Strength of Beam-column Joint in Soft First Story of RC

Buildings Part 2: Design Equations

Sefatullah Halim , SusumuTakahashi , Toshikatsu Ichinode , Masaomi Teshigawara , Takashi Kamiya,
Hiroshi Fukuyama
Journal of Advanced Concrete Technology, volume 12 ( 2014 ), pp. 146-157

Cross-Sectional Damage Index for RC Beam-Column Members Subjected to Multi-Axial Flexure

Satoshi Tsuchiya, Koichi Maekawa

Journal of Advanced Concrete Technology, volume 4

( 2006 ), pp. 179-192

Nonlinear FEM analysis of RC beam-column joint strengthened by cast in-situ joint expansion

Amorn Pimanmas, Preeda Chaimahawa

Journal of Advanced Concrete Technology, volume 7

( 2009 ), pp. 307-326

Strength of Beam-column Joint in Soft First Story of RC Buildings Part 1: Experiment

Susumu Takahashi, Sefatullah Halim, Toshikatsu Ichinose , Go Kotani, Masaomi Teshigawara,

Takashi Kamiya, Hiroshi Fukuyama
Journal of Advanced Concrete Technology, volume 12 ( 2014 ), pp. 138-145

146

Journal of Advanced Concrete Technology Vol. 12, 146-157, May 2014 / Copyright 2014 Japan Concrete Institute

Scientific paper

Strength of Beam-column Joint in Soft First Story of RC Buildings

Part 2: Design Equations
Sefatullah Halim1, Susumu Takahashi2, Toshikatsu Ichinose3, Masaomi Teshigawara4,
Takashi Kamiya5 and Hiroshi Fukuyama6
Received 26 June 2013, accepted 19 February 2014

doi:10.3151/jact.12. 146

Abstract
Part-1 of this study presents the experimental results of beam-column joints in soft-first story buildings, in which the
first-story column depth is twice that of the upper stories. In Part-2, strut-and-tie models (STMs) are developed for the
specimens. The flow of internal forces shown by the developed strut-and-tie models agreed with the observed cracks.
The developed models for the joints are different from those for the usual joint: the main difference is that, in the
specimens large struts were extended into the wall panel. STMs are also developed for the total frame to understand the
overall equilibrium. Based on the STMs, new design equations are proposed, where the joint strength is evaluated as the
sum of the flexural strengths of the beam and the second-story column. In I-type joint, the effect of the wall panel and
beam stirrups is also considered. In O-type joint, the effect of the hoops in the joint is also considered.

1. Introduction

boundary beam and column.

The strength of the beam-column joint is one of the

more discussed issues in structural engineering (Shiohara 2012). Many researchers proposed models to portray the failure mechanism of the usual beam-column
joints (e.g., Paulay et al. 1978; Shiohara 2001, 2004).
However, the application of these models to the joints in
soft-first story building is difficult because as discussed
in Part-1, the bending moments acting on the joints are
opposite to those on usual joints.
This Part-2 paper includes three topics. The first is
the development of strut-and-tie models (STMs),
(Schlaich et al. 1987) for the specimens described in
Part-1 to understand the flow of the forces in the specimens. The STMs are compared with those for usual
knee joints and exterior tee joints, which are discussed
by Liang (2005) and Sritharan (2005). The second topic
is the development of STMs for the total frame to understand the overall equilibrium. The third topic focuses
on the development of new design equations to compute
the strength of such joints considering the effects of the

2. Overview of test results of O-1 and I-1

specimens
Two specimens are reviewed here: Specimens O-1 and
I-1 with first-story column extended toward outside and
inside of the frame, respectively. The black and red lines
in Fig. 1 show the envelopes of the observed loaddisplacement relationship of Specimen O-1 and Specimen I-1, respectively. The purple lines show the analytical strength assuming that the beam and the joint panel
have infinite rigidity. Figure 1 shows that the observed
strengths of the specimens in the joint opening direction
are less than two-third of the analytical strengths. The
analytical strength is based on the flexural capacity of
first-story column as discussed in Part-1. In the closing
direction, the observed strength of Specimen O-1 is almost equal to the analytical strength, whereas the observed strength of Specimen I-1 is smaller. In the following sections STMs are developed to understand the
behavior and failure mechanisms of the specimens.
800

Specimen O-1

Specimen I-1
Analytical Strength
Maximum Strength

400

Lateral Load (kN)

Ph.D Student, Dept. of Architectural Engineering,

Nagoya Institute of Technology, Nagoya, Japan.
E-mail:sefat_halim@yahoo.com
2
Assistant Professor, Dept. of Architectural Engineering,
Nagoya Institute of Technology, Nogoya, Japan.
3
Professor, Dept. of Architectural Engineering, Nagoya
Institute of Technology, Nogoya, Japan.
4
Professor, Department of Architecture, Nagoya
University, Nogoya, Japan.
5
Chief Research Engineer, Yahagi Construction Co.,
Ltd., Nogoya, Japan.
6
Director, Dept. of Structural Engineering, Building
Research Institute, Tsukuba, Japan.

