Engineering Structures 32 (2010) 472–482

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

An analytical model for crack spacing of thick reinforced concrete plates

N. Dawood a , H. Marzouk b,∗
a
Faculty of Engineering & Applied Science, Memorial University of Newfoundland, Canada
b
Faculty of Engineering, Architecture and Science, Ryerson University, Toronto M5B 2K3, Ontario, Canada

article info abstract

Article history: A simple, practical, and new analytical model that is capable of predicting the crack spacing of
Received 3 July 2009 orthogonally reinforced concrete plate elements is developed. The new model can account for the
Received in revised form influence of the transverse reinforcement on the cracking behavior. The major parameters influencing the
26 October 2009
cracking behavior of reinforced concrete members, such as the tensile strength of concrete, reinforcement
Accepted 26 October 2009
Available online 20 November 2009
ratio, bar diameter and spacing in both longitudinal and transverse directions, are also considered. A
reinforced concrete plate element is studied using equilibrium and compatibility equations to formulate
Keywords:
the average stresses of steel and concrete. A series of verification comparisons with the experimental
Crack spacing investigation conducted in this study, as well as the previous experimental work which involved various
Crack width types of structural elements and loading conditions, are performed to measure the validity and reliability
Concrete plates of the model. Also, this model for the crack spacing can be utilized to predict the value of crack width for
Reinforced concrete various loading stages. In all of the studied cases, the model performs satisfactorily with respect to the
Transverse reinforcement measured responses from the experimental work.
Bond stress © 2009 Elsevier Ltd. All rights reserved.
Cracking behavior
Equilibrium
Offshore structures
Nuclear containments

1. Introduction and fracture energy properties of high strength concrete. An

Concrete is reinforced with steel to deal with its inherent
weakness in tension, where steel reinforcement is used to carry the
tensile forces across the cracks. Once the concrete has exceeded
its tension capacity, the steel reinforcement can prevent crack
widening, and avoid premature and brittle failure in structural
members subjected to tensile stresses. However, there are certain
types of structures, such as offshore platforms, containment
structures for nuclear power plants, and water tanks where tensile
cracks can cause very serious problems. Cracks can cause loss of
load-bearing capacity or increase in permeability of the concrete,
which in turn, causes an unacceptable flow of liquid through the
structure.
There are many existing research reports on the cracking
behavior of reinforced concrete members. Two extensive and
independent programs were conducted to study the cracking
behavior of reinforced concrete segments subjected to pure
uniaxial tension with normal strength concrete by Rizkalla et al.
[1,2]. Marzouk and Chen [3,4] studied the cracking behavior of
concrete specimens under direct uniaxial tension loading and
recommended a suitable softening and tension stiffening model for
high strength concrete that considers the post-cracking behavior
∗ Corresponding author. Tel.: +1 416 979 5000x6451; fax: +1 416 979 5122.
E-mail address: hmarzouk@ryerson.ca (H. Marzouk).
0141-0296/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2009.10.009
important technical report was prepared by Williams [5]; the
main objective of the report was to compare the results obtained
from an experimental program on reinforced concrete panels
with the existing design code and other formulae to produce a
viable method of predicting the cracking behavior of these panels.
Ghali et al. [6] discussed the control of cracking of various
reinforced concrete members under different loading conditions.
Many equations and methods were suggested for calculating the
crack width and spacing. Cho et al. [7,8] conducted tension tests
with six half-thickness concrete wall elements as part of a Korea
Atomic Energy Research Institute (KAERI) program, to provide a
test-verified analytical method for estimating the capacity of a
concrete reactor containment building under internal pressure.
A model equation for the stress–strain relationship of concrete
was developed. Fields and Bischoff [9] preformed an experimental
program on a uniaxial tension member of both normal and high
strength concrete, and proposed an empirical model to predict the
average tensile response of concrete after cracking. The cracking
behavior for various reinforced concrete members under flexural
and uniaxial loading were examined both experimentally and
analytically by Gilbert [10,11].
While the majority of concrete structures are reinforced in
two directions, most of the present research work and existing
analytical model equations for predicting cracking behavior take
only the effect of the reinforcement in the loading direction
 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482 473

