Download as pdf or txt
Download as pdf or txt
You are on page 1of 55

Shear Behaviour of Large, Lightly-Reinforced

Concrete Beams and One-Way Slabs

CHAPTER 6:

Control of Crack Widths

CONTROL OF CRACK WIDTHS

Surface cracking is inevitable but, with proper structural design and detailing, the cracks are very
narrow and barely perceptible. -A.M. Neville

This chapter describes a series of experiments designed to investigate the effects of bar
diameter of skin reinforcing steel on crack widths in the webs of large flexural elements.
This chapter also describes experiments to investigate the effects of skin reinforcement on
the shear behaviour of flexural elements, and whether the shear strength is related to the
vertical spacing between skin reinforcing bars. It is found that the diameter of skin
reinforcing bars has a clear effect on crack widths, and that the ACI skin reinforcement
provisions should specify a minimum bar diameter. It is also found that the shear
strength of large members is not entirely related to the vertical spacing between skin
reinforcement, and a modified method by which the SMCFT should calculate sx is
recommended, based on the effective depth of the member.

6.1

General

The low tensile strength of concrete relative to its compressive strength means that most
non-prestressed concrete in service is cracked to some degree. In zones of tension, the
steel reinforcement is engaged primarily when a crack occurs, and design of reinforced
concrete structures is carried out based on the fact that significant portions of the
structure are cracked. However, the widths of these cracks must be limited for
appearance, durability and structural integrity.
It is important to limit crack width so as to ensure adequate shear behaviour. As crack
widths increase, their ability to transfer shear stresses by aggregate interlock decreases.
Members in which there is insufficient reinforcement to control crack widths are at risk
of developing wide cracks that may result in a premature shear failure. This is of
particular concern for very thick members without stirrups, as cracks widths within the
web can be considerably greater than those at the level of the steel. It is an aspect that
has typically not been addressed by previous studies on crack widths in reinforced
232

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

concrete, as the focus has generally been on the effects of crack widths on durability and
appearance.
The 1995 version of the ACI-318 code included provisions for crack control based on
crack width limits of 0.4mm (0.016in.) and 0.33mm (0.013in.) for interior and exterior
applications, respectively. However, ACI Committee 318 is now of the opinion that
crack width is not directly related to long-term durability, with cover depth and concrete
quality being of greater importance (ACI Committee 224, 1993). Furthermore, ACI
Committee 318 now believes that, given the inherent variability of crack widths in
concrete structures, it can be misleading to use a design method that purports to
effectively calculate crack widths. Hence, crack control requirements in the ACI code
have evolved over ten years, with the 2005 crack control requirements representing a
considerable departure from the 1995 requirements.
A particular aspect of the 2005 requirements that is worthy of further study is the skin
reinforcement requirements. Skin reinforcement is provided within the web of thick
members so as to control the width of flexural cracks as they extend above the tension
steel. See, for example, the three 20M bars provided on each face of the transfer girders
described in Figure 1-4. However, as discussed in the following sections, the 2005 ACI
318 code no longer requires a minimum bar diameter for skin reinforcement, based on
research suggesting that spacing of skin reinforcing bars is the primary variable affecting
flexural crack widths in the webs of thick members. It is thus possible to use, for
example, D4 wires in place of No. 5 skin reinforcing bars, at the same spacing, and still
meet the 2005 ACI 318 skin reinforcement requirements. This is despite the fact that the
area of steel has been reduced by 87%.
Based on the above discussion, the intention of this chapter is to investigate the skin
reinforcement requirements of the 2005 ACI-318 code. The effects of crack control
reinforcement on the shear behaviour of thick slabs will also be investigated, and the
ability of the 2004 CSA A23.3 code to account for these effects will be assessed.
233

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.2

Control of Crack Widths

Crack Control in the ACI-318 Code

6.2.1

Crack Control at the Level of the Tensile Steel

In the 1995 ACI-318 code, crack control requirements at the level of the tensile
reinforcement were based on the well-known Gergely-Lutz expression (Gergely and Lutz,
(1968)), which was derived from regression analyses on data from several crack width
studies:
w b = 0.0763 t b A Rf s

(6.1)

where
wb = crack width on the bottom (soffit) of the member,
tb = cover from bottom of member to centre of lowest level of steel= dc,
R = (h-kd)/((1-k)d) = factor to account for strain gradient (ratio between strain at bottom
of member and strain at level of reinforcement)
fs = steel stress
A = 2b(h-d)/m = effective area of concrete in tension surrounding the reinforcement
b = width of member at centroid of steel
m = number of tensile reinforcing bars
In their analysis of flexural crack widths at the level of the reinforcement and on the
bottom face of the member, Gergely and Lutz found that:
1.
2.
3.
4.
5.
6.

The steel stress is the most important variable.


The cover thickness is an important variable but not the only consideration.
The bar diameter is not a major variable.
The size of the side crack width is reduced by the proximity of the compression
zone in flexural members.
The bottom crack width increases with the strain gradient.
The major variables are the effective area of concrete, Ac, the number of bars, m, the
side or bottom cover, and the steel stress.

Point 3 appears to be counterintuitive, but the effect of bar diameter is, in fact, taken into
account by other parameters. Reducing the bar diameter by using a large number of
small diameter bars (at a reduced spacing) would be expected to produce smaller crack
widths than would the use of a small number of large diameter bars at the same steel area
234

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

and steel stress. This effect is accounted for by the variable m. The bar diameter also
affects crack widths when the total steel area is reduced by using smaller diameter bars,
and this effect is accounted for by use of the steel stress term fs, which will increase.
Equation (6.1) was developed to calculate the most probable crack width on the bottom
of the flexural member. A second expression was also derived to calculate the most
probable crack width on the side face of the member at the level of the reinforcement.
The ACI implementation of the Gergely-Lutz expression used R=1.2 and required the
calculation of a z-factor as outlined below, in which z was limited to 175kips/in for
interior exposure and 145kips/in for exterior exposure. These limits correspond to crack
widths of 0.016 and 0.013in (0.4 and 0.33mm). The CSA-A23.3 code also uses the zfactor for controlling crack widths at the tensile reinforcement, with z being limited to
30,000N/mm and 25,000N/mm for interior and exposure conditions respectively.
z = fs 3 dcA

(6.2)

A challenge posed by the z-factor is that it promotes the use of smaller covers below the
level of the reinforcement so as to reduce dc. Yet, it is generally understood that larger
covers are very effective at improving long-term durability, possibly even more effective
than limiting crack widths (ACI Committee 224 (1993)). Furthermore, while a range of
dc values from 0.75-3.31in. (19 84mm) was used by Gergely and Lutz to derive their
expression for bottom-face cracking, there were only three data points with a cover
greater than 2.5in (64mm). As such, it can be difficult to meet the requirements of Eq.
(6.2) at covers exceeding 2in (50mm). The commentary to Clause 10.6.1 in the CSA
code suggests that in situations with large covers, it is not necessary to use a value for
clear cover greater than 50mm when calculating dc and A. In these situations, it is better
to allow thicker covers at the expense of wider surface crack widths. In these cases,
crack widths at the level of the steel will remain small, with the wider surface crack
widths therefore becoming essentially an aesthetic issue. This simple solution was not
implemented in the ACI code.
235

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Frosch (1999) noted the difficulty in meeting the requirements of the z-factor for larger
covers, and developed a new approach to crack control at the level of tension
reinforcement. Reviewing the work of Broms (1965), Frosch noted that the spacing of
cracks depends on the concrete cover, and calculates the crack spacing as follows:
Sc = s d *

(6.3)

where
Sc = crack spacing,
d* = controlling cover distance (Figure 6-3)
s = crack spacing factor
=1.0 for minimum crack spacing
=1.5 for average crack spacing
=2.0 for maximum crack spacing
Figure 6-1: Controlling Cover Distance
Noting that the crack width at the level of the reinforcement wc=sSc, Frosch derived an
equation for the maximum crack width on the bottom of the beam as follows:
wc = 2

fs
d c2 + (s/2) 2
Es

(6.4)

where
Es = Youngs Modulus of steel,
= equivalent to Gergely-Lutz R-value
= 1.0 + 0.08dc as a simplification.
Frosch rearranged this equation to solve for the permissible bar spacing, s, as a function
of the permissible maximum crack width, wc:
w E
s = 2 c s
2f s

d c2

(6.5)

A permissible crack width of between 0.016 and 0.021in (0.4-0.53mm) was chosen by
Frosch, a service load steel stress of 0.6fy was assumed, and simplified design curves
generated as shown in Figure 6-2 (Frosch Design Curves).
236

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Eq. 6-5, wc=0.016in.


Eq. 6-5, wc=0.021in.

Reinforcement Spacing, s (in.)

16

Frosch Design Curve

12
10
8
ACI-318 Design Curve
Eq. 6-6, No. 8 Bar

Grade 75 Steel
Eq. 6-5, wc=0.016in.

16

14

18

Grade 60 Steel

Reinforcement Spacing, s (in.)

18

Control of Crack Widths

Eq. 6-5, wc=0.021in.

14

Frosch Design Curve

12
10
8
6
ACI-318 Design Curve
Eq. 6-6, No. 8 Bar

4
2

2
0

3
4
Concrete Cover, dc (in.)

3
4
Concrete Cover, dc (in.)

Figure 6-2: Rebar Spacing Requirements Eq. 6-5 and Simplified Design Expressions
In implementing the design recommendation of Frosch (1999), ACI Committee 318
chose more conservative design curves as shown in Figure 6-2 (ACI Design Curves), and
the expression for calculating minimum bar spacing is shown below (in ksi, inch units).
The ACI expression is formulated in terms of the clear cover, cc, and entered use in the
1999 ACI-318 design code.
s=

540
2.5c c 12(36 f s )
fs

(6.6)

where
s = centre-to-centre spacing of tension reinforcement
fs = calculated stress in longitudinal reinforcement at service loads (in ksi). In lieu of
direct calculation, it is permitted to take fs=60% of the specified yield strength
cc = clear cover from surface to tensile steel.
ACI Committee 318 now believes that, given the inherent variability of crack widths in
concrete structures, it can be misleading to use a design method that purports to
effectively calculate crack widths (ACI Committee 224, 1993). Frosch (1999), for
example, notes that crack spacing (and, hence, crack widths) can vary by a factor of 2.
The 1978 CEB-FIP code (CEB 1978) suggests that the 95th percentile of crack widths is
equal to 1.7 times the average crack width. A distinction is no longer made between
interior and exterior exposure conditions, as the committee has accepted that crack widths
are not directly related to durability.
237

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Due to changes in -factors and load combinations in the 2002 version of the ACI code,
Equation (6-6) was reformulated for the 2005 version of the code to take into account the
higher service load stresses in flexural steel:
s=

15(40)
2.5c c 12(40 f s )
fs

(6.7)

where
fs = calculated stress in longitudinal reinforcement at service loads (in ksi). In lieu of
direct calculation, it is permitted to take fs=2/3 of the specified yield strength
Despite an increase in service load stresses of 10%, the required spacing of the tensile
reinforcement was not changed. For the case of a 2in. (50mm) clear cover, the required
spacing in both versions of the expression is 10 in. (254mm). The maximum spacing for
tensile reinforcement is 12 in. (300mm).

