Materials and Structures (2016) 49:3459–3476

DOI 10.1617/s11527-015-0732-1

ORIGINAL ARTICLE

Strut-and-tie modelling for the analysis and design of RC

beam-column joints
Wael Kassem

Received: 22 June 2015 / Accepted: 5 October 2015 / Published online: 11 October 2015
 RILEM 2015

Abstract Beam-column joints have been recognised the proposed model provides the best fit with the
as one of the potentially weaker elements of reinforced measured shear strength. Further, a set of simple code-
concrete moment-resisting frame structures when oriented design formulae, with conservative predic-
subjected to seismic lateral loading. Prior earthquake tions, is proposed to facilitate practical engineering
reconnoitring reported the notable damage that can design.
develop from inappropriate beam-column joint
design. Design formulae available in building codes Keywords Reinforced concrete  Exterior, interior 
and literature differ considerably in functional form, Beam-column joints  Shear  Strut-and-tie model
do not consider all of the key variables that affect the
joint response, and broadly predict quite different
values of shear strength from one another. A mathe-
matical model based on the strut-and-tie model is 1 Introduction
developed to estimate the shear strength of exterior
and interior reinforced concrete beam-column joints. Reinforced concrete moment-resisting frame struc-
The proposed model leads to an explicit single closed- tures (RCMRFS) are commonly used as the primary
form expression for computing the shear strength and lateral load resisting system in low- to moderate-rise
accounts for the shear stress contributions provided by buildings in seismically-active regions. During earth-
the diagonal concrete strut and the shear reinforcement quake ground motions, these structures are often
(including both joint hoops and column intermediate subjected to considerable load reversals and it is
bars), while considering those variables that signifi- important that the connecting members are able to
cantly affect the behaviour of such joints. Model develop their full strength before joints suffer any
parameters were calibrated using the experimental severe loss of strength or stiffness. Assessment of
results of 454 exterior and interior joints available in building damage which has occurred during earth-
the literature. When compared to a number of quake strikes indicates that joint failure may result in
predictive expressions for the shear strength of joints, structural failure [1]. A beam-to-column joint is
frequently defined as the part of column within the
depth of framing beams. Depending on the loading
W. Kassem (&) conditions and the anticipated deformations of the
Division of Construction Engineering, College of
connected frame members when resisting lateral
Engineering at Al-Qunfudah, Umm Al-Qura University,
P.O. BOX 288, Al-Qunfudah 21912, Saudi Arabia loads, Joint ACI-ASCE Committee 352 [2] classifies
e-mail: [email protected] beam-column joints (BCJ) into two separate types.
3460 Materials and Structures (2016) 49:3459–3476

Type 1 is a moment-resisting connection designed on some questions still prevail over the understanding of
the basis of strength and Type 2 is a joint that has joint shear behaviour, and a physically-based rational
members that are required to dissipate energy through model is still needed to address all parameters that
reversals of deformation into the inelastic range. significantly influence the behaviour of such joints [5].
Over the last five decades, the BCJ have been Addressing these points motivated the undertaking of
recognised as potentially one of the weaker elements the current study to develop a simple closed-form
of RCMRFS when subjected to seismic lateral load- expression for shear strength estimation of BCJs under
ing. During seismic attack, both horizontal and lateral loading.
vertical shear forces are applied to the joint region In the current study, a strut-and-tie based model is
with magnitudes that are significantly higher than proposed to solve the problem of exterior and interior
those developed within the nearby beams and columns BCJs shear strength prediction by means of a single
[3]. Consequently, BCJs are more susceptible to the expression more accurate and consistent (uniform in
brittle shear failure which should not be permitted to the prediction) than existing formulas. The proposed
occur if a ductile behaviour of the frame is required. approach accounts for parameters that significantly
Moreover, previous experimental testing has shown influence the behaviour of such joints, including
that BCJs encounter substantial shear distortions under concrete compressive strength, in-plane geometry,
earthquake loading and this can contribute signifi- dimensions of the framing beams and columns, joint
cantly to building story drift [4]. Therefore, under- hoops ratio, bond condition of beam longitudinal
standing joint shear behaviour is essential in reinforcement and the level of column axial stress.
controlling the overall performance of reinforced Model parameters are calibrated using a large exper-
concrete BCJs and frames. imental dataset available in the literature. This formula
BCJs have been the subject of intense experimental is then compared with other formulas available for the
and numerical studies as well as debate in recent shear strength of joints to evaluate the performance of
decades. There is still no general agreement in the proposed model.
research communities about the mechanism of joint
resistance to the induced seismic shear forces that
often results in different amounts and arrangements of 2 Research significance
shear reinforcement in the joint [5]. To some extent,
design provisions represent one position in the debate The current research presents a new method for the
which can be easily identified by comparing different prediction of the shear strength of type 2 BCJs which
design codes (ACI Committee [6]; Eurocode, EC8 [7]; are required to dissipate energy through reversals of
Architectural Institute of Japan, AIJ [8]; New Zealand deformation into the inelastic range. The proposed
code, NZS 3101 [9] and Canadian code, CSA A23.3 analysis method consists of the application of the strut-
[10]). and-tie model (STM) and accounts for shear stress
Examining various theoretical models currently contributions provided by concrete strut and shear
used in code provisions and literature showed that less reinforcement. A database of results from tests of 515
attention was paid to the design of BCJs than to the exterior and interior joints was compiled and used to
design of structural members such as beams and develop a simple code-oriented design formula suit-
columns [11]. It was pointed out that they generally able for practical engineering design.
lead to significantly different predictions, cannot
reproduce results from physical tests and showed
substantial scatter when compared to experimental 3 Development of the strut-and-tie model
results, as they have been originally calibrated on a
rather limited number of experiments [12]. Further- Discontinuity regions such as deep beams, corbels,
more, the majority of design code provisions employ squat walls and BCJs, where shear stresses and
empirical models originally developed based on deformations are the dominant influence, are fre-
simplified rules of mechanics and/or statistical regres- quently designed using the STM approach. Based on
sion analyses carried out on a limited number of the flow of internal forces, the STM idealises the
experimental results. These remarks revealed that member as a truss with concrete struts and
Materials and Structures (2016) 49:3459–3476 3461

