Engineering Structures 109 (2016) 1–15

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Pre-earthquake and post-earthquake retrofitting of poorly detailed

exterior RC beam-to-column joints
George I. Kalogeropoulos a, Alexander-Dimitrios G. Tsonos a,⇑, Dimitrios Konstandinidis b,
Stylianos Tsetines a
a
Department of Civil Engineering, Aristotle University of Thessaloniki, GR-54-124 Thessaloniki, Greece
b
Department of Civil Engineering T.E., Alexander Technological Educational Institute of Thessaloniki, GR-57-400, Greece

a r t i c l e i n f o a b s t r a c t

Article history: In the present study the effectiveness of a proposed scheme for the retrofitting of substandard exterior
Received 27 March 2015 reinforced concrete (RC) beam-to-column joints was experimentally and analytically investigated. The
Revised 5 October 2015 joint sub-assemblages were poorly detailed, representing structural members of old RC buildings,
Accepted 11 November 2015
designed prior to the 1970s. Thus, the columns of the specimens had sparse transverse reinforcement,
Available online 11 December 2015
the beam-column joint region was unconfined and the longitudinal reinforcement of the adjacent beam
had straight anchorage of insufficient length in the joint. The retrofitting technique proposed herein com-
Keywords:
bines the improvement of the beam longitudinal reinforcing bars’ anchorage in the joint, by using exten-
Columns
Beam-to-column joints
sion bars and steel plates and the RC jacketing of the columns and the joint region. The strengthened
Retrofitting specimens conform to the standards of modern design codes for RC structures. To evaluate the effective-
Anchorage ness of the proposed retrofitting scheme the lateral performance of one pre-earthquake and one
Pre-earthquake post-earthquake retrofitted specimen was compared to that of an original, control specimen and that
Post-earthquake of a retrofitted one only by RC jacketing. The experimental results revealed a substantial improvement
Weld of the overall seismic behaviour of the sub-assemblages retrofitted according to the proposed technique,
RC jacket with respect to the lateral performance of the control specimen. Moreover, the specimen retrofitted only
by RC jacketing showed poor seismic performance under the incremental displacement amplitudes of the
lateral loading and developed a failure mode dominated by bond-slip, similar to that of the control
specimen.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction RC buildings. Furthermore, due to the lack of capacity design

approach, the adjacent beams were usually stronger than the col-
Recent earthquakes worldwide have caused significant dam- umns of older structures. Thus, the seismic damages were in most
ages to many reinforced concrete (RC) structures designed and of the cases localized in the over-stressed area of the joints, or in
constructed according to the recommendations of older codes. the lower part of the columns, just above the floor slab, where defi-
These structures, built prior to the 1960s and the 1970s, were cient lap splices of the column longitudinal reinforcement are usu-
poorly detailed and had serious structural deficiencies such as ally located [1]. Experimental tests on RC beam-column joints
low confinement of the columns and the joint region, deficient designed for gravity only, showed that the lack of capacity design
lap splices of the column longitudinal bars, straight anchorages considerations combined with structural inadequacies, typical for
of insufficient length of the beam longitudinal reinforcement (par- old buildings, was a source of significantly brittle damage mecha-
ticularly in the case of exterior joints) and flexural strength ratio nisms especially in the case of exterior joints [2]. Moreover, joint
MR < 1.20. As a result the available ductility was inadequate, partic- failures caused by seismic actions result in severe partial damages
ularly in the case of old RC framed buildings. During the seismic or even collapse of many old RC buildings.
lateral loading substantial moments are concentrated in the The beam-to-column joint failure causes substantial degrada-
small-sized structural members of the beam-to-column joints of tion of the overall lateral performance of RC structures, while the
load carrying capacity of the columns is significantly affected.
Hence, in seismic prone areas pre-earthquake and/or post-
⇑ Corresponding author.
earthquake repair and strengthening of substandard beam-to-
E-mail addresses: [email protected] (G.I. Kalogeropoulos), [email protected].
gr (A.-D.G. Tsonos), [email protected] (D. Konstandinidis). column joints of old structures are required, in order to improve

http://dx.doi.org/10.1016/j.engstruct.2015.11.009
0141-0296/Ó 2015 Elsevier Ltd. All rights reserved.
2 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

the seismic behaviour of the joints by shifting the failure from the 200 1050
joint area to the adjacent beam(s). For this purpose various mate-
rials and rehabilitation schemes have been developed and used
during the last decades [3,4]. Tsonos [5] used the epoxy pressure
injection technique for the rehabilitation of two seismically dam-
aged beam-to-column sub-assemblages that incorporated full seis- 500
mic details of the current building Codes. After the repair
interventions the specimens were imposed to the same lateral dis-
placement history as the original ones. The technique proved to
effectively restore the strength, stiffness and energy dissipation
capacity of the joint sub-assemblages. Karayannis and Sirkelis [6] 300 300
examined the behaviour of exterior beam-to-column joints
strengthened with a combination of epoxy resin injections and
carbon-fiber-reinforced plastics sheets. The enhanced specimens
showed a significantly improved lateral performance with respect
to the original specimens. The effectiveness of shotcrete and cast-
500
in place concrete jacketing of RC building columns and beam-to-
column joints in improving the shear and flexural performance
was investigated by Tsonos [7]. Both types of jackets found equally
satisfactory in strengthening existing old frame structures.
Karayannis et al. [8] used thin reinforced concrete jackets to locally Fig. 1. Reinforcement details of the original beam-to-column sub-assemblages
retrofit earthquake damaged beam-to-column joints. The strength- (dimensions in mm).
ened specimens were retested and the experimental results indi-
cated the success of the jackets in restoring and improving the
lateral performance of the joint specimens. Pampanin et al. [9] inclined main reinforcement to prevent brittle failure in RC beam-
experimentally and analytically investigated the efficiency of CFRP to-column joints. Tests on external beam-to-column joints demon-
laminates in the retrofitting of poorly detailed existing RC build- strated that highly ductile behaviour was achieved by reinforcing
ings designed only for gravity loads (with beam bars anchored with these structural elements with inclined bars [11,12]. Tsonos
endhooks in the joint, deficient lap splices with hooked anchorages [13,14], also used collar stirrups and steel reinforcement in the
and lack of joint transverse reinforcement), typical of the construc- form of inclined bars bent diagonally across the joint core of
tion practice in the Mediterranean countries in the 1950s–1970s strengthened reinforced concrete beam-to-column joints with very
period. Two exterior knee joints, two exterior T-joints and two favourable results. The effectiveness of cross-inclined bars when
interior joint sub-assemblages were subjected to quasi-static cyclic used as the only joint shear reinforcement and in combination with
loading, as well as a three-storey three-bay frame system, while a common stirrups or vertical bars was also examined by Chalioris
simplified analytical procedure was presented to evaluate the et al. [15]. The behaviour of steel jacketing, used as a seismic
sequence of events using a M–N performance domain. The exterior strengthening system for reinforced concrete frame structures
CFRP strengthened sub-assemblages showed desirable ductile and designed solely to gravity loads under strong beam-weak column
dissipating hysteresis behaviour with the formation of the plastic concept, was experimentally investigated by Ruiz-Pinilla et al.
hinges in the beam, while the interior joints exhibited an accept- [16]. It was concluded that the axial load and the strengthening
able and controlled minor cracking in the joint panel zone. A par- techniques applied on columns significantly affect the lateral per-
tial retrofitting strategy using CFRP laminates was adopted in the formance of the joints. Filiatrault et al. investigated the use of
case of the three-storey three-bay frame structure, which proved steel-fiber reinforced concrete to improve the seismic behaviour
to be very satisfactory in improving the lateral behaviour and pre- of beam-column interior joints [17]. Tsonos [18] proposed a new
venting brittle failure of the exterior joints and the formation of innovative technique for the rehabilitation of exterior beam-
soft-storey mechanism. Mahmoud et al. [10] experimentally exam- column connections combining the use of a steel fiber ultra-high-
ined the behaviour of defected beam-column joints with absence strength concrete jacket for the strengthening of the joint region
of transverse reinforcement in the joint, insufficient bond length with CFRP jackets for the strengthening of the columns.
of the beam longitudinal reinforcement and inadequately spliced The experimental research related to the importance of improv-
implanted column, when strengthened using Carbon Fiber Rein- ing the beam reinforcement anchorage in exterior joints of older
forced Polymer (CFRP) strips, CFRP sheets and near surface structures is rather poor. Bedirhanoglu et al. [19] improved the
mounded CFRP strips. Tsonos [11] first introduced the idea of using anchorage of the beam longitudinal reinforcement in the joint by

