Crack and Deflection

Control of Pretensioned
Prestressed Beams
Edward G. Nawy
Professor of Civil Engineering
RutgersThe State University
of New Jersey
Piscataway, New Jersey

P. T. Huang
Senior Structural Engineer
Gibbs & Hill
Consulting Engineers
New York, N.Y.

Based on a series of laboratory tests on

precast prestressed T and I beams, the
authors propose crack width and deflection
formulas for evaluating the serviceability of
such members. Several numerical examples
show the applicability of the recommended
crack width equations.

he primary objective of this investigation was to study the serviceability of pretensioned prestressed I and T
beams through an analytical and experimental investigation of their flexural
cracking and deflection behavior up to
failure.
Available experimental data on cracking in prestressed concrete members is

30

limited. 1 - 5 Formulas on crack width

prediction from previous researches, in
general, were based on two different
concepts. The first concept, proposed
by Ferry-Borges, 6 Nawy-Potyondy,2 -3
Holmberg,7, 8 and CEB9 is to relate
crack width to the stress or strain in the
reinforcement.
Another concept, originated by

Abeles 1 proposed the crack width to

be related to the flexural tensile stress
in the concrete. The analytical equations proposed in this investigation are
based on the first concept.
Tests on four series were conducted
on 20 simply supported beams of 9-ft
span and four two-span continuous
beams of effective 9-ft spans. The major controlling parameters were the variations in the steel reinforcement percentages of the prestressing tendons
and the non-prestressed reinforcement.
The prestressing tendons were 1/4 in.
nominal diameter 7-wire strand, 250
ksi elements. The mild steel reinforcement was either #3 or #4 deformed
high strength bars of yield strength
varying between 79 and 84 ksi and ultimate strength between 100 and 110
ksi. The total steel percentage was varied from 0.17 to 1.08 percent.
This paper proposes analytical expressions for evaluating the crack
widths in such members at working
and overload loading levels in terms of
the controlling parameters. The investigation also correlates the deflection re-

suits with the expected computed deflections. Several numerical examples

are included to show the applicability
of the proposed crack width formulas.

Test Program
This section describes the properties
of the materials used in the experimental program and gives the details of fabrication of the beam specimens and the
testing procedure.

Materials
The mix was proportioned for a nominal 28-day compressive strength of approximately 4000 psi (281 kg/cm 2 ). The water-cement ratio varied between 5.4 and
7.1 gal. per sack of cement. The coarse
aggregate used was crushed stone of a/s in.
(9.35 mm) maximum size; while the fine
aggregate was natural local sand. The
slump varied between 5 and 71/2 in. (12.7
and 19.1 cm) as given in Table 1.
Uncoated, stress-relieved, 250 ksi sevenwire 1/4 -in, strands were used for prestressing. The strand nominal area was
0,0356 sq in. (0.234 cm2) and the material had a unit elongation of 0.65 percent

YIELD STRENGTH 8Y
0.2% SET METHOD

250

20C
%

AREA 0.0356 IN2

ULT. STRENGTH=9600 LBS

y 15C
U)
w
rc
U)
IOC

50

0
0.000

Eg
V.VVO

V.V I'

u.V V

STRAIN (IN/IN)

Fig. 1. Typical stress-strain relation of prestressing steel.

PCI JOURNAL/May-June 1977

31

Table 1. Properties of concrete in beam specimens.

Beam

Compressive
Strength (psi)

Age at
Test
(days)

3 Days

B1-64
B5-B6
B7-B8
69-610
611-612
B13-B14
615-B16
B17-B18
619-620
B21
B22
B23
B24

28
29
31
33
35
27
29
34
34
37
36
39
33

At Test

Water-Cement
Ratio
(gals/sack)

4780
4100
4350
2915
4740
4200
4130
4925
3010
4720
4350
4150
4385

650
545
555
470
705
650
660
695
470
730
660
670
620

5.4
6.2
6.3
7.1
5.8
6.4
6.6
5.5
7.1
5.8
6.4
6.6
5.5

3080
2920
2920
1850
3100
2900
2670
2650
1850
3100
2900
2670
2655

Tensile
Splitting
Strength
(psi)

at 70 percent of the ultimate. The material

satisfied ASTM A-416 specifications and
had a typical stress-strain relation as shown
in Fig. 1.
Non-prestressed #3 and #4 deformed
bars were used as supplementary reinforcement at the tension side in all specimens
except Beams B-1 to B-6. A typical stressstrain diagram for the deformed bars is
shown in Fig. 2. Table 2 gives the details
of the reinforcement used for the beams in
this test program, and Fig. 3 gives typical
cross sections of the test beams.

Avg.
Slump
(in.)

5
6i
6',.
9
5
7
7
7z
9
5
7
7
7%,

Fabrication and testing

Twenty single-span and four continuous
beams were fabricated i and tested. The
simply supported single-span beams were
as follows:
(1) Beams B-1 to B-6 were T sections
with pretensioned prestressing tendons
only.
(2) Beams B-7 to B-18 were also T
sections reinforced with both pretensioned
prestressed tendons and non-prestressed
mild steel.

Table 2. Geometrical properties of beam specimens.

Beam Size

Beam

Sect.

Total
Depth

(in.)

B-1
B-2
B-3
B-4
B-5
B-6
B-7
B-8
6-9
0-10
B-11
B-12
B-13
8-14
B-15
B-16
0-17
B-18
B-19
B-20
8-21
B-22
8-23
B-24

32

T
T
T
T
T
T
T
T
T

10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10

T
T
T
T
T
T
T
T
T
I
I
I
I
I
I

Eff.
Depth

(in.)

8.00
7.75
7.30
7.00
6.57
6.25
8.80
8.80
8.65
8.65
8.37
8.37
8.46
8.46
8.22
8.22
8.00
8.00
9.00
9.00
9.00
9.00
9.00
9.00

Flange
Width

(in.)

