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

net/publication/259042472

Behaviour of RC beams with inadequate lap splices, before and after

upgrading by welding of reinforcement

Conference Paper · June 2012

CITATION READS

1 697

3 authors, including:

Petros M. Chronopoulos Konstantinos Trezos

National Technical University of Athens National Technical University of Athens
7 PUBLICATIONS   12 CITATIONS    39 PUBLICATIONS   168 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Self compacting concrete View project

RadarNtua View project

All content following this page was uploaded by Petros M. Chronopoulos on 21 May 2014.

The user has requested enhancement of the downloaded file.

Behaviour of RC elements with inadequate lap splices,
before and after upgrading by welding of reinforcement
P. Chronopoulos, M.Sc./Ph.D.c, C. Trezos, Assistant Professor &
M. Chronopoulos, Sen. Res. Eng.
Lab. of RC, National Technical University of Athens, Greece

ABSTRACT: A first set of RC elements (beams, w/o any axial load), with inadequate splices of their tension
reinforcement, were tested under monotonic load (deformation controlled) and retested after “strengthening”
by welding of failed splices, w/o any repair of the cracks or recovery of the spalled concrete. Based on the
first results of this on-going theoretical and experimental programme on splices of RC elements, and besides
various “satellite” findings regarding certain aspects and details, it seems that the relevant provisions of mod-
ern Codes in force are rather conservative : Shorter (than the normative) lap splice lengths could be effective
and safe, at least for lengths more than a “minimum”, depending on execution details. The main parameters
for these tests were (i) the details (the arrangement) of splices within the element’s section, (ii) the degree of
passive confinement (offered mainly by hoops), and (iii) the spliced length, equal to 15 or 30 bar diameters.

1 INTRODUCTION RC/NTU of Athens/GR, aiming on calibration of

relevant Code provisions (past and recent ones), as
For practical purposes, there is always a need for well as on retrofitting. Tests and evaluation cover
bars to be spliced in almost every RC structure; beams and columns, with adequate or inadequate lap
therefore, splicing of reinforcement is always a chal- splices. In this paper, the first results of four (4)
lenge for structural engineering. In general, splices beams, tested and retested (after welding) under
of reinforcement should not be located in highly monotonic loading are presented and discussed.
stressed areas, e.g. in potential plastic hinge regions
of EQ resistant elements, unless cyclic inelastic test-
ing of the connection details (in realistically sized 2 BOND AND ANCHORAGES
specimens and under conditions compatible with the
2.1 General, bond
foreseen ductility) has established that ductility will
not be impaired; to this end, tests on connections in- The recognition, development and application of the
dicating unimpaired bar strength (or bar break) are “local” bond stress–slip (τ-s) relationship was a rev-
inadequate to establish satisfactory performance. olutionary step in the history of the research of bond,
Nevertheless, splicing of reinforcement in critical for both anchorages and splices, as it is “imprinted”
regions (dissipative zones) of RC elements is still in a series of relevant successive and detailed edi-
the rule at least for conventional new buildings, not tions (see, e.g., the previous Int’l Symposiums on
to mention existing ones, designed and constructed Bond – BICs, the relevant Bulletins of CEB-FIP and
on the basis of past knowledge (and Codes). Lessons fib, the Model Codes, etc.). In addition, a further
learnt from earthquakes lead to the conclusion that critical step was the recognition and differentiation
bad detailing or execution deficiencies associated of the two (2) possible bond failure modes, i.e. that
with splices in columns (and walls) could be cata- of splitting (and spalling), SP, or that of pull-out (or
strophic; in addition, an increase of splice lengths, -through), PO, for bars in unconfined (or passively
w/o adequate detailing and transverse clamping rein- confined) or actively confined concrete, respective-
forcement, is not likely to ensure satisfactory per- ly. To this end, considerable differences and modifi-
formance (i.e. splices tend to unzip, prematurely, cations are presented in the MC 2010 as compared to
even at low levels of cyclic loading). those in the MC 1990 for both PO and (especially)
A theoretical and experimental programme espe- SP types of bond failure. These alterations cover
cially on splices of RC members under monotonic significant aspects such as the relationship itself, the
and cyclic loads (including critical regions of EQ re- differences between bond conditions, the influence
sistant elements) is under completion at the Lab. of of passive confinement offered by concrete cover
or/and stirrups, the influence of bar size effects and Table 1.1 Characteristics of the (τ-s) law, MC 2010.
bar yielding, and the characteristic sudden drop of For bar yielding, and
the bond stress before a plateau is reached (brittle- BOND CONDITIONS
GOOD POOR
ness). Among others, a “new” relationship is pro- COVER STIRRUPS COVER STIRRUPS
posed for , as a product of three (3) terms, rep- Mean
resenting the influence (i) of concrete strength, (ii) 7.0 8.0 5.0 ( 70%) 5.5 ( 70%)
of bar size effects, and (iii) of the degree of passive For
confinement, as follows: 0.30 0.45 1.35 1.70