Table 1
Retrofitting scheme applied to the sub-assemblages and cross-section details.

Original specimens Retrofitted specimens

K1 L1 M1 SK1 SL1 SM1
Beam-to-column sub-assemblages
Retrofitting


Post-earthquake Pre-earthquake Pre-earthquake
type
Retrofitting


(1) (RC jacketing and anchorage (1) (RC jacketing and anchorage (2) (RC jacketing, No anchorage
scheme improvement) improvement) improvement)
RC jacketing


Columns and joint region Columns and joint region Columns and joint region
Beam section 200  300 200  300 200  300 200  300 200  300 200  300
(mm)
Column section 200  200 200  200 200  200 300  300 300  300 300  300
(mm)
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 3

(a) Step 1: The cover concrete of the columns (b) Step 2: Longitudinal reinforcement of the RC
and the joint was chipped away. jacket. Extention bars were used to extend the
beam logitudinal reinforcement.

(c) Step 3: Closed ties were used for the transverse (d) Joint of the retrofitted specimens SK 1 and SL 1
reinforcement of the RC jacket, while the after the implementation of the retrofitting
anchorage of the beam reinforcement was scheme (1).
achieved via steel plates and bolts.

Fig. 2. Retrofitting scheme (1) implemented to sub-assemblages SK1 and SL1.

welding the hooks of the top bars to the hooks of the bottom bars. and found that even the joints conforming to the seismic detailing
A 35 percent increase of the specimens’ strength was observed standards of modern Codes may exhibit excessive damage. In the
although the column and the beam did not reach their nominal common practice the deficiency of the beam reinforcement
strength due to limitation of the capacity of the specimens by anchorage is usually underestimated during the retrofitting of
the damage in joint core. Hakuto et al. [20] experimentally investi- old buildings. Nevertheless, even when additional external con-
gated the seismic behaviour of an exterior beam-to-column joint finement is applied to the columns and the joint via jacketing,
with reinforcement details typical of pre-1970 buildings, namely the inadequate development length and the absence of ninety-
the low confinement of the columns and the joint region and hooks degree hooks of the beam bars in the joint region still cause bond
at the ends of the beam bars that were bent up in the case of the failure between the steel bars and concrete under reversed lateral
top beam bars and down in the case of the bottom beam bars. loading, resulting in the pullout of the bars.
The performance of this joint specimen was subsequently com- In this study the effectiveness of a proposed technique for the
pared to that of another joint specimen with the hooks at the pre-earthquake and post-earthquake retrofitting of exterior RC
end of the beam bars bent into the joint core. It was concluded that beam-to-column joint specimens with low confined columns,
the exterior joint sub-assemblage in which the hooks of the beam unconfined joint area and primarily with short straight anchorages
reinforcement were bent out of the joint core behaved unsatisfac- of the beam reinforcement in the joint is experimentally and ana-
torily during a severe earthquake. Tsonos [21] tested six RC exte- lytically investigated. According to the proposed retrofitting
rior beam-to-column joint specimens typical of modern scheme the anchorage of the beam longitudinal bars was improved
structures that have been designed according to the recent Codes by extending them via additional threaded-end bars to the back
4 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

(a) Transverse reinforcement of the RC jacket and (b) Longitudinal and transverse reinforcement of
extension bars (specimens SL1 and SK1). the RC jacket and improvement of the beam
bars’ anchorage in the joint region (specimens
SL1 and SK1).

Fig. 3. Construction of the RC jackets of sub-assemblages SL1 and SK1.

Fig. 4. Welding of the bars according to the CSRTC [24]. 1: welding starts 10 mm from the bar end, 2: welding direction, 3: space between the welds.

face of the exterior joint, where the anchorage was achieved using while the residents of the strengthened building would be in fact
steel plates. A high strength four-sided RC jacket with additional unaware of the danger.
longitudinal and transverse reinforcement was subsequently con-
structed around the columns and the joint.
2. Experimental investigation
Ultimately, the test results showed that improving the anchor-
age of the beam longitudinal reinforcement in exterior joints sig-
Beam-to-column joints are part of the vertical supporting sys-
nificantly improves the behaviour of the joints when imposed to
tem of structures hence, if they are significantly damaged the duc-
seismic displacement reversals and prevents the premature pull-
tility of the columns is strongly affected, while the safety of the
out failure of the beam reinforcing bars from the joints. Moreover,
structures is jeopardized. Modern codes for the design of RC struc-
it was clearly demonstrated that a wrong interpretation of the fail-
tures specify construction details to avoid the formation of plastic
ure mode, caused by the pullout of the beam reinforcement from
hinges in the joint region. However, the existing buildings possess
the joint region as flexural-type failure, results in the implementa-
structural deficiencies, which may cause severe joint damages or/
tion of retrofitting schemes that are inadequate to improve the lat-
and collapse under reversed lateral displacements of moderate
eral performance of the joint sub-assemblage and therefore, the
earthquakes. Despite the wealth of experimental research on the
exterior beam-to-column connections of pre-earthquake and/or
retrofitting of beam-to-column joint structural members the
post-earthquake retrofitted buildings would still be prone to pre-
research relating to the significance of improving the beam rein-
mature pullout failure of the beam reinforcement when imposed
forcement anchorage in the joint region (particularly in exterior
to cyclic lateral loading. Hence, despite the cost of the rehabilita-
beam-to-column joints) is rather poor. In this paper the efficiency
tion interventions the safety of the structure would be jeopardized,
of a retrofitting method for beam-to-column joint sub-
assemblages, representative of those found in structures built prior
to the 1970s, was experimentally and analytically investigated. Six
Table 2 exterior beam-to-column joint specimens of 1:2 scale were
Design of the column RC jackets in shear. designed and constructed. Three of them were identical original
Shear design parameter values (Eurocode 2 and 8) specimens, typical of beam-to-column joints found in older exist-
ing buildings. These specimens, named K1, L1 and M1, possessed
Vsda (kN) VRD,max (kN) VRd,s P Vsd (kN) Transverse reinforcement – RC
jacket of the columns the properties of substandard beam-to-column joints with widely
182.79 609.74 > Vsd For VRd,s P Vsd Closed B500C Ø8/70 mm ties spaced ties in the columns and the joint region and deficient
sb 6 450 mm straight anchorages of the beam top and bottom longitudinal rein-
a
Vsd: Shear force corresponding to the moment resistance of the retrofitted forcement (see Table 1). Reinforcement details of the original spec-
columns. imens are shown in Fig. 1. For both the longitudinal and transverse
b
s: Spacing of the stirrups. reinforcement of the sub-assemblages B500C steel bars were used.
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 5

Table 3
Column critical height and spacing of the transverse reinforcement according to Eurocode 2 and 8.

