Engineering Structures 174 (2018) 792–813

Contents lists available at ScienceDirect

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

Analytical prediction of seismic behavior of RC joints and columns under T

varying axial load
⁎
Javad Shayanfara, Habib Akbarzadeh Bengara, , Azadeh Parvinb
a
Department of Civil Engineering, The University of Mazandaran, Babolsar, Iran
b
Department of Civil and Environmental Engineering, The University of Toledo, Toledo, USA

A R T I C LE I N FO A B S T R A C T

Keywords: In this study, an analytical model for predicting nonlinear behavior of exterior reinforced concrete (RC) beam-
RC beam-column joints column joints under varying axial load was developed. The main focus was given on the assessment of the effect
Analytical model of axial load variations on the response of RC joints and columns. During seismic actions, overturning moments
Joint shear failure are produced by lateral loads that are translated as axial loads in the columns. It leads to compressive axial force
Column shear failure
on one side of the structure along with tensile on the opposite. It can overwhelm nonlinear behavior associated
Axial load variations
with axial, flexural, shear stresses of RC columns and joints. To simulate and evaluate these nonlinearities, a
beam- column joint model consisted of rotational springs were developed. The characteristics of joint spring
could be computed using principle tensile stress-joint rotation relation (pt versus θj) in the joint core depending
on the type of the beam bar anchorage. Therefore, for the joints with various beam bar anchorage details, pt
versus θj relations in the joint core were proposed. A new theoretical methodology was also developed to
consider the effect of the axial load variations in determining characteristics of rotational springs. To assess the
accuracy and reliability of the analytical model, it was compared through experimental data available in the
literature. The results showed that the proposed analytical model could predict the experimental response of
poorly detailed RC beam-column joints under varying or constant axial loads with reasonable precision.
Furthermore, parametric studies were carried out to highlight the overwhelming effect of axial load variations
on RC beam-column joints and columns. The simple analytic procedure would make the model sufficiently
suitable for practical applications.

1. Introduction complexity of implementing a fluctuating axial force during testing. The

Beam-column joint elements are of paramount importance in
seismic performance of RC structures, especially non-seismically de-
signed ones. Columns tend to transfer vertical forces from roof and
stories to foundations, while joints transmit moments and shears of
beams into the columns. Due to seismic loads, such as earthquake
loading or wind, columns and joints are subjected not only to the effects
of gravity loads but also to variable moment, shear and axial loads.
Axial load produced by lateral loads into columns causes compressive
axial force on one side of the structure along with tensile on the op-
posite. Accordingly, columns as well as the joint core would experience
variable axial load levels during lateral loading.
Significant effects of axial load on inelastic behavior of RC joints
were confirmed by experimental studies [1–4]. On the other hand, to
assess the seismic performance of RC joints under varying axial loads
were experimentally investigated by few researchers [5–13], due to the
Experimental results showed that seismic response of RC beam-column
joints, especially poor detailed ones, were considerably influenced by
axial load variations so that a combination of lateral and varying axial
loads resulted in a severe reduction in strength and ductility capacities.
To consider the effects of axial load variations in the calculation of the
RC beam-column joint response, several analytical studies have been
conducted mostly based on hierarchy of strength method [6–13]. In this
method, carrying out a bending moments versus axial force (M-N)
performance domain, consisting of the internal hierarchy of strengths in
an RC beam-column joint along with the lateral demand in accordance
with axial load variations resulted by lateral loads, the joint response
could be evaluated. Although useful information about seismic perfor-
mance of RC joints can be obtained by this approach, it needs an
iterative procedure to develop M-N domain. Accordingly, considering
the fact that for each stage of rotation, θj, M-N domain should be de-
termined, the method seems not to be enough practical to model RC

⁎
Corresponding author.
E-mail address: h.akbarzadeh@umz.ac.ir (H.A. Bengar).

https://doi.org/10.1016/j.engstruct.2018.07.103
Received 24 February 2018; Received in revised form 20 July 2018; Accepted 30 July 2018
0141-0296/ © 2018 Elsevier Ltd. All rights reserved.
J. Shayanfar et al.
joints under varying axial load. It is well known that to perform a
realistic nonlinear analysis, nonlinearities in the joint core should be
considered in the modeling simulation. Several analytical and numer-
ical joint models have been provided in the literature [14–32]. In
general, these models need large computational efforts or special pro-
grams to model elements with various springs. Besides, these models do
not consider the effects of axial load variations and work virtually well
under constant axial load. Consequently, developing a practical model
which considers the effects of the axial load variations would be ne-
cessary for accurate and realistic evaluation of RC beam-column joints.
One of the most severe deficiencies in existing RC structures which
makes them vulnerable to earthquakes might be inadequate column
shear capacity resulting in brittle shear failure, and subsequently, shear-
axial failure in such columns ([33–40]). Experimental studies
([33,34,41–44]) proved that the shear capacity of RC columns would
significantly be overwhelmed by their inelastic flexural deformations.
Accordingly, some models were developed to compute shear capacity of
RC columns as a function of ductility demand ([44–52]). However, it
should be considered that the models in literature are generally suitable
for RC columns with constant axial load. On the other hand, experi-
mental and analytical studies ([34–36,53–64]) on RC columns under
varying axial load confirmed that the seismic response of RC columns in
terms of strength, deformation capacity and stiffness was noticeably
different from those under constant axial load. Therefore, a compre-
hensive model seems to be necessary to predict their shear and flexural
behavior while considers the effects of fluctuation on axial load.
In the present study, an analytical joint model consisting of a ro-
tational spring in the centre of the joint core was proposed in an at-
tempt to practically simulate seismic response of RC beam-columns
joints under varying axial load. To consider the remarkable effects of
the combination of lateral and varying axial loads, a new theoretical
approach was also developed so that the model could be capable of the
calculation of the joint response in each level of the axial load varia-
tions. The nonlinear characteristics of the joint rotational spring were
determined according to principles of mechanics (Mohr circle’s theory)
and a semi-empirical analytical model. Accordingly, through providing
a relatively large database of test specimens, pt versus θj relations de-
pending on the anchorage type of beam bars were proposed.
Furthermore, for calculating shear and flexural capacities of RC col-
umns subjected to varying axial load, a simple procedure was also
proposed based on moment-curvature analysis. Overall, the analytical
beam-column joint model is suitable to be carried out by hand calcu-
lations to predict joint shear capacity with no special software program
requirement and accordingly, can easily be employed in practical ap-
plications.
2. Proposed analytical beam-column joint model
In this section, nonlinear behavior of RC beam-column joint under
combined lateral cyclic and varying axial loads is modeled. Fig. 1 il-
lustrates the mechanics of an exterior RC beam-column joint when
subjected to lateral loading. Lb and Lc are the distance of the critical
section to the point of contra-flexure in beam and storey height, re-
spectively. The other parameters were defined in the figure. During
seismic loading, beam-column joints are subjected to axial and shear
stresses in the joint core. These stresses would lead to principle tensile
and compressive stresses which result in diagonal cracking or concrete
crushing in the joint core. To simulate nonlinearities in the beam-col-
umns joints, several analytical and numerical joint models have been
provided in the literature [14–32]. In general, these models do not seem
to be practical and suitable enough to be used by engineers for pre-
dicting nonlinear behavior of RC beam-column joints, while practical
models are known to be scarce. One of the main reasons is that these
models need large computational efforts or special programs to model
elements with various springs. Besides, these models do not consider
the effects of axial load variations and work virtually well under
793
Engineering Structures 174 (2018) 792–813
constant axial load. In the present study, an analytical beam–column
joint model was developed to simulate the joint mechanism under
varying axial load (Fig. 2). As can be seen, the model includes a non-
linear rotational spring in the joint core to simulate joint post-cracking
shear deformation (Fig. 2(b)). According to a new theoretical approach,
the effects of axial load variations were considered in the calculation of
the joint rotational spring. Finite dimensions of the joint zone were also
taken into account using rigid elements. The relation between pt and θj
due to the joint shear deformation as well as slippage of beam/column
longitudinal bars was assumed to be converted into a moment-rotation
relationship in the joint core, which will be discussed in more details in
the following sections. As shown in Fig. 2(c) and (d), to consider beam
and column nonlinear behavior, rotational springs, adjacent to the joint
core, were also assigned based on lumped plasticity approach [65]. For
a comprehensive nonlinear analysis, distinct behavior i.e. shear and
flexural behavior of beam and column elements should be taken into
account. As in an RC beam/column, the shear capacity tends to reduce
when the inelastic flexural deformation increases. Thus, the definition
of the interaction between flexural and shear capacities is required to
determine the characteristics of beam and column rotational springs.
After which, the total response of the beam-column joint, in terms of
load versus displacement curve, can be calculated by summing up the
displacement contribution of each element for an applied load stage.
While the contribution of each element in joint response is evaluated
depending on the boundary conditions, other elements are assumed to
behave as rigid elements (see Fig. 2).
3. Flexural and shear behavior of beam and column elements
This section addresses the calculation of moment versus rotation
relation of beam/column rotational spring. To determine flexural ca-
pacity of an RC beam/column, the moment-curvature analysis of cross-
section can give useful information. After which, it can be converted
into the flexural moment-rotation relation based on the plastic hinge
method recommended by Priestley et al. [65] to be assigned to the
rotational springs. Accordingly, the rotation, θi, corresponding to cur-
vature, φi, can be calculated as:
for column element (Fig. 2(c))
φi Leff
θi = for φi ⩽ φy
2 (1a)
θi = θy + (φi−φy ) Lp for φi > φy (1b)
for beam element (Fig. 2(d))
φi Leff
θi = for φi ⩽ φy
3 (2a)
⎛ Lp2 ⎞
θi = θy + (φi−φy ) ⎜Lp− for φi > φy
⎝ 2Leff ⎟⎠ (2b)
in which (Paulay and Priestley [66])
Lp = 0.08L + 0.022f y db ⩾ 0.044f y db (Paulay and Priestley[66]) (3)
Leff = L + 0.022f y db (4)
where φy, φu, θy are defined as the yield curvature, ultimate curvature
and yield rotation, respectively; Lsp and Lp are the strain penetration
length and the plastic hinge length, respectively. fy is the yield stress of
longitudinal reinforcement; db is the diameter of the longitudinal re-
inforcements. L is the shear span (distance from maximum moment
section to inflection point). Using the aforementioned equations, the
flexural characteristics of beam/column rotational springs can be ob-
tained.
One of the most severe deficiencies in existing RC structures which
makes them vulnerable to earthquakes might be inadequate column
shear capacity resulting in brittle shear failure, and subsequently, shear-
J. Shayanfar et al.
Cs Cc Tc
Tb
pc
Vjh
pt
Vjv
Tc Cc Cs
Fig. 1. Mechanics of beam-column joint subjected to lateral loading.
axial failure in such columns ([33–40]). Some models were developed
to compute shear capacity of RC columns as a function of ductility
demand ([41–52]). Park et al. [47] came up with a shear-strength de-
gradation model as a function of ductility curvature to predict shear and
deformation capacities of RC members subjected to cyclic lateral
loading. In the current study, to consider the effects of shear in de-
termining the characteristics of beam/column elements, the shear
model developed by Park et al. [48] was followed, along with some
modifications. According to the model, at a location in the member's
cross-section, based on the Rankine’s failure criteria (Chen [67]), the
shear stress capacities can be presented as follows:
νc (z i ) = fcc [fcc −σ (z i )] for εc ⩽ ε0 controlledbycompression
νc (z i ) = 0 for εc > ε0
νt (z i ) = λs ft [ft + σ (z i )] controlled by tension
in which [48,68]
ft′ σ a combination of gravity load and an induced column axial force,
= 1 + 0.5
ft fcc
λs = 1.2−0.2L ⩾ 0.65 [48, 68] (Lin m)
where f′t defines the concrete tensile strength reduced by transverse
compressive stress (σ(zi)). ft is the direct tensile strength of concrete
(equal to 0.292/√f′c, MacGregor et al. [69]). v defines the shear stress at
a compression fiber of the cross-section. fcc is the concrete compressive
strength (peak strength) which was calculated based on Mander et al.
[70]. zi is the distance from the neutral axis; λs is the parameter for
considering the size effect.εc and ε0 are the normal strain at the extreme
compression fiber of the cross-section and the compressive strain cor-
responding to the peak compressive strength. The governing shear
stress capacity, v(yi) can be derived as the smaller value of vc(zi) and
vt(zi) (v(zi) = min [vt(zi) and vc(zi)]). Hence, the concrete contribution in
shear capacity can be determined summing up vc(zi) in all locations of
the compression region of the section. Accordingly, for the each level of
εc, the shear strength of an RC member can be determined by:
Asv f ys ds
Vn = ∑ v (z i) × b × st + s (9)
where st defines the strip thickness in the cross section. Vs is the
transverse reinforcement contribution in the shear strengths [46]. Ast
Engineering Structures 174 (2018) 792–813
N = Ng K Vc
Vc
Vb
ȕ Lc
d hb
Cc Lb bb
Cs
Vc
bc
hc
and fys are the total area and yield stress of the transverse reinforce-
ments, respectively; s is the spacing of the hoops. ds is the depth of
concrete core measured to outside of the transverse reinforcements. To
compare shear and flexural capacities, shear strength-rotation co-
ordinate can be converted into equivalent shear moment-rotation co-
ordinate through multiplying shear strength by shear span.
It is worth noting that for an RC column, the model works if sub-
jected to constant axial load. However, at structural level, the effect of
axial load variations during a seismic event needs to be considered in
the flexural and shear models.
3.1. The effect of applied axial load variations on the column
(5a)
(5b) In this section, flexural and shear behavior of RC columns while
considering the effects of axial load variations are addressed. Due to the
(6) fact that the axial load induced by lateral loading at the structural level
would directly relate to the shear force developed in the column, Vc. In
other words, the total axial load on the column, N, can be assumed to be
(7) ΔN = K × Vc. Accordingly, the pattern of axial load variations during
seismic actions can be regarded as [6–13,64]:
(8)
Mf
⎧ N = Ng + K L For columns with flexural behavior
⎪
N = Ng + KVc = N = N + KV For columns with shear
⎨ g n
⎪ /flexural-shear behavior
⎩
(10)
where Ng is the gravity load value in the column; Mf is the flexural
moment of resistance of the section corresponding to axial load, N; K
defines the axial load factor which can be derived by a preliminary
pushover analysis on frame system (calculating the ratio of column
axial load to column shear force merely due to lateral loading); L is the
distance from maximum moment section to inflection point, equal to
((Lc − hb)/2). Although the most of shear strength models ([41–52]) are
reasonably accurate (more accurate for certain types of columns, less
for others), according to Del Vecchio et al. [40], these models work
relatively well just under constant axial load and in case of varying
axial load, iterative procedures are also needed to be used in calcula-
tions based on these models which might drive them not to be simple
and practical enough. In the present study, a procedure compatible with
moment-curvature analysis was developed to compute flexural or shear
794
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