tensile strength of concrete, reinforcement ratio, bar diameter

Notations and spacing in both longitudinal and transverse directions.
The proposed model shows a favorable agreement with the
Pc Load sustained by concrete section experimental results of the average crack spacing for different
Ps Load sustained by reinforcement reinforced concrete elements.
Pcr Cracking load
Ec Modulus of elasticity of concrete 2. Experimental investigation
Ac Cross section area of concrete
Act Effective tensile area of concrete In this experimental investigation, a total of six reinforced
Ast Reinforcing steel area concrete panels were fabricated and the effects of the tensile
Smax Maximum crack spacing strength and reinforcement ratio were investigated. The selected
Sm Average crack spacing sizes of the tested panels were 600 × 600 × 190 mm, as shown in
db Bar diameter of the outer layer of the bars Fig. 1. The experimental work was conducted by using a special
ρtNS Effective reinforcement ratio according to NS-S474E test frame that was designed and fabricated in the structural
laboratory at MUN [14]. The main function of this setup was to
code
apply direct axial tension loads in one and/or two perpendicular
ρtCEB Effective reinforcement ratio according to CEB-FIP
directions, and simulate plane uniaxial and biaxial stress states.
1990 code
This test setup consisted of three main parts; namely, the fixed
C Concrete cover, mm
reaction frame, four moving walls, and eight hydraulic jacks.
S Bar spacing of the outer layer of the bars
The hydraulic jacks were placed in between the fixed reaction
k1 Coefficient that characterizes bond properties frame and moving walls to apply forces on the moving walls,
k2 Coefficient to account for strain gradient see Fig. 2. All the specimens in the experimental program were
τb Bond stress at steel–concrete interface reinforced in two perpendicular directions with deformed bars
φ1 Longitudinal bar diameter used in the proposed in two layers, and the longitudinal reinforcements were spaced
model at 150 mm. The main variables considered in the test program
φ2 Transverse bar diameter used in the proposed were the concrete strength, and reinforcement ratio of the axial
model loading in the north–south and east–west directions. Two levels
S1 Longitudinal bar spacing of compressive strength were used to cast specimens. The average
S2 Transverse bar spacing concrete compressive strength fc0 was found equal to 35 MPa
ft0 Tensile strength of concrete for normal strength concrete specimens, and 75 MPa for high
fbb Bearing stress strength concrete specimens. Three control cylinders having a
ρt Reinforcement ratio (Ast /Act ) nominal diameter of 100 mm and height of 200 mm were cast
λ Factor represents the ratio between the stabilized at the same time as the specimens to relate specimen strength
crack spacing at high steel stress and the maximum to standard material test results. The reinforcement used in the
crack spacing at the primary cracks stage segments consisted of deformed bars, and the steel bars used were
wm Average crack width 15 mm and 20 mm in diameter with experimentally measured
εsm Average increase in strain of reinforcement relative yield stress and ultimate tensile strength of 410 MPa and 650 MPa,
to the adjacent concrete. respectively. The steel bars had a modulus of elasticity of 200
εs2 Calculated strain in the steel calculated for a GPa. Two reinforcement ratios were used for the tested panels,
transformed section 1.2%, and 2%, as presented in Table 1. The reinforcement ratio was
ζ Strain reduction factor to allow for the tension equal in both longitudinal and transverse directions. The tension
load was applied in uniaxial and biaxial directions. For the biaxial
stiffening effect of the concrete.
case, the loading was equally applied in the orthogonal directions,
σs2 Steel stresses in a cracked section under the applied
where the loading ratio was equal to 1:1. The concrete cover to bar
load
diameter (C /db ) ratio was fixed equal to 2.50, as shown in Table 1.
σsr Steel stresses in a cracked section at initial crack
All specimens were moist-cured for at least three weeks and then
formation
air-dried until testing.
β1 Factor characterizing the bond quality of the steel Also, Table 1 summarizes the cracking loads Pcr(ex) and concrete
(β1 is 0.5 for smooth bars, otherwise is 1.0) tensile strength ft0 for different concrete specimens, obtained from
β2 Factor representing the effect of load type (β2 is 0.5 the direct uniaxial tension tests conducted in the experimental
for cyclic load or long term, otherwise is 1.0). investigation. The cracking load was selected at the point where
the slope of the stress–strain curve abruptly changed before
reinforcement yielding, and this change could be observed in every
into consideration and tend to ignore the influence of transverse specimen.
reinforcement. Thus, an extensive experimental program was
recently conducted at the Memorial University of Newfoundland 3. Different approaches for estimating the crack spacing
(MUN) to study the cracking behavior of reinforced concrete panels
with different concrete strength under tension (uniaxial and/or Extensive investigations have been carried out with respect
biaxial) in the presence of transverse reinforcement to provide a to crack spacing in reinforced concrete members. Leonhardt [15]
clear understanding of the response of cracked concrete [12,13]. assumed a possible value for the average crack spacing at the
The first phase of this program involves the testing of six reinforced stabilized cracking stage (Sm ). The stabilized crack pattern is the
stage at which no more cracks are expected to occur to the concrete
concrete panels with different concrete strength. During the tests,
as the applied load increased. The value of Sm can be expressed as:
the loads, strains and crack widths were measured.
The main focus of the research presented in this paper aims to 1
Sm = L0 + Lt (1)
develop a rational methodology for predicting the crack spacing 2
for different reinforced concrete members under various loading where, Lt is the transfer length which represents the length of
conditions. It considers all the main factors influencing the active bond stress, and L0 is the length of no bond stresses, and
cracking behavior of reinforced concrete structures, such as the Leonhardt proposed a value for L0 based on the experimental
474 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482

b
Fig. 1. Configuration and dimension of panels. (a) Panels reinforced using a 15 mm bar diameter, (b) Panels reinforced using a 20 mm bar diameter.

Table 1
Reinforced concrete panel segments of the experimental program.
S Specimen symbols Specimen dimensions (mm) Concrete strength fc0 (MPa) Bar dia. (mm) C /db ratio Rft. ratio (%) Pcr(ex) (kN) ft0 (MPa)

1 NS-U-15-2.5-6 40 15 2.5 1.2 240 2.1

2 HS-U-15-2.5-6 90 15 2.5 1.2 400 3.2
3 NS-B-15-2.5-6 35 15 2.5 1.2 220 1.92
600 × 600 × 190
4 HS-B-15-2.5-6 75 15 2.5 1.2 310 2.72
5 HS-U-20-2.5-6 75 20 2.5 2 360 3.1
6 HS-B-20-2.5-6 75 20 2.5 2 335 2.96
NS: Normal strength concrete, HS: High strength concrete, U: Uniaxial tension loading, B: Biaxial tension loading, 6: No. of bars in each direction.

