DeStefano, Evans, Tadros, Sun 2003 ISHPC

FLEXURAL CRACK CONTROL IN CONCRETE BRIDGE STRUCTURES

Ralph J. DeStefano, PE, Pennsylvania Department of Transportation, Hollidaysburg, PA

Jack Evans, PE, Florida Department of Transportation, Tallahassee, FL
Maher K. Tadros, PhD, PE, Dept of Civil Engineering, Univ. of Nebraska, Omaha, NE
Chuanbing Sun, Dept of Civil Engineering, University of Nebraska, Omaha, NE

ABSTRACT

Mitigation and control of cracking is important for obtaining acceptable

appearance and for long-term durability of concrete bridge structures,
especially those subjected to saltwater spray and deicing chemicals. In the
current AASHTO LRFD Specifications and in the ACI 318 Code editions prior
to 1999, flexural crack control requirements were based on the so-called Z-
factor method developed by Gergely and Lutz in 1968. Their work was based
on extensive statistical evaluation of laboratory experiments. ACI 318
decided in the 1999 edition to greatly simplify crack control requirements due
increased evidence suggesting a reduced correlation between crack width and
reinforcement corrosion. The ACI formula relates maximum bar spacing to
steel stress under service loads and clear concrete cover. This paper presents
the background of previous crack control methods and comparison of their
impact on design of bridges and culverts. A new equation is proposed for
adoption in the AASHTO LRFD Specifications that is intended to be simple,
yet account for differences between bridge and building structures.

Keywords: Concrete, crack control, serviceability, reinforcement, corrosion

1
DeStefano, Evans, Tadros, Sun 2003 ISHPC

INTRODUCTION

With the introduction of the AASHTO LRFD specifications in 1994, several states have
made the observation that crack control requirements tend to govern the design of flexural
steel in reinforced concrete members more frequently than under the LFD Standard
Specifications. While it is believed the primary reason for crack control governing more
often under the LRFD specification is due changes made on the load side of the equation
(load factors and reduced number of service limit states), designers and researchers have also
taken the opportunity to re-evaluate the resistance side of the equation, which is essentially a
direct carry over from the Standard Specifications.

Additional crack control research conducted in the 1980's and 1990's, which has failed to
show any kind of definitive correlation between crack width and long-term durability of
reinforced concrete structures has led specification organizations to re-evaluate the
underlying design provisions. They began to question the logical behind requiring detailed
computations for a phenomenon with wide inherent scatter and questionable end results. In
1999, ACI 318 adopted a much simpler provision to design for crack control. Now,
AASHTO is investigating adoption of a similar simplified provision.

The intent of this paper is to provide background on the development of the existing
AASHTO LRFD crack control specifications and to provide a discussion on the proposed
revisions.

BACKGROUND

With the inception of ultimate strength design methodology and higher strength Grade 60
steel reinforcement, researchers and designers recognized the need for providing a
mechanism by which crack widths would be minimized through reinforcement detailing. It
was observed that the use of fewer, larger reinforcement bars at wider spacings resulted in
larger cracks than the use of more smaller reinforcement bars at tighter spacings for the same
amount of total reinforcement area. Controlling crack widths was believed important from a
long-term durability standpoint (corrosion of the reinforcement) as well as from an aesthetics
standpoint.

Research was undertaken in the 1960's to quantify the above concept and to develop design
tools. Out of this research came the well-known Gergely-Lutz equation1. The empirical
equation was developed using statistical analysis techniques on experimental data from
several different researchers. The original equation for predicting the width of a flexural
crack on the tension face of a reinforced concrete member is:

wc = 0.000076 β f s 3 d c A (1)

where,
wc = crack width, in.;

2
DeStefano, Evans, Tadros, Sun 2003 ISHPC

β = factor relating the strain at the tension face to the strain at the centroid of the
reinforcement;
fs = stress in steel reinforcement, ksi;
dc = distance from tension face to centroid of nearest reinforcement layer, in.;
A = average effective concrete area per bar of the flexural tension reinforcement, in2. For
a single layer of reinforcement of constant spacing, the term A simplifies to 2 dc s,
where s = bar spacing.

Another well-known crack width predictive model developed in the same era was the Kaar-
Mattock equation2. Similar to the Gergely-Lutz equation, it was developed based on a
statistical analysis of experimental data. The Kaar-Mattock equation for predicting the width
of flexural cracks at the tension face of a reinforced concrete flexural member is:

wc = 0.000115 β f s 4
A, (2)

where the variables are the same as defined for the Gergely-Lutz equation.

AASHTO adopted the Gergely-Lutz equation for controlling flexural cracking, but in a
slightly rearranged form3. The crack width variable and the β factor were consolidated into a
single Z-factor, and the equation was written in terms of allowable stress. Using an
approximate limiting crack width of 0.016 in. and an average β factor of 1.2 results in the
current equation found in the AASHTO LRFD specifications:

Z
f sa = ≤ 0.6 f y (3)
3 dc A

where,
fsa = allowable reinforcement stress, ksi'
Z = 170 for moderate exposure conditions,
= 130 for severe exposure conditions,
= 100 for precast box culverts,
155
= for cast-in-place box culverts;
β
with the remaining terms as defined previously.