reinforcement ties. Several analytical models are idealised as tension and compression forces carried by
available in the literature to estimate the shear strength the longitudinal reinforcement. The compression
of BCJs including empirical models [13, 14], strut- forces are carried by longitudinal steel and concrete
and-tie based models [15, 16] and rotating angle while the shear is carried by the concrete only. Load
softened truss based models [17]. resisted by longitudinal reinforcement is transferred
Despite the widespread use and respect gained by into the joint core through bond (white arrows in
the STM as a rational method for the design of Fig. 1b), while concrete compression and shear forces
discontinuity regions, very few studies have addressed are shown in black and grey arrows in Fig. 1b.
STMs of beam column joints including those proposed Internal forces transmitted from adjacent members
by Hwang and Lee [15, 16], Ortiz [18], Vollum [19] to the joint, shown in Fig. 1a, result in joint shear
and Park and Mosalam [20]. However, a few of these forces in both the horizontal and vertical directions.
models have been validated using an extensive data set These shear forces lead to diagonal compression and
and each of these models has its own conceptual tension stresses in the joint core. The latter will usually
limitations. For example, the softened strut-and-tie result in diagonal cracking of the concrete core.
based model developed by Hwang and Lee [15, 16] Compression forces introduced to a joint by a beam
employs an iterative process and satisfies equilibrium, and a column at diagonally opposite corners of the
compatibility, and the constitutive laws of cracked joint, and combined into a single diagonal compres-
reinforced concrete which make it unsuitable for sion force Sstrut are shown in Fig. 2a. The figure shows
design purposes. Also, the Park and Mosalam [20] the main shear resistance mechanism of BCJs without
model is suitable only for unreinforced joints (joints horizontal reinforcement (hoops). In the present study,
without hoops). In the current study, a general STM it is assumed that failure always occurs from the
suitable for shear strength estimation of exterior and crushing of the diagonal compressive strut (dotted
interior BCJs is introduced. band in Fig. 2a), whose formation is revealed at
increasing loads by the appearance of inclined cracks
on the concrete core. The shear strength component
provided by the strut-and-tie mechanism due to the
4 Mechanisms of shear resistance in beam-column
diagonal compression strut, Sstrut , is mathematically
joints
derived by considering equilibrium of the internal and
external forces applied to the joint. This component is
It is widely accepted that the two principal mecha-
a function of the biaxial strain state that depends on
nisms mobilised in resisting BCJ shear failure are the
overall dimensions, web reinforcement, concrete
compression strut mechanism and the truss mecha-
compressive strength and failure mode. The diagonal
nism [21, 22]. In the compression strut mechanism, the
strut angle of inclination is defined as [23]:
concrete transfers the compression forces across the  
joint by bearing from concrete compression zones of 1 Hb
/ ¼ tan ð1Þ
framing beam(s) and column(s) while member longi- Hc
tudinal reinforcement in tension is anchored in this
where Hb is the beam depth and Hc is the column
compression force field. The truss mechanism
depth, refer to Fig. 1a. Although the diagonal strut
accounts for the potential contribution of joint hoops
normally has a bottle-shaped form, it is assumed that
and columns intermediate rebars in which the tension
the strut has a prismatic form with a uniform width of
and compression forces are transferred via the bond
Ls for simplicity. The strut width, Ls , can be
between joint reinforcement and core concrete.
determined as [16]:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4.1 Modelling strut mechanism Ls ¼ L2b þ L2c ð2Þ

Figure 1 depicts the idealised load distribution in the where Lb and Lc are the depths of the compression
vicinity of a typical exterior BCJ under gravity and zones in the beam and column (see Fig. 2a), respec-
seismic effects. Beam and column forces may be tively, defined as [24]:
3462 Materials and Structures (2016) 49:3459–3476

(a) (b)

Fig. 1 Idealised load distribution in the vicinity of a typical exterior beam-column joint: a gravitational and seismic acting loads on the
joint and framing members; b loads imposed on the concrete core

(a) (b) (c)

Fig. 2 Shear transfer mechanism in beam-column joints by means of: a diagonal mechanism only; b joint hoops only; c column
intermediate bars only
Materials and Structures (2016) 49:3459–3476 3463

Asb fyb depth of the column, Hc , and twice the smaller

Lb ¼ ð3Þ
0:85bb fc0 perpendicular distance from longitudinal axis of beam
to column side as suggested by the ACI 318 code [6].
and The effective compressive strength of concrete in
  strut, rc , can be specified as a function of concrete
0:85Nc
Lc ¼ 0:25 þ Hc ð4Þ compressive strength and calculated using Eq. (10):
Ac fc0
fcd ¼ wfc0 ð10Þ
where Asb is the area of tensile reinforcement of the
beam, fyb is the yield stress of beam bars, bb is the where w is a nondimensional function proposed by the
beam width, fc0 is the concrete compressive strength, European code, EC8 [7], expression:
Nc is the axial force acting on the column and Ac is the  
f0  0 
cross-sectional area of the framing column. Assess- w¼a 1 c fc in MPa ð11Þ
ment of BCJ behaviour showed that column axial 250
loading made a better force flow mechanism to resist where a ¼ 0:6 for interior BCJ and a ¼ 0:48 for
joint shear demand [25]. As suggested by Eq. (4), the exterior BCJ. Combining Eqs. 6, 8, 9 and 10, the
axial load enlarges the cross-sectional area of the average horizontal shear force in the joint due to
compression strut due to the increase in the compres- diagonal concrete strut may be assumed equal to:
sion zones developed in columns and thus increase the  
shear resistance. Hwang and Lee [15] suggested that Vch ¼ wfc0 cosð/Þ Lc bj ð12Þ
the spalling of the compression zone in the beam The nominal horizontal shear stress in the mid-
hinges adjacent to the column faces produces a small height of the BCJ core, mch , due to diagonal concrete
compression zone in the beam and hence the length Lb
strut may be assumed equal to:
could be neglected. Accordingly, the strut width, Ls ,
can be assumed as Vch
mch ¼ ð13Þ
H c bj
Ls ¼ Lc ð5Þ
Equation 13, using Eq. (4) and (12), gives
Provided that the compression stresses in the diagonal
strut are not excessive, which is usually the case, the mch ¼ wkfc0 cosð/Þ ð14Þ
mechanism in Fig. 2a is efficient [3]. The contribution
with
of this strut to joint shear strength may be quantified by:
 
Vch ¼ Sstrut cosð/Þ ð6Þ 0:85Nc
k¼ 0:25 þ ð15Þ
Ac fc0
Vcv ¼ Sstrut sinð/Þ ð7Þ
where Vch and Vcv are the contributions of the concrete 4.2 Modelling truss mechanism
strut mechanism to joint shear strength in the
horizontal and vertical directions, respectively. The The concept of the truss mechanism was first intro-
diagonal force Sstrut can be estimated using the cross- duced to investigate the shear resistance of BCJ under
sectional area of the inclined strut, Astrut , and the seismic loading by Paulay et al. [22]. They recognised
effective compressive strength of concrete in strut, rc that the truss mechanism alone is not capable of
predicting the shear strength of the joint and a
Sstrut ¼ Astrut rc ð8Þ
contribution from the strut mechanism is required.
and Under reversed cyclic loading, diagonal tension cracks
along a potential corner to corner failure plane may
Astrut ¼ Lc bj ð9Þ
develop. Therefore, both horizontal and vertical joint
where bj is the joint effective width, taken equal to the reinforcement are required to prevent this failure. Joint
overall width of the column, bc . If a beam joining into reinforcements allow a diagonal compression field to
a wider column, the effective joint width shall not be emerged and thus provides an appropriate load path
exceed the smaller of beam width, bb , plus overall for both horizontal and vertical shearing forces [3].
3464 Materials and Structures (2016) 49:3459–3476