results as follows: where C is the concrete cover, db is the diameter of the main
reinforcement, and ρt is the reinforcement ratio.
fs2,cr An attempt was made by Rizkalla et al. [1,2] to refine the crack
L0 = db (fs2,cr in N/mm2 ) (2)
45 spacing expression by using Beeby’s expression in Eq. (3) where
the value of the length L0 is evaluated. Rizkalla recommended an
where db is the diameter of the main reinforcement, and fs2,cr is the
expression based on the experimental results as follows:
stress in the steel at the cracking stage.
Sm = 5(db − 7.2) + 1.33C + 0.08db /ρt (4)
Beeby [16] proposed an approach for the value of the average
crack spacing Sm as shown in the following equation: where Sm is the average crack spacing mm; C is the concrete cover
to the surface of the bar; ρt is the reinforcement ratio; and db is the
Sm = 1.33C + 0.08db /ρt (3) diameter of the main reinforcement.
 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482
(a) Elevation.
(b) Plan.
Fig. 2. Test setup.
4. Crack spacing approach in the different structures’ codes
4.1. CSA S474 2004; and NS 3473 E 1992
Both the Canadian offshore code CSA-S474-04 [17] and
Norwegian code NS-S474E [18] provide similar approaches for
calculating the value of crack spacing. According to CSA-S474-04,
the average crack spacing of cracks normal to the reinforcement,
Sm , may be calculated using the following equation:
Sm = 2.0(C + 0.1S ) + k1 k2 dbe heff b/Ast
where C = concrete cover, mm (not greater than 55 mm); S =
bar spacing of the outer layer of the bars, mm; k1 = coefficient
that characterizes bond properties of bars (=0.4 for deformed
bars, and =0.8 for plain bars); k2 = coefficient to account for
strain gradient = 0.25 (ε1 + ε2 )/2ε1 , ε1 and ε2 are the largest;
and smallest tensile strains in the effective embedment zone;
dbe = equivalent bar diameter of the outer layer of the bars, mm;
heff = effective embedment thickness, mm, taken as the greater
of (a1 + 7.5dbe ) and (a2 + 7.5dbe ), but not greater than the
tension zone or half the shell thickness (see Fig. 3); b = width of
the section, mm; As = area of reinforcement within the effective
embedment thickness, mm2 .
The crack spacing presented in Eq. (5) includes two terms. The
first term [2.0(C + 0.1S )] is expressed in terms of the concrete
cover and bar spacing, and the second term (k1 k2 dbe heff b/Ast ) is
composed of the type of bar, diameter, effect of bond stress, and
effect of strain gradient. Based on the CSA S474 2004 code equation,
475
in case of using two different bar diameters in the same concrete
section, an equivalent bar diameter should be considered to
calculate the value of the average crack spacing according to Eq. (5).
NS-S474E provides the following equation for the calculation of
the average crack spacing:
Sm = 2(C + 0.1S ) + k1 k2 db /ρtNS (6)
where C = concrete cover, mm; S = bar spacing of the outer
layer of the bars, mm; k1 = factor reflects the effect of the
environment class on the reinforcement corrosion (0.4 for ribbed
(5) bar); k2 = coefficient that takes account of the strain distribution
over the cross section = 0.25 (ε1 + ε2 )/2ε 1 , ε1 and ε2 are the
largest; and smallest tensile strains in the effective embedment
zone; ρtNS = the effective reinforcement ratio according to
NS-S474E code = Ast /Act , Act = the effective concrete area in
accordance with Fig. 4, the part of the tensile zone that is assumed
to contribute effectively to resist tensile forces transferred from
the reinforcement to the concrete by bond, and Ast = the area of
reinforcement of the tensile reinforcement.
It is obvious that both the CSA and NS codes present a similar
equation for calculating the crack spacing, with a small difference
attributed to the difference in the strain gradient and calculation
of the effective reinforcement ratio.
4.2. Eurocode EC2 provisions
In the Eurocode EC2 (2004) [19], the average crack spacing
formula includes the concrete cover as a parameter in the first term
of the formula. The average stabilized mean crack spacing Sm is
476 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482

Fig. 3. Effective embedment thicknesses (CSA S474 2004 Code).

Fig. 4. Guidelines for the determination of an effective concrete area for calculation of crack spacing (shaded) [NS 3473 E 1992 Code].

evaluated from the following expression: average bond stress, N/mm2 = 1.8fctm(t ) ; and fctm(t ) = the mean
value of the concrete tensile strength at the time that the crack
db Act
Sm = 2C + k1 k2 (7) forms. ρtCEB = effective reinforcement ratio according to CEB-FIP
4Ast
1990 code, Ast /Act ; Ast = area of tension reinforcement, mm2 ; and
where C = concrete cover; db = bar diameter mm; Ast /Act = ef- Act = effective concrete area in tension, mm2 , see Fig. 5; and for the
fective reinforcement ratio (see Fig. 5) k1 = 0.8 for deformed bars sake of simplicity, (1 + αe ρtCEB ) can be set equal to 1.
and 1.6 for plain bars; and k2 = 0.5 for bending and 1.0 for pure
tension. In this approach, the EC2 code takes only the effect of the
concrete cover and ignores the influence of the bar spacing of the 5. Analytical model formulation for the crack spacing
outer layer of the reinforcing bars.
Past experience with offshore concrete structures indicated
4.3. CEB-FIP 1990 provisions many different factors affecting concrete durability performance
in marine environment. Where, serious deterioration of concrete
The crack spacing in CEB-FIP code [20] is presented in different structures starts after aggressive materials in the surrounding
expressions from previous codes (CSA, NS, and EC2). In this code, environment are allowed to penetrate into the interior of concrete.
the value of the crack spacing is stated by the following expression: Thus, using HSC thick plates with thick concrete cover (50–70 mm)
2 are commonly used for offshore applications.
Sm = ls,max (8) For a reinforced concrete member that is subjected to axial force
3
or a bending moment; the member is free from cracks as long as the
db stress in the concrete never exceeds its tensile strength. When the
ls,max = For stabilized cracking, (9)
3.6ρtCEB tensile strength in concrete is exceeded, primary cracks are formed
in a region of maximum tensile stresses when the external loads
σs2 1
ls,max = db For single crack formation (10) reach the cracking load. At the location of a crack, the tensile stress
2τb 1 + αe ρtCEB is assumed to be resisted completely by the reinforcement.
where ls,max is the length over which slip occurs between the steel When a segment of reinforced concrete panel is subjected to
reinforcement and concrete; σs2 = reinforcement stress at the axial tension loading that is greater than the cracking load Pcr , as
crack location, N/mm2 ; db is the bar diameter; τb = the value of the illustrated in Fig. 6(a), the idealized cracked panel is assumed to
 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482 477

