See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/312307773

CRACK CONTROL - AN ADVANCED CALCULATION MODEL - PART l: REVIEW OF

CLASSIC TESTS

Article · December 2016

CITATIONS READS

5 165

1 author:

Andor Windisch
Budapest University of Technology and Economics
52 PUBLICATIONS   217 CITATIONS   

SEE PROFILE

All content following this page was uploaded by Andor Windisch on 13 January 2017.

The user has requested enhancement of the downloaded file.

CRACK CONTItOL.- AM ADVANCED
CALCULATION MODEL -
PART l: OF CLASS1C TESTS
Dedicated to Prof. Gallus Rehm on the occasion ofhis 90"1 birthday

Andor Windisch

Crack control is a fundamental part ofdimensioning: in many cases it governs the amount ofreinforcement
in R/'C members. In Part I after the overview of development of calculation methods in international codes
since 1989, the data, records and notations offour classical papers with excellently documented tests on
tensile members and beams will be studied and evaluated. Secondary crachs get a new Interpretation.
Keywords: Cracking, Goto-cracks, pnmary and secondary cracks, bond, concrete cover, Cracking model

l. INTRODUCTION influencing factors are taken into consideration at the proper

Position.
Cracking is an inherent characteristic of structural concrete.
Since over 60 years intensive research is made on this topic. 2. DEVELOPMENT OF CALCULA-
National (e.g. ACI, DIN) and international standardization
bodies (e.g. CEB,fib) directand summarize the achievements TION METHODS SINCE 1989 IN
J?,. - ... t: ... . •, - *.:.-. - r» ----- i1__ il. - •* r - J _ 1 /-*_ . ____ -
J - i O l V i LlliiC U; Lliilfc. AV^Vfcllti^ UiC IVJLUU&i ^\JU&^\Jl\J

published. In a paper Baläzs et al. (201 3) an overview is giv-

\±,*JL4*) WaD
INTERNATIONAL CODES
en about the backgrounds of the design rules for serviceabil- The calculation method of the (characteristic/design) crack
ity limit states in MC 2010 and showed calculation examples. width made the following development during the last 25
The current model assumes that: years: in Table l the main formulas are summarized.
- The bond strength is constant and independent of the slip Some comments:
• Comparing the crack width formulas of MC1990 (1993)
- Each rebar in tension is a tensile member within A cef with EC2 (2003) yields l smax = 1.7 sm. The definition of
around. (Note: according to the definition of Ac rf in the äs the distance between the two sections with the point of
codes, the tensile member is eccentrically loaded.) zero slip is theoretically correct, nevertheless in the figure
— The concrete strain is mentioned in the basic formula but explaining the stabilized Cracking Situation = srmax is de-
neglected at the calculation picted: these two distances have no correlation with each
— The influence of the concrete cover is taken into account äs other, hence from this point the theoretical correct char-
extension of the transmission length of bond acter of the deduction vanishes. In MC1990 the constant
— No distinction between RC members in pure tension or in term, 50, does not occur. In MC2010 the position of the
flexure 2 * ls,max symbol
*
is correct: it refers to the two sides of the
— No special rules for slabs with remote rebars are given. crack, nevertheless a) l smax is only a part of the length over
In this paper on the contrary which slip between concrete and steel occurs, the length
— The effective tension area of concrete, Ac ef, presents a fair- between the crack and the section without slip is longer,
iy differem reievance. 's,max *s suPPoseclto be the load transfer length to cause the
— The concrete elongation around the reinforcement is con- cracking of the concrete cross section of Agrfr cross sec-
sidered directly. tion, b) a comparison of the formulas reveals that instead
— The influence of concrete cover and rebar distance, resp. of 1.7*50, in MC2010 2*c occurs, whereas the second
are considered in a theoretically correct way term is identical with the lsinax term of MC1990. In the k*c
- The difference in crack width of RC members under pure term k is an empirical parameter to take into consideration
J* ' 'J J ' j
the influence of the concrete cover. In the background pa-
— The impact of shrinkage on crack width is discussed per (Baläzs et al. (2013)) the k*c term refers to the load
In Part I - instead of inscrutable databanks - the data, records transfer of bond forces from the rebar surface to the con-
and notations of four classical papers with excellently docu- crete surface. We returned to the ancient problem of the
mented tests on tensile members (Broms and Lutz (1965), researchers: "ancient" data bank comparisons revealed
Goto (1971), Scott and Gill (1987)) and beams (Rusch and that the 0s/psef parameter is weak. Already Ferry Borges
Rehm (l 963)) will he studied and evaluated (1966) wrote: "It must be emphasized that if the influ-
In Part II of present article the well-known formula of the de- ence of the cover were omitted, a very weak correlation
sign value of crack width gets an advanced Interpretation. All between s and 0/A ceff_ would be obtained."