The truss mechanism adapted in the current study Vth ¼ Sth ¼ Ajh fyh þ kAb fyb ð16Þ
comprises horizontal and vertical load paths (Fig. 2b,
c). The horizontal load path represents the effect of where Ab is the total cross-sectional area of top and
joint hoops and includes one horizontal tie, made up of bottom beam reinforcements, fyb is the yield stress of
the horizontal shear reinforcement and part of the beam reinforcement and k is a bond factor, which is a
beam longitudinal reinforcement, and two flat struts. nominally dimensionless measure of bond demand
The vertical mechanism, composed of one vertical tie associates the magnitude of stresses in the longitudinal
and two steep struts, is proposed to account for the beam reinforcement to the average hoops stresses. The
forces carried out by the column vertical reinforce- value of k relys on the bond behaviour of beam bars; a
ments; see Fig. 2c. The vertical tie contains only the phenomenon which even independently is not well
intermediate longitudinal steel bars of the column understood [32]. Thus, this value will be estimated
passing through the joint (excluding the edge bars). based on experimental results. The BCJ horizontal
The main function of the horizontal shear rein- shear stress associated with horizontal load path, vth , is
forcement (hoops) in the joint is to confine the core given by Eq. (17):
concrete, to constrain the width of crack and to resist a Ajh fyh þ kAb fyb
part of the joint shear force. Prior experimental results vth ¼ ð17Þ
Hc bj
showed uneven contribution of the joint hoops in
resisting shear forces. Moreover, investigating hoops Due to strength requirements, a considerable
strain profiles denoted that the hoops strain near the amount of longitudinal bars is frequently provided in
mid height of the concrete core was much higher than columns and passing through the joint region. The
those near the beam bars [26]. Accordingly, it may influence of column intermediate bars on the behaviour
reasonably be assumed that the average tensile force in of BCJ is not entirely understood. Some researchers
the horizontal hoops is equal to Ajh fyh , where Ajh is the agreed that the presence of column intermediate bars
total area of joint hoops and fyh is their corresponding has very little influence on the joint shear strength [33,
yield stress. 34]. On the contrary, others reported that the increase in
Bakir and Boduroğlu [27] concluded that increas- column intermediate bars ratio increased the joint shear
ing the percentage of beam longitudinal reinforcement strength in their tests [17, 35–37]. A similar contra-
has a beneficial effect on increasing the joint horizon- diction on the influence of column intermediate bars on
tal joint shear force with less deterioration of bond the behaviour of BCJ can be found in design codes.
resistance around the beam longitudinal bars, resulting While the ACI 318 [6] and AIJ [8] methods for the
in a wider diagonal strut to support the larger design of BCJ ignore this effect and only account for
horizontal joint shear force. Bakir and Boduroglu the concrete contribution provided by the diagonal
[28] and Kim and LaFave [25] showed that the bond mechanism, the Eurocode, EC8 [7] and the New
conditions of beam longitudinal reinforcement has a Zealand Code NZS 3101 [9] imposed a minimum limit
significant effect on the shear behaviour and strength on the amount of column intermediate bars provided to
of the joint. Moreover, based on the patterns of strain resist the joint vertical shear.
histories of beam longitudinal reinforcement at the Contrary to the joint hoops, the column intermediate
column centreline, Attaalla [29] showed that these bars would not be subjected to considerable tensile stresses
strains were somewhat similar to that obtained from since the column is supposed to remain elastic. Therefore,
joint hoops, implying that beam steel within the joint they are expected to contribute to the vertical tie
partially participated in resisting the horizontal shear mechanism of the joint shear resistance [35]. For simplic-
force. Furthermore, bond deterioration of beam lon- ity, it may be reasonably assumed that the average tensile
gitudinal reinforcements inside the joint core is force in the column intermediate reinforcement, Stv , can
expected and, thus, the stresses in beam bars would be assumed equal to Ajv fyv , where Ajv is the cross-
not vanish under applied loading [5, 30, 31]. Conse- sectional area of column intermediate bars excluding
quently, the contribution of the hoops and beam edge bars and fyv is their corresponding yield strength.
longitudinal reinforcements in the horizontal load path From statics, the horizontal shear force carried by the
may be assumed as follows: vertical tie is equal to [see details H in Fig. 2]:
Materials and Structures (2016) 49:3459–3476 3465

Vtv ¼ 2Stv cotð/2 Þ ¼ Stv cotð/Þ ¼ Ajv fyv cotð/Þ ð18Þ Fig. 3a, b. In order to estimate the contribution of
diagonal and truss mechanisms to joint horizontal
and the corresponding joint horizontal shear stress shear strength, an attempt was made to calculate the
associated with column intermediate bars, vtv , is given average force associated with each mechanism with
by Eq. (19):
respect to the joint aspect ratio, Hb=Hc , using
Ajv fyv cotð/Þ Eqs. (6), (16) and (18), see Fig. 4. The forces corre-
vtv ¼ ð19Þ
H c bj spond to the elastic forces in the respective mecha-
nism, assuming that the total shear force is carried by
4.3 Interaction between strut and truss each of the three mechanisms. A careful examination
mechanisms of the results shown in Fig. 4 shows that increasing the
joint aspect ratio results in increasing the elastic forces
To satisfy the basic equilibrium requirement of in the diagonal strut as well as column intermediate
Vjh ¼ Vch þ Vth þ Vtv , it is interesting to estimate the bars. However, experimental observations showed
contribution of the shear force associated with the that the force carried by the diagonal resistance
diagonal strut,Vch , horizontal,Vth , and vertical, Vtv mechanism in a BCJ is inversely proportional to its
load paths. The fraction of the total shear resistance aspect ratio [23, 38]. Moreover, based on observations
associated with each of the different mechanisms from a large data sets of literature test results, a similar
depends on several parameters, including the amount conclusion was drawn by Vollum [19] and Bakir and
and distribution of reinforcement in the joint, com- Bouroğlu [39]. Consequently, the results from the
pressive strength of the concrete, bond conditions, simple statically determinate STMs may be consid-
level of column axial loading and the joint geometrical ered inconsistent with the experimentally observed
dimensions. behaviour, resulting in reinforcement configurations
The STM with either horizontal or vertical load that are not appropriate for the particular stress field
path is a statically determinate truss system, see [40].

Fig. 3 Idealised strut-and- (a) (b) (c)

tie models for: a horizontal
mechanism; b vertical
mechanism; c combined
truss mechanism; and
d combined diagonal and
truss mechanisms

(d)
3466 Materials and Structures (2016) 49:3459–3476

estimated as the sum of the components assigned to

each of the three basic load transfer mechanisms,
represented by their corresponding forces, multiplied
by a correction coefficient:
Vjh ¼ cc Sstrut þ ch Sth þ cv Stv ð20Þ
where cc , ch and cv are the correction factors for the
diagonal mechanism and horizontal and vertical load
paths, respectively. Accordingly, the joint horizontal
shear strength:
  Ajh fyh þ kAb fyb
vjh ¼ cc wkfc0 cosð/Þ þ ch
H c bj
Ajv fyv cotð/Þ
þ cv ð21Þ
H c bj
Fig. 4 Percentage of shear force resisted by diagonal strut,
horizontal, vertical load paths and statically indeterminate strut- Equation (21) can be rearranged by using Eq. (1) and
and-tie model  
the reinforcement ratio q, where qjh ¼ Ajh Hc bj and
qb ¼ Ab =ðHb bb Þ and qc ¼ Ajv ðHc bc Þ
Furthermore, a combined STM comprising the
 
diagonal mechanism and the truss mechanism (in- vjh ¼ cc wkfc0 cosð/Þ  
cluding both horizontal and vertical load paths) was bb
proposed. The model is a fourth statically indetermi- þ ch qjh fyh þ kqb fyb tanð/Þ
  bj
nate STM while support configurations were selected bc
so that the overall truss would be externally statically þ cv qc fyv cotð/Þ ð22Þ
bj
determinate, see Fig. 3d. In order to estimate the
member forces, a stiffness method was employed by Further, Eq. (22) can be nondimensionalised by
assuming that the axial rigidity of all members are using the definition of the reinforcement index x,
equal. Applying this assumption obviously simplified where xh ¼ qjh fyh fc0 , xb ¼ qb fyb fc0 and
the analysis of the proposed combined STM. Figure 4 xv ¼ qc fyv fc . Taking out the common factor, fc0 in
0

presents the percentage of the shear force acting on the Eq. (22) and using the nondimensionalised variables,
vertical and the horizontal members of the statically Eq. (22) becomes:
indeterminate truss model. Obviously, the results in    
bb
the figure proved that the interaction between mech- vjh ¼ fc0 cc ½wkcosð/Þ þ ch xh þ kxb tanð/Þ
anisms in the investigated statically indeterminate bj
   