a b

Fig. 5. Effective area, Aceff : (a) Beam; (b) Slab; (c) Member in tension (MC-90 or EC2).

dσ c 4ρt S1 φ2
 
consist of the effective tensile area of concrete Act , and reinforcing
=− τb + fbb (15)
steel area Ast . In a section situated between two cracks, the bond dx φ1 S2 Act
between the concrete and reinforcing bars restrains the elongation
4ρt S1 φ2
 
of the steel, and the compatibility between the reinforcing bars dσc = − τb · dx + fbb · dx (16)
and the surrounding concrete is assumed, thus, a part of the φ1 S2 Act
tensile force in the reinforcement at the crack is transmitted to the
where fbb is the bearing stress due to the presence of the transverse
concrete situated between the cracks.
reinforcement extended in Y -direction, τb is the bond stress at
The equilibrium of a segment of the longitudinal chord
steel–concrete interface, ρt is the effective reinforcement ratio
representing the tensile zone between two consecutive cracks
(ratio of the tensile reinforcement area to the area of the effective
subjected to an axial tensile force P is studied; the length of the
concrete in tension), φ1 is the longitudinal bar diameter, φ2 is the
segment S represents the crack spacing.
transverse bar diameter, S1 is the longitudinal bar spacing, and S2
The free body diagrams of the steel reinforcement and concrete
is the transverse bar spacing. For simplicity, it is assumed that on
elements are shown in Fig. 6(b). The equilibrium for the free body
of differential steel of length dx is: average, this bearing stress fbb on all bars placed in both directions
X and Y is half of the tensile strength concrete [21].
Fs + τb (π φ1 dx) = Fs + dFs (11) Gilbert [10] proposed that τb = α1 α2 ft0 , where α1 depends
dFs on the steel stress at the crack (and varies from 3.0 at low stress
= π φ1 τb levels to 1.0 at high stress levels); and where α2 = 1.0 for short-
dx
term calculations and α2 = 0.5 for long-term calculations. Short-
dσs π φ1 τb 4τb
= = . (12) term calculations are considered is the present investigation to
dx π φ1 /4
2
φ1 study the cracking response of the concrete plates. The value of
Fig. 6(c) shows a part of the concrete panel and the layout of the the bond forces transferred to the concrete section increase from
reinforcement in the directions X and Y . As a result of presence 0, at location of the crack (x = S /2), to maximum bond force at
of the reinforcement in two-way perpendicular directions and the midway between two consecutive cracks (x = 0), as assumed
considering the firm connection between the longitudinal and in the proposed model, see Fig. 5(a). Thus, the expected location
transverse reinforcements, when the load is applied in the for the maximum concrete stress is at x = 0. Concrete tensile
X -direction and the stretching of longitudinal bars and concrete stresses σc cannot be greater than the tensile strength ft0 , so the
matrix surrounding them in the X -direction are considered, the requirement should be satisfied for the maximum crack spacing
transverse bars in the perpendicular direction can be assumed S = Smax , which will occur when the concrete stress σc = ft0 as
to bear against the surrounding concrete [21]. Another crack presented in the following equations:
may form when over the X -direction a sufficient bond force x =0
4τb ρt x =0
fbb φ2 S1
Z Z 
is developed which, together with the bearing stress along the σc = − dx + dx ≤ ft0
transverse bars is just sufficient to induce a maximum tensile φ1x =S / 2 x=S /2 Act S2
stress equal to the tensile strength of concrete ft0 at the section 
4τb ρt S
 
fbb φ2 S1 S
 
at a distance X = S (the spacing between cracks). Considering = + ≤ ft0 . (17)
uniform stress distribution along the concrete cross section [21], φ1 2 Act S2 2
the equilibrium of forces acting on the concrete segment can be Therefore, the maximum crack spacing may be expressed as:
written as follows:
2τb ρt fbb φ2 S1
 
S1 Smax = ft /0
+ . (18)
Fc = Fc + dFc + (π φ1 dx)τb + fbb φ2 dx (13) φ1 2Act S2
S2
  As loading is increased, additional cracks will form and the
dFc S1
= − (π φ1 dx)τb + φ2 fbb (14) number of cracks will be stabilized when the stress in the concrete
dx S2 no longer exceeds its tensile strength at further locations regard-
dσc (π φ1 ) S1 φ2
 
less of load increase. This condition basically produces the abso-
=− τb + fbb lute minimum crack spacing that can occur at high steel stresses;
dx Act S2 Act
478 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482
b c
Fig. 6. Distribution of bond and bearing stresses over a concrete section.
termed here as stabilized crack spacing [11]. In this research, the
stabilized crack spacing is assumed to be expressed as:
2τb ρt fbb φ2 S1
 
