EARTHQUAKE ENGINEERING AND STRUCTURAL DYNAMICS

Earthquake Engng Struct. Dyn. 2008; 37:881–897

Published online 4 February 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/eqe.793

Reinforced concrete beam–column joints with crossed inclined

bars under cyclic deformations

Constantin E. Chalioris∗, †, ‡, § , Maria J. Favvata‡ and Chris G. Karayannis‡, ¶

Department of Civil Engineering, Democritus University of Thrace, Xanthi 67100, Greece

SUMMARY
This experimental study investigates the effectiveness of crossed inclined bars (X-bars) as joint shear
reinforcement in exterior reinforced concrete beam–column connections under cyclic deformations. Test
results of 20 joint subassemblages with various reinforcement ratios and arrangements including X-bars
in the joint area are presented. The X-type, non-conventional reinforcement is examined as the only
joint reinforcement and in combination with common stirrups or vertical bars. The experimental results
reported herein include full loading cycle curves, energy dissipation values and a categorization of the
observed damage modes. Based on the comparisons between the overall hysteretic responses of the tested
specimens, it is deduced that joints with X-bars exhibited enhanced cyclic performance and improved
damage mode since a distinct flexural hinge was developed in the beam–joint interface. Further, the
combination of crossed inclined bars and stirrups in joint area resulted in enhanced hysteretic response
and excellent performance capabilities of the specimens. However, in some specimens with X-bars as
the only joint shear reinforcement, the deformations of the bent anchorage of the beam’s bars caused
considerable damages at the back of the joint area. Discussion for a potential replacement of the joint
stirrups with X-type reinforcement in some cases of exterior joints is also included. Copyright q 2008
John Wiley & Sons, Ltd.

Received 4 October 2007; Revised 5 December 2007; Accepted 16 December 2007

KEY WORDS: beam–column joints; cyclic test; inclined bars; reinforced concrete

1. INTRODUCTION

The behaviour of beam–column joints has long been recognized as a significant factor that
frequently becomes critical for the overall behaviour of reinforced concrete (RC) framed structures

∗ Correspondence to: Constantin E. Chalioris, Department of Civil Engineering, Democritus University of Thrace,
12 V. Sofias Street, Xanthi 67100, Greece.
†
E-mail: chaliori@civil.duth.gr
‡
Civil Engineer.
§ Lecturer.
¶ Professor.

Copyright q 2008 John Wiley & Sons, Ltd.

882 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

subjected to seismic excitations. Inadequate shear reinforcement in joint regions of existing RC

structures proved to be the cause of brittle failure and catastrophic collapse during earthquake
events [1, 2]. Experimental research is the main tool for the investigation of the parameters that
influence and improve the joint performance [3, 4]. It is also noted that detailed design recommen-
dations for RC beam–column joints were first published in the last two or three decades in Europe
and the U.S.A.
Attempts for the improvement of the seismic properties of these members are usually focused
on the use of non-conventional reinforcement, such as steel fibres [5], composite materials (FRPs)
[6–8], spiral reinforcement [9] and inclined bars. The first published work for the use of crossed
inclined bars (X-bars) in RC joints was in 1984 [10]. Further experimental studies indicated that
exterior beam–column joints with common stirrups and inclined bars as supplementary shear
reinforcement in joint area exhibited significantly improved behaviour with respect to joints with
only conventional reinforcement, such as stirrups and vertical bars [11, 12]. The advantages of
the crossed inclined bars as a feasible solution for joint reinforcement have also been emerged
in the theoretical study of Bakir [13], although in this study the increase in the joint shear
strength due to X-bars proved to be dependent on the joint aspect ratio due to geometrical
constrains.
First experimental attempts to use crossed inclined bars as the only shear reinforcement in the
joint area have been reported by the authors in 1998 [3] and 2007 [14]. The experimental study
described in this paper builds on these previous works and extends them with new test results in
order to extensively investigate the effectiveness of crossed inclined bars in exterior beam–column
connections subjected to cyclic loading. The possibility of full replacement of the joint stirrups
with X-type reinforcement in joint body is also addressed in this study.
It is obvious that the conducted research in this field of study is very limited and the influ-
ence of the crossed inclined bars on the hysteretic response of exterior beam–column joints has
not yet been quantified and incorporated into code recommendations. Thus, the great impor-
tance of beam–column joints on the overall seismic response of RC framed structures, the indu-
bitable need to prevent brittle failure in joint regions, the growing interest for the use of non-
conventional reinforcement and the lack of relative studies are the main motives behind this
effort.
Twenty RC beam–column joint specimens with various reinforcement arrangements that include
X-bars in the joint area were tested under cyclic loading. The main variables examined herein
are the type of the joint shear reinforcement (stirrups, X-bars, vertical bars and combination
of them), the joint steel reinforcement ratio and the joint shear stress. Comparisons between
the measured hysteretic responses in terms of full loading cycle curves and absorbed energy
values along with the observed damage modes of the tested specimens are also presented and
discussed.