Eurocode 2 and 8
Column critical height (m) EC 8 – 5.4.3.2.2.(4) minØw: 8 mm
scl;t;max (mm) scl;t;max (mm)
Spacing of the ties along the critical height Spacing of the ties outside of the critical height
n o n o  
max hc ; l6cl ; 0:45 m min 8  dbL;min ; b2o ; 175 mm min 20  dbL;min ; bc ; 400 mm

0.50 80 200

bo : width of the confined core (to the centerline of the hoops) bo ¼ b
2  ðc þ £w =2Þ, where £w denotes the diameter of ties and c is the width of the cover concrete, dbL;min :
minimum diameter of the longitudinal steel bars.

Table 4
Reinforcement details of the original and the strengthened beam-to-column joint
sub-assemblages.

Reinforcement of the original and retrofitted specimens

Specimen Beam Column Joint
Longitudinal reinforcement (mm) – B500C
K1 6Ø10 4Ø10 4Ø10
L1 6Ø10 4Ø10 4Ø10
M1 6Ø10 4Ø10 4Ø10
SK1 6Ø10 4Ø10 (RC jacket) 4Ø10 (RC jacket)
SL1 6Ø10 4Ø10 (RC jacket) 4Ø10 (RC jacket)
SM1 6Ø10 4Ø10 (RC jacket) 4Ø10 (RC jacket)
Transverse reinforcement (mm) – B500C
K1 Ø8/100 (crit.)/Ø8/150 Ø6/200

L1 Ø8/100 (crit.)/Ø8/150 Ø6/200

M1 Ø8/100 (crit.)/Ø8/150 Ø6/200

SK1 Ø8/100 (crit.)/Ø8/150 Ø8/70 (RC jacket) Ø8/50 (RC jacket)
SL1 Ø8/100 (crit.)/Ø8/150 Ø8/70 (RC jacket) Ø8/50 (RC jacket)
SM1a Ø8/100 (crit.)/Ø8/150 Ø8/70 (RC jacket) Ø8/50 (RC jacket)
Beam bars anchorage improvement
SK1, SL1 6Ø10 B500C straight bars, steel plates and bolts
a
Specimen SM1 was strengthened only by RC jacketing, while no measures for
the beam reinforcement anchorage improvement were undertaken.
Fig. 5. Test setup. Detail of the original beam-column sub-assemblage. (1): Two-
way actuator (1000 kN compression, 600 kN tension), (2): Hydraulic jack (300 kN),
The concrete compression strength of the specimens was mea- (3): hinge arrangements, (4): load cell.

sured on 150  300 mm cylinders (see Table 5). Specimen K1 was

imposed to cyclic lateral deformations to failure and used as the
respectively and were seismically tested according to the same
control specimen. After the seismic test the anchorage of the beam
loading history as K1. The lateral behaviour of the strengthened
top and bottom longitudinal steel bars in the joint region was
sub-assemblages SK1 and SL1 was compared to that of the control
improved, while the joint and the columns of K1 were subse-
specimen, K1, to evaluate the effectiveness of the retrofitting tech-
quently retrofitted by RC jacketing. To improve the anchorage six
nique applied to SK1 and SL1 in both the cases of post-earthquake
straight B500C steel bars of 10 mm diameter each were inserted
and pre-earthquake enhancement of the joint specimens. The
to six holes drilled from the backside to the front side of the joint
importance of improving the beam bars’ anchorage in the joint
and were welded-up to the longitudinal bars of the beam via seg-
region was evaluated by comparing the lateral performance of
ments (see Fig. 2(b)). The extension bars had a threaded end and
specimen SM1 with that of the control specimen K1 and the beha-
were anchored in the backside of the joint via steel plates and
viour of the retrofitted joints SK1 and SL1.
bolts. The beam-to-column joint sub-assemblage L1 was pre-
earthquake retrofitted by the same technique as specimen K1,
while in the case of M1 the pre-earthquake retrofitting did not 3. Retrofitting of the sub-assemblages SK1, SL1 and SM1
involve the improvement of the beam bars’ anchorage in the joint
region. The retrofitting schemes applied to the specimens are The retrofitting of both the pre-earthquake and post-
shown in Table 1. After the interventions to K1, L1 and M1 earthquake strengthened sub-assemblages SL1 and SK1 involved
the strengthened specimens were designated SK1, SL1 and SM1 the improvement of the beam longitudinal bars’ anchorage in

Table 5
Design parameters of the beam-to-column joint sub-assemblages.

Specimen MR a
ðM0R : ACI
ASCE 352R
02Þ c a (MPa) 0
f c (MPa)

K1 1.62 (1.40) 0.650 (1.00) 22.44

L1 1.58 (1.40) 0.680 (1.00) 20.56
M1 1.59 (1.40) 0.613 (1.00) 25.09
SL1 2.82 (1.40) 0.300 (1.00) 58.30 (RC jacket)
SK1 2.91 (1.40) 0.288 (1.00) 58.30 (RC jacket)
SM1 3.06 (1.40) 0.272 (1.00) 60.58 (RC jacket)
a
Numbers outside the parentheses are the provided values, numbers inside the parentheses are the required values by the ACI-ASCE Committee 352R-02.
6 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

