ACI 224.

2R-92
(Reapproved 1997)

Cracking of Concrete Members in Direct Tension

Reported by ACI Committee 224

David Darwin* Andrew Scanlon* Grant T. Halvorsen

Chairman Peter Gergely* Secretary
Subcommittee
Co-Chairmen

Alfred G. Bishara Fouad H. Fouad Milos Polvika

Howard L. Boggs Tony C. Liu Lewis H. Tuthill*
Merle E. Brander LeRoy Lutz* Orville R. Werner
Roy W. Carlson Edward G. Nawy Zenon A. Zielinski
William L. Clark, Jr.*
--``,`,-`-`,,`,,`,`,,`---

* Members of the subcommittee who prepared this report.

Committee members voting on this minor revision:

Grant T. Halvorsen Randall W. Poston

Chairman Secretary

Florian Barth Peter Gergely Keith A. Pashina

Alfred G. Bishara Will Hansen Andrew Scanlon
Howard L. Boggs M. Nadim Hassoun Ernest K. Schrader
Merle E. Brander William Lee Wimal Suaris
David Darwin Tony C. Liu Lewis H. Tuthill
Fouad H. Fouad Edward G. Nawy Zenon A. Zielinski
David W. Fowler Harry M. Palmbaum

This report is concerned with cracking in reinforced concrete caused

CONTENTS
primarily by direct tension rather than bending. Causes of direct tension
cracking are reviewed, and equations for predicting crack spacing and
Chapter 1-Introduction, pg. 224.2-2
crack width are presented. As cracking progresses with increasing load,
axial stiffness decreases. Methods for estimating post-cracking axial stiffness
are discussed. The report concludes with a review of methods for
Chapter 2-Causes of cracking, pg. 224.2-2
controlling cracking caused by direct tension. 2.1-Introduction
2.2-Applied loads
2.3-Restraint
Keywords: cracking (fracturing); crack width and spacing; loads
(forces); reinforced concrete; restraints; stiffness; strains; stresse . Chapter 3-Crack behavior and prediction equations, pg.
tensile stress; tension; volume change. 224.2-3
3.1-Introduction
3.2-Tensile strength
ACI Committee Reports, Guides, Standard Practices, and
Commentaries are intended for guidance in designing, plan- The 1992 revisions became effective Mar. 1, 1992. The revisions consisted of
removing year designations of the recommended references of standards-pro-
ning, executing, or inspecting construction and in preparing
ducing organizations so that they refer to current editions.
specifications. Reference to these documents shall not be Copyright 0 1986, American Concrete Institute.
made in the Project Documents. If items found in these All rights reserved including rights of reproduction and use in any form or by
documents are desired to be part of the Project Documents any means, including the making of copies by any photo process, or by any elec-
they should be phrased in mandatory language and in- tronic or mechanical device, printed, written, or oral, or recording for sound or
corporated into the Project Documents. visual reproduction or for use in any knowledge or retrieval system or device,
unless permission in writing is obtained from the copyright proprietors.
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 224.2R-2 ACI COMMITTEE REPORT

3.3-Development of cracks Concrete members and structures that transmit loads

3.4-Crack spacing primarily by direct tension rather than bending include
3.5-Crack width bins and silos, tanks, shells, ties of arches, roof and
bridge trusses, and braced frames and towers. Members
Chapter 4-Effect of cracking on axial stiffness, pg. such as floor and roof slabs, walls, and tunnel linings may
224.2R-6 also be subjected to direct tension as a result of the
4.1-Axial stiffness of one-dimensional members restraint of volume change. In many instances, cracking
4.2-Finite element applications may be attributed to a combination of stresses due to
4.3-Summary applied load and restraint of volume change. In the fol-
lowing sections, the effects of applied loads and restraint
Chapter 5-Control of cracking caused by direct tension, of volume change are discussed in relation to the for-
pg. 224.2R-9 mation of direct tension cracks.
5.1-Introduction
5.2-Control of cracking caused by applied loads 2.2-Applied loads
5.3-Control of cracking caused by restraint of volume Axial forces caused by applied loads can usually be
change obtained by standard analysis procedures, particularly if
the structure is statically determinate. If the structure is
Notation, pg. 224.23-10 statically indeterminate, the member forces are affected
by changes in stiffness due to cracking. Methods for est-
Conversion factors-S1 equivalents, pg. 224.2R-11 imating the effect of cracking on axial stiffness are
presented in Chapter 4.
Chapter 6-References, pg. 224.2R-11 Cracking occurs when the concrete tensile stress in a
6.1-Recommended references member reaches the tensile strength. The load carried by
6.2-Cited references the concrete before cracking is transferred to the rein-
forcement crossing the crack. For a symmetrical member,
the force in the member at cracking is
CHAPTER l-INTRODUCTION

Because concrete is relatively weak and brittle in

--``,`,-`-`,,`,,`,`,,`---
tension, cracking is expected when significant tensile in which
stress is induced in a member. Mild reinforcement and/or
prestressing steel can be used to provide the necessary A, = gross area
tensile strength of a tension member. However, a number = steel area
of factors must be considered in both design and con- f t' = tensile strength of concrete
struction to insure proper control of cracking that may n = the ratio of modulus of elasticity of the steel
occur. to that of concrete
A separate report by ACI Committee 224 (ACI 224R) p = reinforcing ratio = ASIA,
covers control of cracking in concrete members in gen-
eral, but contains only a brief reference to tension After cracking, if the applied force remains un-
cracking. This report deals specifically with cracking in changed, the steel stress at a crack is
members subjected to direct tension.
Chapter 2 reviews the primary causes of direct tension
cracking, applied loads, and restraint of volume change. fs= f =($ - 1 +rJ)fi (2.2)
Chapter 3 discusses crack mechanisms in tension mem-
bers and presents methods for predicting crack spacing For n = 10, fi’ = 500 psi (3.45 MPa). Table 2.1 gives
and width. The effect of cracking on axial stiffness is the steel stress after cracking for a range of steel ratios
discussed in Chapter 4. As cracks develop, a progressive p, assuming that the yield strength of the steel& has not
reduction in axial stiffness takes place. Methods for been exceeded.
estimating the reduced stiffness in the post-cracking
range are presented for both one-dimensional members Table 2.1-Steel stress after cracking for various steel
and more complex systems. Chapter 5 reviews measures ratios
that should be taken in both design and construction to
control cracking in direct tension members. 1
- - l + n
LT*
Ll D ksi (MPa)