The equation has been a source of contention among designers for the some time for the
following reasons:

• the equation has a built-in paradox in that increasing the concrete cover is detrimental to
crack control even though it is well established that increasing cover is beneficial in
enhancing long-term durability. In fact, concrete covers over 3 inches are essentially
impractical to construct using the results from the equation.

3
DeStefano, Evans, Tadros, Sun 2003 ISHPC

• computation of the effective concrete area term, A, can be cumbersome for members with
mixed bar sizes and bundled bars;
• precast box culverts are treated differently from cast-in-place box culverts.

AASHTO addressed the first issue by allowing designers to use 2 inches maximum for the
concrete cover in the crack control computations regardless of the actual cover with the
rationale that any additional cover is sacrificial. However, such arbitrary manipulation of the
equation has generated suspicion as to the overall accuracy of the methodology.

A NEW CRACK WIDTH EQUATION

Frosch4 observed that a significant shortcoming of the Gergely-Lutz and Kaar-Mattock

equations is that they were both developed empirically using statistical analysis techniques
on experimental data that was limited in the range of concrete covers investigated. In fact, he
noted that only three test specimens had concrete covers greater than 2.5 inches.

Frosch developed the following simple, theoretically-derived equation to predict crack

widths that could be used regardless of the actual concrete cover:

2
⎛s⎞
f
wc = 2 s β (d c )
2
+⎜ ⎟ (4)
Es ⎝ 2⎠

where,
wc = crack width, in.;
fs = stress in steel reinforcement, ksi;
Es = modulus of elasticity of steel reinforcement, ksi;
β = factor relating the strain at the tension face to the strain at the reinforcement layer;
dc = distance from tension face to centroid of nearest reinforcement layer, in.;
s = reinforcement bar spacing, in.

The equation can be re-written to solve for the maximum permitted reinforcement bar
spacing for a limiting crack width, as follows:

2
⎛wE ⎞
s = 2 ⎜⎜ c s ⎟⎟ − d c 2 (5)
⎝ 2 fs β ⎠

The equation is based on the following premises:

• the crack width is a direct function of the strain in the reinforcement times the crack
spacing;
• the maximum crack spacing is estimated to be two times the controlling cover distance
given by the expression, d* = [dc2 + (0.5s)2]1/2;

4
DeStefano, Evans, Tadros, Sun 2003 ISHPC

• the concrete between cracks does not resist any tension (which will result in a more
conservative estimate of crack width.)

Frosch compared the crack widths predicted with his theoretical cracking model to the widths
predicted from the Gergely-Lutz and Kaar-Mattock empirical equations. The results showed
very good correlation.

NEW ACI 318 CRACK CONTROL EQUATION

Based on the work by Frosch5 and additional crack control research6,7 performed in the
1980’s and 1990’s that failed to demonstrate a credible link between crack width and
problems with long-term durability of reinforced concrete members, the ACI 318 Code
switched from the long-held Gergely-Lutz equation and adopted a simplified version of the
Frosch's cracking model8.

The final form of the cracking model equation adopted by ACI 318 in 1999 is:

540 432
s = − 2.5cc ≤ (6)
fs fs

where,
cc = clear concrete cover on reinforcement nearest the tension face, in.;
with the other variables as defined previously.

This simplified equation of the cracking model is based on the following assumptions:

β = 1 + 0.08 dc;
wc = 0.016 in. limiting crack width, in.;
dc = cc + 0.5 in., (i.e., #8 average bar size.)

The following should be noted about the approach taken by ACI:

• in lieu of explicit computation of fs, ACI permits use of an assumed value of 0.6fy, or
36 ksi for Grade 60 reinforcement.
• emphasis is placed on limiting the reinforcement spacing and not limiting the allowable
stress;
• no distinction is made between interior and exterior exposure conditions;
• no upper limit is defined for the stress in the reinforcement, fs.

If a value of 36 ksi is assumed for fs, the ACI equation yields maximum permitted bar
spacings of 11.25", 10", 7.5" and 5" for concrete covers of 1.5", 2", 3" and 4", respectively.

5
DeStefano, Evans, Tadros, Sun 2003 ISHPC

PROPOSED AASHTO LRFD CRACK CONTROL EQUATION

Based on a review of past research, parametric studies comparing various crack width
predictive methods, and results from actual design example problems, an equation very
similar to the Frosch cracking model adopted by ACI 318 is also proposed for consideration
by AASHTO. The proposed AASHTO equation is as follows:

700γ eγ r
f sa = ≤ 0 .8 f y (7)
β ( s + 2d c )

where,
fsa = allowable service level stress in the reinforcement, ksi;
γe = exposure factor,
= 1.0 for Case 1,
= 0.75 for Case 2;
γr = reinforcement factor,
= 0.75 for smooth weld-wire fabric,
= 1.00 for all other types of reinforcement;
dc
β = 1+ ;
0 .7 ( h − d c )
with the remaining variables as defined previously.