V c Vc

L' Column element Rigid

Rigid internal
interface plane Rigid element

Beam element Vb
Joint
core
Rigid
Joint rotational spring
Vb Rigid

L'
Column element

Vc
Vc

(b)
(a)

Vc
Mb = Vb Lb Elastic element
Elastic
element Column rotational
spring
Beam rotational
M c = V c L'
Vc spring
Vb Vb
(d)
Vc
M c = V c L'

Elastic Column rotational

element spring

Vc
(c)
Fig. 2. Contributions of components in exterior beam–column joint subassemblage (a) beam-column joint; (b) joint core; (c) column; (d) beam.

capacities under varying axial load in a simple and practical way. It is Mf = ∑ fc (x ) b × st × xi + ∑ Fsi (cp−di) (11)
noteworthy that although N is a function of column flexural or shear
capacities according to Eq. (10), but at the same time, as it is well
known, it also plays a key role in the determination of these capacities 6- Compute shear strength, Vn, by Eq. (9) and the corresponding mo-
as an input parameter. Accordingly, N and the column capacities should ment, Mv = Vn × L
be calculated simultaneously. For this, in this study, a simple procedure 7- Compute total axial load of column, N, as:
was proposed based on moment–curvature analysis: Mf
N = Ng + K for Mf ⩽ Mv
L (12a)
1- Assume a value of concrete strain at the extreme compression fiber,
εc N = Ng + KVn for Mf > Mv (12b)
2- Assume a value of neutral axis depth, c, and the corresponding
curvature, φi. 8- Check the force equilibrium. If ∑ fc (x ) bst + ∑ fsi is close to N ob-
3- Compute total compressive force in concrete, ∑ fc (x ) × b × st . tained from step 7, then the assumed c is correct. Otherwise go to
4- Compute forces at different levels of reinforcement bars∑ Fsi . step 2 and assume a new value of neutral axis depth.
5- Compute the moment of resistance, Mf, by taking the moments of all 9- Repeat steps 1 to 8 for a range of εc.
internal forces about the centroid of the section:

795
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

10- Determine moment–curvature relation. factors. As can be seen in Fig. 3(a), increasing the axial load factor, K,
from −6 to 6, initial stiffness of the column considerably increased and
where xi defines the distance from the centroidal axis and cp is the contrary to this parameter, the column ultimate displacement drama-
distance from the centroid to the extreme compression fiber of the tically decreased. This would confirm that the axial load variations can
cross-section. overwhelm the column response from the point of view of initial stiff-
Since the pattern of axial load of variations will be changed ac- ness and ductility. Based on Fig. 3(b), there is a direct link between
cording to Eq. (10) if failure mode changes, the determination of failure column shear strength, as a function of curvature ductility, and the
mode of an RC member under varying axial load is of paramount im- axial load factor due to the fact that axial load can positively influence
portance. Using the developed procedure, the failure modes, namely, shear strength. It also proves the effect of the axial load variations on
flexural, shear, and flexural-shear failures of the RC column can easily the shear strength of RC column. On the other hand, the maximum
be determined. shear strengths of the column for K equal to −6, 0 and 6 was de-
It should be noted that although the classic M-N interaction diagram termined as 242, 342 and 303 kN. Accordingly, without considering
can be followed to calculate the seismic response of the column sub- axial load variations, the flexural capacity would be on the non-con-
jected to varying axial load, for computing relations such as moment servative side. Increasing the axial load of the column to the balanced
versus curvature/rotation/displacement curves, an M-N interaction point could improve the flexural capacity, since maximum moment of
diagram should be drawn for each curvature level. Then, using lateral the column was positively influenced by the axial load. After the bal-
demand in accordance with axial load variations, flexural and shear ance point, however, the effect of axial load was negative on flexural
capacities can be determined. Accordingly, it is not unreasonable to say moment. Since in the mentioned column, the balance point approxi-
that the proposed procedure is more practical than the classic M-N mately corresponded to the initial axial load, the maximum flexural
interaction diagram, which would correspond with the main focus of capacity virtually occurred at K = 0. As a result, ignoring the effect of
the study. varying axial load (K = 0), the response of the column would lead to
A parametric analysis was performed to highlight the effect of axial misleading predictions in terms of flexural and shear strengths, ductility
load variations on flexural and shear capacities as shown in Fig. 3. For and initial stiffness.
this, a cantilever column with 400 × 400 mm2 cross-sectional dimen-
sions and 1,400 mm height was assumed. It was reinforced with 8No. 9
3.2. Application of column model
(db = 28.8 mm) bars evenly distributed around the column perimeter.
Moreover, 8-mm transverse reinforcements spaced at 150 mm along the
To demonstrate the applicability of the flexural model, the columns
column height were considered to reinforce the column. A constant
tested by Sezen [34] were simulated in terms of load-displacement re-
axial load equal to 0.3f′c Ag which was due to gravity load was assumed
lation. Sezen [34] performed cyclic tests on an RC column under
as the applied load on the column. The compressive strength of concrete
varying axial loading histories. The selected specimen was Specimen 3,
was considered as 20 MPa. The yield stress of steel bars was 420 MPa.
in which the axial load factors, K, at positive and negative directions
Fig. 3 shows the response of the column due to various axial load
were chosen as 5.83 and −4.67, respectively. A constant axial load

375 900
K=6
non-conservative 750 K=0
300
Conservative K = -6
Flexural capacity (kN)

Shear capacity (kN)

600
225
450

150
300
K=6
75 K=0 150 Non-conservative
K = -6
0 0
0 50 100 150 200 0 30 60 90 120 150
Lateral displacement Lateral displacement
(a) (b)
500
K= -6
K= 0
400 K= 6
Non-conservative Shear failure
Ultimate displacement
Shear force (kN)

300

200

Shear behavior
100
Flexural behavior

0
0 10 20 30 40 50 60
Lateral displacement
(c)
Fig. 3. Effect of axial load variation on (a) flexural behavior; (b) shear behavior; (c) load-displacement relation.

796
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

450 1- Compute the beam moment versus the moment arm through a
moment–curvature analysis.
Specimen-3
300 2- Assume a value of βi.
3- Compute Mb using Eq. (18) and corresponding beam moment arm,
βj, using the relation obtained in step 1.
150 4- Iterate the assumed βi until βi ≃ βj.
Shear force (kN)

0
In this study, a simplified procedure was developed to estimate
-120 -80 -40 0 40 80 120 beam moment arm corresponding to maximum joint capacity. Since the
minimum value of β in an RC section generally occurs when beam
-150 tensile bars are yielded, beam moment arm corresponding to yield
Experiment moment would be an appropriate prediction to calculate β. It was es-
with modeling K timated using the proposed semi-empirical analytical model as:
-300
with no modeling K α1 γ1 fc′ bb c y (0.5γ1 c y−d′)
β = d−d′−
-450 Asb f yb (19)
Displacement (mm) in which
Fig. 4. The predicted results of analytical modeling with the experimental re-
cy ρb′ f yb ⎞−1 hb 2 ρb f yb
sults of specimens 3. = 0.443 ⎛⎜1 + ⎟ ⎛ ⎞ + 0.12 < 0.85
d ⎝ fc ⎠ ⎝ d ⎠ fc (20)
equal to 0.25f′c Ag was assumed as the applied load on column due to
εcy × 106
gravity load. Complete details of the columns were given in Sezen [34]. α1 γ1 = (0.006−εcy )
12 (21)
Fig. 4 compares the predicted results through the analytical model with
the test results of Specimen 3, describing the envelope of column load 0.004−0.5εcy
versus displacement relationship. As can be seen, considering the ef- γ1 =
0.006−εcy (22)
fects of axial load variations, although initial lateral stiffness in the pull
direction of loading was not virtually correctly estimated by the ana- where εcy = εsy/(d/c − 1), εsy = the yielding strain of the steel;
lytical procedure, it could predict the load versus displacement re- εcy = the strain at the extreme compression fiber corresponding to the
lationship, especially the maximum strength and displacement values, yield moment; Asb = the total area of the beam reinforcements in ten-
with reasonable accuracy. However, without considering the effect of sion; d′ = the depth to the compression reinforcement; ρb and ρ′b = the
axial load variations (K = 0), the analysis led to misleading results of ratio of the beam longitudinal reinforcements in tension and com-
the column response in terms of strength and ductility. pression (Asb/bb d and A′sb/bb d), respectively. cy = the neutral axis
depth corresponding to the yield moment.
4. Analytical computation of the moment-rotation relation in joint A database of 86 tests on unreinforced exterior beam-column joints
core is presented and summarized in Table 1. Sixty-four tests (joints with
90°-hook) exhibited joint shear failure (with and without beam re-
This part is allocated to determining the joint spring characteristics inforcement yielding); in 14 tests (joints with 180°-hook), development
of exterior RC beam-column joints. Fig. 1 indicates the mechanism of of “concrete wedge” mechanism was observed as critical failure mode;
RC beam-column joints under seismic actions. Using equilibrium in the in 8 tests (joints with a straight anchorage), bond failure occurred be-
joint zone, column shear force, Vc, can be derived as: fore fully activating the diagonal compressive strut. Concrete com-
Vc = Tb−Vjh (13) pressive strengths of the mentioned specimens were in the range
8.05–46.2 MPa. Joint aspect ratios and axial load ratios were in the
where Vjh is the horizontal shear force acting on the joint core; Tb is the range 0.89–2.00 and 0.00–0.44, respectively. In order to derive the
tensile force in beam reinforcements; Using the external equilibrium, Vb experimental joint shear stress, vjhexp, corresponding to maximum beam
as a function of Vc can be determined: shear force, Vbexp, Eqs. (15)–(18) were adopted. Experimental beam
Lc moment arm, βexp, corresponding to Vbexp was calculated using mo-
Vb = Vc ment–curvature analysis when Mbexp = Vbexp × Lb.
Lb + 0.5hc (14)
It should be noted that to estimate βexp, several simplified equations
Substituting Eq. (14) into Eq. (13), and rearranging, the joint spring have been adopted by investigators as shown in Table 2. In the table,
moment was calculated as: the reliability of the proposed semi-empirical analytical model and the
ηLc existing models was compared. As can be seen, with respect to values of
Mj = Vjh
Lc −η (15) the mean and the coefficient of the variation, CoV, the proposed model
was able to predict the beam moment arm in the adjacent of joint core
in which corresponding to maximum joint capacity with reasonable accuracy
h compared to the other models. Furthermore, βexp was relatively well
η = β × ⎛1 + c ⎞ ⎜ ⎟
predicted by Akguzel and Pampanin [13]'s equation which can be fol-
⎝ 2Lb ⎠ (16)
lowed as an alternative for calculating β due to having the advantage of
Mb being very simple.
β=
Tb (17)

Mj Lb 4.1. Determination of pt versus θj relation in joint core

Mb = Vb Lb =
Lb + 0.5hc
where β is the ratio of beam moment to tensile force adjacent to the
joint core which depends on beam nonlinear characteristics. To calcu-
late β corresponding to Vjh, following iterative procedure was adopted:
(18)
It is well known that nonlinear characteristics of the core of exterior
non-seismically detailed beam-column joints highly depend on re-
inforcement detailing. For joints with 90°-hook, after initial shear
cracking in the joint panel, a hardening behavior can be expected due to

797
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Table 1
Database of experimental tests on unreinforced exterior beam-column joints.
Test ID Type bb (mm) hb (mm) bc (mm) hc (mm) f′c (MPa) ALRa ξ (mm) fyb (MPa) ρ b Vbexp (kN) βexp (mm) vjhexp pt / fc′
(MPa)

Clyde et al. [2] 2# Ab-CALe 305 406 305 457 46.2 0.11 2178 454.4 0.021 289.8 301.6 7.54 0.55
4# A-CAL 305 406 305 457 41.0 0.24 2178 454.4 0.021 293.6 298.1 7.74 0.47
5# A-CAL 305 406 305 457 37.0 0.28 2178 454.4 0.021 267.8 292.0 7.24 0.45
6# A-CAL 305 406 305 457 41.1 0.10 2178 454.4 0.021 275.8 298.7 7.26 0.57

Pantelides et al. [3] Unit 1 Bc-CAL 406 406 406 406 33.1 0.10 3914 458.5 0.016 92.1 297.1 2.74 0.18
Unit 1 A-CAL 406 406 406 406 33.1 0.10 3914 458.5 0.016 194.8 295.0 5.83 0.49
Unit 2 B-CAL 406 406 406 406 30.2 0.27 3914 458.5 0.016 125.9 295.6 3.76 0.20
Unit 2 A-CAL 406 406 406 406 30.2 0.27 3914 458.5 0.016 189.9 294.0 5.71 0.37
Unit 3 A-CAL 406 406 406 406 34.0 0.10 3914 458.5 0.016 187.7 295.6 5.61 0.46
Unit 4 A-CAL 406 406 406 406 31.6 0.27 3914 458.5 0.016 211.3 294.0 6.35 0.42
Unit 5 A-CAL 406 406 406 406 31.7 0.10 3914 458.5 0.016 193.9 294.5 5.82 0.51
Unit 6 A-CAL 406 406 406 406 31.0 0.26 3914 458.5 0.016 197.5 294.1 5.93 0.39

Wong [17] BS-L A-CAL 260 450 300 300 30.9 0.15 2600 520.0 0.008 101.0 358.5 3.76 0.22
BS-L-600 A-CAL 260 600 300 300 36.4 0.15 2600 520.0 0.006 133.0 504.3 3.29 0.14
BS-LL A-CAL 260 450 300 300 42.1 0.15 2600 520.0 0.008 127.5 361.1 4.71 0.22
BS-U A-CAL 260 450 300 300 31.0 0.15 2600 520.0 0.008 109.1 358.4 4.06 0.24
BS-L-LS A-CAL 260 450 300 300 31.6 0.15 2600 520.0 0.008 110.0 358.6 4.09 0.24
JA-NN03 A-CAL 260 400 300 300 44.8 0.03 2600 520.0 0.006 81.3 328.2 3.35 0.23

ANA [71] C1 A-CAL 200 300 200 200 15.6 0.07 1078 585.0 0.008 31.3 234.0 2.30 0.24
C2 A-CAL 200 300 200 200 19.0 0.06 1078 585.0 0.008 31.1 235.3 2.27 0.22

Garcia et al. [72] JA A-CAL 260 400 260 260 32.0 0.07 2192 551.0 0.008 57.0 325.5 3.02 0.20
JB A-CAL 260 400 260 260 31.3 0.07 2192 551.0 0.008 58.0 325.0 3.08 0.21
JC A-CAL 260 400 260 260 32.0 0.07 2192 551.0 0.008 54.5 325.5 2.89 0.19