CHAPTER 2-CAUSES OF CRACKING

2.1-Introduction *Assumes f, < f,.

Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 CRACKING OF CONCRETE MEMBERS IN DIRECT TENSION 224.2R-3

Table 3.1-Variability of concrete tensile strength: Typical results5

Mean 1 Standard deviation 1 Coefficient

Type of test
I strength,
psi
(MPa)
I
within batches,
psi
(MPa)
I
of
variation,
percent
Splitting test 405 (2.8) 20 (0.14) 5
Direct tensile test 275 (1.9) 19 (0.13)
Modulus of rupture 605 (4.2) 36 (0.25) :
Compression cube test 5980 (42) 207 (1.45) 3 1/2
--``,`,-`-`,,`,,`,`,,`---

Table 3.2-Relation between compressive strength and tensile strengths of concrete6

Strength ratio
Compressive Direct tensile Direct tensile
strength Modulus of rupture* strength to strength to
of cylinders, to compressive compressive modulus of
psi (MPa) strength strength rupture*
1000 (6.9) 0.23 0.11 0.48
2000 (13.8) 0.19 0.10 0.53
3000 (20.7) 0.16 0.09 0.57
4000 (27.6) 0.15 0.09 0.59
5000 (34.5) 0.14 0.08 0.59
6000 (41.4) 0.13 0.08 0.60
7000 (48.2) 0.12 0.07 0.61
8000 (55.1) 0.12 0.07 0.62
9000 (62.0) 0.11 0.07 0.63
*Determined under third-point loading.

For low steel ratios, depending on the grade of steel, restraint. This point is demonstrated in Tam and Scan-
yielding occurs immediately after cracking if the force in lon’s numerical analysis of time-dependent restraint force
the member remains the same. The force in the cracked due to drying shrinkage.3
member at steel yield is A&

2.3-Restraint CHAPTER 3-CRACK BEHAVIOR AND

When volume change due to drying shrinkage, thermal PREDICTION EQUATIONS
contraction, or another cause is restrained, tensile
stresses develop and often lead to cracking. The restraint 3.1-Introduction
may be provided by stiff supports or reinforcing bars. This chapter reviews the basic behavior of reinforced
Restraint may also be provided by other parts of the concrete elements subjected to direct tension. Methods
member when volume change takes place at different for determining tensile strength of plain concrete are
rates within a member. For example, tensile stresses discussed and the effect of reinforcement on devel-
occur when drying takes place more rapidly at the ex- opment of cracks and crack geometry is examined.
terior than in the interior of a member. A detailed
discussion of cracking related to drying shrinkage and 3.2-Tensile strength
temperature effects is given in ACI 224R for concrete Methods to determine tensile strength of plain con-
structures in general. crete can be classified into one of the following cate-
Axial forces due to restraint may occur not only in gories: 1) direct tension, 2) flexural tension, and 3) in-
tension members but also in flexural members such as direct tension. 4.4. Because of difficulties associated with
floor and roof slabs. Unanticipated cracking due to axial applying a pure tensile force to a plain concrete spec-
restraint may lead to undesirable structural behavior such imen, there are no standard tests for direct tension.
as excessive deflection of floor slabs1 and reduction in Following ASTM C 292 and C 78 the modulus of rup-
buckling capacity of shell structures.2 Both are direct ture, a measure of tensile strength, can be obtained by
results of the reduced flexural stiffness caused by testing a plain concrete beam in flexure. An indirect
restraint cracking. In addition, the formation of cracks measure of direct tensile strength is obtained from the
due to restraint can lead to leaking and unsightly con- splitting test (described in ASTM C 496). As indicated in
ditions when water can penetrate the cracks, as in Reference 4, tensile strength measured from the flexure
parking structures. test is usually 40 to 80 percent higher than that measured
Cracking due to restraint causes a reduction in axial from the splitting test.
stiffness, which in turn leads to a reduction (or relax- Representative values of tensile strength obtained
ation) of the restraint force in the member. Therefore, from tests and measures of variability are shown in
the high level stresses indicated in Table 2.1 for small Tables 3.1 and 3.2.
steel ratios may not develop if the cracking is due to ACI 209R suggests the following expressions to esti-
Copyright American Concrete Institute
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 224.2R-4 ACI COMMITTEE REPORT

