22 TRANSPORTATION RESEARCH RECORD 1301

Flexural Cracking in Concrete Structures

EDWARD G. NAWY

The state-of-the-art in the evaluation of the flexural crack width Although the macrocracking aspects of cracking behavior
development and crack control of macrocracks is described. It is are emphasized, it is also important to briefly discuss micro-
based on extensive research over the past 50 years in the United cracking.
States and overseas in the area of macrocracking in reinforced
and prestressed concrete beams and two-way-action slabs and
plates. Control of cracking has become essential to maintain the
integrity and aesthetics of concrete structures. The trends are MICROCRACKING
stronger than ever-toward better use of concrete strength, use
of higher-strength concretes including superstrength concretes of
over 20,000-psi compressive strength, use of more prestressed Microcracking can be mainly classified into two categories:
concretes, and increased use of limit failure theories-all re- (a) bond cracks at the aggregate-mortar interface, and (b)
quiring closer control of serviceability requirements of cracking paste cracks within the mortar matrix. Interfacial bond cracks
and deflection behavior. Common expressions are discussed for are caused by interfacial shear and tensile stresses caused by
the control of cracking in reinforced-concrete beams and thick early volumetric change without the presence of external load.
one-way slabs; prestressed, pretensioned, and posttensioned flanged Volume change caused by hydration and shrinkage could cre-
beams; and reinforced-concrete, two-way-action, structural floor
slabs and plates. In addition, recommendations are given for the ate tensile and bond stresses of sufficient magnitude to cause
maximum tolerable flexural crack widths in concrete elements. failure at the aggregate-mortar interface (1). As the external
load is applied, mortar cracks develop because of increase in
compressive stress, propagating continuously through the ce-
Presently, the trend is stronger than ever-toward better use ment matrix up to failure. A typical schematic stress-strain
of concrete strength, use of higher-strength concretes includ- diagram (Figure 1) shows that the nonlinear relationship de-
ing superstrength concretes of 20,000-psi (138-MPa) com- veloped early in the stress history and started with bond mi-
pressive strength and higher, use of high-strength reinforce- nocracking. Although extensive work exists in the area of
ment, use of more prestressed concretes, and increased use volumetric change cracking, the need is apparent for addi-
of limit failure theories-all requiring closer control of ser- tional work on creep effects on microcracking and also for
viceability requirements in cracking and deflection behavior. the development of a universally acceptable fracture theory
Hence, knowledge of the cracking behavior of concrete ele- to interrelate the nonlinear behavioral factors resulting in
ments becomes essential. crack propagation.
Concrete cracks early in its loading history. Most cracks It appears that the damage to cement paste seems to play
are a result of the following actions to which concrete can be a significant role in controlling the stress-strain relationship
subjected: in concrete. The coarse aggregate particles act as stress raisers
that decrease the strength of the cement paste. As a result,
1. Volumetric change caused by drying shrinkage, creep microcracks develop that can only be detected by large mag-
under sustained load, thermal stresses including elevated tem- nification. The importance of additional work lies not only in
peratures, and chemical incompatibility of concrete compo- the evaluation of the microcracks, but also in the evaluation
nents. of their significance for the development of macrocracks that
2. Direct stress caused by applied loads or reactions or generate from those microcracked centers of plasticity.
internal stress caused by continuity, reversible fatigue load,
long-term deflection, camber in prestressed systems, and en-
vironmental effects including differential movement in struc-
FLEXURAL CRACKING
tural systems.
3. Flexural stress caused by bending.
External load results in direct and bending stresses, causing
flexural, bond, and diagonal tension cracks. Immediately after
Although the net result of these three actions is the for-
the tensile stress in the concrete exceeds its tensile strength,
mation of cracks, the mechanisms of their development can-
internal microcracks form. These cracks generate into macro-
not be considered to be identical. Volumetric change gen-
cracks propagating to the external fiber zones of the
erates internal microcracking that may qevelop into full
element.
cracking, whereas direct internal or external stress or applied
Immediately after the full development of the first crack in
loads and reactions could either generate internal microcrack-
a reinforced-concrete element, the stress in the concrete at
ing, such as in the case of fatigue caused by reversible load,
the cracking zone is reduced to zero and is assumed by the
or flexural macrocracking leading to fully developed cracking.
reinforcement (2). The distributions of ultimate bond stress,
Department of Civil and Environmental Engineering, Rutgers Uni- longitudinal tensile stress in the concrete, and longitudinal
versity, Piscataway, N.J. 08855. tensile stress in the steel are shown in Figure 2.
Nawy 23