El Amoury [73] T-S1 B-CAL 250 400 250 400 30.6 0.20 2545 477.0 0.012 61.4 318.3 2.82 0.15
T-S1 A-CAL 250 400 250 400 30.8 0.19 2545 477.0 0.012 116.0 329.0 5.13 0.36

Shafaei et al. [25] J3 B-CAL 220 250 250 250 24.7 0.14 1591 460.0 0.008 21.0 195.0 2.01 0.13
J3 A-CAL 220 250 250 250 24.7 0.14 1591 460.0 0.011 41.3 191.5 4.04 0.35

SHR [74] UCI A-CAL 300 450 300 300 25.6 0.08 2269 532.0 0.013 83.2 356.3 3.21 0.25

Melo et al. [75] TPA-1 A-CAL 250 400 250 250 24.2 0.13 2813 405.0 0.003 28.2 334.1 2.23 0.16
TPA-2 A-CAL 250 400 250 250 25.8 0.12 2813 405.0 0.003 29.4 323.4 2.41 0.17
TPB-1 A-CAL 250 400 250 250 15.8 0.20 2813 405.0 0.003 27.75 339.3 2.16 0.19
TPB-2 A-CAL 250 400 250 250 27.3 0.12 2813 405.0 0.003 29.55 318.0 2.47 0.17
TP A-CAL 250 400 250 250 21.5 0.15 2813 465.0 0.003 31.2 343.3 2.39 0.19
TPC C-CAL 250 400 250 250 23.8 0.13 2813 405.0 0.003 27.3 340.4 2.11 0.15

De Risi et al. [76] 1 A-CAL 300 500 300 300 28.8 0.10 3117 487.0 0.008 74.0 417.3 2.82 0.17
2 A-CAL 300 500 300 300 28.8 0.10 3117 459.0 0.003 58.3 447.6 2.04 0.11

YUR [77] EJ-R A-CAL 250 500 250 500 8.1 0.10 2555 292.5 0.010 100.1 392.6 2.07 0.45

Tsonos [78] O1 A-CAL 200 300 200 200 16.0 0.25 1260 485.0 0.010 43.1 228.8 3.47 0.29
O2 A-CAL 200 300 200 200 16.1 0.25 1260 485.0 0.010 43.9 228.8 3.53 0.30

Tsonos [79] O3 A-CAL 200 300 200 200 9.0 0.44 1267 485.0 0.010 35.1 225.7 3.04 0.45

Parvin et al. [80] U.S.2 B-CAL 300 500 300 300 25.0 0.31 1760 420.0 0.004 42.7 412.9 1.45 0.05
U.S.2 A-CAL 300 500 300 300 25.0 0.31 1760 420.0 0.006 82.1 406.7 2.85 0.15
U.S.3 B-CAL 300 500 300 300 25.0 0.31 1760 420.0 0.004 44.8 412.9 1.52 0.05
U.S.3 A-CAL 300 500 300 300 25.0 0.31 1760 420.0 0.006 84.3 406.6 2.92 0.15
U.S.4 B-CAL 300 500 300 300 25.0 0.16 1760 420.0 0.004 37.3 412.9 1.27 0.06
U.S.4 A-CAL 300 500 300 300 25.0 0.16 1760 420.0 0.006 80.0 406.8 2.77 0.19

KAR [81] A1 A-CAL 200 300 200 200 36.4 0.05 1364 574.0 0.003 22.0 254.9 1.75 0.10
A2 A-CAL 200 300 200 200 36.4 0.05 1364 574.0 0.003 21.5 254.9 1.71 0.10

Karayannis et al. [82] A0 A-CAL 200 300 200 200 31.6 0.06 1364 580.0 0.003 23.7 249.3 1.94 0.12
B0 A-CAL 200 300 200 300 31.6 0.04 1304 580.0 0.008 58.3 245.4 3.22 0.30
C0 A-CAL 200 300 200 300 31.6 0.04 1304 580.0 0.008 64.2 244.9 3.55 0.34

Chalioris et al. [83] JA-0 A-CAL 200 300 200 300 34.0 0.03 1304 580.0 0.008 64.7 246.7 3.54 0.33
JCa-0 A-CAL 100 200 100 200 20.6 0.10 1364 580.0 0.008 12.0 160.0 3.31 0.35
JCb-0 A-CAL 100 200 100 200 23.0 0.10 1364 580.0 0.012 15.3 158.8 4.25 0.44

KAR2 [84] JO A-CAL 100 200 100 200 20.8 0.10 1545 525.0 0.008 14.0 164.3 3.81 0.41

Del Vecchio et al. T_C1 A-CAL 300 500 300 300 12.6 0.20 3117 470.0 0.007 62.7 427.6 2.32 0.21
[85] T_C1 A-CAL 300 500 300 300 12.6 0.20 3117 470.0 0.004 52.1 440.7 1.86 0.16
T_C2 A-CAL 300 500 300 300 16.4 0.20 3117 470.0 0.007 80.5 428.7 2.97 0.24
T_C2 A-CAL 300 500 300 300 16.4 0.20 3117 470.0 0.004 65.0 441.1 2.32 0.17
T_C3 A-CAL 300 500 300 300 16.3 0.20 3117 470.0 0.007 82.7 428.5 3.05 0.25
T_C3 A-CAL 300 500 300 300 16.3 0.20 3117 470.0 0.004 69.7 440.9 2.49 0.19

Kaya et al. [86] CS A-CAL 250 500 250 250 13.5 0.40 1855 360.0 0.005 27.8 408.5 1.36 0.07
CS A-CAL 250 500 250 250 13.5 0.40 1855 360.0 0.006 31.3 404.0 1.55 0.09

(continued on next page)

798
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Table 1 (continued)

Test ID Type bb (mm) hb (mm) bc (mm) hc (mm) f′c (MPa) ALRa ξ (mm) fyb (MPa) ρ b Vbexp (kN) βexp (mm) vjhexp pt / fc′
(MPa)

Genesio [87] JT1-1 A-CAL 300 400 350 300 25.4 0.00 2968 554.0 0.007 76.9 324.2 3.68 0.39
JT5-1 A-CAL 300 400 350 300 28.2 0.00 2968 540.0 0.004 51.0 343.5 2.29 0.23
JT2-1 Cd-CAL 300 400 350 300 28.2 0.00 2968 309.0 0.007 41.5 326.4 1.97 0.20
JT3-1 B-CAL 300 400 350 300 27.5 0.00 2968 560.0 0.007 38.0 326.3 1.81 0.18

Helal [88] JA-1 A-CAL 260 400 260 260 22.5 0.10 2157 590.0 0.008 45.1 323.2 2.54 0.19
JA-3 A-CAL 260 400 260 260 31.4 0.12 2157 590.0 0.008 55.7 325.0 3.12 0.18
JB-1 A-CAL 260 400 260 260 28.6 0.08 2157 590.0 0.008 50.8 324.6 2.85 0.20
JC-1 A-CAL 260 400 260 260 28.6 0.08 2157 590.0 0.008 49.6 324.6 2.78 0.19

Liu [89] RC-1 A-CAL 200 330 230 230 19.4 0.07 1849 323.8 0.007 29.7 275.2 2.62 0.29

Kim et al. [90] EN A-CAL 350 480 350 350 20.0 0.00 2874 550.0 0.012 47.5 362.6 2.53 0.35
EN A-CAL 350 480 350 350 20.0 0.00 2874 550.0 0.006 35.0 372.9 1.80 0.25

Ricci et al. [91] 1 C-CAL 300 500 300 300 28.8 0.10 3117 343.7 0.008 79.2 417.3 3.01 0.19
2 C-CAL 300 500 300 300 28.8 0.10 3117 316.1 0.004 52.8 428.8 1.95 0.11

Akguzel and 2DB2 C-VALf 230 330 230 230 17.9 0.21 1849 430.0 0.004 24.8 286.2 1.95 0.15
Pampanin [11] 2DB2 C-VAL 230 330 230 230 17.9 0.05 1849 430.0 0.004 19.5 287.1 1.53 0.18

Pampanin et al. [9] T1 C-VAL 200 330 200 200 23.9 0.13 2184 357.3 0.005 15.9 284.7 1.82 0.12
T1 C-VAL 200 330 200 200 23.9 0.08 2184 357.3 0.005 13.6 284.7 1.56 0.12
T2 C-VAL 200 330 200 200 23.9 0.13 2184 357.3 0.005 16.0 284.7 1.83 0.12
T2 C-VAL 200 330 200 200 23.9 0.08 2184 357.3 0.005 13.8 284.7 1.58 0.12

Kam [10] NS-O1 C-VAL 230 330 230 230 17.3 0.22 1849 362.0 0.004 24.5 277.4 2.00 0.16
NS-O1 C-VAL 230 330 230 230 17.3 0.05 1849 362.0 0.004 19.3 267.4 1.64 0.19
S-O1 C-VAL 230 330 230 230 15.1 0.24 1849 335.0 0.004 21.9 276.1 1.80 0.15
S-O1 C-VAL 230 330 230 230 15.1 0.06 1849 335.0 0.004 18.5 266.3 1.58 0.20

Note: ANA = Antonopoulos and Triantafillou [71]; SHR = Shrestha et al. [74]; YUR = Yurdakul and Avsar [77]; KAR = Karayannis and Sirkelis [81];
KAR2 = Karayannis et al. [84].
a
ALR = Axial load ratio (N/(bc hc f′c)).
b
Joints with 90°-hook.
c
Joints with straight anchorage into the joint core.
d
Joints with 180°-hook.
e
Varying axial load.
f
Constant axial load.

Table 2 these relations are a function of the value of maximum principal tensile
Beam moment arm prediction by various equations. stress, pt, max, in the joint core. It should be noted that joint rotational
Mean CoV β Authors
characteristics were assumed as the sum of the joint shear deformation,
γj, and the rotation due to column and beam bar slip, θslip. Accordingly,
0.94 0.031 0.85 × d Lowes and Altoontash [15] assigning separate slip springs is not required in the beam-column joint
0.97 0.030 0.875 × d Celik and Ellingwood [20] and Jeon et al. [26] model.
0.93 0.057 0.75 × hb Bousselham [92] and Del Vecchio et al. [93]
1.00 0.030 0.9 × d Akguzel and Pampanin [13] and Hassan [94]
In general, to compute pt, max, it was recommended to be just a
0.99 0.057 0.8 × hb Park and Mosalam [95] function of concrete compressive strength. On the contrary, Genesio
0.99 0.016 Eq. (19) Proposed equation [87] suggested the limit states corresponding to pt, max including the
effects of several factors i.e. concrete compressive strength, detailing
and amount of beam longitudinal bars, column axial load and geo-
the fact that in this type of the anchorage of beam longitudinal bars, an metric aspect ratio. Besides, using the existing test results shown in
effective node point can be provided to develop diagonal compression Table. 1, joint shear capacity might not be merely influenced by con-
strut mechanism. For joints with a short embedded length, bond me- crete compressive strength, while the effects of other factors are highly
chanism forms prior to fully developing the diagonal compressive strut, likely to be noticeable as shown in Fig. 6. As can be observed, in-
and consequently, the critical principle tensile stresses corresponding to creasing joint aspect ratio, vjh would remarkably decrease by which the
first diagonal cracking and peak load are strongly influenced. For joints overwhelming effect of this factor can be confirmed. It is because that
with 180°-hook, the first joint diagonal cracking corresponds to severe the angle of the concrete diagonal strut with horizontal becomes higher
damage in the joint core and peak load because of the concrete wedge by increasingα and consequently, the horizontal portion of the diagonal
mechanism. strut, resisting the tensile forces in beam longitudinal bars, will be less.
In the current study, to model failure mechanism of RC joints with Therefore, more compression force in the concrete diagonal strut is
an anchorage type except for 90°-hook, it was assumed that it could be needed to be capable of sustaining the equilibrium in the joint core, for
simulated through determining the principal tensile stress corre- the same tensile force. Accordingly, a reduction in joint shear capacity
sponding to ultimate state (peak load). Some recent studies have de- would occur due to the enhanced demand in the concrete diagonal
veloped relations to determine the value of principal tensile stress for strut. Based on Fig. 6, joint shear strength, vjh, seems to be virtually
each level of joint shear deformation/rotation, depending on detailing variable by increasing hb/bb, nonetheless, the shown trend confirms that
of beam longitudinal bars (Priestley [96], Pampanin [6], Sharma et al. vjh/√f′c might slightly decrease in indirect proportion to this factor.
[22], Akguzel and Pampanin [13], Shayanfar and Akbarzadeh [29], Considering the fact that joint shear failure corresponds to the strut
Shayanfar et al. [30]). In the present study, the curves recommended by failure starting at the anchorage of beam longitudinal bars, it is argu-
Shayanfar et al. [30] with some modifications in terms of joint rota- able that increasing amount of beam longitudinal bars, the concrete
tional characteristics were followed as shown Fig. 5. As can be seen,

799
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

pt pt
pt,max
0.95 pt,max pt,max
0.85 pt,max
1
0.7 pt,max
40 (MPa / Rad)

0.3 pt,max
0.2 pt,max

1.5 5 10 șj >25 șj ×10-3 0.1 15 30 șj ×10-3

(a) (b)
pt

50 (MPa / Rad)
pt,max 1
1
35 (MPa / Rad)
0.7 pt,max
0.6 pt,max
1
12 (MPa / Rad)
0.2 pt,max

1.5 șj ×10-3
(c)
Fig. 5. Proposed principal tensile stress-joint rotation relation for exterior beam-column joints with (a) 90°-hooks (b) 180°-hooks (c) short embedment length.