2016 41
 rc-ns is related to the mean value of the tensile cracking a construction defect? Why concrete cracks; Con-
strength. This gives the impression that ~zr\s ri-ns has crete Cracks Occur....; Industry Guidance; When is Cracking
a strength character, which is not the case. Expected? ASCC (American Society of Concrete Contrac-
• EC2 specifies that the calculated crack width refers with- tors) Position Statements; Managing Owner Expectations;
in a region close to the bonded reinforcement (i.e. within Principles and Precedents. Part 2: Case law on concrete crack-
the effective concrete area in tension). Note: äs the ef- ing; Legal Principles; Cracking is "Normal and Expected".
fective tension area includes the concrete surface hence
the calculated crack width seems to be constant across
the concrete cover and does not follow the curvature in 3. LESSONS FROM CLASSICAL EX-
members in flexure.
• In MC 1990 the effective concrete area in tension (Acef) PERIMENTS
accounts for the non-uniform normal stress distribution In this section the records and data of four excellently docu-
by bond forces into the concrete cross-section at the end mented reports are evaluated and discussed. The conclusions
of the transmission length. "By means of the method the serve the understanding of the cracking procedure and the im-
design crack width within the effective tension area may provement of the crack width formula.
be calculated. It should be noted that outside this region,
larger cracks may occur." This is a correct comment but 3.1 Tensile specimens of Broms et al.
does not assist the designer.
• The average steel strain terms in MCI 990 and MC2010 Broms and Lutz (l 965) tested seven ~2.2 m long and six short
are identical. tensile specimens reinforced with bars in various arrange-
• The ß = 0.6 value is in contradiction with the assumption ments. The lengths of the short tensile members were chosen
of the uniform bond stress distribution (the correspond- equal to the mean value of primary crack spacings observed
ing value would be 0.5). for the long members. The dimensions of the specimens
• The formulas ignore the tensile deformation of the con- with rectangular cross section were 76 x 240 mm, of those
crete along the 2 * l smw distance. This results in a calcu- of quadratic cross section 135 x 135 mm. T-RC6 specimen
lated crack width on the safe side, nevertheless it yields was reinforced with one 025 (#8) rebar (fR = 0.054, c = 25
not economical amounts of reinforcement. mm), whereas the others with four 012.7 mm (#4) rebars (fR
The standardization body of the US handles the crack con- = 0.043, c = 19 mm), hence the As/Ac values were identical.
trol problem in a different way: The relevant ACI commit- The average compressive strengths measured on 150 x 300
tee publishes reports (2001), (2013), where in addition to the mm concrete cylinders were 30 to 40 N/mm2 and the average
principal causes of cracking- recommended crack-control tensile strengths determined on split cylinder teste were ~3
procedures in flexural members are presented. In a later re- N/mm2. The crack distances and the crack widths were mea-
port performance-based details that can mitigate and control sured on each long side of the cross section along five lines
cracking are reviewed. Detailing rules for two-way and one- and in the middle of the short side, at three stress levels: 314,
way slab Systems and those of columns are presented. 471 and 628 N/mm2 (in case of T-RC6 at 593 N/mm2), resp.
The legal aspects of cracking are regularly considered in the (Note: the reported max. crack widths are the averages of the
relevant US literature. The chapter's headings of Coleman two widest relevant cracks!)
(2013) are well representative: Studying the results the development of the cracking pro-
Cracking... Defect or Normal? Part 1: When is concrete cess in a member under centric tension can be understood and

Table 1: Calculation formulas of the last 25 years in the internationai codes

EC2 [6] MC 1990 [7] MC20IO(8]
design crack wk - ß s™ £w Wj =
widlh p =1.7 **Hj«fe» -*«**.) 2 /,_<*.„-*„-*.>
average crack srm 1t , M li \ — f '• ^ -*
spacing
for stabüized
- 50 + 0.25 kl k2 0/pr /
|WM*
- **
- 3.6ps,ef n C T

cracking 4 r^ pifi
for Mndc *$.max
t """
—

crack
a*2 0 l
formal ion 2rfek * 1 + afps.ef
average strain *sm = £,m ~ £.-m ~ £c, ~
of
reinforcement
K"**©1) £sm ~ £cm ~ £s2 ~ P£$r2
°>-ß'°~
'
*-•
.,
' 'fr C,l,

