Alexandria Engineering Journal (20 12) 51 , 2 11-220

Alexandria University

Alexandria Engineering Journal

www.elsevier.comflocatefaej
FACULTY OF ENGINEERING www .sciencedirect.com
ALEXANDRIA UNIVER~TY

ORIGINAL ARTICLE

Crack width evaluation for flexural RC members

Said M. Allam *, Mohie S . Shoukry, Gehad E. Rashad, Amal S. Hassan

Structural Engineering Department, Faculty of Engineering, Alexandria University, Alexandria, Egypt

Recei ved 18 February 20 12; revised 8 May 20 12; accepted 9 May 20 12

A vailable o nline 20 July 20 12

KEYWORDS Abstract Some building code equations and equations developed by resea rchers arc used for the
Flex ura l cracks; calculation of the crack width in reinforced concrete flexu ral members. To investigate codes' pro-
Crack width; visions beside some equations found in the literature concerning the crack wid th calculation of rein-
Cover; fo rced concrete members subjected to flexure, fi ve reinforced concrete rectangular models were
Serviceabili ty; investigated theoretically. The models incl ude different pa rameters such as reinforcement steel ratio,
Reinforced concrete; steel rcbar a rrangement a nd reinforcement grade. Also, to verify the accuracy of the building code
Codes provisions equations a nd the equations developed by researchers a comparison against some experimenta l da ta
available in the literature was ca rried out. The experimental data include some variables affecting
the crack wid th such as steel stress, concrete cover, flexura l reinforcement ratio a nd rcbar arrange-
ment. The study showed a large scatter among the di fferent code equations, however, most of the
code equations overestimate the effect of concrete cover o n t he calculated values of the crack width .
Also, the Egyptian code equation should limit the value of t he mean steel stress as given by Euro-
codc equation to overcome the underesti mated values obtained in the case of sections having low
steel ratio. Moreover, the reinforcement detailing (bars distribution) is an important factor affecting
the crack width .
© 20 12 Faculty of" Engineering, Alexandria University. Production and hosting by Elsevier B.V.
All rights reserved.

I . Introduction cracking is important for obtaining acceptable a ppeara nce

C rack width calculation is o ne of the servicea bility req uire-
ments in the structura l concrete elements. The occurrence of
cracks in re info rced concrete elements is expected under service
loads, due to the low tensile strength of concrete. Control of
* Corresponding author. Tel.: + 20 0 111 2244066.
E-mail address: Sa_allam@yahoo.com
Peer review under responsibili ty of Faculty of Engineering Alexand ria
University.
and for long-term durability of concrete structures, especially
those s ubjected to aggressive environmen ts. Excessive crack
width m ay reduce the service life of the structure by p ermitting
more rapid penetra tion of corrosive factors such as high
humidity, repea ted satura ti on with mo isture, vapor, salt-wa ter
spra y and gases with chemicals, to reach the reinforcement.
G ene ra lly, c racking should not induce reinforcement steel cor-
rosio n or spoil t he appeara nce of the structure. In additio n,
cracking in reinfo rced concre te structures has a n effect on
structural perfo rmance including stiffness, energy absorptio n,

~I
capacity, a nd d uctility. Consequently, there is an increased
interest in the control o f cracking by building codes a nd
ELSEVIER Production and hosting by Elsevier scienti fic o rgan izations. With the use of ultimate strength

