 N

t ~1~'~
1;40~ll~l. 1051
•
.1

~~ ! ~~
N

t ~~ !
N N

~ ~ ~

4001 - 400j
! ~o~ -1 = liEit::1
N

I;) _
_......-...-.. , . A _ ......__
~ log - . l-~j.l_i -+_
1 ~ 0 1

1- 0.0o.ll
on -0.1
-------+-,-1- 0'1j-4
0.0
'" -0.1
....
, -----....- f-on 0.1f-#
0.0
-0.1
-1----.'. -1 0'1~ -
OJ
0.0
-0.1

i
N'I
E

.. -6
3
_~ ,I iMrl- I
------

HI o
i : I
0.4 0.8 1.2 1.8 2.0 2.4
I ,
{ ~~I ~
::L 1 90 0.4 0.8
,
1.2
I ,
1.8 2.0 2.4
I
1~ EIg b,
0
I "
0.4 0.8 1.2 1.8 2.0 2.4
~: L: fo~l:fv\A
0
~, . ,
-L...::..l..
0.4 0.8 1.2 1.8 2.0 2.4
z[m) z[m) z[m) z[m)

FIG. 11. Numerical Outcomes: (a) Incipient Cracking; (b) Formation of First Crack; (c) Formation of Second Crack; (d) Final Crack
Pattern

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/1045

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 a) b) c)
3000 """""TT"T'"'I""T""T""'l'""1"'T"T""T'"T"T"'I""T"'T"T"T'.,..,....r-n 3000 """"""""''''''''''''''''''''''"T"T"T""T'"'''''''''''''''''''''''''''''''''''' 3000 rr"'"T"T',....,r-r-r"""'T"T""I"""'''''''''''''''''''"T''T'T"1
I!I ID 121 131 1m I!IID 121 131 1m I!IID 121 131 1m
2500
...:aL-:-1....:-1_ _....t
t,--...:.I-:OQiC"ia
2500
r 'liM' • 2500
•
RI \
'local' block stage II ~gell 8~1I
: stage I : stage I
2000 2000
. 2000

E E E
~ 1500 ~ 1500 ~ 1500
::::E ::::E ::::E

1000 1000 1000

...
.' - - 'General moder .. - - 'General model'
" 'General model"
500 ___ Experimental 500 500
___ experimental ___ Experlmenllli

o. 0 . 0 L.'L...L-J'-'-l........................l...l,...........L-l.....................J...J...'-'-L..J
0.0 5.0 10 15 20 25 0.0 5.0 10 15 20 25 0.0 5.0 10 15 20 25
1/rloe [10 ~ Imm] 1/rloc [10 ~ Imm] 1/r [10~ Imm]
1oo
FIG. 12. Comparison between Experimental (Gelfl and Glurlanl1982) and Numerical M-1/rloc Diagrams: (a) Local Block 2 (b) Local
Block 3; (c) Local Block 4

a) b) c)
3500 3500

3000 3000

2500 2500
M
2000 2000
E E
~ ~
::::E 1500 ::::E 1500
Mer

1000 1000
••' _ Experimental •• ' _ Experimental
500 .•' - - Crack pattem n.1 500 .•' - Crack pattem n.1
.,' - - Crack pettem n.2 .. ' - - Crack pattern n.2
.. - - - - Crack pattem n.3 .. - - - - Crack pattem n.3
o l.:..,'....L..-.................L..................o.......L-I..-'-.!-..IL....L-I.....Jo-...J
4.0 8.0 12 18 0 4.0 8.0 12 18
1/r [10~ Imm] 1/r 1/r [10~ Imm]
FIG. 13. Uncertainty of Crack Pattern: (a) Three Possible Outcomes; (b) Range Model; (c) Capability of Range Model to Include Nu-
merical Curves and Experimental Data

