Flexural Cracks in Reinforced Concrete Beams : by Michael Chit Arthur F. Kirsteint
Flexural Cracks in Reinforced Concrete Beams : by Michael Chit Arthur F. Kirsteint
Flexural Cracks in Reinforced Concrete Beams : by Michael Chit Arthur F. Kirsteint
54-48
INTRODUCTION
It is well known that the design load of a reinforced concrete flexural member
will produce cracks in the tensile zone of the concrete. While hairline cracks
are considered inevitable, cracks of appreciable width are intolerable because
they may be a source of danger due to the possible corrosion of the reinforcing
steel. Furthermore, cracks wide enough to be visible tend to destroy the
aesthetic value of the structure.
In order to limit the width of tensile cracks the usual design practice is to
"over-reinforce," and current building codes limit the allowable stress to about
20,000 psi regardless of the steel strength. Even though steel manufacturers
have long been able to produce steel with ductility equal to and strengths
greater than those of the steels used in establishing the 20,000-psi limit, it is
felt by some that any increase of the allowable stress would cause excessive
cracking and deflection. With this limitation the higher strength steels have
not been used.
Since it is recognized that crack width increases in a nearly linear relation-
ship with the stress in the steel, restrictions imposed by the building codes
are not without basis. However, it has also been established that many other
factors also influence the crack width. Conceivably, with proper control of
these factors crack widths can be kept within a tolerable margin even with
steel stresses much higher than 20,000 psi. If this can be accomplished there
should be no objection to permitting the use of a greater allowable stress in
these high strength steels. The saving in steel resulting from such control
should justify the extra effort on the- part of the designer.
*Received by the Institute May 13, 1957. Title No. 54-48 is a part of copyrighted JouRNAL OF THE AMERICAN
CoNCRETE INSTITUTE, V. 29, No. 10, Apr. 1958, Proceedings V. 54. Separate prints are available at 50 centa each.
Dlscu11lon (copies in triplicate) should reach the Institute not later than July 1, 1958. Address P. 0. Box 4754,
Redford Station, Detroit 19, Mich.
tMember American Concrete Institute, Structural Engineer, Structural Engineering and Engineering Mechanics
Sections, National Bureau of Standards, Washington, D. C.
tMember American Concrete Institute, Structural Engineer, Structural Engineering Section, National Bureau
of Standards, Washington, D. C.
865
866 jOURNAL OF THE AMERICAN CONCRETE INSTITUTE April1958
NOTATION
A, = area of reinforcing Bteel p = ratio of the steel area to the concrete
A, = concrete area area
A1 = area of concrete affected by the ex- T = tensile force in reinforcing steel at a
tension of the steel cracked section
D = diameter of reinforcing bars T' = tensile force in reinforcing steel mid-
d = effective depth of specimens way between successive cracks
E, = modulus of elasticity of the steel Uo = bond strength of concrete
e = average minimum spacing of cracks
W"'"" = maximum width of orack at surface
!/ = tensile strength of concrete
of concrete
f, = computed steel stress (based on
Wa, = average width of crack at surface of
linear theory)
concrete
f,o = steel stress just prior to crack forma-
tion in concrete
h = total depth of beam specimen w. h- kd
kd = depth of compressive concrete in the
specimens d- kd
L' = span length of specimen subjected
to pure flexure = average width of crack at the rein-
= factor determining the diameter of forcing steel
the concrete area affected by the cJ> = ratio of the assumed effective area to
extension of the reinforcing steel the fully developed area of concrete
N = number of cracks that occur in the •· = steel strain
concrete Lo = perimeter of reinforcing steel
HISTORICAL BACKGROUND
4. Crack width decreased with an increase in the surface roughness of the reinforcing
bars.
5. Variation of the concrete cube strength did not exert any noticeable influence on the
width of cracks.
TEST SPECIMENS
Test specimens are described in Table 1 along with the amount of reinforcement anr! the
concrete compressive strengths as determined by tests of 6 x 12-in. cylinders. SpPcimens of
each size which have the same amount of reinforcement ditier only in the compressive strength
of concrete.
The specimens were cast in ~teel molds and remained in the molds for 5 days in a damp
~ondition. Then they were subjected to moist curing until 2 days prior to testing at 28 to
30 da~•s. Three 6 x 12-in. cylinders were cast with each specimen and cured under the same
conditions as the specimen to determine compressive strength of the concrete.
The reinforcing bars used in this test series were of intermediate grade steel which met the
requirements of ASTM A 15-54T for billet. steel bars and the deformation of these bars complied
with the ASTM A 305-53T requirements.