ß(inEC2ß ; ) 0.6 0.6 0.6

42 2016
 explained. Fig. l shows the development of the number of area is less than 70 around the rebar. (It must be admitted
the crack intervals (length of the specimen divided by mean that the crack width formula in MC2010 is valid for structural
crack spacing). members with concrete covers of max. 75 mm, in slabs the
The test results of specimen T-RC6 indicate that: bar spacing shall not exceed the lesser of 1.2 times the slab
- Up to os = 314 N/mm2 at positions L3 and R3 three times thickness and 300 mm. Especially this latter seems to be a bit
more crack intervals developed than at position TC (27 vs. 7). too big.)
- At position L3: with increasing steel stress the number of A comparison of the number of crack intervals in case of
crack intervals increased considerably. From 27 pcs. at os the quadratic specimens T-RC10 and T-RC11 reveals the se-
= 314 N/mm2 to 36 pcs. at os = 604 N/mm2, whereas at quence of the Steps how the crack pattern develops. In prin-
position TC the number of crack intervals did not change. ciple two processes can be distinguished:
(Note: at os = 604 N/mm2 the calculated tensile stress in 1. Increasing the tensile load primary cracks develop at the
the concrete member would be 14.5 N/mm2!) This means weakest cross sections of the R/C member. These primary
that the tensile Stresses due to the transfer of bond forces cracks are "through" cracks and are situated outside the ac-
do not reach the fiber äs far from the rebar surfäce. tual transmission lengths of the already developed primary
In case of specimen T-RC5 cracks. The transmission lengths increase with the increas-
- At os = 314 N/mm2 the number of crack intervals were at ing load. Remarkable: at all specimens the occurrence of
positions L3/R3 -34, whereas at position TC only 7. about seven primary cracks can be observed.
- At os = 628 N/mm2 at position L3/R3 the number of crack 2. Starting from the primary cracks, due to the tensile Stresses
intervals increased to -44, at position TC to 8 only induced by the bond between the rebar and the concrete
On contrary, in case of specimen T-RC8 around and the high steel strains the well-known Goto-
- At os = 314 N/mm2 at positions L1/L5/R1/R5 25 - 33 cracks develop. The stress state caused by bond forces is
crack intervals developed, at TC 38, whereas at positions quite Symmetrie to the bar's axis. Depending on the slip
L3/R312-16only. and the local concrete tensile strength the Goto-cracks have
- Increasing the steel stress to os = 628 N/mm at positions different extents. Depending on the distance of the rebar
L1/L5/R1/R5 2 the number of crack intervals increased to from the concrete surfäce these Goto-cracks may arrive at
33 - 40, at position TC to 54, whereas at positions L3/R3 the concrete surfäce and be recognized äs secondary cracks
to 19-20 only! and let decrease the crack distances. As a secondary crack
In case of T-RC7 the measuring line Ll is about in the same does not cross the whole cross section like primary cracks
distance t from the next rebar surfäce äs the line L3. Along do hence the steel stress in a secondary crack is always less
line L3 more secondary cracks developed than along Ll. than in the primary crack. As we are interested in the upper
Along L3 the tensile Stresses due to the bond Stresses of 5% fractile value of the crack width, the width of all these
two adjacent rebars affected the development of secondary secondary cracks are out of our interest (nevertheless, the
cracks. Note: some authors and codes tackling the slab-like secondary cracks "perturb" the development of the mean
reinforcement pattern proposed a term s/2 in the mean crack value of the crack distance and the mean rebar elongation!).
distance formula. If the distance of the rebars is less than - 2* 50 mm and the
At all of these specimens it can be perceived that the number concrete cover is less than 25 mm then it may occur that the
of crack intervals shows a pattern of secondary cracks along secondary cracks of these rebars "meet" each other and a füll
eight intervals. The numbers of crack intervals at position TC crossing crack develops which behaves like a primary crack.
in case of T-RC5 and T-RC6 reveal that at the contrary to Comparing the number of crack intervals at the positions
the proposals of Leonhardt (1987) and the codes: rebars can TC of T-RC6 and T-RC5, äs well äs at positions L3/R3
control the crack width within a distance of 70, but (obvi- of the specimens T-RC7 and T-RC8 it can be diagnosed
ously independent of the rebar diameter) less than -75-100 that beyond t - 100 mm (in case of 025 rebar) and t -
mm only! Moreover, "control" should mean: to be able to 75 mm (in case of 012.7 rebars) the Goto-cracks cannot
cause concrete elongation, therefore, the affected/controlled cause secondary cracks on the concrete surfäce, even if the

Flg. l: - Development of numbers of crack intervals at 314 N/mm2, 471 N/mm2 and 628 N/mm2 steel Stresses (in case of T-RC6 at 604 N/mm2; in
columns from the ieft to the right) of Broms and Lutz f 1965).