mate tensile strength as a function of compressive same stress level. Clark and Spiers14 estimated that the
strength first major crack forms at about 90 percent of the aver-
age concrete tensile strength and the last major crack at
modulus of rupture: f, = gr [w, (fc’])]“’ (3.1) about 110 percent of the average tensile strength. Soma-
yaji and Shah15 used a bilinear stress-strain diagram for
direct tensile strength: fi’ = gt [wc Oc,‘)]” (3.2) concrete in tension to model the formation of cracks
along the member at increasing load levels. They as-
where sumed that the tensile strength beyond first cracking was
= unit weight of concrete (lb/ft3) a function of the strain gradient in the concrete along the
;:: = compressive strength of concrete (psi) length of the bar.
gr = 0.60 to 1.00 (0.012 to 0.021 for wc in Induced tensile stresses caused by restrained concrete
kg/m3 and& in MPa) shrinkage affect the amount of cracking that is visible at
gt = 0.33 (0.0069) a given tensile force. This has been made apparent by
tensile tests conducted to compare the performance of
Both the flexure and splitting tests result in a sudden Type I cement and Type K shrinkage-compensating)
failure of the test specimen, indicating the brittle nature cement in concrete specimens.16 Specimens placed under
of plain concrete in tension. However, if the deformation the same conditions of environment and loading had
of the specimen is controlled in a test, a significant markedly different cracking behavior.
descending branch of the tensile stress-strain diagram can When specimens made with Type I cement had fully
be developed beyond the strain corresponding to maxi- developed external cracks, the specimens made with Type
mum tensile stress. Evans and Marathe7 illustrated this K cement exhibited fewer and narrower external cracks.
behavior on specimens loaded in direct tension in a test- The Type K specimens exhibited first cracking at a higher
ing machine modified to control deformation. Fig. 3.1 load than the Type I specimens, and in some tests no
shows tensile stress-strain curves that include unloading visible cracking was evident in Type K specimens.
beyond the maximum tensile stress. More recent work by The compressive stress induced in the concrete by the
Petersson8 shows that the descending branch of the curve restrained expansion of the Type K cement was appar-
is controlled primarily by localized deformatiou across ently responsible for increasing the loads both at first
individual cracks, indicating that there are large dif- cracking and at which cracking was fully developed. Thus,
ferences between the average strain (Fig. 3.1) and local efforts to compensate for concrete shrinkage also appear
strains. to help reduce cracking.

3.3-Development of cracks 3.4-Crack spacing

When a reinforced concrete member is subjected to As a result of the formation of cracks in a tension
tension, two types of cracks eventually form (Fig. 3.2). member, a new stress pattern develops between the
One type is the visible crack that shows at the surface of cracks. The formation of additional primary cracks con-
the concrete, while the other type does not progress to tinues as the stress increases until the crack spacing is
the concrete surface. Broms9 called cracks of the first approximately twice the cover thickness as measured to
type primary cracks and those of the second type se- the center of the reinforcing bar.”
condary cracks.
Each of the two types of cracks has a different geo-
metry. The primary or external cracks are widest at the
500
surface of the concrete and narrowest at the surface of mix w/c age

reinforcing bars.10-12 ” The difference in crack width be- (4)_ 1:1:2 0.45
(5)_1::4 0.60
65 days
270 "
tween the concrete surface and the reinforcing bar is (6)_1:3:6 0.90 70 "

small at low tension levels (just after crack formation),

and increases as the tension level increases; therefore,
the crack width at the reinforcing bar increases more
slowly than the width at the concrete surface with an
increase in load. The deformations on the reinforcing
--``,`,-`-`,,`,,`,`,,`---

bars tend to control the crack width by limiting the slip

between the concrete and the steel.
The secondary, or internal, cracks increase in width I I I I I
with distance away from the reinforcement before nar- 800 1200 1600 2000 2400
longitudinal tensile strain x 106
rowing and closing prior to reaching the concrete surface.
More detail on internal crack formation is presented in
Reference 13.
Because of the variability in tensile strength along the Fig. 3.1-Tensile stress-strain diagrams for concrete7
length
Copyright ofConcrete
American a tension
Institute member, cracks do not all form at the (includes unloading portion)
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 CRACKING OF CONCRETE MEMBERS IN DIRECT TENSION 224.2R-5

cover) at high steel stress by the average strain in the

reinforcement. When tensile members with more than
one reinforcing bar are considered,” the actual concrete
cover is not the most appropriate variable. Instead, an
effective concrete cover t, is used. t, is defined as a
function of the reinforcement spacing, as well as the
concrete cover measured to the center of the reinfor-
--``,`,-`-`,,`,,`,`,,`---

cement. 11 The greater the reinforcement spacing, the

greater will be the crack width. This is reflected as an
increased effective cover. Based on the work of Broms
and Lutz,11 the effective concrete cover is

2,
(3.3)

in which d, = distance from center of bar to extreme

tension fiber, in., and s = bar spacing, in.
The variable t, is similar to the variable 3/o used in
the Gergely-Lutz crack width expression for flexural
members,17 in which A = area of concrete symmetric
with reinforcing steel divided by number of bars (in.2).
Using t,, it is possible to express the maximum crack
width in a form similar to the Gergely-Lutz expression.
Due to the larger variability in crack width in tension
members, the maximum crack width in direct tension is
T expected to be larger than the maximum crack width in
flexure at the same steel stress.
The larger crack width in tensile members may be due
Fig. 3.2-Primary and secondary cracks in a reinforced to the lack of crack restraint provided by the compression
concrete tension member zone in flexural members. The stress gradient in a flex-
ural member causes cracks to initiate at the most highly
There is, of course, a considerable variation in the stressed location and to develop more gradually than in
spacing of external cracks. The variability in the tensile a tensile member that is uniformly stressed.
strength of the concrete, the bond integrity of the bar, The expression for the maximum tensile crack width
and the proximity of previous primary cracks, which tend developed by Broms and Lutz” is
to decrease the local tensile stress in the concrete, are
the main cause of this variation in crack spacing. For the Wmax = 4 ES t, = O.l38f, t, x 10 -3 (3.4)
normal range of concrete covers, 1.25 to 3 in. (30 to 75
mm), the average crack spacing will not reach the limit- Using the definition of t, given in Eq. (3.3), Wmax may be
ing value of twice the cover until the reinforcement stress expressed as
reaches 20 to 30 ksi (138 to 200 MPa).11
The expected value of the maximum crack spacing is
about twice that of the average crack spacing.11 That is, Wmnn = 0.138 f,d, 1 + -& 2 x 1O-3 (3.5)
the maximum crack spacing is equal to about four times I-7-T c
the concrete cover thickness. This range of crack spacing
is more than 20 percent greater than observed for flex- p; yyainet,; 3/C for a single layer of reinforcement
ural members. 2s/d, see Fig. 3.3), which is approximately equal
The number of visible cracks can be reduced at a to 1.35d, Jl+(s/4d,)2 for S/d, between 1 and 2. Thus,
given tensile force by simply increasing the concrete for tensile cracking
cover. With large cover, a larger percentage of the cracks
will remain as internal cracks at a given level of tensile Wmax = 0.10 f, “p x 10-3 (3.6)
force. However, as will be discussed in Section 3.5, in-
creased cover does result in wider visible cracks. Eq. (3.6) can be used to predict the probable maximum
crack width in fully cracked tensile members. As with
3.5-Crack width flexural members, there is a large variability in the
The maximum crack width may be estimated by mul- maximum crack width. One should fully expect the maxi-
tiplying the maximum crack spacing (4 times concrete
Copyright American Concrete Institute
mum crack width to be 30 percent larger or smaller than
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 224.2R-6 ACI COMMITTEE REPORT