Closing

436 kN

436 kN

Opening
Specimen I-1 Specimen O-1

-400
-800

-950 kN

-1200
-40

Closing direction Opening direction

-1065 kN
-30
-20
-10
0
10
20
30
40
Defrmation (mm)

Fig. 1 Load-Deformation relationship.

147

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

3. Strut-and-tie analysis of specimen O-1 to

opening load
Figure 2a shows the STM for O-1 specimen. Solid blue
lines indicate the bars in the tensile yielding, while
dashed lines indicate below the yield point. Red zones
indicate struts in compression.
The STM for the opening direction was determined
assuming the following:
(1) Three layers of the column main bars (TAD in Fig.
2a and 9-D19 in Fig. 2c) resist the tension1. The
beam bottom bars, the vertical bars in the wall
panel, and the stirrups also resist the tension. Each
tie is located at the centroid of the reinforcement.
(2) The effective compression strength of concrete to
determine depth of strut is assumed to be 85% of
the concrete strength (Yun and Ramirez 1996).
Struts width is determined as follows:

Wstrut =

Cstrut
0.85 f c ' b

depth of the first-story column was considered effective.

In the STMs, node B is located at the outermost point of
the hooked beam bottom bars (Fig. 2d) resulting in a
smaller distance between the compressive strut and the
tensile reinforcement. This is equivalent to the reduced
effective depth of the first-story column. The other reason is the location of the critical section. In the conventional analysis, the critical section is assumed at the
level of the beam bottom face. In the STM, node B is
located at the level of the beam bottom reinforcement,
which is equivalent to a longer shear span (M/Q) than
usual.
The anchorage details of the outer corner of the specimens are similar to the usual knee joint. The observed
cracks in Fig. 2a are also similar to those in knee joint
(Wight and MacGregor 2012). To compare O-type joint
with knee joints from the view of the force-resistance
mechanism, a STM is constructed for a knee joint as
shown in Fig. 4. The procedure and assumptions to con-

(1)

where Cstrut is the compression force of a strut, fc' is the

concrete strength, and b is the width of each element
that the strut goes through. Because the thickness of the
wall panel was 100 mm and 1/4 of the width of the
beam (400mm), the width of each strut widens by four
times at the boundary of the beam and the wall panel.
Similar changes also occur at the boundary of the beam
and the column.
(3) Node B is located at the centroid of the beam bottom bars such that the outer edge of the nodal zone
coincides with the edge of the outermost hooked
beam bottom bars (Fig. 2d). Node C is located at
the re-entrant corner considering the geometric restrictions. The inclinations of the struts KJ, LI, and
EH are simply assumed as 45 because the effects
of the inclinations on the strength are negligible.
(4) The compressive force of the column reinforcement is neglected because it is located outside node
B.
(5) The stresses in the reinforcement are determined
within the yield strength so that the lateral load capacity is maximized.
(6) The effect of the hoop reinforcement is neglected
in the joint because the hoop reinforcement ratio is
very small (0.17%).
As shown in Fig. 2a, the strut distribution agreed
with the crack pattern observed in the test. The load
carrying capacity based on STM (Q) agreed with the
test result with error less than 10% (Fig. 3). This
strength is closer to the observed strength than that of
the computed capacity based on conventional methods
(Fig. 1). There are two reasons for this. The main reason
is that in the conventional analysis method, the full

Fig. 2 Strut and tie model for O-1. (Opening).

If we assume two or four layers instead of three, we obtain a

smaller horizontal force Q.

Fig. 3 Observed strengths vs. STM.