1110-0168 © 2012 Faculty of Engineering, Alexandria University. Production and hosting by Elsevier B.V. All rights reserved.
http://dx.doi.org/ I 0.10 16/j.aej.20 12.05.00 I
2 12
methodo logy and high strength reinforcement steel, research-
ers a nd designers recognized the need fo r providing a mecha-
nism by which crack width would be minimized . Therefore,
researches were undertaken in 1960s to q ua ntify the a bove
concep t a nd develop design tools [I ].
The crack width of a flex ura l member is obtained by multi-
plying the maximum crack spacing by the mea n stra in of the
flex ura l steel reinforcement. Therefore, t he crack width de-
pends o n the nature and the arra ngement o f the reinfo rcing
steel crossing the cracks and the bo nd between the steel bars
fo und in the tension zone of concrete. Many research work
found in the literature predicted the crack width o f a flexura l
member based o n t heoretical models a nd experimenta l d ata.
Saliger [2] and T omas [3] used Bond-Slip model, Borms [4]
and Base et al. [5] used No-Slip model, however Welch and
Janjua [6] a nd Leonhardt [7] used L oca lized Bo nd-Slip mod el
to predict the crack width. G ergely and Lutz [8] used the re-
sults of experimental d ata to fo rmulate an equa tion to calcu-
late the crack wid th. Based o n the experimental wo rk, O h
and K ang [9] p roposed a fo rmula tion fo r predicting the max-
imum crack width. Frosch [10] developed a simp le theoretical
equa tio n to predict the crack width based on a physica l mod el.
Besides the research work carried o ut for crack width fo rmula-
tio n, other resea rch work experimenta lly investiga ted the fac-
tors a ffecting t he crack width . M a khlo uf and M alhas [II]
investiga ted the effect o f t hick concrete cover on the maximum
flex ura l crack wid th under service load . Beeby [12] and Nawy
and Blair [ 13] showed tha t the transverse reinfo rcement had a
strong influence o n the crack spacing. G ilbert and Nejadi [14]
tested six beams and six o ne-way slabs with different flexura l
reinforcement ratio and bar a rra ngement including va rious
concrete cover.
In this paper, a stud y is carried o ut to investiga te severa l
fo rmulas suggested by different building codes fo r the calcula-
tion o f the crack width in reinfo rced concrete flexural mem-
bers. Also, the formulas proposed by other researchers were
also in vestigated a nd compared with codes' equations. In add i-
tion, the most prevalen t building codes' equa tio ns are exam-
ined and tested against some experimental da ta availa ble in
the literature. M oreover, a comparison was carried out among
the equatio ns to discuss the va rio us factors and parameters
affecting the crack width.
2. C rack width prediction according to some building codes
provisions
2.1. Eurocode2 1992-1 (2001 )
E urocode2 [15] gives the fo llowing equatio n fo r predicting the
crack width o f flexu ral members
( I)
where Wk = the design crack width, mm.
The mea n tensile strain F.sm - F.cm) is given by the followi ng
eq ua tion:
(F.sm- f.cm) =
(!;_K, ~'""<;,~;'P£frl)) ;;, 0.6-
f,
(2)
£, E,
where esm is the mea n strain in the reinforcement under the rel-
eva nt com binatio n of loads, including t he effect o f imposed
deformations a nd taking into acco unt the effects o f tensioning
S.M . A llam et a l.
stiffening. Only the addition al tensile strain beyond zero strain
in the concrete is considered; F-cm is the mea n strain in concrete
between cracks. K, = facto r expressing the duration of load-
ing: K, = 0.6 for short term loading and K, = 0.4 fo r long
term load ing,]; = the stress in the tensio n reinforcement com-
puted on the basis of a cracked section, n = t he modular ratio
-/!:;;,, f~terr = the mea n val ue o f tensile strength of the concrete
effecti ve at the time when the cracks may first be expected to
occur,
A,
Pcrr=--
A ccff
Acerr = effective tension a rea, is the a rea of concrete surrou nd-
ing the tensio n reinfo rcement. S r,max = t he maximum crack
spacing, mm and is given by the fo llowi ng equatio n
(3)
where c = concrete clear cover, k 1 = coefficient tha t takes
into account the bond properties of t he bonded reinforcement
and eq uals to 0.8 for high bond reinforcing bars, a nd eq ua ls to
1.6 for plain reinforcing bars, 4> = the bar d iameter, mm; in
case o f using various d iameters, the average d iameter sha ll
be used, k 2 = coefficient t ha t takes into account t he strain dis-
tribution and is equa l to 0.5 for sections su bjected to p ure
bending and eq uals to 1.0 fo r sections subjected to pu re axial
tension.
2.2. Egyptian code; ECP 203-2007
T he Egyptian code ECP 203-2007 [ 16] gives the crack width by
the fo llowing equation:
(4)
where Wk = coefficient for checking crack width conditio n,
mm; Srm = average sta bilized crack spacing, mm; F.sm = mean
steel strain under releva nt combination of loads and a llowing
for the effect such as tension stiffening or shrinkage; fJ = coef-
ficient relating the average crack wid th to the design value:
fJ = 1.7 for cracks ind uced by loading and fo r cracking in-
d uced by restraining t he deformation for cro ss section having
width or depth (whichever smaller) in excess of 800 mm, and
fJ = 1.3 fo r cracking ind uced by restraining the deformation
fo r cross section having width or depth (whichever sma ller) less
tha n 300 mm.
The mea n steel strain ~>sm is given as
2
f.
sm E.t (1 -{J{J (/scr2)
= /,
1 2 ~
) (5)
where/, = stress in the tensio n reinfo rcement calculated on
the basis of a cracked sectio n, N jmm 2 ;/,.0 , 2 = stress in the ten-
sio n lo ngit udinal reinforcement computed o n the basis o f a
cracked sectio n under loadi ng conditio ns that cause the first
crack, N jmm2 ; {1 1 = a coefficient accounting fo r bar bond
characteristics, and is equal to 0.8 for deformed bars and 0.5
fo r plain smooth bars, {12 = a coefficien t accounting for load