1.4 1.4
90°-hooks 90°-hooks
1.2 1.2
180°-hooks 180°-hooks
1.0 Straight anchorage 1.0 Straight anchorage
vjh/√f 'c

vjh/√f 'c

0.8 0.8

0.6 0.6

0.4 0.4

0.2 0.2

0.0 0.0
0.75 1.25 1.75 2.25 0.75 1.25 1.75 2.25 2.75

h b / hc h b / bb
1.4 1.4

1.2 1.2

1.0 1.0
vjh/√f 'c

0.8 0.8
vjh/√f 'c

0.6 0.6

0.4 90°-hooks 0.4 90°-hooks

0.2
180°-hooks 0.2
180°-hooks
Straight anchorage Straight anchorage
0.0 0.0
0.00 0.25 0.50 0.75 0.00 0.20 0.40 0.60 0.80

N /Aj×f 'c b fyb /√f 'c

Fig. 6. Variations of joint shear strength to the geometry and material characteristics of experimental beam-column joints;

800
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

diagonal strut would experience the ultimate failure at a higher exterior joint subjected to varying axial load, the vertical shear stress in
strength, which it can be proven by the results shown in Fig. 6. It is the joint core can be calculated as:
caused by a considerable increase in the contribution of tensile forces
αVjh
transferred by bond mechanism along with being more efficiently σ= + fv = αvjh + fv
bj hc (25)
supported by the bending of the beam bars [87]. As can be also seen,
there is a direct relation between joint shear capacity and axial load in which ([97,98]) ([95])
applied on column. It confirms that axial load might play an important
role in nonlinear behavior of RC joint. Accordingly, the variations of KVc
fv = σa +
this parameter could bring about remarkable changes in joint response. bj hc (26)
In light of these discussions and based on the experimental evidence
(Table 1), an equation to calculate pt, max was proposed using regression Vjv = αVjh ([97, 98]) (27)
analysis as:
bb + bc
aωb + c bj = ([95])
pt ,max = 2 (28)
αSetup (23)
where α is joint aspect ratio; fv is axial compressive stress on joint core;
in which σa is axial compressive stress applied on joint due to gravity load (Ng/bj
C hc); Vjv is vertical joint shear force; On the other hand, pt in the joint
C C
h 1 h 2 bj 3
C ρb f yb ⎞ 4 Ldh C5 ′C6
ω = ⎛ b ⎞ ⎛ b ⎞ ⎛ ⎞ ⎜⎛1 +
⎜ ⎟ ⎜ ⎟ ⎜ ⎟
⎟
⎛ ⎞ f
⎜ ⎟ core can be determined using Mohr‘s circle ([98]):
⎝ hc ⎠ ⎝ bb ⎠ ⎝ hc ⎠ ⎝ fc′ ⎠ ⎝ hc−c′ ⎠ c (24)
σ 2 2 σ
where ρb = Asb/bb hb; Asb = the total area of beam reinforcement in pt = ⎛ ⎞ + v jh −
⎝2⎠ 2 (29)
tension; Ldh = the length of the straight anchorage in the joint core.
c′ = the concrete cover of the column, and a, b, c, C1, C2, C3, C4, C5 and where vjh is horizontal shear stress in an element of the joint core.
C6 are the non-dimensional empirical coefficients which are given in Substituting Eq. (25) into Eq. (29), pt can be written:
Table 3. In general, two different setups were followed for testing ex-
2
terior beam-column joints in which (1) a concentrated load (Vb) is σ 2 σ −fv ⎞ σ
pt = ⎛ ⎞ +⎛ −
⎜ ⎟

vertically applied to the beam tip and the columns are hinged at their ⎝2⎠ ⎝ α ⎠ 2 (30)
tip (type 1); (2) the joint is horizontally loaded at the top column tip
which is free to move in the horizontal direction. Furthermore, the Rearranging Eq. (30) leads to:
beam and bottom column are hinged (type 2). Although the static σ 2−(α 2pt + 2fv ) σ + (f v2 −α 2pt2 ) = 0 (31)
systems are quite equivalent in terms of the loading of beam–column
joint, the deformation shapes are quite different. In general, the results Solving the equation for σ and simplifying, vertical shear stress can
extracted from experiment (i.e. pt) are taken into account identical for be written as:
both the test setups. However, based on the finite element study con-
ducted by Genesio [87], nonlinear characteristics of an RC joint could α 2pt αpt f α 2pt α2 f
σ = fv + + α 2 + 4 v + 4 = fv + + αpt 1 + + v
be noticeably affected by test type. Hence, αsetup was proposed to 2 2 pt 2 4 pt
transform results from one test setup to another. In the current study, (32)
αsetup = 1.18 ([87]) was used in converting the characteristics of joints
from test setup 1 to test setup 2 (αsetup = 1 for test type 2). It is note- Substituting Eq. (26) into Eq. (32), we get:
worthy that the boundary condition in the test 1 is closer to moment V
α 2pt σa + K b hc
resisting frame reality in terms of deformation shape. Accordingly, to α2 j c
σ = fv + + αpt 1 + +
use Eq. (23) in determining the nonlinear characteristics of the joint 2 4 pt (33)
core, it should be divided by αsetup = 1.18 at structural level.
Using Eq. (25), horizontal shear force can be written as:
4.2. Derivation of joint shear capacity under varying axial load V
⎛ σa + K b hc ⎞
α α2 j c
Vjh = pt bj hc ⎜ + 1+ + ⎟
This section addresses the calculation of the horizontal shear ⎜2 4 pt ⎟
strength with regard to the effects of axial load variations. An induced ⎝ ⎠ (34)
axial load in column ΔP = KVc is applied to the exterior joints due to
The substitution of Eq. (15) in Eq. (34) gives:
seismic loading. Therefore, the varying axial load in the joint as well as
column expected during the earthquake is explained as a function of the ⎛ ηK
⎞
σa + V
bj hc (Lc − η) jh
lateral load. Accordingly, since joint capacity mainly depends on the α α2
Vjh = pt bj hc ⎜ + 1+ + ⎟
level of axial load applied on the column ([1–4]), the effects of axial ⎜2 4 pt ⎟
load variations should be considered in derivation of joint shear capa- ⎝ ⎠
city. Furthermore, to calculate the characteristics of the joint rotational ⎛α α2 σ ηK ⎞
= pt bj hc ⎜ + 1+ + a + Vjh
spring, pt versus θj relation should also be converted into Mj versus θj 2 4 pt pt bj hc (Lc −η) ⎟
⎝ ⎠ (35)
relation.
According to the joint mechanism during seismic actions, for an in which

Table 3
Empirical coefficients to determine pt,max.
Anchorage C1 C2 C3 C4 C5 C6 a b c

90°-hooks −2.42 −0.13 0.1 3.31 0 0.58 0.25 0.56 0

180°-hooks −1.14 −0.48 0.2 0.30 0 0.00 0.32 1.26 0
Short embedment anchorage −1.17 −0.28 0 0.15 1 0.15 −0.0035 −1.78 0.327

801
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Lc ηLc c = 2(η−Lc )(1 + A1 )−αηK (47c)

Mj = Vb (Lb + 0.5hc ) = Vc (Lb + 0.5hc ) = Vjh → Vc
Lb + 0.5hc Lc −η
η As a result, by solving Eq. (46) for x, parameter m as an input
= Vjh parameter in Eq. (44) can be computed. Therefore, using Eq. (43), joint
Lc −η (36)
spring moment can be calculated as a function of the geometry and
α2 σa ηK
Assuming A1 = + and A2 = , we get: material properties, the axial load factor and the value of pt in the joint
4 pt pt bj hc (Lc − η)
α core corresponding to θj.A concise scheme of the procedure to calculate
Vjh = pt bj hc ( + 1 + A1 + A2 Vjh ) joint spring moment under varying axial load was presented as:
2 (37)
In order to solve the equation for Vjh, binomial expansion can be 1- Assume a value for beam moment arm, βi.
followed as shown in Eq. (38): 2- Calculate pt, max in the joint core (Eq. (49)) and for the obtained pt,
n (n−1) 2 n (n−1)(n−2) 3 max, determine the value of pt and corresponding θj based on the pt
(1 + b)n = 1n + nb + b + b + ... versus θj relation (Fig. 5).
2 3! (38)
3- Calculate m by solving Eq. (46).
It is important to note that binomial expansion is valid at any 4- Calculate joint moment using Eq. (43).
number n, positive or negative and integer or fractional, with |b| < 1. 5- Determine the beam moment arm, βj, corresponding to Mb (Eq. (18))
Therefore, if it is higher than one, the result obtained from binomial based on the developed iterative procedure.
expansion will not be accurate enough. Accordingly, to use binomial 6- Iterate the assumed βi until βi ≃ βj.
expansion, m was defined so that for n = 0.5, b = A1 + A2 Vjh will be 7- Repeat steps 1 to 6 for a range of pt.
equal to 0.001 for each level of column axial load. Hence, the multi- 8- Determine the characteristic of moment–rotation relation of joint
plication of b with m yields: rotational spring.
m
0.5
α m
Vjh = pt bj hc ⎧ + ⎛1 + (A1 + A2 Vjh) ⎞ ⎫ It should be noted that the procedure can be simplified using Eq.
⎨
⎩2 ⎝ m ⎠ ⎬ ⎭
0.5
(19) to calculate the beam moment arm. Fig. 7 shows joint moment-
⎧α 1 (A1 + A2 Vjh) ⎞ ⎫ rotation domain for different values of the axial load factor, K. A con-
= pt bj hc + m ⎜⎛ {1−1} + + ⎟
⎨2 ⎝ m m ⎠ ⎬ (39) stant axial load of 115 kN due to gravity load was assumed as applied
⎩ ⎭
load on the column. Compressive strength of concrete was considered
Rearranging Eq. (39) gives:
17.9 MPa, Furthermore, the yield stress of steel bars was assumed as
1 + (A1 + A2 Vjh)−m ⎞0.5⎫ 430 MPa. As can be seen in Fig. 7, neglecting the effect of varying axial
⎧α
Vjh = pt bj hc + m ⎜⎛1 + ⎟
load, the calculation of the joint spring characteristics led to con-
⎨2 ⎝ m ⎠ ⎬ (40)
⎩ ⎭ servative and non-conservative results at both positive and negative
Using binomial expansion (see Eq. (38)), horizontal shear force can values of K, respectively.
be written as:
4.3. Application of joint model
α 1 + (A1 + A2 Vjh)−m ⎫
Vjh = pt bj hc ⎧ + m (1 + )
⎨
⎩ 2 2m ⎬
⎭ To assess the accuracy of the proposed analytical procedure for
1 + α m + m + A m A 2 Vjh predicting the capacity of RC beam-column joints, the results were
= pt bj hc ⎧ 1
+ ⎫
⎨
⎩ 2 m 2m ⎬
⎭ (41) compared to those extracted from the experiments available in the lit-
erature. Table 4 compares the results obtained from the proposed
Rearranging Eq. (41) gives:
analytical model with those extracted from tests. As can be seen, using
1 + α m + m + A1 the proposed iterative procedure to calculate β, the experimental results
Vjh = × bj hc pt (Lc −η)
2 m (Lc −η)−ηK (42) were predicted more accurately than the simplified model (Eq. (19))
and the models presented in Table 2. Considering the values of the
Substituting Eq. (42) into Eq. (15), and rearranging, joint spring
mean and CoV, it does not seem unreasonable to say that vjh and the pt
moment as a function of axial load factor can be calculated as:
were estimated with acceptable precision. It would also lead to pro-
Mj = X × Lc bj hc pt (43) viding a very good level of accuracy in predicting maximum shear force
applied to the beam tip, even though the failure mode of some speci-
in which
mens was determined by analytical analysis as beam failure.
1 + α m + m + A1 For the verification of the capability of the proposed analytical
X=
2 m ( −1)−K
Lc
η (44)
model in predicting the results of tests conducted under varying axial

As can be seen in Eq. (44), it is a function of the axial load factor. On K=8
the other hand, to correctly follow the binomial expansion (|b| < 1), K=4
the parameter m can be derived based on Eqs. (38), (40) so that b will K=0
be determined equal to 0.001: K = -4
K = -8
1 + (A1 + A2 Vjh)−m
b= = 0.001 → A2 Vjh = 1.001m−1−A1
m (45)
Substituting Eq. (42) into Eq. (45), and rearranging Eq. (46) leads
to:
ax 2−bx −c = 0 → m = x 2 (46)
in which
0
a = 2.002(Lc −η) (47a) Joint rotation (Rad)

b = −2.001ηK (47b) Fig. 7. Variation of joint characteristics with axial load factor of K.

802
 Table 4
Comparison of obtained and experimental joint capacity.
exp vjhanal . 2 ptanal . Vbanal . 1 Vbanal . 2
Test ID βexp (mm) βanal. 1 (mm) βanal. 2 (mm) βanal . 1 βanal . 2 h (MPa) vjh anal. (MPa) ptexp (MPa) ptanal. (MPa) Vbexp (kN) Vb anal. 1 (kN) Vb anal.2 (kN)
. . . . .
βexp βexp vjhexp ptexp Vbexp Vbexp
J. Shayanfar et al.

Clyde et al. [2] 2# 301.6 295.0 295.3 0.98 0.98 7.54 6.73 0.89 3.71 3.21 0.87 289.8 252.1 252.4 0.87
4# 298.1 296.6 293.6 0.99 0.98 7.74 7.92 1.02 3.03 3.13 1.03 293.6 298.3 294.8 1.02 1.00
5# 292.0 295.3 292.2 1.01 1.00 7.24 7.88 1.09 2.71 3.07 1.13 267.8 295.6 292.0 1.10 1.09
6# 298.7 293.9 293.6 0.98 0.98 7.26 6.38 0.88 3.68 3.13 0.85 275.8 237.8 237.5 0.86 0.86

Pantelides et al. [3] Unit 1 297.1 297.1 294.7 1.00 0.99 2.74 2.72 1.00 1.05 1.05 0.99 92.1 91.6 90.8 0.99 0.99
Unit 1 295.0 295.8 294.7 1.00 1.00 5.83 5.12 0.88 2.84 2.42 0.85 194.8 171.5 170.8 0.88 0.88
Unit 2 295.6 295.6 293.3 1.00 0.99 3.76 3.69 0.98 1.08 1.05 0.97 125.9 123.6 122.6 0.98 0.97
Unit 2 294.0 293.4 293.3 1.00 1.00 5.71 6.35 1.11 2.03 2.37 1.17 189.9 210.8 210.8 1.11 1.11
Unit 3 295.6 296.0 295.1 1.00 1.00 5.61 5.17 0.92 2.69 2.43 0.90 187.7 173.3 172.7 0.92 0.92
Unit 4 294.0 293.9 294.0 1.00 1.00 6.35 6.45 1.01 2.34 2.39 1.02 211.3 214.4 214.4 1.01 1.01
Unit 5 294.5 295.3 294.0 1.00 1.00 5.82 5.04 0.87 2.86 2.39 0.84 193.9 168.6 167.8 0.87 0.87
Unit 6 294.1 293.8 293.7 1.00 1.00 5.93 6.29 1.06 2.19 2.38 1.09 197.5 209.2 209.1 1.06 1.06

Wong [17] BS-L 358.5 358.9 355.9 1.00 0.99 3.76 3.61 0.96 1.20 1.13 0.95 101.0 97.2 96.3 0.96 0.95
BS-L-600 504.3 504.8 498.4 1.00 0.99 3.29 3.12 0.95 0.82 0.76 0.93 133.0 126.2 124.2 0.95 0.93
BS-LL 361.1 361.4 360.6 1.00 1.00 4.71 4.27 0.91 1.45 1.26 0.87 127.5 115.8 115.5 0.91 0.91
BS-U 358.4 359.0 356.0 1.00 0.99 4.06 3.62 0.89 1.33 1.13 0.85 109.1 97.4 96.4 0.89 0.88
BS-L-LS 358.6 359.1 356.2 1.00 0.99 4.09 3.65 0.89 1.33 1.14 0.86 110.0 98.4 97.5 0.89 0.89
JA-NN03 328.2 328.5 329.4 1.00 1.00 3.35 3.27 0.98 1.51 1.47 0.97 81.3 79.4 79.7 0.98 0.98