0 0.4 0 0.4
At
, (1) * *
* clear distance between consecutive bar ribs, depending on Ø
where for “good” conditions, or
or for “poor” conditions and SP or PO mode, Table 1.2 Characteristics of the (τ-s) law, MC 1990.
respectively; or for unconfined concrete THERE IS NO REFERENCE FOR BAR YIELDING OR SIZE EFFECTS
or in presence of stirrups; and . BOND CONDITIONS
The main differences between the MCs, especial- GOOD POOR
ly for ribbed bars and a brittle splitting (and spalling) THERE IS NO REFERENCE FOR THE TWO MECHANISMS
mode of bond failure, are presented in the following Mean
9.0 4.5 (50%)
Tables, for a concrete class . For
Nevertheless, such local (τ-s) laws are only statis- 0.60
tically acceptable mean curves, applicable as an av-
erage formulation for a broad range of cases; for a 0.15
given slip value, the coefficient of variation of the At
corresponding local bond stress may amount to ap- 1.00 2.50
prox. 30% (or even more). After all, design recom-
mendations and rules of Codes are conservatively  In Chapter 6 (bond), two (2) bond strengths (av-
based on low bond “strengths”, roughly correspond- erage over the bond length) are given, the basic
ing to a slip of approx. 0.1 mm, incorporating partial ( ) and the design one, with
safety factors as well. According to the MCs (and to
more recent results), the ultimate bond stress of
ribbed bars corresponds to a slip of approx. 0.5 mm , (3)
(or even higher, see Tables 1.1 and 1.2), depending
on bond conditions as well. In addition, the concept , (4)
of “bond strength” as a material property of the con- where and (see § 2.3) are factors reflecting
crete (for a given bar in a given position with respect the influence of passive confinement from concrete
to casting), although simple and convenient, is not cover and from transverse reinforcement ,
representative of reality; therefore, overall effects, in excess of their respective “permissible” minima;
depending on actual detailing, could govern. or , for bond conditions (as per the MC
According to the EC 2, or to the Greek Code, still 1990); for ; and
in force, for ribbed bars with a relative rib area (or depending on
“bond index”) As a simplification, the following expressions
for for could be used:
and for or
√ (for a value of for , (5)
should be used), the design value of the , with . (6)
ultimate bond strength could be estimated as :
To this end, it is interesting to note that for com-
, (2) mon concrete classes
where and .
(MPa), for a concrete class  In Chapter 13 (detailing), a simplified expression
or , for “good” or “poor” conditions; is proposed (see § 2.3), as follows:
and for . , (7)
The MC 1990 contains similar expressions, alt-
hough the mean concrete tensile strength, is with , for a concrete class
based on instead on (with disregarding the influence of and .
almost equal values). The MC 2010 contains certain Here again, it is interesting to note that these
differentiations, as follows (for ribbed bars): values according to the MC 2010 are at least 10%
lower than those according to the MC 1990.
2.2 Bar size effects where, according to the EC 8, for column (not beam)
bars, contributing to the flexural strength of critical
Size effects depending on concrete (on , the nom.
regions, or , and
max. aggregate size) are not taken into account into
bond models, even at large slip values (i.e. fractur-
ing). On the other hand, results for bar size effects
are contradictory, especially for bars in tension, , (9.1)
since for bars in compression the bar diameter is where or for bars in tension or com-
taken into account, due to end bearing (see § 2.3). pression, respectively; and are factors ac-
Nevertheless, almost all Codes contain certain clear counting for the concrete cover and the transverse
rules for bar size effects, as in the following Figures. reinforcement; or for bars in te-
nsion or for bars in compression; and
.
To this end, where
is the min. of the cover for bond (not for durabil-
ity), with for
mm) or the half clear distance between parallel bars
with
and where (
Figure 1.1. Bar size effects, EC 2 and MCs. ) , with the cross sectional
area of the largest anchored bar, and K (for the ef-
fectiveness) = 0.10 or 0.05, for fully “supported”
corner bars or “partially” supported (“unsupported”)
intermediate bars (away from stirrup corners), re-
spectively. In this respect, the transverse reinforce-
ment ( , along the relevant length), to resist
transverse tension (bursting) forces, shall be enclos-
ing longitudinal bars, placed perpendicular to them
and between them and the free surface of the RC el-
Figure 1.2. Bar size effects (indirectly), based on the
RILEM/CEB-FIP Beam Tests for a bond length of 10
ement; otherwise, K = 0. It is interesting to note that
, according to the EC 2/Annex C (normative). in order to reach a value of (i.e.
), values of more than 2 and
Applying statistical regression to the results of a much more than 1 are needed, see § 3.
large number of previously performed anchorage For the MCs, the following are valid:
and splice tests, Orangun et al. (1977) developed an  MC 1990
empirical formula exhibiting no bar size effects; Identical expressions and values are given.