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

Fig. 4 Strut and tie model of knee joint. (Opening).

struct the STM are the same as before and the reinforcement details are similar to those of the test specimens. The biggest difference is that the wall panel and
the second-story column are removed.
The strength of the knee joint is approximately threefourth of that of the test specimen. The capacity of the
knee joint is governed by beam flexural capacity. The
difference of the capacities in the opening load is attributable to the difference of the location of node C. Note
that strut BC in Fig. 2a is less inclined than that in Fig.
4. This leads to a larger compressive force in strut AB
and a larger shear force in the column. Moreover, note
that the inclination of strut CD in Fig. 2a is opposite to
that of strut AB, indicating that the sign of the shear
force in the second-story column is opposite to that of
the first-story column. The cracks observed in the test
also confirm this tendency. The negative shear force is
compensated by strut HE in the wall panel (Fig. 2a).

4. Strut-and-tie analysis of specimen I-1 to

opening load
Figure 5a shows the STM for Specimen I-1. This model
is developed based on the same assumptions made for
Specimen O-1, except that the first and third assump-

148

tions are modified as follows:

(1') Ties TAK and TAD in Fig. 5a represent 9-D19 and
2-D19 in Fig. 5d, respectively.
(3') Inclination of struts LN and JO is 45. Strut KM
is steeper (51) considering the prominent crack appeared in the test.
Node K is located at the centroid of the two layers of
the main bars of the first-story column such that the
upper edge of the nodal zone coincides with the upper
edge of the hooked bars (Fig. 5e).
As shown in Fig. 5a, the strut distribution agreed
with the crack pattern appeared in the test. The computed tensile forces in the ties agreed with the observed
values. For example, in the test the beam bottom bars
yielded and the observed strain of the first-story column
main bars was about 90% of its yield strain as was computed in STM (Fig. 5a). The strength based on the STM
agreed with the observed strength with an error less than
5% (Fig. 3). Recall that the observed strength is smaller
than that obtained by the conventional analysis (Fig. 1).
The reason for this decrease is the different failure
mode: the beam failure of Specimen I-1 was associated
with the yielding of the beam reinforcement (tie BM)
and stirrup (tie ML).
Figure 5c shows the detail of node B. the depth of
strut AC, wAC, is narrower than the gap between the
column edge and the anchorage point; instead, it is determined by the equilibrium of the axial and shear
forces and the tensile forces in the ties. If the axial force
of the first story column is in tension and extremely
large, wAC may disappear.
If we carefully examine Fig. 5a, we find that the flexural cracks in the second story disappear around node D,
which contradicts the STM showing the constant tensile
stress in tie AD. This contradiction is attributable to the
negligence of the tensile stress of concrete, which would
gradually bend strut BC to the left absorbing the tensile
stress of tie AD.
Figure 6 shows an STM for an exterior joint. Procedure, assumptions, and reinforcement details are similar
to those for Specimen I-1. The beam axial force is assumed to be zero.
The strength of the exterior joint based on STM (230
kN) is approximately 55% of the specimens strength
(425 kN). This difference is attributable to the location
and orientation of the strut and tie in the second-story
column. In Fig. 5a, tie AD is located to the left of strut
BC, producing a bending moment opposite to that of
strut ED and tie BC in the usual exterior joint (Fig. 6).

5. Strut-and-tie analysis of specimen O-1 to

closing load

Fig. 5 Strut and tie model for I-1. (Opening).

Figure 7a shows the skeleton of the STM of Specimen

O-1 for the closing direction. Among the six assumptions for the opening direction, the first, third, and
fourth are changed as follows:
(1') The three layers of the column main bars (TAC

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

149

Fig. 8 Strut and tie model for knee joint. (Closing).

Fig. 6 Strut and tie model for exterior joint. (Opening).

Fig.7 Strut and tie model for O-1. (Closing).

and TAD in Fig. 7d) resist the tension. The beam top bars
(TCH) also resist the tension.
(3') The locations of nodes B, I, F, and H are determined so that the strength is maximized.