ANA [71] C1 234.0 234.4 229.8 1.00 0.98 2.30 2.40 1.04 0.95 1.00 1.05 31.3 32.7 31.9 1.04 1.02
C2 235.3 235.6 230.8 1.00 0.98 2.27 2.46 1.09 0.94 1.03 1.10 31.1 33.8 32.9 1.09 1.06

Garcia et al. [72] JA 325.5 325.6 321.3 1.00 0.99 3.02 2.99 0.99 1.14 1.13 0.99 57.0 56.4 55.6 0.99 0.97
JB 325.0 325.5 321.1 1.00 0.99 3.08 2.97 0.97 1.17 1.12 0.96 58.0 56.1 55.2 0.97 0.95
JC 325.5 325.6 321.3 1.00 0.99 2.89 2.99 1.03 1.08 1.13 1.04 54.5 56.4 55.6 1.04 1.02

803
El Amoury [73] T-S1 318.3 320.6 318.9 1.01 1.00 2.82 2.92 1.04 0.82 0.87 1.06 61.4 64.2 63.8 1.05 1.04
T-S1 329.0 328.7 318.9 1.00 0.97 5.13 5.28 1.03 2.00 2.09 1.04 116.0 119.3 115.3 1.03 0.99

Shafaei et al. [25] J3 195.0 195.0 194.0 1.00 1.00 2.01 1.99 0.99 0.63 0.62 0.98 21.0 20.8 20.6 0.99 0.98
J3 191.5 197.0 191.6 1.03 1.00 4.04 4.55 1.13 1.71 2.01 1.17 41.3 48.1 46.6 1.17 1.13

SHR [74] UCI 356.3 356.4 353.1 1.00 0.99 3.21 3.25 1.01 1.28 1.29 1.01 83.2 87.3 83.2 1.05 1.00

Melo et al. [75] TPA-1 334.1 333.0 330.8 1.00 0.99 2.23 2.26 1.01 0.79 0.80 1.02 28.2 28.4 28.2 1.01 1.00
TPA-2 323.4 327.5 331.7 1.01 1.03 2.41 2.31 0.96 0.88 0.83 0.94 29.4 28.5 28.9 0.97 0.98
TPB-1 339.3 339.4 325.5 1.00 0.96 2.16 1.99 0.92 0.75 0.67 0.89 27.8 25.6 24.4 0.92 0.88
TPB-2 318.0 322.7 332.6 1.01 1.05 2.47 2.35 0.95 0.91 0.85 0.93 29.6 28.5 29.5 0.97 1.00
TP 343.3 330.8 328.2 0.96 0.96 2.39 2.20 0.92 0.87 0.78 0.89 31.2 27.5 27.3 0.88 0.87
TPC 340.4 332.2 330.6 0.98 0.97 2.11 1.86 0.88 0.73 0.61 0.83 27.3 23.4 23.3 0.86 0.85

De Risi et al. [76] 1 417.3 417.5 413.3 1.00 0.99 2.82 2.88 1.02 0.93 0.96 1.03 74.0 75.9 75.0 1.03 1.01
2 447.6 -a 424.1 – 0.95 2.04 2.56 1.25 0.61 0.82 1.35 58.3 65.7 65.7 1.13 1.13

YUR [77] EJ-R 392.6 392.6 397.5 1.00 1.01 2.07 2.07 1.00 1.28 1.28 1.00 100.1 99.9 101.4 1.00 1.01

Tsonos [78] O1 228.8 229.2 227.0 1.00 0.99 3.47 3.22 0.93 1.16 1.05 0.90 43.1 40.1 39.6 0.93 0.92
O2 228.8 229.2 227.0 1.00 0.99 3.53 3.22 0.91 1.19 1.05 0.88 43.9 40.1 39.6 0.91 0.90

Tsonos [79] O3 225.7 229.2 225.5 1.02 1.00 3.04 3.22 1.06 0.97 1.05 1.08 35.1 37.9 37.2 1.08 1.06

Parvin et al. [80] U.S.2 412.9 413.0 410.6 1.00 0.99 1.45 1.58 1.09 0.24 0.28 1.16 42.7 46.5 46.1 1.09 1.08
U.S.2 406.7 405.7 403.8 1.00 0.99 2.85 3.12 1.10 0.73 0.84 1.15 82.1 89.7 89.1 1.09 1.09
U.S.3 412.9 413.0 410.6 1.00 0.99 1.52 1.58 1.04 0.26 0.28 1.06 44.8 46.5 46.1 1.04 1.03
U.S.3 406.6 405.7 403.8 1.00 0.99 2.92 3.12 1.07 0.76 0.84 1.11 84.3 89.7 89.1 1.06 1.06
U.S.4 412.9 413.0 410.6 1.00 0.99 1.27 1.20 0.94 0.30 0.28 0.91 37.3 35.2 35.0 0.94 0.94
U.S.4 406.8 406.5 403.8 1.00 0.99 2.77 2.49 0.90 0.97 0.84 0.86 80.0 71.9 71.3 0.90 0.89

(continued on next page)

Engineering Structures 174 (2018) 792–813
 Table 4 (continued)
exp vjhanal . 2 ptanal . Vbanal . 1 Vbanal . 2
Test ID βexp (mm) βanal. 1 (mm) βanal. 2 (mm) βanal . 1 βanal . 2 h (MPa) vjh anal. (MPa) ptexp (MPa) ptanal. (MPa) Vbexp (kN) Vb anal. 1 (kN) Vb anal.2 (kN)
. . . . .
βexp βexp vjhexp ptexp Vbexp Vbexp
J. Shayanfar et al.

KAR [81] A1 254.9 – 256.9 – 1.01 1.75 2.72 1.55 0.62 1.07 1.74 22.0 25.5 25.5 1.16 1.16
A2 254.9 – 255.7 – 1.00 1.71 2.72 1.58 0.60 1.07 1.79 21.5 25.5 25.5 1.19 1.19

Karayannis et al. [82] A0 249.3 – 253.5 – 1.02 1.94 2.58 1.33 0.70 1.00 1.43 23.7 25.4 25.4 1.07 1.07
B0 245.4 252.0 242.5 1.03 0.99 3.22 3.71 1.15 1.70 2.00 1.18 58.3 69.4 66.2 1.19 1.14
C0 244.9 – 242.8 – 0.99 3.55 3.67 1.04 1.90 1.98 1.04 64.2 67.8 65.8 1.06 1.02

Chalioris et al. [83] JA-0 246.7 – 243.2 – 0.99 3.54 3.75 1.06 1.90 2.03 1.07 64.7 68.0 67.3 1.05 1.04
JCa-0 160.0 161.4 158.5 1.01 0.99 3.31 3.55 1.07 1.58 1.72 1.09 12.0 13.0 12.7 1.08 1.06
JCb-0 158.8 159.0 157.9 1.00 0.99 4.25 4.17 0.98 2.09 2.04 0.98 15.3 15.0 14.9 0.98 0.98

KAR2 [84] JO 164.3 160.2 159.1 0.98 0.97 3.81 3.47 0.91 1.87 1.67 0.89 14.0 12.4 12.3 0.89 0.88

Del Vecchio et al. [85] T_C1 427.6 427.9 429.0 1.00 1.00 2.32 2.25 0.97 0.75 0.72 0.96 62.7 60.9 61.1 0.97 0.97
T_C1 440.7 440.6 435.2 1.00 0.99 1.86 2.01 1.08 0.56 0.62 1.11 52.1 56.2 55.4 1.08 1.06
T_C2 428.7 429.7 429.0 1.00 1.00 2.97 2.51 0.85 0.96 0.77 0.80 80.5 68.3 68.2 0.85 0.85
T_C2 441.1 441.3 434.6 1.00 0.99 2.32 2.30 0.99 0.69 0.68 0.99 65.0 64.4 63.2 0.99 0.97
T_C3 428.5 429.7 429.0 1.00 1.00 3.05 2.50 0.82 1.00 0.77 0.77 82.7 68.1 68.0 0.82 0.82
T_C3 440.9 441.3 434.6 1.00 0.99 2.49 2.29 0.92 0.76 0.68 0.89 69.7 64.2 63.0 0.92 0.90

Kaya et al. [86] CS 408.5 404.1 402.9 0.99 0.99 1.36 2.00 1.47 0.26 0.48 1.84 27.8 40.3 40.1 1.45 1.44
CS 404.0 404.1 399.1 1.00 0.99 1.55 2.09 1.35 0.32 0.51 1.59 31.3 42.2 41.5 1.35 1.33

Genesio [87] JT1-1 324.2 325.5 320.0 1.00 0.99 3.68 2.36 0.64 1.97 1.26 0.64 76.9 49.4 48.5 0.64 0.63
JT5-1 343.5 331.4 329.2 0.96 0.96 2.29 2.18 0.95 1.23 1.17 0.95 51.0 46.7 46.4 0.92 0.91
JT2-1 326.4 326.5 328.2 1.00 1.01 1.97 1.85 0.94 1.06 0.99 0.94 41.5 39.0 39.2 0.94 0.94
JT3-1 326.3 326.4 320.7 1.00 0.98 1.81 1.60 0.89 0.97 0.86 0.89 38.0 33.7 33.0 0.89 0.87

Helal [88] JA-1 323.2 323.1 317.5 1.00 0.98 2.54 2.80 1.10 0.92 1.04 1.13 45.1 49.6 48.6 1.10 1.08

804
JA-3 325.0 325.0 320.3 1.00 0.99 3.12 3.38 1.08 1.02 1.14 1.11 55.7 60.2 59.2 1.08 1.06
JB-1 324.6 324.6 319.4 1.00 0.98 2.85 2.95 1.03 1.06 1.10 1.04 50.8 52.5 51.5 1.03 1.01
JC-1 324.6 324.6 319.4 1.00 0.98 2.78 2.95 1.06 1.03 1.10 1.07 49.6 52.5 51.5 1.06 1.04

Liu [89] RC-1 275.2 275.7 276.9 1.00 1.01 2.62 1.97 0.75 1.29 0.91 0.71 29.7 22.4 22.5 0.75 0.76

Kim et al. [90] EN 362.2 363.1 363.1 1.00 0.99 2.53 2.20 0.87 1.50 1.36 0.87 47.5 41.3 41.3 0.87 0.87
EN 372.9 372.9 378.0 1.00 0.99 1.80 1.73 0.96 1.07 1.07 0.96 35.0 33.5 34.0 0.96 0.97

Ricci et al. [91] 1 417.3 417.4 421.3 1.00 1.01 3.01 2.10 0.70 1.02 0.63 0.62 79.2 55.1 55.7 0.70 0.70
2 428.8 417.4 427.0 0.97 1.00 1.95 2.07 1.06 0.57 0.62 1.09 52.8 54.4 55.8 1.03 1.06

Akguzel and Pampanin [11] 2DB2 286.2 286.3 282.5 1.00 0.99 1.95 2.04 1.05 0.63 0.67 1.07 24.8 25.9 25.6 1.04 1.03
2DB2 287.1 287.2 282.5 1.00 0.98 1.53 1.43 0.93 0.75 0.67 0.90 19.5 18.2 17.5 0.93 0.90

Pampanin et al. [9] T1 284.7 284.8 284.3 1.00 0.99 1.82 1.80 0.99 0.59 0.58 0.99 15.9 15.7 15.7 0.99 0.99
T1 284.7 284.9 284.3 1.00 0.99 1.56 1.57 1.00 0.58 0.58 1.01 13.6 13.7 13.6 1.00 1.00
T2 284.7 284.8 284.3 1.00 0.99 1.83 1.80 0.98 0.59 0.58 0.98 16.0 15.7 15.7 0.98 0.98
T2 284.7 284.9 284.3 1.00 0.99 1.58 1.57 0.99 0.59 0.58 0.99 13.8 13.7 13.6 0.99 0.99

Kam [10] NS-O1 277.4 277.3 283.9 1.00 1.00 2.00 2.01 1.00 0.65 0.66 1.04 24.5 24.6 25.3 1.00 1.03
NS-O1 267.4 267.5 283.9 1.00 1.04 1.64 1.43 0.87 0.81 0.66 0.83 19.3 16.8 17.4 0.87 0.90
S-O1 276.1 276.0 284.3 1.00 1.01 1.80 1.91 1.06 0.57 0.62 1.12 21.9 23.2 23.9 1.06 1.09
S-O1 266.3 266.4 284.3 1.00 1.05 1.58 1.37 0.86 0.77 0.62 0.83 18.5 16.0 16.7 0.86 0.90

Mean 1.00 0.99 1.00 1.02 0.99 0.98

CoV 0.010 0.016 0.149 0.203 0.119 0.117

Note: ANA = Antonopoulos and Triantafillou [71]; SHR = Shrestha et al. [74]; YUR = Yurdakul and Avsar [77]; KAR = Karayannis and Sirkelis [81]; KAR2 = Karayannis et al. [84].
anal.1 denotes the analytical analyses using bema moment arm calculated using proposed procedure; anal.2 denotes the analytical analyses using bema moment arm calculated using Eq. (19).
a
Failure mode obtained from analytical model was as beam failure.
Engineering Structures 174 (2018) 792–813
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Table 5
Comparison between the responses of test specimens with and without considering axial load factor (by following the proposed principle tensile stress).
Test ID Experimentally determined parameters With considering axial load factor Ignoring axial load factor

a
K ALR vjh (MPa) Vb (kN) ALR vjh (MPa) Vb (kN) ALR vjh (MPa) Vb (kN)

Akguzel and Pampanin [11] 2DB2 4.63 0.21 1.95 24.8 0.21 2.04 25.9 0.12 1.72 22.0
2DB2 −4.63 0.05 1.53 19.5 0.05 1.43 18.2 0.12 1.72 22.0

Pampanin et al. [9] T1 2.44 0.13 1.82 15.9 0.13 1.80 15.7 0.10 1.68 14.7
T1 −2.44 0.08 1.56 13.6 0.08 1.57 13.7 0.10 1.68 14.7
T2 2.44 0.13 1.83 16.0 0.13 1.80 15.7 0.10 1.68 14.7
T2 −2.44 0.08 1.58 13.8 0.08 1.57 13.7 0.10 1.68 14.7

Kam [10] NS-O1 4.63 0.22 2.00 24.5 0.22 2.01 24.6 0.13 1.70 19.9
NS-O1 −4.63 0.05 1.64 19.3 0.06 1.43 16.8 0.13 1.70 19.9
S-O1 4.63 0.24 1.80 21.9 0.25 1.91 23.2 0.14 1.62 18.9
S-O1 −4.63 0.06 1.58 18.5 0.07 1.37 16.0 0.14 1.62 18.9