Longitudinal

Web
Thickness

No.
of
1/4-in.
Tendons

No.
of
#3
Bars

Compres.
Steel

Tensile Steel

No.
of
#4
Bars

As

As,
p

(in.)

Size
of
Bar

(5)
(s9
in.)

Web
Reinforcement

Spacing

(5)
(s9
in.)

(in.)

10

11

12

13

14

15

8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
8
6
6
6
6
6
6

3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3

3
4
5
6
7
8
3
3
4
4
5
5
6
6
7
7
8
8
3
4
3
3
4
4

2
2
2
2
2
2
2
2
2

.108
.144
.180
.216
.252
.288
.328
.328
.364
.364
.400
.400
.616
.616
.652
.652
.688
.688
.328
,544
,328
.328
,544
.544

0.17
0.23
0.31
0.39
0.48
0.58
0.47
0.47
0.53
0.53
0.60
0.60
0.91
0.91
0.99
0.99
1.08
1.08
0.61
1.01
0.61
0.61
1.01
1.01

.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22
.22

.34
.35
.38
.39
.42
.44
.31
.31
.32
.32
.33
.33
.33
.33
.33
.33
.34
.34
.41
.41
.41
.41
.41
.41

#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3
#3

6
6
6
6
4
4
6
6

2
2
2
2
2
2
2
2
2
-

6
4
4
4
4
4
4
4
4
4
4
4
4
4

0)
Y
y
W
h

Fig. 2. Typical stress-strain relation of non-prestressing steel.

(3) Beams B-19 and B-20 were I sections with both prestressed and non-prestressed reinforcement.
Beams B-21 through B-24 were I sections continuous on two spans reinforced
with both pretensioned prestressed tendons
and mild steel reinforcement. Straight
strand profiles were used in all beams.
For continuous beams, strands were inserted into plastic hoses in the compression
regions to achieve zero bond between con-

(a)

T - SECTION

crete and strands. Before testing, those

strands in the compression regions were
cut through to eliminate the effectiveness
of strands in these zones.
All beams were 10 in. (25 cm) deep as
shown in Table 1. They were all over-designed to resist diagonal tension. Deformed
closed stirrups (#3 bars) at 6 in. (15 cm)
center-to-center were used throughout the
span for beams with a low tensile steel
reinforcement percentage and 4 in. (10

(b)

I - SECTION

Fig. 3. Geometric beam cross -sectional dimensions for T and I sections.

PCI JOURNAL/May-June 1977

33

cm) for beams containing a high percent-

Readings were taken of the change in

strain at all the necessary stages of pmstressing and loading. In addition, demountable mechanical gages were used to
measure the variation of strain on the concrete faces of the beams.
Mechanical dial gages having 2 in. (5
em) travel and 0.001. in, (0.025 mm) accuracy were used to measure the change in

age of tensile reinforcement (see Table 4).

Pretensioning frames previously designed
for the work reported in References 2 and.
3 were used in this investigation. Each
frame provided an independent unit with
60,000 lb (13,6 tons) axial capacity.
All reinforcement was instrumented with
electric strain gages at critical locations.

Table 3, Measured and calculated cracking and ultimate

moments of simply supported beam specimens.
Meas. M cr

Calc, M

Mcr

Meas. M.

Calc. Mu

Meas. M.

--r
Ca
(in.-kips)

(in.-kips)

B-1
B-2
B -3
8 -4
B-5
6 -6
8-7
B-8
B-9
B-10
6-95

137.0
164.0
178.2
189.0
1,83.0
193,2,
126.0
136.4
130.0
126.5

136.9.
159.7

B-13
B-14
0-15
B -16
B-17
B-1@
B-19.
8-08

994.0

173.0

195.0
205.9.
201.$.
216.01
253,0
126.0
127.3

(in. -kips)

183.6
186.7

1.00
1.03
1..04
7.03
0,98.

258.3
310.8
344.4
390.6
4.17.9,

137.6
140.0
136..2
138.4
777,3:
177.3

0.92
0.98.
0.96
0.92
0913
0.98

336.0
336.0
336.0
327.6
445.2
443.1

1.04.
1.03

558.6
562.8
554.4
600.6.
592.2
308.,7
3003.

103,4;

979,0

6-12

(in.-kips)
4

170..6

1.,03

182.3
187.5

1.02

1.00
0.97
0.92,

0.75

1.175
1.121
1.149.
1.181

935,999;

Tension
Tension
Tension
: Tension
Compression
Compression
Tension
Tension
Compression
Compression
Compressin.
I Compression
Compression
Compression
Compression
Compress.i; Qn.
Compression

1..240
1.240

345.59
296.10
296.78
329.11
328.21
379.62
380.14
546.79
547.94
513.15

588.0

219.74
277.30
299.73.
330.68

430.5.

1.06

199.0
197..5.
215.0
219.1
737.3
167.9

Mode of
Failure

Cal c.Mu

1.135
1.132
1.021
0.998
1.172

1.166

1.075"
1;01.9
1.092

510,36.

1,999

580.65
584,90
293.57
454,11

1.030

0ompres$ son.

1..015
1.052

Compression.
Shear

0.660

Table 4. Measured stabilized crack spacirugs versus their theoretical values.

Beam

Meas.
Stab. min.
Crack
Spacing
acs

min.

Theo.
- Stab. min.
Crack
Spacing
acs

a
Csmin.
ateCSein.

Meas.
Stab. max.
Crack
Spacing
acs

min.

max.

a
Stab. max.
Crack
Spacing

Meas.
Stab.
Mean
Crack
Spacing.

csmax.
aT
osmax.

acs
max.

a
(in.)
1
B-1
B-2
B-3-

8-4

B-5.
6-6
0-7
B-8.
B-9
B-10
8-11
B-12
0-13
0-14.
B-15
B -16
8-17"
8-18
0-19.
B-20
Mean
5.0.