Sm = λSmax = λft /
0
+ (19)
φ1 2Act S2
where λ is a factor which represents the ratio between the stabi-
lized crack spacing at high steel stress value and the maximum
crack spacing at the primary crack formation stage and it is as-
sumed to be equal to 0.67 [11].
The second term in the parentheses in Eq. (18) represents the
effect of the transverse reinforcement on the cracking behavior,
and it is found that the influence of the transverse bar represents
approximately 10% of the first term of the denominator between
the parentheses.
6. Comparison with experimental results
A series of verification tests, involving various types of struc-
tural elements and loading conditions, is conducted to measure
the reliability of the model. The selection of the test series was
based so that the test must allow the verification of the model in
various loading conditions for thick normal and high strength con-
crete structures which is the case for offshore and nuclear power
plants structures. The selected tests involve direct tension tests,
and laterally loaded thick slabs.
6.1. Direct tension tests
An experimental program was recently conducted at MUN
to study the cracking behavior of reinforced concrete thick
plates under tension [12,13]. To enable an accurate crack spacing
measurements to be obtained, the tested panels were carefully
inspected at each load step. The cracks were marked manually
on all specimens at each load increment. Crack mapping of the
specimen was depicted by means of photographs at each stage
of loading throughout the experiment. These photographs were
imported into a computer-aided AutoCAD, tools and the spacing
was measured and averaged using the software.
 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482 479

Table 2
Measured and calculated average crack spacing of reinforced concrete plates under axial loading.
Specimen Specimen Measured average crack spacing Beeby Sb equation Leonhard SL equation Rizkalla equation Proposed
serial number (mm) (mm) (mm) (mm) model S (mm)

1a NS-B-15-2.5-6 151 139.9 203.3 178.9 155.9

2a NS-B-15-2.5-6 144 139.9 186.1 178.9 148.5
3a HS-U-15-2.5-6 152 139.9 238.9 178.9 166.9
4a HS-B-15-2.5-6 151 139.9 211.9 178.9 157
5a HS-U-20-2.5-6 150 133.2 172 197.2 156.6
6a HS-B-20-2.5-6 148 133.2 159.5 197.2 147.2
7b Slab #1 96.01 86.9 142.8 102.36 96.91
8b Slab #2 103.89 86.9 139.2 99.31 123.3
9b Slab #3 137.16 112.3 163.6 131.32 121.2
10b Slab #4 101.85 112.3 162.1 129.54 111
11b Slab #5 155.7 112.3 162.6 126.24 112.5
12b Slab #6 129.79 112.3 162.1 126.49 116.8
13b Slab #7 123.7 112.3 162.1 123.19 133.3
a
Experimental program conducted by Dawood and Marzouk [12,13].
b
Experimental program conducted by Rizkalla et al. [1,2].

Table 3
Measured and calculated average crack spacing of reinforced concrete plates under axial loading.
Specimen serial Specimen number Measured average crack spacing (mm) Sm(ex) /Sm(Beeby) Sm(ex) /Sm(Leonhard) Sm(ex) /Sm(Rizkalla) Sm(ex) /Sm(Proposed)
a
1 NS-B-15-2.5-6 151 1.08 0.74 0.85 0.97
2a NS-B-15-2.5-6 144 1.03 0.77 0.8 0.96
3a HS-U-15-2.5-6 152 1.09 0.64 0.85 0.91
4a HS-B-15-2.5-6 151 1.08 0.71 0.844 0.96
5a HS-U-20-2.5-6 150 1.13 0.87 0.76 0.95
6a HS-B-20-2.5-6 148 1.21 0.93 0.75 1.01
Mean (µ) for Sm(ex) /Sm(tho) 1.103 0.776 0.809 0.96
Standard deviation (σ ) for Sm(ex) /Sm(tho) 0.061 0.107 0.046 0.032
Coefficient of variance (σ /µ) for Sm(ex) /Sm(tho) 5.5% 13.7% 5.6% 3.3%
a
Experimental program conducted by Dawood and Marzouk [12,13].

Table 4
Measured and calculated average crack spacing of reinforced concrete members loaded laterally.
Specimen Specimen Measured average crack spacing Beeby Sm equation Leonhard Sm equation Rizkalla equation Proposed
serial number (mm) (mm) (mm) (mm) model S (mm)

1a Slab NSC1 134 99.52 161.56 188.52 121

2a Slab NSC2 223 173.82 218.42 212.82 235.3
a
3 Slab NSC3 239 221.2 275.48 235.2 252.2
4a Slab HSC1 171 126.12 172.7 215.12 142.1
5a Slab HSC2 185 139.42 174.19 228.42 154.2
6a Slab HSC3 163 119.47 204.71 208.47 161.4
7a Slab HSC4 172 139.26 243.18 228.26 165.5
a
8 Slab HSC5 120 96.15 184.12 135.15 112.3
Mean (µ) for Sm(ex) /Sm(tho) 1.28 0.86 0.85 1.02
Standard deviation (σ ) for Sm(ex) /Sm(tho) 0.094 0.149 0.123 0.06
Coefficient of variance (σ /µ) for Sm(ex) /Sm(tho) 0.073 0.172 0.145 0.059
a
Experimental program conducted by Hossin and Marzouk [22].