bending moment but is also observed in RC tensile members conditions and that of the block produced by its cracking.
(Avalle et aJ. 1994) and, in general, in all the situations where Therefore, the deformability of the member is within these two
the initial crack pattern is uncertain. Such situations surely limits. For these reasons, the identification of the so-called
include also members subjected to variable bending moment. range can be achieved through the study of deformability of
In fact, in such members the crack pattern is no longer affected an individual block in Condition A (incipient cracking) and in
by the same uncertainty associated with a constant moment Condition B [after the opening of the crack; Fig. l3(b)]. With
beam. However, the presence of stirrups or the load history reference to Fig. 14(a), Condition A is characterized by a block
may alter the theoretical regular evolution of the cracks (Riz- length L = 21" (1" being the so-called transfer length) so that
kalla and Hwang 1984). To deal with this random nature of CJ'e, = fc, in the midspan of the block. Conversely, in Condition
the phenomenon, instead of pursuing a probabilistic approach, B, the previous block is assumed to crack just in the midspan,
all of the solutions due to different possible crack evolution and the deformability is computed for a half-block length L =
hypotheses could be included in a well-defined field or l,r subjected to the same M. Therefore the identification of the
"range." This range, which will be defined in terms of a mo- range assumes the the block length L = 21" varies continuously
ment-curvature relationship, may be delimited by the curves as a function of the bending moment M.
of maximum and minimum deformability. The M-lIr relation- With reference to the generic point A, it is possible to con-
ships obtained from all of the possible crack patterns must sider a block with length 21,r [Fig. 14(a)] comprised between
therefore fall within this range. The definition of the range, on two adjacent cracks and subjected to M 2: Mer. The behavior
the basis of all the infinite possible solutions, cannot be pro- of such a block can be described through the same equilibrium
posed. However the observation of the beams suggests that, and compatibility conditions used for the "general model"
for a given moment M, the deformability of the generic block (2)-(8), but with different boundary conditions. In particular,
falls within the deformability of the block in incipient cracking for the cracked sections [denoted with ~ in Fig. 14(a)] Stage
1046 / JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
II conditions are assumed to apply. For the midspan section ing with the integration of the differential equations from the
the incipient cracking conditions requires (T e' = !C" while s = 0 midsection and shifting the cracked section until all the bound-
due to symmetry. From a mathematical viewpoint, this prob- ary conditions are satisfied.
lem represents a particular boundary value problem, whose On the other hand, point B [Fig. 13(b)] is determined as-
solution requires the determination of the unknown length lIT' suming that, when the previous incipient crack occurs, two
This length is computed with a trial-and-error algorithm start- separate blocks of known length lIT are formed. Under these
circumstances, the boundary conditions in the cracked sections
a) CONDITION A (denoted with ®) are still those of Stage II. Instead, in the
midspan section of the new block, S = 0 and the stress (Tet is
® still unknown. The boundary conditions, relating to half the
Mz:Mcr block, define a two-point boundary value problem, which can
( be solved through the same shooting method described for the
general model. The range identified in this way is sufficient to
encompass the numerous possible solutions [Fig. 13(c)]. For
the previous constant moment example, the M-l/r diagrams
\. produced by the test and the three general analyses [Fig. 13(a)]
are depicted in Fig. 13(c).
The range, obtained through the local analysis of the same
beam, is able to contain the M-l/rlo e diagrams, both in the
constant moment case illustrated previously and in the case of
variable moment. Fig. 15(a), for instance, shows that the local
diagrams of the variable moment beam (previously analyzed
by the general model) fall within the range. Fig. 15(a) also
shows the M-lIr curve according to the CEB (1993) model.
Except for the initial portion, this curve, obtained by means
of average observations, falls within the range, although the
latter has been obtained according to a local approach. Thus,
FIG. 14. Block Considered to Define Range: it can be observed that, in this particular case and for values
Cracking; (b) Formation of Crack of M slightly higher than Mat the indications given in CEB
(1993) are not reliable when dealing with all those structural
a) conditions that require a local evaluation of the curvature. As
3OO0,.,..,.........,........,,...........................,.,"T"'"...,.,..,
a final consideration, it should be pointed out that, in the com-
aHD 12I1311SJ mon block model approaches, the crack pattern evolves step-
2500 rRI I' I I I ~
,....:-
~
'J wise with increasing applied moment, that is, by halving grad-
:stege I 1"..0::-:;;)<' ually the length of each cracked portion [e.g., Creazza and Di
2000 : ~.:''' .. ' / .'slage II Marco (1993)]. Conversely, by means of the range model, a
/.. . '"
. continuous variation of the block length L as a function of the
.: l~·."
: ....:.:... /
, .
.~
... .. curvature is obtained [Fig. 15(b)]. However the stepwise ev-
olution, which reflects the physical behavior, is still contained
in the range. Fig. 15(b) illustrates the stepwise development
of the transfer length llr and the corresponding shift of the
500 bending moment from M I _2 to M 3 • Diagrams substantially sim-
ilar to the one shown in Fig. 15(b) have also been proposed
for tensile members by BaZant and Oh (1982), Floegl and
0.0 5.0 10 15 20 25 lIrtoc Mang (1982), Tanabe and Yoshikawa (1987), and Wu et al.
1/rloc 110-6lmml (1991).
In the determination of both the deformability range and the
FIG. 15. M·1/r Relationships Relating to Beam Examined by
Gelfl and Glurlani (1982): (a) Comparison between Range general model, the solution of the differential problem gives
Model, General Model, and CES (1993); (b, c) Continuous and the distribution of the slip s, whose values identify the width
Discrete Development of Thansfer Length I" w of the cracks. In this way, it was possible to plot the w-M