2. EXPERIMENTAL PROGRAM

2.1. Specimen characteristics

The program was designed to investigate the efficiency of steel crossed inclined bars as shear
reinforcement in the joint area. It includes 20 exterior beam–column joint subassemblages sorted
into four groups (A, B, Ca and Cb). Groups A and B comprise four two-third-scale specimens

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 883

each, whereas Ca and Cb groups include six one-third-scale specimens each. Each group includes
one control specimen without stirrups or X-bars in the joint body (JA-0, JB-0, JCa-0 & JCb-0).
The geometry of all the specimens of A and B groups is the same; total columns’ height and
cross-sectional dimensions are 1800 mm and 300×200 mm, respectively, whereas the beam length
and cross-sectional dimensions are 1100 mm and 200×300 mm, respectively (Figure 1(a) and (b)).
The geometry of the specimens of Ca and Cb groups is also the same; total columns’ height and
cross-sectional dimensions are 1900 mm and 200×100 mm, respectively, whereas the beam length
and cross-sectional dimensions are 1100 mm and 100×200 mm, respectively (Figure 1(c)). Most
of the specimens of Ca and Cb groups came from a previous experimental study carried out by
the authors [3].
The joint shear reinforcement of 10 specimens: two specimens in group A (JA-X12 and JA-X14),
two in group B (JB-X10 and JB-X12), three in group Ca (JCa-X10, JCa-s1-X10 and JCa-s2-X10)
and three in group Cb (JCb-X10, JCb-s1-X10 and JCb-s2-X10) include two pairs of inclined bars
which form the X-type reinforcement, as shown in Figure 1. In six of them, the X-bars are the
only shear reinforcement in the joint area (JA-X12, JA-X14, JB-X10, JB-X12, JCa-X10 and JCb-
X10). Further, there are six specimens with closed stirrups as joint reinforcement (JA-s5, JB-s1,
JCa-s1, JCa-s2, JCb-s1 and JCb-s2) and four of them (JCa-s1, JCa-s2, JCb-s1 and JCb-s2) are
used as reference specimens to those which have stirrups and X-bars in the joint area (JCa-s1-X10,
JCa-s2-X10, JCb-s1-X10 and JCb-s2-X10, respectively).
The beam and the columns of the specimens of group A were designed according to the
specifications of ACI 318-02 [15] and ACI-ASCE 352-02 [16] for Type 2 exterior connections.
Further, the joint area of the specimen JA-s5 of group A was also designed according to the ACI
Design Codes. Therefore, five deformed steel stirrups with diameter 8 mm (5⭋8 or ⭋8/50 mm)
were provided as the required amount of joint transverse reinforcement.
Reinforcement arrangements, deformed steel strengths and geometrical details of all tested
specimens are summarized in Table I and shown in Figure 1, respectively. It is noted that the
code name of each specimen in Table I is characterized by the group (A, B, Ca and Cb) and the
reinforcement in the joint area as follows: 0 denotes the control specimens (without stirrups or
inclined bars); X10, X12 and X14 indicate specimens with X-bars with diameter ⭋10, ⭋12 and
⭋14, respectively; s1, s2 and s5 specify the numbers of the stirrups added in joint body (diameter
of used stirrups is equal to 8 mm or ⭋8).

(a) (b) (c)

Figure 1. Geometry and reinforcement characteristics of the beam–column joint specimens.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
884 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

Table I. Characteristics of the beam–column joint specimens.

Specimen Column reinforcement Beam reinforcement Joint area reinforcement

code name Bars Stirrups Bars Stirrups Stirrups Vertical bars X-bars
Group A
JA-0 4⭋14 ⭋8/50 4⭋12 (top)+ ⭋8/65-130 — 2⭋10 —
+2⭋10 (middle) 4⭋12 (bottom)
JA-s5 4⭋14 ⭋8/50 4⭋12 (top)+ ⭋8/65-130 5⭋8 (⭋8/50) 2⭋10 —
+2⭋10 (middle) 4⭋12 (bottom)
JA-X12 4⭋14 (+4⭋12∗ ) ⭋8/50 4⭋12 (top)+ ⭋8/65-130 — 2⭋10 2X⭋12
+2⭋10 (middle) 4⭋12 (bottom)
JA-X14 4⭋14 (+4⭋14∗ ) ⭋8/50 4⭋12 (top)+ ⭋8/65-130 — 2⭋10 2X⭋14
+2⭋10 (middle) 4⭋12 (bottom)

Group B
JB-0 4⭋10 ⭋8/150 6⭋10 (top) + ⭋8/150 — — —
6⭋10 (bottom)
JB-s1 4⭋10 ⭋8/150 6⭋10 (top) + ⭋8/150 1⭋8 (⭋8/150) — —
6⭋10 (bottom)
JB-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 6⭋10 (top) + ⭋8/150 — — 2X⭋10
6⭋10 (bottom)
JB-X12 4⭋10 (+4⭋12∗ ) ⭋8/150 6⭋10 (top) + ⭋8/150 — — 2X⭋12
6⭋10 (bottom)