34

2
5.00
4.13
4.31
3,7.3
2.38
2.50
2.1,3.3

1.94.

2.13
2.50
1.75
1.75
1.44
1.38
1,25
1.75.
1.63
1.50.
1,88
1.69.

(in.)
3
4.08
3.93
3.73
3.26
2.77
2.41
2.13
1.98
1.90
1.84
1.63.
1.65
1.58
1.51.
1.39
1.50
1.50
1.37
2.09.
1.85.

4
1.23-

1.05:

1.1.6
0.96
0.86
1.04
1.00
0.93.
1.12
1.36

1.07

1.06
0.91
0.91
0.90:.

1.17
1.08
1.09.
0.90
0,91
1.04:
0.13,

(in.)

(in.),

7.25
7.63
6.63
6.38:
5.00,
5.13
3.94
4.06.
3.75.
5.00
3.13
3.38
3,56
3.63
3.73
3.00
3.25.
3.06-.
4.25
3.75

8.15:
7.86
7.4,6.
6.52
5.53,
4.51
4.26.
3.94:
3.80
3.68.
3.26
3.30.
3.16
3.03
2.773.00.
3..00
2.754.18
3.70

Corre-.
sponding
Initial
M/M
n

mean
(in.)

7
0..89
0,97
0.89
0.98
0.90
1..07
0.92
7.03
0.97'
1.36
0..96
1.02
1.13
1.20
1.13
1.00
1.08
1.11
1.02
1.03
1.03
0.11.

6.12
5.90
5.60
4.89
4.15
3.61
3.20
2.97
2.85
2.76
2.44
2.47
2.37
2.27
2.082.25.
2.25
2.06
3.732.78.

a
Theo.
Mean
Crack
Spacing

csmean
a
csmean

a csmean
(in.)

9
0..59
0.55
0.67
0.67
0.70
0..66
9,.75
0.75
0.73
0.76:
0.66
0.66
0.68
0.71
0..75
0.76
0.70
0.71
0.72
0.70

10..
6.33
5.34
5.13
4.75.
4.39
3.92
2.97
2.97
2.72
2.72
2.67
2.67
2.21
2.21.
2.46-.
2.46
2.29
2.29
3.52
2.57

11
0.97
1.11
1.09
1.03
0.95
0.92
1.08
1.00
1.05
1.02
0.91
0.93
1.07
1.03
0.84.
0.91
0.98
0.90
0.89
1.08
0.99.
0.08

deflection due to loading. Crack widths

were measured with illuminated 25-power
microscopes having a 0.05 mm accuracy.
Crack spacings of all the developing cracks
were also accurately recorded as well as the
crack penetration of the principal cracks.
For most of the beams, eight to nine
increments of load were applied to failure. At the conclusion of each beam test,
the concrete control cylinders were tested
for both compressive and tensile failure.

from 138.0 to 148.0 ksi (9701 to 10404

kg/cm 2 ) after deducting anchorage,
elastic shortening and creep and shrinkage losses.
The measured external moments, Mc,.,
producing the first hair crack, is compared to the calculated cracking moment in Table 3. This table also gives
the measured and calculated ultimate
moments.

Flexural cracking

Test Results

Maximum crack widths were measured at the reinforcement level and at

the bottom tensile face of the concrete.
The spacings of the cracks were meassured an both faces of each beam at
each loading stage. These spacings
were summarized for each test specimen and the mean crack spacing was
calculated.
Table 4 gives the mean stabilized
crack spacing for load ratios of 50 to
70 percent of the ultimate load. Table
5 gives the measured crack width of the
stabilized cracks at the reinforcement
levels of the steel closest to the outer
fibers for the various stress levels.

This section summarizes the test results for effective prestress force, moment capacity, flexural crack widths,
and load-deflection data.

Effective prestress and

moment capacities
A measurement of the effective prestressing force is necessary to determine
the net steel stress for each loading
state. The initial prestress varied between 174.1 and 180.6 ksi (12,240 and
12,696 kg/cm2 , respectively). The effective prestress correspondingly varied

Table 5. Observed versus theoretical maximum crack width at steel level.

Net Steel Stress afs
30 ksi
Beam

W obs.

W theo,

40 ksi
W
Obs.

obs,

W #heo,

W theo.

obs.

Wis.

W #+eo,

the,,.

B-1
B-2
B-3
0-4
6-5
B-6
B-7
B-8
B-9
B-10
B-11
B-12
8-13
B-14
B-15
8-16
B-17
B-18
B-19
B-20

.0089
.0071
.0062
,0053
.0047
.0042
.0040
.0039
,0077
.0041
.0036
.0035
.0030
,0032
.0025
.0026
.0027
.0021
.0041
.0026

.0089
.0075
.0072
,0067
,0065
.0058
.0041
,0041
,0035
.0035
.0043
,0043
.0034
.0034
.0040
.0040
.0437
,0037
.0049
.0036

1.00
0.95
0.86
0.79
0.72
0.72
0,98
0.95
1.06
1.17
0.84
0.81
0.86
0,94
0.63
0.65
0.73
0.57
0.84
0.72

.0122
.0104
.0090
.0075
.0067
,0058
.0055
.0054
.0050
.0056
.0050
,0052
.0045
,0047
.0042
.0039
.0042
.0031
.9957
,0039

Mean
S. 0.

60 ksi

_ --- 6

0,84
0,18

PCI JOURNAL/May-June 1977

,0118
.0100
.0096
.0089
.0087
.0078
.0055
.0055
.0047
.0047
.0058
.0058
.0046
.0046
.0053
,0053
.0049
.0049
,0065
.0047

Wig,

obs,

rhea,

Wtheo.