T-RC6 T-RC5 T-RC7 T-RC8

s $s «It 16 U 17 'S 4S 55
tt u ie TC
9 12 2« U- -W » U 17 1 U » U- -Kl g 10 22 1» 24 31 U- *> 17 23 30 .M 37 39 "- •• K 32 41 41
21 25 2» U- *) 21 23 2* 25 A3 » U- K.' 2? 5? 58 2» 31 55 «- •• •» 3S 45 48 24 23 3« U- 4C 21 25 2*
2t Jl J7 U. O JU 27 32 3» 35 4} 47 U. ** JL« « 42 42 25 2« 31 U- *» 2« 3« 33 l? 29 2» "• *i O 21 21
U jj jft u. Ju 22 27 30 25 13 * U. •« 29 M 3t 37 3» 47 U- « a JU st 5l 53 23 23 2* "- ** 1» 27 2»
« u n "- * » 17 21 » U 12 tf- •» 11 14 25 22 24 2« "- » 21 2» 33 2» 3« 5l u •• * 2« 32 «

T-RC10 T-RC11
l? 25 2t 3« 31 5?
n. u n n._K_tK
u « n u. •n I« It U 4* 47 45 U- • • -»l M 4S 51
W %* ** L3 * * •K U 18 23 3« 3« y> U- * 31 M 3'
** ** ••
15 2« 22 U- -W 15 17 21 35 43 51 U- • • JU j? 52 52

2016 43
 calculated uniformly distributed concrete tensile stress in cordingly the corresponding data in the databanks are not
cross section was 15 N/mm2! This shows that in the cross correct.
sections along the transmission lengths füll of Goto- and - Internal cracks usually start at a steel stress less than 100
secondary cracks, a quite complicated, not uniform stress N/mm2, shortly after primary cracks are formed. They first
state prevails. develop around ribs near the primary cracks then with
After injection and hardening of resin into the short specimens increase in steel stress or with repetition of load, at ribs
under loading they were sawed. The crack pattems reveal that progressively farther from the primary cracks. An adjacent
the secondary cracks around one rebar are not at symmetri- (inclined) internal crack sometimes grows to such an ex-
cal positions, the appearance of the secondary cracks is not tent that it appears at the inner face of the primary crack.
imperative. The positions of the Goto- and secondary cracks - Internal cracks need not necessarily occur.
show no regularity along the transmission lengths and no - The width of both, the Goto- and secondary cracks arises
symmetry to the rebar's axis either. This observation explains not äs slip from both sides of the crack but the section be-
die different number of secondary cracks along tlie measur- tween the primary crack and the Goto-crack moves in the
ing lines in identical positions. Near to the primary cracks direction of the primary crack hence the direction of the
the Goto-cracks show a -45° inclination, farther they become slip does not change.
steeper.
As preliminary conclusion it can be stated that 3.3 Tensile members of Scott and Gill
- In a tensile member with a given ratio of 0/pef and given (1987)
rebar pattern in the cross section very different numbers of
crack intervals (i.e. primary cracks) with accordingly very Short-term longitudinal reinforcement strain distributions in
difFerent rnean and maximum crack widths develop. reinforced concrete tension members were measured using
- Secondary cracks with very different distances and widths electric resistance strain-gauges. The 2.5 mm long gauges
may occur along a single transmission length of the rebar. were installed at 12.5 mm centers in a 1000 mm long duct
milled longitudinally through the center of the rebar. Twelve
3.2 Goto's fundamental tests specimens with square cross-sections ranging from 70 mm x
70 mm up to 200 mm x 200 mm were tested, all reinforced
In his fundamental tests Goto (1971) investigated the crack with a single rod positioned centrally in the cross-section.
formation by providing narrow holes parallel to reinforcing One specimen was of 100 mm x 300 mm dimensions. The
bars for ink injection in axially loaded specimens. The cylin- rebars were either of 12 mm or 20 mm diameters, both plain
der compressive strength and the tensile Splitting strength of mild steel (R) or ribbed high yield (Torbar, T) rods. The strain
the concrete were about 30 N/mm2 and 2.8 N/mm2, resp. The distributions for five specimens (100T12, 100T20, 200T20,
side length of quadratic cross sections was 100 mm for 019 300/1OOT20 and 140R20) were published.
mm rebars and 120 mm for 032 mm rebars. The steel stress at Note:
injection was 300 N/mm2 (yield strength 390 N/mm2). In each - The measured strains show the steel elongation and the
case the maximum crack spacing was assumed approximate- direct concrete contribution hence the slip and crack width
ly to 250 mm. Shallow notches were cut in the specimens values refer to the steel surface, the crack width on the
beforehand such that primary cracks would first form at spac- concrete surface is difFerent.
ings of 220, 240, 260 and 280 mm. resp. (The 2-1 ^ values - Due to the milled ducts along the rebars, their longitudinal
äs proposed by MC2010 are 247 mm for specimens with 032 tensile stiffness is less than in normal cases.
rebars and 270 mm for 019 rebars, resp.) It is a pity that Goto The perceptions are äs follow:
did not report on the development of the crack widths. - The crack distance and the distance between the two points
Observations, preliminary conclusions: of no slip on both sides of a crack are not equal. This was
- The primary cracks - which occur mostly along the re- theoretically proven by Windisch (1989).
gions of the member where still no bond slip occurred and - The rate of concrete contribution increases with increas-
develop from the most tensioned outer fibers of the mem- ing crack distance. The tensile stress in the concrete cross-
ber according to the Bernoulli-Navier hypothesis - are section reaches its max. intensity around the max. fractile
mostly perpendicular to the member's axis, whereas the value of the tensile strength just before a new crack devel-
secondary cracks - which develop outward from the inner ops. This means that in case of the mean crack distance
Goto-cracks - are inclined. the tensile stress in the concrete cross section is definitely
- The development of the crack patterns is of random char- (much) less than the f te value. Even in case of the char-
acter: A new primary crack and/or secondary cracks need acteristic length, lsjiiax, the bond stress to be taken into ac-
not occur within the longest crack distance i.e. on any side count is not any bond strength, but depends on the steel
of the widest crack. In case of the specimen with 019 stress level in the crack, os. (Note: depending on the dead
mm rebar primary cracks occurred at a steel stress of 150 to live load ratio the steel stress in the primary crack to be
N/mm2 between the cracks in 280 mm distance, at a steel taken into account in SLS can be difFerent.)
stress of 185 N/mm2 between the cracks in 260 mm dis- - In case of a given crack distance the concrete contribution
tance, whereas secondary cracks at 300 N/mm2 steel stress increases with increasing loading (increasing steel stress/
between the cracks in 240 mm distance, near to the cracks. slip in the crack). The rate of increase slows down. Never-
- The secondary cracks developed in case of both rebar di- theless, when a new crack develops within this crack dis-
ameters from Goto-cracks at a steel strain in the crack of tance then the contribution drops and the increase begins
0.0015, and Start from rebar ribs in (3-3.5)0 distance in afresh.
case of 019 mm rebars and in 20 distance in case of 032 - The course of the steel strains is near to linear (which cor-
mm rebars. The secondary cracks have an inclination -60° responds to the assumption of uniformly distributed bond
to the rebar's axis. This means that their position on the Stresses along the rebar). Hence the ß-value could be taken
concrete surface differs from that at the rebar surface. Ac- toO.5.