a) b) c) d)
0.5 0.5 0.5 0.5 r--"'-----'--r--r-,..........,
f Ii, , I tRI 1£'1 I • 13l
0.4 RI 0.4 0.4 ~RI
- 'General' model - "General' model
E E ... Experimental
E *
Experimentel
§. 0.3 §. 0.3 §. 0.3
CRACKf2)

:7
~ ~ ~
0.2 0.2 0.2 0.2
~ .
..:.; Models:
0.1 0.1 0.1 0.1
..:.' : - - 'Range'
~. ;-----'Gener~'
0.0 L -..........._-'--..L. ....I
0.0 0.0 0.0 '---,-._-,-...J.._L.-....J
o 1000 2000 3000 0 1000 2000 3000 0 1000 2000 3000 0 1000 2000 3000
M [Nm] M [Nm] M [Nm] M [Nm]
FIG. 16. Evaluation of Crack Width was Function of Moment M: (a-c) Comparison between General Model Curves and Test Data
(Gelfl and GlurlanI1982); (d) Comparison between General Model Curves, Range Model Curves and Test Data (Gelfl and Glurlanl1982)

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/1047

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
curves, with reference to the structural member shown in Fig. cei, Accademia Nazionale dei Lincei, Rome, Italy, 9(2), 249-258 (in
10. In particular, in Figs. 16(a-c) the experimental data relat- Italian).
ing to three different cracks and the numerical results obtained Creazza, G., and Di Marco, R. (1993). "Bending moment-mean curvature
relationship with constant axial load in the presence of tension stiff-
with the general model are compared. ening." Mat. and Struct., Paris. France, 26, 196-206.
Together with the "defonnability range," it is also possible Eligehausen, R., and Li, L. (1992). "Rotation capacity of prestressed
to obtain the "crack width range" shown in Fig. 16(d), which concrete members." Bond in Concrete, Proc., Int. Conf, Comite Euro-
nearly encompasses the test data and the numerical results of International du Beton (CEB), ed., Riga, Latvia, 2.58-2.67.
the general model. A similar range for the crack widths was Espion, B., and Halleux, P. (1988). "Moment curvature relationship of
proposed, in the case of RC tensile members, by Somayaji and reinforced concrete sections under combined bending and normal
force." Mat. and Struct., Paris, France, 21, 341-351.
Shah (1981). The "range approach" seems to be able to quan- Ferretti, D. (1996). "On short term behaviour of RC beams in bending,"
tify the "model uncertainties" arising from the randomness of PhD thesis, 11, Politecnico di Torino, Turin, Italy (in Italian).
equipotential crack patterns, characterized by the same con- Ferretti, D., lori, I., and Vallini, P. (1995). "On the serviceability behavior
crete tensile strength instead of its statistical variation along of reinforced concrete elements in bending." Studie e Ricerche, Scuola
the member. di Specializzazione per Ie Costruzioni in C. A., Elli Pesenti, Politecnico
di Milano, Milan, Italy 16, 209-229 (in Italian).
Floegl, H., and Mang, M. (1982). "Tension stiffening concept based on
CONCLUSIONS bond slip." J. Struct. Div., ASCE, 108(12),2681-2701.
Gelfi, P., and Giuriani, E. (1982). "Local moment-curvature relationship
Some important features of the defonnability of RC beams
for rc beams under bending and shear. Experimental analysis by means
subjected to constant or variable bending moment can be ex- of moire method." Afti del X Convegno Nazionale A.I.A.S., Universita
plained by the proposed general model. A few considerations della Calabria, Cosenza, Italy, 625 -640 (in Italian).
can be finally remarked. Gilbert, R. I., and Warner, R. E (1978). "Tension stiffening in reinforced
Fig. 15(a) confinns that the moment-curvature models pro- concrete slabs." J. Struct. Div., ASCE, 104(12), 1885-1900.
posed in CEB (1993) are average-type models. The reason for Giuriani, E., and Plizzari, G. (1989). "Propagation and distance of cracks
in R.C. beams with a bending moment gradient." Studi e Ricerche,
this should probably be sought in the direct dependence of the Corso di Perfezionamento per Ie Costruzioni in C. A., Elli Pesenti,
CEB guidelines on the early experimental observations con- Politecnico di Milano, Milan, Italy, 11, 61- 106.
ducted by testing methods focusing on average characteristics. Giuriani, E., and Ronca, P. (1979). "The geometric moire method tech-
In some situations, the average approach is clear-cut, and, as nique in the study of RC beams in bending." VII Convegno Nazionale
already pointed out by Giuriani and Ronca (1979), it may lead A.IAS., Cagliari, Italy, 6.55-6.68 (in Italian).
to rough structural analyses. Giuriani, E., and Sforza, C. (1981). "Relationships between local and
average moment-curvature for a RC beam under increasing and re-
peated imposed curvatures-Experimental research by means of the
• In these situations the local approach seems to work better moire method with superimposed grids." Studi e Ricerche, Corso di
(Fig. 12). However both approaches are affected by the Perfezionamento per Ie Costruzioni in C. A., Elli Pesenti, Politecnico