the joint region and the RC jacketing of the columns and the 4. Design of the RC jackets
joint of the specimens (Table 1). The beam-to-column joint spec-
imen M1 was strengthened only by RC jacketing of the columns The RC jackets of the retrofitted sub-assemblages were
and the joint, while no further means were undertaken for designed in shear to conform to the requirements of Eurocode 2
improving the anchorage of the beam’s longitudinal reinforce- and 8 and the recommendations of ACI-ASCE Committee 352R-02
ment. During the retrofitting process the cover concrete of the [25]. Eq. (1) was used to calculate the concrete compression
columns and the joints of the specimens was chipped away. In strength of the enhanced columns and joints. The shear design
the cases of sub-assemblages SK1 and SL1 six holes were drilled parameter values of the strengthened specimens are summarized
to the concrete of the joint region in appropriate positions to in Table 2.
allow the extension of the beam longitudinal reinforcing bars. The strengthened columns were designed to conform to the
Six straight 10 mm diameter B500C steel bars with a threaded provisions of Eurocode 2 and 8. Therefore, the control of the con-
end were inserted in the holes from the back to the front side crete compression strut adequacy was made according to Eq. (2).
of the joint and were welded to the beam bars via 8 mm In Table 3 the Eurocode provisions for the shear reinforcement
B500C steel segments. After the welding of the beam reinforce- are presented. The transverse reinforcement demand was finally
ment with the extension bars, the beam section was restored given by Eq. (3), where it is recommended that V Rd;s P V sd . In
by using a premixed, non-shrink, rheoplastic, flowable and Eqs. (1–3) As;ex is the cross-section area of the existing column;
non-segregating mortar of high strength. Moreover, due to the As;j is the cross section area of the RC jacket; f ck;ex is the concrete
extended cross-section dimensions of the RC jacket with respect characteristic strength of the existing column; f ck;j is the concrete
to the dimensions of the original columns, the RC jacket extends 0
characteristic strength of the jacket; f ck is the equivalent concrete
beyond the location of the welds in the beam (see Fig. 2d). Thus,
characteristic strength of the retrofitted column; V Rd;max is the
the strength of the beam was slightly increased. The longitudinal
strength of the concrete compression struts; V Rd;s is the shear force
steel reinforcement of the RC jacket (10 mm B500C) was subse-
resisted by the hoops; Asw is the cross section area of the transverse
quently welded to the longitudinal reinforcement of the original
reinforcement; s is spacing of the stirrups; f ywd is the yield strength
specimens via s-shaped segments. The bars used to extend the
beam reinforcement were anchored to the backside of the joint of the transverse reinforcement; bw is the cross-sectional depth of
by using steel plates and bolts. The plates were placed externally the column; z ¼ 0:9  d; acw is a coefficient taking account of the
in contact with the longitudinal bars of the jacket. For the con- state of the stress in the compression strut, m1 is the strength
struction of the jacket transverse reinforcement in the joint reduction factor for concrete cracked in shear with recommended
region, four holes were drilled into the concrete of the beam in value of 1
f ck ðMPaÞ=250 and h is the angle between the concrete
the conjunction with the joint. Four L-shaped 8 mm B500C bars compression strut and the member axis perpendicular to the shear
were inserted into the holes and suitably bent to form close force (h = 21.8°). Ultimately closed B500C ties of 8 mm diameter
hoops. The ends of the hoops were welded. Closed hoops were spaced at 70 mm were be placed along the critical column height
also used for the transverse reinforcement of the jacket in the of the strengthened sub-assemblages.
columns. The steps of the retrofitting process for specimens SL1 0
ðAs;ex  f ck;ex þ As;j  f ck;j Þ=ðAs;ex þ As;j Þ ¼ f ck ð1Þ
and SK1 are clearly illustrated in Fig. 2. The cement grout jacket
of the retrofitted sub-assemblages was constructed according to V Rd;max ¼ acw  bw  z  v 1  f cd =ðcot h þ tan hÞ ð2Þ
the provisions of Eurocode 2 and 8 [22,23]. In Fig. 3 the con-
struction of the RC jacket reinforcement is presented. The weld- V Rd;s ¼ ðAsw =sÞ  z  f ywd  cot h P V sd ð3Þ
ing of reinforcement steel bars satisfied the requirements of the
Code for Steel Reinforcement Technology for Concrete (CSRTC)  00 0 
Ash ¼ 0:3  ðsh  h  f c Þ=f yh  ðAg =Ac
1Þ ð4Þ
[24] (see Fig. 4). A premixed, non-shrink, rheoplastic, flowable
and non-segregating mortar was used for constructing the jacket 00 0
of the specimens. Ash  0:09  ðsh  h  f c Þ=f yh ð5Þ

(a) (b)
Fig. 6. (a) Lateral displacement history and (b) qualitative deformed shape of the specimens.
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 7

(a) Hysteresis loops of specimens K1 and SM1 (b) Hysteresis loops of specimens K1 and SL1

(c) Hysteresis loops of specimens K1 and SK1 (d) Hysteresis loops of specimen SM1

(e) Hysteresis loops of specimen SL1 (f) Hysteresis loops of specimen SK1
Fig. 7. Hysteretic behaviour of the original and strengthened beam-to-column joint specimens.

The RC jacket of the beam-to-column joint region of the speci- forcement in joints of primary seismic beams with columns should
mens was designed in shear to conform to the recommendations of be not less than that specified for the critical regions of columns.
the ACI-ASCE Committee 352R-02 and the Eurocode 2 and 8. The Eqs. (4) and (5) were used to calculate the transverse reinforce-
expression RM Rc P 1:3  RM Rb should be satisfied according to the ment of the RC jacket of the joint region according to the recom-
weak girder-strong column design philosophy of Eurocode 8, while mendations of the ACI-ASCE Committee 352R-02. Ash is the total
the ACI-ASCE Committee 352R-02 recommends an ultimate cross-sectional area of all legs of hoop reinforcement, including
strength ratio RM Rc =RM Rb ¼ 1:2. According to Eurocode 8 for crossties, crossing a section having core dimension bc22 (bc22 is
DCM (ductility class medium) the horizontal confinement rein- the core dimension of tied column, outside to outside edge of
8 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

transverse reinforcement bars, perpendicular to the transverse improvement is also examined by comparing the seismic beha-
reinforcement area Ash being designed); the factor sh denotes the viour of sub-assemblages SL1 and SK1 with that of SM1.
00
distance between hoops; h is the height of a rectangular column Fig. 7 illustrates the plots of applied shear-versus-displacement
core section; Ag is the column section area; Ac is the column core of the beam-to-column joint specimens. The original sub-
0 assemblage, K1, performed poorly under the incremental displace-
section area; f yh denotes the yield stress of the hoops and f c is
the specified compressive strength of concrete in the connection. ment amplitudes of the seismic loading, showing a sharply degra-
Eventually closed B500C ties of 8 mm diameter spaced at 50 mm dation of the lateral strength from the second cycle of loading and
were placed in the joint region of the strengthened sub- severe deterioration of the energy dissipation capacity and peak-
assemblages. In Table 4 the longitudinal and transverse reinforce- to-peak stiffness, as well. Bond failure between the deficiently
ment of the RC jackets of the retrofitted specimens is presented. anchored steel beam bars and concrete resulted in the premature
The anchorage of the beam longitudinal reinforcement to the pullout of the bars from the joint, while excessive slipping domi-
joint of the strengthened specimens SK1 and SL1 was achieved nated the seismic behaviour of K1 after the second cycle. The
via extension bars, steel plates and bolts according to the provi- pre-earthquake and the post-earthquake strengthened specimens
sions of Eurocode 8. SL1 and SK1 showed a significantly improved lateral response with
respect to that of the original beam-to-column joint, K1. The
spindle-shaped hysteresis loops of SL1 and mainly of SK1 reflect
5. Test setup and loading history

The seismic tests on the beam-to-column joint sub-assemblages

K1, SK1, SL1 and SM1 were conducted in the test frame shown in
Fig. 5. The inflection points of the columns were simulated with
the aid of specific arrangements connected to the test frame and
the free ends of the columns by hinges. Thus, the horizontal and
vertical displacements of the columns’ ends were restrained, while
able to rotate freely. All specimens were subjected to a large num-
ber of inelastic cycles of incremental displacement amplitudes. The
qualitative deformed shape of the specimens is given in Fig. 6(b).
The lateral load was applied to the beam free end by a two-way
actuator and the load values were measured using a load cell. A
constant axial load of 200 kN was imposed to the columns of the
sub-assemblages by a hydraulic jack (see Fig. 4). The load point dis-
placement was measured by a calibrated linear variable differen-
tial transducer (LVDT). Electrical resistant strain gages were
installed on the longitudinal steel bars of the beam and on the ties
of the joint of both the control and the retrofitted specimens. Since Fig. 8. Envelope curves of the sub-assemblages K1, SK1, SL1 and SM1.
the behaviour of sub-assemblages is mainly demonstrated by the
envelope curves, an ascending lateral displacement history with
one cycle per amplitude of displacement was adopted, without
considerable influence in the seismic performance of the sub-
assemblages. All specimens were loaded transversely following
the displacement-controlled schedule shown in Fig. 6(a). A test
specimen similar to the original beam-to-column joint K1 but with
well anchored longitudinal beam reinforcing bars into the joint
region was used to determine the steps of loading and was at first
loaded to its yield displacement. This was measured from the plot
of applied shear-versus-displacement of the specimen for the point
when a significant decrease in stiffness occurred and was also ver-
ified by the yielding of the longitudinal beam reinforcement at the
column face. The loading was continued in the same direction
(push cycles) to 1.5 times the yield displacement and the sub-
Fig. 9. Energy dissipation comparison.
assemblage was subsequently unloaded and loaded in the other
direction (pull cycles) to the same lateral displacement. After the
first cycle of loading, the maximum displacement of each subse-
quent cycle was incremented by 0.5 times the yield displacement
[26–28].