Concrete used to cast these specimens was proportioned for nominal compressive strengths
of 2000 and 6000 psi at 28 days. Type I cement, sand, and gravel were proportioned by
weight at approximately 1:3.9:6.1 with 11 gal. water per sack to produce the 2000-psi concrete,
and l :1.4:2.3 with 5 gal. water per sack to produce the 6000-psi concrete. Both sand and gravel
were siliceous ag;gregates and the maximum size of coarse aggregate was 1 in.
Instrumentation for measuring the width of tensile cracks in the specimens consisted of two
rows of Tuckerman optical strain gages placed equidistant from the longitudinal center line
of the specimen and staggered to overlap each other by 0.5 in. This instrumentation was used
to include every crack that formed within the region of constant bending moment. Electrical
re~iHt.ance wire strain gages were applied in pairs on the reinforcing steel at midspan to deter-
mine the steel strain. The location of the strain gages and the loading method are shown
~chcmatically in Fig. 1.
TESTING PROCEDURE
All of the specimens were placed in the testing machine with the tensile side up and loaded
near the ends (Fig. 1). This arrangement offered an easily accessible flat surface for installation
of the Tuckerman gages. Rockers were used to support the specimens at their quarter-points
and a roller and spherical bearing unit were used to apply the load near the ends. All loads
FLEXURAL CRACKS IN BEAMS 869
--r-------TI--------
\1
\
!" II
6 11 5"( 2 11 11 5t II"
~\.1• •I•I• I•
--~--~......_-...L..-...L..--
LOCATION OF TUCKERMAN
STRAIN GAGES
and reactions were transmitted to the specimens by steel bearing plates 1 in. thick, 3.5 in. wide,
and long enough t.o cover the full width of specimen.
Zero load readings were made after the specimen was preloaded to 500 lb and the load
increments depended on the size of the specimen, the range being from 500 to 5000 lb. Each
increment of load was applied at a rate of 1000 lb per min. All gages were read and the location
and extent of cracks were recorded immediately after application of each increment.
The extension indicated by the Tuckerman gage was taken to represent the width of crack
spanned by the gage, and no attempt was made to correct the crack width for the strain in
the concrete. Although the gages indicated tensile strain in the concrete before the cracks
formed in the gage length, this strain was small compared to the strain indicated after the
<"racks appeared. Furthermore, Watstein and Mathey 9 have reported test results which
definitely establish that compressive strains can exist on the tcn~?ile face of concrete midway
between successive cracks. Therefore, due to the random occurrence of cracks and the un-
certain magnitude of concrete strain between cracks, it was deemed inadvisable to attempt
the correction of the crack width measurements.
TEST RESULTS
This test series of 16 specimens was designed and tested by the late Arthur P.
Clark as a continuation of his previous investigation. 7 Specimens identical
except for compressive strength of the concrete were used to determine the
effect of compressive strength on crack formation. The data presented herein
seemed to indicate no appreciable effect for compressive strengths from 2000
to 6000 psi. If there was any effect it was small enough to be veiled by the
normal scatter as shown in the six graphs of Fig. 2. Each graph represents
the test results of beams of identical design. The black points represent data
.o 10.-----r----r---r---.---, .OJ Or--r--~-,---.---, .OJOr---~---r---~--.---, co
·¥- ......
uz 0
c-
a::- SPECIMENS SPECIMENS SPECIMENS
UILI •- 6-7.J.-4-1,2 a 3 REF. 7 •-6-11-6-1
a
INVESTIGAT;:~· ~
IA..u -o--3}T HIS 2 REF• 7
-o--I}THIS
0~
%~
o- 2 INVESTIGATION .0061- 0 - 4 •
a::
:1:~
-~~~--
... (I) l-
0 0
-ILl
.004t- c70
!~ . .
0
• .J
ILic;; z
.00~
C!lz }>
r
~~ 0"T'1
ILl
~!ii -i
ANALYSIS
The first three assumptions are those normally used in crack investigations.
Assumption 4 can be easily verified by strain measurements as Clark 7 and
Watstein and Mathey 9 have done. As mentioned previously strains on the
cracked tensile face of a beam are quite small compared to the crack width
and may be of either sign. Therefore, these strains were neglected for simplicity.
It has been pointed out that the deformation of the reinforcing bar causes
a shear deformation of the concrete area immediately surrounding the bar.
The deformed area is only a portion of the total concrete area of the tensile
zone, and the authors consider its relative magnitude an important factor in
determining the width and spacing of cracks. The illustration in Fig. 3 (a) is a
greatly exaggerated profile of the reinforcing bar and the deformed concrete
area.