2016
- The crack having the design crack width is always an ex- crack No. l the slips between the positions ca. -400 mm up
tant primary crack. The development of primary cracks is to ca. -200 mm change their sign.
finished when at every point along the rebar-concrete con- • On the right hand side of crack No. l the bond Stresses
tact surface slip occurred. The basic length for determining show a parabolic course. The max. bond stress is not at
the length lsmax is the transmission length 10, which occurs the crack where the slip is the maximum but ca. 100 mm
when the last primary crack occurred (considered statis- away: this shows the development of internal Goto-cracks
tically: in the uncracked cross-section the upper fractile which makes the bond soften. The actual transmission
value of the concrete tensile strength appears). lsmax can be length ends before the transmission length of a crack being
deduced from this length. on the right hand side of the gauged-section.
- When the distance of two primary cracks is just 2*10 or • Increasing the steel stress the bond Stresses near to crack
more, then increasing the loading a further primary crack No. l decrease due to further development of internal Go-
will occur between these two cracks. The published strain to-cracks. As equilibrium must be maintained, the trans-
distributions reveal that when the distance of two primary mission length increases, and the bond stresses there in-
cracks is max. 1.8 10 then an occurrence of a further pri- crease äs well.
mary cracks is not probable, i.e. l smax = 0.910. This relation- • At steel stress of 328 N/mm2 the two increasing transmis-
ship is consistent with the well-kriown relationship: sr max = sion lengths touch each other. At even small slip values
1.7 srjn (even if crack distance and transmission length are quite high bond stresses act which then remain rather con-
not interrelated at all). stant.
• For the calculation of the slip values the curved courses
About the distribution of the bond Stresses (300/1OOT20) of the bond stresses can fairly well substituted with a uni-
(Fig. 2): formly distributed course.
The bond Stresses were caiculated from the measured strains • With increasing steel stress and slip in the crack the inten-
äs shown in the paper of Scott and Grill. The zig-zag course sity of the substituting uniformly distributed bond stresses
results from the small deviations from the planned 12.5 mm even if slightly, but increases.
distances between the strain-gauges. Fig. 2 top shows the Fig. 2 bottom shows the courses of the bond stresses just be-
bond Stresses up to the occurrence of the 2nd crack along the fore and after the occurrence of crack No. 2:
gauge-mounted section, whereas Fig. 2 bottom of those just • Crack No. 2 occurs just outside of the transmission length
before and after the appearance of crack No. 2 there. The leg- of crack No. l. This causes a dramatic change in the course
ends refer to the steel Stresses in the cracks. First the courses of the bond stresses between cracks No. l and No. 2: the
shown in Fig. 2 top are discussed: section with the zero bond stress moves nearer to crack
- The course at steel stress of 244 N/mm2 reveals that al- No. l. On the right hand side of the zero point it can be
ready a crack (No. 0) to the left from the section occurred. observed that due to the hysteretic behavior of the local
Crack No. l occurred at about 70 mm distance outside the bond-slip relationship, the change of sign of the slip re-
transmission length of this No. 0 crack. At development of sults in relative small bond stresses until the relevant ribs