1,03
1.04
0.94
0.84
0.77
0.74
1.00
0.98
1,06
1.19
0.86
0,90
0.98
1.02
0.79
0.74
0.86
0.63
0.88
0.83

.0198
.0172
:,0149
.0127
,0108
.0095
,0095
,0089
.0081
.0090
.0080
.0086
.0077
.0080
.0077
.0067
.0074
.0063
.0090
.0070

.0177
.0150
.0144
.0133
.0130
.0116
.0082
.0082
.0070
,0070
.0087
.0087
.0088
.0068
.0080
.0080
.0074
.0074
.0097
.0071

0.90
0.14

80 ksi.

10
1.12
1.15
1.03
0.95
0.83
0,82
1.16
1.09
1.16
1.29
0.92
0.99
1.13
1.16
0.96
57.84
1.00
0.85
0.93
0.99
1.02
0.14

obs.
then

11
.0285
.0242
.0212
.0184
.0160
.0140
,0135
.0128
10123
.0132
.0123
,0130
.0112
.9117
.0117
.9355
.0112
,0106
.0132
.0110

12

13

.0236
.0198
.0193
.0178
.0174
.0155
,0109
.0109
.0094
.0094
.0115
.0115
.0091
,0091
.0107
.0107
.0098
.0098
.0130
.0096

1.21
1.22
1,10
1.03
0.92
0.90
1.25
1.17
1.31
1.40
1.07
1.13
1.23
1.28
1.09
0.98
1.14
3.08
1.02
1.16
1.13
0.13

35

aI
O
J

0.3

0.0

0.6

0.9

1.2

1.5

1.8

2.1

DEFLECTION (IN)

Fig. 4. Typical load -deflection relation for Beams B-3 and B-4.
Deflection
Deflection behavior of prestressed
concrete structural elements differs
from that of reinforced concrete systems. Initial reverse deflection due to
prestressing, namely camber, has to be
considered. Typical composite net
load-deflection relations due to external
load are shown in Fig. 4.

(a) FOR EVEN DISTRIBUTION OF

REINFORCEMENT IN CONCRETE

These curves essentially show a trilinear relation. The first stage represents the precracking stage, essentially
elastic, the second stage, namely, the
post-cracking stage where the deflection
of the beam increases faster as more
cracks develop, and the third stage denotes the behavior prior to failure.

(b) FOR NON-EVEN DISTRIBUTION OF

REINFORCEMENT IN CONCRETE

Fig. 5. Effective area in tension for distribution of reinforcement.

36

Mathematical Model for

Evaluating Serviceability
This section describes the mathematical model used to evaluate serviceability based on crack spacing, crack width,
and deflection.

Crack spacing
Primary cracks form in the region of
maximum bending moment when the
external load reaches the cracking load.
As loading is increased, additional
cracks will form and the number of
cracks will be stabilized when the stress
in the concrete no longer exceeds its
tensile strength at further locations regardless of load increase.
This condition is important as it essentially produces the absolute minimum crack spacing which can occur at
high steel stresses, to be termed the
stabilized minimum crack spacing. The
maximum possible crack spacing under this stabilized condition is twice
the minimum, to be termed the stabilized maximum crack spacing. Hence,
the stabilized mean crack spacing ad8
is shown to be the mean value of the
two extremes.
The total tensile force T transferred
from the steel to the concrete over the
stabilized mean crack spacing can be
defined as:
(1)
T = f a,, u, , Io
where
f = a factor reflecting the distribution of bond stress
u rn = maximum bond stress which is
a function of \/fa
Jo = sum of the circumferences of
the reinforcing elements
The resistance R of the concrete
area in tension A t can be defined as:
R = A t f't
(2)
By equating Eqs. (1) and (2), the
following expression for a C3 is obtained,
where c is a constant to be developed
from the tests:

PCI JOURNAL/May-June 1977

a08
Cs

(3)

The concrete stretched area, namely,

the concrete area in tension A t for both
the evenly distributed and non-evenly
distributed reinforcing elements is
shown in Fig. 5.
With a mean value of f'tVL = 7.95
in this investigation, a regression analysis of the test data resulted in the following expression for the mean stabilized crack spacing:
a C8 = 1.20 A t /so
(4)
Fig. 6 gives the basic regression
analysis plot for aC8.

Crack width
If Af g is the net stress in the prestressed tendon or the magnitude of
the tensile stress in the normal steel at
any crack width load level in which
the decompression load (decompression
here means f c = 0 at the level of the
reinforcing steel) is taken as the reference point, 3 , 5 then for the prestressed
tendon:
where

ifs = fn.t fd

(5)

f,^ t = stress in the prestressing steel

at any load level beyond the
decompression load
f,t = stress in the prestressing steel
corresponding to the decompression load
The unit strain e, = Of8 /E 3 since it
is reasonable to disregard the unit
strains in the concrete due to the effects of temperature, shrinkage and
elastic shortening. 4 ' 5 Hence, the maximum crack width can be defined as:
2U mx = k a 0, Ea,
(6)
where k and a are constants to be
established by tests.
Alternatively:
w.mam = k' a as(Of,) a
(6a)
Eq. (6a) is rewritten in terms of Of8
so that an analysis of the test data of
all the simply supported test beams
leads to the following expression at the

37

Z 5
d

o
4

to
z

a^ g =0.I5I f+At

a)
3

C.,

rc
C.,

a 2
w

0
w
M

STAND. ERROR OF EST. 0.2698

sv

ov

LV

IV

V 20

Fig. 6. Stabilized mean crack spacing versus

0.1

Ila

0,!

Its

-0.,
z

II4

f tAt/V?,

jo

^
/
'
Wmax.5.85xt9" (Mo ^Dfs
__

0. f12

E
0. 110

*40%

0.