In Figure 2, crack width is a primary function of the de-

formation of reinforcement between adjacent Cracks 1 and
J:
I- 2, if the small concrete strain along the crack interval ac is
C>
z SLOW RAPID neglected. The crack width would hence be a function of the
w CRACK -->---CRACK
er GROWTH GROWTH crack spacing, and vice versa, up to the level of stabilization
1-
(J) of crack spacing (Figure 3).
w 100 The major parameters affecting the development and char-
~ Increase in acteristics of the cracks are percentage of reinforcement, bond
::!:
5:::> 70
------~~~~~~ crocking characteristics and size of bar, concrete cover, and the con-
crete stretched area (namely, the concrete area in tension).
u.. Increase in
0 On this basis, one can propose the following mathematical
1-
z
~ 30
-----t ~~~~--crocking model:

ffiQ. w = a a~EJ (1)

where
UNIT STRAIN - PERCENT
w = maximum crack width,
FIGURE 1 Schematic stress-strain diagram of concrete
in microcracking. a, 13, and -y = nonlinearity constants, and
Es = strain in the reinforcement induced by ex-
ternal load.
Crack spacing ac is a function of the factors enumerated pre-
viously, being inversely proportional to bond strength and
active steel ratio (steel percentage in terms of the concrete
area in tension).
The basic mathematical model in Equation 1 with the ap-
propriate experimental values of the constants a, 13, and -y
can be derived for the particular type of structural member.
Such a member can be a one-dimensional element such as a
beam, a two-dimensional structure such as a two-way slab, or
2 a three-dimensional member such as a shell or circular tank
wall. Hence, it is expected that different forms or expressions
(8J
apply for the evaluation of the macrocracking behavior of
different structural elements consistent with their fundamen-
Tµ tal structural behavior (1-10).

l_
FLEXURAL CRACK CONTROL IN REINFORCED-
CONCRETE BEAMS

(bl Requirements for crack control in beams and thick one-way

slabs in the American Concrete Institute (ACI) Building Code

lT~
(cl
"'cic:"
·c::;
"'c.
T ~"'
0
~

1 u

(d)

FIGURE 2 Schematic stress distributions (a) between two Crack Width, Wmax
flexural cracks for (b) ultimate bond stress, (c) longitudinal
tensile stress in the concrete, and (d) longitudinal tensile stress FIGURE 3 Schematic variation of crack
in the steel. width with crack spacing.
24 TRANSPORTATION RESEARCH RECORD 1301

(ACI 318) are based on the statistical analysis of maximum

crack width data from a number of sources. The following
general conclusions were reached:

1. The steel stress is the most important variable;

2. The thickness of the concrete cover is an important var- 400
iable, but not the only geometric consideration; 300
3. The area of concrete surrounding each reinforcing bar
is also an important geometric variable; 200
4. The bar diameter is not a major variable; and 150
5. The size of the bottom crack width is influenced by the 100
amount of strain gradient from the level of the steel to the 80
tension face of the beam. 60 :{
)
40
The simplified expression relating crack width to steel stress
is as follows (4):
20
(2)

where 10
8
lsreinforcing steel stress, kips/in. 2 (ksi);
=
6
A area of concrete symmetric with reinforcing steel di-
vided by number of bars, in. 2 ; ----._r1"1e 4
de = thickness of concrete cover measured from extreme
tension fiber to center of bar or wire closest thereto,
r-~~-:-:--~~~~~~---,-~---
--INTERIOR EXPOSURE
--
w • 0 0 16- --
3
2 -- - --- EXTERIOR EXPOSURE w•OOIJ" 2
in.; and
h1 = distance from neutral axis to the reinforcing steel, in.; 50 60 70 80
h2 = distance from neutral axis to extreme concrete tensile Steel Yield Stress, ksi