(4') The first layer of the first-story column main bars

resists the compression (CAE in Fig. 7d) because it is
continuous throughout the first- and second-story columns.
Figure 7b shows the STM with the compression
zones, which agrees with the cracks observed in the test.
The computed tensile forces in the ties also agree with
the observed strains. The beam top bar (tie CH) yielded
as it was observed in the test. The stresses in the firststory column bars (tie BC) were 95% of the yield
strength, as was observed (Fig. 7e). The strength based
on STM is 88% of the observed strength (Fig. 3). The
STM and the cracks appeared in the test show that a
large amount of forces is transferred into the wall panel.
To investigate the effect of the wall panel, an STM is
developed for a knee joint, as shown in Fig. 8. Virtually,
no axial force is applied both in the opening and closing
directions, except those required for the equilibrium.
The axial force of 2,250 kN (see Part-1), which was
applied in the closing direction, cannot be applied to the
knee joint because the beam cannot resist a shear force
of 2,250 kN. The orientation of struts IH and IF in Fig.
7 is different from that of IH and ID in Fig. 8; this difference can be attributed to the presence of the secondstory column and the wall panel in the test specimen. In
Fig. 7, most of the shear force is transferred to the shear
wall panel, which leads to a larger capacity of the test
specimen than the usual knee joint.
On the other hand, there are important similarities between the test specimens and the knee joint. In the opening load, as discussed above, node B (Fig. 2d) is located
inside the hook of the beam bars, limiting the capacity
of the first-story column, as will be discussed later. In
the closing load, node C (Fig. 7b) is resisted by the tensile forces of the column and beam bars, which indicates
that beam top bars in O-type joint should be anchored,
according to the provisions for the knee joint making
the tail length long enough (O-1t, Fig. 5 of Part-1).

6. Strut-and-tie analysis of specimen I-1 to

closing load
Figure 9a shows the skeleton of an STM of I-1 specimen for the closing load. The STM is developed similar
to that of Specimen O-1, except that the first layer of the

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

150

first-story column main bars does not resist the compression force because it terminates in the joint panel.
Figure 9b shows the STM with compression zones. The
strut distribution agrees with the crack patterns observed
in the test. The computed forces in the ties also agree
with the observed strains (e.g., the beam top bars as well
as the first-story column main bars yielded as was observed). The strength based on STM is 80% of the observed strength (Fig. 3). The STM as well as the observed cracks show that a large shear force is transferred
to the wall panel in the second story. As concluded in
paper-1 the additional beam top bars provided in I-1t
prevented the compressive failure of the wall panel.
This phenomenon can also be discussed in term of the
strut and tie model (Fig. 9b). Increasing tensile force in
tie TCH would result in larger forces in strut IC and IH,
and smaller compressive force in strut IE. Similarly,
considerable amount of vertical force can be resisted by
the beam stirrups making a truss model along the beam
length which is similar to Fig. 5a but upside down. The
truss is not shown in Fig. 9b because the effect in specimen I-1 is much smaller than those of the other struts.
Figure 10 shows the STM for the closing load of the
exterior joint. The procedure and assumptions are similar to those for Specimen I-1. The strength of the exterior joint (110kN) is approximately 15% of the specimens strength (755 kN). The strength in Fig. 10 is governed by the flexural capacity of the beam. This difference is attributable to the following reasons:
(a) Struts IH and IE in Fig. 9 are much larger than
struts BE and BF in Fig. 10 because they extend the
forces into the wall panel.
(b) The vertical tie BD in Fig. 9 is located on the right
of the strut CD, producing a bending moment opposite to that of the usual joint (Fig. 10).

7. Strut-and-tie model for a prototype frame

ASTM is developed for the frame of a full-scale structure to understand the distribution of the inner forces in
a real structure under applied external loads. The building structure is assumed to be a five-story single bay
apartment building (Fig. 11) with a span length of 12m
in both directions and gravity load of 12.5 kN/m2, which
leads to a total gravity load of 12.5 5 12 12 =
9000 kN. For simplicity, the load is applied as concentrated load at two points to depict its distribution and
ease constructing the strut and tie model. The resultant
force of the two point loads acts at the center of gravity.
The first-story columns are extended toward outside
similar to that for Specimen O-1. The dimensions of the
beams, columns, and wall panels are twice of those of
the specimen. The reinforcement ratio and details are
the same as those of the specimen. Recalling that the
computed strengths of Specimen O-1 to the opening and
closing loads were 400 kN and 940 kN, respectively, the
earthquake force is assumed to be (400 + 940) 4 =
5360 kN and is applied at two-thirds of the building

Fig. 9 Strut and tie model for I-1. (Closing).

Fig. 10 Strut and tie model for exterior joint. (Closing).

height. The assumed vertical and horizontal forces cause

the axial forces in the right and left columns in the first
story to be 0 and 9000 kN, which are four times the ax-