..-

---

.- _______

-40%

108

ac
g 0.
rc

a1 W4

Q 0,

STAND. ERROR OF EST, 0.00125'"

0 )o
0

'SC)

Ivy

13v

Gvv

c:v

avv

(At)Af6
fig. 7, Linearized maximum crack width versus net steel stress.

38

aav

crack widths at the reinforcing steel

level as well as at the tensile concrete
faces of the beams at net stress levels
t f s of 30, 40, 60, and 80 ksi.

reinforeement level::
(7).
= J.4 X W 'i 4
g S (Of4)1..31
A 25 percent band of sc3tt-cv envelopes all the data for the expression.
ir> Eq.. (7) for Af 9 20 to 80 ksi.
Linearizing Eq. (7) for easier use by
the design engineer leads to the following simplified expression of the
maximur crack width at the reinforcing steel level:
= 5.85 X 10- 5
(8a)

w.

Deflection
Deflection computations under service load conditions will usually be
necessary to ensure deflection serviceability in addition to crack control serviceability. Since the deflection which
concerns the design engineer most is a
service load condition, both the nncracked and cracked section properties
are needed11,12 for the computations.
An estimate of the magnitude of deflection can be made from the following equation:

(A7 Io) Ofs

The maximum crack width (in.) at

the tensile face of the concrete is given
by.:,
tp Max = 5.85>< 1 0-5 (R4)(At/o)^s (8b)
where R4 is a distance ratio as defined
in the notation.
A plot of the data and the best fit
expression for Eq (8a) is given in Fig. 7
with a 40 percent spread which is reasonable in view of the randomness of
crack development and the linearization of the original expression of Eq.

S c ML ?

(9)

where c is a constant depending on the

loading and support conditions.
Branson's generally accepted11,13,14
expression for the effective moment of
inertia Ie is as follows:

(7).
Tables 5 and 6 give the relation between the observed and theoretical

Table 6. Observed versus theoretical maximum

cr k width at tensile face of beam.
let
40" ksi

39 ksi.
Beam

W obs.thea.
W
.

W obs.

W obs-

Wibeo.

c.

1
B-1
6-2
B-3
B-4
B-5
6-6
B-7
B-8
6-9
B-10
8-11
B-12
B-13
B-14
B-15
B-16
B-17
B-18
B-19
B-20
Mean
S.D.

.0111
.0127
.0131
.0097
.0091
.0124
.0052.
.0049
.0051
.0058
.0054
.0048
.0043
.0052
.0039
.0038
.0039
.0030
.0057
.0034

.0731
.0110
.0128
.0130
.0147
.0148.
.0051
.0051
.0045
.0045
.0059
.0059
.0046
.0046
.0057
.0057.
.0056
.0056
.0061
.0045

Steel Stress.4fs.

0.847
1.079
1.022
0.742
0.619
0.835
1..015
0,956
1.130
1.285
0,917
0,815
0.937
1.133
0.682
0.664
0.698
0.537
0.931.
0.750

.0151
.0175.
.0204 ' .0157
.0166:
.0172
.0174
.0158
.0117
.0197.
.0181
.0199
.0068
.0069
.0061
.0069
.0064
.0061
.0082
.0061
.0069
.0079
.0076
.0079
.0058
.0062
.0062
.0059
.0061
.0076
.0057
.0076
.0060
.0074
.0045
.0074
.0085
.0081.
.0045:
.0059,

0.8800.196.

,
PCI JOURNAL/May-June 1977

80 ksi

60 ksi
W obS.
Wtheo.

Webs.

0,865
1.299
0.971
-. 0.910
0.595
0.906
0.989
0.887
1.056,
1.352
0.868
0.956
0.934
0.950
0.805
0.752
0.811
0,608
1.046.
0.760
0.916
0.187

WtheO.

Wobs.

Webs

Web s .

Wtheo.

.0261
" .0275
.0304
: .0226
.0205
.0213
.0117
.3111
.0107
.0134
.0112
.0134
.0105
.0103
.011.5
.0093
.0098
.0086
.0129
.0089

9
..0262
.`9236.
.0256,
.0259
.0294
.0297:
.0103:"
.01.03
.0090
.0090
.01.19
.0119
.0092
.0092
.0114
.0114
.0112
.0112
.0121
.0089

10
1.000
1.108
1.166
:' 0.871
0.698
! 0.717
1.141
1.Q63
1..185:

1,484

0.940
1.124
1 144
1.122
1.005
0.81.3
0.877
0,770
1.064
1.000
1.017
0.190

Wtheo.
11

12.

13

.0400
.0409
.0382
.0304
.0320
.0364
.0188
.0146
.0165
.0185.
.0172.
.0192
( .0138
.0145
.0181
.0160
.0159
.0147
.0202
.0139

.0349:.
.0313
.0344
.0347
.0393,
.0397
.0137
.0137
.0121
.0121
.0158
.0158
.0123
.0123
.0153
.0153
.0148
.0148
.0163
.0119

1.1.45.
1.309
1.112
0.876
0.814
0.917
1.380
1.072
1.361
1.526
1.092
1.219
1.1.23
1.180
1.183
1.046
1.074
0.993
1.243
1.161
1.141
0.174

39

1 ' = (Mcr

3 1,

[i_(Mcr)]1^^

(10)

The cracking moment M0,. of a prestressed concrete beam can be computed from the following expression
for evaluating le:

M c,.=Fe-}

AI9
9y t

+frl.9 (11)
yt

where Ier and I9 can be calculated from

the properties of the section. The concrete modulus of rupture, f, is equal to

7.5a/f..

Eqs. (10) and (11) were applied to

the deflection test results of 19 prestressed beams, 13 of which had additional non-prestressed tension steel. The
range of application was from the
cracking load level to 90 percent of
the ultimate load.