indetennination of the initial crack pattern. di Milano, Milan, Italy, 3, 177 - 204 (in Italian).
• For this reason a range approach, able to encompass all Jones, R., Swamy, R. N., Bloxham, J., and Bouderbalah, A. (1980).
the possible defonnabilities has been proposed [Fig. "Composite behavior of concrete beams with epoxy bonded external
reinforcement." Int. J. Cement Composites, Lancaster, England, 2(2),
B(c)]. 91-107.
Maldague, J. C. (1965). "Etablissement des lois moment-courbures." An-
It should be observed that the introduction of the range nales de l'ITBTP, Paris, France, 213, 1170-1218 (in French).
could not be considered as a direct tool for design practice, Monnier, T. (1970). "The moment-curvature relation of reinforced con-
which is usually carried out by means of simplified models. crete." Heron, 17(2), 1-101.
Ngo, D., and Scordelis, A. C. (1967). "Finite element analysis of rein-
The identification of a range of possible real situations, how-
forced concrete beams." ACI J., 64(3), 152-163.
ever, enables the designer to check the validity of the adopted Plauk, G., and Hees, G. (1981). "Finite element analysis of reinforced
simplified M-lIr relationship and to quantify the uncertainty concrete beams with special regard to bond behaviour." Proc., Ad-
of the parameters [e.g., coefficient "y in (1)]. vanced Mech. of Reinforced Concrete, IABSE Colloquium, 655-670.
Prakhya, G. K. v., and Morley, C. T. (1990). "Tension-stiffening and
APPENDIX I. REFERENCES moment-curvature relations of reinforced concrete elements." ACI
Struct. J., 87(5), 597 -605.
Alwis, W. A. M. (1990). "Trilinear moment-curvature relationship for Press, W. H., Teukolsky, S. A., Vetterling, W. T., and Flannery, B. P.
reinforced concrete beams." ACI Struct. J., 87(3), 276-283. (1992). Numerical recipes in C: The art of scientific computing. Cam-
Avalle, M., Ferretti, D., lori, I., and Vallini, P. (1994). "On the deform- bridge University Press, Cambridge, U.K.
ability of reinforced concrete members in tension and bending." Com- Priestley, M. J. N., Park, R., and Lu, E P. S. (1971). "Moment-curvatures
putational Modeling of Concrete Struct., Proc., Int. Conf EURO-C relationships for prestressed concrete in constant-moment zones."
1994. H. Mang, N. Bicanic, and R. de Borst, eds., Pineridge Press, Mag. of Concrete Res., 23(75), 69-78.
Innsbruck, Austria, 723 - 734. Rao, P. S., and Subrahmanyam, B. V. (1973). "Trisegmental moment-
Bahnt, Z. P., and Oh, B. H. (1982). "Spacing of cracks in reinforced curvature relations for reinforced concrete members." J. Am. Concrete
concrete." J. Struct. Engrg. Div., ASCE, 109(9), 2066-2085. Inst., 70(5), 346-351.
Beeby, A. W. (1968). "Short-term deformations of reinforced concrete Rizkalla, S. H., and Hwang, L. S. (1984). "Crack prediction for members
members." Tech. Rep. TRA 408, Cement and Concrete Assn., London, in uniaxial tension." ACI J., 81(6), 572-579.
England. Sato, R., Ujike, I., Minato, H., and Dilger, W. (1992). "Basic bond equa-
Branson, D. E. (1966). "Deflection of reinforced concrete flexural mem- tions in a reinforced concrete flexural element." Bond in Concrete,
bers." J. Am. Concrete Inst., ACI, 63(6), 637-674. Proc., Int. Conf, Comite Euro-International du Beton (CEB), ed., Riga,
"Building code requirements for reinforced concrete." (1989). ACI 318- Latvia, 2.89-2.98.
89, Am. Concrete Inst. (ACI), Detroit, Mich. Somayaji, S., and Shah, S. P. (1981). "Bond stress versus slip relationship
Cohn, M. Z., and Riva, P. (1987). "A comprehensive study of the flexural and cracking response of tension members." ACI J., 78(3), 217 -225.
behaviour of structural concrete elements." Studi e Ricerche, Corso di Tanabe, T., and Yoshikawa, H. (1987). "Constitutive equations of a
Perfezionamento per Ie Costruzioni in C. A., Elli Pesenti, Politecnico cracked reinforced concrete panel." Proc., Computational Mech. of
di Milano, Milan, Italy, 9, 365-413. Concrete Struct. -Advances and Applications, IABSE Colloquium,
Comite Euro-International du Beton (CEB). (1993). "CEB-FIP model 17-34.
code 1990." CEB Bulletin d'Information 213-214, Thomas Telford Vasiliev, P. I., and Belov, V. V. (1993). "A study of the flexural behavior
Service Ltd., London, England. of RIC cracked beams, by means of a 3D 'Block Model.' " Studi e
Consenza, E., Greco, C., and Manfredi, G. (1991). "Theoretical evalua- Ricerche, Scuola di Specializzazione in Costruzioni in C.A., Elli Pe-
tion of inelastic rotations and displacements in the reinforced concrete senti, Politecnico di Milano, Milan, Italy, 14,435-452.
monodimensional elements." Atti della Accademia Nazionale dei Lin- Westergaard, H. M. (1933). "Stresses at a crack, size of the crack, and