surface, in.; and FIGURE 4 Steel reinforcement strength J,. versus ratio of
s
= h2fh1. concrete area in tension to reinforcement area for stress level
/, = 0.6J,..
A plot relating the reinforcement strength to the ratio of the
concrete area in tension to the reinforcement area is shown
stressed and reinforced concrete can be summarized as fol-
in Figure 4 for all bar sizes.
lows:
In the ACI code, when the design field strength lY for ten-
The mean crack width wm in beams is expressed in terms
sion reinforcement exceeds 40,000 psi, cross sections of max-
of the mean crack spacing s,m, such that
imum positive and negative moment have to be so propor-
tioned that the quantity z given by
(5)
(3)
:S 0.4 ls (6)
does not exceed 175 kips/in. for interior exposure and 145 Es
kips/in. for exterior exposure. Calculated stress in the rein-
where
forcement at service load ls (ksi) shall be computed as the
moment divided by the product of steel area and internal Esm = average strain in the steel,
moment area. In lieu of such computations, it is permitted to ls = steel stress at the crack,
take ls as 60 percent of specified lengthly· ls, = steel stress at the crack caused by cracking forces at
When the strain Es in the steel reinforcement is used instead the tensile strength of concrete, and
of stress l,, Equation 3 becomes x = bond coefficient (1.0 for ribbed bars, reflecting influ-
ence of load repetitions and load duration).
(4)
The mean crack spacing is
Equation 4 is valid in any system of measurement.
The cracking behavior in thick one-way slabs is similar to
that in shallow beams. For one-way slabs having a clear con- (7)
crete cover in excess of 1 in. (25.4 mm), Equation 4 can be
adequately applied if S = 1.25 to 1.35. where
c = clear concrete cover;
Committee Eurolnternationale du Beton (CEB)
s = bar spacing, limited to l5db;
Recommendations
x2 = coefficient that is 0.4 for ribbed bars;
Crack control recommendations proposed that the European X3 = coefficient that depends on the shape of the stress
Model Code for Concrete Structures (9) apply to both pre- diagram, 0.125 for bending;
Nawy 25

QR = AJA,; and The total tensile force T transferred from the steel to the
A, = effective area in tension. concrete over the stabilized mean crack spacing can be defined
as
Depending on arrangement of bars and type of external forces,
A, is limited by a line c + 7db from the tension face for beams
(9a)
(in the case of thick slabs, not more than halfway to the neutral
axis).
where
A simplified formula can be derived for the mean crack
width in beams with ribbed bars. -y = a factor reflecting the distribution of bond stress;
µ = maximum bond stress, which is a function of f~ 112 ;
acs = mean stabilized spacing; and
(8)
~o = sum of reinforcing element circumferences.
The resistance R of the concrete area A, in tension can be
A characteristic value of the crack width, presumably equiv- defined as
alent to the probable maximum value, is given by 0.7wm.
R = A,J; (9b)

FLEXURAL CRACK CONTROL IN where f; = tensile splitting strength of the concrete. By equating
PRESTRESSED,PRETENSIONED,AND Equations 9a and 9b, the following expression for acs is ob-
POSTTENSIONED BEAMS tained:

The increased use of partial prestressing, allowing limited

(lOa)
tensile stresses in the concrete under service and overload
conditions while allowing nonprestressed steel to carry the
tensile stresses, is becoming prevalent because of practicality where c is a constant to be developed from the tests. The
and economy. Consequently, an evaluation of the flexural concrete stretched area, namely the concrete area in tension
crack widths and spacing and control of their development A, for both the evenly distributed and nonevenly distributed
become essential. Work in this area is relatively limited be- reinforcing elements, is shown in Figure 5. With a mean value
cause of the various factors affecting crack width development off,' 1(!~) 112 = 7 .95 in this investigation, a regression analysis
in prestressed concrete. However, experimental investiga-
tions support the hypothesis that the major controlling pa-
rameter is the reinforcement stress change beyond the de- b
compression stage. Nawy et al. have undertaken extensive
research since the 1960s on the cracking behavior of pre-
stressed, pretensioned, and posttensioned beams and slabs
because of the great vulnerability of the highly stressed pre-
stressing steel to corrosion and other environmental effects
and the resulting premature loss of prestress (11,12). Ser-
viceability behavior under service and overload conditions can
c . ~. s .
be controlled by the design engineer through the application
of the criteria presented in this section.