Considering a semi-infinite elastic body subjected to a uniform load over a
portion of its surface, the maximum deflection in the body is not a constant but
CXl
-..J
to
No. cracks spacmg, m. m. 1n. cracks spacmg, m. tn. m. cracks spacmg, m. In. In. 0
c70
f, ~ 15,000 psi f, ~ 20,000 psi f, ~ 25,000 psi z
)>
1 2 20.0 0.00083 0.00091 4 10.0 0.00148 0.00191 9 4.5 0.00197 0.00256 r
2 - - - - 9 3.6 0.00133 0.00247 12 3.3 0.00205 0.00460
3 7 5.1 0.00206 0.00285 9 4.1 0.00286 0.00469 10 3.8 0.00371 0.00659 0
.,
4 7 5.3 0.00265 0.00375 7 5.0 0.00405 0.00535 10 4.0 0.00404 0.00669
5 10 8.3 0.00170 0.00205 10 5.9 0.00283 - 12 4.9 0.00319 0.00355 -i
6 10 5.7 0.00170 0.00250 13 4.6 0.00250 0.00460 n 4.6 0.00349 0.00499 I
7 9 5.8 0.00181 0.00323 9 5.8 0.00269 0.00456 10 5.3 0.00367 0.00582 m
8 16 4.2 0.00184 0.00335 16 4.2 0.00295 0.00522 17 3.8 0.00378 0.00706
9 1 - 0.00141 - 7 7.8 0.00210 0.00334 10 5.6 0.00328 0.00541 )>
10
11 -
5 7.9
-
0.00131
-
0.00272
-
10
-
5.9
-
0.00238
-
0.00497
-
11
10
5.5
7.0
0.00342
0.00300
0.00736
0.00428
s:m
12 12.3 0.00228 12 0.00527 70
5 0.00179 10 6.4 0.00284 0.00416 5.6 0.00400
13
14
-
4
-
31.6
-
0.00205
-
0.00286
10
12
7.2
5.7
0.00298
0.00364
0.00610
0.00550
10
13
7.2
5.3
0.00484
0.00468
0.00878
0.00735
n
)>
15
16
15
12
5.5
5.3
0.00191
0.00169
0.00282
0.00319
16
17
5.1
4.2
0.00330
0.00222
0.00469
0.00502
17
18
4.8
4.0
0.00437
0.00286
0.00673
0.00659
z
()
f, ~ 30,000 psi f, ~ 35,000 psi f, ~ 40,000 psi 0
1 10 3.7 0.00283 0.00353 10 3.7 0.00366 0.00449 10 3.7 0.00446 0.00546
z
()
2 12 3.3 0.00253 0.00621 12 3.3 0.00317 0.00783 12 3.3 0.00393 0.00945 70
3
4
11
10
3.5
3.8
0.00429
0.00464
0.00805
0.00796
11
10
3.5
3.8
0.00512
0.00549
0.00941
0.00929
12
10
3.2
3.8
0.00605
0.00631
0.01071
0.01062 ~
m
5 15 4.2 0.00400 0.00683 15 4.2 0.00490 0.00840 16 3.9 0.00543 0.01022
6
7
16
11
3.8
4.8
0.00374
0.00420
0.00630
0.00698
17
-
3.6
-
0.00440
-
0.00800
-
17
-
3.6
-
0.00510
-
0.01000
-
z
(/)
8 17 3.8 0.00478 0.00895 17 3.8 0.00585 0.01091 17 3.8 0.00674 0.01257 -i
9 10 5.6 0.00440 0.00739 10 5.6 0.00583 0.00991 10 5.6 0.00682 0.01166 =i
10 11 5.5 0.00454 0.00954 11 5.5 0.00558 0.01170 - - - - c- i
11 11 5.8 0.00336 0.00538 12 5.7 0.00456 0.00672 12 5.7 0.00547 0.00811 m
12 12 5.6 0.00510 0.00637 12 5.6 0.00622 0.00775 - - - -
13 10 7.2 0.00657 0.01108 10 7.2 0.00800 0.01298 10 7.2 0.00956 0.01566
14 13 5.3 0.00567 0.00903 14 4.9 0.00650 0.01068 14 4.9 0.00772 0.01235
15 18 4.5 0.00517 0.00863 18 4.5 0.00622 0.01049 18 4.5 0.00737 0.01274
16 18 4.0 0.00360 0.00821 18 4.0 0.00439 0.00991 18 4.0 0.00506
-
)>
-g_
......