nevertheless, such effects were and are implied by  MC 2010
development length provisions of the ACI Different expressions and values are provided, based
Codes, strong ones (e.g. ACI 318-77 or ACI 318-89) on the previous Equations (3) and (4), see § 2.1.
or weak ones (e.g. ACI 318-95 or ACI 318-08), de- First of all, factors and (as a sum, with
spite the absence of any revolutionary new findings , and not as a product) are directly in-
(Sener 1999, Ichinose et al. 2004). The relevant cur- fluencing bond strengths and not lengths. In addi-
rent ACI rules are based on – factors, multipliers tion, for at
of the , a sort of a discontinuous bar size effect, least
with for bars and for at least one
or for bars . piece of transverse reinforcement positioned within
the relevant length and no further than from
2.3 Anchorages the end of the bar, where:
The anchorage length is an input for the estimation
of the splice length of bars (see § 3), both in tension is the side cover, is the bottom or top cover
and compression, based on the bond strength and
(§§ 2.1 and 2.2). According to the EC 2, the basic
required ( ) and the design anchorage is the diameter of the smaller bar
length for straight bars (with , in any case) is the number of anchored or spliced bars
are given by the following expressions: is the number of legs in a cross-section
,
is the cross sectional area of one leg
( ), (8) is the longitudinal spacing
 is the “density” factor, with tance between parallel and lapped bars of the same
pair should be otherwise, a
k is the “effectiveness” factor, with for correction is needed, see § 3.2 .
helical transverse reinforcement or for “support- A third issue, a “controversial” one, is that of the
ed” corner bars (provided no anchored bar or pair arrangement of lapped bars (pairs) either in full con-
of spliced bars are further than the lesser of tact (side-by-side) or at a certain distance (spaced
or from where the leg crosses the split- apart). The older provisions and application rules, or
ting plane), or for “unsupported” bars even design guidelines, were in favour of distant
“confined” by straight legs within the cover bars. On the other hand, the latest recommendations
thickness (provided clear distance of bars is at are in favour of contact bars, in part due to the
least , with ; otherwise, recognition that SP modes of failure precede PO
Finally, the design bond length, for anchorages or ones, not to mention other contributing mechanisms
splices in tension, may be calculated from: (e.g. interlocking of the ribs). According to the
Codes in force, bars may be allowed to touch one
another, while a stronger relevant statement may be
( ), (9.2) found in the MC 2010: Overlapping bars should be
placed in contact, while in elements other than slabs
where is a function of the percentage ( ) of lon- they should be tied together as well.
gitudinal reinforcement anchored or spliced within The fourth issue, also a “controversial” one, is
0.65 from the centre of the relevant length con- that of the arrangement of the pairs of bars so that
sidered, with they are either face parallel or perpendicular, i.e. so
, see § 3.2. As a simplification, an expression that the two planes (that of the lapped bars and that
for could be used, with of the element’s surface – of the legs of transverse
( ) and see reinforcement) are either parallel or normal, since
Equ. (7). the arrangement (and the possible failure) is directly
For bars in compression, end bearing could be di- affecting the amount of transverse reinforcement
rectly taken into account, by means of the expres- (see § 3.3), not to mention problems due to concrete
sion: casting (see, e.g., Jirsa et al. or Cairns and Jones in
( ) , Bartos 1982, Cairns 2006). Here again, the main is-
sue is the expected mode of failure, i.e. that of burst-
with , (10) ing/splitting (and spalling), in the plane of the bars,
or that of crushing and sleeving (shearing-off, fric-
where or 0 for or , tion slipping), inbetween the bars. Of course, in the
with the end cover (measured parallel to the bar relevant technical literature considerable differences
axis) from an unsupported face, under the provision could be found regarding the recommended ar-
of a minimum transverse reinforcement. rangements of the pairs of bars (see, e.g., Paulay
1982, Paulay and Priestley 1992, Fardis 2009).
In general, all bars in compression may be lapped
3 LAP SPLICES in one section, while for bars in tension the “permis-
3.1 Arrangement sible” percentage is 100% or 50% where the bars are
in one or in several layers, respectively.
In general, laps should normally or preferably (i) be
located away from highly stressed areas, (ii) be stag-
gered, and (iii) be symmetric. 3.2 Length
Laps in the critical regions of beams could be The differences regarding bond strengths, bar yield-
(and are) avoided, but they are practically unavoida- ing, bar size effects, anchorage, transverse rein-
ble in columns (and walls), especially at the base of forcement, etc., as presented in §2, are directly in-
the building. In such cases, particular attention fluencing splices. Therefore, uncertainties related to