dura tion; is equa l to 1.0 for single short-term loading and
0.5 fo r susta ined or cyclic loading; £ , = Mod ulus o f elasticity
o f t he reinforcement, N jmm 2 .
The Egyptia n code gives the average sta bilized mea n crack
spacing by the following equation:
(6)
Crack width evaluation for flexural RC members
However, the items of the above equation are simila r to
the items of the equa tion given by the Eurocode2 [1 5]
with the assumption that the clear concrete cover is equal to
25mm.
2.3. ACI 3 18
Prio r to 1999, flexural crack con trol requirements in AC T were
based o n the so-called z-factor method developed be Gergely
a nd Lutz [8]. Their work was based on extensive sta tistica l
a na lysis techniques o n experimental data from several
resea rc he rs. The equation proposed by the early version of
ACT 3 18-95 (1 7] took the following form:
(7)
where fJ = ~::; is t he ratio of distance between ne utral axis
and extreme tensio n face to dista nce between neutral axis
and cent roid of reinforcing steel; fJ = 1.20 in beams may
be used to compare the crack widths obtained in flexure
and axia l tensio n. Ao = the area of concrete su rrounding
each rein fo rcing ba r = Aefnh, Ae = the effective a rea of con-
crete in tension, Ae can be defined as the area of concrete
having the full width o f the beam a nd ha ving the sa me cen-
troid of t he main reinforcement; Ae = 2 deb, nh = the num-
ber of tension reinforci ng bars. de = the distance measured
from the centroid of tensile steel to the extreme tensioned
fiber
The flex ural crack width expression in the a bove equation,
with ~::; = 1.2 , is used in ACT 318-95 in the following form:
(8)
A maximum value of z = 3064.5 N fmm is permitted for inte-
rior ex posure, corresponding to a limi ting crack width of
0.41 mm. ACT 3 18-95 a lso limits the value of z to 2539.2 f
mm for ex te rior exposure, corresponding to a crack width o f
0.33 mm. When structures are subjected to very aggressive
exposure o r designed to be watertigh t, ACT committee 350
(18], limits the va lue of z to 1700 N jmm corresponding to a
crack width of 0.20 mm.
In the 1999 edition, ACT decided to greatly simplify crack
contro l requirements due to increased evidence suggesting are-
duced co rrelation between c rack width and reinforcement cor-
rosion . Beeby [ 19] and [20] showed tha t corrosion does no t
co rrelate with surface c rack widths in the ra nge normally
fo und with reinforcement stresses at service load levels. There-
fo re, AC T introd uced changes to the crack rules in which a
maximum bar spacing, rather than a z-factor (rela ted to crack
width) is prescribed.
ACT 318-05 [2 1], ACI 31808 [22] proposed the fo llowing
equatio n for crack control:
s = 380(280/J.;) - 2.5c::; 300(280/.J:) (9)
where s = maximum spacing of reinfo rcement closest to the
tension face, mm; c = least distance from surface of reinforce-
ment to tension face, mm.
H owever the equation does not make a distinctio n between
inte rior a nd exterior exposure, i.e. the expos ure condi tions
dependence was eliminated. A lso, the equation is indirectly
tied to a crack width eq ua ls to 0.4 mm. The value of f.~ a t ser-
vice load sha ll be computed o n the basis of service moment.
ACT permits the use off, = 0.67 J;,.
213
2.4. British standards BS 8/10- / 997
According to BS 8 110- 1997 [23], the width of flex ural cracks at
a particular point on the surface of a member depends primar-
ily on three factors:
(a) The proximity to the point considered of reinforcing
bars perpendicular to the cracks; concrete cover.
(b) The proximity of the ne utral axis to the point consid-
ered; h- x .
(c) The average su rface stra in at the point considered.
BS 8110- 1997 recommends that the strain in the tension
reinforcement is lim ited to 0.8 J~/ E., (i.e. 0.8 *steel yield strai n)
and the design surface c rack width should not exceed the
appropriate va lue. Cracking should not lead to spoil appear-
ance. So for members that are visible, the calculated maximum
crack width should not exceed 0.30 mm. Also, cracking should
not lead to steel corrosion, so for mem bers in aggressive envi-
ronment the calc ulated max imum crack width should not lead
to a loss of the performance of the structure.
BS 8 110-97 provisions are based on Beeby [12] empirical
equations,
3 * acr *F.,
Design surface crack wid th W" =
1
+ 2 (~) ( I0)
h- x
where acr = distance from the point considered to the surface
of the nearest longitudinal bar; 8111 = a verage strain at t he level
where the cracking is being considered; Cmin = m inimu m cover
to the tension steel; h = overall depth of the member;
x = depth of neutral axis.
For cracked section, the value of e111 is expressed as:
b(h- x )(d - x )
e, = e1 - 3£ .fA.f (d - x) ( II )
where e 1 = strain at the level considered, calcula ted ignoring
the stiffening effect of the concrete in the te nsion zone,
b = width of the section at the centroid of the tension steel,
d = distance from the compression face to the point a t which
the crack width is being calculated.
Accord ing to BS 8 1 I 0- 1997 [23], in assessi ng the strains, the
modulus of elasticity of the concrete should be taken as half
the instantaneous va lues.
3. Some significant formulas for crack width given by
researchers
3.1. Gergely and Lutz
Gergely and Lutz [8) used test resu lts from H ognestad [24],
Kaar and Mattock [25], Kaa r a nd Hognestad [26], C lark
[27], and Rusch and Rehm [28], to conclude their equation
for the calculation of crack widths at the tension surface.
As stated by Frosch [10], the maximum concrete cover; de
used in the tests a na lyzed by Gergely a nd Lutz (8] was
84 mm and only three specime ns of 6 12 o bservations had clear
covers greater than 64 mm .
The o rigina l equatio n developed by Gergely a nd Lutz [8] is
as follows:
W, = 0.0 11 y'(c + 4J/ 2)A0 (f; - 34.45) * 10- 3 ( 12)
214 S.M. Allam et a l.