6.2.2

Skin Reinforcement

It has long been recognized that flexural crack widths can increase in width as the cracks
extend into the web of a deep member (Figure 6-3), and it is argued in Chapter 5 that this
is the primary cause of the size effect in shear. The 1977 version of the ACI-318 code,
for example, required that an area of steel equal to 10% of the tensile reinforcement be
distributed along the side faces of deep members to control crack widths in the web.

Figure 6-3: Side-Face Cracking in Large Beams (adapted from Frantz and Breen (1980))

238

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The logic of this requirement in the 1977 ACI code is weak, however, as identified by
Frantz and Breen (1976, 1980a,b). For a given factored moment, increasing the effective
depth of a member will result in a lower area of tensile steel and a lower area of skin
reinforcement, where in fact a deeper section would logically require at least the same, or
probably additional, skin reinforcement. Thus, Frantz and Breen carried out an extensive
series of tests in which crack widths in deep flexural members with various skin
reinforcement configurations were measured and analyzed, and recommended a design
procedure for skin reinforcement. They found a very clear relationship between the skin
reinforcement ratio, sk, and the maximum crack width in the web. They recommend that
the expression described in Figure 6-4 be used when determining the required amount of
skin reinforcement.

For 30 < d < 100inches, sk = 0.00024 (d-30)


d > 100inches, sk = 0.011 + 0.000058d

Figure 6-4: Frantz and Breen (1980a) Skin Reinforcement Requirements


ACI 318-02
The results of Frantz and Breens study formed the basis of the ACI skin reinforcement
requirements up until the 2002 ACI-318 design code. As shown in Figure 6-5, it is
reflected in the requirement that that the spacing, ssk, between skin reinforcing bars not
exceed 1000Ab/(d-30), where Ab is the area of an individual bar. In previous versions of
the code this requirement was formulated such that the area of skin reinforcement per
foot height of web per side exceed 0.012(d-30) in2/ft. Both expressions, however, are

239

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

mathematically equivalent. No distinction was made between interior and exterior


exposure conditions, unlike the z-factor for the flexural steel.
10.6.7 If the effective depth d of a beam or joist
exceeds 36in., longitudinal skin reinforcement shall
be uniformly distributed along both side faces of the
member for a distance d/2 nearest the flexural
tension reinforcement. The spacing ssk between
longitudinal bars or wires of the skin reinforcement
shall not exceed the least of d/6, 12in. and
1000Ab/(d-30). It shall be permitted to include such
reinforcement in strength computations if a strain
compatability analysis is made to determine stress in
the longitudinal bars or wires. The total area of
longitudinal skin reinforcement in both faces need
not exceed one-half of the required flexural tensile
reinforcement.

Figure 6-5: ACI 318-02 Skin Reinforcement Requirements


The ACI 318-02 maximum spacings, ssk, are shown in Figure 6-6a) as a function of the
beam depth, d, for various bar diameters. The resulting skin reinforcement ratios, sk, are
shown in Figure 6-6b) for a value of cc+0.5db=2in, where sk is calculated as Abar/(ssk x
(2cc+db)), as opposed to the method by which Frantz and Breen calculate sk. This is the
method used to calculate sk in the CSA code, and sk calculated using this method can be
converted to Frantz and Breens sk by multiplying it by the ratio (No. of bars per side
/(No. of bars per side + 1). The Frantz and Breen expressions are also shown in Figure
6-6b), and have been modified for the different method of calculating sk.
Inspection of Figure 6-6a) will show that, for all bar sizes, the maximum spacing initially
increases as a function of the depth by virtue of the d/6 spacing limit, until the point at
which the 1000Ab(d-30) limit governs. At this point, the required spacing decreases.
The d/6 limit thus serves to prevent designs with both large bar spacings and large bar
diameters at effective depths close to 36in. An efficient use of steel would result by
using No. 3 bars as skin reinforcement for depths from 36-48in., No. 4 bars for depths
from 48-60in., and No. 5 bars for depths exceeding 60in. (Figure 6-6b)).
240

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

12

1000

c c <1.2in.

Required Spacing of Skin Reinforcement, s

(in.)

10

2000

(mm)

3000

1000Ab
Limit
(d-30)

d/6 Limit

c c =2in.

300

200
No. 5

c c =3in.
ACI-318-05

No.4

150

No. 3

100

D10

2
Skin Rft.
not req'd in
either code

0
0

1.8

2000

1.6

ACI-318-02

1.4
1.2

No. 5
CSA Exterior

1.0
0.8

CSA Interior

0.6

Range of sk
for d=55in.

No. 4

0.4
D10

0.2

(in.)

No. 3

D4

d=55in.

Effective Depth, d

3000

(mm)

Frantz&Breen -No.5

50

D4

0
12 24 36 48 60 72 84 96 108 120

a)

1000

Frantz&Breen -D4
(mm)

250

ACI-318-02

0
2.0

Skin Reinforcement Ratio, sk (%)

Control of Crack Widths

d=55in.

0.0
0

b)

12

24

36

48

60

72

84

96 108 120

Effective Depth, d

(in.)

Figure 6-6: ACI 318-02 and ACI-318-05 Skin Reinforcement Requirements


ACI 318-05
In an attempt to unify the flexural reinforcement spacing requirements of clause 10.6.4
and the skin reinforcement spacing requirements of clause 10.6.7, Frosch (2002)
undertook a review of Frantz and Breens study. In the review, Frosch suggested that the
spacing of skin reinforcing bars has a far greater effect on the crack widths in the web of
flexural elements than does the diameter of the bars. In deriving a new design method for
skin reinforcement, Frosch (2002) predicts that the bar size does not significantly affect
crack width and that any size bar can be used successfully. Frosch also quotes
Gergely and Lutz, who found that the bar diameter was not major variable. The clear
effect of sk found by Frantz and Breen was thus suggested by Frosch to be a result
primarily of changes in the bar spacing rather than the bar diameter.

241

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Based on further analytical work, Frosch found that the crack widths on the side face of
deep flexural elements are related to their distances from the bars. As shown below,
Frosch developed an equation for the side face crack width, ws, at a distance x below the
neutral axis (Figure 6-7):
w s = s x d *x

(6.8)

where
ws = crack width located a distance x below the neutral axis,
x = longitudinal strain at a distance x below the neutral axis
dx* = controlling cover distance at a distance x below the neutral axis
= (x' )2 + ds 2
s =1.0 for minimum crack spacing, =1.5 for average spacing, =2.0 for max spacing
x` =vertical distance from point x to nearest reinforcing bar

The depth of the neutral axis, c, is calculated based on an elastic analysis of the
transformed section.

Figure 6-7: Effect of Skin Reinforcement According to Frosch (2002)


The effect of side face steel on crack widths is modelled by the reduction of the
controlling cover distance, dx*. Cracks are thus predicted to be locally wider at s/2 from
a skin reinforcing bar, and locally narrower directly beside a bar (where x=0). Noting
that side face cracks exhibit considerable variability in widths, it was suggested that the
largest crack width can be calculated using s=2, and the narrowest crack width
calculated using s=1. Showing generally good agreement between predicted and
242

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

experimental crack widths from the Frantz and Breen study, Frosch recommended that
Equation (6-7) be applied to the design of skin reinforcement in addition to flexural
reinforcement. Clause 10.6.7 was thus rewritten for the 2005 version of the ACI code
based on the work of Frosch (2002). See Figure 6-8.
Code

Commentary

10.6.7 Where h of a beam or joist


exceeds 36in., longitudinal skin
reinforcement shall be uniformly
distributed along both side faces of the
member. Skin reinforcement shall extend
for a distance h/2 from the tension face.
The spacing s shall be as provided in
10.6.4, where cc is the least distance from
the surface of the skin reinforcement or
prestressing steel to the side face. It shall
be permitted to include such reinforcement
in strength computations if a strain
compatibility analysis is made to
determine stress in the individual bars or
wires.

R10.6.7 For relatively deep flexural members, some


reinforcement should be placed near the vertical faces
of the tension zone to control cracking in the web.10.20,
10.21
(See Fig. R10.6.7.) Without such auxiliary steel, the
width of the cracks in the web may exceed the crack
widths at the level of the flexural tension reinforcement.
This section was modified in the 2005 edition to make
the skin reinforcement spacing consistent with that of
the flexural reinforcement. The size of the skin
reinforcement is not specified; research has indicated
that the spacing rather than bar size is of primary
importance.10.21 Bar sizes No. 3 to No. 5 (or welded
wire reinforcement with a minimum area of 0.1 in.2
per foot of depth) are typically provided. (emphasis
added)

Figure 6-8: ACI 318-05 Skin Reinforcement Requirements


The code clause no longer specifies a minimum sk or bar diameter, since ACI
Committee 318 believes, as identified in the commentary, that bar size is, at best, of
secondary importance.
The skin reinforcement bar spacings based on the 2005 ACI code are shown in Figure
6-6a), where it can be seen that, for example, for a 2in. cover, a spacing of 10in. is
required. However, unlike the 2002 code provisions, this spacing is independent of the
bar size, db, and thus the skin reinforcement ratio, sk. In the case of a 55in. deep beam
with a clear cover of 2in., the 2002 code would have required an unreasonably small
maximum spacing of 1.6in for a D4 deformed wire (Abar=0.04in2), while the 2005 code
requires a maximum spacing of 10in. In this case, the 2005 code suggests that, using D4
skin reinforcement, a skin reinforcement ratio of about 15% that recommended by the
2002 code will sufficiently control crack widths.
243

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

While the commentary suggests that No. 3 to No. 5 bars are generally used, this is not a
mandatory code requirement. Furthermore, the commentary also suggests that welded
wire reinforcement with an area of 0.1in2/foot of depth is also typical. For a 55in. deep
beam with a cover of, say, 1.2in., the 2002 code would have required a spacing of 4in. for
D10 bars, resulting in an sk of 0.9%. For the same beam using D10 bars as skin
reinforcement, the 2005 code requires a spacing of 12in., resulting in an sk of 0.3%. In
this case, the 2005 code suggests that one-third of the steel that was required in the 2002
is sufficient.