Group Ca
JCa-0 4⭋10 ⭋8/150 2⭋10 (top)+ ⭋8/150 — — —
2⭋10 (bottom)
JCa-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 2⭋10 (top) + ⭋8/150 — — 2X⭋10
2⭋10 (bottom)
JCa-s1 4⭋10 ⭋8/150 2⭋10 (top) + ⭋8/150 1⭋8 (⭋8/100) — —
2⭋10 (bottom)
JCa-s1-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 2⭋10 (top) + ⭋8/150 1⭋8 (⭋8/100) — 2X⭋10
2⭋10 (bottom)
JCa-s2 4⭋10 ⭋8/150 2⭋10 (top) + ⭋8/150 2⭋8 (⭋8/67) — —
2⭋10 (bottom)
JCa-s2-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 2⭋10 (top) + ⭋8/150 2⭋8 (⭋8/67) — 2X⭋10
2⭋10 (bottom)

Group Cb
JCb-0 4⭋10 ⭋8/150 3⭋10 (top) + ⭋8/150 — — —
3⭋10 (bottom)
JCb-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 3⭋10 (top) + ⭋8/150 — — 2X⭋10
3⭋10 (bottom)
JCb-s1 4⭋10 ⭋8/150 3⭋10 (top) + ⭋8/150 1⭋8 (⭋8/100) — —
3⭋10 (bottom)
JCb-s1-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 3⭋10 (top) + ⭋8/150 1⭋8 (⭋8/100) — 2X⭋10
3⭋10 (bottom)
JCb-s2 4⭋10 ⭋8/150 3⭋10 (top) + ⭋8/150 2⭋8 (⭋8/67) — —
3⭋10 (bottom)
JCb-s2-X10 4⭋10 (+4⭋10∗ ) ⭋8/150 3⭋10 (top) + ⭋8/150 2⭋8 (⭋8/67) — 2X⭋10
3⭋10 (bottom)
∗ X-bars in joint area extend to the full column height.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 885

The steel reinforcement ratios of the column, the beam and the joint body of the tested specimens
are listed in Table II. The same table also includes the values of the column to beam flexural yield
y y
strength ratio
= MR,c /MR,b for each specimen. Three different expressions for the reinforcement
ratios have been used:
• Volumetric ratio:
Vsteel
vol.
s = (1)
Vconcrete

• ACI 318 [15] ratio:

Ash
ACI = (only for stirrups) (2)
s
sh bc

Table II. Reinforcement ratios of the tested specimens.

Column reinforcement Beam reinforcement Joint area shear
ratios (%) ratios (%) reinforcement ratios (%)
Specimen y
MR,c
code name Longitudinal∗ Transverse Longitudinal Transverse vol.
s ACI
s s
∗ = y
MR,b

Group A
JA-0 1.29 (1.29) 1.01 1.51 0.77 0.26 — 0.26 0.88 (0.88)
JA-s5 1.29 (1.29) 1.01 1.51 0.77 1.33 1.26 1.10 0.88 (0.88)
JA-X12 2.04 (1.29) 1.01 1.51 0.77 1.13 — 0.64 1.36 (0.88)
JA-X14 2.31 (1.29) 1.01 1.51 0.77 1.44 — 0.77 1.50 (0.88)

Group B
JB-0 0.52 (0.52) 0.34 1.57 0.34 — — — 0.52 (0.52)
JB-s1 0.52 (0.52) 0.34 1.57 0.34 0.25 0.36 0.17 0.52 (0.52)
JB-X10 1.05 (0.52) 0.34 1.57 0.34 0.65 — 0.26 0.87 (0.52)
JB-X12 1.28 (0.52) 0.34 1.57 0.34 0.94 — 0.38 1.01 (0.52)

Group Ca
JCa-0 1.57 (1.57) 0.67 1.57 0.67 — — — 1.21 (1.21)
JCa-X10 3.14 (1.57) 0.67 1.57 0.67 1.84 — 0.79 2.18 (1.21)
JCa-s1 1.57 (1.57) 0.67 1.57 0.67 0.64 1.17 0.50 1.21 (1.21)
JCa-s1-X10 3.14 (1.57) 0.67 1.57 0.67 2.48 1.17 1.29 2.18 (1.21)
JCa-s2 1.57 (1.57) 0.67 1.57 0.67 1.29 1.75 1.00 1.21 (1.21)
JCa-s2-X10 3.14 (1.57) 0.67 1.57 0.67 3.12 1.75 1.79 2.18 (1.21)

Group Cb
JCb-0 1.57 (1.57) 0.67 2.36 0.67 — — — 0.83 (0.83)
JCb-X10 3.14 (1.57) 0.67 2.36 0.67 1.84 — 0.79 1.50 (0.83)
JCb-s1 1.57 (1.57) 0.67 2.36 0.67 0.64 1.17 0.50 0.83 (0.83)
JCb-s1-X10 3.14 (1.57) 0.67 2.36 0.67 2.48 1.17 1.29 1.50 (0.83)
JCb-s2 1.57 (1.57) 0.67 2.36 0.67 1.29 1.75 1.00 0.83 (0.83)
JCb-s2-X10 3.14 (1.57) 0.67 2.36 0.67 3.12 1.75 1.79 1.50 (0.83)
∗ Valuesof longitudinal reinforcement ratios and
inside the parentheses have been calculated not taking into
account the X-bars in the column cross-section.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
886 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