truss model resulted in a notable reduction in the bc
þcv xv cotð/Þ ð23Þ
magnitude of the forces in the ties (Sth , Stv ). bj
where /, w and k are calculated using the expressions
given by Eqs. (1), (11) and (15), respectively. The
5 Closed-form expression for beam-column shear
values of parameters cc , ch and cv appearing in
strength
Eq. (23) can be obtained by calibration with test data
available in the literature.
To overcome the discrepancies between the experi-
mentally observed behaviour of BCJs and the values of
the internal forces calculated using the elastic truss 6 Experimental database
models, correction factors were proposed based on the
experimental results. The corrections factors denote With the purpose of calibrating model parameters and
the contribution of the shear force associated with the assessing the accuracy of the existing expressions that
diagonal strut,Vch , horizontal,Vth , and vertical, Vtv load are used in BCJ design, an extensive experimental
paths. The total horizontal shear force of the BCJ is database spanning a wide range of design parameters
Materials and Structures (2016) 49:3459–3476 3467

was constructed. The database comprises the experi- joint reinforcement ratios, as summarised in Fig. 5
mental results of 515 reinforced concrete BCJs (319 which shows the frequency distribution of different
exterior joints and 196 interior joints), and was parameters in the specimens included in the analysis.
collected from (in chronological order): Alva and de Since the exact value of the joint experimental
Cresce El Debs [41]; Attaalla [29]; Bindu and Jaya shear strength, Vtest , was not explicitly reported in
[42]; Chalioris and Favvata [43]; Chun and Shin [38]; several tests, the experimental joint shear force was
Chun and Kim [44]; Chun and Lee [45]; Chun and Oh calculated from force equilibrium and a free-body
[46]; Chutarat and Aboutaha [47]; Clyde and Pan- diagram at mid-height of the joint panel at overall joint
telides [48]; Dhakal and Pan [49]; Durrani and Zerbe maximum story shear as explained in Kim and LaFave
[50]; Ehsani and Wight [31]; Ehsani and Alameddine [88] ENREF 12. Afterwards, the joint shear stress,vtest ,
[51]; Ehsani and Wight [52]; Ehsani and Moussa [53]; was estimated as the maximum joint shear demand
Fernandes and Melo [54]; Ganesan and Indira [55]; divided by the effective joint shear area as per the ACI
Gençoǧlu and Eren [56]; Hakuto and Park [57]; Hamil 318 code [6].
[58]; Hassan [59]; Hegger and Sherif [60]; Hwang On the basis of the analysis of the 454 BCJs
et al. [35]; Hwang and Lee [61]; Ibrahim [62]; Idayani considered, the values of model parameters cc , ch , k
[63]; Karayannis and Sirkelis [64]; Karayannis and and cv , appearing in Eq. (23), have been iteratively
Sirkelis [65]; Kim and LaFave [66]; Kotsovou and changed. With the aim to achieve the best prediction
Mouzakis [67]; Kuang and Wong [68]; Kusuhara and uniformity of shear strength, the fmincon function
Shiohara [69]; Lee and Ko [70]; Lehman and Stanton available in MATLAB [89] was employed. The
[71]; Li and Pan [72]; Li and Tran [73]; Li and function searches for the minimiser of a scalar
Kulkarni [74]; Lu and Urukap [75]; Masi et al. [11]; function of multiple parameters, within a zone spec-
Murty and Rai [76]; Pampanin and Calvi [77]; ified by linear constraints and limits. The constraint
Pantelides and Hansen [78]; Park and Mosalam [79]; utilised in the current study was that the average (Avg)
Parker and Bullman [80]; Scott [81]; Shiohara and value of the ratio of shear strength measured in the test
 
Kusuhara [82]; Teng and Zhou [83]; Tsonos and Tegos and the calculated shear strength mtest mjh was set
[84]; Tsonos and Tegos [85]; Wallace and McConnell equal to 1.0. No lower or upper limits were applied on
[86]; Wong [17] and Wong and Kuang [87]. the model coefficients. Further, the unknowns were
BCJs considered in the analysis were limited to calculated by minimising the coefficient of variation
those which (1) were subassemblies of RCMRFS, with (Cov), which is calculated as the ratio between the
at least 1/3 scale; (2) were tested under either standard deviation (s) and the average (Avg), associ-
monotonically or cyclically increased lateral loading; ated with the ratios of experimental to calculated shear
(3) experienced joint shear failure (either in conjunc- strengths.
tion with beam joint failure [BJ] or joint failure A search for the coefficients that satisfy the
[J] without yielding of beam reinforcement); (4) none constraints and minimise the variation leads to the
of the specimens in the constructed database had out- set of cc = 0.21, ch = 0.09, k = 3.47 and cv = 0.22
of plane members in the form of transverse beams or for exterior BCJs, and cc = 0.26, ch = 0.44, k = 1.39
slabs. Further, the database contains only specimens and cv = 0.07 for interior BCJs. Substituting these
with conventional types of reinforcement anchorage constants into Eq. (23), the proposed closed-form
(no headed bars); longitudinal beam and column expression is as follows:
reinforcement are either anchored by hooks or pass
continuously through the joint panel, according to the • For exterior beam-column joints
in-plane geometry. vjh ¼ ð0:21½wkcosð/Þ
Among the considered 515 joints, only 454 joints  
bb
[283 exterior joints and 171 interior joints] satisfy the þ0:09 xh þ 3:47xb tanð/Þ
bj
proposed criteria and the results of 61 joints results   
bc
were excluded from the database. The assembled þ 0:22 xv cotð/Þ fc0 ½MPa; ð24a1Þ
database covers a wide range of different variables bj
comprising the geometrical properties, concrete com- • For interior beam-column joints
pressive strength, the column axial force level and the
3468 Materials and Structures (2016) 49:3459–3476

(a) (b) (c)

(d) (e) (f)

(g) (h) (i)

Fig. 5 Frequency distribution of main parameters in the g joint horizontal reinforcement ratio; h vertical shear rein-
compiled database: a column depth; b beam depth; c joint forcement ratio; and i column coexisting axial stress
aspect ratio; d beam width; e column width; f concrete strength;

vjh ¼ ð0:26½wkcosð/Þ and exterior joints computed according to the current

  design codes [2, 6–8, 10] and shear strength expres-
bb
þ0:44 xh þ 1:39xb tanð/Þ sions [5, 13, 14, 90, 91] address the same improvement
bj
   in shear strength. Another important observation is
bc that the value of model parameter cv , which represents
þ 0:07 xv cotð/Þ fc0 ½MPa ð24b1Þ
bj the potential contribution of the vertical load path to
joint shear strength, was small and significantly
A careful inspection of model parameters cc , ch ; k smaller than the corresponding value of parameters
and cv appearing in Eq. 24 reveals that their values are ch and k, which reflects the potential contribution of
higher for interior joints than the corresponding values the horizontal load path to joint shear strength. This
for exterior joints, which implies that an interior joint observation implies that the effect of column inter-
possesses a higher strength than a similar exterior one. mediate bars on joint shear strength is modest in
This may be attributed to the fundamental differences comparison with the effect of the joint hoops and bond
in the mechanisms of beam bar anchorages [3]. It is condition of beam longitudinal bars. This observation
interesting to notice that the shear strengths of interior concurs with the evaluation of available data [27] and
Materials and Structures (2016) 49:3459–3476 3469