where W, = the most probable maximum crack width a t level 3.3. Frosch
of steel, mm; f-. = the reinforcement steel stress, N f mm2 ;
A 0 = the area of concrete symmetric with reinforcing steel di- Frosch [10] stated that Eq. (7) proposed by ACI-95 [17] is valid
vided by number of bars, mm2 . for a relatively na rrow ra nge of concrete covers (i.e. up to
To obtai n the maxi mum crack width a t the extreme ten- 63 mm). The use of thicker concrete covers is increasing be-
sioned fiber, Eq. ( 12) is multiplied by a factor II= ::::. ca use research and experience ha ve indicated tha t the use of
thicker covers can increase durability. Therefo re, Frosch [ I 0]
3.2. Oh and Kang developed the following simple, theoretically-derived eq uation
to predict crack widths that could be used regard less the actual
Oh and Kang [9] proposed a formula tio n for predicting the concrete cover.
maximum crack width; W max and average crack spacing; S
in reinforced concrete flexural members. They tested five ( 19)
reinforced concrete beams with design variables including the
concrete cover, diameter of steel bars, reinforcement ra tios, where tr is the controlling cover distance and is taken the
spacing of steel bars and steel stress. Based on an energy
approach, Oh and Kang concluded a new definitio n of
the effecti ve area of concrete in tension and suggested the fol-
lowing equation for the calculation of the maxim um crack
greater of { F },
y~ +G)2
II= 11•::x and d, is the side cover.
<X

width:
4.Comparison between some code equations and other available
W max = ¢ao(r., - 0.0002)R ( 13) equations for predicting the crack width of some theoretical
In which, models