10481 JOURNAL OF STRUCTURAL ENGINEERING 1 SEPTEMBER 1998

Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org
 the bending of reinforcement concrete." J. Am. Concrete [nst., 5(6), w = crack width;
93-102. y = vertical coordinate;
Wu, Z., Yoshikawa, H., and Tanabe, T. (1991). "Tension stiffness model z = horizontal coordinate;
for cracked reinforced concrete." J. Struct. Engrg., ASCE, 117(3),
715-732. 1Ir = curvature;
o = diameter of reinforcing bar;
£ = strain;
APPENDIX II. NOTATION 1.1. = cross-sectional curvature;
The following symbols are used in this paper: (J" = axial stress;
T = bond stress; and
A = area; <I> = rotation.
c cover;
SUbscripts
d effective height of beam;
E = modulus of elasticity; c = concrete in compression;
f = strength of material; ct = concrete in tension;
L = length of element or block; 'et = concrete in tension surrounding reinforcing bar;
l" = transfer length; I = stage I;
M = bending moment; II = stage II;
Mer cracking moment; loc = local observation;
n number of integration points; s = steel in tension;
p perimeter of reinforcing bar; ,s steel in compression; and
s slip or displacement; y = yielding.

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 1998/1049

View publication stats Downloaded 18 Nov 2010 to 130.34.85.94. Redistribution subject to ASCE license or copyright. Visit http://www.ascelibrary.org