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

ial forces applied to Specimen O-1 during the opening

(0 kN) and closing load (2250 kN), respectively.
The red zones and the blue lines in Fig. 11 represent
the compressive and tensile forces in the frame. There is
one difference between Figs. 11 and 2. Strut AH in Fig.
11 is supported by tie AG, which represents the tensile
forces in the vertical reinforcement in the wall panel and
the stirrups in the beam. Strut HE in Fig. 2 is supported
by tie HG, which represents the tensile forces in the longitudinal reinforcement in the vertical stub. Otherwise,
the stress distribution near the opening joint in Fig. 11 is
identical with that in Fig. 2. Furthermore, there are no
differences between Fig. 7 and Fig. 11 for the closing
side. These similarities indicate that the vertical stub
constructed for the test specimen represents the rest of
the frame if the wall panel of the actual frame is provided with appropriate amount of vertical reinforcement,
which is indicated by ties AG and BH in Fig. 11. The
required quantity is discussed by Shobu et al (2012).If
we assume a frame taller than that the in Fig. 11, the
location of the horizontal force moves upward causing
tensile axial force in the column. This will reduce the
amount of the compressive strut in the tensile column
but will not change the strut and tie system drastically.
Similar strut-and-tie model can be obtained for I-type
frame.
The diagrams at the bottom of Fig. 11 show the axialforce, shear-force, and bending moment of the boundary
beam. The forces in the struts in the beam are decomposed into vertical and horizontal components to calculate the beam axial and shear forces in each section. The
beam moment diagram is calculated as the moment of
the forces in the struts and ties about the centerline of
the beam. In the axial force diagram, graph above the
horizontal axis represents compressive force. In the
opening joint (right), the beam is subjected to a tensile
axial force, leading to a smaller flexural capacity of the
beam, while in the closing joint (left), it resists the compressive axial force with lager bending moment capacity.
The shear force in the opening joint (right) is resisted by
the stirrups, while the shear force in the closing joint
(left) is resisted by the concrete strut. The bending moment diagram indicates that the beam bottom reinforcement needs to be provided all along the span,
whereas the top reinforcement needs to be provided
only near the end regions. In the proposed design equations, the above mentioned effects of the stirrups and the
beam axial force are considered to determine the beam
moment capacity.

8. Simplified equations
On the basis of the observed failure modes and the results of the strut-and-tie analysis, simplified equations
are proposed. The equations are exclusively utilizing
conventional flexural analyses.
The observed failure modes can be classified into two
groups: (1) the failure associated with horizontal cracks

151

Fig. 11 Strut and tie model for prototype frame.

at the top of the first-story column (Figs. 10b and 13b

of Part-1) and (2) the failure associated with inclined
cracks in the beam or the joint (Figs. 7b, 8b, 16b, and
17b of Part-1). The strength of each specimen Qu (or the
lateral force which acts on the first story column) is assumed to be the smaller value of the first story column
strength Qc and the joint strength Qj.
Qu = min Qc , Q j

(2)

(a) First-story column strength Qc

As discussed with Fig. 10b of Part-1, the critical cracks
of the column failure in the O-type specimen caused by
the opening load tend to propagate toward the point of
anchorage or node B in Fig. 2a. Therefore, the critical
section is assumed as shown in Fig. 12a; where dco is
the distance between the bottom of the second story
beam and the centroid of the beam bottom bars, and L is
the length between the point of contra flexure and the
top of the first story (column shear span). The first-story
column strength Qc can be obtained considering the
equilibrium of the moments around the white circle
shown in Fig. 12. For simplicity, however, we assume
that Qc can be calculated by the following equation:

Qc =

M c1
L + d co

(3)

where Mc1 is the moment capacity of the first-story column

Mc1 must be computed considering that the concrete
outside the point of anchorage may not resist the compression, as assumed in Fig. 2d. Such a phenomenon

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

Deff

152

Dc1

Dc1

dco

Tv

Cv

dco

Cv

Tv

This distance is
almost dco

Qc

a) Opening load (O-type)

Qc

Qc

b) Opening load (I-type)

Tv

Cv

c) Closing load (O-and-I-types)

Fig. 12 Critical sections in column.

was prominent in Specimen O-1, where the hoop reinforcement ratio in the joint was 0.17%; the inclined
cracks propagated toward the point of anchorage (Fig.
16b in Part-1). In Specimen O-1t, where the hoop reinforcement ratio was 0.80%, the hoop reinforcement in
the joint reduced the opening of such cracks (Fig. 21b
in Part-1). Figure 13a shows the assumed strut created
by the contribution of the hoops in the joint, where
awy represents the horizontal force of the hoops in the
joint and Cout represents the compressive force acting
outside the point of anchorage. This assumption leads to
Cout =

Db
aw y
2(lb Dc 2 )

(4)

where Db is depth of the beam, lb is embedment length

of beam bottom bars and Dc2is the depths of the secondstory column. The depth of the strut outside the anchorage point is
C
Db
D = out =
aw y
f c ' b 2 f c ' b(lb Dc 2 )