'()
U1
CXl
FLEXURAL CRACKS IN BEAMS 873
~
is conceivable that the deformed area
will be distorted when it intercepts the
exterior surface of the beam or the
At•ASSUMED
88
FULLY DEVELOPED
boundary of a similar deformed area EFFECTIVE AREA AREA • m2 A9
around another reinforcing bar in close
(c)
proximity. This effect can be taken
into account approximately by de- Fig. 3-Effective concrete area
ducting the overlapping areas from
the total deformed or fully developed area as indicated in Fig. 3 (c). The
ratio of the assumed effective area to the fully developed area is designated as
4> and written as
A,
<t>=-
m2A,
The values of 4> for the test specimens reported herein are given in Table 1.
A typical portion of a beam subjected to pure bending is shown (with the
width of the cracks exaggerated) in Fig. 4 (a). It is assumed that the flexural
stress in the concrete between successive cracks is negligible. Therefore, the
bond stress between the steel and the concrete is one of the determining factors
in this investigation. It is evident from the free body diagram in Fig. 4 (b)
that the bond stress was assumed to vary linearly from zero at Section 2-2 to
the maximum at Section 1-1, and that the tension in the steel T' at Section 2-2
is smaller than T at Section 1-1 by the amount of the shear forces along the
bar or
1 1 Uoel:o
7 - T = - 4 - ........................... (1)
The concrete tensile stress at the midsection of the segment of length e due
to the direct tension of the bar which is transmitted through bond is
,. aL aL
t. L J
T-' c€8"o
I. -- .,~--~~
-T
(b)
FREEBODY DIAGRAM OF REINFORCING
(a) STEEL BETWEEN SECTS. I- I a 2-2
Fig. 4-Flexural member subjected to uniform bending moment over portion of span
4f,' A,
e ... -:;;,:- Lo ............................ (3)
and by incorporating the definition 4> = A 1/m 2A. we have
nW,
-u=~
f,.- f,o
............................ (5)
and when n is large compared to 1
By substituting I/ from Eq. (6) and e from Eq. (4) into Eq. (5), we have
Eq. (4) and (7), as derived through this simple semi-empirical approach,
are in such a form that the constants are easily evaluated empirically. The
value of m = 4 was estimated through successive approximations, and f,' /u.
and j,. were determined from straight lines fitted to the data shown in Fig. 5
and 6. These data are from the 16 beams reported herein, and from Clark's
FLEXURAL CRACKS IN BEAMS 875
~
~
a: LEGEND:
0
u.. •- CLARK'S DATA 7
0
o- DATA FROM THIS INVESTIGATION
~
o_
~:!!:
(/)....
11.1
>
0
~cv
a:
11.1
~
~
:::1
:::E
z
i
20
0
LEGEND:
•- CLARK'S DATA 7
o- DATA FROM THIS INVESTIGATION 0
0 •
..J
w
.....
...
Ill
• •
...
Cl
•
:11::
u
Cl
a:
u
1&.
0
J:
t-
o
3t
16 t Dfs
Es
Fig. 6-Experimental results used to determine f/ ju. and f•• for Eq. (7)
f,' • D (f. - / •• )
W .=-mrj>
Uo E.
876 jOURNAL OF THE AMERICAN CONCRETE INSTITUTE April1958
and
f,- 21)~]
W, = 5q,D [ / ....................... (!))
Note from Eq. (8) and Fig. 5 that the average minimum spacing of cracks is
directly proportional to the product ¢D. This indicates that the arrangement
and the diameter of the steel are of equal importance. This is illustrated by
Clark's data 7 on beams listed in Table 3. These beams differed in thickness
of concrete cover from their respective companion beam specimens. This
change produced appreciable effects on the observed crack widths. These
effects can be explained by this analysis. Since the fully developed area
(m 2A,) remains unchanged for each set of beams, then from¢ = A 1/(m 2 A,) a
change in concrete cover causes a corresponding change in ¢. From Eq. (8)
and (9) it can be seen that the change of concrete cover produces a correspond-
ing change in the computed values of spacing and width of cracks. A com-
parison between the observed and computed values of spacing and width of
cracks is shown in Table 3. The results indicate that the theory predicts the
spacing and width of cracks for various thicknesses of concrete cover.