should be paid in design and execution to ensuring the behaviour of laps (under monotonic and, espe-
(with an increased reliability) integrity of the whole cially, cyclic actions) are higher than those of an-
area and no disproportional damage. In addition, chorages. First of all, and according to the EC 2 (and
“staggering”, for a distance 1.3 or even 0.65 is the EC 8), the anchorage length for straight ribbed
not practical, except for slabs. bars is favourably affected by the “passive” con-
A second relevant issue is that of distances be- finement, based on the coefficients and , with
tween bars in laps. To this end, the clear distance be- (see § 2.3), especially for corner (“sup-
tween parallel pairs of bars should be ported”) bars, while for intermediate bars the effec-
different than the relevant gen- tiveness factor K is 0.05 instead of 0.10. As it is ob-
eral provision (see the values for the estimation vious from the Figures 2 and 3, an excessive cover
of and , Equ. (9.1), §2.3), while the clear dis- or/and proper transverse reinforcement is needed in
order to achieve a reduction of the anchorage and lap of the square root of , multiplying the length of the
length. In addition, lap lengths should be increased anchorage to obtain the lap length, as follows:
in case of large spaces between conjugated bars, see  MC 1990 (and Greek Code),
Figure 4, although such an increase is of a limited
practical significance. √ (10.1)
Laps do not generate twice the splitting (bursting)
force as compared to that of an “equivalent” single  EC 2 (and EC 8),
bar anchorage, in any of the two (2) characteristic √ (10.2)
planes of the pair of bars; experimental and theoreti-
cal evidence shows that the stresses in all directions  MC 2010,
around a lap are closer to uniform. Therefore, the
strength of a lap is not half that of the anchorage but √ (10.3)
between 70% (or 80%) and 100% of the strength of According to these values it is obvious that for
the single bar anchorage. Nevertheless, for an equiv- the last 20 (or 30) years, the relevant coefficient
alent amount of “confinement”, lap lengths need to for laps is divided by a factor approx. equal to √
be longer than the corresponding anchorage lengths, almost every 10 years (!).
especially for bars in tension. However, what appears under monotonic loading
In the last 30 or 40 years, a lot of relevant re- as a wide safety margin is absolutely necessary in
search was performed and calibrated, as imprinted in EQ resistant structures, because cycling of the bond
the technical literature, with a lap considered as an stress causes a large drop in the effective bond resi-
almost “double” (in the past) to an almost “single” stance, in terms of both strength and stiffness (Fardis
anchorage (recently), influenced by the percentage 2009). To this end, if sufficient length combined
( ) of longitudinal reinforcement lapped in a “sec- with “passive” confinement is provided, the peak
tion”. The width of this “section” from the centre of bond demand is low and accumulation of slip is lim-
the relevant length considered, along the length of ited, i.e. the “bond damage threshold” is higher and
the RC elements, is according to the MC the “disorganization” of laps could be reduced to ac-
1990 (and the Greek Code, still in force) or half of ceptable levels. Moreover, seismic loading does not
that according to the MC 2010 and the cause full reversals of bond in real RC structures.
EC 2. Finally, a factor was and is applied, a function A related problem is that actual lap lengths are
longer (or much longer) than the lengths of critical
regions, a reality which is not properly examined.
Finally, the relevant practical provisions for the
assessment and retrofitting of existing RC structures,
according to the EC 8 or to the new Greek Code on
Structural (Assessment and) Interventions (nGCSI
2012), should be mentioned here, as follows (for
straight ribbed bars):
1) For overlapping bars in compression, full
Figure 2. Influence of concrete cover, yielding is to be expected, under a minimum of an-
, according to the EC 2, see § 2.3, with . tibuckling transverse reinforcement; therefore, the
value of the compression reinforcement area is dou-
bled in laps over the value applying outside them.
2) For overlapping bars in tension, almost full
yielding (but not hardening) is achieved for
√ (mm, MPa/mean values),
while for the steel stress is
proportional to ( ) and for
the steel stress is practically zero.
Figure 3. Influence of transverse reinforcement, – ,
3) Based on the previous approximations and
for corner bars , according to the EC 2 (and the EC modifications, local ductility of laps could be as-
8), see § 2.3, with . sessed in terms of the chord rotation at member
ends, based on rational although complex relations
regarding both parts of the rotation capacity, the
elastic ( , a sum of 3 terms) and the plastic or the
total one ( or a product of 6 terms).
4) The reduced cyclic shear resistance (as con-
Figure 4. Increased lap length in the case of distant bars; for trolled by the stirrups), decreased with the ductility
with the Code length. demand, is practically not affected due to the laps