ao = 159 (--t
h-) 4.5
+ 2.83
(A- ) 0 1/ 3
( 14)
In this section, the above mentio ned building codes and some
h -x Av~ equations given by researchers a re applied to five reinforced
concrete sections having different values of reinforcement ra-
R =(h- x)/(d- x) ( 15) tio, bar distribution, and ba rs grade.
Fig. I shows the layout of the five models. The Models I , 2,
where A 0 is the effecti ve area of tensioned concrete around
and 3 ha ve the same reinforcement (strength; /y, type: de-
each reinforcing bar; Ao = Ae/nh and Ae = b h 1, h 1 is the depth
formed bars, number a nd diameter), section width; b, concrete
of eq uiva lent a rea,
strength; feu· These models va ried in section depth; d and con-
hi = (h- x)J sequently have va riable reinforcement ratios; fL. Val ues of J.L
( 16) were: 0.565%, 0.87%, and 1.21 % fo r the three model respec-
3(d - x) 2
ti vely. M odels 2 and 4 have identical properties (same steel
reinforcement, same section dimensions) and they varied in
o.236*1o-6 )
Sm = ¢ ( Co + r.; ( 17) ba r surface conditio n and yield stress of steel. Pla in mild steel
bars were used fo r M odel 4, while high strength deformed bars
In which, were used for M odel 2. Models 2 and 5 have the same section
dimensions, approximately, same area of tension steel but var-
(A o) 1/ 3
c0 = 25.7(h _
lh
J 4.5
+ 1.66 A.,l ( 18)
ied in bar distribution (number and d iameter). T he val ues of
the crack width were calculated at the same steel stress level

--

r-
E
E F E F EE
E - E - "1 E
"' E
'DO
E E "'
'DOE t-o
..;-o E E ..;-o V"lo
~0
V) V) "<tV)

("') t-
("')

~~16 E 2 # 25 -~
i i ~ ~ gz~-
1
• 250 mm - I I· 250 mm. l I· 250 mm -I I• 250 mm - I 1· 250mm -I "<~"
model! model2 model3 model 4 model 5
f..l =0.565% f..l = 0.87% f..l = 1.21% J..l =0.87% f..l = 0.858%

Figure I Dimensions and reinforcement of the studied models.

Crack wid th evaluation for flexu ra l RC members
Table I Va lues of the crack width for the stud ied models.
Eq uation Crack width (mm)
Model I Model 2
Eurocode2 0.1 27 0.1 27
ECP-2007 0.1 07 0.1 29
ACl 3 18-95 0.1 24 0.1 29
BS 81 10-97 0.079 0.087
Gergely a nd Lutz [8] 0.1 02 0.1 07
F rosch [10] 0.1 01 0.1 06
Oh a nd Kang [9] 0.1 47 0.1 35
(j, = 200 N jmm 2 for models I , 2, 3 and 5 andj~ = 140 N jmm2
fo r model 4). Concrete strength was considered as feu = 30 N /
mm 2 a nd the clea r cover = 30 mm.
Table I shows the values of crack widths calcula ted fo r t he
studied models for bo th o f code equatio ns a nd equations given
by Gergely a nd Lu tz [8], F rosch [10] and Oh and K a ng [9]. T he
fo llowing was o bserved concerning the results given in the Ta-
ble 1:
1. Genera lly, values of crack width predicted by the code
eq uatio ns showed a large scatter among the d ifferent
code equations and those o btained by previo us resea rch-
ers. Values of crack width, predicted using BS 8 110-97,
were less tha n those predicted by other bui lding codes
o r by equations developed by researchers, especia lly
for M odel 4, reinforced with plain mild steel bars. Val-
ues o f the crack wid th predicted by the Egyptian code
were very close to t hose predicted by ACT 3 18-95 and
E urocode2 equatio ns except fo r model I with low rein-
forcement ra tio; ft. T he Egyptian code equatio n pred icts
values o f crack width fo r sectio ns with low reinfo rce-
ment ratio smaller than those pred icted by ACT or Euro-
code2. Also, the ta ble indica tes that the equatio n of O h
and Kang [9] overestima ted the va lues of the crack
widths in the case of low steel ratio (M odel 1), however
ECP-2007 overestima ted the va lue of the crack wid th in
the case of plain bars o f model 4.
11. Comparing the results of the crack widths calculated for
the models I ,2 a nd 3 ind icated that as the percentage of
the flexura l steel ratio, Jl, increased the calculated values
of the crack width increased. T his is given by code eq ua-
tions of ECP-2007, ACI-3 18-95 a nd other equations
given by Gergely and Lutz [8] a nd Frosch [10]. W ith
the increase o f Jl, the concrete contribution in tension
decreases and the mean steel strain increases, and then
the crack width consequently increases. H owever,
according to Eurocode2, the increase of J1. d oes not affect
the crack width. O n contrary, the eq uation o f Oh and
K ang [9] is found to give low va lues of crack wid th with
the increase o f t he flex ural steel ra tio, Jl.
111. T he effect of using different numbers of steel ba rs with
eq ual area on the crack width is investigated by compar-
ing the results o f models 2 and 5. It is well esta blished
tha t the bar d istributio n is an important factor affecting
crack wid th. Wi th well cho ice of bar arrangement (la rger
number, smaller diameter), better bond between con-
crete a nd steel occurs and thus reducing the crack spac-
ing. T he reduction in crack wid th due to better bar
215
Model 3 Model 4 Model 5
0. 127 0.1 07 0. 16
0. 137 0. 123 0. 170
0. 136 0.1 19 0. 19 1
0 090 0.056 0. 162
0. 11 3 0.0682 0. 158
0. 111 0.0742 0.23 1
0. 13 1 0.0845 0.2 18
distribution ranged from 26% to 118%, the lowest value
was obta ined using the E urocode2 eq uation and t he
highest va lue was obtained according to Frosch [10]
equatio n. T he equation developed by Frosch [1 0] ind i-
cates tha t crack spacing (and consequen tly crack width)
depends ma inly o n the d istance between reinfo rcement
bars.
iv. T he effect of bar surface cond ition (pla in or deformed)
o n crack width was obtained by comparing the results
of models 2 and 4. The use o f plain mild steel bars (with
low steel stress at service loads) results values o f crack
width less by about 5- 56% compared with model rein-
fo rced with the same area o f high strength deformed
bars. T he lowest value was obtained using the ECP-
2007 eq ua ti on a nd the highest va lue was obtained by
Oh and K ang equatio n. The effect of bar surface defor-
matio n on the calculation of crack width no t only affects
the crack spacing, but a lso affects the mea n strain. W hen
the bond between the concrete a nd the steel increased,
more tension force is transferred to the concrete between
cracks. With the increase in the concrete contribution in
tension, less slip between concrete and steel occurs,
hence less va lue of total elo ngatio n between them
(Bsm - Bcm); and consequently resulting in less crack
width. However, the effect of bar surface deformation
is not considered in code fo rm ulas except the Egyptia n
code and Eurocode2.
5. Application of some code equations to some availa ble
experimental data
To asses the accuracy o f code equations and the most common
formulas fo r predicting crack width in reinfo rced concrete
members, a comparison is carried o ut with the experimental re-
sults reported by Makhlouf and Malhas [II ]. T he comparison
includes two factors a ffecting the crack width such as t he steel
stress and concrete cover.
5.1. Makhlouf and Malhas
Makhlou f and M alhas [ II] repo rted results of tests on I I
beams reinfo rced with high strength deformed bars. T he max-
im um crack width was recorded at the reinforcement level. T he
specimens consisted of two groups. Group " A" consisted of
eight beams fo ur of them were with 22 mm concrete clear cover
and the other fo ur were with 52 mm concrete clear cover.
Gro up "A" mainly intended to investigate the effect o f
216 S.M. A llam et a l.