Fig. 8 gives a plot of the computed

versus measured deflections for the
short-term loadings applied in this investigation. It is noted that the degree
of scatter is within a 20 percent band
which can be considered fully adequate.

Discussion of
Test Results
It is observed from this investigation
that the initial flexural cracks started at
a relatively low net steel stress level
between 3 and 8 ksi. These initial
cracks formed in a rather random manner and with an irregular spacing. All
major cracks usually developed at a net
steel stress level of 25 to 30 ksi.
At higher stresses the existing cracks
widened and new cracks of narrow
width usually formed between major
cracks. A visibile stabilized cracking
space condition was generally reached

z
0)
z
0
t-

U
W
J
W
W
O
O
W
F-

a.

0
U

MEASURED DEFLECTION (IN.)

Fig. 8. Computed versus measured deflection.

40

8.0

2
E 6.0
z

a'' 4.0

U
U

V
Z
W 2.0
2

0.0
5.0

150

25.0

35.0

45.0

55.0

65.0

75.0

STEEL STRESS ofy (KS I)

Fig. 9. Mean crack spacing versus net steel stress.

at 0.5 to 0.7 of the ultimate load.

This investigation established that
the maximum crack spacings after stabilization were close to twice the minimum possible spacings having a mean
value of 2.02 and a standard deviation
of 0.29. The effect of the variation of
percentages of the non-prestressed
steel was significant both on the crack
spacing and the crack width.
For beams with non-prestressed steel,
the number of flexural cracks was almost twice as many as those with no
mild steel. These cracks were more
evenly distributed, with considerably
less spacing and finer widths. This behavior can be attributed to the fact
that the bond of the mild steel to the
surrounding concrete played a pronounced role in crack control.
A typical plot of the effect of the
various steel percentages on the crack
spacing at the various stress levels Of,
is given in Fig. 9. It is seen from this
plot that the crack spacing stabilized at
a net stress level of 36 to 40 ksi. The
influence of the various parameters,

PCI JOURNAL/May-June 1977

particularly, the variation in the steel

reinforcement percentages of the prestressing tendons and the non-prestressed steel on cracking and deflection
of prestressed concrete T beams is given in Table 7.
It is also observed that it is advantageous to locate the non-prestressed steel
below the prestressed tendons. This is
due to the fact that mild steel has larger diameters than the prestressing reinforcement, hence a larger bond area of
contact with the surrounding concrete.
Also, by placing the mild steel close to
the tensile concrete face, cracks will be
more evenly distributed, hence crack
spacing and consequently crack width
will be smaller.
The effect of the spacing of the stirrups on the crack spacing was not pronounced. It was found that the final
crack spacing and crack pattern did not
necessarily follow the vertical shear
reinforcement. Even though the first
few cracks usually started at the stirrups, the vertical legs of the stirrups
served only as initial weak areas of

41

Table 7. Influence of various parameters on cracking

and c.1 eftection of prestressed concrete T beams.
Beam

Tenon

NonPrestressed
Steel

Percentage
of Tension
Steel

B -1
8-7
64

3-1/4'
3-1/4"
31/4"

None
243
-2=0

0.17
0.47
9.47

6=2
8'-9
8=10

4.1/4"
4-1/4"
4-1/4"

588'e

2=#3
2-#3

B-3

8-12

5 -1/4"
5-1/4"
5 -1/4'

B-4
8-- 13

6-1/4"

2-#4
2-#4

B-11

8-14

8-5

B-15
8 -16
8-6
B-17

8-18

of steel*
(in. from Top
Fiber)

Crack
Spacing

Max. Crack Width

at Level of Steel
at f 5 = 60 ksi

No. of
Flexural
Cracks

8.52
8.75
8,75

6.12
3.20
2,37

.0198
.0095
.0089

10

.395

0:23
0,53
0.53

8,43
8,71
8.71

5,90
2.85
2.76

.0172
.0081

10

.390
.342
.350

None
2=93
2-03

0.31
0,60

0.60

8,22
8.63
8.63

5.60
2.44
2.47

,00$0
,0286

11

6
11

.320
.299
.288

6689

0.39
0.91
0.91

8.06
8.65
8.65

4,89
2,37
2,27

.0127
.0077
.0080

7
11
12

.255
.228
.247

7 -1/4"
7-1/4"

None
2-#4
2-#4

0.48

7.74
8,57

4.15
2.08

.0108

8
12

.234
.211

0-1/4"

Non
2- 64
2-#4

0.58
1.08
1.08

7,44
8.49
8,49

3.61
2.25
2,06

.0095

8
12
13

,176

6-1/4"

6-1/4'

7=1/4"

8-1/4"

8-1/4"

0.99
0.99

C.G.

8.57

2.25

.0090

10

.0149

.0077
.0067

.0074
.0093

12

Midspan
Deflection
at P = lOk
.950
.400

.207

.258
.192

Distance from compression fiber to the center of lAytr of mild reinforcing steel closest to tension fiber.

stress concentration. In most cases, the

stabilized mean crack spacings were
smaller than the spacings of stirrups.
Based on these observations and the
analytical results of this investigation
it can be said that the proposed equations for crack and deflection control
can be reasonably applied by the design engineer for maintaining the serviceability of pretensioned partially
prestressed beams and girders under
working load and overload conditions.
Once the allowable crack width is established for the prevailing environmental conditions, the proper percentage of non-prestressing reinforcement
can be determined to ensure serviceable behavior. Four design examples
are given in the Appendix to demonstrate the applicability of the proposed
crack width equations.