themselves.
3.3 Transverse reinforcement
Various expressions have been proposed for the ef-
fectiveness of transverse reinforcement “offering”
clamping forces to anchorages and laps (see, e.g.,
Paulay and Priestley 1992), although research find-
ings cannot be readily “translated” into relatively
simple and straight forward design recommenda-
tions. Among them, and as they have been adopted
by Codes, the expressions of the EC 2 (and the EC
8) regarding the factor have
been already presented in §§ 2.3 and 3.2 as well as
in Figure 3, while those of the MC 2010 regarding
the factor the (Equ. 1) and
the (Equ. 12), including a term of , are
discussed in §§ 2.3 and 5.4. Figure 5. Transverse reinforcement arrangement at laps, EC 2.
To this end, certain conclusions of general validi-
ty are shortly presented here below:
 Transverse reinforcement provided because of
other requirements may be included in while
its full “activation” is still an issue.
 Corners of stirrups are much more effective than
“unsupported” legs of transverse reinforcement for
intermediate bars.
 In general, for anchorages while
for laps (especially for EQ resistant ele-
ments), see Figure 3.
 The larger the amount of or (mainly) the
closer the spacing of multiple (or overlapping) stir- Figure 6. depending on the arrangement of laps, EC 8.
rups, the less important is concrete cover as a factor
influencing bond and lengths. and on (the cross section area of the larger
 The amount of above which there is no fur- lapped bar) for face perpendicular laps, see Figure 6.
ther or little benefit is larger under cyclic than under In this respect, it is interesting to note that the
monotonic loading. provision for (one leg) (for
 Longer anchorages and laps exhibit mixed and is directly based on a proposal by Paulay
more complex failure modes, depending on overall (Paulay 1982, Paulay and Priestley 1992).
structural behaviour as well.
 Multiple or overlapping stirrups are capable of of-
fering a reliable and adequate degree of concrete in- 4 THE PRESENT RESEARCH
tegrity and “confinement”, also preventing the buck-
ling of bars after splitting and spalling of the cover. The on-going experimental and theoretical campaign
According to the EC 2, should be concentrated at Laboratory of RC/NTUA is focusing on the over-
at the outer parts of the lap at distances all behaviour (including deformation capacity) of
(a min. of 2 or 3 legs), see Figure 5. Nevertheless, RC elements with spliced bars, based on almost full
under cyclic loading (and damage penetration) a uni- scale tests on beams and columns.
form spacing of legs provides a higher effectiveness, Three (3) characteristics have to be mentioned:
unlike under monotonic loading where legs near the 1) Since the propagation of damage from the end
ends are much more effective. (or the ends) of a lap (unzipping) is critical, it may
According to the EC 8, three (3) main rules are be assumed that moderate levels of shear force – and
given additional to those of the EC 2, for both the therefore unequal bending moments at the ends, will
medium and the high ductility classes, as follows: improve the performance of laps, especially when
1) , where b is the subjected to cyclic loading, although a lap failure in
minimum cross sectional dimension of the RC ele- the max. V region seems to exhibit stronger size ef-
ment (i.e. ); fects and higher brittleness than a lap failure in the
2) ( ) applied for max. M region of an element. Therefore, for all
each lap/pair of bars, where is the diameter of lapped elements, two (2) identical laps were applied
lapped bars; and (see Figs. 7 to 9), under a constant shear force com-
3) should be based on (the sum of cross bined with a bending moment gradient.
section areas of all lapped bars) for face parallel laps 2) The bond conditions for all bars and laps of all
elements were good, irrespectively of their position
during concreting; “top bar effects” was not an issue.
In addition, for a limited number of specimens, den-
sification of the concrete matrix was applied by
means of revibration (approx. after 15 minutes). Ag-
gregate size ( ) and gradation has a limited role;
the surface of the bars gets always in contact with
the finest aggregates of the concrete matrix.
3) All laps were tied with single or even double
mild steel wires in addition, for a
limited number of specimens nylon ties were used,
especially for tying lapped bars together. The posi-
Figure 7. General arrangement and loading (for N=0); 1 and 2:
tion of bars and laps, as well as the thickness of con- strong and stiff steel reaction frame; 3: RC element, upside
crete cover, was secured by means of closely spaced down, monotonic loading;
plastic spacers.
One of the first series of simple tests on RC
beams (w/o axial load) with inadequate laps under
monotonic loading is presented and discussed; in the
following Figures 7 to 9, the general testing and
loading arrangement as well as the details of the four
(4) beams with a and a shear span to
depth ratio are shortly presented.
The main characteristics of the materials are as Figure 8. Details of the RC beams Δ1.1 and Δ1.2, with