The proposed AASHTO equation is essentially the same as the ACI 318 equation with the
following modifications:

• the limiting crack width is approximately 0.017 in., slightly larger than the value of
0.016 in. used previously;
• addition of an exposure factor, γe, is included to maintain a distinction between different
environmental conditions;
• addition of a reinforcement factor, γr, to recognize the fact that research suggests smooth
welded-wire fabric has reduced bond properties, most notably for wider wire spacings,
compared to other reinforcement types;
• the equation is written in terms of allowable stress to be consistent with past practice,
although rearranging the equation to be written in terms of maximum permitted bar
spacing, s, can be done easily;
• the β-factor is made an integral part of the equation rather than an assumed "average"
value;
• the dc term is retained in the formal development of the cracking model instead of using
the clear cover with an assumed "average" rebar size of #8;

A larger limiting crack width of 0.017 inch is used because of the proposal to retain the
distinction between exposure conditions. By providing two different levels of cracking
tolerance, a slightly larger crack width is considered acceptable for the less severe exposure
condition (Case 1.)

6
DeStefano, Evans, Tadros, Sun 2003 ISHPC

An upper limit of 0.8fy on the allowable stress is proposed to provide a factor of safety
against permanent yielding of the reinforcement under service loads. This provision is
similar to that stipulated for steel flexural members in Article 6.10.5.2 of the AASHTO
LRFD specifications3.

Additional discussion is presented in later sections of the paper on the β-factor and the
exposure factor being proposed for the AASHTO equation.

COMPARISON OF RESULTS

Figures 1, 2 and 3 show a comparison of results between the current AASHTO LRFD
equation (Gergely-Lutz), the Kaar-Mattock equation, the ACI 318-99 equation, the cracking
model by Frosch and the proposed AASHTO LRFD equation for concrete covers of 2", 3"
and 4", respectively. In these figures, a constant value of 1.2 is assumed for β. Furthermore,
Case 1 exposure conditions ('moderate' exposure) are evaluated as well as a constant
reinforcement bar size of #8. It is important to note that for the current AASHTO equation,
the 2" maximum concrete cover stipulation permitted by the specification is used.

The following observations are made concerning the results:

• The simplified ACI 318-99 equation fits the theoretical cracking model well for all
concrete cover cases. It permits slightly higher allowable stresses for the 2-inch cover
case and slightly lower allowable stresses for the 3-inch and 4-inch cover cases.
• The proposed AASHTO equation permits slightly higher allowable stresses compared to
the ACI 318-99 and the crack model equation, which is to be expected since the proposed
AASHTO equation is based on a slightly larger limiting crack width.
• The Kaar-Mattock equation is permits the highest levels of allowable stresses compared
to all of the other equations, and for stress levels less than approximately 40 ksi, it
permits significantly higher allowable stresses.
• The current AASHTO equation, in general, permits higher allowable stresses at wider bar
spacings and lower allowable stresses at tighter bar spacings. It is interesting to note,
though, that if the comparisons are limited to a stress level less than 36 ksi, which is the
upper limit stipulated in the AASHTO specifications, then the current AASHTO equation
is less conservative (permits higher allowable stresses) than the cracking model, the ACI
318-99 equation and the proposed AASHTO equation. However, what is important to
note is that these comparisons are limited to a single layer of reinforcement consisting of
#8 bars (i.e., dc = 2.5 inches.) Whenever situations are encountered in which additional
depth must be accounted for in the computation of the dc term, then the current AASHTO
equation can be significantly more conservative. This fact is demonstrated in the design
example for the footing reinforcement at the end of this paper.

7
DeStefano, Evans, Tadros, Sun 2003 ISHPC

2" CONCRETE COVER FOR CASE 1 EXPOSURE (#8 BAR)

24.0

β = 1.2
21.0

KAAR-MATTOCK
18.0
MAXIMUM BAR SPACING (in.)

ACI 318-99
15.0

PROPOSED AASHTO
12.0

9.0

CRACKING
6.0 MODEL

3.0 CURRENT AASHTO LRFD

(GERGELY-LUTZ)

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 1: Allowable Stress vs. Bar Spacing for Case 1 Exposure — 2" Concrete Cover

3" CONCRETE COVER FOR CASE 1 EXPOSURE (#8 BAR)

24.0

β = 1.2
21.0
CURRENT AASHTO LRFD
(Based on 2" Max. Concrete Cover)
18.0
MAXIMUM BAR SPACING (in.)

KAAR-MATTOCK
15.0

CRACKING MODEL
12.0

PROPOSED AASHTO LRFD

9.0

6.0

ACI 318-99
3.0

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 2: Allowable Stress vs. Bar Spacing for Case 1 Exposure — 3" Concrete Cover

8
DeStefano, Evans, Tadros, Sun 2003 ISHPC

4" CONCRETE COVER FOR CASE 1 EXPOSURE (#8 BAR)

24.0

β = 1.2
21.0
CURRENT AASHTO LRFD
(Based on 2" Max. Concrete Cover)
18.0
MAXIMUM BAR SPACING (in.)

KAAR-MATTOCK
15.0

12.0
PROPOSED AASHTO LRFD

9.0
CRACKING MODEL
6.0

ACI 318-99
3.0

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 3: Allowable Stress vs. Bar Spacing for Case 1 Exposure — 4" Concrete Cover

DISCUSSION OF β−FACTOR

One of the primary differences between the ACI 318-99 equation and the equation proposed
for AASHTO involves explicitly including the β-factor in the crack control equation. The
ACI equation does consider the β-factor, but in an implicit manner. The β-factor built into
the development of the equation is:

β = 1 + 0.08 d c (8)

which was developed by Frosch based on a review of the sections from the crack control test
data.