(a)
Mathematical Model Formulation for Serviceability
Evaluation

Crack Spacing

Primary cracks form in the region of maximum bending mo-

ment when the external load reaches the cracking load. As ...."
loading is increased, additional cracks will form and the num-
ber of cracks will be stabilized when the stress in the concrete C.G.S.
no longer exceeds its tensile strength at further locations re-
gardless of load increase. This condition is important as it
essentially produces the absolute minimum crack spacing that (b)
can occur at high steel stresses, to be termed the stabilized
minimum crack spacing. The maximum possible crack spacing FIGURE 5 Effective concrete
area in tension (a) for even
under this stabilized condition is twice the minimum, to be distribution of reinforcement in
termed the stabilized maximum crack spacing. Hence, the concrete, and (b) for noneven
stabilized mean crack spacing acs is deduced as the mean value distribution of reinforcement in
of the two extremes. concrete.
26 TRANSPORTATION RESEARCH RECORD 1301

of the test data resulted in the following expression for the Expressions for Pretensioned Beams
mean stabilized crack spacing:
Equation 12a is rewritten in terms of tlfs so that analysis of
acs = l.20A,!lo (lOb) the test data of all the simply supported test beams in this
work leads to the following expression at the reinforcement
level:
Crack Width
(in.) (13)
If tlf, is the net stress in the prestressed tendon or the mag-
nitude of the tensile stress in the normal steel at any crack Linearizing Equation 13 for easier use by the design en-
width load level in which the decompression load (decompres- gineer leads to the following simplified expression of the max-
sion here means fc = 0 at the level of the reinforcing steel) imum crack width at the reinforcing steel level:
is taken as the reference point, then for the prestressed tendon
Wmax = (5.85 X 10- 5 ) ;~ (tlfs) (14a)
(11)
and a maximum crack width (in.) at the tensile face of the
where
concrete:
fnr

fd
=

=
stress in the prestressing steel at any load beyond the
decompression load, and
stress in the prestressing steel corresponding to the
w~ax = (5.85 x 10- 5 ) R; :I
.:..0
(tlf,)
.
(14b)

decompression load.
A plot of the data and the best-fit expression for Equation
The unit strain Es = tlfs/ Es. It is logical to disregard as insig- 14a are shown in Figure 6 with a 40 percent spread (which is
nificant the unit strains in the concrete caused by the effects reasonable in view of the randomness of crack development
of temperature, shrinkage, and elastic shortening. The max- and the linearization of the original expression in Equa-
imum crack width as defined in Equation 1 can be taken as tion 13).

(12a)
Expressions for Posttensioned Beams
where k and a are constants to be established by tests, or
The expression developed for the crack width in posttensioned
bonded beams that contain mild steel reinforcement is
(12b)

where k' is a constant in terms of constant k. (6.51 x 10-s) A, (tlf) (15a)

wmax
LO s

I KN/mm I
x10'~~~~.....:..;
•0;..._~~~~~~~.....:..3o
__~~~~~~~~s~o;..._~~~~~
0
0

16 0 ,4
0
z 0

:i::
.... l'l.
0
-~
E
E
..""u
a: 8 O.'l
u
.,,
::I!
::I!

..
;;:
::I!
4

(KIP/1111)

FIGURE 6 Linearized maximum crack width versus (A,ILo) D..J; for pretensioned beams.
Nawy 27

(KN/ m11 I
•J 10 30 50
x10
/
/
14 / 0.6
z /
/
::c /
I-
c 18
~ 0.4
..
u
c
E
E
a:
u 12
::E
:>
::E 0.2
j(
c
::E 6

(KIP/IN.I

FIGURE 7 Linearized maximum crack width versus (A,!Lo) il.f, for posttensioned
beams.

for the width at the reinforcement level closest to the tensile in Figure 8. In this plot, crack spacing stabilizes at a net stress
face, and level of 36 ksi.

Wmax = (6.51 X 1Q- 5 )R; :~ (6.fs) (15b)

Other Work on Cracking in Prestressed Concrete

at the tensile face of the concrete lower fibers. On the basis of the analysis of results of various investigators,
For nonbonded beams, the factor 6.51 in Equations 15a Naaman (8) produced the following modified expression for
and 15b becomes 6.83. partially prestressed pretensioned members
A plot of the data and the best-fit expression for Equation
15a are shown in Figure 7.
A typical plot of the effect of the various steel percentages
on the crack spacing at the various steel levels 6-fs is shown
(wmax) = [ 42 + 5.58:~ (6.Js) J X 10-s (in.) (16)

STEEL STRESS I MPa I

100 200 300 400 500

16 40
i
P= o.25%
CJ
z
0 12 30
c
0..