• Effective ratio:

 stirrups

As , AsX-bars , Avert.
s
bars
s = (3)
bc h c

where Vsteel and Vconcrete are the volumes of total steel shear reinforcement and concrete,
stirrups
respectively, in the joint area; As , AsX-bars and Avert.bars
s are the areas of the shear rein-
forcement due to stirrups, X-bars and vertical bars, respectively, in the joint area; Ash is the
area of the transverse reinforcement [16]; sh is the centre-to-centre spacing of stirrups; bc and
h c are the column cross-sectional dimensions; and bc is the core dimension of tied column,
outside-to-outside edge of transverse reinforcement bars, perpendicular to the transverse
reinforcement area Ash .
The values of the concrete mean cylinder compressive strength at the age of 28 days, f c , the
ultimate joint shear stress, u , and the level of the horizontal shear stress in terms of the parameter
 as defined by ACI-ASCE 352-02 [16] for each group of specimens are presented in Table III.
The nominal joint shear stress is calculated based on the relationship:

Vjh
= (4)
bc h c

where Vjh is the joint horizontal shear force that can be expressed as Vjh = Tb − Vcol ; Tb is the tensile
force of the beam reinforcement that can be evaluated by Tb = Abs1 f s ; Abs1 and f s are the area and
the developed stress of the tensile longitudinal reinforcement of the beam, respectively; Vcol is the
horizontal shear force in the column that can be expressed as Vcol = (Vc,up + Vc,lo )/2 = Vc = Mb /
c ;
Vc,up and Vc,lo are the horizontal shear forces in the upper and lower columns, respectively, Mb is
the beam bending moment;
c is the mean value of the heights of the upper and lower columns:

c = (
c,up +
c,lo )/2.
Therefore, the ultimate value of joint shear stress, reported in Table III, is deduced when f s
becomes maximum (ultimate longitudinal steel strength f su ) and Mb reaches the value of the beam
y
flexural yield strength MR,b as follows:

Abs1 f su − MR,b /
c

y
u = (5)
bc h c

Table III. Horizontal joint shear stress.

Concrete compressive Ultimate shear stress Level of horizontal shear stress∗
Group of specimens strength (MPa) (u , MPa) ()
A 34.0 4.1 8.5
B 31.6 4.3 9.2
Ca 20.6 3.6 9.6
Cb 23.0 5.4 16.1
∗ The maximum allowable joint shear stress by ACI-ASCE 352-02 [16] in terms of the parameter  is equal to
12 (for connection Type 2).

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 887

It is also mentioned that the level of horizontal joint shear stress of the tested specimens, as
reported in Table III, has been determined by the parameter  [16]:
u =1 u
=  −→  =  (6)
·0.083 f c 0.083 f c
The values of  in Table III show that only the horizontal joint shear stress of group Cb specimens is
higher than 12. Value 12 expresses the maximum allowable joint shear stress defined by ACI-ASCE
352-02 [16] for Type 2 exterior connections.

2.2. Test setup and loading sequences

Test setup and instrumentation details are shown in Figure 2(a). Supports that allow rotation were
used to simulate the inflection points assumed to occur at a point of the columns in a laterally
loaded frame structure. All specimens were subjected to full cycle deformations imposed near the
free end of the beam by a pinned-end actuator.

2.2.1. Column axial load. Column axial load equal to Nc = A g f c , where  = 0.05 for the speci-
mens of groups A and B, and  = 0.10 for the specimens of groups Ca and Cb, was applied during
the tests. The application of this axial load was controlled in order to maintain a constant value
during the entire testing procedure. The influence of the column axial load on the shear capacity of
the joints is considered to be favourable since the developed principal tensile stresses in the joint
area are reduced due to the application of compressive axial load, whereas they reach maximum
values when the column axial load is null or neglected. However, any possible beneficial effect of
the column axial compressive load on the joint capacity is not clearly represented in this study,
since the horizontal joint shear strength governs the joint shear capacity.

2.2.2. Loading histories. Tested specimens of group A suffered a loading history of five full
loading steps with maximum displacements ±6, ±20, ±40, ±60 and ±80 mm, while each loading

(b)

(a) (c)

Figure 2. Test setup and loading histories.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
888 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