The β-factor simply represents the increase in crack width at the tension face of the member
compared to the crack width at the centroid of the outer most layer of steel reinforcement
assuming linear strain distribution. The theoretical definition of the β factor is:

ε2 h - c
β= = (9)
ε1 d - c

9
DeStefano, Evans, Tadros, Sun 2003 ISHPC

A sketch showing the various parameters defining β is shown in Figure 4.

c
N.A.

d
h
ε1

ε2

dc
Reinforcement
Fig. 4: β-Factor Definition

Figures 5, 6 & 7 show the significant effect of the β-factor because of the wide range of
flexural member depths encountered in bridge construction. Box culvert slabs can be on the
order of 8" thick while pier cap depths and footing thicknesses can be on the order of 7 feet
and more. The β-factor effect is fairly dramatic for the 2-inch concrete cover case, but
diminishes as the concrete cover increases primarily because thin, shallow flexural members
are uncommon for the thicker cover cases. Take, for example, the case of 2" concrete cover
for an 8"-thick flexural member in which β = 1.65. The allowable stress for reinforcement
spaced at 12" is only 25 ksi (based on Case 1 exposure condition.) However, for a 72"-thick
member in which β = 1.05, the allowable stress increases to approximately 39 ksi — a 56%
increase just by considering the depth of the flexural member.

The effect of β has long been recognized by AASHTO as an important parameter in the crack
control equation. However, since the inception of the provision there has been a general
reluctance to explicitly include the term in the general equation even though the variables
needed to compute the parameter are minimal and readily available. Instead, special cases
were used. For example, the AASHTO equation has always had β = 1.2 implicitly included.
But, in the 1982 Interim AASHTO Standard Specifications (Article 1.5.6), a Z-factor of 98
was introduced specifically and solely for box culverts. The commentary provided in the
interim specification stated, "The expression for maximum service load stress for crack
control more nearly reflects the Gergely and Lutz recommendations (SP-20 A.C.I. 1968, pp.
87-117) as confirmed by Lloyd, Rejali and Kesler (A.C.I. Journal No. 5, V. 66, May 1969,
pp. 366-376) for one-way slabs than does the formula in Section 5." By applying the ratio of
a β = 1.2 divided by β = 1.6 to the severe exposure condition value of Z = 130, the value of
98 is obtained.

In the AASHTO LRFD specifications, a value of Z = 100 was selected for box culverts.
Unfortunately, this value for Z applied to all cases was recognized as being overly
conservative for many cast-in-place box culvert applications in which slab and wall
thicknesses typically exceed 12 inches. It is for this reason that subsequent to the 1st edition
of the AASHTO LRFD specifications released in 1994, an expression for β was introduced.
However, use of this equation was stipulated for cast-in-place box culverts only. Z = 100
continued to be specified for precast box culverts.

10
DeStefano, Evans, Tadros, Sun 2003 ISHPC

The expression for Z given in the current AASHTO LRFD specifications is:

155 dc
Z= , where β = 1 + (10)
β 0.7d

Substituting β = 1.2 in the above equation results in the typical value of Z = 130 that has
been used for severe exposure conditions since inception of the provision. The expression
for β can be derived easily using an assumed value of 0.3 for 'k' in the familiar 'kd' term (the
location of the neutral axis) from working stress design.

2" CONCRETE COVER FOR CASE 1 EXPOSURE (#8 BAR)

24.0

BETA EFFECT
21.0

PROPOSED AASHTO
18.0 72" Member Depth,
MAXIMUM BAR SPACING (in.)

β = 1.05

15.0
PROPOSED AASHTO
β = 1.2
12.0

9.0

6.0

PROPOSED AASHTO
3.0 8" Member Depth, β = 1.65

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 5: Allowable Stress vs. Bar Spacing Demonstrating β Effect — 2" Concrete Cover

11
DeStefano, Evans, Tadros, Sun 2003 ISHPC

3" CONCRETE COVER FOR CASE 1 EXPOSURE (#8 BAR)

24.0

BETA EFFECT
21.0

18.0
MAXIMUM BAR SPACING (in.)

PROPOSED AASHTO
72" Member Depth,
15.0
β = 1.07

12.0
PROPOSED AASHTO
β = 1.2
9.0

6.0

3.0 PROPOSED AASHTO

12" Member Depth, β = 1.59
0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 6: Allowable Stress vs. Bar Spacing Demonstrating β Effect — 3" Concrete Cover
4" CONCRETE COVER FOR CASE 1 EXPOSURE (#8 BAR)
24.0

BETA EFFECT
21.0

18.0
MAXIMUM BAR SPACING (in.)

15.0
PROPOSED AASHTO
72" Member Depth,
12.0
β = 1.10

9.0
PROPOSED AASHTO
β = 1.2
6.0

3.0 PROPOSED AASHTO

18" Member Depth, β = 1.48
0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 7: Allowable Stress vs. Bar Spacing Demonstrating β Effect — 4" Concrete Cover

12
DeStefano, Evans, Tadros, Sun 2003 ISHPC

DISCUSSION OF EXPOSURE FACTOR

Another significant variation from the ACI 318-99 equation being proposed for AASHTO is
continuing to differentiate between exposure conditions. ACI 318 decided to eliminate the
previously held distinction between interior and exterior exposure conditions. Maintaining
some form of distinction between different conditions was considered important for bridge
construction because:

• Not all engineers are convinced there is no correlation between crack width and long-
term durability.