tributed between the reinforcing steel and concrete in

proportion to their respective stiffnesses. Total load at
strain E is given by
2d x 2s
c = 2d$
2 P = PC + P, = (Ed, + nE& e
= EP, (1 - p + n p) E = (EA)uCe (4.1)

in which EC = modulus of elasticity of concrete. Loads

carried by the concrete and reinforcing steel are, re-
spectively
Fig. 3.3-3p parameter in terms of bar spacing

the value obtained from Eq. (3.6). PC’ 2-- P

i 1 + np I
The maximum flexural crack width expression17
and
Wma* = O.O76fif, “&& 10 -3 (3.7)
p,” np p (4.3)
in which B = ratio of distance between neutral axis and ( 1 + np 1
tension face to distance between neutral axis and cen-
troid of reinforcing steel = 1.20 in beams may be used Cracking occurs when the strain E corresponds to the ten-
tocompare the crack widths obtained in flexure and ten- sile strength of concrete. If the ascending branch of the
sion. tensile stress-strain curve is assumed to be linear E = Et
Using a value of @ = 1.20, the coefficient 0.076fl in = &‘/EC, in which frl’ is the tensile stress causing the first
Eq. (3.7) becomes 0.091 compared to a coefficient of 0.10 crack. The total load at cracking PCr is carried across the
in Eq. (3.6) for tensile cracks. This indicates that for the crack entirely by the reinforcement. If the applied force
same section and steel stress f, the maximum tensile remains unchanged, steel stress after the crack occurs&
crack will be about 10 percent wider than the maximum is given by
flexural crack.
The flexural crack width expression in Eq. (3.7), with
13 = 1.2, is used in the following form in ACI 318 f,, = % = fr; (f - 1 + n] (4.4)

=f,3@ The load carried across the crack by the reinforcement

Z (3.8)
is gradually transferred by bond to the concrete on each
A maximum value of z = 175 kips/in. (30.6 MN/m) is side of the crack. As the applied load increases, addi-
permitted for interior exposure, corresponding to a tional cracks form at discrete intervals along the member
limiting crack width of 0.016 in. (0.41 mm). ACI 318 as discussed in Chapter 3. The contribution of concrete
limits the value of z to 145 kips/in. (25.4 MN/m) for between cracks to the net stiffness of a member is known
exterior exposure, corresponding to a crack width of as tension stiffening. The gradual reduction in stiffness
0.013 in. (0.33 mm). To obtain similar crack widths for Axial Load
tensile members, the z-factors of 145 and 175 for flexural
--``,`,-`-`,,`,,`,`,,`---

members should be multiplied by the ratio of coefficients

in Eq. (3.7) and (3.6) (= 0.91). Using the same definition
of z for both tensile and flexural members, this produces
z-values of 132 and 160, respectively, for tensile members.
Rizkalla and Hwang18” reported recently on tests in
steel plus concrete
direct tension and presented an alternative procedure for
computing crack widths and crack spacing based on ex-
pressions given by Beeby19 and Leonhardt.20

shaded area represents

CHAPTER 4-EFFECT OF CRACKING contribution of concrete
to overall stiffness
ON AXIAL STIFFNESS

4.1-Axial stiffness of one-dimensional members Axial Strain

When a symmetrical uncracked reinforced concrete
member is loaded in tension. the tensile force is dis-
Copyright American Concrete Institute
Fig. 4.1-Tensile load versus strain diagram
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 CRACKING OF CONCRETE MEMBERS IN DIRECT TENSION 224.2R-7

due to progressive cracking is referred to as strain Other methods for determining E, are reviewed by
softening. Moosecker and Grasser.21
The stiffening effect of the concrete between cracks An alternative approach is to write the effective stiff-
can be illustrated by considering the relationship between ness (EA)c in terms of the modulus of elasticity of the
the load and the average strain in both the uncracked concrete and an effective (reduced) area of concrete, i.e.
and cracked states. A tensile load versus strain curve is
shown in Fig. 4.1. In the range P = 0 to P = P c r , the P=E/I,c, (4.12)
member is uncracked, and the response follows the line
OA. The load-strain relationship [Eq. (4.1)] is given by
This approach is analogous to the effective moment of
inertia concept for the evaluation of deflections de-
P = EP, (I - p + np) E = (EA),,e (4.5) veloped by Branson and incorporated in AC1 318.
--``,`,-`-`,,`,,`,`,,`---