Conclusions
1. The maximum crack width (in.)
at the level of reinforcement closest to
the tensile face can be predicted from
the expression:
w = 5.85 x 10- 5(A 2/I9)f8
42

The crack width at the outer face of

the concrete is
W '?vax = tom"Rt
where R4 is the distance ratio.
2. Initial flexural cracks randomly
form at irregular spacings at low net
steel stresses of 3 to 8 ksi. All major
cracks usually develop at a net stress of
25 to 30 ksi.
3. Visible crack spacing stabilizes at
50 to 70 percent of the ultimate load.
4. The presence of non-prestressed
steel in the prestressed members has a
significant effect on crack control such
that the cracks become more evenly distributed and the crack spacings and
widths become smaller.
5. An increase in the percentage of
reinforcement decreases substantially
the crack spacings and width in the
partially prestressed beams.
6. For loads above the first cracking
load, the deflections are smaller for
prestressed concrete beams containing
non-prestressing reinforcement than for
a similar beam without non-prestressed
reinforcement.

REFERENCES
1. Huang, P. T., "Serviceability Behavior
and Crack Control in Pretensioned Prestressed Concrete I- and T-Beams,"
PhD Thesis under the direction of the
first author, Rutgers University, June
1975, 338 pp.
2. Nawy, E. C., and Potyondy, J. C.,
"Moment Rotation, Cracking and Deflection of Spirally Bound, Pretensioned
Prestressed Beams," Engineering Research Bulletin No. 51, Rutgers University, November 1970, 96 pp.
3. Nawy, E. G., and Potyondy, J. C.,
"Flexural Cracking Behavior of Pretensioned, Prestressed Concrete I- and
T-Beams," ACI Journal, Proceedings
V. 68, No. 5, May 1971, pp. 355-360.
4. Nawy, E. G., "Crack Control in Reinforced Concrete Structures," ACI Journal, Proceedings V. 65, October 1968,
pp. 825-836.
5. ACI Committee 224, "Control of
Cracking inConcrete Structures, ACI
Journal, Proceedings V. 69, No. 12,
December 1972, pp. 717-752.
6. Ferry-Borges, J. N., "Preliminary Report," Comite Europeen du Beton,
Commission IV aCracking.
7. Holmberg, A., and Lindgern, S.,
"Crack Spacing and Crack Width due
to Normal Force or Bending Moment," Document D2:1970, National
Swedish Building Research.

8. Holmberg, A., "Flexural Crack

Width," Nordisk Betong, 1970.
9. CEB-FIP, "International Recommendations for the Design and Construction of Concrete Structures," Comite
Europeen du Beton/Federation Internationale de la Precontrainte, Paris
(English Edition Published by the Cement and Concrete Association, London), 1970, pp. 1-80, Appendix V,
pp. 1-47.
10. Abeles, P. W., "Design of Partially
Prestressed Concrete Beams," ACI
Journal, Proceedings V. 64, No. 10,
October 1967, pp. 669-677.
11. ACI Committee 318, "Building Code
Requirements for Reinforced Concrete (ACI 318-71)," American Concrete Institute, Detroit, 1971, 78 pp.
12. ACI Committee 318, "Commentary on
the Building Code Requirements for
Reinforced Concrete (ACI 318-71C),"
American Concrete Institute, Detroit,
1971, 96 pp.
13. Branson, D. E., "Instantaneous and
Time-Dependent Deflections of Simple
and Continuous Reinforced Concrete
Beams," Part 1, Report No. 7, Alabama
Highway Research Report, Bureau of
Public Roads, August 1963.
14. PCI Design HandbookPrecast and
Prestressed Concrete, Prestressed Concrete Institute, Chicago, 1971, pp. 1-1
to 12-8.

NOTE
Four numerical examples showing the applicability of the recommended crack width
equations and a notation section, summarizing in alphabetical order the meaning of each
mathematical symbol, appear on the next four
pages.

PCI JOURNAL/May-June 1977

43

APPENDIX-DESIGN EXAMPLES
The following design examples are
prepared solely to illustrate the application of the preceding recommended
crack width equations. In these examples it is assumed that the member
geometry, load condition, stresses in
tendons and mild steel have been defined. Consequently, the detailed moment and stress calculations are
omitted.

Example Al
A prestressed concrete beam has a T
section as shown in Fig. Al. It is prestressed with fifteen 7/16-in. diameter,
7-wire strand of 270-kip grade. The locations of neutral axis and center of gravity
of steel are shown in the figure. f's, = 5000
psi, E, = 57,000V f',, E, = 28,000 psi.
Find the mean stabilized crack spacing
and the crack widths at the steel level as
well as at the tensile face of the beam at
Ofa = 30 ksi. Assume that no failure in
shear or bond takes place.
(a) Mean stabilized crack spacing
A, = 7 X 14 = 98 sq in.
Eo=15,rD
= 15i-(7/16)
= 20.62 in.
at, =1.2(A:/yo)
= 1.2(98/20.62)
=5.7in.
(b) Maximum crack width at steel level
w,naa = 5.85 x 10-s(A, /lo)Ofs
= 5.85 x 10-6(98/20.62)30
= 834.1 x 10 -6 in.
-^ 0.0083 in.
(c) Maximum crack width at tensile face
of beam
_

25 - 10.36

R` 25-10.36-3.5
= 1.31

w' ax = wma.c R,
= 0.0083 X 1.31
= 0.011 in.

Example A2
For the prestressed beam in Example

44

Al except that three additional #6 nonprestressed mild steel bars are added as
shown in Fig. A2.
Find the crack spacing and crack widths
at Af, = 30 ksi.
(a) Mean stabilized crack spacing
A, = 14(3 x 1.75 + 1/ X 7hs+ 13/a)
= 14 x 6.84
= 95.8 sq in.
Eo=20.62+3X2.36
= 27.70 in.
a,, = 1.2(At /mo)
= 1.2(95.8/27.7)
= 4.15 in.
(b) Maximum crack width at steel level
wmaz = 5.85 X 10-5(A,/^o)\f,
= 5.85 x 10-5(95.8/27.7)30
= 606.9 X 10-5
- 0.0061 in.
(c) Maximum crack width at tensile face
of beam
25 - 10.6
R`= 25-10.6-2.75

= 1.24

w'max = wma: R
= 0.0061 X 1.24
= 0.007 in.