• Some states definitely have much harsher exposure and environmental conditions (such
as deicing chemical and salt spray) compared to other states. To eliminate all such
distinction does not seem rational and is certainly inconsistent with past practice.

• Even if corrosion is not highly correlated to cracking, some situations can certainly
tolerate higher levels of cracking compared to others. For example, cracking can be less
tolerated for piers that are highly visible to the traveling public simply from a perception
standpoint, even if it is known from a structural standpoint that the cracking will have no
detrimental effect.

Instead of using the current terminology of 'normal' and 'severe' exposure conditions, the
terms "Case 1" and "Case 2" are proposed. The intent of changing the terminology is to
better reflect the purpose of the factor, which is to simply differentiate between situations in
which more or less cracking can be tolerated by the owner.

The proposed value of 0.75 for the exposure factor, γe, was derived simply by taking the ratio
of the current 130 and 170 Z-factors. Figures 8, 9, 10 and 11 show how the proposed crack
control equation compares against the current AASHTO equation as well as the ACI 318-99
equation for the Case 2 exposure condition for concrete covers of 1.5", 2", 3" and 4"
respectively. As can be seen, the ACI curve is the same as for the Case 1 exposure condition,
again, because ACI does not make any distinction.

The largest deviation between the proposed equation and the current equation occurs for the
2-inch concrete cover case. Only for relatively wide bar spacings (exceeding approximately
13.5") does the current AASHTO equation permit a higher allowable stress than the proposed
equation. For bar spacings less than 9", the proposed equation results in significantly higher
allowable stresses compared to the current equation, yet, the allowable stresses remain
significantly less than the ACI 318 equation.

As the concrete cover increases beyond 2", the current and proposed equations converge only
because of the artificial 2" maximum cover limitation allowed by the current AASHTO
LRFD specifications. Such an arbitrary limitation is not used for the proposed method. The
actual cover is to be always used except for cases in which additional clearance is stipulated
for construction tolerances. (An example of this condition in Pennsylvania occurs for the

13
DeStefano, Evans, Tadros, Sun 2003 ISHPC

bottom mat of reinforcement in footings. Because the concrete is cast directly on the earth
subgrade that may be unevenly prepared, an additional 1" of cover is stipulated as a
precautionary measure to ensure that 3" minimum cover is obtained.)

Only for the 4" concrete cover case does the current AASHTO equation result in a higher
allowable stress than the proposed equation for the full range of bar spacings. However, for
the normal range of bar spacings between 4" and 9", the difference in allowable stress is
relatively small and considered acceptable especially considering the fact that the proposed
equation does not use the arbitrary 2" maximum cover limitation. If the owner considers this
reduction in allowable stress unacceptable, a possible viewpoint that could be taken is that by
providing 4 inches of concrete cover, a Case 1 exposure condition could be used for
computation purposes instead of the Case 2 exposure condition. It is this type of flexibility
in interpretation of the crack control provisions that is considered to be a positive aspect of
the proposed equation.

Under no condition is the allowable stress obtained from the proposed AASHTO equation
greater than the allowable stress obtained from the ACI 318 equation.

1.5" CONCRETE COVER FOR CASE 2 EXPOSURE (#8 BAR)

24.0

21.0

CRACKING MODEL
18.0
MAXIMUM BAR SPACING (in.)

15.0
ACI 318-99

12.0
PROPOSED AASHTO LRFD

9.0

6.0

3.0
CURRENT AASHTO LRFD

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 8: Allowable Stress vs. Bar Spacing for Case 2 Exposure — 1.5" Concrete Cover

14
DeStefano, Evans, Tadros, Sun 2003 ISHPC

2" CONCRETE COVER FOR CASE 2 EXPOSURE (#8 BAR)

24.0

21.0

18.0
MAXIMUM BAR SPACING (in.)

CRACKING MODEL
15.0

ACI 318-99
12.0

PROPOSED AASHTO LRFD

9.0

6.0

3.0
CURRENT AASHTO LRFD
0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 9: Allowable Stress vs. Bar Spacing for Case 2 Exposure — 2" Concrete Cover
3" CONCRETE COVER FOR CASE 2 EXPOSURE (#8 BAR)
24.0

21.0

18.0
MAXIMUM BAR SPACING (in.)

CRACKING MODEL
15.0

ACI 318-99
12.0

PROPOSED AASHTO LRFD

9.0

CURRENT
6.0 AASHTO LRFD

3.0

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 10: Allowable Stress vs. Bar Spacing for Case 2 Exposure — 3" Concrete Cover

15
DeStefano, Evans, Tadros, Sun 2003 ISHPC

4" CONCRETE COVER FOR CASE 2 EXPOSURE (#8 BAR)

24.0

21.0

18.0
MAXIMUM BAR SPACING (in.)

CRACKING MODEL
15.0
PROPOSED AASHTO LRFD
12.0
ACI 318-99
9.0
CURRENT AASHTO LRFD
6.0

3.0

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 11: Allowable Stress vs. Bar Spacing for Case 2 Exposure — 4" Concrete Cover

DISCUSSION OF CULVERTS

Of particular interest in revising the crack control equation is to treat precast and cast-in-
place culverts in a consistent manner. Figures 12 and 13 show comparisons between the
proposed AASHTO equation and the current equation for precast and cast-in-place culverts
for member depths of 8" and 12" and for concrete covers of 1 1/2" and 2", respectively. A
separate exposure condition for precast culverts, as currently implemented in AASHTO, is
not proposed for the new provisions, because the β factor is explicitly included in the general
crack control equation and a special case is considered unnecessary.