Using the same form of the equation as used for the

If the contribution to stiffness provided by the con-
effective moment of inertia, the effective cross-sectional
crete is ignored, the response follows the line OB, and
area for a member can be written as
the load-strain relationship is given by

P = EfiSe = n E, pA,cl = (EA),E (4.6)

Ae =Ag($r + A$ -(+'I (4.13)
For loads greater than Pcr,, the actual response is inter-
mediate between the uncracked and fully cracked limits,
and response follows the line AD. At Point C on AD, where A9 = gross cross-sectional area and Acr = nA,.
where P is greater than P,,, a relationship can be de- The term Ag could be replaced by the transformed
veloped between the load P and average strain in the area Ae to include the contribution of reinforcing steel to
member E,,, the uncracked system [At = AC + n& = Ag + (n - 1)
AsI*
The load-strain relationships obtained using the CEB
expression and the effective cross-sectional area [Eq.
The term (m), can be referred to as the effective (4.13)] compare quite favorably, as shown in Fig. 4.2.
axial cross-sectional stiffness of the member. This term A third approach that has been used in finite element
can be written in terms of the actual area of steel As, and analysis of concrete structures involves a progressive
an effective modulus of elasticity Esln of the steel bars reduction of the effective modulus of elasticity of con-
crete with increased cracking.
p= EsmAs% (4.8)
or 4.2-Finite element applications
Extensive research has been done in recent years on
f, the application of finite elements to modeling the be-
Esm =; (4.9) havior of reinforced concrete and is summarized in a
m
report of the ASCE Task Committee on Finite Element
Several methods can be used to determine E,. For Analysis of Reinforced Concrete.23 Two basic ap-
example, the CEB Model Code gives proaches, the discrete crack approach and the smeared
crack approach, have been used to model cracking and
tension stiffening.
(4.10) In the discrete crack approach, originally used by Ngo
and Scordelis,244 individual cracks are modeled by using
separate nodal points for concrete elements located at
in which f,, is given by Eq. (4.4), f, = PIA,, E; = LIE,, cracks, as shown for a flexural member in Fig. 4.3. This
and k = 1.0 for first loading and 0.5 for repeated or sus- allows separation of elements at cracks. Effects of bond
tained loading. degradation on tension stiffening can be modeled by
Combining Eq. (4.9) and (4.10) linear24 or nonlinear 255 bond-slip linkage elements con-
necting concrete and steel elements.
The finite element method combined with nonlinear
fracture mechanics was used by Gerstle, Ingraffea, and
Esm =

The CEB expression is based on tests of direct tension

i 01
l_kf”’
f,
(4.11) Gergely26 to study the tension stiffening effect in tension
members. The sequence of formation of primary and se-
condary cracks was studied using discrete crack modeling.
A comparison of analyses with test results is shown in
members conducted at the University of Stuttgart.20
Copyright American Concrete Institute Fig. 4.4.
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 224.2R-8 ACI COMMITTEE REPORT

240

--- CEB
- - - - - - Effective Area, A ,
- - - Steel Alone

180
Lood P
(kips)

120
--``,`,-`-`,,`,,`,`,,`---

l200

Axial Strain (millionths)

Fig. 4.2-Tensile load versus strain diagrams based on CEB and effective cross-sectional area expressions

300 Elastic analysis;

secondary cracking
‘;; ignored
g 250
-Analysis including ’ ’

STEEL ELEMENT

Fig. 4.3-Finite element modeling by the discrete crack

24
approach

In the smeared crack approach, tension stiffening is

modeled either by retaining a decreasing concrete
Elongation, (mm)
modulus of elasticity and leaving the steel modulus
unchanged, or by first increasing and then gradually Fig. 4.4-Steel stress versus elongation curves for tension
decreasing the steel modulus of elasticity and setting the specimen based on nonlinear fracture mechanics ap-
26
concrete modulus to zero as cracking progresses. Scanlon proach
and Murray’ introduced the concept of degrading con-
crete stiffness to model tension stiffening in two-way same general approach can be used for members in
slabs. Variations of this approach have been used in direct tension.
finite element models by a number of researchers.28-31
Gilbert and Warner31 used the smeared crack concept 4.3-Summary
and a layered plate model to compare results using the Several methods have been proposed in the literature
degrading concrete stiffness approach and the increased to estimate the axial stiffness of cracked reinforced
steel stiffness approach. Various models compared by concrete members. The CEB Model Code approach in-
Gilbert and Warner are shown in Fig. 4.5. Satisfactory volves the modification of the effective steel modulus of
results were obtained using all of the models considered. elasticity and appears to be well-established in Europe.
However, the approach using a modified steel stiffness An alternative approach, suggested in this report,
was found to be numerically the most efficient. More involves an expression for the effective cross-sectional
recent works32 has shown that the energy consumed in area that is analogous to the well-known effective
fracture must be correctly modeled to obtain objective moment of inertia concept. Both of these approaches
finite element results in general cases. While most of appear to be acceptable for the analysis of one-
these
Copyright models
American have been applied to flexural members, the
Concrete Institute dimensional members.
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 CRACKING OF CONCRETE MEMBERS IN DIRECT TENSION 224.2R-9

effects. This chapter is intended to provide guidance to

- Layer Containing the Tensile Steel assist in achieving that objective.
-.-.- Layer Once Removed from the Steel
1 - - - Layer Twice Removed from the Steel
The recommendations contained in ACI 224R apply
where applicable. This chapter deals more specifically
with members loaded in direct tension.