step includes two full loading cycles (Figure 2(b)). Specimens of group B were also tested using
similar loading history with four loading steps. The joint specimens of groups Ca and Cb were
subjected to increasing full multi-cyclic loading sequence that included a first cycle with maximum
deformation equal to approximately 1.3 times the yielding deformation and later to constantly
increasing deformation with increment step equal to 2.5 mm per cycle (Figure 2(c)). Specimens
suffered this loading sequence until the measured maximum cycle load decreased to approximately
40% of the yield load level as measured in the first cycle or when the maximum deformation
reached the value of 100 mm.
The choice of the testing program and loading history should be guided by the objective to
maximize information and minimize complexities that will complicate test evaluation. Basic seismic
capacity parameters for a structural element are strength, stiffness, inelastic deformation capacity
(ductility) and cumulative damage capacity parameters such as energy dissipation capacity. All
these parameters are expected to deteriorate as the number of damaging cycles and the amplitude
of cycling increase. In order to draw conclusions for the ultimate limit state and to evaluate
performance under arbitrary seismic excitations, a multi-cyclic loading sequence with constantly
increasing displacement and large inelastic excursions has been chosen for specimens of groups
Ca and Cb (Figure 2(c)). It is also emphasized that there is significant dependence of the demand
parameters on the natural period of the structure of which the element is a member. For generic test
specimens, the need exists to base these parameters on short period structures with the understanding
that their values may be very conservative for long period structures. Further, the cumulative
damage resulting from numerous inelastic cycles has also been proposed to be taken into account
[17]. Since small amplitude cycle loadings are frequent in the response of natural excitations, the
adopted loading program of specimens of groups A and B includes two full loading cycles per
step with the same deformations (Figure 2(b)). It is a fact that the adopted loading sequence leads
to large drift values. This can be considered as an evaluation factor for the behaviour beyond the
common seismic design targets. Nevertheless, an enhanced performance of beam–column joints
has to include a reliable post-reparable branch.
The imposed load was measured by a load cell with accuracy equal to 0.025 kN and the
displacements of the beam at its end were measured by Linear Variable Differential Transformer
with accuracy equal to 0.01 mm. Linear Variable Differential Transformers were also placed at
each end of the column part of the specimens, as shown in Figure 2, in order to check the supports
during the test.

3. TEST RESULTS AND DISCUSSION

3.1. Hysteretic responses

The hysteretic responses in terms of full loading cycle curves (load versus deformation diagrams)
of the beam–column subassemblages of groups A and B are presented and compared with the
performance of the corresponding control specimens in Figure 3. Similarly, for the subassemblages
of groups Ca and Cb, in Figure 4 the hysteretic responses of specimens with X-bars in the joint
area are compared with their corresponding control and reference specimens.
Further, the dissipated hysteretic energy measured in terms of the area of the full load-
deformation envelopes of the specimens of groups A and B is demonstrated in Figure 5. Energy dis-
sipation is an indication of the specimen capacity and defines the energy that could be dissipated

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 889

Figure 3. Hysteretic responses of specimens of groups A and B.

Figure 4. Hysteretic responses of specimens of groups Ca and Cb.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
890 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

(a)

(b)

Figure 5. Energy dissipation: (a) specimens of group A and (b) specimens of group B.

before the loss of system stability. Furthermore, the ultimate load capacities per loading cycle
of the beam–column specimens of groups Ca and Cb that reflect the strength degradation of the
examined joints are displayed for comparison reasons in Figure 6. From Figure 6 it is deduced that
the flexural strength ratio of column to beam of the specimens significantly affects the observed
response, as it was expected. Joint specimens of group Ca with
= 1.21 exhibited improved perfor-
mance and higher load capacities with respect to the corresponding specimens of group Cb with

= 0.83.
The nominal principal tensile stress (diagonal tensile stress) developed in the beam–column
joint region is calculated by the expression:

p 2p
t = + +2 (7)
2 4

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 891

15

JCa-s2 Group Ca joints

JCa-s2-X10 JCa-0
10
JCa-s1
JCa-s1-X10
JCa-s2

5 JCa-X10
JCa-0 JCa-X10
JCa-s1 JCa-s1-X10
Load (kN)

Loading cycle JCa-s2-X10

0
5 10 15 20 25 30 35
JCa-0 JCa-s2-X10
JCa-s1 JCa-X10
-5

JCa-s2 JCa-s1-X10
-10

(a) -15

20
Group Ca
Cb joints
joints
15 JCb-0
JCb-s1
10 JCb-s2
JCb-s2-X10
JCb-X10
5 JCb-X10
JCb-s1-X10 JCb-s1-X10
Load (kN)

JCb-0 JCb-s1 JCb-s2-X10

JCb-s2 Loading cycle
0
5 10 15 20 25 30 35
JCb-0 JCb-s1 JCb-s2
JCb-s2-X10
-5
JCb-X10
JCb-s1-X10
-10

-1 5

(b) -20

Figure 6. Ultimate load capacities per loading cycle: (a) specimens of group
Ca and (b) specimens of group Cb.

where  p is the axial compressive stress in the joint area that can be expressed as  p =
(Nc + P)/(bc h c );  is the nominal joint shear stress; Nc is the column axial compressive load; and
P is the imposed cyclic load at the end of the beam.
Figure 7(a) displays the values of t for the specimens JA-0, JA-s5 and JA-X14 of group A
versus the imposed deformations for the first and second loading cycles. It is mentioned that, based
on the test results, joint cracking
 can be expected when the joint nominal principal tensile stress
exceeds the value of 0.29 f c [18]. This value is pointed out in the graphs of Figure 7(a). Further,
Figure 7(b) builds on the results of Figure 7(a) and represents the tensile stresses of the X-bars tx
for the specimen JA-X14, as the differences in the t values of the joints JA-X14 and JA-0. From

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
892 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

Figure 7. Principal tensile stresses: (a) nominal principal tensile stresses of specimens of group A and (b)
tensile stresses of X-bars of specimen JA-X14.