test observations [17]. Moreover, the bond factor k Vjh , have been selected for comparison with respect to
takes the value of 0.31 and 0.61 for exterior and the experimental results of the 454 BCJs.
interior BCJs, respectively. Typical values for this
parameter ranges from zero for perfect bond to 1 for 7.1 Hwang and Lee [90]
poor bond [28].
After careful consideration of an extensive database VjhHwang ¼ kffc0 Astr cosðaÞ ð25Þ
of BCJs, Kim and LaFave [25] showed that a robust
model to predict the shear strength of BCJs should where k is a factor accounting for the contribution of
account for the most significant influence parameters, horizontal and vertical joint shear reinforcement, f is
including the concrete compressive strength, fc0 , in- the concrete softening coefficient, f ¼ p ffiffiffi0  0:52 Astr
3:35
fc
plane geometry (interior or exterior joints), dimen- is the effective area of the diagonal strut depending on
sions of the framing beams and columns
(bc ; bb ; Hc ; Hb ), joint hoops and beam reinforcement column axial load and dimensions of the joint, Astr ¼
ratios. Besides the aforementioned parameters, bond 0:25 þ 0:85N
Ac f 0 Hc ; and a is the angle of inclination of
c

condition and the level of column axial stress are also c

known to have an impact on joint shear strength. All the diagonal compression strut with respect to the
Hb
these parameters are satisfied in the proposed closed- longitudinal direction of the beam, a ¼ tan1 Hc .
form expression (Eq. 24). Moreover, explicitly con-
sidering the influence of joint hoops and column 7.2 Wang et al. [91]
intermediate bars represents a major feature of the
proposed analysis method.  2 
sin a sin2 a
The calculated shear strength,vjh , versus the shear VjhWang ¼ b 1   0:8 0 ry
ft fc
strength reported in the test, vtest , is plotted in Fig. 6 for  
1
exterior and interior BCJ considered in this investiga- þ 0:8 fc0 sin 2a Aj ð26Þ
tion. Based on the optimisation process, the average ft
ratio between vtest and vjh is equal to one and the Cov, where b is the a factor equal to 1.0 or 0.8, depending on
was found equal to 0.20 and 0.18 for the exterior and the in-plane geometry; a is the inclination of the
interior BCJs, respectively, with overall value of 0.19. diagonal strut, ft is the nominal tensile strength of
concrete with contributions from joint shear reinforce-
ment; and ry is the column axial stress.
7 Comparison with existing shear strength
formulas 7.3 Kim and LaFave [13]

To assess the accuracy of the proposed model, five sets  0:75

Vjhkim ¼ 1:31at bt ft ðJIÞ0:07 ðBIÞ0:25 fc0 Ajh ð27Þ
of recent expressions for estimating the joint strength,

Fig. 6 Experimental joint

shear strength, vtest , versus
developed joint shear
strength, vjh (Eq. 24)
3470 Materials and Structures (2016) 49:3459–3476

where at and bt are parameters accounting for in-plane strength by means of the proposed closed-form
and out-of-plane geometry, respectively, ft is a factor expression and different methods versus measured
accounts for beam eccentricity, JI is the joint trans- shear strength. Moreover, Avg and Cov, values are
verse reinforcement index (accounts for the volumet- also reported in the figure and summarised in Table 1.
ric joint shear reinforcement ratio) and BI is the beam The results in Fig. 7 and Table 1 demonstrate that
reinforcement index. the considered expressions perform differently
depending on the in-plane geometry of the joint.
However, in all cases the proposed closed-form
7.4 Attaalla [5] expression can be considered as the best performing
2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi expression and the largest strength ratio and coefficient
   ffi 3
qt fytNb =bb hb  qt fytNc =bc hc of variation are for the Hwang and Lee [90] model
VjhAtta ¼ 0:45;ffc0 4 5Aj
followed by Wang et al. [91] model. For exterior joints,
qt fyt þ ql fyl  ðNb =bb hb þ Nc =bc hc Þ
the mean values of the shear strength ratios indicate
ð28Þ that all the expressions except Hwang and Lee’s [90]
and Wang et al.’s [91] overestimate the shear strength.
where ; is a parameter accounting for in-plane
In the case of interior joints, the equations proposed by
geometry; f is the concrete softening coefficient
  Kim and LaFave [13] and Tran et al. [14] overestimate
110fc0
3  
defined as f ¼ 0:4 1 þ 69  1 fc0 in MPa ; the shear strength, and none of the equations produces
accurate estimates of shear strength. Overall, it is quite
qt and ql are the joint reinforcement ratio in transverse
evident that the proposed closed-form expression
and longitudinal direction with their corresponding
yields the most uniform results with a minimum value
yield stresses fyt and fyl , respectively, while Nb and Nc
of coefficient of variation and average shear strength
are the axial forces in beam and column assumed
ratio equal to one. It is also important to stress that the
positive in tension; and Aj is the effective joint area.
model reliability depends, to an extent, on the
completeness and robustness of the test database used
7.5 Tran et al. [14]
in model development. Moreover, the best efficiency
  could be conditioned to the coincidence of the test set
pffiffiffiffi Nc
VjhTran 0
¼ f c c1 þ þ 1:2nb Ajh considered for coefficient evaluation and efficiency
 bc hc fc0  assessment.
þ c2 Asjh fjhy þ Asjv fjvy ð29Þ

where c1 ¼ 0:81 and c2 ¼ 0:14 for interior joints and

c1 ¼ 0:34 and c2 ¼ 0:22 for exterior joints, nb is the 8 Design formula
beam bar index that accounts for beam bar diameter to
column depth ratio, Asjh and Asjv are the cross- Despite the superiority achieved by the proposed
sectional areas of joint horizontal and vertical rein- model in matching the experimental shear strength, as
forcements and fjhy and fjvy are their corresponding expressed by the minimum coefficient of variation and
yield stress. average strength ratio of unity, an unmediated use of
For all the considered models, the joint shear the expression as a code-oriented design equation is
not possible. The reason for not allowing the unmedi-
strength, vmodel
jh , was estimated as the joint shear force,
ated use is that the model does not account for any
Vjhmodel divided by the effective joint shear area as per safety factor in the predictions besides the unsafe
the ACI 318 code [6]. A statistical analysis by using predictions in 227 specimens out of the 545 considered
means of average and dispersion in the ratios between joints, refer to Fig. 7a. In order to avoid these
the experimental, vtest , and calculated joint shear undesirable and unconservative predictions of shear
stress, vmodel
jh , provides insight into the reliability of the strength, Eq. (24) may be multiplied by a factor that
proposed closed-form expression. Accuracies of the can increase the average strength ratio without altering
proposed Eq. (24) is assessed in light of the other the coefficient of variation. The factor value is
expressions introduced earlier using the experimental determined on a statistical base by setting 5 %
results in Fig. 7 which represents the calculated shear probability that the design shear strength is on the
Materials and Structures (2016) 49:3459–3476 3471

(a) (b) (c)

(d) (e) (f)

Fig. 7 Calculated shear strength by means of the proposed closed-form expression and different methods versus measured shear
strength

Table 1 Statistical analysis  

Strength ratio vtest vjh Exterior joints Interior joints All joints
of shear strength ratio using
different sets of equations Avg s Cov Avg s Cov Avg s Cov