follows: ; concrete cover to stirrups
Concrete (at the time of testing/retesting, standard deviations less than 5 mm; stirrups
cylinders) (roughly corresponding to
a concrete class C20-C25),
√ deviations less than 10%.
Steel, bars B500C (TEMPCORE),
with
where is the
mean height of the ribs (over the circumference),
and is the mean spacing Figure 9. Details of the RC beams Δ2.1 and Δ2.2, with
of the ribs (centre to centre), . ; concrete cover to stirrups
Steel, stirrups, Beams Δ1, within deviations less than 5 mm; stirrups .
B500C (TEMPCORE), with similar characteris-
tics to those of the main longitudinal bars, although 9 , w ≈ 2 /3, t ≈ w/2 ≈ /3, Aw ≈ tℓt ≈ 3 2). All
strengths were slightly higher and values were welded laps performed well under monotonic retest-
lower, with . ing, w/o any sings of distress, cracking, etc.
Steel, stirrups, Beams Δ2, within In the following Figures 10 and 11, the experi-
S220 (or SAE 1010), plain/smooth wires, mental monotonic response of all tested and retested
actually not pe- beams is presented, together with the expected max-
rmitted for RC structures but only as “secondary” re- imum applied load, based on simple RC mechanics,
inforcement (in cages, wire meshes, etc.). i.e. (i) for a tension longitudinal reinforcement of
After initial testing and failure of laps (in either B500C, with an “activated” of for
side of the beam), an upgrading/strengthening by Beams Δ1 and 0 for Beams Δ2 (see the follow-
welding of the failed laps was applied, w/o any re- ing remarks and notes), and (ii) for an internal lever
pair of cracks (major or minor ones) or recovery of arm .
the spalled concrete areas in the region of the failed The experimental values of the displacements (d,
laps, followed by retesting; therefore, a total of at the points of the two (2) equal loads P) and of the
tests were performed. limited apparent rotation capacities (of the “hinging”
Welding of laps was executed according to the laps) correlate well with the theoretical predictions,
EC 2, the EN 10080 (2004), the EN ISO 17660 based on the relevant mechanics and on the expres-
(2006) and the Greek Code on technological aspects sions according to the EC 8 and the nGCSI (2012)
of reinforcement (2008), but not fully respecting the regarding “inadequate” lap splices, with lengths
relevant rules mainly due to the presence of stirrups. shorter than a minimum ( ), see also § 5 and
A manual SMAW method was applied (under in- Note (1) here bellow.
door conditions), with all-purpose low strength coat- NOTE (1): For good bond conditions,
ed electrodes (E6013/TiO2, AWS, 2.5 mm, I ≈ 150 (mean
A) and one-sided 2 or 3 interrupted fillet welds (ℓt ≈ values), the mean
 (see §§ 2 and 3). Therefore, the clamping forces), since
lengths of “adequate” anchorages and lap and
splices – for a full activation of – are: In addition,
the “effectiveness” factor is the same for all of the
( ) beams, since all lapped bars are corner ones.
, and (11.1)
(for contact laps) = (EC 2), or 5 THE FIRST FINDINGS AND RESULTS,
(MC 2010), (11.2) CALIBRATIONS AND CONCLUSIONS
where and (for ). 1) All ends of laps towards the beam cross section
Therefore, the values of Beams Δ1 under the max. action-effects acted as crack inducers
and Δ2 are clearly inadequate, even for across the lap zone (Figure 12), causing large “pri-
and . mary” transverse cracks (at the areas of higher bond
According to the EC 8 and the nGCSI (2012), stress) which in turn or simultaneously trigger split-
√ ting cracks, different (in general) for the two (2) sets
NOTE (2): It is interesting to note the following, of beams, with or final-
regarding the transverse (“confining”) reinforcement ly a “mixed” failure was observed, a kind of splitting
along the lap length induced pull through, leaving clear imprints of the
 In terms of the cross section area: bar ribbed surface on the concrete matrix.
for Beams Δ1 and 2) To this end, Beams Δ1 /Figure 12.1
for Beams Δ2, where is failed under a generation of full longitudinal split-
the area of (lapped longitudinal bar). ting cracks (and spalling), at both the two sides
 In terms of the clamping force, and if full yield- (throughout the cross section) and the most ten-
ing of stirrups takes place: sioned face, while Beams Δ2 /Figure
for Beams Δ1, and 12.2 failed under a partial generation of splitting
for Beams Δ2. cracks only at the two sides of the specimen
According to the MC 2010 provisions, and if they (throughout the cross section). In addition, the fail-
are interpreted correctly by the authors, the differ- ure mechanism of all Δ1 was almost identical and
ences between the two (2) sets of beams regarding much more “stable” than that of Δ2 (with differences
the so called “density” factor are much between the 4 tests, before and after welding).
lower (than those based on cross section areas or on 3) As it is already reported, all laps were in full con-
80 tact, with “interlocking” transverse ribs/lugs, while
70
2Pth = 72.50 kN all of them were well tied as well. Limited meas-
60
urements of stirrup strains in the lap zones do not
Δ1 Δ1W permit a reliable evaluation of their contribution, in
Load 2P (kN)