The first six specimens presented in Table 2 show a comparison
of the average crack spacing of thick plates, between the results
of the experimental work conducted in the present study and
the analytical model proposed in Eq. (19) next to the previously
proposed equations in Eqs. (1), (3) and (4). The ratios of
experimental to calculated crack spacing (Sm(ex) /Sm(cal) ) are shown
in Table 3. Statistical comparison study is presented in Table 3
between the proposed model in Eq. (19) and the previously
developed model by Beeby, Leonhard, and Rizkalla, at the stabilized
crack pattern.
Furthermore, the proposed Eq. (19) is used to predict the
value of the crack spacing for previous experimental work on
uniaxially loaded slabs conducted by Rizkalla et al. [1,2] which are
represented by specimens 7 through 13 in Table 2. This is to ensure
the suitability of the proposed model for evaluating the value of the
average crack spacing for both thick and thin reinforced concrete
members under axial loading conditions.
6.2. Reinforced thick concrete members under punching load
A series of two-way reinforced thick concrete slabs were
tested under punching load by Hossin and Marzouk [22]. This
experimental testing program was designed to examine the
influence of the concrete cover and bar spacing of normal and
high strength concrete on the cracking behavior. A total of eight
concrete slabs were tested. Five high strength concrete slabs (HSC)
and three normal strength concrete slabs (NSC) were selected
for the experimental investigation of the cracking behavior study
as presented by slabs 1 through 8 as shown in Table 4. The
variables considered in that experimental investigation were the
concrete cover, slab thickness, and bar spacing for normal and high
strength concrete. The selected dimensions for the experimental
testing were typical of those for possible use in Canadian offshore
applications.
Tables 2 and 4 show a comparison between the proposed model
in Eq. (19) versus different approaches for calculating crack spacing
480 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482

Table 5
Measured and calculated average crack spacing by different codes’ approach of reinforced concrete plates under axial loading.
Specimen serial Specimen number Measured average crack spacing (mm) CSA (mm) NS (mm) EC2 (mm) CEB-FIP (mm) Proposed model S (mm)

1a NS-B-15-2.5-6 151 210 230 310 230 155.9

2a NS-B-15-2.5-6 144 210 230 310 230 148.5
3a HS-U-15-2.5-6 152 210 230 310 230 166.9
4a HS-B-15-2.5-6 151 210 230 310 230 157
5a HS-U-20-2.5-6 150 230 230 280 185 156.6
6a HS-B-20-2.5-6 148 230 230 280 185 147.2
7b Slab #1 96.01 92.26 89.39 119 88.9 96.91
8b Slab #2 103.89 92.62 89.39 119 88.9 123.3
9b Slab #3 137.16 130.4 127.5 211.7 125.9 121.2
10b Slab #4 101.85 130.4 127.5 211.7 125.9 111
11b Slab #5 155.7 130.4 127.5 211.7 125.9 112.5
12b Slab #6 129.79 130.6 127.5 211.7 125.9 116.8
13b Slab #7 123.7 130.6 127.5 211.7 125.9 133.3
a
Experimental program conducted by Dawood and Marzouk [12,13].
b
Experimental program conducted by Rizkalla et al. [1,2].

Table 6
Measured and calculated average crack spacing by different codes’ approach of reinforced concrete members loaded laterally.
Specimen serial Specimen number Measured average crack spacing (mm) CSA (mm) NS (mm) EC2 (mm) CEB-FIP (mm) Proposed model S (mm)

1a Slab NSC1 134 127 127 93 78.2 121

2a Slab NSC2 223 203 203 163 189.8 235.3
3a Slab NSC3 239 223 223 195 143.3 252.2
4a Slab HSC1 171 167 167 137 69.3 142.1
5a Slab HSC2 185 187 187 118 67.6 154.2
6a Slab HSC3 163 148 148 109 109.5 161.4
7a Slab HSC4 172 170 170 123 130.9 165.5
8a Slab HSC5 120 139 139 111 83.3 112.3
a
Experimental program conducted by Hossin and Marzouk [22].

of reinforced concrete structures. This is to check the suitability
of such expressions for evaluating the crack spacing of offshore
structures which require using high strength concrete sections and
thick concrete covers.
Based on the comparison presented in Tables 2 and 4 it is noted
that the model proposed by Leonhardt in Eq. (1) overestimates
the value of the crack spacing. However the approach suggested
by Beeby in Eq. (3) underestimates the value of the crack spacing
when it is compared with the experimental results. Thus, it might
be concluded that neither of these two expressions is accurate
enough to predict the value of the crack spacing of the offshore
structures which include the presence of transverse reinforcement
and a thick concrete cover. The model recommended by Rizkalla in
Eq. (4) offers reasonable results as the reinforced structures have a
small bar diameter, such as 10 mm and 15 mm. However, when a
higher bar diameter, such as 20 mm and 25 mm is used, the model
starts to overestimate the value of the crack spacing, since the first
part in Eq. (4) [5(db − 7.2)] is sensitive to the effect of increasing
the bar diameter. Thus, when a higher bar diameter is used, the
final value of the crack spacing is relatively high in comparison to
the experimental value. For all of the studied cases, the proposed
model in Eq. (19) in this study has a satisfactory performance as it
is comparable with the measured crack spacing.
7. Comparison of experimental versus different codes’ ap-
proach for crack spacing
One of the main objectives for this investigation is to check
the applicability of the different approaches in the different codes’
approach, by comparing the calculated value of the crack spacing
according to different code equations and the proposed model
with the experimental results for various concrete members. The
test results will be evaluated with regards to available codes, such
as: Canadian offshore CSA code S474 (CSA-04), Norwegian Code
3473E (NS-92), European Committee for Standardization Eurocode
2 (EC2-2004), and the European CEB-FIP model code (CEB-90). The
average crack spacing measured at each experiment is considered
and presented in Tables 5 and 6 along with numerical estimations
of other international codes.
Tables 5 and 6 show that both CSA and NS code approaches for
calculating the crack spacing fail to take the effect of the transverse
reinforcement into consideration, and that is the reason for giving
the same value for crack spacing for some slabs with different
transverse bar spacing.
With regards to the EC2 code, in spite of its simplicity in
calculating the value of the crack spacing, it is obvious in Table 6
that this code equation underestimates the calculated value for
the crack spacing. EC2 code equation [Eq. (7)] recommends the
use of double concrete cover without considering the effect of
the spacing between the reinforcing bars as recommended by
CSA-S474-04 and NS-S474E in Eq. (5). That is why the EC2 code
equation underestimates the value of the crack spacing, as shown
in Table 6.
The CEB-FIP code equation gives reasonable values for both
pure tension and flexural specimens with a normal concrete cover
of 20–35 mm. However, for specimens with thick concrete covers,
the CEB-FIP equation underestimates the value of the crack spacing
where the effect of the concrete thick cover is not taken into
consideration, as shown in Table 6.
In general, as both the concrete cover and bar spacing increase,
the crack spacing increases theoretically and experimentally. All
of the previous code equations ignore the effect of the transverse
reinforcement on the cracking behavior of the reinforced concrete
structures.
8. Average crack width
Many of the offshore structures and nuclear power containment
are suffering from serious deterioration problems due to the
excessive cracking that will eventually result in a shortened
service life. The eventual construction and refurbishment of such
structures may be in excess of millions of dollars. Thus, it is
desirable to provide a clear understanding of the cracking response
of concrete under tensile axial stresses, as the durability of
reinforced concrete structures and the protection of reinforcement
 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482 481