Closed form, Eq. (24) 1.00 0.20 0.20 1.00 0.18 0.18 1.00 0.19 0.19
Hwang and Lee [90] 1.46 0.36 0.25 2.46 0.88 0.36 1.83 0.78 0.42
Wang et al. [91] 1.30 0.61 0.47 1.94 0.83 0.43 1.53 0.76 0.50
Kim and LaFave [13] 0.84 0.26 0.31 0.97 0.31 0.32 0.89 0.29 0.32
Attaalla [5] 0.81 0.43 0.53 1.67 0.76 0.46 1.11 0.70 0.63
Tran et al. [14] 0.87 0.45 0.51 0.90 0.47 0.52 0.88 0.46 0.52
Design formula, Eq. (30) 1.48 0.30 0.20 1.45 0.26 0.18 1.46 0.27 0.19
ACI 318 [6] 1.25 0.32 0.25 2.03 0.75 0.37 1.55 0.65 0.42
Eurocode, EC8 [7] 1.72 0.58 0.34 2.75 1.04 0.38 2.11 0.93 0.44
AIJ [8] 1.28 0.30 0.23 2.10 0.69 0.33 1.59 0.63 0.39
NZS 3101 [9] 1.60 0.82 0.51 2.53 1.29 0.51 1.91 1.09 0.57
CSA A23.3 [10] 1.23 0.32 0.26 2.05 0.70 0.34 1.54 0.64 0.41

unsafe side (i.e. mtest mjh \1) as suggested by the multiplication factor were found equal to 0.672 and
Eurocode [92]. Based on the experimental results of 0.705 for exterior and interior BCJs, respectively.
the considered 454 joints, the values of the Consequently, Eq. (24) can be converted into a
3472 Materials and Structures (2016) 49:3459–3476

closed-form that could perhaps be conveniently used 8.2 European code, EC8 [7]
as a practical design expression, that is:
rffiffiffiffiffiffiffiffiffiffiffiffiffi
td
• For exterior beam-column joints VjhEC8  gfc0 1  hjc bbc ð32Þ
g
Vjhd ¼ ð0:14½wkcosð/Þ f0
  where g ¼ a 1  250 c
and g ¼ 0:6 for interior joints
bb and 0.48 for exterior joints; td is the normalised axial
þ0:06 xh þ 3:47xb tanð/Þ
bj force in the column above the joint; hjc is the distance
  
bc between the extreme layers of column reinforcement;
þ 0:15 xv cotð/Þ fc0 ½MPa;
bj and bbc is the effective joint width.
ð30a1Þ
• For interior beam-column joints 8.3 Japanese design guidelines, AIJ [8]

Vjhd ¼ ð0:18½wkcosð/Þ VjhAIJ ¼ kuFj bj Dj ð33Þ

 
bb
þ0:31 xh þ 1:39xb tanð/Þ where k is a factor accounting for the joint in-plane
bj
   geometry (1.0 for interior joints, 0.7 for exterior
bc
þ 0:05 xv cotð/Þ fc0 ½MPa; joints), u is a factor accounting the effects of out-of-
bj
plane geometry (1.0 for joints with transverse beams
ð30b1Þ and 0.85 for other cases), Fj is the standard value of the
 0:7
joint shear strength; Fj ¼ 0:8 fc0 and bj and Dj are
In spite of the extensive experimental and theoretical the effective joint width the effective column depth,
investigations on BCJ behaviour, there is still no respectively.
consensus between the research communities regarding
the mechanism of joint resistance to the induced seismic 8.4 NZS 3101 Code [9]
forces that often results in varying configurations of
transverse reinforcement in the join. BCJs design meth-   
fc0 Asjh fyj
ods provided in different countries vary considerably VjhNZS ¼ h1 b1  0:20fc0 h1 b1  10h1 b1
6a As fyb
since they were developed entirely on the empirical
interpretation of different experiments. The primary ð34Þ
difference in the design philosophy of BCJs is whether where a is a parameter accounting for column axial
the truss mechanism is considered or disregarded. stress; Asjh and fyj are the total cross-sectional area of
Using the database, an assessment is now made of joint hoops and their corresponding yield stress,
the design relationships provided by the current study respectively; As and fyb are the greater of the area of
and five code shear strength computing models, top or bottom beam reinforcement passing through the
including the ACI 318 [6], the European code, EC8 joint (excluding bars in an effective tension flange)
[7], the Architectural Institute of Japan design guide- and their yield stress, respectively; h1 is the column
lines AIJ [8], the New Zealand Code NZS 3101 [9] and thickness; and b1 is the effective joint width.
the Canadian code CSA A23.3 [10].
8.5 CSA A23.3 [10]
8.1 ACI 318 [6]
pffiffiffiffi
pffiffiffiffi VjhACSA ¼ c fc0 Aj ð35Þ
VjhACI ¼ c fc0 Aj ð31Þ
where c is equal to 2.2, 1.6, or 1.3, depending on the
where c is equal to 1.7, 1.2, or 1.0, depending on the geometry and the level of confinement exerted by the
geometry and the level of confinement exerted by the beams framing into the joint, and Aj is the effective
beams framing into the joint, and Aj is the effective joint area defined as the minimum cross-sectional area
joint area; Aj ¼ Hc bj . within a joint in a plane parallel to the axis of the
Materials and Structures (2016) 49:3459–3476 3473

reinforcement generating the shear in the joint, equal 9 Summary and conclusions
to the lesser of the gross area of the column or 2bb Hc .
Comparisons of the ratio between the joint measured Building design codes and literature proposed sev-
shear stress, vtest , and predictions obtained from code eral predictive expressions for estimating the shear
provisions VjhCode and the proposed design formula strength of BCJs. These expressions considerably
[Eq. (30)] versus concrete strength are shown in differ in functional form, do not consider all of the
Fig. 8. Further, the Avg and Cov values of the strength key variables that influence the response of joints,
. and broadly predict quit different values of shear
ratio, vtest VjhCode , are summarised in Table 1. The
strength from one another. A closed-form expression
comparison results in the table clearly confirm the that accounts for the shear strength contributions
supremacy of the proposed design formula as provided by the diagonal concrete strut and shear
expressed by the least scatter in predicting BCJ shear reinforcement (including both joint hoops and
strength over the whole range, while the EC8 [7] and column intermediate bars) is presented, while con-
NZS 3101 [9] code procedures lead to over-conserva- sidering those variables that significantly affect the
tive predictions with high values of coefficient of behaviour of such joints. Model parameters were
variations. Furthermore, Fig. 8 visually shows and calibrated using a large number of published exper-
confirm the finding that the proposed closed-form imental results of exterior and interior joints avail-
design formula is the most reliable at predicting joint able in the literature. From the above provided
shear strength when compared with considered design analysis and considerations relevant to BCJs, it can
models. be concluded that:

(a) (b) (c)

(d) (e) (f)

Fig. 8 Variation of experimental test results to design shear strength ratios with respect to concrete compressive strength, fc0
3474 Materials and Structures (2016) 49:3459–3476