6.2.3

Skin Reinforcement CSA Code

The CSA code has also used the work of Frantz and Breen for its skin reinforcement
requirements. However, the code requires a constant sk of 0.8% for interior conditions
and 1% for exterior conditions in flexural members where h exceeds 750mm (29.5in.)
(Figure 6-9). The CSA code further limits the maximum spacing of skin reinforcing bars
to 200mm (7.9in.). The CSA values are plotted in Figure 6-6b). The 2002 ACI required
sk values exceed the CSA interior requirement at d=62in. (1575mm), and exceeds the

CSA exterior requirement at d=71in.

Figure 6-9: CSA Skin Reinforcement

244

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.3
6.3.1

Control of Crack Widths

Skin Reinforcement Study


General

The previous discussion has shown that the ACI-318-05 skin reinforcement requirements
have been considerably changed from earlier versions. A particularly significant change
has been the elimination of a minimum bar diameter, and reformulation based solely on
bar spacing. Yet common sense would dictate that it is, at best, counter-intuitive that the
bar diameter would have no effect on crack width at a constant bar spacing and cover. It
was thus decided to investigate the effects of these changes by adding skin reinforcement
to a series of specimens in the test program described in Chapter 5, as well as construct a
series of specimens to investigate the effect of skin reinforcement on the shear behaviour
of large reinforced concrete flexural elements.

6.3.2

Experimental Program

The design of the skin reinforcement is shown in Figure 6-10. Four closed steel hoops
were added to the midspan region in specimens L-10H, L-10HS, L-50N1 and L-50N2.
The hoops in L-10H and L-50N1 were constructed out of D4 deformed wire, with a
cross-sectional area of 0.04in2 (25.8mm2), and a diameter of 0.226in. (5.7mm). The
hoops in L-10HS and L-50N2 were constructed out of 15M rebar with a cross-sectional
area of 200mm2 (0.31in2) and a diameter of 16mm (5/8in.). The clear cover to the skin
reinforcement in these four specimens was 2in (50mm), thus Clause 10.6.7 of the 2005
ACI code states that a maximum spacing of 10in. (254mm) must be provided. The hoops
measured 60in. (1525mm) long, and were centered below the line of application of the
load. They were suspended in the middle of the formwork using thin steel wire at five
locations along the 60in. length of the hoops (labeled (a) in Figure 6-10). Three D4 bars
were tied to each hoop (labeled (b) in Figure 6-10), with each bar cut long enough to bear
against the sides of the formwork so as to maintain the clear cover of the hoops. The
hoops were tied tightly by steel wire to the bottom longitudinal reinforcement.
245

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

As the crack control steel was contained well within dv=1260mm from the edge of the
loading plate, it can not be expected to have a significant impact upon the shear
behaviour of specimens L-10H, L-10HS, L-50N1 and L-50N1. The crack patterns at
failure are consistent with this.
The 90o bends had an inside radius of 2in (50mm) in the 15M bars and an inside radius of
1in (25mm) in the D4 bars. The side clear cover to the flexural steel was 30mm in L10HS and 55mm in the other specimens. The skin reinforcement ratio for the specimens
reinforced with D4 deformed wire was 0.095%, and it was 0.67% for the specimens
reinforced with 15M bars. These ratios are calculated by the CSA method for calculating
sk=Abar/((db+2cc)s). They correspond to 0.07% and 0.5% based on Frantz and Breens

method of calculating sk=3Abar/((db+2cc)dsk), where dsk=4s.


An additional large specimen (L-20D) was constructed, as well as two small specimens
(S-20D1 and S-20D2). Specimen L-20D was constructed with four 30M bars at an
effective depth of 1450mm, and with 2-10M longitudinal bars spaced at 225mm over the
height of the specimen. The clear cover to this skin reinforcement was 50mm, and the
skin reinforcement extended the full length of the specimen. The skin reinforcement ratio
for this specimen was 0.4% based on the CSA method, and 0.3% based on Frantz and
Breens method. This full length skin reinforcement was hung using thin steel wires.
Taking into account the skin reinforcement in the bottom h/2 of the specimen results in
an effective depth of 1371mm and a longitudinal reinforcing ratio w=0.84% for
Specimen L-20D. The specimen as-built width was 295mm, and it was tested like the
other large specimens in this study at a span of 8100mm, resulting in an a/d ratio of 2.95.
Specimens L-20D and S-20D1/D2 will be described in more detail in Section 6.6, but
some of the results from L-20D will be used in this section.

246

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

(a)

(b)

Support of Skin Reinforcement


in Formwork

Figure 6-10: Design of Skin Reinforcement


247

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The crack control steel in the mid-span regions of L-10H, L-10HS, L-50N1 and L-50N2
was designed to investigate the effects of the bar diameter and sk, at a constant clear
cover, on flexural crack widths. Thus, crack widths in the middle 48in. (1220mm) of the
specimens were measured at various stages as they were loaded to shear failure. They
were measured visually using a crack comparator gauge with a precision of 0.05mm
(0.002in.) The zurich targets were used to confirm the visually measured horizontal
crack widths at midheight when only one crack occurred between adjacent targets. No
special adjustment was made to account for the slope of the crack. Crack widths were
calculated from the zurich target data assuming all strain measured between targets
occurred at the crack.
According to clause 12.5 of the ACI code, the development length for the 90o 15M hooks
is 6.9in. (175mm), while it is 6in. (150mm) for the D4 hooks. Cracks located outside the
middle 48in were thus not considered, so as to avoid any issues with regards to
inadequately developed reinforcement. This also reduced the likelihood that locally
increased tensile stresses at the locations of the bends would affect the crack widths.
Since the hoops were supported by steel wire, considering only the middle 48in.
eliminates any inaccuracies in the exact placing of the hoops, in case they shifted slightly
during the concrete pour.

6.3.3

Experimental Results

A summary of measured crack widths is provided in Table 6-1. Diagrams of the cracks
on the south faces of the specimens are presented in Figure 6-11 for mid-span steel
stresses of between 209-237MPa (30.3-34.4ksi). These steel stresses are reasonably
constant, hence comparison of crack patterns between specimens can be made. These
stresses are 51-52% of the specified yield stress, and thus represent a probable service
load stresses. The steel stress is calculated based on an elastic analysis of the transformed
section, in which the depth to the neutral axis, c, is calculated as:

248

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

c = kd =

(( 2

wn

Control of Crack Widths

))

+ ( w n) 2 - w n d

where w = As/bwd

n=Es/Ec

(6-9)
E c = 57,000 f c' (psi units) (clause 8.5.1, ACI-318)
E c = 4734 f c'

(MPa units)

While the actual yield strength of the flexural reinforcement was 452 MPa, (65.5ksi), in
this section, the US specified yield stress of 414MPa (60ksi) will be used.
Specimens L-10N1 and L-10N2, without skin reinforcement, exhibited wide flexural
cracking at this steel stress, with midspan cracks of up to 0.4 to 0.5mm in width (0.0160.020in). Aesthetically, these cracks are far too wide, and skin reinforcement is required.
Provision of D4 wires at 10in reduced the maximum crack widths to about 0.3 to 0.35mm
(specimens L-10H and L-50N1). While this is a decrease in maximum crack width of
about 25% and represents some improvement, in the opinion of this author, these cracks
are still too wide. This is particularly the case when considering that the stress in the
steel represents only about 3/4 of the maximum service load steel stress. Increasing the
steel stress by about one-third will increase the crack widths by a similar amount,
resulting in cracks of about 0.4 to 0.45mm.
Provision of 15M rebars at 10in. has resulted in much narrower cracks. The widest
flexural crack in the midspan region of L-10HS and L-50N2 is 0.15mm, representing a
decrease in maximum crack width of about 67% versus no skin reinforcement. The bar
diameter thus appears to have had a considerable impact on flexural crack widths.

249

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Figure 6-11: Crack Widths in Middle of Specimens at fs Ranging


from 209 to 237MPa (30.3-31.3ksi)
250

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Table 6-1: Crack Width Data


Specimen

App.
Load

Ec

(kN) (MPa) (GPa)


225 38.4 29.3
350
L-10N1
388
No skin rft. 430
475
500
160 40.3 30.0
225
L-10N2
350
No skin rft.
420
450
187 73.6 40.6
L-10H
225
D4@10in.
350
sk=0.09%
420
450
450 71.2 39.9
L-10HS
15M@10in. 700
942
sk=0.67%
1200
193
41 30.3
L-50N1
225
D4@10in.
350
sk=0.09%
420
225 40.1 30.0
L-50N2
350
15M@10in. 420
sk=0.67%
450
500
225 35.8 28.3
L-20D
350
10M@9in. 450
sk=0.39%
550
650

6.3.4

Max. Crack
Width (mm)

Crack Widths at Mid-Height (mm)


f'c

Es/Ec
6.82

(%)
0.83

0.285 0.90

6.66

0.83

0.282 0.91

4.92

0.83

0.248 0.92

5.01

1.33

0.305 0.90

6.60

0.83

0.281 0.91

6.67

7.06

0.83

0.84

0.282 0.91

0.290 0.90

fs
(MPa)
122
180
197
216
237
249
93
122
179
211
225
104
121
177
209
222
142
214
283
357
108
122
179
211
122
179
211
225
248
129
189
237
285
334

Based on Zurich Targets


Crack Widths
- - 0.217 - - - 0.347 - - - 0.373 - - - 0.418 - - - 0.436 - - - 0.480 - ---- - 0.249 - 0.171 0.339 - 0.188 0.360 - 0.228 0.396 - - - 0.100 (1)
- - 0.100 - 0.150 0.100 - 0.198 0.150 - 0.242 0.150 - - - 0.094 0.039
- - 0.169 0.077
- - 0.196 0.113
0.039 0.211 0.152
-0.058 0.038 - 0.202 0.202 - 0.258 0.302 - 0.000
-0.110 0.134 - 0.119 0.180 - 0.147 0.197 - 0.137 0.214 - ---0.099 0.125 0.145
0.136 0.192 0.207
0.195 0.282 0.287
0.204 0.336 0.341

Avg.
0.217
0.347
0.373
0.418
0.436
0.480
-0.249
0.255
0.274
0.312
0.100
0.100
0.125
0.174
0.196
0.067
0.123
0.155
0.134
0.048
0.202
0.280
0.122
0.150
0.172
0.176
-0.123
0.178
0.255
0.294

Measured Visually
Crack Widths
-0.20
--0.30
--0.35
--0.40
--0.40
--0.45
----0.15
-0.10 0.20
-0.10 0.25
-0.20 0.30
--0.10
--0.10
-0.15 0.10
-0.20 0.15
-0.25 0.15
-0.10 0.05
-0.10 0.05
-0.15 0.05
-0.20 0.10 0.05
-0.10 0.05
-0.20 0.20
-0.20 0.25
-0.05
-0.10 0.10
-0.10 0.10
-0.10 0.15
-0.10 0.15
----0.05 0.10 0.05
0.05 0.15 0.10
0.10 0.20 0.20
0.20 0.20 0.25