against corrosion should be considered as imperative requirement

as the strength of the structural members.
Based on the experimental investigation conducted in the
present study [12,13], the HSC panels of 70 MPa show a significant
increase for the tension stiffening response compared with the
similar NSC panels of 35 MPa. Tension stiffening phenomenon
can be defined as the increase in stiffness in a reinforced
concrete member due to the interaction between concrete and
reinforcement, as it compared with the bare bar response. Thus,
any the reduction in crack width for a given member is attributed
mainly to the increase in tension stiffening in terms of the
concrete contribution between cracks. In Eq. (18), the concrete
characteristic is represented by the tensile strength of concrete
and the bond stress between the concrete section and the
reinforcement.
The theoretical model for predicting the crack spacing in
Eq. (19) can be used to predict the value of crack width. The average
Fig. 7. Change in crack widths with steel bar stress for panel NS-U-15-2.5-6.
crack width can be estimated by the following expression in
Eq. (20) for the crack runs in a direction that is perpendicular to
the tensile axial stresses in members subjected to an axial force.
In this equation, it is assumed that the elongation of the concrete
between two adjacent cracks is small enough to be neglected
as it is compared with the elongation of the reinforcing bars.
Then, the average crack width can be evaluated by the following
equation [6]:
wm = Sm εsm . (20)
The average crack width wm is equal to the average crack spacing
Sm , which can be evaluated based on Eq. (19), times the average
increase in strain of reinforcement relative to the adjacent concrete
(εsm ). The average strain increase in steel can be estimated as
follows:
εsm = ζ εs2 (21)
" 2 #
σsr

ζ = 1 − β1 β2 (22)
σs2 Fig. 8. Change in crack widths with steel bar stress for panel HS-U-20-2.5-6.

where εs2 is the calculated strain in the steel calculated for a
transformed section in which the concrete in tension is ignored
(state 2) and neglecting the contribution of concrete tension
stiffening. Variable ζ is a strain reduction factor to allow for the
tension stiffening effect of the concrete. Variables σs2 and σsr are
the steel stresses in a cracked section under the applied load
and at initial crack formation, respectively. Variable β1 is a factor
characterizing the bond quality of the steel (β1 is 0.5 for smooth
bars, otherwise is 1.0). Variable β2 is a factor representing the effect
of load type (β2 is 0.5 for cyclic load or long term, otherwise is 1.0).
Figs. 7 and 8 present a comparison of the values of crack width
at different loading stages between measured experimental results
conducted in this investigation, and evaluated values using Eq.
(20), for normal and high strength concrete panels, respectively.
Figs. 7 and 8 show a convenient match between the analytical and
the experimental results for the crack width for both normal and
high strength concrete panels. For panel NS-U-15-2.5-6 at a steel
stress of 270 MPa, which represents 2/3 of the yield stress of the
reinforcement (steel stress at the service load) [23], the measured
crack width is 0.19 mm, and the calculated value for the crack
width based on Eq. (20) is found equal to 0.211 mm, as shown in
Fig. 7. For panel HS-U-20-2.5-6, at service load level the measured
crack width is found to be 0.089 mm, and the calculated value is
0.121 mm, as shown in Fig. 8.
9. Summary and conclusion
In this paper, a rational crack spacing model is developed by
considering the equilibrium and compatibility equations of rein-
forced concrete elements. The influence of the main parameters
affecting the cracking behavior of reinforced concrete structures,
such as the tensile strength of concrete, reinforcement ratio, and
longitudinal bar diameter φ1 and spacing S1 are taken into consid-
eration. Moreover, the effect of transverse reinforcement is incor-
porated into this model in terms of the transverse bar diameter φ2
and spacing S2 . Hence, the proposed analytical model for maximum
crack spacing can be expressed as the following:
2τb ρt fbb φ2 S1
 