Mean (Analytical/Experiment) 1.05 0.97 0.97 1.32 0.98 0.97

CoV (Analytical/Experiment) 0.07 0.06 0.06 0.58 0.09 0.10

Note: Mean and CoV for pt,model/pt,test.

a
: ALR = Axial load ratio (N/(bc hc f′c));

Table 6 obtained at any level of loading (Vc,i):

Comparison between the responses of test specimens with and without con-
Δtot = Δb, i + Δc, t , i + Δc, b, i + Δj, i (48)
sidering axial load factor (by following the experimental principle tensile
stress). where Δb,i, Δc,t,i, Δc,b,i and Δj,i are the displacement at the column tip due
Test ID K Vb exp
(kN) Vba (kN) Vbb (kN) to beam, top column, bottom column and joint contributions corre-
sponding to Vc,i which was calculated by Eqs. (49)–(51), respectively:
Akguzel and Pampanin 2DB2 4.63 24.8 24.8 20.92
[11] 2DB2 −4.63 19.5 19.5 23.72 Δb, i = θb, i Lc (49)
Pampanin et al. [9] T1 2.44 15.9 15.9 14.86
−2.44
Lc −hb ⎞
T1 13.6 13.6 14.65 Δc, t , i = Δc, b, i = θc, i ⎛ for θc, i ⩽ θc, y
T2 2.44 16.0 16.0 14.94 ⎝ 3 ⎠ (50a)
T2 −2.44 13.8 13.8 14.86
Lc −hb ⎞ Lc −hb−Lp, i ⎞
Kam [10] NS-O1 4.63 24.5 24.5 19.81 Δc, t , i = Δc, b, i = θc, y ⎛ + θc, p, i ⎛
⎜ ⎟ for θc, i > θc, y
NS-O1 −4.63 19.3 19.3 24.25 ⎝ 3 ⎠ ⎝ 2 ⎠ (50b)
S-O1 4.63 21.9 21.9 17.87
S-O1 −4.63 18.5 18.5 23.13 Δj, i = θj, i Lc (51)
Mean (Analytical/ 1.00 1.02
The total response of the joint is schematically illustrated in Fig. 8.
Experiment)
CoV (Analytical/ 0.00 0.17 Using the proposed analytical model in each component, the load–-
Experiment) displacement as well as failure mode RC beam-column joints were de-
termined. In beam-column joints subjected to a concentrated load at
Note: Mean and CoV for pt,model/pt,test. beam end, the total response of beam-column joint in terms of load-
a
Analytical analysis by considering K. displacement at the beam tip was determined using the proposed pro-
b
Analytical analysis with K = 0.
cedure explained in Appendix A

load, it was applied to specimens tested by Akguzel and Pampanin [11],

Pampanin et al. [9], and Kam [10] as shown in Table 5. As can be seen,
considering the effect of axial load variations, the predicted results (N,
vjh and Vb) were in good agreement with the experimental counterparts.
It is noteworthy that the differences in applied axial loads to the joint
core between the theoretical and experimental results are because of
the difference in pt,max between the theoretical and experimental re-
sults. Accordingly, following the experimental pt, N and vjh as well as Vb
were exactly predicted, as shown in Table 6.
On the contrary, as Tables 5 and 6 show, ignoring the effect of axial
load variations, misleading results were estimated which could be
really unsafe. It is due to the fact that applied axial load to the joint core
was inappropriately estimated.
5. Determining the total response of RC beam-column joint
After determining moment–rotation relation of rotational springs in
both beam and column as well as the joint core corresponding to the
proposed model, the total response of beam-column joint in terms of
calculating load-displacement curves at the column tip can be de-
termined by summing up the displacement contributed by joint core,
beam and column elements. Hence, the total response of the joint was
6. Validation of the proposed analytical model with experiments
To assess the accuracy of the proposed analytical procedure in
predicting the behavior of RC beam-column joints, the results of the
analytical model were compared to the results extracted from experi-
ments by Akguzel and Pampanin [11], Pampanin et al. [9], and Kam
[10]. In each beam-column joint, two nonlinear analyses were per-
formed in which (1) the effects of joint nonlinearity and axial load
variations were considered; (2) the joint core was assumed as rigid
however, the effects of axial load variations were considered when
calculating the characteristics of column rotational springs.
Akguzel and Pampanin [11] conducted a study on exterior joints
with considering the effect of the varying axial load applied to the
column. Lateral load application was applied to the tip of the column.
Beam longitudinal bars with 180°-hooks were anchored into the joint
core. Compressive strength of concrete was considered 17.9 MPa. The
yield stress of steel bars was assumed 430 MPa. A constant axial load of
115 kN due to gravity load was assumed as applied load on the column.
The value of axial load factor in specimen 2DB2 was taken equal to
4.63. Further details can be found in Table 1. Fig. 9 compares the
load–displacement curve of beam-column joint obtained from the
analytical analysis with that reported from the experiment. The results

805
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Vc Joint failure Vc Beam failure

Displacement Displacement
(a) (b)

Vc Column failure

Column contribution
Beam contribution
Joint contribution
Total displacement

Displacement
(c)
Fig. 8. Developed beam–column joint model to determine the joint response with failure mode as (a) joint failure; (b) beam failure; (c) column failure.

45 negative values of K, the same has a descending branch. Increasing

applied axial load on column, tensile strains in column longitudinal
2DB2 30 bars is reduced which delays the column flexural yielding and column
bar yield penetration into the core of the joint as well. Therefore, it can
positively influence the joint shear capacity. On the other hand, the
15
axial load enhances the compression demand in the strut. However, the
strut width would simultaneously increase by the enlargement of the
Vc (kN)

0 compression zone in the column, below and above the core of the joint.
-150 -100 -50 0 50 100 150
This behavior also induces an enhancement in ultimate joint strength.
-15 Experiment Due to the discussed mechanics, at positive values of K, an increase in
the level of column axial load (N > Ni) and at positive values of K
Analytical model
(N < Ni), the joint capacity is improved and reduced, respectively. As
-30 Joint contribution Fig. 10 shows, not considering the relationship between the column
Beam contribution axial load and joint core capacity, the analytical analysis would lead to
-45 conservative results in the ascending branch but non-conservative re-
Column contribution
sults in the descending branch, being unsafe and dangerous. It would
-60 confirm the importance of considering the effects of axial load varia-
Displacement (mm) tions on joint capacity. On the other hand, increasing the value of K,
initially, column capacity rises and a ductile failure for column part can
Fig. 9. Comparison of prediction by the analytical model, with experimental
be expected (tension-controlled column). However, when the applied
results reported by Akguzel and Pampanin [11].
axial load on column reaches balance axial load, Nb, the column ca-
pacity decreases and column failure mode changes to a brittle failure
showed that considering the nonlinear behavior of the joint core and mode (compression-controlled column). The failure mode of beam-
the effects of the axial load variations on the column, the proposed column joint is also dependent on the proportion coefficient value of K.
analytical procedure could accurately simulate the response of the Ignoring the actual variations of the axial load, the expected mechanism
beam-column joint, while, considering the joint core as rigid, the ana- is shear failure in joint core prior to beam and column failure. However,
lysis results would lead to higher capacity than the ones from the ex- at low values of K, a minor change in the sequence of events can be
periment. It should be noted that the application of the step-by-step observed. While, at high values of K, major changes in failure mode can
analytical procedure to determine the response of specimen 2DB2 was be seen. It is important to note that at structural level, joints and col-
illustrated in Appendix B. umns experience both positive and negative values of K depending on
For a more extensive assessment of the effect of the varying axial seismic actions. Hence, ignoring the effect of axial load variations, the
load on nonlinear behavior of beam-column joints, relationships of analysis may lead to the non-conservative side results (due to the ne-
maximum capacity of beam, column and joint core elements versus gative value of K). As a result, the proper understanding of this concept
axial load factor are presented in Fig. 10. As can be seen, at positive seems to be essential in properly assessing the expected seismic per-
values of K, joint core capacity has an ascending branch while at formance of RC beam-column joints. This will translate in decisions

806
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Change in failure mode Ductile behavior Brittle behavior

Column capacity with K=0 Conservative

Non-conservative Non-conservative

Beam capacity
Conservative

Joint capacity with K=0

Non-conservative
Joint capacity
N = Ni N = Nb

Axial load factorK

Fig. 10. Effect of axial load factor K on nonlinear behavior of beam-column joints.

making related to selecting an appropriate design or strengthening
strategy.
Pampanin et al. [9] and Kam [10] carried out experimental studies
on non-ductile exterior joints under varying axial load. Lateral load
application was applied to the tip of the column. In the RC joints tested
by Pampanin et al. [9] and Kam [10], the values of K were 2.44 and
4.63, respectively. Beam longitudinal bars were anchored with 180°-
hooks into the joint zone. Further details can be found in Table 1. In
Fig. 11, load–displacement curves of the joints obtained from the ana-
lytical analyses were compared to the results reported from the
experiments. The results would confirm that considering nonlinear
behavior of the joint core and the effects of applied axial load variations
on the column, the proposed procedure could correctly predict the joint
response. Ignoring nonlinear behavior of the joint core, on the other
hand, the analytical analysis predicts a more misleading capacity than
the ones from the experiment.
To further verification of the proposed beam-column joint model,
load–displacement responses of test specimens under constant axial
load were also compared to the predictions derived from the model.
The details of specimens are available in Table 1. As can be seen in

27 27

T1 T2
18 18

9 9
Vc (kN)
Vc (kN)

0 0
-120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120

-9 -9
Experiment Experiment
Joint model Joint model
-18 Rigid joint -18 Rigid joint

-27
-27
Displacement (mm)
Displacement (mm)
(a) (b)
30 30

NS-O1 S-O1
20 20

10 10
Vc (kN)

Vc (kN)

0
0
-150 -100 -50 0 50 100 150
-150 -100 -50 0 50 100 150

-10
Experiment -10
Joint model
Experiment
Rigid joint Joint model
-20
-20 Rigid joint

-30
Displacement (mm) -30
Displacement (mm)
(c) (d)
Fig. 11. Verification of the proposed model with the test specimen (a) T1 [9], (b) T2 [9], (c) NS-O1 [10], (d) S-O1 [10].

807
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

150 36 36 36

EN TD
100
TPA-2 TPB-2 24 24
24

50 12 12 12
Vc (kN)

Vc (kN)

Vc (kN)

Vc (kN)
0 0 0 0
-210 -140 -70 0 70 140 210 -210 -140 -70 0 70 140 210 -210 -140 -70 0 70 140 210 -210 -140 -70 0 70 140 210

-50 -12 -12 -12

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-100 -24 Rigid joint -24
Rigid joint -24 Rigid joint Rigid joint

-150 -36 -36 -36

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

150 150 150 180

U.S.2 U.S.3 U.S.4 BS-L

100 100 100 120

50 50 50 60

Vb (kN)
Vc (kN)
Vc (kN)

Vc (kN)

0 0 0 0
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150

-50 -50 -50 -60

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-100 -100 Rigid joint
Rigid joint -100 Rigid joint -120 Rigid joint

-150 -150 -150 -180

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

270 180 180 180

BS-L-600 BS-LL BS-U BS-L-LS

180 120 120 120

90 60 60 60
Vb (kN)

Vb (kN)
Vb (kN)

Vb (kN)

0 0 0 0
-120 -80 -40 0 40 80 120 -150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150

-90 -60 -60 -60

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-180 -120 -120 -120 Rigid joint
Rigid joint Rigid joint Rigid joint

-270 -180 -180 -180

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

120 120 120 120

JA-3 JB-1 JC-1

JA-NN03 80 80 80
80

40 40 40 40
Vb (kN)
Vb (kN)
Vb (kN)
Vb (kN)

0 0 0 0
-150 -100 -50 0 50 100 150 -120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120 -120 -80 -40 0 40 80 120

-40 -40 -40 -40

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-80 -80 Rigid joint -80 Rigid joint -80 Rigid joint
Rigid joint

-120 -120 -120 -120

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

120 120 120 90

T_C1 T_C2 T_C3 C1

80 80 80 60

40 40 40 30
Vb (kN)
Vb (kN)
Vb (kN)

Vb (kN)

0 0 0 0
-150 -100 -50 0 50 100 150 -125 -75 -25 25 75 125 -150 -100 -50 0 50 100 150 -120 -80 -40 0 40 80 120

-40 -40 -40 -30

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-80 Rigid joint -80 Rigid joint -80 Rigid joint -60 Rigid joint

-120 -120 -120 -90

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

Fig. 12. Verification of the proposed model with the test specimen.

Fig. 12, a good agreement confirms the reliability of the proposed The proposed model was compared to those from existing five
analytical model to simulate total response of exterior RC beam-column models recommended by Priestley [96], Sharma et al. [22], Akguzel
joints with various anchorages of beam longitudinal reinforcements. On and Pampanin [13], and Sharma [99]. The aforementioned models used
the contrary, ignoring nonlinear behavior of the joint core, the analy- the principle tensile stress in the joint core as damage indicator to
tical analysis would estimate more misleadingly in terms of strength evaluate the nonlinear behavior of poorly detailed exterior beam-
and ductility capacity than the ones from the experiment. column joints. Priestley [96] suggested a value for the limit state of

808
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

90 360 360 360

C2 Unit-1 Unit-2 Unit-3

60 240 240 240

30 120 120 120

Vb (kN)

Vb (kN)

Vb (kN)

Vb (kN)
0 0 0 0
-120 -80 -40 0 40 80 120 -300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 300

-30 -120 -120 -120

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-60 Rigid joint -240 Rigid joint -240 -240 Rigid joint
Rigid joint

-90 -360 -360 -360

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

360 360 360 510

Unit-4 Unit-5 Unit-6 #2

240 240 240 340

120 120 120 170

Vb (kN)

Vb (kN)

Vb (kN)
Vb (kN)

0 0 0 0
-300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 300 -300 -200 -100 0 100 200 300 -210 -140 -70 0 70 140 210

-120 -120 -120 -170

Experiment Experiment Experiment Experiment
Joint model Joint model Joint model Joint model
-240 Rigid joint -240 -240 -340 Rigid joint
Rigid joint Rigid joint

-360 -360 -360 -510

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

510 510 510 150

#4 #5 #6 JC2
340 340 340 100

170 170 170 50

Vb (kN)

Vb (kN)
Vb (kN)
Vb (kN)

0 0 0 0
-210 -140 -70 0 70 140 210 -210 -140 -70 0 70 140 210 -210 -140 -70 0 70 140 210 -180 -120 -60 0 60 120 180

-170 -170 -170 -50

Experiment Experiment Experiment
Experiment
Joint model Joint model Joint model
-340 -340 -340 -100 Joint model
Rigid joint Rigid joint Rigid joint
Rigid joint

-510 -510 -510 -150

Displacement (mm) Displacement (mm) Displacement (mm) Displacement (mm)

Fig. 12. (continued)

ultimate failure of exterior joints with 90°-hook. According to the those computed by various models. As can be seen, with regard to the
model, pt = 0.42/√f′c corresponds to severe damage and diagonal values of mean, CoV and mean absolute percentage error, MAPE, it can
cracking in the joint core. Recently, Sharma et al. [22] followed this be concluded that the proposed analytical model would lead to accurate
value to compute shear capacity of RC joints with 90°-hook. In beam- and uniform estimation of RC exterior joint capacity. For joints with
column joints where beam longitudinal bars terminates in the core with 90°-hook, the model recommended by Priestley [98] and Sharma [99]
a short embedment length, in terms of bond failure mechanism, Sharma would significantly overestimate pt and vjh, although kα seems to be
et al. [22] also recommended pt = 0.19/√f′c using an indirect approach. relatively effective in improving the model by Priestley [98]. For joints
Akguzel and Pampanin [13] proposed the upper limit of pt = 0.2/√f′c with 180°-hook, the model suggested by Akguzel and Pampanin [13]
for exterior joints with plain round bars and 180°-hook details, re- lead to consistent overestimation of the joint capacity, although it can
spectively. Sharma [99] proposed that in order to consider the effect of be virtually regarded as an upper limit for first diagonal cracking in the
joint aspect ratio on the principal tensile stress recommended by joint core. For joints with a short embedment length, adopting the
Priestley [98] and Sharma et al. [22], it could be multiplied by the model recommended by Sharma et al. [22] generally induces a sig-
kα = 1/α. nificant overestimation of pt and vjh, although in this case, kα can be also
Table 7 compares the pt and vjh extracted from experiments and efficient.