Example A3
A prestressed concrete beam has an I
section as shown in Fig. A3. It is prestressed with twenty-four 7/16-in. diameter, 7-wire strand of 270-kip grade. The
locations of neutral axis and center of
gravity of steel are shown in the figure.
f', = 5000 psi,

E, = 57,000-v f' Es = 28,000,000 psi.

Find the mean stabilized crack spacing
and the crack widths at the steel level as
well as the tensile face of the beam at
-Af, = 20 ksi. Assume that no failure in
shear or bond takes place.
(a) Mean stabilized crack spacing
A,=7X 18 = 126 sq in.
Eo = 24irD
= 247r(7/16)
= 32.99 in.
ass =1.2(A:/ xo)
= 1.2(126/32.99)
=4.58 in.

14"

C..Ci.S_N

15 '0
7-WIRE STRAND

^j

^I

+ ._ _ C.^S. M
I5
7-WIRE STRAND

3-s

l
14'

FIG. At

FIG. A2

ri

_
co

5.5'

5.5'

N.A.

-_

MM
^

24 - 16'D
7-WIRE STRAND

O1

N.A.

r-

FIG. A 3

20 6 i
7-WIRE
4 -STRAND
-7

FIG. A 4

Fig. A. Geometric details of sections in Examples Al to A4.

PCI JOURNAL/May-June 1977

45

(b) Maximum crack width at steel level

w..a, = 5.85 X 10-5(At/Eo)ofs
= 5.85 x 10-5(126/32.99)20
= 446.9 x 10-

== 0.0045 in.

(c) Maximum crack width at tensile face

of beam

(b) Maximum crack width at steel level

= 5.85 X 10-5(A,/Xo)Of,
= 5.85 x 10`5(122.06/38.49)20
= 371.0 x 10 -5
0.0037 in.
(c) Maximum crack width at tensile face
of beam

_ 36 18.72
R`- 36-18.72 3.5

36 19.23
_
R` 36-19.23-2.79

= 1.25
w'mz = w ar R,
= 0.0045 X 1.25

= 1.2
w' as = w,naz R,
=0.0037>< 1.2
= 0.004 in.

= 0.006 in.

Example A4
For the prestressed beam in Example

A3 except that four 7 /16 -in. diameter,

7 -wire strand of 270-kip grade at the
bottom row are replaced by four #7 nonprestressed mild steel bars as shown in
Fig. A4.
Find the crack spacing and crack widths
at \f s = 20 ksi.
(a) Mean stabilized crack spacing
At=18X(3X1.75+ 1/2 X 7/16+1%6)
= 122.06 sq in.

= 20^D + 4 X 2.75
=207rX7/6+4x2.75
= 38.49 in.
a g = 1.2(At/2o)
1.2(122.06/38.49)
= 3.8 in.

NOTE: From the comparison of

crack width values at the tensile face
of the concrete, it is noted that even
at the high net stress p fs = 30 ksi, it is
possible to reduce the crack width considerably for partially prestressed or
overloaded beams with the addition of
a few mild steel reinforcing bars.
For normal levels of net stress Af 8
under working load conditions up to
15 to 20 ksi, the crack width level that
can develop at the tensile face using
the criteria developed in this paper
can become negligible.

NOTATION
Ag
As
At

a,s
E,

E5
F

46

= gross area of concrete, sq in.

= area of steel reinforcement,
sq in.
= concrete stretched area, namely, concrete area in tension,
sq in.
= stabilized mean crack spacing,
in.
= modulus of concrete, psi

= 57,000 f',_ psi for normal

weight concrete
= Young's modulus of steel, ksi
=, prestressing force in tendon,
kips, lb

f'0
f't
fa

ft

= a factor reflecting distribution

of bond stress
= compressive stress in concrete,
psi
= cylinder compressive strength
of concrete, psi
= tensile splitting strength of
concrete, psi
= stress in prestressing steel corresponding to decompression
load; ksi
= stress in prestressing steel at
any load level beyond decompression load, ksi

= (f, - J,) = net stress in pre-

stressing steel, or magnitude

of tensile stress in non-prestressing steel at any load level
= modulus of rupture of conJr
crete, psi, ksi
Icr = cracked moment of inertia, in.4
= effective moment of inertia,
in.4
= gross moment of inertia, in.4
'U
= effective beam span, ft, in.
Ma = maximum service load moment
in span, in.-lb, ft-kips
Urr = cracking moment, in.-lb, ftkips
= sum of reinforcing element cir10
cumferences
Ri = ratio of distance from neutral
axis of beam to concrete outside tension face (h2) to distance from neutral axis to steel
reinforcement centroid (h1);
value has a range in this investigation from 1.25 to 2.56,

ma,dmum bond stress as a

function of \ff; psi
maximum flexural crack width
at steel level, in,
maximum flexural crack width
at tensile face of concrete, in,
= unit strain in reinforcement, in.
per in,
deflection, in.

Acknowledgment
This investigation is part of a continuing research program on the behavior of
prestressed concrete beams and slabs, initiated by the first author and conducted at
the Concrete Research Laboratories of the
Department of Civil and Environmental
Engineering of Rutgers UniversityThe
State University of New Jersey. The paper
is based in part on the PhD thesis of the
second author under the direction of the
first.

Discussion of this paper is invited.

Please forward your comments to
PCI Headquarters by November 1, 1977.
PCI JOURNAL/May-June 1977

47