(5)

where f'c is the concrete strength and b is the column

width. Therefore, considering the contribution of the
hoops in the joint panel, the effective depth of the firststory column is computed as follows:
Deff = lb + D = lb +

Db
aw y Dc1 (6)
2 f c ' b(lb Dc 2 )

In case Dc1 equals to Dc2, the strut outside the anchorage point would be straight as shown in Fig. 13b and
therefore Deff would be equal to Dc1.
Computing the moment capacity of the first-story column, Mc1, the moment due to eccentricity of the axial
force should also be considered because the center of
the effective section does not coincide with that of the
full section of the first-story column. Therefore,

M c1 = M ef f N i

Dc1 Def f
2

(7)

where Deff is the effective depth of the first-story column,

Fig.13 Contribution of hoops in the joint.

Dc1 is the actual depth of the first-story column, and Meff

is the moment capacity of the column computed by conventional flexural analysis considering the effective
depth.
For I-type joints, the moment capacity of the first
story column, Mc1, is approximated to the moment capacity of the full section of the first-story column (Fig.
12b); because the strut AC (Fig. 5a) can resist vertical
compressions without any help from the hoop. In case
the tensile axial force is extremely large and the strut
AC in Fig. 5 disappears, the strength will be slightly
overestimated.
In the closing direction, as discussed with Fig. 12b
of Part-1, the critical cracks tend to propagate toward
node I (Figs. 7a and 9a), which is close to the centroid
of the beam bottom reinforcements. Therefore, Eq. 3 is
again assumed to apply and the column section is assumed to be fully effective (Fig. 12c).
(b) Strength associated with joints, Qj, for O-series
specimens
As discussed in Part-1, the critical crack of the beam
failure caused by the opening load tends to start from
the inside corner and propagate toward node C in Fig.
2a. Therefore, the critical section is assumed as shown

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

in Fig. 14a. Ch and Cv in Fig. 14a represent the horizontal and vertical components of the strut acting on nodes
C, respectively. Th and Tv represent the tensile forces of
the beam and column reinforcement, respectively. Thj
represents the tensile forces of the hoops in the joint
assuming that the crack in the joint penetrates till about
1/3 of the beam depth (height of the joint panel).
Thj = Pj y bc

Db
3

(8)

where pj is the hoop reinforcing ratio, y is the yield

strength of the hoops, and bc is the width of the joint
panel.
The strength associated with the joint, Qj in Eq. 2, can
be obtained considering the equilibrium of the moments
around the white circle shown in Fig. 14. In other words,

Qj =

Mj

(9)

L + Db / 2

where Mj is the moment caused by the inner forces in

Figs. 14a and 14b (Ch, Cv, Th, Thj, and Tv).
Referring to Fig. 16a of Part-1, we decompose Mj as
follows:
Mj = M c2 + M b

(10)

where Mc2' is the moment caused by Cv and Tv (regarded

as the moment capacity of the second-story column
around the center of the first-story column) and Mb is
the moment caused by Ch, Th, and Thj (regarded as the
moment capacity of the beam).
The critical crack of the beam failure caused by the
closing load is similarly assumed, as shown in Fig. 14b.
The following paragraphs are devoted to discuss the
calculation of Mc2' and Mb.
1) It is assumed that the magnitude and the location of
Cv and Tv in Fig. 14 are similar to the compressive
and tensile forces obtained by conventional flexural
analysis of the second-story column using the axial
force of the first-story column. Thus, Eq. 9 is approximated as

Ch
Cv

Cv

CC
Tv
Ch

.Db

Dc2

Ch

Cv
C
Db

Thj

Th

CI
Th
Cv

Tv
I

Ch

.Db

Qj

Qj

Db
2

(11)

where Mc2 is the moment capacity of the second-story

column around its center and the second term is the
moment due to the axial force around the center of the
first-story column. The sign of the second term is negative and positive for opening and closing loads, respectively.
2) The moment capacity of the beam, Mb, caused by
the horizontal inner forces is assumed as
M b = Ch Db + (d

Db
) (Th Thj )
2

(12)

where Db represents the distance of the compressive

force Ch from the centerline of the beam (Figs. 14a and
14b). Based on the STMs, is assumed to be 0.4 and
0.2 for opening and closing loads, respectively (Table
1). For simplicity, the location of Thj is assumed to be
the same as that of Th. On the other hand, the equilibrium of horizontal forces leads to Ch = Th + Thj Q j
and Ch = Th + Thj + Q j for opening and closing loads,
respectively. Thus, we get
D