Smax = ft /0
+ .
φ1 2Act S2
To validate the proposed model, the predicted value of the crack
spacing based on the proposed model is compared to the results of
various experimental investigations conducted on thick reinforced
concrete panels loaded axially in tension, thick reinforced concrete
slabs subjected to lateral loads, and transversely loaded full scale
beams. Also, this model is employed to predict the value of the
crack width of thick concrete plates subjected to in-plane axial
loading. In all of the studied cases in this research, the model pre-
dicts the value of crack spacing satisfactorily. This new approach
shows a favorable performance in predicting the values of the crack
spacing and width for thick reinforced concrete members under
different loading conditions. Thus, this model can be used as a prac-
tical aid for calculating an accurate value for the crack spacing in
designing thick heavy reinforced concrete structures, such as off-
shore oil platforms, and containment structures for nuclear power
plants.
Acknowledgements
This research is granted by the Natural Science and Engineering
Council of Canada (NSERC) and received with appreciation. Sincere
thanks are due to the members of the Technical Services at the
482 N. Dawood, H. Marzouk / Engineering Structures 32 (2010) 472–482
Memorial University for their assistance during conduction of the
experimental program. The authors express their gratitude and
appreciation to the technical staff of the Structural Engineering
Laboratory at the Memorial University for their assistance during
testing, especially M. Curtis and S. Organ. Sincere thanks are
extended to Capital Ready Mix Ltd. of St. John’s, Newfoundland for
donating the concrete for this project.
References
[1] Rizkalla SH, Hwang LS, El-Shahawi M. Transverse reinforcement effect on
cracking behavior of R.C. members. Canad J Civil Eng 1983;10(4):556–81.
[2] Riakalla SH, Hwang LS. Crack perdition for members in uniaxial tension. ACI
Struct J 1984;81(4):572–9.
[3] Marzouk H, Chen ZW. Nonlinear analysis of normal- and high-strength
concrete slabs. Canad J Civil Eng 1993;20(4):696–707.
[4] Marzouk H, Chen ZW. Fracture energy and tension properties of high-strength
concrete. J Mater Civil Eng 1995;7(2):108–16.
[5] Williams A. Tests on large reinforced concrete elements subjected to direct
tension. Cement and Concrete Association. Technical report 562. 1986. p. 1–55.
[6] Ghali A, Favre R, Elbadry M. Concrete structures: Stresses and deformations.
3rd ed. New York: Spon Press; 2002. p. 380–554.
[7] Cho Jae-Yeol, Kim Nam-Sik, Cho Nam-So, Choun Young-Sun. Stress–strain
relationship of reinforced concrete subjected to biaxial tension. ACI Struct J
2004;101(2):202–8.
[8] Cho Jae-Yeol, Kim Nam-Sik, Cho Nam-So, Choi In-Kil. Cracking behavior of
reinforced concrete panel subjected to biaxial tension. ACI Struct J 2004;
101(1):76–84.
[9] Fields K, Bischoff PH. Tension stiffening and cracking of high-strength
reinforced concrete tension members. ACI Struct J 2004;101(4):447–56.
[10] Gilbert RI. Time-dependent cracking and crack control in reinforced concrete
structures. ACI Struct J 2005;225(15):223–44 [Special Publication].
[11] Gilbert RI. Design for flexural crack control-recent amendments to the Aus-
tralian standard AS3600. In: 1st international structural specially conference,
ST-048-1- ST-048-10. 2006.
[12] Dawood N, Marzouk H. Cracking behavior of high strength concrete panels
under uniaxial and biaxial loading. In: 1st international structural specialty
conference. 37th annual conference for the canadian society for civil
engineering. 2008. p. 1–10.
[13] Dawood N, Marzouk H. Effective stress–strain relationship of reinforced
concrete panels under uniaxial and biaxial loading. In: 8th international
symposium on utilization of high strength and high-performance concrete.
2008. p. 1–8.
[14] Sabrah TB, Marzouk H, Hussein A. Development of a new test facility to
examine Reinforced concrete panels under uniaxial and biaxial direct tension.
J. Test Eval ASTM 2009;37(6). http://www.astm.org.
[15] Leonhardt F. Crack control in concrete structures. IABSE surveys, no. S-4/77,
IABE periodical 3/1977. In: International association for bridge and structural
engineering. 1977. p. 26.
[16] Beeby AW. A study of cracking in reinforced concrete members subjected
to pure tension. Technical Report 42.468. London: Cement and Concrete
Association; 1972.
[17] Canadian Standard Associations. S474 concrete structures. Mississauga
(Ontario, Canada): Canadian Standard Associations; 2004.
[18] Norwegian Council for Building Standardization. NS34773E Concrete struc-
tures design rules. Norway: Norwegian Council for Building Standardization;
1992.
[19] EC2, Eurocode 2. Design of concrete structures—Part I: General rules and rules
for buildings. Paris: European Committee for Standardization; 2004.
[20] Comite Euro-International Du Beton. CEB-FIP model code 1990. Lausanne
(Swittzerland): Comité Euro-Internationale du Beton et Federation Interna-
tionale de la Procontrainte.
[21] Desayi P, Kulkarni AB. Determination of maximum crack width in two-way
reinforced concrete slabs. Proc Inst Civil Eng (London), Part 2. Res Theory 1976;
61:343–9.
[22] Hossin MA, Marzouk H. Offshore concrete crack analysis. In: 1st international
structural specially conference, ST-119-1-ST-119-10. 2006.
[23] ACI Committee 318. Building code requirements for structural concrete (ACI
318-05) and commentary (ACI 318R-05). Reported by ACI Committee 318.
Farmington Hills (MI): American Concrete Institute; 2005.