50
40
terms of yielding or not; nevertheless, close visual
30
examination after testing and retesting reveal an in-
teresting and positive effect: Slip and dilatancy (due
20
to the rib interlock) activated well tied stirrups in the
10
lap zone forcing them to offer a kind of a dowel ac-
0
0 5 10 15 20 25 tion as well. This phenomenon, which deserves a
Displacement d (mm) more detailed study, was more pronounced in Beams
Figure 10. Beams Δ1, with
Δ2 (with stirrups than in Beams Δ1
and . (with stirrups ), see Figure 13.
4) An interesting issue is that of the mean “activat-
ed” stress of tensioned lapped bars, , de-
pending on the mechanical and structural details of a
lap zone. Therefore, a calibration was performed,
based on the relevant provision of the MC 2010 for
, as a product of four (4) terms, as follows:
(12.1)
where the 4th term, the influence of “passive” con-
finement is a sum of two (2) subterms, i.e. the con-
tribution of the concrete cover and of the transverse
reinforcement, as follows:
Figure 11. Beams Δ2, with
and . . (12.2)
Figure 12.1. Beams Δ1 , longitudinal splitting
cracks at both the sides and the face of the elements.

Figure 12.2. Beams Δ2 , longitudinal splitting Figure 13. Bending and kinking/cranking of well tied stirrups,
cracks only at the sides of the elements (throughout). less pronounced for Beams Δ1 (with stirrups )/top
than for Beams Δ2 (with stirrups )/bottom. Slip was
This expression is a semi-empirical recent one, approx. equal to 10 ÷ 16 mm for Beams Δ1 and to 2 ÷ 8 mm
calibrated against experimental results from hun- for Beams Δ2.
dreds of tests, valid for short bond lengths, as it is
evident from the conjugate expression of the MC Δ1,
2010 for the local bond stress (Equ. 1, §2.1).
All the prerequisites for the application of this Δ2,
expression are fulfilled, except those for a percent-
,
age of lapped bars less than 50% and for at least one
leg of transverse reinforcement positioned within the meaning that , “neglecting” the
lap no further than 3Ø from both ends of the lapped absolute values.
bars; for the specimens and tests of the present re- Nevertheless, and as it is already reported (§4),
search (a total of ) all tension reinforcement the “activated” bar stresses for all tests were
was lapped and the position of the first stirrup was for and for
much further inside the lap, at a distance of approx. , with .
7.5Ø for Beams Δ1 and 5.0Ø for Beams Δ2. To this end, the following are reported:
The first two terms of the Equ. 12.1 are the same  Regarding the MC 2010, a further calibration is
for Δ1 and Δ2, with a value (as a product) equal to probably needed, if its provisions are interpreted
approx. 1.225. Also, the 4th term (the influence of correctly by the authors; after all, for the 4th influ-
“confinement”) has practically the same value for encing term (i.e. that for the “confinement”), the
Δ1 and Δ2, since: Code itself gives stronger relations regarding
 The subterm for the concrete cover (see Equ. for both the subterm of concrete cover ( ), with su-
12.2) has a value of approx. 1.25, while perscripts 1.5 times higher than those for the ,
 The subterm for the transverse reinforcement (see and the subterm of transverse reinforcement ( ),
Equ. 12.2) amounts to (see also §4) with an effectiveness factor k = 15.0 or 7.5, instead
Δ1: , of an overall k = 8 for the , with
Δ2: , .  It is interesting that the limited results according
Therefore, what is influencing is the 3rd term, i.e. to this research are in very good correlation with the
that of the lap length, leading to the following values relevant provisions for laps according to the recent
of (MPa): Codes on assessment and retrofitting of existing RC
structures. According to them and to the present re- EC 2, EN 1992-1-1: 2004.
sults, the “activated” could be based on EC 8-1, EN 1998-1: 2004; EC 8-3, EN 1998-3:2005.
Einea, P. et al. 1999. Lap splices in confined concrete. ACI Str.
with √ , by means of J., No. 96-S104.
a “discontinuous” relationship as in Figure 14. Eligehausen, R. et al. 1982. Local τ-s relationship of deformed
bars under generalized excitations. 7 ECEE. Athens.
Eligehausen, R. & Lettow, S. 2007. Background documents for
the new fib MC 2010 (see Fardis 2009).
Fardis, M.N. et al. 2004. Seismic response of RC columns with
lap splices. Res. Rep., The Univ. of Patras (in greek).
Fardis, M.N. 2009. Seismic design, assessment and retrofitting
of concrete buildings (based on EC 8). Dordrecht: Springer.
fib (CEB-FIP) Bull. no. 10. 2000. State of the art report on
bond of reinforcement in concrete. Lausanne.
fib (CEB-FIP) MC 2010 (first complete draft). Bull. nos 55 and
56. Lausanne.
Figure 14. The proposed relationship between and for Gambarova, P.G. et al. 1989. Steel-concrete bond after con-