..
"'
u
..
E

c 20
a:
0

z p = o.60 %
c
w
::E 4

0
5 25 45 65 ala
STEEL STRESS IKSll

FIGURE 8 Effect of steel percentage on mean crack spacing in prestressed beams.

28 TRANSPORTATION RESEARCH RECORD 1301

This regression expression is close to Equation 14. When Flexural Cracking Mechanism and Fracture
plotted against the experimental results of the various re- Hypothesis
searchers, it gives a best fit as shown in Figure 9.
The author's equations and the CEB-Federation Interna- Flexural cracking behavior in concrete structural floors under
tionale de la Precontrainte (FIP) equations can be compared two-way action is significantly different from that in one-way
using similar notations. members. Crack control equations for beams underestimate
the crack widths developed in two-way slabs and plates, and
do not tell the designer how to space the reinforcement.
Nawy:
Cracking in lwo-way slabs and plates is rnnlrolled primarily
by the steel stress level and the spacing of the reinforcement
in the two perpendicular directions. In addition, the clear
CEB-FIP: concrete cover in two-way slabs and plates is nearly constant
(%in. (19 mm) for interior exposure], whereas it is a major
variable in the crack control equations for beams. The results
These equations are similar assuming 1/lo = i\( cp/p,), where
from extensive tests on slabs and plates by Nawy et al. dem-
<P is the diameter of the bar, ,\ is a multiplier, k and b are
onstrate this difference in behavior in a fracture hypothesis
experimental parameters, and a 0 and l:::..uP,11:::..u/ are terms in
on crack development and propagation in two-way plate ac-
the CEB-FIP expression not of major significance that are
tion. As shown in Figure 10, stress concentration develops
accounted for by using I,; = 6.51 for the posttensioned beams
initially at the points of intersection of the reinforcement in
in the author's expressions.
the reinforcing bars and at the welded joints of the wire mesh,
The study by Meier and Gergely (10), concentrating on the
that is, at grid nodal points, thereby dynamically generating
area of concrete in tension and the nominal strain in the
fracture lines along the paths of least resistance, namely, along
concrete at the tensile face, does not yield a reliable prediction
A 1 B 1 , A 1 A 2 , A 2 B 2 , and B 2 B 1 • The resulting fracture pattern
of the crack width. In particular, it does not account for the
is a total repetitive cracking grid, provided that the spacing
actual stress in the steel reinforcement and depends on mea-
of the nodal points Al> B,, A 2 , and B 2 is close enough to
surements of strain at the concrete surface that are difficult
generate this preferred initial fracture grid of orthogonal cracks
to reliably evaluate.
narrow in width as a preferred fracture mechanism.
If the spacing of the reinforcing grid intersections is too
large, the magnitude of stress concentration and the energy
FLEXURAL CRACK CONTROL IN TWO-WAY- absorbed per unit grid is too low to generate cracks along the
ACTION SLABS AND PLATES reinforcing wires or bars. As a result, the principal cracks
follow diagonal yield-line cracking in the plain concrete field
Flexural crack control is essential in structural floors where away from the reinforcing bars early in the loading history.
cracks at service load and overload conditions can be serious, These cracks are wide and few.
such as in office buildings, schools, parking garages, industrial This hypothesis also leads to the conclusion that surface
buildings, and other floors where the design service load levels deformations of the individual reinforcing elements have little
exceed those in normal-sized apartment building panels and effect in arresting the generation of the cracks or controlling
also in all cases of adverse exposure conditions. their type or width in a slab or plate of two-way action. In a

KN/mm I
_, 10 30
•IO.-~~~~~......;,;;-~~~~~~...-~~~~~~=,=-~~~~~~...-~--.