Flg. 2: The course of the bond stresses in the specimen 300/1OOT20 of Scott and Gill (1987) at different steel stress levels; top: before, bottom: after
the development of the 2"" crack. (Horizontal axis: 1000 mm lenqth with qauqes, vertical axis: caiculated bond stresses.l
r ! ! ' ' ~

:vwv
i 7
*~„x*i*±
v 1^ «, _ ** V .Völ
-500 100 300\
__A

500

2016 45
 of the rebar return to their original positions and load the • For crack width control the primary cracks are of interest.
concrete lugs between them from the opposite direction. • The mean value of crack distance and crack width are
• Between cracks No. l and No. 2 -having the distance close quite inert even for the most important influencing factors,
to the transmission length of crack No. l- the bond Stresses like concrete class, whereas the upper fractile values re-
are of moderate magnitude hence the concrete Stresses be- veal strong dependence.
tween these cracks remain small, too, therefore no further • In many cases the max. crack width was wider than the
cracks can develop there. 95% upper fractile value (which is normal). The mean
• On both sides of crack No. 2 no development of intemal value of the crack width is substantially influenced by the
Goto-cracks can be detected. width of the secondary cracks which are out of interest.
• The application of more thin rebars influences very much
The misleading character of the databank, the accentuation of both the crack pattern and the max. crack width.
the mean values and the Substitution of the slip length with Table 3 gives an impression about the change of the maximum
the crack distance can be shown on the results of Scott and crack widths while changing the diameter of the deformed
Gill, see Table 2: (Nori) rebars. (Note: these crack widths are measured on the
beam surface at the level of the rebars. The crack widths on
Table 2: Comparison of crack distances and distances of points of no the rebar surface are smaller.)
slip