laps, at least under monotonic actions. crete splitting. Material and Structures, RILEM.
Gambarova, P.G. and Rosati, G.P. 1996. Bond and splitting in
 Additional tests and calibrations are needed and RC. Materials & Structures, RILEM,
are under completion at Lab. of RC/NTUA, for both Gergely, P. & White, R. 1980. Seismic design of lapped splices
monotonic and cyclic loads on fully lapped bars in in RC. 7 WCEE. Istanbul.
Giuriani, E. 1981. Experimental investigation on the bond-slip
highly stressed zones of RC elements; some of the law of deformed bars in concrete. IABSE Coll., DELFT.
issues “deserving” further study have been already Ichinose, T. et al. 2004. Size effect on bond strength of de-
mentioned in the previous paragraphs of this paper. formed bars. Construction and Building Materials. ELSE-
VIER, sciencedirect.
Lucose, K. et al. 1982. Behaviour of RC lapped splices for ine-
lastic cyclic loading. ACI J., No. 79-63.
REFERENCES Lura, P. et al. 2002. 3D FE modelling of splitting crack propa-
gation. Mag. of Concrete Res., 54 (6).
ACI Com. 408 Reports; 408.1R, 408.3R. Melek, M. & Wallace, J. 2004. Cyclic behavior of RC columns
ACI 318R-95 & -08, Chapters 12 and 21. with short lap splices. ACI Str. J., No. 101-S79.
Balázs, G.L. 1991. Fatigue of bond. ACI Mat. J., 88 (6) nGCSI, 2012. The new Greek Code on Structural Assessment
Biskinis,D.E. & Fardis, M.N. 2007. Effect of lap splices on the and Interventions. EPP/GR, Athens.
resistance and cyclic deformation capacity of RC members. Orangun, C.O. et. al. 1977. A revaluation of test data on devel-
Beton und Stahlbetonbau, Sonderheft English, 102. opment length and splices. ACI J., 74 (3).
Bond in Concrete (BIC), 1st Int’l Symp., Paisley, 1982, P. Bar- Paulay, T. 1982. Lapped splices in EQ – resisting RC columns.
tos (ed.), 2nd Int’l Symp., Riga, 1992, A. Shudra (ed.), 3 rd ACI Str. J., 78 (6).
Int’l Symp., Budapest, 2002, G. Balázs et al. (ed.) Paulay, T. & Priestley, M.J.N. 1992. Seismic design of rein-
Bousias, S. et al. 2004. Seismic retrofit of RC columns with lap forced concrete and masonry buildings. New York: John
splices. 13 WCEE. Vancouver. Wiley & Sons.
Cairns, J. & Arthur, P. 1979. Strength of lapped splices in RC Rehm, G. 1968. The basic principles of the bond between steel
columns. ACI J., No. 76-14. and concrete. Cement and Concrete Association, No. 134
Cairns, J. & Jones, K. 1995a. The splitting forces generated by (in english).
bond, Mag. of Concrete Res., 47(171). RILEM/CEB/FIP Rec. 1982 and 1983. Bond test for reinforc-
Cairns, J. & Jones, K. 1995b. Influence of rib geometry on ing steel. RC5, Beam Test, and RC6, Pull-out Test. RILEM.
strength of lapped joints. Mag. of Concrete Res., 47(172). Sener, S. et al. 1999. Size effects on failure of bond splices of
Cairns, J. 2006. Background documents for the new fib MC steel bars in concrete beams. ASCE J. Str. Div., 125 (6).
2010 (see Fardis 2009). Shima, H. et. Al. 1987. Micro and macro models for bond in
CEB Bull. d’ Info. no. 131, 1979; no. 151, 1981; no. 161, 1983; RC. J. of the Fac. of Engg, 39 (2), The Univ. of Tokyo.
no. 210, 1991. Sozen, M. and Moehle, J. 1990. Development and lap-spliced
CEB-FIP MC 1990 (final draft), 1991. Bull. d’ Info. nos lengths for deformed reinforcing bars in concrete. A Report
203÷205. Lausanne. to PCA & CRSI.
Cho, J. – Y. & Pincheira, J.A. 2004. Nonlinear modelling of Tassios, T.P. 1983. Physical and mathematical models for re-
RC columns with short lap splices. 13 WCEE. Vancouver. design of damaged structures. IABSE Symp. Venezia.
Choi, E. et al. 2009. Seismic retrofit of RC columns by NiTi Tepfers, R. 1973. A theory of bond applied to overlapped ten-
and NiTi Nb SMA wire jackets. 8 ESOMAT. Prague. sile reinforcement splices for deformed bars. Calmers Univ.
Choi, O. et al. 2003. Bond performance of highly ribbed rein- of Tech., 73 (2). Goeteborg.
forcing bars to concrete under cyclic loading. Fib Symp. on Tepfers, R. 1979. Cracking of concrete cover along anchored
RC structures in seismic regions. Athens. deformed reinforcing bars. Mag. of Concrete Res., 31
Ciampi, V. et al. 1982. Hysteretic behavior of deformed rein- (106).
forcing bars under seismic excitations. 7 ECEE. Athens. Tepfers, R. 1980. Bond stress along lapped reinforcing bars.
Darwin, D. & Graham, E. 1993. Effect of deformations height Mag. of Concrete Res., 32 (112).
and spacing on bond strength of reinforcing bars. ACI Str. Yannopoulos, P.J. & Tassios, T.P. 1991. RC axial elements
J., 90 (6). under monotonic and cyclic actions. ACI J., 88 (1).
Darwin, D. et al. 1996. Development length criteria for con- Zuo, J. & Darwin, D. 2000. Splice strength of conventional and
ventional and high relative rib area reinforcing bars, ACI high relative rib area bars in normal and high strength con-
Str. J., 93 (3). crete. ACI Str. J., No. 97-S65.

View publication stats