Avg.
0.20
0.30
0.35
0.40
0.40
0.45
0.15
0.15
0.18
0.25
0.10
0.10
0.13
0.18
0.20
0.08
0.08
0.10
0.12
0.08
0.20
0.23
0.05
0.10
0.10
0.13
0.13
-0.07
0.10
0.17
0.22

Zurich Visual
0.298
0.428
0.460
0.520
0.538
0.589
0.088
0.296
0.427
0.499
0.528
-----0.169
-----0.28
0.35
0.089
0.148
0.174
0.207
0.205
0.127
0.213
0.319
---

0.20
0.35
0.40
0.50
0.50
0.50
0.10
0.25
0.35
0.40
0.40
0.10
0.10
0.25
0.30
0.35
0.15
0.15
0.20
0.25
0.05
0.15
0.25
0.30
0.10
0.15
0.15
0.15
0.20
0.10
0.15
0.25
0.30
0.30

Crack Widths as a Function of Steel Stress

Maximum Crack Widths

The maximum single crack widths measured at any point in the 48in. wide midspan
regions of specimens L-10N1 + L-10N2 (no skin reinforcement), L-50N1 (D4@10in.), L50N2 (15M@10in.) and L-20D (10M@8.9in.) are plotted in Figure 6-12a) versus the
steel stress at midspan. The maximum single crack widths measured at any point in the
48in. wide midspan regions of specimens L-10H (D4@10in.) and L-10HS (15M@10in.)
are plotted in Figure 6-12b). The crack widths in Figure 6-12 were measured visually
using a crack comparator gauge at load stages. The steel stresses were calculated based
on an elastic analysis of the section at midspan. Since all specimens in Figure 6-12 other
251

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

than L-20D and L-10HS failed prior to the steel stresses reaching the maximum service
load steel stress of 67% of fy, linear extrapolations have been added to the data. Also
shown is the ACI 318-99 maximum crack width of 0.016in. (0.4mm) for interior
exposure. ACI 318-99 required that crack widths not exceed this width at service loads.

Figure 6-12 Maximum Crack Widths in 48in. Wide Midspan Region, Measured Visually
It can be clearly seen in Figure 6-12 that unacceptably wide cracks occurred in L-10N1
and L-10N2 under service loads. For example, at the maximum service load steel stress
(0.67fy) the widest crack in L-10N1 would be expected to be about 0.6mm (0.024in.).
Indeed, the maximum crack widths in L-10N1 and L-10N2 reached 0.4mm (0.016in.) at a
steel stress of about 50% of fy. Clearly, members of this depth are in need of additional
reinforcement to control crack widths at service loads.
However, the provision of D4 bars spaced at 10in. did not provide adequate control of
crack widths. Based on the linear extrapolations in Figure 6-12, it is expected that the
maximum crack width in both L-50N1 and L-10H would reach 0.45mm at a steel stress
252

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

of 67% of the yield stress. Unacceptably wide cracks thus occurred in the midspan
regions of L-50N1 and L-10H, even though these regions contained skin reinforcement
designed to conform to ACI 318-05.
Using 15M bars spaced at 10in. successfully controlled crack widths in specimens L50N2 and L-10HS. Compare the crack widths measured in L-50N2 to those measured in
L-50N1, and the crack widths measured in L-10HS to those measured in L-10H. The
maximum measured crack width in L-10HS was 0.25mm at a steel stress of 31.8ksi
(357MPa), representing 86% of the yield stress. Using 10M bars at 8.9in. also
successfully controlled crack widths in L-20D. Any differences in the crack widths
between L-20D and L-50N2 fall within the precision of the crack comparator gauge.
Crack Widths at Midheight

The averages of the crack widths at the midheight (h/2) of the 48in. wide midspan regions
are plotted in Figure 6-13 and Figure 6-14 versus the midspan steel strains. The averages
of the crack widths measured using the comparator gauge are plotted in Figure 6-13,
while the average based on the zurich target data are plotted in Figure 6-14.
The maximum crack widths in the middle 48in. of the specimens did not occur at midheight, but, rather, were found to occur at between 450mm and 670mm (18-26.5in.) from
the bottom face. Nevertheless, the average crack widths at the midheight of the beams
without skin reinforcement were unacceptably wide. Provision of D4 skin reinforcement
reduced the average crack widths at mid-height, though the 15M skin reinforcement was
far more effective.
The analysis of crack widths plotted in Figure 6-12, Figure 6-13 and Figure 6-14 has
shown that that the D4 skin reinforcement had only a small impact on flexural crack
widths. Crack widths were still unacceptably wide in the specimens with D4 skin
reinforcement, even though this skin reinforcement met the requirements of ACI 318-05.
Skin reinforcement consisting of 15M bars at the same spacing and bar cover, however,
was extremely effective at controlling crack widths.
253

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Figure 6-13 Average Crack Width at Midheight of Midspan Region, Measured Visually

Figure 6-14 Average Crack Widths at Midheight Midspan Region, -Zurich Target Data
254

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.3.5

Control of Crack Widths

Crack Widths at Maximum Service Load Steel Stress

While crack widths below the maximum service load are useful to investigate, the crack
widths at the ACI maximum service load steel stress of 0.67fy are the most useful. Since
a number of the specimens shown in the previous figures failed at loads below those
corresponding to a steel stress of 0.67fy, it is necessary to estimate the crack widths at
fs=0.67fy had the beams not failed. This was accomplished using the extrapolations
shown in Figure 6-12, and the resulting estimated maximum crack widths at fs=0.67fy are
shown in Figure 6-15. Extrapolated crack widths for specimens L-20N1, L-20N2, L40N1 and L-40N2 are also shown. Using an extrapolation is an appropriate method for
estimating the maximum crack width at fs=0.67fy since a stable crack pattern formed in
the specimens, in which existing cracks widened with increasing load, with few new
cracks forming.
Figure 6-15 clearly shows that at a constant clear cover and a constant spacing of 10in.,
decreasing the skin reinforcement ratio by decreasing the bar diameter can result in
unacceptably wide cracks at service loads. At a constant spacing and cover, the bar
diameter thus has a very clear effect on flexural crack widths.

Maximum Expected Crack Width at fs=0.67fy

0.70

0.028

L-10N Series
L-20N Series
L-40N Series

(mm)

0.60

(in.)

0.024
Range in Crack
Widths at sk=0

0.50

0.02

c=2in (50mm)

0.40

0.016

0.30

0.012

0.20

0.008

0.10

Normal f'c
Series
(L-50N Series, L-20D)

0.00
0

0.4

0.004

High f'c
Series
(L-10H, L-10HS)
0.8

1.2

0
1.6

Skin Reinforcement ratio, sk

Figure 6-15: Expected Crack Widths at fs=0.67fy


255

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The concrete strengths in the specimens without skin reinforcement ranged from 28.141MPa (4075-5945psi), resulting in neutral axis depths of 393-426mm (15.5-16.8in.).
Thus, some of the scatter in crack widths at sk=0 might be expected to have occurred due
to differences in the concrete strength. An increase in concrete strength is associated
with a decrease in neutral axis depth, resulting in slightly larger longitudinal strains
below the neutral axis and consequent crack widths at otherwise constant crack spacings
and steel strains. While the differences in concrete strengths between the specimens
would result in only very small differences in longitudinal web strains, this effect is
evident in the data at sk=0 and sk=0.095%, in that there is a general trend of increasing
crack widths with increasing concrete strengths. An opposite effect is noted, however, at
sk=0.67%. Other researchers, however (for example, Hognestad (1962)) have not found

that concrete strength affects crack widths beyond the initial cracking load.
In specimens of similar concrete strength, the maximum expected crack width decreases
by about 20% in specimen L-50N1 (reinforced with D4 wires) relative to specimens L10N1 and L-10N2 (with no skin reinforcement). Nevertheless, it is instructive to note
that the crack widths in the specimens reinforced with D4 wire fall within the scatter of
maximum crack width data from specimens without skin reinforcement. It thus
interesting to note that provision of D4 skin reinforcement at a spacing of 10in. was about
as effective at reducing crack widths as reducing the concrete strength by about 30%.

6.3.6

Predictions of Web Crack Width

Crack Width Profiles

The widths of the single widest flexural crack in the six specimens shown in Figure 6-11
are plotted in Figure 6-16. Using the cracking model developed by Frosch (2002) and
described by Equation 6-8 and in Figure 6-7, it is possible to predict the measured crack
width profiles shown in Figure 6-16.

256

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The predicted widest (=2) and narrowest (=1) cracks in specimens L-10N1 and L10N2 are shown in Figure 6-16a) and b). It can be seen that Froschs model is generally
good at calculating the overall width profile of the widest cracks in these specimens.
While the maximum crack width was observed to occur lower in the section than what is
predicted by Froschs model, this maximum width was calculated reasonably accurately
using =2 in Equation 6-8.
The predicted crack width profiles for the specimens with D4 reinforcement are shown in
Figure 6-16c) and d). The predicted crack width profiles for the specimens with 15M
reinforcement are shown in Figure 6-16e) and f). Two sets of predictions have been
generated. The first set, shown in solid lines, is for the case where the controlling cover
distance, dx* in Equation 6-8, is calculated based on skin reinforcement provided at
10inch spacing. Since all four specimens in Figures 6-18c)-f) have skin reinforcement
present, these are the predictions that apply to these beams. The second set, shown in the
dashed lines, was generated neglecting the skin reinforcement, and resemble the crack
width predictions for L-10N1 and L-10N2. The predicted effect of reducing the
controlling cover distance, dx*, by providing skin reinforcement can be clearly seen by
comparing these two sets of predictions. It is predicted (Figure 6-7) that cracks will
locally widen between skin reinforcing bars, but the overall width is predicted to be
considerably reduced.
It can be seen in Figure 6-16e) and f) that the measured crack widths in the specimens
with 15M skin reinforcement are considerably smaller than those measured in L-10N1
and L-10N2 (Figure 6-16a) and b)). For these two beams, calculating dx* based on the
presence of the skin reinforcement resulted in accurate predictions of the maximum crack
width from Froschs model. While the precision of the crack comparator gauge did not
allow for the detailed mapping of the variation in crack widths over the height of
specimens L-10HS and L-50N2, the maximum crack width was accurately calculated
using s=2.