6. Experimental results and failure mode of the specimens

The overall lateral performance of the original (K1) and

strengthened (SL1, SK1 and SM1) beam-to-column joint sub-
assemblages is subsequently compared, to evaluate the effective-
ness and suitability of the implemented retrofitting schemes for
the pre-earthquake and post-earthquake retrofitting of the speci-
mens. Thus, the energy dissipation capacity, peak-to-peak stiffness
and lateral strength in each cycle of the reversed lateral loading are
compared. Moreover, the significance of the beam bars’ anchorage Fig. 10. Peak-to-peak stiffness of the original and strengthened sub-assemblages.
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 9

(a) Shear strength – upper half cycles (b) Shear strength – lower half cycles

(c) Applied shear ratios SL1/SM1, SK1/SM1 (d) Applied shear ratios SL1/SM1, SK1/SM1
and SK1/SL1 – upper half cycles and SK1/SL1 – lower half cycles

(e) Applied shear ratios SM1/K1, SL1/K1 (f) Applied shear ratios SM1/K1, SL1/K1
and SK1/K1 – upper half cycles and SK1/K1 – lower half cycles
Fig. 11. Shear strength of the sub-assemblages and strength ratios SL1/SM1, SK1/SM1, SK1/SL1, SM1/K1, SL1/K1 and SK1/K1.

the high values of seismic energy dissipated in the plastic hinges, scheme implemented to SM1 failed to improve adequately the seis-
formed in the adjacent beam of the specimens (see Fig. 9). After mic behaviour of the specimen and shift the damage in the beam.
six cycles of lateral loading, the energy dissipation ratios SL1/K1 Bond-slip still dominated the lateral response of SM1, which per-
and SK1/K1 equaled to 4.53 and 8.29, respectively. Unlike SL1 and formed identically to the original beam-to-column joint, K1. Thus,
SK1 the pre-earthquake retrofitted beam-to-column joint, SM1, the values of lateral strength, peak-to-peak stiffness and energy
was strengthened only by RC jacketing in the columns and the joint dissipation capacity of SM1 were only slightly higher than those
region, while no additional measures to improve the beam rein- of K1, particularly for drift angle, R, beyond 2.38 percent. For drift
forcement anchorage in the joint were undertaken. From the hys- angle R equal to 3.33 percent the beam-to-column joint
teresis loops of the specimen it is obvious that the retrofitting sub-assemblage SM1 showed a 50.1 percent increase in energy
10 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

(a) Specimen K 1 (original). (b) Specimen SL1 (retrofitting scheme (1)).

(c) Specimen SK1 (retrofitting scheme (1)). (d) Specimen SM 1 (retrofitting scheme (2)).
Fig. 12. Failure mode of the original (a) and the retrofitted beam-to-column sub-assemblages (b–d). ⁄Usually the pullout cracks are wrongly considered as flexural-type
cracks. This may lead to dangerous collapses of strengthened buildings during strong and catastrophic earthquakes.

dissipation value with respect to the control specimen, K1, while percent for SK1. The strength of SL1 and SK1 showed a mild reduc-
the energy dissipation ratio values SM1/SL1 and SM1/SK1 were tion during the incremental displacement of the lateral loading and
equal to 0.325 and 0.205, respectively. after six cycles the specimens maintained more than 50 and 70
The peak-to-peak stiffness of the specimens is summarized in percent (in both the upper and lower half cycles) of their initial
Fig. 10. The beam-to-column joint sub-assemblages SL1 and SK1 strength, respectively. Contrariwise, after five cycles of loading
showed a substantial increase in stiffness values with respect to the shear strength of joint specimen SM1 (in both the push and pull
the original specimen. For drift angle, R, equal to 1.43 percent the half cycles) was about a quarter of the initial strength value.
stiffness values of SL1 and SK1 was 2.35 and 2.53 times the corre- The success of the proposed scheme (1) for the retrofitting of
sponding value of K1, respectively. After six cycles of loading an substandard exterior beam-to-column connections of old build-
increase of 287 and 460.82 percent in stiffness of SL1 and SK1 ings, which includes the RC jacketing of the columns and the joint
was observed, with respect to the stiffness of K1. The retrofitted region and the improvement of the deficient beam bars’ anchorage
specimen SM1 showed a smaller increase in stiffness with respect in the joint, is clearly demonstrated by the failure mode of the sub-
to the original sub-assemblage. During the first cycle of the seismic assemblages (see Fig. 12). The original specimen, K1, exhibited an
loading the stiffness ratio SM1/K1 was equal to 1.82, while for load- early pullout failure of the insufficiently anchored longitudinal
point lateral displacement of 35 mm the ratio value was 1.4. steel reinforcement of the beam. A wide opened flexural-type crack
All the retrofitted sub-assemblages showed higher lateral was formed in the beam in the conjunction with the joint of K1,
strength than the original specimen as it can be observed in while the columns remained intact until the end of the seismic test
Fig. 11(a) and (b) and from the envelope curves of the specimens, and only few hairline cracks were formed in the joint of the spec-
presented in Fig. 8. The shear forces resisted by the strengthened imen (see Fig. 12(a)). Despite the retrofitting of specimen SM1 with
beam-to-column joint SM1 however, were significantly lower than RC jacketing, the failure mode of the sub-assemblage was similar
those resisted by specimens SL1 and SK1 (see Fig. 11(c) and (d)), to that of the original beam-to-column joint, K1 (see Fig. 12(d)).
due to the inadequacy of the beam reinforcement anchorage in Thus, the retrofitting scheme implemented to SM1 failed to prevent
the joint, which resulted in excessive slipping and early pullout the pullout failure and to shift damage in the beam. Consequently,
of the beam bars. Hence, a 42.3 (push half cycle) and 27.8 (pull half if the improvement of the beam bars’ anchorage in exterior beam-
cycle) percent decrease in strength of specimen SM1 was observed to-column joints is ignored during the retrofitting of old buildings,
during the second cycle of the seismic loading, while the corre- a usual common practice, the safety of the joints is jeopardized.
sponding values in the cases of SL1 and SK1 were 19 and 13.6 per- Moreover, if the flexural-type crack pattern of specimens K1 and
cent (decrease) for SL1 and 2.13 (decrease) and 2.28 (increase) SM1 is underestimated, severe partial damages or even collapse
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 11

(a) Specimen K 1: No1 was located in the top beam (b) Specimens SL 1 and SK1: No1 was located in
longitudinal bar and No2 in the tie of the joint. the top beam longitudinal bar, No2 in the
extension bar and No3 in the tie of the RC
jacket in the joint region.