Because of the complex nature of the problem the theory can be expected
to apply only to average values of crack width. Although a comprehensive
statistical study of the dispersion of the data on widths of cracks was not
made, one measure of the variability of the widths of cracks in a beam was
obtained by determining the ratios of the maximum to average widths of
FLEXURAL CRACKS IN BEAMS 877
cracks for 70 beams at each of six stress levels (15,000 to 40,000 psi computed
steel stress). The lengths over which the cracks were measured for these
beams ranged from 36 to 66 in. and the ratio of maximum to average width
of crack ranged from 1.08 to 2.64. * The average value for this ratio was found
to be 1.63, which agrees with the value of 1.64 reported by Clark. 7 The
relationship does not necessarily apply to the longer members used in practice.
SUMMARY
1. A new concept is introduced into the analysis of the problem of crack
formation in a portion of a beam under pure bending. After some initial cracks
have orcurred the tensile force in the conc-rete is resisted by an "effective area"
of concrete immediately surrounding the steel, which is less than the total
area of the ronerete in the tensile zone of the beam. Further cracking of the
concrete is attributed to the direct tension developed through bond and due
to the strain in the steel.
2. The minimum average spacing of cracks was found to be proportional to
the product of the diameter of the reinforcing bar and the parameter c/J, which
is dependent on the general arrangement and diameter of reinforcement. The
analysis indicates that the constant of proportionality is a measure of the
mtio f,'/uo. Watstein and Seese 4 and Brice 11 have shown that a linear cor-
rection can be applied to the ratio ft' /uo to extend the solution to cover a
variety of different deformed bars. This correction is determined from the
relative bond values of the bars.
3. The average width of the cracks at the steel is giyen by the product ot
the average minimum spacing of the cracks and a function of the computed
steel strain.
4. For the lengths of span subjected to pure flexure in these tests (from 36
to 6-fi in.) the ratio of the maximum to average width of crack was found to
range from 1.08 to 2.64 with an average valu·e of 1.63.
5. Test results reported herein indicate that concrete strength in the range
from 2000 to 6000 psi has practically no effect on the formation of cracks.
*Beams containing only one or two cracks at a given stress level were not ineluded in these calculationa.
878 jOURNAL OF THE AMC.I<ICAN CONCRETE INSTITUTE April 1958
ACKNOWLEDGMENTS
The test data were accumulated by the late Arthur P. Clark while he was a
research associate of the American Iron and Steel Institute Fellowship at the
National Bureau of Standards. Appreciation is also expressed to C. Burton,
student trainee, who performed the major portion of numerical computations
in this investigation.
REFERENCES
1. Watstcin, D., and Parsons, D. E., "Width and Spacing of Tensile Cracks iu Axially
Reinforced Concrete Cylinders," Journal of Research, National BureaL! of Standard~, V. 31,
RP1545, 1943.
2. Bornemann, E., "Increasing the Tensile Strength and Avoidance of Formation of Cracks
in Concrete," Proceedings, Second Congress, International Association of Bridge and Struc-
tural Engineering, 1936.
3. Colonnetti, G., "The Tensile Strength of Stressed Parts in Reinforced Concrete," Pro-
ceedings, Second Congress, International Association of Bridge and Structural Engineering,
1936.
4. Watstein, D., and Seese, N. A., Jr., "Effect of Type of Bar on Width of Cracks in Rein-
forced Concrete Subjected to Tension," ACI JoURNAL, Feb. 1945, Proc. V. 41, pp. 2\J3-304.
5. Bjuggren, Ulf, discussion of "Bond and Anchorage" by T. D. Mylrea, ACI JouRI-IAL,
Part 2, Dec. 1948, Proc. V. 44, pp. 552-1-552-4.
6. Saliger, R., "High-Grade Steel in Reinforced Concrete." Proceedings, Second Congress,
International Association of Bridge and Structural Engineering, 1936.
7. Clark, Arthur P., "Cracking in Reinforced Concrete Flexural Members," ACI JouRNAL,
Apr. 1956, Proc. V. 52, pp. 851-862.
8. Wli.stlund, G., and Jonson, P. 0., "Investigation on Formation of Cracks in Reinforced
Concrete Structures," Preliminary Publication, Third International Congress of Bridge and
Structural Engineering, 1948.
9. Watstein, D., and Mathey, R. G., "Evaluation of Width of Cracks in Concrete at the
Surface of Reinforcing Steel by Means of Tensile Bond Specimens," RILEM Symposium on
Bond and Crack Formation in Reinforced Concrete, Stockholm, June 1957.
10. Timoshenko, S., and Goodier, N., Theory of Elasticity, McGraw Hill Book Co., Inc.,
New York, 1951, p. 368.
11. Brice, L; P., "Theorie de Ia Fissuration des Pieces en Beton Arme," Annales, l'Institut
Technique du Batiment. et des Travaux Publics, (Paris) June 1952.