257

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Figure 6-16: Widest Crack in Midspan Region and Crack Width Predictions by Eq. 6-8
258

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Figure 6-16c) and d) indicate that calculating dx* based on the presence of skin
reinforcement did not accurately predict the crack widths and overall crack width profiles
in the two specimens with D4 skin reinforcement. For L-10H and L-50N1, the measured
crack widths fell well outside the band of expected crack widths defined by 1<s<2. For
example, the maximum measured width in these specimens was 0.3mm, and in both this
maximum width was observed to occur directly beside a skin reinforcing bar, where
Froschs model predicts a local narrowing of the crack. By observation, in fact, it can be
seen that the predicted crack width profiles generated neglecting the presence of the D4
bars (the dotted lines) more accurately predict the measured crack width profiles.
When using Froschs model to predict the flexural crack widths in the two beams with
D4 skin reinforcement, it appears that it is more accurate to neglect the skin
reinforcement than it is to consider it in the model.
Experimentally Determined s Values

At each load stage, it is possible to calculate an experimentally-determined s value. It


can be calculated by dividing the maximum crack width measured at any point on the
crack by the largest width on the predicted crack width profile generated using s = 1.
Separate experimental s values can be determined by both considering and neglecting
the skin reinforcement.
To further examine how the crack widths measured in L-10H and L-50N1 lie outside the
range of crack widths predicted by Froschs model, Figure 6-17 was created. Each data
point represents the average of experimentally-determined s values for load stages
where the steel stress exceeded one-third of the yield stress. This limit on the steel stress
was chosen to ensure a stable crack pattern had formed. This figure shows that the
maximum measured crack widths in specimens with sk=0.4% and 0.67% lie within the
scatter band defined by Frosch, with experimental s values ranging from 1.3 to 1.6.
Likewise, the maximum measured crack widths for sk=0% lie within the scatter band.
259

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The maximum measured crack widths in the specimens reinforced with D4 wire lie well
outside the expected range of s values, with s values of 2.4 and 2.5. However, the key
to this figure is to note that the maximum measured crack widths in these specimens
actually fall within the range of s values generated assuming no skin reinforcement. In
this case, the experimental values of s are both 1.23. Thus, when using Froschs model
to predict the crack widths in the specimens with D4 skin reinforcement, assuming that
no skin reinforcement is present produces more accurate predictions of the maximum
crack width than assuming skin reinforcement to be present.

Experimental Crack Spacing Factor, s

4.0

3.5
Assuming Skin Reinforcement Provided
3.0

2.5
2.4

2.5

Range of s in Frosch Model

2.0

1.5

1.23

1.0

0.5
Assuming no Skin Reinforcement Provided
0.0
0.0

0.2

0.4

0.6

0.8

1.0

Skin Reinforcement Ratio, sk (%)

Figure 6-17: Experimentally Determined Values of s


260

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.3.7

Control of Crack Widths

Another Look at Frantz and Breen

Average and Maximum Crack Widths

Frantz and Breen tested a total of 44 inverted-T specimens, studying numerous


combinations of web depths, flange depths, reinforcement ratios w and sk, web widths,
bar diameters and clear covers. The majority of them were reduced segment specimens,
which were 72in. (1830mm) long, and which were loaded using hydraulic rams to
directly apply tension to the main reinforcing bars and compression above the neutral
axis. The location of the compression force was determined by an analysis of the cracked
transformed section. The specimens were thus loaded in pure moment, and were tested
upside-down to facilitate observations of the cracks.
Despite the impressively large number of specimens in the test program, only one series
of five specimens was tested in which all variables were kept constant, with only db being
varied. Results from this series of tests are summarized in Figure 6-18, in which average
and maximum web crack widths are plotted as a function of steel stress. In these
specimens, the centre-to-centre bar spacing, clear cover and section dimensions were kept
constant, while the skin reinforcement was varied from Swedish Grade 77 deformed
6mm bars (Specimen A-7) to No. 3 bars and No. 4 bars (Specimens A-8 and A-9). These
bars resulted in sk values of 0.34%, 0.81% and 1.34%. Two specimens, A-1 and A-2,
were tested without skin reinforcement.
In reviewing the data summarized in Figure 6-18, Frosch (2002) notes that there was
essentially no difference between the measured crack widths in the specimens reinforced
with No. 3 bars and No. 4 bars, and thus concludes that any bar size can be successfully
used to control crack widths. The considerably wider cracks in the specimen reinforced
with 6mm bars were attributed by Frosch to the fact that the bars were Swedish, with
deformations that were less pronounced than the No. 3 and 4 bars, and which, according
to Frosch, did not conform to the ASTM A615 standard on deformations. Thus, the
increase in crack widths was attributed by Frosch to differing bond properties between
the Swedish and US bars. The fact that smaller deformations can be used on smaller bars
261

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

because of the larger ratio of perimeter to cross-sectional area (and thus inherently better
bond properties) is not addressed. Also note that the accuracy of the gauge used to
measure the crack widths in Frantz and Breens experimental program was 0.001in.
Hence, the maximum crack widths in A-9 (No. 4 bars) were larger than those in A-8 (No.
3 bars) by only one increment on the gauge.
250

(MPa)

290

170
20
-3

0.50

250

(mm)

(x10 in.)

(mm)

210

(MPa)

290
20
No Skin Rft.

-3

210

(x10 in.)

170

0.50

16

16
0.40

No Skin Rft.
sk=0%

12

0.30

6mm Bars
sk=0.34%

0.20

Crack Width

Crack Width

0.40

6mm Bars

12
0.30
#3 Bars

8
0.20
#4 Bars

0.10

#3 Bars
sk=0.81%

#4 Bars
sk=1.41%

b) Maximum Web Crack Width

a) Average Web Crack Width

0.00

0
25

30

35

Bar Stress

0.10

40 (ksi)

45

Crack Width Gauge


Resolution=0.001in.

0.00

0
25

30

35

40 (ksi)

45

Bar Stress

Figure 6-18: Effect of Bar Diameter on Crack Width in Web (Frantz and Breen, 1976)
Crack Magnification Ratio and Extension of Cracks into the Web

The ratios of the average web crack width (wweb) to the average crack width at the level
of the steel (wsteel) for all of Frantz and Breens specimens are plotted in Figure 6-19a) for
a steel stress of 35ksi. Frantz and Breen refer to wweb/wsteel as the crack magnification
ratio. The crack magnification ratios for specimens A-1, A-2, A-7, A-8 and A-9 have
been identified with open symbols. The crack magnification ratios for specimens A-7, A8 and A-9 are plotted in Figure 6-19b) at steel stresses of 25, 30, 35 and 40ksi. Figure
6-20 plots the percentage of the cracks at the level of the steel that extended into the web
in specimens A-1, A-2, A-7, A-8 and A-9, and crack diagrams have been reproduced.
262

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

It can be seen in these figures that, in fact, using #4 bars instead of #3 bars affected
cracking in the webs in two important ways. Firstly, the crack magnification ratio was
lower in A-9 than in A-8 for all steel stresses. Although crack widths in the webs of A-8
and A-9 were similar, cracks at the level of the steel in A-9 were, on average, about 15%
wider than they were in A-8. The use of #4 bars reduced their widths in the web to
widths that were similar to the widths measured in A-8, resulting in the reduced crack
magnification ratios. It is interesting to note that a curve of best fit drawn through the
cloud of data points in Figure 6-19a) might be drawn almost right on top of the curve
connecting the points for A-7, A-8, A-9 and the average of A-1 and A-2.
Secondly, and more importantly, a considerably greater number of cracks extended into
the web in specimen A-9 than in A-8. Frantz and Breen report that twenty-eight cracks
occurred at the level of the steel in specimens A-7, A-8 and A-9 (counted on both sides of
the specimen), while twenty-six were counted in A-2 and nineteen in A-1. Sixteen of
these cracks extended into the web in A-9, versus ten in A-8, nine in A-7 and six in
specimens A-1 and A-2. The crack patterns reproduced in Figure 6-20 clearly indicate
how closely spaced the cracks were in A-9 as compared to other specimens. This
behaviour corresponds with Frantz and Breens findings with regards to all of their
specimens, whereby the percentage of flexural cracks that extended into the web
increased in direct proportion to the total area of skin reinforcing bars provided.
The fact that A-9 had 60% more cracks in the web than A-8, yet had web cracks that
were similar in width, leads to the conclusion that the longitudinal strain in the web was
considerably greater than it was in A-8. Specimens A-1 through A-9 were all loaded in a
similar manner, with the location of the hydraulic ram modified based on an elastic
analysis of the cross-section. Thus any differences in the strain profiles would be due to
differences in the locations of the neutral axes due to differences in the concrete strengths.
The concrete strength of A-9 was 5231psi (36.1MPa) and was greater than the concrete
strength of A-8, which was 4580psi (31.6MPa). However, this difference resulted in a
neutral axis depth in A-9 that was 97% of the neutral axis depth in A-8, an
inconsequential difference.
263

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

4.0

Control of Crack Widths

2.5

a) fs=35ksi

wweb
wsteel
3.5

b) Beams A-7, A-8 A-9

wweb
wsteel

No Skin Rft. (A-1, A-2)

Other
Specimens

6mm Bars
(A-7)

2.0

3.0
6mm bars
(A-7)

#3 Bars
(A-8)

1.5

Cross-Section

2.5

1.0

#4 Bars
(A-9)

#3 Bars (A-8)

2.0

#4 Bars (A-9)

0.5

1.5

Note: Frantz and Breen (1976) present


crack width data for fs=25, 30, 35 and 40ksi.
Data not available for A-9 at fs=35ksi.
0.0

1.0
0

0.5

1.5

2.5

20

25

Skin Reinforcement Ratio, sk (%)

30

35

40

45

Steel Stress, fs (MPa)

Figure 6-19: Effect of Bar Size on Crack Magnification Ratio (Frantz and Breen, 1976)

Percent of Cracks at Level of Steel Extending Into Web

80%

fs=35ksi

70%

#4 Bars

60%
#3 Bars

50%
6mm bars

40%

30%

20%

10%
0.00

Vetical spacing
between bars = 4.125in.

No Skin Rft.