5.2-Control of cracking caused by applied loads

The main concern of crack control is to minimize
b) Gradually Unloading Response maximum crack widths. In the past, tolerable crack
After Cracking widths have been related to exposure conditions (ACI
224R). However, at least in terms of protecting rein-
forcement from corrosion, the effect of surface crack
width appears to be relatively less important than be-
lieved previously (ACI 224.1R). For severe exposures, it
is preferable to provide a greater thickness of concrete
cover even though this will lead to wider surface cracks.
Tolerable crack widths may also be related to aesthetic
or functional requirements. Based on experience using
the z-factor for flexural cracking (ACI 318), a crack width
Alternative Stress-Strain Diagram for Concrete in Tension of 0.016 in. (0.4 mm) may be acceptable for appearance
in most cases. Functional requirements such as liquid
storage (ACI 350R) may require narrower crack widths
such as 0.008 in. (0.2 mm) for liquid-retaining structures.
Eq. (3.6) and (3.8) for members in direct tension may
be used to select and arrange reinforcement to limit
crack widths.
Since crack width is related to tensile stress in rein-
forcement, cracks attributed to live loads applied for
short periods may not be as serious as cracks due to
sustained load, since the cracks due to live load may be
a ) I f Ey>E, b) I f Ey’C, expected to close or at least decrease in width upon
removal of the load. If acceptable crack control cannot
Material Modelling Law :
be achieved by the use of mild reinforcement alone, pre-
E,I cz I ca I &b El I E(I 1 c,
kr 1%5.&j 3.&J S.&j 8.a ll.CJU.C, stressing can be used to reduce tensile stresses in a
El E2 E2 EC Es Es ,
structure. Shrinkage-compensating concrete placed in
~0.~27Es~20.E&6.E&15E&Y+ accordance with ACI 223 can also be effective.
While measures can be taken at the design stage to
control cracking, it is equally important to apply proper
Fig. 4.5-Tension stiffening models proposed for smeared construction procedures to insure the intended perfor-
crack finite element approach31 (a, = tensile strength of mance of the structure. This requires avoiding over-
concrete, fv = yield strength of reinforcment, u = stress, loading the structure during construction. Careful
and E = strain) placement of reinforcement is also essential, including
provision for properly designed lap splices (ACI 318).
For more complex systems, finite element analysis pro-
cedures have been used successfully to model the be- 5.3-Control of cracking caused by restraint of volume
havior of cracked reinforced concrete, using a variety of change
stiffness models. Cracking due to the restraint of volume change during
the early life of a structure can be minimized by pro-
tecting new concrete for as long as practical from drying
CHAPTER 5-CONTROL OF CRACKING or temperature drop of the surface of the member that
CAUSED BY DIRECT TENSION would cause tensile stress in the member greater than its
early age strength.
5.1-Introduction Drying shrinkage at early ages can be controlled by
The three previous chapters emphasized predicting be- using proper curing procedures. Shrinkage-compensating
havior when cracking due to direct tension occurs in re- concrete can be very effective in limiting cracks due to
inforced concrete members. However, a major objective --``,`,-`-`,,`,,`,`,,`---
drying shrinkage (ACI 223).
of design and construction of concrete structures should Temperature changes related to the heat of hydration
be
Copyright to Concrete
American minimize
Institute and/or control cracking and its adverse can be minimized by placing concrete at lower than nor-
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 224.2R-10 ACI COMMITTEE REPORT

--``,`,-`-`,,`,,`,`,,`---
ma1 temperatures (precooling). For example, placing con- that of concrete
crete at approximately 50 F (10 C) has significantly P = axial load
reduced cracking in concrete tunnel linings.33 It should PC = axial load carried by concrete
be noted that concrete placed at 50 F (10 C) tends to Pu = axial load at which cracking occurs
develop higher strength at later ages than concrete PS = axial load carried by reinforcement
placed at higher temperatures. s = bar spacing, in.
Circumferential cracks in tunnel linings (as well as t, = effective concrete cover, in.
cast-in-place conduits and pipelines) can be greatly wc = unit weight of concrete, lb/ft3
reduced in number and width if the tunnel is kept bulk- w mar = most probable maximum crack width, in.
headed against air movement and shallow ponds of water = factor limiting distribution of reinforcement
are kept in the invert from the time concrete is placed ; = ratio of distance between neutral axis and
until the tunnel goes into service (see Fig. 35 of tension face to distance between neutral axis
Reference 34). and centroid of reinforcing steel = 1.20 in.
To minimize crack widths caused by restraint stresses, beams
bonded “temperature” reinforcement should be provided. Em
= average strain in member (unit elongation)
As a general rule, reinforcement controls the width and ES = tensile strain in reinforcing bar assuming no
spacing of cracks most effectively when bar diameters are tension in concrete
as small as possible, with correspondingly closer spacing P = reinforcing ratio = A#$
for a given total area of steel. Fiber reinforced concrete
may also have application in minimizing the width of
cracks induced by restraint stresses (ACI 544.1R). CONVERSION FACTORS-SI EQUIVALENTS
If tensile forces in a restrained concrete member will
result in unacceptably wide cracks, the degree of restraint 1 in. = 25.4 mm
can be reduced by using joints where feasible or leaving 1 lb (mass) = 0.4536 kg
empty pour strips that are subsequently filled with con- 1 lb (force) = 4.488 N
crete after the adjacent members have gained strength 1 lb/in.2 = 6.895 kPa
and been allowed to dry. Flatwork will be restrained by 1 kip = 444.8 N
the anchorage of the slab reinforcement to perimeter 1 kip/in.2 = 6.895 MPa
slabs or footings. When each slab is free to shrink from
all sides towards its center, cracking is minimized. For Eq. (3.5)
slabs on ground, contraction joints and perimeter
2
supports should be designed accordingly (ACI 302.1R-
80). Frequent contraction joints or deep grooves must be Wmax = 0.02 fSdc x 10-3
provided if it is desired to prevent or hide restraint
cracking in walls, slabs, and tunnel linings [ACI 224R-80
(Revised 1984), ACI 302.1R-80]. Eq. (3.6)
Wmax = 0*0145f, ‘&X x 10-3