1. Using the introduced STM, a single closed-form 14. Tran TM, Hadi MN, Pham TM (2014) A new empirical
expression, which is more accurate and reliable model for shear strength of reinforced concrete beam–col-
umn connections. Mag Concr Res 66(10):514–530
than other predictive equations available in the 15. Hwang S-J, Lee H-J (1999) Analytical model for predicting
design codes and literature, for predicting the shear strengths of exterior reinforced concrete beam-col-
shear strength of BCJs is provided; umn joints for seismic resistance. ACI Struct J 96(5):
2. The closed-form expression proposed herein was 846–858
16. Hwang S-J, Lee H-J (2000) Analytical model for predicting
compared with the measured shear strength of 454 shear strengths of interior reinforced concrete beam-column
exterior and interior BCJs available in the liter- joints for seismic resistance. ACI Struct J 97(1):35–44
ature, and a favourable correlation was found; 17. Wong HF (2005) Shear strength and seismic performance of
3. To facilitate practical engineering design, a set of non-seismically designed reinforced concrete beam-column
joints. PhD thesis, Hong Kong University of Science and
simple code-oriented design formula, with con- Technology, Hong Kong
servative predictions, is proposed. 18. Ortiz R (1993) Strut and tie modeling of reinforced concrete
short beam and beam-column joints. PhD thesis, University
of Westminster
19. Vollum RL (1998) Design and analysis of exterior beam
References
column connections. PhD thesis, Imperial College of Sci-
ence Technology and Medicine, University of London
1. Mitra N, Lowes L (2007) Evaluation, calibration, and ver- 20. Park S, Mosalam K (2009) Shear strength models of exterior
ification of a reinforced concrete beam-column joint model. beam-column joints without transverse reinforcement.
ASCE J Struct Eng 133(1):105–120 Pacific Earthquake Engineering Research Center
2. Joint ACI-ASCE Committee 352 (2002) Recommendations 21. Mazzoni S, Moehle JP (2001) Seismic response of beam-
for design of beam-column connections in monolithic column joints in double-deck reinforced concrete bridge
reinforced concrete structures (ACI 352R-02). American frames. ACI Struct J 98(3):259–269
Concrete Institute Farmington Hills, Michigan 22. Paulay T, Park R, Preistley M (1978) Reinforced concrete
3. Paulay T, Priestley M (1992) Seismic design of reinforced beam-column joints under seismic actions. ACI J Proc
concrete and masonry buildings. Wiley, New York 75(11):585–593
4. Canbolat BB, Wight JK (2008) Experimental investigation 23. Park S (2010) Experimental and analytical studies on old
on seismic behavior of eccentric reinforced concrete beam- reinforced concrete buildings with seismically vulnerable
column-slab connections. ACI Struct J 105(2):154–162 beam-column joints. PhD thesis, University of California,
5. Attaalla SA (2004) General analytical model for nominal Berkeley
shear stress of type 2 normal-and high-strength concrete 24. Hong S, Lee S (2004) Strut-and-tie models for deformation
beam-column joints. ACI Struct J 101(1):65–75 of reinforced concrete beam-column joints dependent on
6. American Concrete Institute (2011) Building code plastic hinge behavior of beams. In: Proceedings of 13th
requirements for structural concrete (ACI 318-11) and world conference on earthquake engineering vancouver,
commentary (ACI 318R-11). Farmington Hills, Michigan BC, Canada, pp 1–6
7. European Committee for Standardization (CEN) (2004) 25. Kim J, LaFave JM (2007) Key influence parameters for the
Design of structures for earthquake resistance, part 1: gen- joint shear behaviour of reinforced concrete (RC) beam–
eral rules, seismic actions and rules for buildings. Eurocode column connections. Eng Struct 29(10):2523–2539
8, BS EN 1998-1, BSI British Standards, London 26. Park R, Milburn J (1983) Comparison of recent New
8. Architectural Institute of Japan (AIJ) (1999) AIJ standard Zealand and United States seismic design provisions for
for structural calculation of reinforced concrete structures. reinforced concrete beam–column joints and test results
Tokyo, p 168 from four units designed according to the New Zealand
9. Standards New Zealand (2006) Concrete structures standard code. Bull N Z Natl Soc Earthq Eng 16(1):3–24
(NZS 3101-1). Standards Association of New Zealand, 27. Bakir P, Boduroğlu H (2002) A new design equation for
Wellington predicting the joint shear strength of monotonically loaded
10. CSA Committee A23.3 (2004) Design of concrete struc- exterior beam-column joints. Eng Struct 24(8):1105–1117
tures. Canadian Standard Association, p 232 28. Bakir P, Boduroglu H (2006) Nonlinear analysis of beam–
11. Masi A, Santarsiero G, Lignola GP, Verderame GM (2013) column joints using softened truss model. Mech Res Com-
Study of the seismic behavior of external RC beam–column mun 33(2):134–147
joints through experimental tests and numerical simula- 29. Attaalla SAA (1997) Seismic shear capacity of beam-col-
tions. Eng Struct 52:207–219 umn joints in multistory reinforced concrete-frame build-
12. Lima C (2011) Capacity models for beam-to-column joints ings. PhD thesis, University of Southern California
in RC frames under seismic actions. PhD thesis, University 30. Leon RT (1990) Shear strength and hysteretic behavior of
of Salerno, Italy interior beam-column joints. ACI Struct J 87(1):3–11
13. Kim J, LaFave JM (2008) Probabilistic joint shear strength 31. Ehsani MR, Wight JK (1985) Exterior reinforced concrete
models for design of RC beam-column connections. ACI beam-to-column connections subjected to earthquake-type
Struct J 105(6):770–780 loading. ACI J 82(4):492–499
Materials and Structures (2016) 49:3459–3476 3475