0.20

0.40

0.60

0.80

1.00

1.20

1.40

Skin Reinforcement Ratio, sk (%)

Figure 6-20: Effect of Bar Size on Crack Extension into Web


264

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

It is not entirely clear why, with considerably more cracks in the web of A-9 than in A-8,
the average web crack widths were measured to be about the same. There is an inherent
scatter associated with crack widths in the webs of concrete members, and it is a
considerable challenge to measure crack widths accurately and uniformly. However, it is
a far more straight-forward exercise to simply visually identify the locations of cracks
and to mark them with pens. The crack diagrams shown in Figure 6-20 can thus be
regarded as a very accurate representation of the effect of increasing the skin
reinforcement bar diameter to a No. 4 from a No. 3, without the inherent variability
associated with visual crack width measurements.
There is a considerable amount of scatter in the data presented in Figure 6-19a), due in
part to the large number of variables considered in the experimental program. There is
even a considerable amount of scatter in crack magnification ratios between nominally
identical specimens tested with sk =0.22%, and this is illustrative of the type of scatter
inherent in the phenomenon of cracking in concrete. However, the data in this figure
indicates that a small amount of crack control reinforcement may have a significant
impact on web crack widths, but that this effect is unreliable. As the value of sk is
increased, web crack widths decrease, but at a decreasing rate. The behaviour of the
cracks exhibited in specimen A-9, in which the majority of cracks at the level of the steel
extend well into the web, is far preferable to the behaviour of the cracks in the other
specimens, including those in A-8. In the other specimens, most of the cracks at the level
of the steel coalesced into a smaller number of wider cracks in the web.
The data in Figure 6-19a) is reproduced in Figure 6-21 based on the CSA method of
calculating sk. It can be seen that beyond a skin reinforcement ratio of about 1%, cracks
in the webs tended to stay narrower than 1.7 times the widths of the cracks at the flexural
steel. One notable exception is specimen A-6, which contained skin reinforcement
consisting of one No. 6 (19mm diameter) rebar on either side of the specimens spaced at
13.375in. (340mm) from the tensile steel. The spacing of this single bar exceeded the
maximum spacing of 12in. (300mm) allowed in the ACI 318-05 skin reinforcement
265

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

provisions by only 1.375in (35mm), yet considerably wider cracks occurred in the web
than at the level of the steel. Commenting on the cracks in this specimen, Frantz and
Breen note that the bars in A-6 are located too far away from the crack development
zoneto significantly influence the crack formation.
4.0

wweb
wsteel

sk =

Abar
(2c + db )ssk

3.5

A-6

3.0

2.5

sk>1%
2.0

1.5

1.0
0

0.5

1.5

2.5

Skin Reinforcement Ratio, sk (%)

Figure 6-21: Effect of sk on Crack Magnification Ratio

6.3.8

Suggested Modifications to ACI Code

The experimental program described in this chapter has shown that the bar size has a
clear effect on crack widths in the webs of thick flexural elements. It seems that, at the
very least, a minimum bar diameter must be specified in clause 10.6.7 of the ACI code
(Figure 6-8). Describing commonly-used bar sizes in the commentary is not sufficient.
The crack widths measured in the current experimental program were successfully
controlled using both 10M bars at 225mm (about equivalent to #3 bars at 9in.), resulting
in an sk value of 0.4%, and 15M bars at 255mm (about equivalent to #5 bars at 10in.),
266

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

resulting in an sk value of 0.68%. Furthermore, Froschs model generally was accurate


at predicting the maximum crack widths for these specimens.
Referring to Figure 6-21, it can be seen that specimens in the Frantz and Breen study
exhibited relatively narrower cracks in the webs for sk values exceeding about 1%, so
long as the vertical spacing between the skin reinforcing bars was not too great. Based
on specimen A-6, it appears that the maximum spacing of 12 in. for skin reinforcing bars
may be too large.
It is suggested that the ACI 318-05 code require at least a No. 5 rebar for skin reinforcing
bars, and that the vertical spacing between layers of skin reinforcement, s, be calculated
as follows, with the maximum spacing being 10 inches (254mm) for a steel stress of
40ksi, rather than 12 inches (300mm):
s=

15(40)
- 2.5c c 10(40 f s )
fs

(6-10)

267

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.4
6.4.1

Control of Crack Widths

Effect of Crack Control Steel on Shear Strength


General

It has been suggested that the ACI codes inability to account for the size effect in shear
can be partially or fully addressed by the use of skin reinforcement. For example, narrow
beams with sufficiently large skin reinforcing bars such that cracks remain narrow and
closely spaced within the web have been shown to exhibit higher shear strengths (Collins
and Kuchma (1996)). However, the preceding sections have shown that deep sections
with skin reinforcement designed in conformance with the ACI code may still exhibit
wide cracks, since the code does not mandate a minimum bar diameter or skin
reinforcement ratio. Furthermore, wide slabs with a large width to depth ratio may still
experience wide cracking well away from the skin reinforcement. The 1978 CEB-FIP
code, for example, suggests that a rebar controls crack widths only within 7.5db from the
centre of the bar.
As discussed in Chapter 2, the 2004 CSA code suggests that the crack spacing parameter,
sz, can be taken to be equal to the vertical spacing between layers of distributed
longitudinal reinforcement, if the total area of steel in a layer (Alayer) exceeds 0.3% of bwsz.
The reduced value of sz can be used instead of 0.9d or 0.72h, and is designed to model the
effect that distributed longitudinal reinforcement has on the spacing of cracks in the web.
Specifically including bw when calculating the required area of steel requires the designer
to specify larger bars for wider members if it is desired to use the reduced value of sz. To
be effective, the bars in each layer must be spaced closer than 600mm (24in.) horizontally
(Lubell et. al. (2004)). In the context of calculating sz, this distributed longitudinal
reinforcement is referred to as crack control reinforcement, and this is a more accurate
description of the role that this reinforcement plays.
Previous tests of beam specimens with crack control steel reported by Collins and
Kuchma (1999) involved values of Alayer/bwsz of between 0.61% and 1.19%, and there are
no tests to be found in the literature on thick beams with only just the minimum quantity
268

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

of crack control reinforcement to use the reduced value of sz. Furthermore, a number of
the larger specimens with crack control steel reported by Collins and Kuchma failed at
ratios of vexp/vCSA of less than 1. The average value of vexp/vCSA for the eight tests
(including three repeated tests) with an effective depth of 920mm or greater was 0.9, with
the lowest value being 0.77.
It was thus decided to investigate the shear behaviour of a large beam specimen with
distributed longitudinal steel designed to just meet the 0.003bwsz requirement. It was also
decided to investigate whether this quantity of crack control steel could control crack
widths to such an extent that the size effect could be essentially eliminated. In order to
simplify discussion, it is proposed that a new parameter, d, be defined, where
d=Alayer/bwsz, where Alayer is the area of steel in a layer of crack control reinforcement.

6.4.2

Experimental Program

The specimens are described in Figure 6-10. Specimen L-20D was designed to be
300mm wide, 1510mm high and 9000mm long. It was constructed with 4-30M rebars at
a depth of 1450mm from the top face of the specimen, and six layers of two 10M
longitudinal bars spaced at 225mm over the height of the specimen. The as-built width
was 295mm, resulting in a d value of 200/(295 x 225)=0.301%, and this is almost
exactly the minimum required value of d to use a reduced sz value.
The CSA code and AASHTO-LRFD both specify that the area of flexural tension
reinforcement, As, includes all longitudinal reinforcement in the flexural tension half of
the member. Counting the layers of 10M crack control steel below h/2 results in an
effective depth of 1370mm, an area of steel of 3400mm2, a reinforcement ratio of 0.84%
and an a/d ratio of 2.96. Cross sectional properties were chosen such that the
reinforcement ratio was similar to the ten equivalent specimens that were tested as part of
the aggregate interlock study which had a reinforcement ratio of 0.83%.

269

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The specimen was cast at the same time as L-20L (discussed in Chapters 5 and 7), with
normal strength concrete and a maximum aggregate size of 19mm (3/4in.). Because of
the use of crack control steel, the 2004 CSA code states that sz can be taken as the vertical
spacing between the layers of steel, namely 225mm. The crack control steel was hung in
place with a series of small-gauge steel wires tied to the bars and to supports spanning
over the top of the forms. Photographs of the cage under construction are provided in
Figure 6-22.

Figure 6-22: Photographs of L-20D Cage Under Construction


The 225mm centre-to-centre spacing of the crack control steel is 12.5% greater than the
200mm maximum spacing in the CSA code. However, the ACI code maximum spacing
was 254mm (10in.). To make the specimen applicable to the study of both codes, the
average of the code spacings was used.
Specimen L-20D was designed to be shear critical, and was tested in three point bending
under the Baldwin test frame in the Huggins structures laboratory in the Department of
Civil Engineering at the University of Toronto. The test setup was identical to that used
to test the other L-series of specimens described in Chapter 5 of this thesis, with
270

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

associated zurich targets and LVDTs to measure vertical deflection and shear strains.
After the specimen failed on the east side, this side was clamped together with a series of
externally placed Dywidag bars, and the beam was retested to failure on the west side (L20DR).
Two small specimens, S-20D1 and S-20D2 were constructed at the same time as L-20D
to represent small members with identical effective crack spacing parameters, sxe. They
nominally measured 96mm wide, 280mm high and 1800mm long, and were reinforced
with two 10M rebars at an effective depth of 250mm. They were loaded to failure in
three point bending under a displacement-controlled MTS actuator in a setup similar to
that used for the S-Series of specimens described in Chapter 5, with associated LVDTs to
measure midspan vertical displacements and shear strains (Figure 6-23).
Specimens S-20D1 and S-20D2 were supported on 30mm wide supports spaced 1480mm
apart, had a nominal reinforcement ratio of 0.83% and were constructed with normal
strength concrete and a maximum aggregate size of 19mm (3/4in.). They had an sxe value
of 0.9(250) = 225mm.

Figure 6-23: Test Setup Specimens S-20D1 and S-20D2


Specimens S-20D1, S-20D2 and L-20D were each designed to have identical sxe values
and identical a/d and w values. Thus, the CSA code predicts that each specimen should
fail at identical values of =V/bwd(fc)0.5 and hence that there should be no size effect
between the large and small specimens.
271

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.4.3

Control of Crack Widths

Experimental Results

As built dimensions and key experimental results are presented in Table 6-2. Average
experimental results from the ten equivalent L-series specimens without crack control
steel are provided for comparison. Load vs. mid-span displacement curves are presented
in Figure 6-24. Also shown in Figure 6-24 are the failure loads predicted by the
simplified MCFT and ACI Eq. 11-5.
The average failure shear force of L-20D and L-20DR was 356kN, which is 32% greater
than the average failure shear force of the ten equivalent specimens without the crack
control reinforcement.
The behaviour of L-20D under load was characterized by the formation of cracks in the
web that were narrower and more numerous than those formed in the ten equivalent Lseries specimens without crack control steel. See, for example, Figure 6-25, which
compares the cracking patterns of L-20D and L-20N1. It can be seen that, for L-20N1,
the three widest cracks measured 0.4mm, 0.6mm and 1.0mm when the load was 500kN,
whereas for L-20D the three widest cracks all measured 0.3mm when the load was
650kN. These loads correspond to midspan steel strains of 1250 and 1670
respectively, based on an elastic analysis of the cracked sections. Furthermore, note that
the average longitudinal spacing of the cracks at midheight of the beams decreased from
about 700mm (0.5d) to 400mm (0.3d), and the number of cracks at midheight increased
from seven to thirteen. It can thus be seen that at considerably higher steel strains, the
crack control steel caused narrower, more numerous cracks to form in specimen L-20D.
The crack control steel in L-20D also resulted in final failure cracks that were flatter than
those in the specimens without crack control steel. The crack angle calculated by the
SMCFT (Eq. 2-22) is shown in Figure 6-25, and it can be seen that this equation
provides a good estimate of the actual crack angle in these specimens.