The results show that the proposed AASHTO equation permits higher allowable stresses than
the current equation for both precast and cast-in-place culverts. The increase is particularly
more for precast culverts than for cast-in-place culverts. Such increases in allowable stresses
are not cause for concern, because the allowable stresses currently permitted by the
AASHTO equation are unusually low and do not seem to be warranted. For example, for the
2" cover case, even with a relatively tight bar spacing of 6 inches, the allowable stress is only
24 ksi regardless of the slab thickness. Under the proposed provisions, the allowable stress
would increase to 35 ksi for a 12" thick slab and 30 ksi for an 8" thick slab.

16
DeStefano, Evans, Tadros, Sun 2003 ISHPC

1.5" CONCRETE COVER FOR CASE 2 EXPOSURE (#8 BAR)

24.0

CULVERTS
21.0

CURRENT AASHTO
18.0 CAST-IN-PLACE
MAXIMUM BAR SPACING (in.)

12" Member Depth, β = 1.29

15.0
PROPOSED AASHTO
12" Member Depth, β =
12.0 1 29
CURRENT AASHTO
CAST-IN-PLACE
9.0 8" Member Depth, β = 1.48

PROPOSED AASHTO
6.0 8" Member Depth, β = 1.48

3.0
CURRENT AASHTO
PRECAST, Z = 100
0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 12: Allowable Stress vs. Bar Spacing for Culverts — 1.5" Concrete Cover

2" CONCRETE COVER FOR CASE 2 EXPOSURE (#8 BAR)

24.0

CULVERTS
CURRENT AASHTO
21.0
CAST-IN-PLACE
18" Member Depth, β = 1.23
18.0
MAXIMUM BAR SPACING (in.)

PROPOSED AASHTO
12" Member Depth, β = 1.38
15.0
CURRENT AASHTO
CAST-IN-PLACE
12.0
10" Member Depth, β = 1.48

CURRENT AASHTO
9.0
PRECAST, Z = 100
PROPOSED AASHTO
6.0 8" Member Depth, β = 1.65

3.0

0.0
20 25 30 35 40 45 50 55 60

ALLOWABLE STRESS (ksi)

Fig. 13: Allowable Stress vs. Bar Spacing for Culverts — 2" Concrete Cover

17
DeStefano, Evans, Tadros, Sun 2003 ISHPC

DESIGN EXAMPLES

Several design examples are provided for the most common serviceability computations
encountered in bridge design. The intent is to demonstrate the simplicity of the proposed
method and to provide a direct comparison in the allowable stress results between the current
AASHTO method and the proposed method.

EXAMPLE 1 — CAST-IN-PLACE BOX CULVERT

Given:
Member depth, h = 12"
Concrete cover, cc = 2"
Reinforcement, #6 @ 8"

Current AASHTO Method:

dc = 2" + 0.75" / 2 = 2.375"

Effective depth, d = 12" – 2.375" = 9.625"
155
Z =
β
dc 2.375
β = 1+ = 1+ = 1.35
0.7 d 0.7 x 9.625
155
Z = = 114.8
1.35
A = 2 d c s = 2 x 2.375 x 8 = 38.0 in2
Z 114.8
fsa = = 3 = 25.6 ksi
3 d A 2.375 x 38.0
c

Proposed AASHTO Method:

dc = 2" + 0.75" / 2 = 2.375"

Effective depth, d = 12" – 2.375" = 9.625"
Exposure factor, γe = 0.75 for culverts.
Rebar factor, γr = 1.00 for deformed reinforcement
dc 2.375
β = 1+ = 1+ = 1.35
0.7(h − d c ) 0.7(12 − 2.375)
700 γ e γ r 700 x 0.75 x 1.00
fsa = = = 30.5 ksi
β (s + 2 d c ) 1.35 (8 + 2 x 2.375)

18
DeStefano, Evans, Tadros, Sun 2003 ISHPC

EXAMPLE 2 — PRECAST BOX CULVERT

Given:
Member depth, h = 12"
Concrete cover, cc = 1.5"
Reinforcement, #6 @ 8"

Current AASHTO Method:

dc = 1.5" + 0.75" / 2 = 1.875"

Effective depth, d = 12" – 1.875" = 10.125"
Z = 100
A = 2 d c s = 2 x 1.875 x 8 = 30.0 in2
Z 100
fsa = = 3 = 26.1 ksi
3 d A 1.875 x 30.0
c

Proposed AASHTO Method:

dc = 1.5" + 0.75" / 2 = 1.875"