(c) Specimen SM1: No1 was located in the top

beam longitudinal bar and No2 in the tie of
the RC jacket in the joint region.
Fig. 13. Location of the strain gages – specimens K1, SL1, SK1 and SM1.

may occur during a future earthquake, particularly in RC framed nated by sudden bond-slip failure and premature pullout of the
structures. The beam-to-column joint sub-assemblages strength- beam reinforcing bars from the joint, which was clearly demon-
ened by the proposed scheme (1), SL1 and SK1, both showed a strated in the hysteresis loops of the sub-assemblages. Moreover,
remarkable seismic performance (particularly specimen SK1) and the development length of the beam longitudinal reinforcing bars
exhibited a ductile failure mode with the formation of the plastic into the joint region of sub-assemblages K1 and SM1 was very short
hinges in the beam. The columns and the joints of the specimens and thus, inadequate to ensure steel yielding and prevent the slip-
remained intact during the seismic tests, while considerable slip- ping of the bars. Therefore, the use of ribbed bars instead of plain
ping of the beam longitudinal reinforcing bars was observed for S220 steel reinforcement had only a slight influence in the seismic
drift angle beyond 4.29 percent in the case of specimen SL1 and performance of these beam-to-column joints specimens. From the
may be attributed to partial failure of the welding between the experimental results it is clearly demonstrated that the improve-
beam reinforcement and the extension bars in some cases (see ment of the beam reinforcement anchorage in the joint region
Fig. 12(b) and (c)). However, both specimens SL1 and SK1 showed (specimens SK1 and SL1) is critical and absolutely necessary to
a desirable failure mode with the damage accumulated in the ensure that the failure would be shifted in the beam and thus,
beams. Thus, the retrofitting technique proposed herein found to achieve a substantial improvement of the overall lateral behaviour
be an effective way of significantly improving the inelastic beha- of the strengthened sub-assemblages.
viour of substandard exterior beam-to-column joints found in old
structures.
In this experimental and analytical research B500C ribbed steel
reinforcing bars were used for the construction of both the original 7. Steel micro-strain monitoring
specimens and the RC jackets of the strengthened sub-assemblages
instead of plain S220 or ribbed Stahl III reinforcement, which were The plots of displacement-versus-strain of the steel reinforcing
used in older structures. Given that both S220 and Stahl III steel bars of the original and strengthened beam-to-column joint speci-
bars could be conditionally welded, the difference with B500C mens are illustrated in Fig. 14. The steel strains were measured
reinforcement is mainly focused in the steel tensile strength. This using electrical resistant strain gages. The location of each strain
parameter however, had a minimum impact in the seismic gage is presented in Fig. 13. The original specimen K1 and the
response of the original and strengthened beam-to-column sub- strengthened sub-assemblage SM1 exhibited early pullout failure
assemblages, since the failure mode of the original specimen K1 of the beam reinforcement from the joint. Hence, minimal strain
and the strengthened one only with RC jacketing, SM1, was domi- of the longitudinal beam bars was expected for both specimens.
12 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

Fig. 14. Plots of displacement-versus-strain of the beam-to-column sub-assemblages’ reinforcement. (a, b): Specimen K1, (c, d): specimen SL1, (e): specimen SK1, (f):
specimen SM1.

The plots shown in Fig. 14(a), (b) and (f) indicate that both increase in maximum steel strain during consecutive cycles of
beam-to-column joints performed poorly under the incremental loading indicates the absence of bar slipping and ductile lateral
displacement amplitudes of the seismic loading. Contrariwise, response, while the maintaining or decrease of strain values
the strain values of both the beam longitudinal reinforcement reflects hysteresis deterioration due to the slipping of the bars, as
and the extension bars used for the beam bars’ anchorage improve- long as buckling has not taken place. Thus, it is clearly demon-
ment in the cases of specimens SL1 and SK1 showed a continuous strated (see Fig. 15) that a significant superiority was achieved in
increase and even exceeded the steel yield stress of 2.5‰ (see the hysteresis of sub-assemblages SL1 and SK1 with respect to
Fig. 14(c)–(e)). This increase in strain values indicates that the specimens K1 and SM1.
extension of the beam bars to the backside of the exterior joints Strain values observed in the transverse reinforcement of the RC
of SL1 and SK1 and the use of steel plates and bolts substantially jacket of specimens SL1 and SK1 were higher than the correspond-
and successfully improved the anchorage of the beam reinforce- ing values of SM1, while in the case of the original sub-assemblage,
ment. Thus, excessive slipping and pullout of the bars were pre- K1, the strain values of the tie in the joint were minimal (strain
vented. According to Ehsani and Whight [26] the continuous gage No. 2).
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 13