272

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Table 6-2: As-Built Properties and Experimental Results, L-20D and S-20D Series
Specimen Properties (as-built)
Specimen

bw
d
(mm) (mm)

a/d

Experimental Observations

w f'c(1) ag,eff sxe Pexp Vexp(2) vexp ult/0.5L ult s,max wmax
-3
(%) (MPa) (mm) (mm) (kN) (kN) (MPa) x10-3 x10
(mm)

S-20D1
S-20D2

98
96

252
250

2.94 0.81 40.6


2.96 0.83 40.8

19
19

227
225

49.5
51.6

24.8
25.8

1.00
1.08

4.5
4.3

0.87 2160 0.15


0.43 2340 0.15

L-20D
L-20D-R

295
295

1370 2.96 0.84 35.8


1370 2.96 0.84 35.8

19
19

225
225

668
689

350
361

0.87
0.89

2.8
--

0.60 1620 0.30


0.61 1660 0.35

L-Avg(3)

300

1400 2.89 0.83 39.5 0-51

1071509
2756

270

0.64

2.1

0.65 1255 0.53

Notes:
(1) day of test
(2) Calc. at d from face of loading plate, incl. self-weight (vexp = Vexp/bwd)
(3) Average of L-10H, 10N1, 10N2, 20N1, 20N2, 40N1, 40N2, 50N1, 50N2, 50N2-R

60
ACI -D1

ACI

800

ACI -D2
2-10M @ 225

700

SMCFT
2-10M

Failure Load
L-20DR

w=0.82%

50

4-30M

w=0.84%

600

Applied Load (kN)

Applied Load (kN)

40
500
L-20D

400

300

SMCFT -D1
SMCFT -D2
S-20D2

30
S-20D1

20

200

10
100
L/480

L/800

0
0

10

Mid-Span Deflection, (mm)

12

0.0

1.0

2.0

3.0

4.0

Mid-Span Displacement, (mm)

Figure 6-24: Load vs. Midspan Displacement Curves, L-20D, S-20D1, S-20D2

273

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Figure 6-25: Failure Crack patterns in L-20D and L-20N1


274

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

It is interesting to note that the ratio of /0.5L to the shear strain, ult, in L-20D was 21%,
while the average for the specimens without crack control steel was 31% (with the lowest
value being 23%). Less of the total mid-span displacement was caused by shear strains in
L-20D and it exhibited a higher shear stiffness than the specimens without crack control
reinforcement. See Figure 6-26, in which the measured shear stress vs. shear strain
curves at the quarterspans are plotted for specimens L-20DR and L-20N2.
1.0
L-20D Failure
on East End

0.9

L-20DR

Shear Stress, v (MPa)

0.8
0.7
L-20N2
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0

0.2

0.4

0.6

0.8

Shear Strain (mm/m)

Figure 6-26: Shear Stress vs. Shear Strain for Specimens L-20DR and L-20N2,
Measured at Quarterspan
The stiffer response of L-20D compared to specimens without crack control steel can also
be seen by comparing the midspan deflections. At an applied load of 500kN, the
measured midspan deflection for L-20D was 7.2mm, and this is 22% less than the
average measured midspan deflection for L-20N1 and L-20N2 at the same load. Yet the
transformed moment of inertia for L-20D was 94,300x106mm4, versus an average value
of 95,500x106mm4 for L-20N1 and L-20N2.
275

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.4.4

Control of Crack Widths

Effect of Dowel Action and Aggregate Interlock

It is suggested that the considerable increase in shear strength exhibited in L-20D and L20DR was primarily a result of the formation of narrower cracks that were better able to
transfer shear due to aggregate interlock. A portion of the increase can also be attributed
to increased dowel resistance.
Bhide and Collins (1987) note that a large number of small size bars offers a larger dowel
resistance than a small number of large bars, for the same reinforcement ratio and dowel
displacement, where the dowel displacement is calculated as the component of the
crack width in the direction at right angles to the reinforcing bar under consideration.
Put another way, a large number of small bars offers a stiffer response than does a small
number of large bars. Thus, beams in which some of the longitudinal steel area
concentrated in the bottom has instead been distributed over the height would be
expected to exhibit a stiffer shear response, and this behaviour was observed in L-20D
and L-20DR. It would also be expected that a greater proportion of the total shear at a
section in such a beam would be carried by dowel action, for three reasons:
1) the cracks in the web are at a flatter angle than at the level of the main reinforcement,
with a larger component of the crack width oriented in the vertical direction. Analysis
of the cracks at the level of the main reinforcement in all the large specimens tested
indicated a wide range of crack angles ranging from perfectly vertical to identical to
the angle of the crack in the web. In general, however, the cracks in the web were
observed to be flatter than the cracks at the level of the main reinforcement.
2) cracks in the web are wider than they are at the level of the main steel. This would
only be the case if the longitudinal crack spacing in the web was not reduced to a
degree such that crack widths reduced from a maximum at the steel to zero at the tips
of the cracks.

276

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

3) reduced longitudinal stiffness in the main reinforcement results in wider cracks at the
level of the reinforcement, with the components of the widths in the vertical direction
increasing as a consequence.
At a load of 650kN, a total of eight 10M crack control bars crossed the critical shear
cracks in L-20D and L-20DR. Combined with the four 30M bars in the bottom of the
specimen, this resulted in a total cross-sectional area of 3600mm2 of steel crossing the
crack that was available to resist shear by dowel action. This cross-sectional area is 3%
greater than the area of steel crossing the crack in the specimens without crack control
steel, but more importantly there was a total of twelve bars crossing the crack as opposed
to five.
Bhide and Collins (1987) derived a series of expressions to estimate the dowel force in a
bar crossing a crack of width w. The dowel force, Fd, versus dowel displacement, ,
relationships for a 10M and a 30M bar are shown in Figure 6-27, in which the cracking
pattern in L-20D at P=650kN has been reproduced. The measured crack widths at a
single crack are shown as well. This crack was inclined at about 60o, and measured
0.25mm wide over most of its height. At the level of the main reinforcement, it was
about 0.15mm wide. Based on comparisons with crack patterns in the other L-series of
specimens, had this beam not contained crack control steel, it likely would have failed at
this crack at a load that was considerably smaller than 650kN.

Figure 6-27: Analysis of Dowel Action in Specimen L-20D


277

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

The component of the crack widths in the vertical direction were 0.125mm and 0.075mm
in the web and at the main reinforcement, respectively. These dowel displacements
resulted in dowel forces of 2.5kN and 5.9kN in the 10M bars and 30M bars, respectively,
and a total of about 8 x 2.5 + 4 x 5.9 = 44kN of shear being transferred by dowel action at
the section. This represents 13% of the total shear at the section. Had five 30M bars
been placed in the bottom layer without crack control steel, and the same crack pattern
formed with identical inclinations and crack widths, the shear transferred by dowel action
would have been about 5 x 5.9 = 30kN, representing 9% of the total shear at the section.
Indeed, a slightly narrower crack might be expected to have formed at the level of the
steel due to the presence of an additional 30M bar, reducing the dowel forces slightly.
We thus see that increased dowel action may have accounted for about 44kN30kN=14kN of the total additional shear of (0.5x650 0.5x500) = 75kN. This represents
19% of the total increase in shear force. The geometry of the cracking patterns in L-20D
and L-20DR generally resembled the geometry of the cracking patterns in L-10N2, in
which about 24% of the total shear force was transferred in the uncracked compression
zone, and it is unlikely that crack control steel significantly alters this proportion.
We are therefore left with the conclusion that the majority of the additional shear strength
exhibited by L-20D and L-20DR, about (100%-24%-19%)=57%, was due to enhanced
aggregate interlock capacity along the narrower cracks.

6.4.5

Code Estimates of the Shear Strength

It can be seen in Figure 6-24 that both the ACI and SMCFT generally provided accurate
estimates of the failure loads of the small beams, but both methods were somewhat
unconservative when estimating the failure loads of the two large specimens. For these
large beams, the average ratio of vexp/vACI was 0.86, while the average ratio of vexp/vSMCFT
was 0.91.

278

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Providing skin reinforcement spaced closer than that required by ACI 318-05 thus did not
completely eliminate the size effect in shear. Had bars of a diameter smaller than that of
a 10M rebar been used as skin reinforcement, such as #3 US rebar or small deformed
wires, the shear strength would have been reduced. Application of the ACI skin
reinforcement provisions, which do not require a minimum bar diameter, does not,
therefore, ensure that the size effect will be adequately accounted for when designing
slender beams and slabs without shear reinforcement.
The failure shear stresses of L-20D, L-20DR, S-20D1 and S-20D2 are plotted in Figure
6-28 versus the effective crack spacing, and have been normalized by (fc)0.5 and the CSA
strain effect term. The CSA size effect term, 1300/(1000+sze), has been plotted as well,
along with the experimental points from the other S- and L- series specimens. This figure
shows that the size effect has not been completely eliminated through the use of crack
control steel with d=0.3%. Interestingly, the difference between the normalized shear
strengths of S-20D1/2 and L-20D/R resembles the difference between the normalized
shear strengths of the small and large specimens tested with stirrups (see Chapter 7).
Based on these observations, it seems that, while the CSA shear provisions perhaps need
to be improved for members with distributed reinforcement, it is unclear whether
increasing d will adequately address the situation.
Table 6-3 compares experimental results from fifteen tests of continuous and simplysupported beams of various depths reported by Collins and Kuchma, all of which
contained crack control steel with d values well in excess of 0.3%, with the four tests in
the D series of this thesis. Examination of the vexp/vsmcft values indicates that a number of
the specimens failed at shear strengths that were less than those predicted by the SMCFT.
BHD100, BND100 and SE100B-45, for example, failed at shears that were 78%, 83%
and 77%, respectively, of the SMCFT predicted failure shears. With the exception of
SE50B-45, generally the deeper specimens exhibited reduced shear strengths. These
specimens had d values considerably higher than that used in L-20D as a result of both
smaller vertical spacings between the layers of crack control steel and larger bar
279

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

diameters, yet failed at lower ratios of vexp/vsmcft. Furthermore, these specimens were
shallower than L-20D. Thus, requiring a higher d value before sz can be set equal to the
vertical spacing between layers of crack control steel does not appear to be the correct
modification.