Effective depth, d = 12" – 1.875" = 10.125"
Exposure factor, γe = 0.75 for culverts.
Rebar factor, γr = 1.00 for deformed reinforcement
dc 1.875
β = 1+ = 1+ = 1.26
0.7(h − d c ) 0.7(12 − 1.875)
700 γ e γ r 700 x 0.75 x 1.00
fsa = = = 35.5 ksi
β (s + 2 d c ) 1.26 (8 + 2 x 1.875)

EXAMPLE 3 — ABUTMENT STEM, REAR FACE STEEL

Given:
Member depth, h = 30"
Concrete cover, cc = 3"
Reinforcement, #10 @ 12"

Current AASHTO Method:

AASHTO Art. 5.7.3.4 stipulates using a maximum concrete cover of 2" regardless of
the actual cover when computing dc.
dc = 2.0" + 1.27" / 2 = 2.64"
Z = 130 (severe exposure condition)
A = 2 d c s = 2 x 2.64 x 12 = 63.4 in2
Z 130
fsa = = 3 = 23.6 ksi
3 d A 2.64 x 63.4
c

19
DeStefano, Evans, Tadros, Sun 2003 ISHPC

Proposed AASHTO Method:

dc = 3.0" + 1.27" / 2 = 3.64" (Note: the actual concrete cover is used.)

Exposure factor, γe = 0.75 (Case 2 condition selected.)
Rebar factor, γr = 1.00 for deformed reinforcement
dc 3.64
β = 1+ = 1+ = 1.20
0.7(h − d c ) 0.7(30 − 3.64)
700 γ e γ r 700 x 0.75 x 1.00
fsa = = = 22.7 ksi
β (s + 2 d c ) 1.20 (12 + 2 x 3.64)

Change Reinforcement Design:

Try #7 @ 6"

Current AASHTO Method:

dc = 2.0" + 0.875" / 2 = 2.44"

A = 2 d c s = 2 x 2.64 x 6 = 29.3 in2
Z 130
fsa = = 3 = 31.3 ksi
3 d A 2.44 x 29.3
c

Proposed AASHTO Method:

dc = 3.0" + 0.875" / 2 = 3.44"

dc 3.44
β = 1+ = 1+ = 1.19
0.7(h − d c ) 0.7(30 − 3.44 )
700 γ e γ r 700 x 0.75 x 1.00
fsa = = = 34.3 ksi
β (s + 2 d c ) 1.19 (6 + 2 x 3.44)

EXAMPLE 4 — PIER CAP

Given:
Member depth, h = 96"; Cap width = 48"
Concrete cover, cc = 2" + 5/8" stirrup = 2.63"
Reinforcement,
Row 1 → 6 – bundled (2 bar) #10's
Row 2 → 6 – single #10's with 3" clear distance from Row 1.

Current AASHTO Method:

dc = 2" max cover + 5/8" + 1.27 = 3.90"

Centroid of reinforcement:
y = 12 x 3.90" + 6 x (3.90" + 1.27 + 3" clr + 1.27/2) / 18 = 5.54"

20
DeStefano, Evans, Tadros, Sun 2003 ISHPC

Effective concrete area per bar:

For bundled bars, use 1.4 equivalency factor.
A = 48" cap width x (2 x 5.54") / (6 bundles x 1.4 eq. factor + 6) = 36.9 in2
Z 130
fsa = = 3 = 24.8 ksi
3 d A 3.90 x 36.9
c

Proposed AASHTO Method:

dc = 2.0" + 5/8" stirrup + 1.27" = 3.90"

Exposure factor, γe = 0.75 (Case 2 exposure condition selected.)
Rebar factor, γr = 1.00 for deformed reinforcement.
Reinforcement spacing,
48" wide cap
s = = 8"
6 bar bundles in row nearest face
dc 3.90
β = 1+ = 1+ = 1.06
0.7(h − d c ) 0.7(96 − 3.90)
700 γ e γ r 700 x 0.75 x 1.00
fsa = = = 31.3 ksi
β (s + 2 d c ) 1.06 (8 + 2 x 3.90)

EXAMPLE 5 — PIER FOOTING

Given:
Member depth, h = 48"
Concrete cover, cc = 3" (4" actual but 1" considered construction tolerance)
Reinforcement in bottom mat,
Bottom layer → #10 @ 9"
Top layer → #10 @ 9"

Bottom Layer — Current AASHTO Method:

AASHTO Art. 5.7.3.4 stipulates using a maximum concrete cover of 2" regardless of
the actual cover when computing dc.
dc = 2.0 + 1.27" / 2 = 2.64"
Z = 170 (normal exposure condition)
A = 2 d c s = 2 x 2.64 x 9 = 47.5 in2
Z 170
fsa = = 3 = 34.0 ksi
3 d A 2.64 x 47.5
c

Bottom Layer — Proposed AASHTO Method:

dc = 3.0" + 1.27" / 2 = 3.64"

Exposure factor, γe = 1.00 (Case 1 exposure condition selected.)
Rebar factor, γr = 1.00 for deformed reinforcement

21
DeStefano, Evans, Tadros, Sun 2003 ISHPC

dc 3.64
β = 1+ = 1+ = 1.12
0.7(h − d c ) 0.7(48 − 3.64)
700 γ e γ r 700 x 1.00 x 1.00
fsa = = = 38.4 ksi
β (s + 2 d c ) 1.12 (9 + 2 x 3.64)