Figure 7(b) it can be deduced that the contribution of the crossed inclined bars is significantly
inferior to their yield capacity.
The comparison of the seismic performance between the specimens of groups A and B
(Figures 3, 5 and 7) indicated that the joint specimens reinforced with X-bars (without stirrups)
exhibited significantly improved behaviour with respect to the corresponding control specimens
JA-0 and JB-0 (without stirrups). Load capacity and absorbed energy values of all the X-type
reinforced joints were higher than those of the control specimens in every loading step of the load
history. This was expected, since control specimen JB-0 had no shear reinforcement in joint area
and JA-0 had only two vertical bars, which proved to be quite inadequate. Comparing the hysteretic
responses of specimens JB-X12 with two pairs of crossed inclined bars ⭋12 as the only joint shear

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 893

reinforcement and JB-s1 with one stirrup ⭋8 in joint area (Figures 3 and 5(b)), it can be deduced
that the first one demonstrated higher strength and energy dissipation capabilities. However, this
fact is also related to the higher reinforcement ratio of the specimen JB-X12 with X-bars which
equals s = 0.38%, in comparison with the reinforcement ratio s = 0.17% of the conventionally
reinforced joint with one stirrup (JB-s1). Nevertheless, the experimental results of group A show
a considerable advantage to the X-type reinforced joints since specimen JA-X14 (with 2 X-bars
⭋14 in joint area) with s = 0.77% demonstrated, more or less, similar cyclic behaviour with the
specimen JA-s5 (with 5⭋8 stirrups in joint area) with s = 1.10%, which meets the requirements
of ACI codes [15, 16] for Type 2 exterior connections (see also Figures 3 and 5(a)). This very
alike behaviour of specimens JA-X14 and JA-s5 is also revealed from the comparison of graphs
in Figure 7(a) that display the relationships between the nominal principal tensile stress and the
imposed deformations of the examined joints.
Regarding the effectiveness of the inclination, , of the X-bars (see also Figure 1 for notation)
in resisting the principal tensile stress in the joint, the analytical study of Bakir [13] has shown that
crossed inclined bars are effective only in the horizontal direction. Therefore, for high -values the
effectiveness of X-bars in increasing the horizontal joint shear stress is reduced. Since the increase
in the joint shear strength is offset by the increase in the principal tensile strain [13], it can be
deduced that higher values of inclination  result in increasing the efficiency of the crossed bars in
resisting the principal tensile stress in the joint. This can be observed in the experimental results of
specimens JB-X12 with  = 50◦ and max t = 3.2 MPa and JA-X12 with  = 58◦ and max t = 3.3 MPa.
However, the inclination  is more or less the same in the tested specimens since the ratio of h b /h c
is equal to 1 for all joints; therefore, safe remarks cannot be concluded.
Concerning the performance of specimens with X-bars and stirrups of groups Ca and Cb
(Figures 4 and 6), the observed hysteretic responses and load capacities were considerably improved
in comparison with those of the corresponding reference specimens with stirrups only. It is also
noted that test results of specimens JCa-s1-X10 and JCa-s2-X10 indicated that the combination
of crossed inclined bars and stirrups is essential for the joint safety since these joint specimens
showed excellent performance and high strength capacities due to the presence of stirrups along
with the X-bars. The minimum stirrup reinforcement that would be required when crossed bars
are used is rather difficult to be rationally estimated considering the small amount of the available
experimental works in the literature. However, based on the test results of the present work and
despite the limited test data, it could be estimated that the presence of one stirrup ACI
s = 1.17% has
been proved adequate for joints with u = 3.6 MPa and
= 1.21 (joints of group Ca), whereas joints
with increased u = 6.4 MPa and lower
= 0.83 seem to need more stirrups (two): ACI s = 1.75%
(joints of group Cb).
Specimens with X-bars only, JCa-X10 and JCb-X10, exhibited significantly enhanced cyclic
behaviour compared with the control specimens without shear reinforcement, JCa-0 and JCb-0,
respectively, as it was expected. It is emphasized that specimen JCb-X10 (with 2 X-bars ⭋10
only in the joint area and s = 0.79%) displayed higher load capacities throughout the entire cyclic
loading sequence than specimen JCb-s2 (with 2⭋8 stirrups in joint area and s = 1.00%) (see
also Figure 6). This fact, in combination with the results of specimens of group A (joints JA-X14
and JA-s5), gives the possibility of total or partial replacement of the joint stirrups with X-type
reinforcement in some cases.
The effectiveness of the inclined crossed bars as joint shear reinforcement is influenced by
various complex interacting phenomena such as the joint shear stress, the beam’s bars anchorage,
the reinforcement ratios of the column and the beam, the column-to-beam flexural strength ratio,

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
894 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

the presence of stirrups, etc. Thus, in order to develop a safe and accurate rule to determine the
amount of the required inclined crossed reinforcement for an enhanced joint response, more test
results would be needed. However, based on the experimental results of the present work, it could
be estimated that the addition of X-bars in an effective ratio s
0.38% leads to a satisfactory
improvement in the cyclic response for joints with u = 4.2 MPa and
= 0.88.