NOTATION Eq. (3.7)

Wmax = O.OlIf, ‘g-x x 10-3
A area of concrete symmetric with reinforcing
steel divided by number of bars, in.’
effective cross-sectional area of concrete, in.’ W,nar in mm, f, in MPa, dc in mm, A in mm2, and s in
A,
gross area of section, in.2 mm.
A,
As area of nonprestressed tension reinforcement,
in.2
dc distance from center of bar to extreme tension CHAPTER 6-REFERENCES
fiber, in.
EC modulus of elasticity of concrete, ksi 6.1-Recommended references
Es modulus of elasticity of reinforcement, ksi The documents of the various standards-producing or-
Eml effective modulus of elasticity of steel to that ganizations referred to in this report are listed below with
of concrete their serial designation.
f, modulus of rupture of concrete, psi
fs stress in reinforcement, ksi American Concrete Institute
fSW steel stress after crack occurs, ksi
fc’ compressive strength of concrete, psi 209R Prediction of Creep, Shrinkage, and Temper-
fi’ tensile strength of concrete, psi ature Effects in Concrete Structures
n the ratio of modulus of elasticity of steel to
Copyright American Concrete Institute
223 Standard Practice for the Use of Shrinkage-
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 CRACKING OF CONCRETE MEMBERS IN DIRECT TENSION 224.2R-11

Compensating Concrete Engineering Report No. 122, Department of Civil En-

224R Control of Cracking in Concrete Structures gineering, University of Alberta, Edmonton, Dec. 1984,
224.1R Causes, Evaluation, and Repair of Cracks in 126 pp.
Concrete Structures 4. Neville, Adam M., Hardened Concrete: Physical and
302.1R Guide for Concrete Floor and Slab Construc- Mechanical Aspects, ACI Monograph No. 6, American
tion Concrete Institute/Iowa State University Press, Detroit,
318 Building Code Requirements for Reinforced 1971, 260 pp.
Concrete 5. Wright, P.J.F., “Comments on Indirect Tensile Test
350R Environmental Engineering Concrete Struc- on Concrete Cylinders,” Magazine of Concrete Research
tures (London), V. 7, No. 20, July 1955, pp. 87-96.
544.1R State-of-the-Art Report on Fiber Reinforced 6. Price, Walter H., “Factors Influencing Concrete
Concrete Strength,” ACI JOURNAL, Proceedings V. 47, No. 6, Feb.
1951, pp. 417-432.
7. Evans, R.H., and Marathe, M.S., “Microcracking
ASTM and Stress-Strain Curves for Concrete in Tension,”
Materials and Structures, Research and Testing (RILEM,
C 78 Standard Test Method for Flexural Strength of Paris), V. 1, No. 1, Jan.-Feb. 1968, pp. 61-64.
Concrete (Using Simple Beam with Third- 8. Petersson, Per-Erik, “Crack Growth and De-
Point Loading) velopment of Fracture Zones in Plain Concrete and
C 293 Standard Test Method for Flexural Strength of Similar Materials,” Report No. TVBM-1006, Division of
Concrete (Using Simple Beam with Center- Building Materials, Lund Institute of Technology, 1981,
Point Loading) 174 pp.
C 496 Standard Test Method for Splitting Tensile 9. Broms, Bengt B., “Crack Width and Crack Spacing
Strength of Cylindrical Concrete Specimens in Reinforced Concrete Members,” ACI JOURNAL , Pro-
ceedings V. 62, No. 10, Oct. 1965, pp. 1237-1256.
10. Broms, Bengt B., “Stress Distribution in Rein-
Co&e’ Euro-International du Be’tonlFbd&ation Inter- forced Concrete Members With Tension Cracks,” ACI
nationale de Ia Pre’contrainte JOURNAL, Proceedings V. 62, No. 9, Sept. 1965, pp. 1095-
1108.
CEB-FIP Model Code for Concrete Structures 11. Broms, Bengt B., and Lutz, Leroy A., “Effects of
Arrangement of Reinforcement on Crack Width and
Spacing of Reinforced Concrete Members,” ACI
These publications may be obtained from: J OURNAL , Proceedings V. 62, No. 11, Nov. 1965, pp.
1395-1410.
American Concrete Institute 12. Goto, Yukimasa, “Cracks Formed in Concrete
PO Box 19150 Around Deformed Tension Bars,” ACI JO U R N A L,
Detroit, MI 48219 Proceedings V. 68, No. 4, Apr. 1971, pp. 244-251.
13. Goto, Y., and Otsuka, K., “Experimental Studies
ASTM on Cracks Formed in Concrete Around Deformed Ten-
1916 Race St. sion Bars,” Technology Reports of the Tohoku University,
Philadelphia, PA 19103 V. 44, No. 1, June 1979, pp. 49-83.
14. Clark, L.A., and Spiers, D.M., “Tension Stiffening
Comite Euro-International du B&on in Reinforced Concrete Beams and Slabs Under Short-
EPFL Case Postale 88 Term Load,” Technical Report No. 42.521, Cement and
CH 1015 Lausanne, Switzerland Concrete Association, Wexham Springs, 1978, 19 pp.
15. Somayaji, S., and Shah, S.P., “Bond Stress Versus
Slip Relationship and Cracking Response of Tension
6.2-Cited references Members,” ACI JOURNAL , Proceedings V. 78, No. 3, May-
1. Scanlon, Andrew, and Murray, David W., “Practical June 1981, pp. 217-225.
Calculation of Two-Way Slab Deflections,” Concrete In- 16. Cusick, R.W., and Kesler, C.E., “Interim Report-
ternational: Design & Construction, V. 4, No. 11, Nov. Phase 3: Behavior of Shrinkage-Compensating Concretes
1982, pp. 43-50. Suitable for Use in Bridge Decks,” T. & A.M. Report No.
2. Cole, Peter P.; Abel, John F.; and Billington, David 409, Department of Theoretical and Applied Mechanics,
P., “Buckling of Cooling-Tower Shells: State-of-the-Art,” University of Illinois, Urbana, July 1976, 41 pp.
Proceedings, ASCE, V. 101, ST6, June 1975, pp. 1185- 17. Gergely, Peter, and Lutz, Leroy A., “Maximum
1203. Crack Width in Reinforced Concrete Flexural Members,”
3. Tam, K.S.S., and Scanlon, A., “The Effects of Causes, Mechanism, and Control of Cracking in Concrete,
--``,`,-`-`,,`,,`,`,,`---