M b = Db + d b
2

(Th + Thj ) Db Q j

(13)

The sign of the second term is negative and positive

for opening and closing loads, respectively.
(c) Strength associated with joint, Qj, for I-series
specimens
It is assumed that Eqs. 8 and 9 also apply to I-series
specimens. However, Mc2 in Eq. 10 needs to be evaluated differently because of the different failure modes.
Figure 15a shows the forces around the joint of I-series
specimens subjected to opening load. As shown in Fig.
7b of Part-1, the prominent crack in the opening load
starts from the beam and proceeds to the joint above the
hook of the first-story column main bars in the extended
area. In the STM, the struts BK and MK (Fig. 5a) are
anchored inside the hook. Thus, the value for the
opening load is smaller (0.3) than that for the O-Series
(0.4), as shown in Fig. 15a and Table 1. The moment
capacity of the beam is calculated using Eq. 14.

Th Db Q j

(14)

Figure 16 shows the contribution of the beam stirrups.

As discussed with Fig. 7a of Part-1 paper, the prominent

b
B L+ 2

Table 1 Variation of coefficient .

Load Cases

(a) Opening load

(b) Closing load
Fig. 14 Critical sections in O-type joint.

Dc1 Dc 2
2

M b = Db + d b
2

Thj
C

M c2 = M c 2 N i

153

O-Type
I-Type

Opening
0.4

Closing
0.2
0.3

154

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

Dc2
Cv

.Db

Cv

Tv C v

Ch Tw
Tst1

L=(Dc1 -Dc2 )+0.8Db

Ch

Th

45

Db d

C
Tst2

Db
2

L+

Db
2

Ch

.Db
Qj

Tv

Th

0.4Db

Qj

(a) Opening load

(b) Closing load
Fig. 15 Critical sections in I-type joint.

crack proceeds to point A in Fig. 16 (the top of the

hooks of vertical bars in the extended area of the first
story column); and assuming that the crack is inclined at
45, the tensile force of the beam stirrups Tst1 in the
opening case (Fig. 16) is then equal to the tensile force
of the beam stirrups in the length of ld2dcr (Fig. 16).
ld is the embedment length of the vertical bars of the
first-story column anchored in the joint, dc is the concrete cover of the beam and first story column and r is
radius of the hook.
Tst1 = pw y bb (ld 2d c r )

(15)

where pw is the stirrup reinforcing ratio, and bb is width

of the beam. Though the stirrups were neglected in Fig.
9b for specimen I-1, stirrups in specimen I-1t considerably contribute to the strength.
Moreover, the wall vertical reinforcement (Tw) also
contributes to the joint strength in the opening case. The
force Tw acting at node K is the resultant force of the
wall vertical reinforcement above the joint. Because the
direction of Tst1and Tw is vertical, their contribution is
added to the moment capacity of the second-story column; thus, we get
M c2' = M c2 + N

Dc1 + Dc 2
+ Tst1 lst1 + Tw lw
2

(16)

where Mc2 is the moment capacity of the second-story

column around its center. lst1 is the distance between Tst1
and Cv in Fig. 15a because Tst1 increases Cv, and is approximated as lst1 = 0.5ld dc + 0.9 Dc1 . lw is the distance between Tw and the Cv, and is approximated as
lw = 0.6 Dc1 .
In Specimen I-1 subjected to the closing load, the inclined crack in the beam and the crushing of the wall

panel concrete were prominent (Fig. 8 of Part-1). Therefore, the critical section is assumed as shown in Fig.
15b. Ch and Cv in Fig. 15b represent the horizontal and
vertical components of the strut acting on node I, respectively. For simplicity, Ch is assumed to act at the
beam section and Cv is assumed to act at the T-section
composed of the wall panel and the second-story column. The length of the wall panel L is assumed as the
sum of the length above the joint (Dc1-Dc2) and 80% of
the beam depth.
L = ( Dc1 Dc 2 ) + 0.8Db

(17)

where 0.8Db represents the projection of the beam depth

above the centroid of the beam compression zone. It is
further assumed that the moment created by Cv and Tv
can be approximated by the flexural strength of the T-

lst1=0.5ld dc+0.9Dc1

Tw

Cv

Tst1

ld
45
dc

ld 2dcr
dc

Fig. 16 Contribution of beam stirrups in opening direction. (I-type joint).

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

155

Table. 2 Procedure to compute lateral strength of joint.