3.2. Damage modes and cracking patterns

An essential parameter that evaluates the efficiency of the examined X-type reinforcement in the
joint area is its ability to shift the failure from the joint area to the beam. For this purpose,
a categorization of the observed damage modes of the examined specimens is included herein.
Cracking patterns of the specimens at the end of their loading history are illustrated in the photos
of Figure 8 (groups A and B) and Figure 9 (groups Ca and Cb). Based on these photographs,
which illustrate the final cracking modes and the cracking propagation during the cyclic loading
procedure of each tested specimen, the following damage modes can be reported:
(a) Damage mode ‘A’: Joint area failure. Diagonal crossed cracks are concentrated at the region
of the joint from the early loading steps. Subsequent loading cycles result in a gradual increase in
the width of the cracks across the joint region. Additionally, the deformations of the bent anchorage
of the beam’s bars (anchorage failure) in combination with the absence of stirrups in the joint area
cause significant damage to the concrete cover at the back side of the joint area. Consequently,
joint body is seriously damaged and concrete of this area eventually becomes totally fragmented.
Failure of all control specimens JA-0, JB-0, JCa-0 and JCb-0 without stirrups and without X-bars in
joint area and specimens JB-s1, JCb-s1 and JCb-s2 with inadequate shear reinforcement (stirrups)
in the joint area was characterized by this mode.
(b) Damage mode ‘B’: Mixed failure mode; damages in joint and the beam. Cracks are initially
formed both at the joint region and at the end of the beam conjuncted to the joint. In subsequent
loading cycles, cracks and concrete spalling are located at the joint body and a distinct flexural

Figure 8. Damage and cracking mode of specimens of groups A and B.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 895

Figure 9. Damage and cracking mode of specimens of groups Ca and Cb.

hinge is also developed in the beam–joint interface. Especially in specimens without stirrups,
the deformations of the bent anchorage of the beam’s bars cause serious spalling to the concrete
cover at the back of the joint area. The specimens JCa-s1 and JCa-s2 with stirrups in the joint
area along with most of the specimens with X-bars, JA-X12, JA-X14, JB-X10, JB-X12, JCa-X10,
JCa-s1-X10, JCb-X10, JCb-s1-X10 and JCb-s2-X10, displayed this damage mode.
(c) Damage mode ‘C’: Plastic hinge in the beam while the joint area remains intact. The main
cracking is localized at the end of the beam in beam–joint conjunction area where severe vertical
cracks are formed from the early loading deformations. Subsequent loading cycles result in a
gradual increase in the width of one or two vertical cracks at the beam region where eventually
a distinct flexural hinge is developed. The joint body remains practically undamaged. Specimens
JA-s5 and JCa-s2-X10 exhibited this damage mode.
Based on the previously mentioned observations, it can be summarized that
• From the 10 specimens of the test program that have X-bars in the joint area as shear reinforce-
ment, nine of them exhibited improved damage behaviour with respect to the corresponding
control and reference specimens. Only specimen JCa-s1-X10 demonstrated the same damage
mode ‘B’ with its reference specimen JCa-s1.
• Crossed inclined bars shift the damage outside the joint area and demonstrated considerably
improved behaviour and damage mode characteristics in comparison with the corresponding
control specimens, since distinct plastic hinges were formed in the beams in all the examined
cases. None of the specimens with X-type reinforcement exhibited damage mode ‘ A’ (joint
area failure).
• The combination of X-bars and stirrups proved to be essential for the overall capacity of the
joints. Stirrups not only increased the joint shear capacity but also restrained the deformations
of the bent anchorage of the beam’s bars that caused severe damage at the back of the joint
area in the other specimens without stirrups. Specimen JCa-s2-X10 with 2 X-bars ⭋10 and
2⭋8 stirrups demonstrated excellent damage response since joint area remained intact and
plastic hinge was formed in the beam (damage mode ‘C’). On the other hand, in specimens

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
896 C. E. CHALIORIS, M. J. FAVVATA AND C. G. KARAYANNIS

JA-X14 and JB-X12 with X-bars only, although a distinct plastic hinge was formed in the
beam, severe damage was also observed at the back of the joint area due to the deformations
of the bent anchorage of the beam’s bars and the absence of stirrups (damage mode ‘B’).

4. CONCLUDING REMARKS

Cyclic test results of 20 RC exterior beam–column joint subassemblages were presented herein in
order to investigate the effectiveness of the use of crossed inclined bars as shear reinforcement in
the joint area. Based on the hysteretic responses and the cracking modes of the tested specimens,
it can be deduced that the overall seismic performance of the joints with X-type reinforcement
was considerably improved with respect to the cyclic behaviour of the control specimens (without
stirrups) and the reference specimens (with stirrups).
Specimens with X-bars as the only joint shear reinforcement exhibited high values of load
capacity in most of the loading cycles and increased hysteretic energy dissipation practically in
the entire loading sequence. All these specimens also demonstrated improved cracking mode since
a distinct flexural hinge was developed in the beam–joint interface, whereas all corresponding
control specimens failed by brittle joint failure. However, the deformations of the bent anchorage
of the beam’s bars caused considerable damages at the back side of the joint area and substantial
reductions in the potential capability in joint shear strength were reported. Nevertheless, the
hysteretic response of specimens JA-X14 and JCb-X10 (X-reinforcing joints without stirrups)
proved to be very alike or even better in comparison with the behaviour of the specimens JA-s5
and JCb-s2, respectively, which had stirrups with higher reinforcement ratios. These results raise
the questionable possibility of full replacement of the joint stirrups with X-bars in cases where it
is more convenient.
Joints of Ca and Cb groups are one-third-scale specimens. Therefore, only provisional conclu-
sions can be made from these tests and further research is needed in this field of study. There
is also need for more tests at a larger scale in order to further enlighten the effectiveness of the
crossed inclined bars as joint shear reinforcement and to constitute sound recommendations for
inclusion in the codes.
Finally, it is reported that specimens with crossed inclined bars and stirrups showed enhanced
hysteretic response, excellent performance capabilities and the cracking was mainly localized
in the beam–joint interface creating a distinct flexural hinge. Especially, specimen JCa-s2-X10
demonstrated the optimum damage mode ‘C’ in which the joint body remained undamaged.
Experimental results revealed that the combination of X-bars and stirrups is essential for the
safety of the joint since stirrups not only increased the joint shear capacity but also restrained the
deformations of the bent anchorage of the beam’s bars that damaged concrete cover at the back
of the joint area in the specimens without stirrups.