8. Range of application and significance of the proposed retrofit Thessaloniki, as well as using data from similar experiments car-
scheme ried out in the United States, Japan and New Zealand. More details
about the formulation are given in [7,14,21]. In Table 6 the pre-
The retrofit procedure proposed herein (retrofit scheme (1)) can dicted shear values according to the formulation are demonstrated.
be applied to exterior RC beam-to-column joints of both two- When the computed joint shear stress is greater or equal to the
dimension and three-dimension RC frames. In the case of three- joint ultimate capacity ccal P cult the predicted actual value of the
dimension RC frames slight modifications are required in order connection shear stress will be near cult because the connection
to implement the technique. Thus, whilst the interventions for fails earlier than the beam(s). On the contrary when ccal < cult the
the improvement of the beam reinforcement anchorage would be predicted actual value of the connection shear stress will be near
the same as in the case of two-dimension frames, the transverse to ccal because the connection permits its adjacent beam(s) to yield.
reinforcement of the RC jacket in the joint region would consist Eq. (6) was adopted for the representation of the biaxial concrete
of ties, each one of which would be formed using a t-shaped and strength curve. The value of the joint ultimate capacity cult is given
a straight steel bar, bent properly to form a close hoop. The ends by solving the system of Eqs. (6) and (7) for a given value of joint
of the bars should subsequently be welded. Due to the presence aspect ratio, a.
of the slab additional holes should be drilled in the slab to allow
the longitudinal reinforcement of the RC jacket to pass through. ðx þ yÞ5 þ 10y
10x ¼ 1 ð6Þ
The transverse beams and the slab provide additional confinement
to the joint region, which, nevertheless, is inadequate to prevent qffiffiffiffiffiffi qffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
bond-slip failure between steel bars and concrete and the prema- x ¼ ac=ð2 f c Þ; y ¼ ½ac=ð2 f c Þ  1 þ ð4=a2 Þ ð7Þ
ture pullout of the beam reinforcement from the joint, due to the
Due to the poor reinforcement detailing and the structural defi-
extremely short length of the straight anchorages of the beam lon-
ciencies of the beam-to-column joint specimen K1, the original
gitudinal reinforcing bars. Thus, the anchorage improvement into
sub-assemblage developed a premature failure, caused by the pull-
the joint region is still critical and absolutely necessary for the ret-
out of the beam longitudinal reinforcement. Thus, beam-to-column
rofitting process.
joints representative of those found in pre-1970 buildings appear
If other buildings exist nearby minor modifications should also
to perform as ‘‘weak links” in RC frames. The strengthening scheme
be made in order to apply the proposed retrofit scheme, which
implemented on sub-assemblages SL1 and SK1 was designed
combines the RC jacketing of the columns and the joint region with
according to modern Code standards [22,23,25] and also included
the improvement of the beam reinforcement anchorage in the
the improvement of the beam reinforcement anchorage in the
joint. Instead of steel plates and bolts, which were used in the pre-
joint. Test results clearly show that this scheme effectively
sent study for the anchorage of the extended beam reinforcement
improved the overall seismic behaviour of the specimens by shift-
in the backside of the joint, t-shaped steel bars could enclose the
ing the failure to the beam, while the joint and the adjacent col-
joint and the free ends of these bars should be subsequently
umns remained intact and the pullout of the beam steel bars was
welded to the longitudinal beam reinforcement (after the removal
prevented. The retrofitting scheme implemented on specimen
of the cover concrete). Moreover, a three-sided RC jacket could be
SM1 proved to be ineffective and the specimen showed a lateral
constructed instead of a four-sided one, since it can significantly
performance almost similar to that of the original specimen K1
improve the overall seismic performance of beam-to-column joint
(see Figs. 7(a), (d) and 12(a), (d)).
sub-assemblages [29].
The model used in the present study predicts the actual values
In the case of substandard existing complete two-dimension RC
of connection shear stress during an earthquake. Thus, it could also
frames, typical structural inadequacies such as the deficient
predict the shear stress of the adjacent columns of a beam-to-
anchorage of the beam longitudinal reinforcing bars in exterior
column joint sub-assemblage, given that the shear forces acting
beam-to-column joints might significantly deteriorate the hystere-
in the joints significantly exceed those acting in their adjacent col-
sis behaviour and cause substantial strength degradation, due to
umns [31]. The model can be used to predict the actual values of
excessive slipping and eventually pullout of the bars from the joint
shear forces and moments resisted by the beam-to-column joint
region. This extremely dangerous situation strongly affects the
sub-assemblages, provided that the longitudinal reinforcement of
structural integrity and may activate a soft-storey mechanism
the beam is well anchored in the joint region. Hence, it was only
leading to partial or even general collapse of an existing building.
used for the strengthened sub-assemblages SL1 and SK1, since the
In a typical retrofitting strategy usually adopted, the columns will
beam reinforcement anchorage in the cases of specimens K1 and
likely be strengthened whilst the joints left as they are. However,
this might lead to a similar global seismic response, since the exte-
rior joints remain vulnerable and damage in the joints (which are
part of the columns) is likely to occur [30]. Therefore, the proposed
retrofit scheme provides a reliable solution for the earthquake-
resistant rehabilitation of poorly detailed exterior beam-to-
column joints in order to achieve an indisputable superiority in
the overall hysteretic behaviour.

9. Theoretical considerations

To predict the actual values of connection shear stress of the

sub-assemblages with improved anchorage of the beam longitudi-
nal reinforcement in the joint, SL1 and SK1, a formulation published
in recent studies [7,14,21] was used. The validity of the formula-
tion was checked using test data from more than 140 exterior
and interior beam-to-column specimens tested in the Laboratory
of Reinforced Concrete Structures of the Aristotle University of Fig. 15. Maximum steel strain-versus-displacement.
14 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15

Table 6
Experimental and predicted values of the shear strength of specimens SL1 and SK1.

Specimen Joint aspect ratio a ¼ hb =hc ccal cexp cult Predicted shear Observed shear l ¼ sspred
strength spred a strength sexp b
exp

SL1 1.0 0.281 0.262 1.320 1.362 1.270 1.072

SK1 1.0 0.269 0.261 1.320 1.361 1.322 1.029
pffiffiffiffiffi pffiffiffiffiffi
a
For ccal P cult ; cpred ¼ cult and spred ¼ cult  f c (MPa). Forpcffiffiffiffi
calffi < cult ; cpred ¼ ccal and spred ¼ ccal  f c (MPa). An overstrength factor a0 ¼ 1:25 for the beam steel was
included in the pcomputations
ffiffiffiffiffi of joint shear stress scal ¼ ccal  f c (MPa).
b
sexp ¼ cexp  f c .

SM1 was inadequate. According to the model for ccal =cult < 0:5 the  Specimens SL1 and SK1 performed very satisfactorily during the
damage and the formation of the plastic hinges is shifted in the seismic loading sequence, showing substantial improvement of
beam, while the joint region remains intact. However, if the cyclic lateral response. The extension of the beam steel bars
1:0 > ccal =cult > 0:5 a mixed-type failure would occur with more to the backside of the joint, where anchored via steel plates and
excessive shear failure of the joint for values of ccal =cult near 1.0 bolts, prevented the excessive slipping and pullout of the longi-
and more excessive damage accumulation in the beam for values tudinal beam reinforcement from the joint. The RC jacketed col-
near 0.5. The value of ccal of sub-assemblages SL1 and SK1 was umns and the beam-to-column joints of the specimens were
lower than the corresponding value of cult , while also intact at the conclusion of the tests, while the formation of
ccal =cult < 0:5. Hence the specimens developed, as expected, a flex- the plastic hinges was in the beams at the column face.
ural hinge in the beam, while no damage was observed in the joints  Despite being strengthened by RC jacketing, specimen SM1
of the strengthened specimens. In Table 6 the comparison between exhibited an early pullout failure of the beam bars from the
experimental and predicted results by the shear model is summa- joint and performed similarly to the original specimen. Thus,
rized. A particularly close correlation can be observed. the retrofitting scheme implemented to SM1 failed to prevent
at least a partial or a probable general collapse of the strength-
10. Conclusions ened building. This is rather crucial for real structures, when a
false interpretation of the cracking formed in the beam near
An experimental program was conducted on six (three original the joint region is made.
and three strengthened) 1:2 scale exterior RC beam-to-column  Two different types of failure mode were developed by the
joint sub-assemblages to identify the effectiveness and feasibility strengthened beam-to-column joint sub-assemblages. The fail-
of a pre-earthquake and post-earthquake retrofitting technique. ure mode of the enhanced specimen SM1 was similar to that
Two retrofitting schemes were examined. The first scheme, of the original specimen K1 and indicated the poor lateral per-
(scheme 1), included the RC jacketing of the columns and the joint formance of the sub-assemblage. The second failure mode was
of the specimens and the improvement of the beam’s longitudinal developed by the specimens strengthened according to the pro-
reinforcement anchorage in the joint region, as well. The second posed retrofit scheme, SL1 and SK1 and included the shifting of
one (scheme 2) included only the RC jacketing of the columns damage and the formation of the plastic hinge in the beam. This
and the joint and represented the scheme that is applied when type of failure clearly demonstrates the effectiveness of the pro-
engineers make a false interpretation of the failure mode. This posed retrofit procedure and was also verified by the theoretical
occurs when the cracking in the beam near the joint region is con- model adopted.
sidered as a flexural failure, while in fact it was formed by the pull-  The improvement of deficient beam reinforcement anchorage in
out of the beam reinforcement from the joint. The original sub- exterior beam-to-column connections proved to be of high sig-
assemblages were representative of beam-to-column joints found nificance in order to achieve a desirable, ductile behaviour of
in pre1970 buildings, with poor detailing of reinforcement and pri- the beam-to-column joints, when strengthened via RC jackets.
marily with short straight anchorages of the beam longitudinal If the anchorage improvement is underestimated during the
reinforcement in the joint. One specimen was used as the control implementation of retrofitting schemes, the seismic lateral
specimen. The retrofitting scheme (1) was implemented for the response of the strengthened structures is jeopardized. Thus,
pre-earthquake and post-earthquake strengthening of two sub- even structures strengthened according to modern design codes
standard specimens. One more substandard beam-to-column joint may exhibit premature pullout failure of the beam reinforce-
sub-assemblage was pre-earthquake retrofitted using the retrofit- ment from exterior joints, resulting in a possible partial
ting scheme (2). The original and the strengthened sub- collapse.
assemblages were subjected to a large number of reversed lateral  The analytical model successfully predicted the actual values of
displacements to simulate the equivalent of strong earthquakes connection shear stress of the sub-assemblages with the
and the lateral performance of the structures was compared. An improved anchorage of the beam longitudinal reinforcement
analytical formulation published by Tsonos in recent studies in the joint, SL1 and SK1. The observed capacity was predicted
[7,14,21] was also used to predict the actual values of connection to within approximately 7 and 3 percent of that computed using
shear stress of the sub-assemblages with the improved anchorage the joint shear strength formulation for specimens SL1 and SK1
of the beam longitudinal reinforcement in the joint, SL1 and SK1. respectively.
The results of the experimental work are as follows:

 The original beam-to-column joint sub-assemblage, K1, showed References

poor seismic performance, characterized by rapid degradation
of the lateral strength, stiffness and energy dissipation capacity. [1] Kalogeropoulos IG, Tsonos A-DG. Effectiveness of r/c jacketing of substandard
The specimen exhibited premature failure due to the pullout of r/c columns with short lap splices. Struct Monitor Maintenance – Int J 2014;1
(3):273–92.
the beam longitudinal reinforcement from the joint, while the [2] Pampanin S, Calvi GM, Moratti M. Seismic behaviour of R.C. beam-column
load–deflection response of K1 was dominated by bond slip. joints designed for gravity loads. In: 12th ECEE, London, paper no. 726; 2002.
 G.I. Kalogeropoulos et al. / Engineering Structures 109 (2016) 1–15 15

[3] Engideniz M, Kahn LF, Zureick A-H. Repair and strengthening of non- [16] Ruiz-Pinilla JG, Pallarés FJ, Gimenez E, Calderón PA. Experimental tests on
seismically designed RC beam-to-column joints: state-of-the-art. Georgia retrofitted RC beam-column joints underdesigned to seismic loads. General
Institute of Technology, Research report No. 04-4; 2004. approach. Eng Struct 2014;59:702–14.
[4] Engindeniz M, Kahn FL, Zureick A-H. Repair and strengthening of reinforced [17] Filiatrault A, Pineau S, Houde J. Seismic behavior of steel–fiber reinforced
concrete beam–column joints: state of the art. ACI Struct J 2005;102(2):1–14. concrete interior beam-column joints. ACI Struct J 1995;92(5):543–52.
[5] Tsonos A-DG. Seismic repair of reinforced concrete beam-column sub- [18] Tsonos A-DG. An innovative solution for strengthening old R/C structures and
assemblages of modern structures by epoxy injection technique. Struct Eng for improving the FRP strengthening method. Struct Monit Maintenance – Int J
Mech 2002;14(5):543–63. 2014;1(3):323–38.
[6] Karayannis CG, Sirkelis George M. Strengthening and rehabilitation of beam- [19] Bedirhanoglu I, Ilki A, Pujol S, Kumbasar N. Behavior of deficient joints with
column joints using carbon-FRP jacketing and epoxy resin injection. plain bars and low-strength concrete. ACI Struct J 2010;107(3):300–10.
Earthquake Eng Struct Dynam 2008;37:769–90. [20] Hakuto S, Park R, Tanaka H. Seismic load tests on interior and exterior beam-
[7] Tsonos A-DG. Performance enhancement of R/C building columns and beam- column joints with substandard reinforcing details. ACI Struct J 2000;97(1):11–25.
column joints through shotcrete jacketing. Eng Struct 2010;32(3):726–40. [21] Tsonos A-DG. Cyclic load behavior of reinforced concrete beam-column sub-
[8] Karayannis GC, Chalioris CE, Sirkelis GM. Local retrofit of exterior beam- assemblages of modern structures. ACI Struct J 2007;104(4):468–78.
column joints using thin RC jackets – an experimental study. Earthquake Eng [22] EN 1998-1. Eurocode 2: design of concrete structures – Part 1-1: general rules
Struct Dynam 2008;37:727–46. and rules for buildings; 2004.
[9] Pampanin S, Bolognini D, Pavese A. Performance-based seismic retrofit [23] EN 1998-1. Eurocode 8: design of structures for earthquake resistance – Part 1:
strategy for existing reinforced concrete frame systems using fiber- General rules, seismic actions and rules for buildings; 2004.
reinforced polymer composites. Journal of Composites for Construction [24] CSRTC. New code of steel reinforcement technology for concrete, Athens;
2007:211–26. 2008.
[10] Mahmoud MH, Afefy HM, Kassem NM, Fawzy TM. Strengthening of defected [25] ACI-ASCE Committee 352-02. Recommendations for design of beam-column
beam-column joints using CFRP. J Adv Res 2014;5:67–77. connections in monolithic reinforced concrete structures (ACI-ASCE 352R-02).
[11] Tsonos A-DG. Contribution to the seismic behaviour’s improvement of exterior American Concrete Institute; 2002. p. 37.
B/C joints of reinforced concrete frame structures. PhD Thesis. Aristotle [26] Ehsani MR, Wight JK. Exterior reinforced concrete beam-to-column
University of Thessaloniki. Appendix 46, vol. 11; 1990. p. 405 (in Greek). connections subjected to earthquake – type loading. ACI J, Proc 1985;82
[12] Tsonos A-DG. Improvement of the earthquake resistance of R/C beam-column (4):492–9.
joints under the influence of P–D effect and axial force variations using [27] Ehsani MR, Whight JK. Effect of transverse and slab on behavior of reinforced
inclined bars. Struct Eng Mech – Int J 2004;18(4):389–410. concrete beam-to-column connections. ACI J Proc 1985;82(82–17):188–95.
[13] Tsonos A-DG. Lateral load response of strengthened reinforced concrete beam- [28] Durrani AJ, Whight JK. Earthquake resistance of reinforced concrete interior
to-column joints. ACI Struct J 1999;96(1):46–56. connections including a floor slab. ACI J Proc 1987;84(84-S42):400–6.
[14] Tsonos A-DG. Effectiveness of CFRP-jackets and RC-jackets in post-earthquake [29] Tsonos A-DG. Seismic repair of exterior r/c beam-to-column joints using two-
and pre-earthquake retrofitting of beam-column subassemblages. Eng Struct sided and three-sided jackets. Struct Eng Mech 2002;13(1):17–34.
2008;30:777–93. [30] Calvi GM, Magenes G, Pampanin S. Relevance of beam-column joint damage
[15] Chalioris CE, Favvata MJ, Karayannis CG. Reinforced concrete beam-column and collapse in RC frame assessment. J Earthquake Eng 2002;6(1):75–100.
joints with crossed inclined bars under cyclic deformations. Earthquake Eng [31] Paulay T, Priestley MJN. Seismic design of reinforced concrete and masonry
Struct Dynam 2008;37:881–97. buildings. John Wiley and Sons; 1992.