p p
500 mm
,-2 br. ¢ 8mm@

:i n~~0
I SOmm
1 2 - 10 2 .• 10

mm : c = 22 mm
1------'
~lU 460 mm: c =52 mm

..t!!!!!!l!.. ..t!!!!!!l!.. 180 mm

1------1
3200 mm

Group A

p p
500 mm

4br.¢ 8mm@ 1657 r 4 •1o 4 - 10

(__

<!!!!!!!!!!l!l. <!!!!!!!!!!l!l.
1900 mm

Group B

Figure 2 Beam specimens tested by Makhlouf and Malhas [II].

Table 2 Properties of specimens tested by Makhlouf and Malhas [I I].

Specimen h h d Reinforcement Rei nforcement fc' (N/ fy (N/ c
no. (mm) (mm) (mm) steel ratio; J1 (%) mm 2) mm 2 ) (mm)
Group GA ll 180 430 400 20 12 0.31 34 425 22
" A"
GA12 180 430 400 2016 0.56 34 425 22
GA1 3 180 430 400 2020 0.87 34 425 22
GAM 180 430 400 2025 1.36 34 425 22
GA21 180 460 400 2012 0.3 1 34 425 52
GA22 180 460 400 20 16 0.56 34 425 52
GA23 180 460 400 2020 0.87 34 425 52
GA24 180 460 400 2025 1.36 34 425 52
Group Gl 600 400 338 4014 + 2020 0.62 40 430 50
" .B"
G2 600 400 336 6020 0.94 40 430 50
G3 600 400 336 9020 1.40 40 430 50

doubling the concrete clear cover o n the crack width.
Group " B" consisted o f 3 wide beams o f 600 mm width,
400 mm depth a nd clea r concrete cover = 50 mm. The expe r-
imenta l specimens were tested under a load level of 80- 110%
o f service load. Group " B" aimed to assess the m agnitude o f
crack widths in full-size bea ms under differe nt levels o f steel
stress.
F ig. 2 shows the dimensions and loading system o f the
tested bea ms while Table 2 gives prope rties o f a ll specimens
tested by Makhlouf a nd Malhas [ II] for both Group " A"
a nd Group " B" .
Measurements were taken o n both sides of the beam spec-
imens a nd at 10 mm above the lower edge. R ead ings we re ta-
ken at predetermined load levels tha t corresponded to a
specific safety factor rela tive to the u ltima te failure load .
5. 1.1. Effect of steel stress,/,
The main a nd the most important factor affecting the crack
width is the steel strain (st ress) whic h is directly p roportional
with the crack wid t h. Makhlouf a nd Malhas [ II ] measured
the crack width o f t he specimens of Group " B" a t different lev-
els o f steel stress, as given in Table 3. F igs. 3- 5 d isplay the rela-
tionship between the crack width; W max' as calculated using
codes' equa tions, and the steel stress; J,, for the group " B"
specimens: G I, G2, and G3 respectively. A summary of there-
sults obtained from code' s equatio ns is given in Table 3. F rom
the Figures a nd Ta ble t he fo llowing are observed:
1. The Egyptian code gives underestimated values of crack
width fo r sections reinforced with lo w reinfo rcement
ra tios (ft = 0.62%), especially at lo w le vels of steel stress
Crack width evaluation for flexu ral RC members 21 7

(f; = 0.30 /y and j; = 0.37 j~). According to the Egyp -

tian code, at steel stress f-. = 0.30 J;, (specimen G I), the
section is considered uncracked . H owever, the va lues
of crack width as proposed by the Egyptian code corre-
la ted well with the experimental values, at all levels of
steel stress, for sections having percentage of reinforce-
oo - o-. -
o - - ("'")
0\V)Nr-
0 - N N
V C0\0 -.:::t ment (Jl = 0.94%). On the other hand, for sections ha v-
-- N M
0000 0000 0000 ing higher percentage of reinforcement (Jl = 1.4% ) the
Egyptian code calculated crack width is lower than the
crack width was meas ured experimentally. T hat may
imply that a maximum reinforcement ratio for crack
V)Q\\0("') Mr-NO\ control should be determined.
-- N M -- N N
0000 0000 11. The underestimated values of crack wid th did not

appear in E urocode2 predictions at low levels of steel

stress, since Eurocode2 limits the va lue of r., no t to be
0 v 0 00 oo - r-rr~
less than 0.6 JJE,, as appears from Table 3.
N N M - N N M
0000
("<")