i O •AO .. BABU LEAST SQUARE LINE

= (41 +~.SIX }10

4 AAUASWAMY. Al. 1
0 BENNETT .. CHANDAASEKAA
(Wma,)in
0.5

..
::c
a
~ NAWY .. HUANG
a
;J 15
E
E
"
u
c
a:
u
:I
::J
:I
)( 5
c
:I

•• f •
x = x.;6 )KIP/IN.I

FIGURE 9 Reinforcement stress versus crack width (best-fit data or several

investigators).
Nawy 29

T (/)
T RANSV ERS E STEEL

1
CONTROLLING NON-ORTHOGONAL
CRACKING AWAY FROM BARS A1 B 1
OR A 2 B2 WHEN GRID SPACING S1
OR S2 EXCEEDS 7 TO 12 INCHES

FIGURE 10 Grid unit in two-way-action reinforcement.

similar manner, one may conclude that the scale effect on for 90 slabs tested to failure, the following equation for crack
cracking behavior during two-way action is insignificant, be- control emerged:
cause the cracking grid would be a reflection of the reinforce- 112
ment grid if the preferred orthogonal narrow cracking widths = K f d,,,s2 (18)
develop. Therefore, to control cracking in floors with a two- w p s ( Q )
JI
way action, the major parameter to be considered is the re-
inforcement spacing in two perpendicular directions. Con- where
crete cover has only a minor effect, because it is usually a
small constant of value 0.75 in. (19 mm). w = crack width at concrete face caused by flexural load,
For a constant area of steel determined for bending in one in.;
direction, that is, for energy absorption per unit slab area, k = fracture coefficient, in. 2 /lb;
the smaller the spacing of the transverse bars or wires, the 13 = ratio of the distance from the neutral axis to the
smaller should be the diameter of the longitudinal bars. The tensile face of the slab to the distance from the neu-
reason is that less energy has to be absorbed by the individual tral axis to the centroid of the reinforcement grid;
longitudinal bars. When the magnitude of fracture is deter- Is = actual average service load stress level, or 40 percent
mined by the energy imposed per specific volume of rein- of the design yield strength, ksi;
forcement acting on a finite element of the slab, a proper db, diameter of the reinforcement in Direction 1 closest
choice of the reinforcement grid size and bar size can control to the concrete outer fibers, in.;
cracking into preferred orthogonal grids. s2 spacing of the reinforcement in the perpendicular
This hypothesis is important for serviceability and reason- Direction 2, in.;
able overload conditions. In relating orthogonal cracks to As = area of steel per foot of width of concrete, in. 2 ;
yield-line cracks, the failure of a slab ultimately follows the c 1 = clear concrete cover measured from the tensile face
generally accepted rigid-plastic yield-line criteria. of the concrete to the nearest edge of the reinforcing
bar in Direction 1, in.; and
Q;i = active steel ratio, given by A)l2(db 1 + 2c 1).

Crack Control Equation Direction 1 is the direction of the reinforcement closest to

the outer concrete fibers; this is the direction for which the
The basic Equation 1 for relating crack width to strain in the crack control check is to be made.
reinforcement is The quantity whose square root is taken is termed the "grid
index," and can be transformed as follows:
w= aa~ E'J (17)
(19)
where
ac = crack spacing,
in which s 1 is the spacing of the reinforcement in Direction 1.
Es = unit strain in the reinforcement, and
For uniformly loaded, restrained, square slabs and plates
a, 13, '{ = constants.
of two-way action, k = 2.8 x 10-s in. 2 /lb. For concentrated
The effect of the tensile strain in the concrete between the loads or reactions, or when the ratio of short span to long
cracks is neglected as insignificant. span is less than 0.75 but greater than 0.5, k = 2.1 x 10-s
As a result of the fracture hypothesis, the mathematical in. 2 /lb. For a span aspect ratio of less than 0.5, k = 1.6 x
model of Equation 17, and the statistical analyses of the data 10-s in. 2 /lb.
30 TRANSPORTATION RESEARCH RECORD 1301

Although ~ varies in value between 1.20 and 1.35, the 1300 2600 3900 5200 cm•
intermediate value ~ = 1.25 was used to simplify the calcu- mm
lations.
.028 0.7
Subscripts 1 and 2 generally pertain to the two directions
of reinforcement. Detailed values of the fracture coefficients
""
u
.E
for various boundary conditions are presented in Table 1. .,,, .0 24 0.6
A graphical solution of Equation 18 is shown in Figure 11 '
3:
for fy = 60,000 psi (414 MPa) and f, = 0.4 fy = 24,000 psi I
.020 0.5
(165.5 MPa) for rapid determination of the reinforcement size f-
0
and spacing needed for crack control. 3: .0 16
"'
u
<(
Permissible Crack Widths in Concrete Structures ct:
u .012