Distance of points of no Table 3: Max. crack width vs. rebar diameter (in l/l 00 mm)
Specimen Crack distance, min
slip, mm
100T12 304 322 steel stress N/ reinforcement
mm2
393 321 1032 4016 10010
244 200 35 17 7
Mean value 314 322
250 50 20 10
100T20 202 215 OAA CA 1 A 11
~> \J A /
202 215
350 65 35 20
126 174
400 75 50 20
162 133
149 • The 1032 rebar placed in the middle of the 30 cm wide
Mean value 173 177 cross section (certainly an unusual but quite informative
pattern) hardly controls the crack widths at the corners of
Whereas the mean values of crack distance and ls.,max
max are quite
* the cross section
close together, the governing values of sm and are essen- • The primary cracks reveal the increase between the widths
tially different. at the level of the reinforcement vs. on the bottom of the
beam caused by the curvature of the beam.
• The primary cracks are perpendicular to the beam's axis
3.4 BEAM TESTS OF RUSCH AND and run through the tensile zone of the beam, while the
REHM secondary cracks -äs they develop from Goto-cracks- are
short and inclined to the relevant primary crack in their
Rusch and Rehm (1963) tested beams with rectangular (62.5
neighborhood. Note: the Goto-cracks are mostly inclined.
* 30 cm2) cross section, reinforced with plain (7 pcs.) and
• The development of secondary cracks is not imperative on
with different deformed bars (25 pcs.). (Note: In the 1950's
both sides neither of the first nor of the widest primary
the optimal form of the rib pattern for deformed bars was
cracks!
searched for in Germany.) In 16 beams (evaluated in this pa-
• The ratio of the 90% upper fractile value to the mean crack
per) the reinforcements 1032, 4016 or 10010 (each yields
width value changes with the loading level: at 200 N/mm2
804 mm2 steel cross sections) were applied. The concrete
steel stress 1.56 - 2.73, at 300 N/mm2 1.38 - 2.44, i.e. it
compression strength (20 cm cubes) was 15-32 N/mm2. The
slightly decreases.
beams had 4 m span and two concentrated loads. Along the
• The minimum distance of the primary cracks let estimate
200 cm long middle section with constant moment no stirrups
the upper limit value of the bond Stresses (assuming uni-
were placed. Concrete cover was (mostly) 3 cm. The loading
form distribution). Values between 2 and 3 N/mm2 were
was increased causing 50 N/mm2 stress-steps in the tensile
found.
reinforcement. The formation of cracks was observed and
• No strong correlation between the tensile stress at crack-
their width was measured in the level of reinforcement (on
ing of the beams and the compressive strength (cube) was
both sides of the cross section) and on the bottom side of the
found.
beams (mostly below the rebars in the corners). Extreme fiber
• Neither the value, nor the position of the maximum con-
strains of the concrete (near top and bottom) were measured
crete elongation (multiplied by 20 cm gauge length) do
with 20 cm or 50 cm long mechanical gauges. All results of
coincide with the position and width of the widest crack,
the measurements were presented in graphs and summarized
resp.
in tables. The crack widths were evaluated statistically, too,
Fig. 3 shows the measured max. crack widths for the 62.5 *
and presented also äs cumulative frequency curves. Further
30 cm2 beams reinforced with diflerent diameter and rate of
on two more rectangular beams (120 * 45 cm2) were tested
reinforcements of Nori-type deformed rebars: R37 - 1032,
(5026 and 8020) with 400 cm constant moment length
R25 - 4016, Rl 9 - 10010 (all p=0.46), R54 - 2026 (p=0.62),
The side views of the beams with the Steps of crack de-
R21 -1032 + 4016 und R17 - 8016 (p~0.9S).
velopment and the tables lead to profound insights into the
In addition to the pronounced influence of the rebar diameter
Cracking per se:

46 2oi6 - WCRETESTRÜCTURES
 4. CONCLUSIONS
Based on the re-evaluation of results of four test series from
—»-RJ?
the classical literature the following conclusions can be
-»-W5
made:
-»-BIS
- Shortages of international code proposals for calculation
of crack width are presented
——R54 - A clear distinction between primary and Goto-/secondary
cracks is made.
- The design crack width belongs to a "last" primary crack
200 300 400 500 with the longest transfer lengths on its both sides
ttcclttreu. N/mm' - The Goto- and secondary cracks serve the compatibility
slg. 3: Max. crack widths measured on tne "smair deams of Rusch
between the high strains in the rebar and the very limited
and Rehm (1963) with different diameters and amounts of deformed tensile strain of concrete around.
bars - The Goto- and secondary cracks "soften" the bond char-
acteristics of the rebars.
in case of identical rate of reinforcement, the crack width re- In Part II - after a short discussion of different problem-rel-
ducing efFect of increasing rate of reinforcement can be real- evant items - the well-known formula of the design value of
ized. crack width receives an advanced Interpretation All influenc-
A comparison of the crack patterns of the "small" (62.5 *30 ing factors are taken into consideration at the proper position.
cm) beams with those of the "high" beams (120 * 45 cm)
reveals that
• The geometrical rate of reinforcement, n, is a misleading 5. REFERENCES
characteristic: the crack patterns of beams with similar u
Baläzs, L. Gy. et al. (2013): "Design for SLS according tofib Model Code
values but with different b * h sizes are very different. 2010".fib Structural Concrete, 2/2013, June 2013, pp. 99-123.
• In case of the "small" beam with 2026 rebars (beam R 54, Broms, B. B., Lutz, L. A. (l 965): "Effects of Arrangement of Reinforcement
p = 0.62, clear distance between the rebars -180 mm) no on Crack Width and Spacing of Reinforced Concrete Members", ACI
Structural Journal,Nov. 1965,No. 62-77 pp. 1395-1410. + Part II Report.
tendency of occurrence of secondary cracks can be found Goto, Y. (l 971): "Cracks formed in concrete around deformed tension bars",
whereas at the "high" beam with 5026 rebars in one layer ACI Journal, April 1971, pp. 244-251.
(beam R 103, p = 0.51, clear distance between the rebars ~ Scott, R. H., Gill, P. A. T. (1987): "Short term distributions of stand and bond
60 mm) a very pronounced secondary crack pattern devel- stress along tension reinforcement", The Structural Engineer, 1987,65B,
No. 2, pp. 39-48.
oped. This means that the tensile stresses in the concrete Rusch, H., Rehm, ü. (1963): „Versuche mit Betonformstählen", Deutscher
due to bond strongly disperse therefore in case of big bar Ausschuss für Stahlbeton. Heft 140. 182 p.
distances no secondary cracks can develop. EN 1992-1-1. (2003): "Eurocode 2: Design of concrete structures - Part 1-1:
General rules and rules for buildings", CEN April 2003
• At development of the secondary cracks the compression CEB-FTP Model Code 1990 (1993), Comite Euro-International du Beton,
zone has no impact hence the tensile reinforcement with Bulletin d'Information No. 213/214, Lausanne, May 1993
the concrete around behaves independently of the com- fib Model Code for Concrete Structures 2010 (2013), Ernst & Sohn, Wiley,
pression zone. ISBN 978-3-433-03061-5
Ferry Borges, J. (1966): "Cracking and DeformabüUy of Reinforced Con-
• As the secondary cracks originate from Goto-cracks hence crete Beams", IABSE Publication, Zürich, Vol. 26, pp. 75-95.
their width does not result from the slips on both sides of Control of Cracking of Concrete Structures (2001), ACI 224R-01, Reported
the secondary crack but from the slips on the remote side by ACI Committee 224, American Concrete Institute, October 2001, 49
P-
from the relevant primary crack only. The Goto- and -äs Guide to Design Detailing to Mitigate Cracking (2013), ACI 224.4R-13, Re-
a consequence- the secondary cracks do not change the ported by ACI Committee 224, American Concrete Institute, December
direction of the slips along the transfer length belonging to 2013,24p.
Coleinan, J. W. (2013). "Cracking... Defecl or Normal? Part 1: When is
corresponding the primary crack. Accordingly the transfer Cracking a construction defect?", Concrete International, Vol. 35, No. 9,
lengths of this primary crack extend beyond the neighbor- Sept. 2013, pp. 35-38, "Part 2: Case law on concrete Cracking", Concrete
ing secondary cracks on both sides. International, Vol. 35, No. 10, November 2013, pp. 29-35.
Leonhardt, F. (1987): "Cracks and Crack Control at Concrete Structures",
• The steel stress in the secondary cracks is always less than IABSE Proceedings P-109/87, IABSE Periodica 1/1987, pp. 25-44.
in the neighboring primary crack, nevertheless bigger than Windisch, A. (l 989): „Zur Berechnung von kritischen Rissbreiten in Zugsta-
it would result from the concrete contribution along the ben und Biegebalken", Werkstoff und Konstruktion II, Prof. Rehm zum
total transmission length of the primary crack. The steel 65. Geburtstag. 1989 Okt., ibidem-Verlag, Stuttgart, pp. 241-261.
stress must balance the concrete tensile stresses which
caused the secondary crack.
• After occurrence of a secondary crack its width and exten- NOTATIONS
sion develops/increases until equilibrium is achieved. Ac ef effektive area of concrete in tension
• The distance of primary and next secondary crack is dif- Ac concrete area in tension
A. reinforcement area
ferent to the transfer length one considered at development E, Young's Modulus of Elasticity of reinforcement
of the primary cracks äs in the secondary crack the slip c concrete cover, rib distance on rebar
is not zero and the steel stress does not correspond to the fflm mean value of concrete tensile strength
fR relative rib area
concrete tensile strength. k empirical parameter
• The Goto- and secondary cracks reduce the stiffness of the kj, k2 coefficients
concrete around the rebar hence the concrete contribution 's mm length over which slip between steel and concrete occurs
relevant for the width of the primary crack decreases too. 10 bond transfer length
s distance of rebars
sm, sm mean crack distance
wt characteristic crack width
wd design crack width

2016 47
o. ratio Es/Ec o^ steel stress at the crack under forces causing fam
ß empirical factor to assess averaged strain within 1^ TJJ. lower fractile value of the average bond stress
f.m average/mean concrete strain ^ mean bond strength between steel and concrete
ECS relevant concrete shrinkage strain
E^ average/mean steel strain Andor Windisch PhD, Prof. b.c. retired äs Technical Director of DY-
ES steel strain at the crack WIDAG-Systems International in Munich, Germany. He made bis MSc and
s^ steel strain at the crack under forces causing f am PhD at Technical University of Budapest, Hungary, where he served 18 years
0,0 nominal diameter of rebar and is now Honorary Professor. Since 1970 he is member of diöerent com-
P A/A missions ofFIP, CEB and tib. He is autlior of more tlian 120 teclinical papers.
pr, prf effective reinforcement ratio (=As/Ac lf) (Dr. Andor Windisch, Schwarzhölzistrasse 19A, D-85757 Karlsfeld, Andor.
a steel stress at the crack [email protected])

View publication stats