0000 iii. Ta ble 3 gives the va lues of steel strain e., together with
the values of the average strain (r.sm or s,). The results,
o btained from both the Egyptian code and BS 8 1 I 0,
0 ("") r- N 0N \0 -
--- N --- N indicate that with the increase of reinforcement ratio
0000 0000
Jl, the estimated concrete contribution in tension (ten-
"'C>
I
sion stiffening) decreased. In general, the limitation of
(""") co - . 0\. average strain introduced by E urcode2 predicts realistic
.. 0 N
\Or- - - values of crack width than those predicted by the Egyp-
tian code. In addition, with increase of the steel stress,
\00\VO
the contribution of tensioned concrete decreased.
- - N ("")
0000 I V. Generally, the ACI 318-95 equation grea tly overesti-

"'I mated the val ues of the crack width, except at high level
C>
of steel stress and at high values of reinfo rcement ratio
(Jl = 1.4%). It should be noted that ACI 3 18-05
switched from the procedure of calculating the crack
"'I width and adopts a simplified equation for the maxi-
C>

X ~ 6~ ~ mum bar spacing. For values of steel stress 134-284 Nf

J I I mm2 (which are the levels of stress at which crack width
E' were meas ured experimentally), equation 9 of ACI 318-
s N N <'> -\0 V N NV)Ot-
05 yields maximum permitted bar spacing of 625-
:; C>- N
1 000
-- N
0000
M -- N
0000
N
250 mm, for concrete clear cover = 50 mm.
"'I v. The values of crack width predicted using BS 8 11 0-97
C> equations were smaller than the experimental values a t
X N <'> high levels of steel stress. As given in Table 3, high val-
E 0\ r- . ("')
.
.. 0
Of t.nr--- ues of mean strain s, were proposed using BS 8 11 0-97
and this indicates small contribution of tensioned
concrete.
VI. The Oh-Ka ng [9] formula (Eq . ( 13)) correlated well with
N N N N
0000 the experimenta l va lues of crack width fo r most speci-
00 00 00 00
mens used in the compa rison.

5.1.2. Effect of concrete cover, ( c)

N crack width was obtained as a result of increasing the concrete
0 G
T he value of concrete cover may be considered as the second
important factor that affecting the crack wid th, but its effi-
ciency is considered in different ways in the building codes.
Ma khlouf and Malhas [I I ] reported results of tests on eight
beams reinforced with high strength deformed bars as given
by F ig. 2 and Ta ble 2, Group "A" . The maximum crack width
was recorded a t the reinforcement level. Four of the beams
were with 22 mm concrete clear cover and the other fo ur were
with 52 mm concrete clear cover. The a im of these tests was to
investigate the effect of increasing the concrete clear cover on
the crack width. The results showed that a 16% increase in the
cover.
2 18 S.M . A llam et a l.

0.40

- ... - ECP-2007
0.35

eE 0.30
--o-- EuroCode-
2001
0.25 - • - ACI 318-95
~
'0
;: 0.20
- x- Oh and Kan
.:.:: [9]
(.) 0.15
~ - Experimental
u
0.10
--B.S.8110-97
0.05

0.00
100 150 200 250 300
Steel Stress (N/mm2 )

Figure 3 Crack width versus steel stress for G I (11 = 0.62%) Makhlouf and M alhas [I I].

0.40

0.35 _,.
.. - ... - ECP-2007

eE 0.30 --<>-- EuroCode-

2001
0.25
~ - • - ACI318-95
'0
;: 0.20
- - Oh and Kang
.:.:: [9]
(.)
ell 0.15
._
u - Experimantal
0.10
--B.S.8110-97
0.05

0.00
100 150 200 250 300
Steel Stress (N/mm2)

Figure 4 Crack width versus steel stress for G2 (11 = 0.94%) Makhlouf and Malhas [I I] .

0.40

0.35 - -+- - ECP-2007

eE 0.30 --o-- EuroCode-

2001
~ 0.25
- • - ACI 318-95
'0
;: 0 .20
.:.:: - x- Oh and Kang
(.)
0.15 [9]
._
ell
u 0.10 ............ Experimental

0 .05 --B.S. 8110-97

0 .00
100 150 200 250 300
2
Steel Stress (N/mm )

Figure 5 Crack width versus steel stress for G3 (11 = 1.40%) Makhlouf and M alhas [I I].

T a ble 4 gives a nd compares the va lues o f t he crack width the eq uation de veloped by O h a nd K ang [9]. According to
calculated using the equations p roposed by the building codes Oh a nd K ang [9], an increase of t he crack width by about
fo r t he bea ms tested by M akhlo uf a nd Malhas [II ] a nd a lso, 20% is obtained with increase of the concrete cover from
C rack width eval ua tion for flexural RC members 219

Table 4 Application of codes' eq uatio ns to bea ms tested by Makhlouf and Malhas [I I] - Group A.
Egyptian code 2007" Eurocode2-200 Ih BS 81 10-97 ACI 318-95 Oh and Kang [9]
J;cr2 (N/mm2) l:smX 10 4
wk (mm) t:,.X 10 4
~:;,, X 10- 4 wk (mm) W"(mm) Wmax (mm) Wmax (mm)
GAl l 251.2 2.9 0.06 6.2 7.7 0.16 0.12 0.17 0.19
GA2 1 287.44 0.5 7.7 0.33 0.14 0.29 0.23
110.0% 20.0% 72.4% 17.9%
GAI2 144.5 9.5 0.17 8.6 8.6 0.15 0.14 0.17 0.2 1
GA22 165.36 8.5 0.23 5.5 7.7 0.27 0.19 0.29 0.25
35.5% 84.5% 32.1 % 73.3% 19.2%
GAI 3 94.4 11 .4 0.18 9.8 9.8 0.14 0.15 0.17 0.22
GA23 108.1 10.9 0.25 7.8 7.8 0.26 0.20 0.30 0.27
40.6% 79.1% 38.5% 74.3% 20.6%
GAI4 61.9 12. 1 0.1 7 10.5 10.5 0.13 0.14 0.17 0.24
GA24 70.9 12.0 0.24 9.2 9.2 0.27 0.20 0.30 0.29
38.6% 107.8% 45.4% 75.6% 22.4%
Average increase in 38.3% 95.4% 34.0% 73.9% 20.0
crack width
2 4
J,= 255.0 N/mm , 1:-' = J,/£.,. = 12.75 X I0 Uncracked section according to the Egyptian code. Bold values refer to the percentage of the
increase of the crack width value in each two specimens, however italic values refer to the average increase in the crack wid th for all specimens in
each column of the table.