The maximum crack width that a structural element should

.008 0.2
be permitted to develop depends on the particular function
t y =60 KSI
of the element and the environmental conditions to which the 18 •24 KSI (•0.401 )
.004 1 0.1
structure is liable to be subjected. Table 2 from the ACI
Committee 224 report on cracking serves as a reasonable 200 400 600 800 in 2
0
guide on the permissible crack widths in concrete structures
GRID INDEX =db, s2 / p11
under the various environmental conditions that are normally
encountered.
The crack control equation and guidelines presented are
important not only for the control of corrosion in the rein-
FIGURE 11 Crack control reinforcement
forcement but also for deflection control. The reduction of distribution in two-way-action slabs and plates
the stiffness EI of the two-way slab or plate caused by or- for all exposure conditions: fy = 60,000 psi,/,
thogonal cracking when the limits of permissible crack widths = 24,000 psi ( = 0.40/y).
in Table 2 are exceeded, can lead to excessive deflection, both
short-term and long-term. Deflection values several times those
anticipated in the design, including deflection caused by con- crack width can vary considerably in cases of cyclic loading,
struction loading, can be reasonably controlled through cam- such as in bridges, but the width increases at a decreasing rate
ber and control of the flexural crack width in the slab or plate. with time. In most cases, a doubling of crack width after
Proper selection of the reinforcement spacings s 1 and s 2 in the several years under sustained loading can be expected.
perpendicular directions as discussed in this section, and not
exceeding 12 in. center to center, can maintain good ser-
viceability performance of a slab system under normal and CONCLUSIONS
reasonable overload conditions.
In most cases, the magnitude of crack widths increases in With the aid of the expressions summarized in the following,
long-term exposure and long-term loading. The increase in the design engineer and the constructor can limit the flexural

TABLE 1 FRACTURE COEFFICIENTS FOR SLABS AND PLATES

Fracture
c
Loading Slab Boundary Span ratio. coefficient
a conditionb 10-5 K
type shape SIL

A Square 4 edges r 1.0 2.1

A Square 4 edges s 1.0 2.1
B Rectangular 4 edges r 0.5 1.6
B Rectangular 4 edges r 0.7 2.2
B Rectangular 3 edges r, 0.7 2.3
1 edge h
B Rectangular 2 edges r, 0.7 2.7
2 edges r
B Square 4 edges r 1.0 2.8
B Square 3 edges r, 1. 0 2.9
1 edge h
B Square 2 edges r. 1.0 4.2
2 edges h

aLoading type: A, concentrated; B. uniformly distributed.

b
Boundary condition: r, restrained; s, simply supported; h. hinges.
CSpan ratio: s. clear short span; L, clear long span .
Nawy 31

TABLE 2 MAXIMUM PERMISSIBLE FLEXURAL CRACK WIDTHS

Cra ck wid t h
Exposure condition in. mm .
Dry air or protective membrane 0 . 016 0.41
Humidity, moist air, soil 0.012 0.30
De-icing chemicals 0.007 0.18
Seawater and seawater spray; 0.006 0.15
wetting and drying
Water-retaining structures 0 . 004 0.10
(excluding nonpressure pipes)

macrocrack wid th that develops in con rete systems. By lim- Values of coefficient Kare presented in Table 3.
iting the width to wi th in the pennis iblc level · pre ented in Some useful metric unit equivalents are presented below:
Table 2 i.J1 accordance wit h the pr va iling environme ntal c n- Customary Unit Metric Unit
dit ion., it wou ld be possible to preven t r con i.derably mi n-
1 in. 25.4 mm
imize long-term co rr sion deterioration and al main tai n th 1 ft 0.305 m
aesthetic behavior of the various elements of the system. 1 in. 2 645.16 mm 2
1 in. 3 16 387.06 mm 3
1. Reinforced-Concrete Beams and Thick One-Way Slabs 1 in. 4 416.231 mm4
1 psi 6.895 Pa
1 ksi 6.895 MPa
1 lb 4.448 N
lkip 4448 N
or 1 lb/ft 14.594 Ni m
1 kip/ft 14.594 kN/m
1 kig:in 113 N-m
1 VJ; psi 0.083036 Vf:. MPa

where fs is in ksi and z is not to exceed a value of 145 kip/in.