Figure 6-28: Size-Effect Factors for Members with Distributed Longitudinal Steel
280

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Table 6-3: Summary of Experiments of Beams with Crack Control Reinforcement


Specimen

Specimen Properties and Experimental


Observations

ACI Method

bw
sd f'c(1) ag,eff Vexp vexp
vACI
d
w
d
(mm) (mm) (%) (%) (mm) (MPa) (mm) (kN) (MPa) (MPa)

vexp
vACI

SMCFT -Modified sx

SMCFT

sxe vSMCFT vexp


sxe vSMCFT vexp
(mm) (MPa) vSMCFT (mm) (MPa) vSMCFT

Continuous Beams
SE100B-45
SE100B-45R
SE50B-45

295
295
169

920 1.36 1.04 195


920 1.36 1.04 195
445 1.20 0.61 195

50
50
53

9.5
9.5
9.5

281
316
87

1.04
1.16
1.16

1.16
1.16
1.19

0.99
1.00
0.97

268
268
268

1.35
1.35
1.28

0.77
0.86
0.90

547
547
268

1.18
1.18
1.28

0.88
0.99
0.90

SE100B-83
SE100B83R
SE50B-83

295
295
169

920 1.36 1.04 195


920 1.36 1.04 195
445 1.20 0.61 195

86
86
91

0
0
0

365
364
101

1.34
1.34
1.34

1.36
1.36
1.35

0.99
0.99
0.99

427
427
427

1.38
1.38
1.28

0.97
0.97
1.05

872
872
427

1.15
1.15
1.28

1.17
1.17
1.05

3-Point Bending Beams


B100D
BND100
BND50
BND25

300
300
300
300

925
925
460
225

1.19
1.19
1.09
1.48

1.18 170
0.78 170
0.78 85
1.19 40

36
37
37
37

9.5
9.5
9.5
9.5

320
258
163
112

1.15
0.93
1.18
1.66

1.02
1.03
1.02
1.04

1.13
0.90
1.15
1.59

233
233
117
55

1.11
1.12
1.15
1.33

1.04
0.83
1.02
1.25

480
480
119
55

0.99
1.00
1.15
1.33

1.17
0.93
1.02
1.25

BHD100
BHD100R
BHD50
BHD50R
BHD25

300
300
300
300
300

925
925
460
460
225

1.19
1.19
1.09
1.09
1.48

0.78 170
0.78 170
0.78 85
0.78 85
1.19 40

99
99
99
99
99

0
0
0
0
0

278
334
193
205
111

1.00
1.20
1.40
1.49
1.64

1.38
1.38
1.37
1.37
1.40

0.73
0.87
1.02
1.08
1.18

372
372
186
186
88

1.28
1.28
1.35
1.35
1.60

0.78
0.94
1.04
1.10
1.03

764
764
190
190
88

1.09
1.09
1.35
1.35
1.60

0.92
1.11
1.04
1.10
1.03

S-20D1
S-20D2
L-20D
L-20DR

98 252 0.81 0
96 250 0.83 0
295 1370 0.84 0.3
295 1370 0.84 0.3

40.6
40.8
35.8
35.8

19
19
19
19

24.8
25.8
350
361

1.00
1.08
0.87
0.89

1.08
1.08
1.02
1.02

0.93
0.99
0.85
0.87

227
225
225
225

0.99
1.00
0.96
0.96

1.01
1.07
0.90
0.92

227
225
685
685

0.99
1.00
0.78
0.78

1.01
1.07
1.11
1.13

--225
225

Average:
COV:

1.01
17%

0.97
12%

1.06
9%

The values of vexp/vsmcft from Table 6-3 are plotted in Figure 6-29 for the different d and
effective depth values. There is a general trend indicating that vexp/vsmcft decreases as the
effective depth increases. This relationship with the effective depth can also be seen in
Figure 6-30, where generally the CSA size effect term becomes more unconservative as
the effective crack spacing increases. The shape of the experimentally determined size
effect curves for the BND and BHD data, in particular, do not match the shape of the
CSA size effect curve.
Failure cracks in members with crack control steel can be expected to be flatter than in
members without crack control steel (see Figure 6-25). However, as cracks become
flatter, longitudinal crack control steel will become less effective at controlling crack
widths. (This is in contrast to members with stirrups, in which flatter crack angles result
in increased stirrup effectiveness.) It appears that this reduced effectiveness was
sufficient to overcome the beneficial effects of large crack control bar diameters in
members summarized in Table 6-3.
281

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

0
1.4

500

1000

1500

1.4

Vexp
VCSA

Vexp
VCSA

d=445-460mm

1.2

1.2

1.0

1.0

0.8

0.8

(mm)

Tr e n

dline

d=1370mm
0.6

0.6

d=925mm
0.4

0.4
CSA-04 Minimum

0.2

0.2

a)

(in.)

b)

0.0

0.0
0.0

0.3

0.6

0.9

1.2

1.5

2Abar/bwsz (%)

10

20

30

40

50

60

Effective Depth, d

Figure 6-29: Shear Strengths of Members with Crack Control Steel


Vc

Test Series

b w dv
1.6
0.4
'
fc
1 + 1500 x
1.4

L-20D
S-20D
BND
BHD
B100D
SE-45
SE-84

1.2

1.0

0.8

0.6

0.4

0.2

0.0
-200

200

400

600

800

1000

1200

Effective Crack Spacing, sze (mm)

Figure 6-30: CSA Size Effect Term for Members with Crack Control Steel
282

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.4.6

Control of Crack Widths

Suggested Modifications to CSA Code

It is suggested that the method by which sz, the crack spacing parameter, is calculated be
modified to account for the lower values of vexp/vsmcft values at higher effective depths. It
is recommended that it be calculated as follows:
For d < 450mm,
For d 450mm,

sz = sd
d
sd 0.9d or 0.72h, whichever is larger
sz =
450

(6-11)

where sd is the vertical centre-to-centre spacing between layers of crack control steel having a
minimum area of 0.003bwsd per layer. See Figure 6-31.

Figure 6-31: Definition of sd in Eq. 6-11


A number of different factors of varying complexity by which sd could be multiplied
were examined, but it is felt that the (d/450) factor is appropriately simple for design
situations.
The vexp/vsmcft values presented in Figure 6-29 are recalculated and plotted in Figure 6-32
using the modified method to calculate sz. As summarized in Table 6-3, the average
value of vexp/vsmcft for the experiments is 1.09 with a COV of 9% using the modified value
of sz, versus 0.97 and 12% based on sz=sd. Using the modified method to calculate sz
results in a slight upwards trend in the vexp/vsmcft values as the effective depth increases.
As there is a lack of experimental data for members with effective depths exceeding
1400mm with crack control steel, it is appropriate to be conservative when extrapolating
beyond the range of experimental data.
283

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

Referring to Figure 6-33, it can be seen that the experimentally determined values of the
size effect term and the shapes of the experimentally determined size effect curves more
accurately follow the CSA size effect curve. Note, for example, that test series with
formerly one value for sze (the SE series and the S- and L- series reported here) now have
different sze values, with variations in the normalized shear strength that more closely
match the CSA size effect term. While the line for the SE-45 series falls below the CSA
curve, the slope of the line matches the slope of the CSA curve.
The crack spacing parameter in the CSA code, sz, is intended as a measure of the crack
spacing at mid-depth of a member. For members without stirrups or crack control steel,
it is equal to 0.9d, and is analogous to a characteristic crack spacing. Shioya et. al.
(1989) found in their tests that the average crack spacing at midheight was generally
about 0.5d in specimens without stirrups or crack control steel, and similar crack spacings
were found in this experimental program. The CEB-FIP code defines a characteristic
crack width, wk, as the width that 5% of cracks will exceed, and it is equal to 1.7 times
the mean crack spacing, wm. The characteristic crack spacing, sk, therefore, is equal to
1.7 times the mean crack spacing, wm. For a mean crack spacing of 0.5d, the
characteristic crack spacing is 1.7 x 0.5d = 0.85d, which is about equal to the CSA crack
spacing parameter.
In specimen L-20D, the mean measured crack spacing was 400mm, resulting in a
characteristic crack spacing of 1.7x400=680mm. The crack spacing parameter for
specimen L-20D calculated using Eq. 6-11 is equal to 685mm, and thus corresponds to
the characteristic crack spacing.

284

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

Control of Crack Widths

0
1.4

500

1000

1500

1.4

Vexp
VCSA

Vexp
VCSA

1.2

(mm)

1.2
Tre ndline

1.0

1.0

0.8

0.8

0.6

0.6

0.4

0.4
CSA-04 Minimum

0.2

0.2

0.0

0.0

(in.)
0.0

0.3

0.6

0.9

1.2

1.5

2Abar/bwsz (%)

10

20

30

40

50

60

Effective Depth, d

Figure 6-32: Shear Strengths of Members with Crack Control Steel Based on Eq. 6-11
Vc

Test Series

b w dv
1.6
0.4
'
fc
1 + 1500 x
1.4

L-20D
S-20D
BND
BHD
B100D
SE-45
SE-84

1.2

1.0

0.8

0.6

0.4

0.2

0.0
-200

200

400

600

800

1000

1200

Effective Crack Spacing, sze (mm)

Figure 6-33: Size Effect Term for Members with Crack Control Steel Based on Eq. 6-11
285

Shear Behaviour of Large, Lightly-Reinforced


Concrete Beams and One-Way Slabs

6.5

Control of Crack Widths

Concluding Remarks

The experiments described in this chapter have shown the importance of proper structural
detailing in minimizing the negative effects of cracking in reinforced concrete structures.
The skin reinforcement study has shown that it is incorrect to state that any sized rebar
can be used effectively to control flexural crack widths in the webs of large members.
The study on the shear strength of elements with crack control steel has shown that shear
strength can be enhanced through the use of a sufficiently large quantity of well
distributed longitudinal steel.
In tests of reinforced concrete elements subjected to pure tension, in which covers and
bar spacings were kept constant, Williams (1986) showed that small bars were unable to
control crack widths. When a crack first formed, the tensile force in the concrete, Acfcr,
exceeded the yield strength of the small bars. Hence, further straining occurred at the
cracks due to bar yielding rather than by the formation of additional cracks. Even in
elements with larger bars, where yielding did not immediately occur, elements with lower
steel areas exhibited more widely spaced cracks, as the small area of steel crossing the
crack had a low stiffness, AsEs, that was insufficient to restrain widening of the crack.
Likewise, in the case of skin reinforcement, small diameter bars spaced far apart are not
able to restrain the widening of cracks, as the longitudinal stiffness of the steel crossing
the crack will be insufficient. While Frosch quotes Gergely and Lutzs statement that
the bar diameter is not a major variable, (as discussed on page 241) in fact, Gergely and
Lutz state that the steel stress is the most important parameter. Very small bars with a
low sk value are highly stressed at the locations of flexural cracks. The steel stress is,
indeed, the primary variable, and the stress will be increased considerably through the use
of small diameter bars.

286

You might also like