32. Pantazopoulou S, Bonacci J (1992) Consideration of ques- 51. Ehsani MR, Alameddine F (1991) Design recommendations
tions about beam-column joints. ACI Struct J 89(1):27–36 for type 2 high-strength reinforced concrete connections.
33. Park S, Mosalam KM (2010) Analytical and experimental ACI Struct J 88(3):277–291
study of RC exterior beam–column joints without transverse 52. Ehsani MR, Wight JK (1985) Effect of transverse beams
reinforcement. In: Proceedings of the 7th international and slab on behavior of reinforced concrete beam-to-col-
conference on urban earthquake engineering (7CUEE) & umn connections. ACI J Proc 82(2):188–195
5th international conference on earthquake engineering 53. Ehsani MR, Moussa AE, Vallenilla CR (1987) Comparison
(5ICEE), p 803–811 of inelastic behavior of reinforced ordinary- and high-
34. Sengupta P, Li B (2013) Modified Bouc-Wen model for strength concrete frames. ACI Struct J 84(2):161–169
hysteresis behavior of RC beam–column joints with limited 54. Fernandes C, Melo J, Varum H, Costa A (2013) Cyclic
transverse reinforcement. Eng Struct 46:392–406 behavior of substandard reinforced concrete beam-column
35. Hwang S-J, Hung-Jen L, Liao T-F, Kuo-Chou W, Hsin- joints with plain bars. ACI Struct J 110(1):137–147
Hung T (2005) Role of hoops on shear strength of reinforced 55. Ganesan N, Indira PV, Abraham R (2007) Steel fibre rein-
concrete beam-column joints. ACI Struct J 102(3):445 forced high performance concrete beam-column joints
36. Park R, Keong YS (1979) Tests on structural concrete subjected to cyclic loading. ISET J Earthq Technol
beam-column joints with intermediate column bars. Bull N 44(3–4):445–456
Z Natl Soc Earthq Eng 12(3):189–203 56. Gençoǧlu M, Eren I (2002) An experimental study on the
37. Paulay T (1981) Developments in the seismic design of effect of steel fiber reinforced concrete on the behavior of
reinforced concrete frames in New Zealand. Can J Civ Eng the exterior beam-column joints subjected to reversal cyclic
8(2):91–113 loading. Turk J Eng Environ Sci 26(6):493–502
38. Chun S-C, Shin Y-S (2014) Cyclic testing of exterior beam- 57. Hakuto S, Park R, Tanaka H (2000) Seismic load tests on
column joints with varying joint aspect ratio. ACI Struct J interior and exterior beam-column joints with substandard
111(3):693–704 reinforcing details. ACI Struct J 97(1):11–25
39. Vollum RL, Newman JB (1999) The design of external, 58. Hamil SJ (2000) Reinforced concrete beam-column con-
reinforced concrete beam-column joints. Struct Eng nection behavior. PhD thesis, School of Engineering,
77(23–24):21–27 University of Durham
40. Matamoros AB, Wong KH (2003) Design of simply sup- 59. Hassan WM (2011) Analytical and experimental assess-
ported deep beams using strut-and-tie models. ACI Struct J ment of seismic vulnerability of beam-column joints with-
100(6):704–712 out transverse reinforcement in concrete buildings. PhD
41. Alva GMS, de Cresce El Debs ALH, El Debs MK (2007) An thesis, University of California, Berkeley
experimental study on cyclic behaviour of reinforced con- 60. Hegger J, Sherif A, Roeser W (2003) Nonseismic design of
crete connections. Can J Civ Eng 34(4):565–575 beam-column joints. ACI Struct J 100(5):654–664
42. Bindu KR, Jaya KP (2008) Performance of exterior beam 61. Hwang SJ, Lee HJ, Wang KC (2004) Seismic design and
column joints with cross inclined bars under seismic type detailing of exterior reinforced concrete beam-column
loading. J Eng Appl Sci 7:591–597 joints. In: Proceedings of the 13th world conference on
43. Chalioris CE, Favvata MJ, Karayannis CG (2008) Rein- earthquake engineering, Paper No 03972004
forced concrete beam-column joints with crossed inclined 62. Ibrahim HHA (2011) Stud reinforcement in beam-column
bars under cyclic deformations. Earthq Eng Struct Dynam joints under seismic loads. PhD thesis, University of Cal-
37(6):881–897 gary, Canada
44. Chun SC, Kim DY (2004) Evaluation of mechanical 63. Idayani BS (2007) The influence of concrete strength on the
anchorage of reinforcement by exterior beam-column joint behaviour of external beam column joints. PhD thesis,
experiments. Proceedings of the 13th world conference on Faculty of Civil Engineering, Universiti Teknologi
earthquake engineering Malaysia
45. Chun SC, Lee SH, Kang THK, Oh B, Wallace JW (2007) 64. Karayannis CG, Sirkelis GM (2005) Response of column
Mechanical anchorage in exterior beam-column joints and joints with spiral shear reinforcement. In: Proceedings
subjected to cyclic loading. ACI Struct J 104(1):102–112 of the CMEM conference on computational methods and
46. Chun S-C, Oh B, Lee S-H, Naito CJ (2009) Anchorage experimental measurements
strength and behavior of headed bars in exterior beam-col- 65. Karayannis CG, Sirkelis GM (2008) Strengthening and
umn joints. ACI Struct J 106(5):579 rehabilitation of RC beam-column joints using carbon-FRP
47. Chutarat N, Aboutaha RS (2003) Cyclic response of exterior jacketing and epoxy resin injection. Earthq Eng Struct
reinforced concrete beam-column joints reinforced with Dynam 37(5):769–790
headed bars—experimental investigation. ACI Struct J 66. Kim J, LaFave JM, Song J (2009) Joint shear behaviour of
100(2):259–264 reinforced concrete beam-column connections. Mag Concr
48. Clyde C, Pantelides CP, Reaveley LD (2000) Performance- Res 61(2):119–132
based evaluation of exterior reinforced concrete building 67. Kotsovou G, Mouzakis H (2011) Seismic behaviour of RC
joints for seismic excitation. PEER Report 2000/05 external joints. Mag Concr Res 63(4):247–264
49. Dhakal RP, Pan T-C (2003) Characteristics of high-speed 68. Kuang JS, Wong HF (2006) Effects of beam bar anchorage
cyclic test of beam-column joints. ACI Struct J 100(2):188–196 on beam-column joint behaviour. Proc Inst Civ Eng
50. Durrani AJ, Zerbe HE (1987) Seismic resistance of R/C 159(2):115–124
exterior connections with floor slab. ASCE J Struct Eng 69. Kusuhara F, Shiohara H (2008) Tests of R/C beam-column
113(8):1850–1864 joints with variant boundary conditions and irregular details
3476 Materials and Structures (2016) 49:3459–3476

on anchorage of beam bars. In: Proceedings of the 14th 81. Scott RH (1996) Intrinsic mechanisms in reinforced con-
world conference on earthquake engineering crete beam-column connection behavior. ACI Struct J
70. Lee HJ, Ko JW (2007) Eccentric reinforced concrete beam- 93(3):336–346
column connections subjected to cyclic loading in principal 82. Shiohara H, Kusuhara F (2010) An overlooked failure
directions. ACI Struct J 104(4):459–467 mechanism of reinforced concrete beam-column joints. In:
71. Lehman D, Stanton J, Alire D (2011) Seismic evaluation of Proceedings of the 9th NCEE
beam-column joints in older concrete exterior frames. 83. Teng S, Zhou H (2003) Eccentric reinforced concrete beam-
ASCE J Struct Eng:348 column joints subjected to cyclic loading. ACI Struct J
72. Li B, Pan T, Tran C (2009) Effects of axial compression 100(2):139–148
load and eccentricity on seismic behavior of nonseismically 84. Tsonos AG, Tegos IA, Penelis GG (1992) Seismic resis-
detailed interior beam-wide column joints. ASCE J Struct tance of type 2 exterior beam-column joints reinforced with
Eng 135(7):774–784 inclined bars. ACI Struct J 89(1):3–12
73. Li B, Tran C, Pan T (2009) Experimental and numerical 85. Tsonos A, Tegos I, Penelis GG (1993) Seismic resistance of
investigations on the seismic behavior of lightly reinforced type 2 exterior beam-column joints reinforced with inclined
concrete beam-column joints. ASCE J Struct Eng bars. ACI Struct J 89(1):3–12
135(9):1007–1018 86. Wallace JW, McConnell SW, Gupta P, Cote PA (1998) Use
74. Li B, Kulkarni S (2009) Seismic behavior of reinforced of headed reinforcement in beam-column joints subjected to
concrete exterior wide beam-column joints. ASCE J Struct earthquake loads. ACI Struct J 95(5):590–606
Eng 136(1):26–36 87. Wong HF, Kuang JS (2008) Effects of beam-column depth
75. Lu X, Urukap TH, Li S, Lin F (2012) Seismic behavior of ratio on joint seismic behaviour. Proc Inst Civ Eng
interior RC beam-column joints with additional bars under 161(2):91–101
cyclic loading. Earthq Struct 3(1):37–57 88. Kim J (2007) Joint shear behavior of reinforced concrete
76. Murty C, Rai DC, Bajpai K, Jain SK (2003) Effectiveness of beam-column connections subjected to seismic lateral
reinforcement details in exterior reinforced concrete beam- loading. PhD thesis, University of Illinois at Urbana-
column joints for earthquake resistance. ACI Struct J Champaign, USA
100(2):46–105 89. MATLAB (2011) Version 7.12.0, The MathWorks, Inc.,
77. Pampanin S, Calvi GM, Moratti M (2002) Seismic beha- Natick, MA, USA
viour of RC beam-column joints designed for gravity loads. 90. Hwang S-J, Lee H-J (2002) Strength prediction for dis-
In: Proceedings of the 12th European conference on earth- continuity regions by softened strut-and-tie model. ASCE J
quake engineering Struct Eng 128(12):1519–1526
78. Pantelides CP, Hansen J, Nadauld J, Reaveley LD (2002) 91. Wang G-L, Dai J-G, Teng JG (2012) Shear strength model
Assessment of reinforced concrete building exterior joints for RC beam–column joints under seismic loading. Eng
with substandard details. PEER Report 2002/18 Struct 40:350–360
79. Park S, Mosalam KM (2012) Experimental investigation of 92. European Committee for Standardization (CEN) (2004)
nonductile RC corner beam-column joints with floor slabs. Design of concrete structures—Part 1-1: general rules and
ASCE J Struct Eng 139(1):1–14 rules for buildings. Eurocode 2, BS EN 1992-1-1, BSI
80. Parker DE, Bullman PJM (1997) Shear strength within rein- British Standards, London
forced concrete beam-column joints. Struct Eng 75(4):53–57