for exterior exposure or 175 kip/in . for interior exposure.
2. Prestressed, Pretensioned Beams TABLE 3 KV ALVES FOR FULLY RESTRAINED SLABS
a. Steel reinforcement level AN D PLATES
Slab/Plate Conditions" K
Wmax = (5.85 X lQ - 5 )~ (!:;.Js) Un ifo rmly loaded, square 2.8 x 10- 5
At cone ·ntrated loads and columns 2.1 x 10- 5
0.5 < 1,11, < 0.75 2.1 x 10 - 5
b. Tensile face of concrete IA< 0.5 0.6 x 10- 5
"For simply supported slabs multiply these values by 1.6. Interpolate K
w:nax (5.85 X 10 - 5 )R;;~ (!:;.fs) values for intermediate span ratios 1,111 or for partial restraints at the
boundaries such as cases of end and corner panels of multipanel floor
systems. I, and 11 are the short and long spans of the two-way slab or plate,
respectively .
3. Prestressed Post-Tensioned Beams
a. Steel reinforcement level

REFERENCES
b. Tensile face of concrete 1. 8uild/11g Code Req11irtme111 for Reinforced Concrete (ACI 318-
89) and 011111u:111ary (A l 3 L8R· 9) . American Concrete Insti-
tute, Detroit, Mich., 19 9, 89 pp.
2. E. G. Nawy. Reinforced Co11crete - A F11111/ame111al Approach ,
2nd ed. Prentice-Ha ll , Englewood liffs. .J., 1990. 73< pp.
3. A I o mm ittce 224. o ntrol of Cracking in Concrete tructures.
For nonbonded beams, the factor 6.51 becomes 6.83. AC! Joumal Procee<li1ig, Vol. 20, No. 10, Oct. 19 , pp. 35-
4. Two-way Action Structural Slabs and Plates 76.
4. P. Gergely. and L. A. Lutz. Maxim um rack Width in Rein-
Wmax = Kf'>f, ( G,) 112 fo rced o ncretc Flexural Mem()i.:rs. Cf111ses, Meclia11is111, 11111/
Co11 trol of Cmcki11g i11 Co11cret1J , SP-20, Ame rica n o ncrcte In-
stitute, De troit, Mich .• 1968, pp. 87- 117.
where
5. E. G. Nawy and K. W. Blair. Furt her Studies o n Flcxurnl Crack
Con trol in Structu ral Slab Sy tern . In Crncki11g, Defh!ctiou, 1111d
Ultim(l{e Load of Co11cre1e S/flb Sys1ems, SP-30, American Con-
crete lnstiwrc , Detroit , Mich . 1971. pp. 1-4'1.
32 TRANSPORTATION RESEARCH RECORD 1301

6. E.G . awy. Crack ontrol Through Rcinf rccment Distribution 11. E.G . Nawy. Flexural racking Behavior of Partially Prestres cd
in Two-Way Acting Slab. and Platc.s . A I Jo11mnl Proceetli11gs, Prctcntioned nnd Post-Ten ioncd Beam - tate-of-the-Art. In
Vo l. 69, No. 4, April 1972, pp. 217- 219. Cmcki11g iJ1 Pres1ressed Concrete trm;111res, American Concrete
7. E. G. Nawy. rack 01itrol in Beams Reinforced with Bundled I.nstiwtc , Detroit, Mich .. 1990, pp. 1-42.
Bars. AC! loum(I/ Pro eedings, Oct. 1972, pp . 637-639. 12. E. G. Nawy. Prr:stre sed 011r.re1e- A F1111damental Approach.
8. A. E. Naaman 11 nd A . Sirink om . rviccal>ility Ba cd Design Prentice-Hall, Englewood tiffs, N..1 . L9 9 739 p.
of Partially Pre tressed Beam • Part 1-Analysis. P .., Jouma/,
Yo.I. 24, o. 2, March- April 1979, pp. 64-89.
9. CEB-FIP. Model Code for Concrete Structures. Paris, 1978, pp.
1-347.
10. S. W. Meier and P. Gergely. Flexural Crack Width in Prestre · cd
Concrete Beams. Technical Note, ASCE Joumnl, ST2, Feb. 1981, Publication of this paper sponsored by Committee on Mechanical
pp. 429-433. Properties of Concrete.