REFERENCES
1. Pagni CA, Lowes LN. Fragility functions for older reinforced concrete beam–column joints. Earthquake Spectra
2006; 22(1):215–238.
2. Pantelides CP, Clyde C, Reaveley LD. Performance-based evaluation of reinforced concrete building exterior
joints for seismic excitation. Earthquake Spectra 2002; 18(3):449–480.
3. Karayannis CG, Chalioris CE, Sideris KK. Effectiveness of RC beam–column connection repair using epoxy
resin injections. Journal of Earthquake Engineering 1998; 2(2):217–240.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe
 REINFORCED CONCRETE BEAM–COLUMN JOINTS 897

4. Karayannis CG, Sirkelis GM, Chalioris CE. Seismic performance of RC beam–column joints retrofitted using
light RC jacket—experimental study. Proceedings of the 1st European Conference on Earthquake Engineering
and Seismology, Geneva, Switzerland, 2006; PN 136.
5. Shannag MJ, Abu-Dyya N, Abu-Farsakh G. Lateral load response of high performance fiber reinforced concrete
beam–column joints. Construction and Building Materials 2005; 19(7):500–508.
6. Karayannis CG, Sirkelis GM. Effectiveness of RC beam–column connections strengthening using carbon-FRP
jackets. Proceedings of the 12th European Conference on Earthquake Engineering, London, U.K., 2002; PR 549.
7. Ghobarah A, Said A. Seismic rehabilitation of beam–column joints using FRP laminates. Journal of Earthquake
Engineering 2001; 5(1):113–129.
8. Karayannis CG, Sirkelis GM. Strengthening and rehabilitation of RC beam–column joints using carbon-FRP
jacketing and epoxy resin injection. Earthquake Engineering and Structural Dynamics 2008. DOI:
10.1002/eqe.785.
9. Karayannis CG, Sirkelis GM. Response of columns and joints with spiral shear reinforcement. Proceedings of the
12th Computational Methods and Experimental Measurements (CMEM), Malta, Wessex Institute of Technology
Transactions on Modelling and Simulation, vol. 41, 2005.
10. Paulay T, Park R. Joints of reinforced concrete frames designed for earthquake resistance. Research Report 84-9,
Department of Civil Engineering, University of Canterbury, Christchurch, New Zealand, 1984.
11. Tsonos AG, Tegos IA, Penelis GG. Seismic resistance of Type 2 exterior beam–column joints reinforced with
inclined bars. ACI Structural Journal 1992; 89(1):3–12.
12. Tsonos AG. Improvement of the earthquake resistance of R/C beam–column joints under the influence of P–

effect and axial force variations using inclined bars. Structural Engineering and Mechanics 2004; 18(4):389–410.
13. Bakir PG. Seismic resistance and mechanical behaviour of exterior beam–column joints with crossed inclined
bars. Structural Engineering and Mechanics 2003; 16(4):493–517.
14. Chalioris CE, Karayannis CG, Favvata MI. Cyclic testing of reinforced concrete beam–columns joints with crossed
inclined bars. Proceedings of the 13th Computational Methods and Experimental Measurements (CMEM), Prague,
Czech Republic, Wessex Institute of Technology Transactions on Modelling and Simulation, 2007; 623–632.
15. ACI Committee 318. Building Code Requirements for Structural Concrete (ACI 318-02) and Commentary (ACI
318R-02). American Concrete Institute: Farmington Hills, MI, 2002.
16. ACI-ASCE Committee 352. Recommendations for Design of Beam–Column Connections in Monolithic Reinforced
Concrete Structures (ACI 352R-02). American Concrete Institute: Farmington Hills, MI, 2002.
17. Fajfar P. Equivalent ductility factors, taking into account low-cycle fatigue. Earthquake Engineering and Structural
Dynamics 1992; 21:837–848.
18. Priestley MJN, Seible F, Calvi GM. Seismic Design and Retrofit of Bridges. Wiley: New York, 1996.

Copyright q 2008 John Wiley & Sons, Ltd. Earthquake Engng Struct. Dyn. 2008; 37:881–897
DOI: 10.1002/eqe