22 mm to 52 mm. This a mount of increase is very close to t he
value obtained experimentally by Makhlouf and Malhas (II].
H owever, such increase in the values of crack wid th as pre-
dicted by the building codes was greatly overestimated as com-
pared to the experimental results. The maximum p ercentage of
increase was recorded by Eurocode2 followed by ACI 3 18-95
while the Egyptia n code 2007 a nd BS 8 11 0-97 give reasonable
pred ictio ns.
6. Conclusions
From the results obtained from the compa rison of several
building codes' equa tions a nd from the theoretical work pre-
sented in this paper through studying the ma in variables a ffect-
ing crack width, s uch as the concrete cover, steel stress,
reinforcement ratio, bar surface, and reinforcement a rra nge-
ment, the following conclusions may be drawn:
I. Values of crack wid th predicted by the code eq uations
showed a large scatter among the different code eq uations
a nd those obtained by other researchers. The Egyptian cod e
equation predicted va lues very close to those predicted by
ACI 3 18-95 a nd E urocode2 equations except sections with
low reinforcement ratio. For this case, the Egyptia n code
equation predicts va lues of crack widt h smaller tha n those
p redicted by ACI or Eurocode2.
2. Values of crack width, predicted using BS 8 11 0-97, were
less than those predicted by other bu ilding codes or by
equations developed by researchers, especially for sections
reinforced wi th plain mild steel bars. It should be noted tha t
BS 8 11 0 gives rules for bar detailing to ensure control of
cracking rather than giving detailed calculations for crack
width a nd crack sp acing.
3. W ith the increase of reinforcement ra tio Jl, the concrete
contribution in tension decreases, the mea n steel strain
increases, conseq uently the crack width increases. However,
for crack control, it is suggested to limit the reinforcement
ra tio instead of limiting the steel stress.
4. The reinforcement detailing (i.e. the bars distribution) is an
important factor affecting crack width. With the well choice
o f ba r arrangemen t (larger number, smaller diameter), bet-
ter bond between concrete and steel occurs res ulting in a
reductio n in the crack spacing. In addition to BS-81 10 rec-
ommendatio ns for crack control, ACI 3 18-05 switched fo r
the procedu re o f calculating the crack width a nd adopts
simplified equation fo r the maximum bar spacing.
5. The effect of bar surface deformation on the calc ulation of
crack width not only affects the crack spacing, but a lso
affects t he mean strain . When the bond between t he con-
crete and the steel increased, more tension force is t rans-
ferred to the concrete between cracks. With t he increase
in the concrete contribution in tension, less slip between
conc rete a nd steel occurs, hence less val ue of tota l elonga-
tion between them; (i.e. F.sm -<-em); a nd consequently result-
ing in less crack width . H owever, t he effect of bar surface
deformation is not considered in most codes' formulas
except t he Egyptia n cod e and E urocode2.
6. Most equa tions proposed by the building codes overesti-
mate the effect of the concrete cover on the calculated val-
ues of crack width when compared with the exp erimenta l
results. H owever, O h and Kang [9] formula, (Eq. ( 13), cor-
related well with the experimental val ues of crack width for
most specimens used in the comparison.
7. The limitation of the average strain <., not to be less than
0.6 JJ E, introduced by Eurocode2 results in rea listic va lues
of crack width when compared wit h experimental results
tha n those predicted by the Egyptia n code a t low levels of
steel stress for sections reinforced with low reinforcement
ratios.
8. Comparison o f building codes against experimenta l results
revea led that the Egyptian code gives underestima ted val-
ues of crack wid t h for members reinforced with low rein-
forcement ratios especially a t low level of steel stresses.
T he ECP 203-2007 adopted the use of mean strain simila r
to the eq ua tion given by Leo nhardt [7]. The Egyptian code
eq uatio n expresses the tension stiffening as a functio n of the
220
steel stress. For the calculation of the steel stress just after
the first cracking, the Egyptian code assumes that the force
resisted by concrete in tension is completely neglected
immediately after the occurrence of the first crack. This
assumption produces a very large va lue of steel stress j ust
after cracking <!sed for members having low reinforcement
ratios especially, at low levels of steel stress. Values of.fscr2
higher than f, indicate uncracked sections. It should be
noted that by definition .fscr2 should be less than f,. There-
fore, for sections with low percentage of reinforcement, it
is recommended, when using Eq. (5) of the ECP 203-
2007, to limit the va lue .fscr2 to f, (the steel stress under
working load conditions).
References
[I] A.S. Hassa n, Crack Control for Reinforced Concrete Members
Subjected to Flexure, M.Sc. Thesis, Alexa ndria U niversity,
Alexandria, Egypt, 2008, 119pp.
[2] R. Saliger, H igh grade steel in reinforcement concrete, 2nd
Congress, In ternational Association for Bridges and Structural
Engineering, Preliminary Publication, Berlin-Munic h, 1936, pp.
293 315 (cited by MacGregor et at. (1980).
[3] F.G. Tomas, Crac king in reinforced concrete, Structural
Engineer 14 (7) (1936) 298 320 (cited by MacGregor et at.
(1980).
[4] B.B. Broms, Crack width and crack spacing in reinfo rced
concrete members, Journal of American Concrete Institute 62
(10) (1965) 1237 1256.
[5] G.D. Base, J.B. Read , A.W. Beeby, H.P.J . Taylor, An
Investigatio n of the Crack Cont ro l Characteristics of Va rio us
Types of Bar in Rein forced Concrete Beams, Research Report
No. 18, Parts l a nd IT, Cement a nd Concrete Associatio n,
London, 1966 (cited by Nawy ( 1968)).
[6] G. B. Welch, M.A. Ja njua, Width a nd Spacing o f Tensile C racks
in Reinforced Concrete, UNTCIV Report No. R 76, University
of NSW, Kensington, 197 1 (ci ted by Warner and Rangan
( 1977)).
[7] F. Leonhardt, Crack Control in Concrete Structures, TABSE
Surveys, No. S4/77, Tnte rnationa l Association fo r Bridges and
Structural Engineering, Zurich, 1977, 26 pp (cited by Rizkalla
(1984)).
[8] P. Gergely, L.A. Lutz, Maxi mum C rack Width in RC Flexural
Members, Causes, Mechanism and Control of Cracking in
Concrete, SP-20, American Concrete Institute, Detroit, 1968,
pp. 87 11 7.
[9] B.h. Oh, Y.J. K ang, New formulas for maximum crack width
and crack spacing in reinforced concrete flexural members, ACT
Structural Journal 84 (2) ( 1987) I03 112.
[ 10] R.J. Frosch, A nothe r look a t crac king and crack control in
reinforced concrete, ACT Structural Journa l 96 (3) ( 1999) 437
442.
[ II] M.H. Makhlouf, F .A . Mal has, The effect o f thick concrete cover
on the maximum flexural crack width under service load, ACI
Structural Journa l 93 (3) ( 1996) 257 265.
S.M. Allam et a l.
[12] A. W. Beeby, Prediction and Control o f F lexural C racking in
Reinforced Concrete Members, Cracking, Deflection and
U lti mate Load of Concrete Slab System, SP-20, American
Concrete Tnstitu te, D etroit, 1971, pp. 55 75 (cited by Makhlouf
a nd Malhas ( 1996)).
[13] E .G. N a wy, K. W. Blair, Further Studies on Flexural Crack
Control in Structural Slab Systems, Cracking, Deflection, and
Ultima te Load of concrete Slab Systems, SP-20, America n
Concrete Tnstitute, D etroit, 197 1, pp. I 42.
[1 4] R.I. Gilbert, S. Nejadi, A n Experimental Study of Flexural
C racking of Reinforced Concrete Members under Sustained
Loads, UNTC!V Report No. R-435, School of Civil and
Environmental Engineering, University o f New South Wales,
Sydney, Australia, 2004, 59pp (cited by Gilbert (2008)).
[15] Eurocode 2: Design o f Concrete Structures Pa rt I: General
Rules a nd Rules fo r Buildings 1992-1, European Committee for
Standardization, Oc tober 200 I, Belgium.
[16] The Egyptian Code for Design and Construction of Reinforced
Concrete Structures, ECP 203-2007, Ministry of Housing, Egypt.
[17] Building Code Requireme nts for Rein forced Concrete, AC! 31 8-
95 a nd Commentary ACT 318R-95, American Concrete
Tnstitute, Detroit, 1995.
[18] Concrete sanitary engineering structures, a report by ACT
committee 350, ACT Structural J o urnal 80 (1983) 467 486.
[19] A. W. Beeby, Corrosio n of reinforcing steel in concrete and its
relation to cracking, Struct ural Engineer 56A (3) ( 1978) 77 8 1.
[20] A. W. Beeby, C racking, cover, a nd corrosio n of reinforcement,
Conc rete I nternational D esign and Construc tion 5 (2) (1983) 35
40.
[2 1] Building Code Requiremen ts for Reinforced Concrete, ACI 318-
05 a nd Commenta ry ACT 318R-05, American Concrete
I nstitute, Detroit, 2005.
[22] Building Code Requireme nts for Rein forced Concrete, AC! 31 8-
08 a nd Commentary ACT 318R-08, American Concrete
Tnstitute, Detroit, 2008.
[23] Structural Use o f Concrete, Part 2: Code of Practice for Special
C ircumstances BS 8110: P a rt 2: 1997, British Standard
Tnstitution, London, 1998.
[24] E. Hognestad, High strength bars as concrete reinforcement
Pa rt 2: Control of flexural cracking, J ournal Portland Cement
Association Research a nd D evelopment Laboratories 4 (I)
(1962) 46 63 (cited by Park and Paulay (1975)).
[25] P.R . Ka rr, A.H. Mattock, High-strength bars as concrete
reinforcement Part 4 : Control of cracks, J ournal of the PCA
Research Development Laboratories 5 ( I) (1963) 15 38 (cited by
Pa rk and Pa ulay ( 1975)).
[26] P.H. K a rr, E. H ognestad, High-strength ba rs as concrete
reinforceme nt Pa rt 7: Control o f cracks in T-beams flanges,
Journal of the PCA R esearch Development Laboratories 7 ( I)
(1965) 42 53 (cited by Park and Paulay (1975)).
[27] A.P. C la rk, Cracking in reinforced concrete flexural members,
ACI Structural Journal 27 (8) ( 1956) 85 1 862 (cited by Frosch
(1999) and Oh-Kang ( 1987)).
[28] H. Rusch, G . Rehm , Versuche mit Beton formsta hlen,
Deutscher Ausschuss fur Stahlbcton, bulletins, No. 140, Part I,
1963. No. 160 Part 2, 1963; and No. 165, Part 3, 1904 (cited by
Frosch ( 1999)).