Table 7 7. Summary and conclusions

Comparing between the proposed and existing models to predict the char-
acteristics of RC beam-column joints.
A practical model was proposed here for the evaluation of nonlinear
Statistics Priestley [96] Sharma [99] Sharma Akguzel and Proposed behavior of poorly detailed RC beam-column joints under varying axial
et al. Pampanin model
load. According to the model, to simulate nonlinearities in the joint core
[22] [13]
depending the type of beam bar anchorage, a rotational spring along
Principal tensile stress (Analytical/Experiment) with rotational springs for considering beam and column inelastic be-
Mean 1.95 1.36 2.07 1.37 1.02 havior was assigned based on the plastic hinge approach. Based on the
CoV 0.512 0.350 0.582 0.212 0.204 proposed analytical procedure, the total response of an RC beam-
MAPE 98.6% 40.5% 108.0% 36.6% 13.2%
column joint in terms of load versus displacement curve was calculated
Shear horizontal strength (Analytical/Experiment) through summing the displacement contribution of each element at an
Mean 1.65 1.24 1.53 1.25 1.01
applied load stage, with no special software program requirement. At
CoV 0.368 0.250 0.365 0.159 0.148
MAPE 68.5% 28.1% 53.2% 25.2% 9.8% the structural level, axial load on column is varied during seismic
loading. A new methodology was developed to determine the column/

809
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

joint spring characteristics as a function of variable axial load, with no
iterative procedure requirement. The results of the simulation of the
responses of tested RC beam-column joints under varying axial load
confirmed that modeling nonlinear behavior of the joint core and the
effect of axial load variations, the proposed method is capable of pre-
dicting the response of beam-column joints with reasonable accuracy.
However, ignoring nonlinear behavior of the joint core or axial load
variations, the analysis would result in misleading capacity than the
ones obtained from the experiment. The results of the simulation of the
response of tested RC column under varying axial load indicated that
taking into account the effect of axial load variations, the developed
analytical model would correctly predict seismic response of RC
columns while, the analysis might lead to misleading results due to
ignoring this effect.
Parametric analyses were also performed to provide a clear under-
standing of the overwhelming effect of axial load variations on non-
linear behavior of RC beam-column joint. The results showed that from
the point of view of nonlinear behavior of the joint core, ignoring this
effect, the predicted response of joint would be conservative at positive
values of axial load factor, while the same at negative values of axial
load factor would be quite non-conservative and unsafe. For RC col-
umns, neglecting axial load variation effects, the seismic response of RC
column in terms of shear and flexural capacities, initial stiffness and
ductility led to misleading results.

Appendix A

In this section, the proposed analytical procedure to calculation of the total response of RC joints subjected to a concentrated load at beam end
will be addressed. The total response of beam-column joint in terms of load-displacement at the beam tip was determined by summing up dis-
placement contributed by joint core, beam and column elements. Accordingly, the total response of the joint was evaluated at any level of loading,
Vb,i:
Δtot = Δb, i + Δc, t , i + Δc, b, i + Δj, i (A-1)
where Δb,i, Δc,t,i, Δc,b,i and Δj,i are the displacement at the beam tip due to beam, top column, bottom column and joint contributions corresponding to
Vb,i which was computed by Eqs. (A-2)–(A-4), respectively:
2
Δb, i = θb, i Lb for θb, i ⩽ θb, y
3 (A-2a)

2 Lp
Δb, i = θb, y Lb + θb, p, i ⎛Lb− ⎞ for θb, i > θb, y
⎜ ⎟

3 ⎝ 2⎠ (A-2b)

θc, i Lb
Δc, t , i = Δc, b, i =
2 (A-3)

h
Δj, i = θj, i ⎛Lb + c ⎞
⎝ 2⎠ (A-4)
in which:
φb, i Lb
θb, i = for φb, i ⩽ φb, y
2 (A-5a)

θb, i = θb, y + (φb, i−φb, y ) Lp for φb, i > φb, y (A-5b)

Lc −hb ⎞
θc, i = φc, i ⎛ for φc, i ⩽ φc, y
⎝ 3 ⎠ (A-6a)

Lp ⎞
θc, i = θc, y + (φc, i−φc, y ) Lp ⎛1− ⎜ ⎟ for φc, i > φc, y
⎝ Lc −hb ⎠ (A-6b)
where θb,i and θc,i denote the rotation of the beam and column elements; φb,i and φc,i denote the curvature of the beam and column elements; θb,y and
θc,y denote the yield rotation of the beam and column elements; φb,y and φc,y denote the yield curvature of the beam and column elements,
respectively.

Appendix B. Solved example

The specimen 2BD2 tested by Akguzel and Pampanin [11] is an RC exterior beam–column joint, subjected to varying axial load, detailed and built
according to older (pre-1970s) construction practice in New Zealand [100]. Lateral load application was applied to the tip of the column. Beam
longitudinal bars with 180°-hooks were anchored into the joint core. Compressive strength of concrete was considered 17.9 MPa. The yield stress of
steel bars was assumed 430 MPa. A constant axial load of 115 kN due to gravity load was assumed as applied load on the column. The value of axial
load factor in specimen 2DB2 was taken equal to 4.63. Further details can be found in Table 1.
Taking into account the end hooks detailing of beam longitudinal reinforcements, the pt versus θj relation as shown in Fig. 5(b) was adopted. To
calculate the response of the joint, pt corresponding to first joint cracking was considered:
−1.14 −0.48 0.2 0.3
0
ρ f −1.14
330 −0.48 230 0.2 0.0041 × 430 0.3
h
ω = ⎛ b⎞
⎜ ⎟
⎛ hb ⎞
⎜ ⎟
⎛ bj ⎞ ⎜⎛1 + b yb ⎟⎞ f ′ = ⎛ 330 ⎞
⎜ ⎟ ⎛ ⎞ ⎛ ⎞ ⎛1 + ⎞ (17.9)0 = 0.5733
c
⎝ hc ⎠ ⎝ bb ⎠ h
⎝ c⎠ ⎝ f ′
c ⎠ ⎝ 230 ⎠ ⎝ 230 ⎠ ⎝ 230 ⎠ ⎝ 17.9 ⎠

aωb + c 0.32(0.5733)1.26 + 0
pt ,max = fc′ = × 17.9 = 0.569MPa
αSetup 1.18

Regarding the axial load factor K = 4.63 and β = 286.3 mm (this parameter was computed by the proposed iterative procedure, while following

810
J. Shayanfar et al. Engineering Structures 174 (2018) 792–813

Eq. (19), β = 282.5 mm), parameter m can be calculated as:

h 230 ⎞
η = β × ⎛1 + c ⎞ = 286.27 × ⎛1 +
⎜ ⎟ = 309.63mm
⎝ 2Lb ⎠ ⎝ 2 × 1409 ⎠

3.384x 2−2.869x −19.529 = 0 → m = x 2 = (2.863)2 = 8.198

Substituting the obtained values in Eq. (44), parameter X would be:
1 + α m + m + A1 1 + 1.44 × 2.863 + 8.197 + 4.182
X= = = 0.656
2 m ( )−K
Lc
η
−1 2 × 2.863 ( 2000
309.55 )−4.63
−1

and the joint moment:

Mj = XLc bj hc pt = 0.656 × 230 × 230 × 2000 × 0.569 = 39.509 kN·m

Beam moment at the column face corresponding to Mj was determined:

Mj Lb 39.509 × 106 × 1409
Mb = = = 36.53 kN·m
Lb + 0.5hc 1409 + 0.5 × 230
By calculating the moment–curvature analysis in beam section, the moment and tensile force in beam longitudinal bars corresponding to Mb were
36.53 kN·m and 128.83 kN, respectively. The internal forces lever arm in the beam section was 36.53/128.83 × 1000 = 283.55 mm. Since it was
close to β, the computed joint moment seemed to be correct (otherwise, a new value of β would be assumed and iterated). The corresponding θj was
obtained by the relation shown in Fig. 5(b):
θj = 0.0001

and the column shear force:

Mj 39.509 × 106
Vc = = = 19.75 kN·m (Experimental value is18.92 mm)
Lc 2000
The displacement at the column tip due to joint contribution corresponding to Vc is:
Δj = θj × Lc = 0.0001 × 2000 = 0.2 mm

By converting the curvature distribution along the beam element into the displacement at the column tip, Δb due to the beam contribution
corresponding to Vc was calculated as Δb = 9.56 mm. Similarly, Δc due to the top or bottom column contribution corresponding to Vc was computed
as Δc = 2.65 mm. Therefore, the total response of the joint was determined:
Δtot = Δb, i + 2 × Δc + Δj, i = 0.20 + 5.30 + 9.56 = 15.05 mm (The experimental value was17.71 mm)

Using the mentioned procedure at other points of the pt versus θj relation, the response of the joint can be calculated.

Appendix C. Supplementary material

Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.engstruct.2018.07.103.

References [Ph.D. thesis].

[1] Beres A, Pessiki R, White R, Gergely P. Implications of experiments on the seismic
behaviour of gravity load designed RC beam-to-column connections. Earthq Spectra
1996;12:185–98.
[2] Hakuto S, Park R, Tanaka H. Seismic load tests on interior and exterior beam-
column joints with substandard reinforcing details. ACI Struct J 2000;97(1):11–25.
[3] Clyde C, Pantelides CP, Reaveley LD. Performance-based evaluation of exterior
reinforced concrete building joints for seismic excitation. Pacific Earthquake Eng.
Res Center (PEERC); 2000.
[4] Pantelides CP, Hansen J, Nadauld J, Reaveley LD. Assessment of reinforced concrete
building exterior joints with substandard details. Pacific Earthquake Eng. Res Center
(PEERC); 2002.
[5] Tsonos AG, Tegos IA, Penelis GG. Influence of axial force variations on the seismic
behaviour of exterior bean-column joints. J Eur Earthq Eng 1995;3:29–43.
[6] Pampanin S, Moratti M, Calvi GM. Seismic behaviour of RC beam–column joints
designed for gravity loads. In: 12th European conference on earthquake en-
gineering. Paper no. 726. 2002.
[7] Hertanto E. Seismic assessment of pre-1970s reinforced concrete beam-column joint
subassemblies. Christchurch (New Zealand): University of Canterbury; 2006.
[Master of engineering thesis].
[8] Chen T. Retrofit strategy of non-seismically designed frame systems based on a
metallic haunch system. Christchurch (New Zealand): University of Canterbury;
2006. [Master of engineering thesis].
[9] Pampanin S, Bolognini D, Pavese A. Performance-based seismic retrofit strategy for
existing reinforced concrete frame systems using fiber reinforced polymer compo-
sites. J Compos Constr 2007;11(2):211–26.
[10] Selective Kam WY. Weakening and post-tensioning for the seismic retrofit of non-
ductile RC frames. Christchurch (New Zealand): University of Canterbury; 2010.
[11] Akguzel U, Pampanin S. Effects of variation of axial load and bidirectional loading
on seismic performance of GFRP retrofitted reinforced concrete exterior beam–-
column joints. J Compos Constr 2010;14:94–104.
[12] Akguzel U. Seismic performance of FRP retrofitted exterior RC beam-column joints
under varying axial and bidirectional loading. Christchurch (New Zealand):
University of Canterbury; 2011. [Ph.D. thesis].
[13] Akguzel U, Pampanin S. Assessment and design procedure for the seismic retrofit of
reinforced concrete beam-column joints using FRP composite materials. J Compos
Constr 2012;16:21–34.
[14] Pampanin S, Magenes G, Carr A. Modelling of shear hinge mechanism in poorly
detailed RC beam–column joints. In: Proceedings of the FIB 2003 symposium; 2003.
[15] Lowes NL, Altoontash A. Modeling of reinforced-concrete beam–column joints
subjected to cyclic loading. J Struct Eng 2003;129(12):1686–97.
[16] Shin M, LaFave JM. Modeling of cyclic joint shear deformation contributions in RC
beam–column connections to overall frame behavior. Struct Eng Mech
2004;18(5):645–69.
[17] Wong HF. Shear strength and seismic performance of non-seismically designed RC
beam–column joints. Hong Kong University of Science and Technology; 2005.
[Ph.D. thesis].
[18] Mitra N, Lowes NL. Evaluation, calibration, and verification of a reinforced concrete
beam–column joint model. J Struct Eng 2007;133(1):105–20.
[19] Favvata MJ, Izzuddin BA, Karayannis CG. Modelling exterior beam-column joints
for seismic analysis of RC frame structures. J Earthq Eng Struct Dynam
2008;37:1527–48.
[20] Celik OC, Ellingwood BR. Seismic fragilities for non-ductile reinforced concrete
frames–role of aleatoric and epistemic uncertainties. Struct Saf 2010;32(1):1–12.
[21] Niroomandi A, Maheri A, Maheri MR, Mahini SS. Seismic performance of ordinary
RC frames retrofitted at joints by FRP sheets. Eng Struct 2010;32:2326–36.
[22] Sharma A, Eligehausen R, Reddy GR. A new model to simulate joint shear behavior