Top Layer — Current AASHTO Method:

dc = 2.0 + 1.27" + 1.27" / 2 = 3.91"

Z = 170 (normal exposure condition)
A = 2 d c s = 2 x 3.91 x 9 = 70.4 in2
Z 170
fsa = = 3 = 26.1 ksi
3 d A 3.91 x 70.4
c

Bottom Layer — Proposed AASHTO Method:

dc = 3.0" + 1.27 + 1.27" / 2 = 4.91"

Exposure factor, γe = 1.00 (Case 1 exposure condition selected.)
Rebar factor, γr = 1.00 for deformed reinforcement
dc 4.91
β = 1+ = 1+ = 1.16
0.7(h − d c ) 0.7(48 − 4.91)
700 γ e γ r 700 x 1.00 x 1.00
fsa = = = 32.1 ksi
β (s + 2 d c ) 1.16 (9 + 2 x 4.91)

Table 1: Summary of Design Example Results

Member Specifics Reinforcement Current Proposed
Cast-in-place 12" thick slab,
#6 @ 8" 25.6 ksi 30.5 ksi
Box Culvert 2" concrete cover
Precast Box Culvert 12" thick slab,
#6 @ 8" 26.1 ksi 35.5 ksi
1 1/2" concrete cover
Abutment Stem, R.F. 30" thick section, #10 @ 12" 23.6 ksi 22.7 ksi
3" concrete cover #7 @ 6" 31.3 ksi 34.3 ksi
Pier Cap 48" wide, Row 1:
96" deep, 6 – 2 bar #10
24.8 ksi 31.3 ksi
2" concrete cover, Row 2:
#5 stirrups 6 – 1 bar #10
Pier Footing 48" thick, Bottom layer:
3" concrete cover, #10 @ 9" 34.0 ksi 38.4 ksi
bottom mat Top layer:
#10 @ 9" 26.1 ksi 32.1 ksi

22
DeStefano, Evans, Tadros, Sun 2003 ISHPC

CONCLUSIONS

The proposed revised distribution of flexural reinforcement equation satisfactorily addresses

a number of shortcomings identified with the current Z-factor method as implemented in the
AASHTO LRFD specifications. A logical and rational, yet simple, physical cracking model
developed by Frosch, that has been shown to correlate well with the database of existing
experimental data, is used to predict crack width. While additional concrete cover does
reduce the allowable stress, the reduction is not nearly as severe as with the current method.
No longer is the concrete cover arbitrarily limited to a maximum value of 2 inches in order to
obtain reasonable results. Furthermore, precast box culverts are handled the same as cast-in-
place box culverts in that the β effect is an integral part of the equation instead of being
applied to cast-in-place boxes only. Deeper flexural members will also benefit because the β
value is typically 10% – 15% smaller than the assumed value of 1.2 built into the existing
equation.

An exposure factor, γe, defined as Case 1 or Case 2, is incorporated into the equation in order
to account for varying structural applications found in bridge construction, be it from an
aesthetics or corrosion standpoint. The factor provides the opportunity to restrict or relax the
crack control criteria based on a particular application as determined by the owner.

Example designs were performed to compare the allowable stresses between the existing
Z-method and the proposed crack control equation. The results show reasonable increases in
allowable stresses, thus permitting more economical designs without sacrificing long-term
durability or aesthetics.

REFERENCES

1. Gergely, P., and Lutz, L. A., "Maximum Crack Width in Reinforced Concrete Flexural
Members," Causes, Mechanism, and Control of Cracking in Concrete, SP-20, American
Concrete Institute, Farmington Hills, Mich., 1968, pp. 87–117.
2. Kaar, P. H., and Mattock, A. H., "High-Strength Bars as Concrete Reinforcement —
Part 4: Control of Cracking," Journal, PCA Research and Development Laboratories, V.
4, No. 1, January 1962, pp. 46-65.
3. "AASHTO LRFD Bridge Design Specifications," 2nd Edition, American Association of
State Highway and Transportation Officials, Washington, D.C., 1998. p. 5-40.
4. Frosch, R. J., "Another Look at Cracking and Crack Control in Reinforced Concrete,"
ACI Structural Journal, May-June 1999, pp. 437 – 442.
5. Frosch, R. J., “Flexural Crack Control in Reinforced Concrete,” Design and Construction
Practices to Mitigate Cracking, SP 204, American Concrete Institute, Farmington Hills,
Mich., 2001, pp. 135-154.
6. Darwin, D. et. al., "Debate: Crack Width, Cover, and Corrosion," Concrete International,
V. 8, No. 5, May 1985, pp. 20-35.
7. Oesterle, R. G., "The Role of Concrete Cover in Crack Control Criteria and Corrosion
Protection," PCA R&D Serial No. 2054, Portland Cement Association, 1997.

23
DeStefano, Evans, Tadros, Sun 2003 ISHPC

8. ACI Committee 318, "Building Code Requirements for Structural Concrete (ACI 318-99)
and Commentary (ACI 318R-99)," American Concrete Institute, Farmington Hills,
Mich., 1999.

24