Restrained
Copyright American ConcreteShrinkage
Institute on Concrete Slabs,” Structural SP-20, American Concrete Institute, Detroit, 1968, pp.
Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale
 224.2R-1 2 ACI COMMlTTE E REPORT

87-117. ference on Bond in Concrete (Paisely, June 1982), Ap-

18. Rizkalla, S.H., and Hwang, L.S., “Crack Prediction plied Science Publishers, London, 1982, pp. 97-106.
for Members in Uniaxial Tension,” ACI JOURNAL , Pro- 27. Scanlon, A., and Murray, D.W., “An Analysis to
ceedings V. 81, No. 6, Nov.-Dec. 1984, pp. 572-579. Determine the Effects of Cracking in Reinforced Con-
19. Beeby, A.W., “The Prediction of Crack Widths in crete Slabs,” Proceedings, Specialty Conference on the
Hardened Concrete,” Structural Engineer (London), V. Finite Element Method in Civil Engineering, EIC/McGill
57A, Jan. 1979, pp. 9-17. University, Montreal, 1972, pp. 841-867.
20. Leonhardt, Fritz, “Crack Control in Concrete 28. Lin, Cheng-Shung, and Scordelis, Alexander C.,
Structures,” IABSE Surveys No. S-4/77, International “Nonlinear Analysis of RC Shells of General Form,”
Association for Bridge and Structural Engineering, Proceedings, ASCE, V. 101, ST3, Mar. 1975, pp. 523-538.
Zurich, Aug. 1977, 26 pp. 29. Chitnuyanondh, L.; Rizkalla, S.; Murray, D.W.;
21. Moosecker, W., and Grasser, E., “Evaluation of and MacGregor, J.G., “An Effective Uniaxial Tensile
Tension Stiffening in Reinforced Concrete Linear Mem- Stress-Strain Relationship for Prestressed Concrete,”
bers,” Final Report, IABSE Colloquium on Advanced Structural Engineering Report No. 74, University of
Mechanics of Reinforced Concrete (Delft, 1981), Inter- Alberta, Edmonton, Feb. 1979, 91 pp.
national Association for Bridge and Structural 30. Argyris, J.H.; Faust, G.; Szimmat, J.; Warnke, P.,
Engineering, Zurich. and William, K.J., “Recent Developments in the Finite
2 2 . Branson, D . E . , “Instantaneous and Time- Element Analysis of Prestressed Concrete Reactor
Dependent Deflections of Simple and Continuous Vessels,” Preprints, 2nd International Conference on
Reinforced Concrete Beams,” Research Report No. 7, Structural Mechanics in Reactor Technology (Berlin,
Alabama Highway Department, Montgomery, Aug. 1963, Sept. 1973), Commission of the European Communities,
94 pp. Luxembourg, V. 3, Paper H l/l, 20 pp. Also, Nuclear
23. Finite Element Analysis of Reinforced Concrete, Engineering and Design (Amsterdam), V. 28, 1974.
American Society of Civil Engineers, New York, 1982, 31. Gilbert, R. Ian, and Warner, Robert F., ‘‘Tension
545 pp. Stiffening in Reinforced Concrete Slabs,” Proceedings,
24. Ngo, D., and Scordelis, A.C., “Finite Element ASCE, V. 104, ST12, Dec. 1978, pp. 1885-1900.
Analysis of Reinforced Concrete Beams,” ACI JOURNAL., 32. Bazant, Zdenek, and Cedolin, Luigi, “Blunt Crack
Proceedings V. 64, No. 3, Mar. 1967, pp. 152-163. Band Propagation in Finite Element Analysis,”
25. Nilson, Arthur H., “Nonlinear Analysis of Proceedings, ASCE, V. 105, EM2, Apr. 1979, pp. 297-315.
Reinforced Concrete by the Finite Element Method,” 33. Tuthill, Lewis H., “Tunnel Lining With Pumped
ACI JOURNAL , Proceedings V. 65, No. 9, Sept. 1968, pp. Concrete,” ACI JOURNAL., Proceedings V. 68, No. 4, Apr.
757-766. 1971, pp. 252-262.
26. Gerstle, Walter; Ingraffea, Anthony R.; and 34. Concrete Manual, 8th Edition, U.S. Bureau of
--``,`,-`-`,,`,,`,`,,`---

Gergely, Peter, ‘‘Tension Stiffening: A Fracture Reclamation, Denver, 1975, 627 pp.
Mechanics Approach,” Proceedings, International Con-

Copyright American Concrete Institute

Provided by IHS under license with ACI
No reproduction or networking permitted without license from IHS Not for Resale