811
J. Shayanfar et al.
of poorly detailed beam–column connections in RC structures under seismic loads,
part I: exterior joints. Eng Struct 2011;33:1034–51.
[23] Niroomandi A, Najafgholipour MA, Ronagh HR. Numerical investigation of the
affecting parameters on the shear failure of Nonductile RC exterior joints. Eng
Failure Anal 2014;46:62–75.
[24] Favvata MJ, Karayannis CG. Influence of pinching effect of exterior joints on the
seismic behavior of RC frames. Earthq Struct 2014;6(1):89–110.
[25] Shafaei J, Zareian MS, Hosseini A, Marefat MS. Effects of joint flexibility on lateral
response of reinforced concrete frames. Eng Struct 2014;81:412–31.
[26] Jeon JS, Lowes LN, DesRoches R, Brilakis I. Fragility curves for non-ductile re-
inforced concrete frames that exhibit different component response mechanisms.
Eng Struct 2015;85:127–43.
[27] Omidi M, Behnamfar. A numerical model for simulation of RC beam-column con-
nections. Eng Struct 2015;88:51–73.
[28] Borghini A, Gusella F, Vignoli A. Seismic vulnerability of existing R.C. buildings: a
simplified numerical model to analyse the influence of the beam-column joints
collapse. Eng Struct 2016;121:19–29.
[29] Shayanfar J, Akbarzadeh BH. Numerical model to simulate shear behaviour of RC
joints and columns. Comput Concr 2016;18(4):877–901.
[30] Shayanfar J, Akbarzadeh BH, Niroomandi A. A proposed model for predicting
nonlinear behavior of RC joints under seismic loads. Mater Des 2016;95:563–79.
[31] Laterza M, D’Amato M, Gigliotti R. Modeling of gravity-designed RC sub-
assemblages subjected to lateral loads. Eng Struct 2017;130:242–60.
[32] Shayanfar J, Akbarzadeh BH. A practical model for simulating nonlinear behaviour
of FRP strengthened RC beam-column joints. Steel Compos Struct
2018;27(1):49–74.
[33] Lynn A. Seismic evaluation of existing reinforced concrete building columns.
Berkley (CA): University of California at Berkeley; 1999. [Ph.D. thesis].
[34] Sezen H. Seismic behavior and modeling of reinforced concrete building columns.
Berkley (CA): University of California at Berkeley; 2002. [Ph.D. thesis].
[35] Ousalem H, Kabeyasawa T, Tasai A, Ohsugi Y. Experimental study on seismic be-
havior of reinforced concrete columns under constant and variable axial loadings.
In: Proceedings of the annual conference of Japan concrete institute, Tsukuba,
Japan; 2002. p. 229–34.
[36] Shake Elwood K. table tests and analytical studies on the gravity load collapse of
reinforced concrete frames. Berkley (CA): University of California at Berkeley;
2002. [Ph.D. thesis].
[37] Elwood KJ. Modelling failures in existing reinforced concrete columns. Can J Civ
Eng 2004;31(5):846–59.
[38] Elwood KJ, Moehle JP. Axial capacity model for shear-damaged columns. ACI
Struct J 2005;102(4):578–87.
[39] Elwood KJ, Moehle JP. Drift capacity of reinforced concrete columns with light
transverse reinforcement. Earthq Spectra 2005;21(1):71–89.
[40] Del Vecchio C, Del Zoppo M, Di Ludovico M, Verderame GM, Prota A. Comparison
of available shear strength models for non-conforming reinforced concrete columns.
Eng Struct 2017;148:312–27.
[41] Ang BG, Priestley MJN, Paulay T. Seismic shear strength of circular reinforced
concrete columns. ACI Struct J 1989;86(1):45–59.
[42] Ho JCM, Pam HJ. Inelastic design of low-axially loaded high-strength reinforced
concrete columns. Eng Struct 2003;25:1083–96.
[43] Moretti M, Tassios TP. Behaviour of short columns subjected to cyclic shear dis-
placements: experimental results. Eng Struct 2007;29:2018–29.
[44] Aschheim M, Moehle JP. Shear strength and deformability of RC bridge columns
subjected to inelastic cyclic displacements. Rep. no. UCB/EERC-92/04. Berkeley
(CA): Earthquake Engineering Research Center, University of California at Berkeley;
1992.
[45] Priestley MJN, Verma R, Xiao Y. Seismic shear strength of reinforced concrete
columns. J Struct Eng 1994;120(8):2310–29.
[46] Sezen H, Moehle JP. Shear strength model for lightly reinforced concrete columns. J
Struct Eng 2004;130(11):1703–962.
[47] Park H, Choi K, Wight JK. Strain-based shear strength model for slender beams
without web reinforcement. ACI Struct J 2006;103(6):783–93.
[48] Park HG, Yu EJ, Choi KK. Shear-strength degradation model for RC columns sub-
jected to cyclic loading. Eng Struct 2012;34:187–97.
[49] Park HG, Kang S, Choi KK. Analytical model for shear strength of ordinary and
prestressed concrete beams. Eng Struct 2013;46:94–103.
[50] Niroomandi A, Pampanin S, Dhakal R, Soleymani Ashtiani M. Comparison of al-
ternative assessment procedures to predict seismic performance of RC columns. In:
Proceedings of tenth Pacific conference on earthquake engineering, Sydney,
Australia; 2015.
[51] Osorio E, Bairán JM, Marí AR. Analytical modeling of reinforced concrete columns
subjected to bidirectional shear. Eng Struct 2017;138:458–72.
[52] Nouali A, Matallah M. A simplified approach to assess the size effect on the shear-
flexure interaction in RC elements. Eng Struct 2017;144:151–62.
[53] ElMandooh Galal K, Ghobarah A. Flexural and shear hysteretic behaviour of re-
inforced concrete columns with variable axial load. Eng Struct 2003;25:1353–67.
[54] Kreger ME, Linbeck L. Behaviour of RC. Columns subjected to lateral and axial load
reversals. In: Proceedings of third United States national conference on earthquake
engineering; 1986. p. 1475–86.
[55] Li KN, Otani S, Aoyama H. Reinforced concrete columns under varying axial load
and bi-direction horizontal load reversals. In: Proceedings of Pacific conf. on
earthquake engineering, Wairakei, New Zealand, 1987.
[56] Li X, Park R, Tanaka H. Effects of variations in axial load level on the strength and
ductility of reinforced concrete columns. Proceedings of Pacific conference on
earthquake engineering. New Zealand, vol. 1. p. 147–58.
[57] Abrams DP. Influence of axial force variations on flexural behavior of reinforced
812
Engineering Structures 174 (2018) 792–813
concrete columns. ACI Struct J 1987;84(3):246–54.
[58] Saadeghvaziri MA, Foutch DA. Dynamic behavior of R/C highway bridges under the
combined effect of vertical and horizontal earthquake motions. Earthq Eng Struct
Dyn 1991;20(6):535–49.
[59] Esmaeily A, Xiao Y. Behavior of reinforced concrete columns under variable axial
loads. ACI Struct J 2004;101(1):124–32.
[60] Ousalem H, Kabeyasawa T, Tasai A. Evaluation of ultimate deformation capacity at
axial load collapse of reinforced concrete columns. In: Proceedings of 13th world
conference on earthquake engineering, Vancouver (BC), Canada; 2004.
[61] Bechtoula H, Kono S, Watanabe F. Experimental and analytical investigations of
seismic performance of cantilever reinforced concrete columns under varying
transverse and axial loads. J Asian Architect Build Eng 2005;4(2):467–74.
[62] De Stefano M, Mariani V, Tanganelli M, Viti S. The influence of the axial load
variation on the seismic capacity of existing RC buildings. In: Proceedings of 2
European conference on earthquake engineering, Istanbul; 2014.
[63] Rodrigues H, Furtado A, Arêde A. Behavior of rectangular reinforced-concrete
columns under biaxial cyclic loading and variable axial loads. J Struct Eng
2016;142.
[64] Shayanfar J, Bengar HA. Nonlinear analysis of RC frames considering shear beha-
viour of members under varying axial load. Bull Earthq Eng 2017;15(5):2055–78.
[65] Priestley MJNF, SeibleCalvi GM. Seismic design and retrofit of bridges. New York:
John Wiley & Sons; 1996.
[66] Paulay T, Priestley MJN. Seismic design of reinforced concrete and masonry
buildings. New York: John Wiley & Sons; 1992.
[67] Chen WF. Plasticity in reinforced concrete. New York: McGraw-Hill; 1982. p. 474.
[68] Zararis PD, Papadakis GC. Diagonal shear failure and size effect in RC beams
without web reinforcement. J Struct Eng 2001;127:733–42.
[69] MacGregor JG, Sozen MA, Siess CP, Strength and behavior of prestressed concrete
beams with web reinforcement. University of Illinois Civil Engineering Studies,
Structural Research Series 210, Urbana; 1960.
[70] Mander JB, Priestley MJN, Park R. Theoretical stress–strain behavior of confined
concrete. J Struct Eng, ASCE 1988;114(8):1804–26.
[71] Antonopoulos CP, Triantafillou TC. Experimental investigation of FRP strengthened
RC beam-column joints. J Compos Constr 2003;77(1):39–49.
[72] Garcia R, Jemaa Y, Helal Y, Guadagnini M, Pilakoutas K. Seismic strengthening of
severely damaged beam-column RC joints using CFRP. J Compos Constr
2014;18(2):04013048.
[73] El-Amoury T. Seismic rehabilitation of concrete frame beam-column joints. Ontario:
McMaster University; 2004. [Ph.D. thesis].
[74] Shrestha R, Smith ST, Samali B. Strengthening RC beam-column connections with
FRP strips. Proc Inst Civ Eng-Struct Build 2009;162:323–34.
[75] Melo J, Varum H, Rossetto T, Costa A. Cyclic response of RC beam-column joints
reinforced with plain bars: An experimental testing campaign. In: Proceedings of
the 15th world conference on earthquake engineering, Lisbon, Portugal; 2012. p.
24–8.
[76] De Risi MT, Ricci P, Verderame GM, Manfredi G. Experimental assessment of un-
reinforced exterior beam–column joints with deformed bars. Eng Struct
2016;112:215–32.
[77] Yurdakul Ö, Avsar Ö. Strengthening of substandard reinforced concrete beam -
column joints by external post-tension rods. Eng Struct 2016;107:9–22.
[78] Tsonos A-DG. Seismic repair of exterior r/c beam-to-column joints using two-sided
and three-sided jackets. Struct Eng Mech 2002;13(1):17–34.
[79] Tsonos A-DG. An innovative solution for strengthening old R/C structures and for
improving the FRP strengthening method. Struct Monit Maintenance – Int J
2014;1(3):323–38.
[80] Parvin A, Altay S, Yalcin C, Kaya O. CFRP rehabilitation of concrete frame joints
with inadequate shear and anchorage details. ASCE J Compos Constr
2010;14:72–82.
[81] Karayannis CG, Sirkelis GM. Strengthening and rehabilitation of RC beam–column
joints using carbon-FRP jacketing and epoxy resin injection. Earthq Eng Struct Dyn
2008;37(5):769–90.
[82] Karayannis CG, Chalioris CE, Sirkelis GM. Local retrofit of exterior RC beam –
column joints using thin RC jackets – an experimental study. Earthq Eng Struct Dyn
2008;37:727–46.
[83] Chalioris CE, Favvata MJ, Karayannis CG. Reinforced concrete beam-column joints
with crossed inclined bars under cyclic deformations. Earthq Eng Struct Dyn
2008;37:881–97.
[84] Karayannis CG, Chalioris CE, Sideris KK. Effectiveness of RC beam–column con-
nection repair using epoxy resin injections. J Earthq Eng 1998;2(2):217–40.
[85] Del Vecchio C, Di Ludovico M, Balsamo A, Prota A, Manfredi G, Dolce M.
Experimental investigation of exterior RC beam–column joints retrofitted with FRP
systems. ASCE J Compos Constr 2014;18:1–13.
[86] Kaya O, Yalcin C, Parvin, A, Altay S. Repairing of shear-damaged RC joint panel
zone using chemical epoxy injection methodology. In: Proceedings of 14th world
conference on earthquake engineering, Beijing, China; 2008.
[87] Genesio G. Seismic assessment of RC exterior beam–column joints and retrofit with
haunches using post-installed anchors. Stuttgart (German): IWB, University of
Stuttgart; 2012. [Ph.D. thesis].
[88] Helal Y. Seismic strengthening of deficient exterior RC beam-column sub-assem-
blages using post-tensioned metal strips. Sheffield (England): University of
Sheffield; 2012. [Ph.D. thesis].
[89] Liu C. Seismic behaviour of beam–column joint subassemblies reinforced with steel
fibers. Christchurch (New Zealand): University of Canterbury; 2006. [Master of
engineering thesis].
[90] Kim CG, Eom TS, Park HG, Kim TW. Seismic performance of lightly reinforced
concrete beam-column connections for low-rise buildings. J Architect Inst Korea
J. Shayanfar et al.
Struct Constr 2016;32(3):19–30.
[91] Ricci P, De Risi MT, Verderame GM, Manfredi G. Experimental tests of unreinforced
exterior beam-column joints with plain bars. Eng Struct 2016;118:178–94.
[92] Bousselham A. State of research on seismic retrofit of RC beam–column joints with
externally bonded FRP. ASCE J Compos Constr 2010;14:49–61.
[93] Del Vecchio C, Di Ludovico M, Prota A, Manfredi G. Analytical model and design
approach for FRP strengthening of non-conforming RC corner beam– column joints.
Eng Struct 2015;87:8–20.
[94] Hassan WM. Analytical and experimental assessment of seismic vulnerability of
beam–column joints without transverse reinforcement in concrete buildings.
Berkeley (CA): Department of Civil and Environmental Engineering, University of
California; 2011. [Ph.D. thesis].
[95] Park S, Mosalam KM. Parameters for shear strength prediction of exterior beam–-
column joints without transverse reinforcement. Eng Struct 2012;36:198–209.
813
Engineering Structures 174 (2018) 792–813
[96] Priestley MJN. Displacement based seismic assessment of reinforced concrete
buildings. J Earthq Eng 1997;1(1):157–92.
[97] Paulay T, Park R. Joints of reinforced concrete frames designed for earthquake
resistance. Research report 84-9. Christchurch: Department of Civil Engineering.
University of Canterbury; 1984.
[98] Tsonos AG. Cyclic load behavior of reinforced concrete beam-column sub-
assemblages of modern structures. ACI Struct J 2007;104(4):468–78.
[99] Sharma G. Seismic behavior and retrofitting of RC frame structures with emphasis
on beam-column joints—experiments and numerical modeling. Stuttgart (German):
IWB, University of Stuttgart; 2013. [Ph.D. thesis].
[100] NZS95:1955. New Zealand standard - model building by-laws: Part IV and V.
Wellington (NZ): New Zealand Standard Inst.; 1953.