Shear Force

Moment

Qc
(Eq. 3)

M c1

Qj
(Eqs. 9-10)

M c2
Mb

Joint Type
O-Type
I-Type
O-Type
I-Type
O-Type
I-Type

Loading Cases
Opening
Closing
Eqs. 4-7 and Flexural Analysis
with Deff
Flexural Analysis with Full Section
Eq. 11 and Flexural Analysis of Rectangular Section
Eqs. 15-16 and Flexural Analysis
Eqs. 17-19 and Flexural Analysis
of Rectangular Section
of T-Section
Eq. 13 + Table 1
Eq. 14 + Table 1

section in Fig. 15b. Thus, Mc2' in Eq. 9 is computed by

the following equation:
M c 2 ' = M c 2 + N e + Tst 2 lst 2

(18)

where Mc2 is the moment capacity of the T-section

around its centroid and e is the distance between the
centroid and the center of the first-story column, and
corresponds to:
e=

Dc 2 b ( Dc1 Dc 2 ) + L2 t
2( Dc 2 b + L t )

(19)

where b is the width of second-story column and t is the

thickness of wall panel, as shown in Fig. 15b. The last
term in Eq. 18 shows contribution of the beam stirrups
in the closing direction. The tensile force of the beam
stirrups (Tst2), is different from that in the opening direction (Tst1). Tst2 is assumed to be equal to the tensile force
of the beam stirrups in the length equal to 0.8Db and
acting at the distance 0.4Db from the face of the column.

Tst 2 = pw y bb 0.8 Db Tst 2 = pw y bb 0.8Db (20)

lst 2 in Eq. 18 is the distance between Tst2 and Tv in
Fig. 15b because Tst2 increases Tv, and is approximated
as lst 2 = 0.4 Db + 0.9 Dc1 .
Mb in Eq. 10 is computed using Eq. 14, and is assumed to be 0.3 (Table 1) based on Fig. 15.
(d) Summary and verification
The procedure to compute the strength of the joint is
summarized here with the help of Table 2. To compute
the first-story column strength (Qc), the first-story column moment (Mc1) should be computed. For the opening strength of O-type joints, Eqs. 4-7 are used considering the effective depth of the first-story column. For
the other cases (closing strength of O-type joints as well
as opening and closing strengths of I-type joints) full
section of the first-story column should be considered
(Table 2).
To compute the joint strength (Qj), the beam bending
moment (Mb) and second-story column moment (Mc2')
should be calculated. The second-story column moment
(Mc2') of O-type joints should be computed using Eq. 11
both for opening and closing loads. Mc2 in Eq. 11 is the

moment capacity of the rectangular section of the second-story column (Fig. 14). The second story moment
(Mc2') of I-type joints for the opening load should be
calculated from Eqs. 15-16. Mc2 in Eq. 16 is the moment
capacity of the second-story column with rectangular
section (Fig. 15a). For the closing load of I-type joints,
Eqs. 17-19 should be used to find Mc2'. Mc2 in Eq. 18
should be computed by flexural analysis of the T-section
(Fig. 15b).
The beam bending moment Mb is defined by Eq. 13.
The values are given in Table 1 for different joint
types and load cases. Because Eq. 13 includes the joint
strength (Qj), Qj is computed with an iteration method.
Figure 17 shows the comparison between the observed and calculated strengths. Qexp in the vertical axis
is the observed strength, while Qj in the horizontal axis
is the computed joint strength. The denominator Qc is
the computed column strength. Assume that the proposed equations are perfect (Qu= Qexp). Then Eq. 2 leads
to the following:
If Qj<Qc, Qexp= Qj, which is represented by the inclined red line connecting (0, 0) and (1, 1) in Fig. 17.
If Qj>Qc, Qexp= Qc, which is represented by the horizontal black line connecting (1, 1) and (2, 1) in Fig. 17.
These lines constitute the border between the green
and yellow zones. A plot in the green zone indicates that
the observed strength is larger than the expected value
and vice versa. Because most plots are located around
the border, we conclude that the proposed design equations appropriately evaluate the strengths of the speci-

1.5

1.0
I-1
0.5

O-1 I-1t O-1t

I-1t
I-1
O-1t
O-1
Observed failure modes

Hanai-1
Hanai-2

Hanai-3

Joint Column
Load Opening
Case Closing

Beam failure is Column failure is estimated

estimated
0.5
1.0
1.5
2.0
Fig. 17 Estimated and observed strengths.

156

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

mens. The plotted data include the specimens of Hanai

et al. (2009). The vertical broken line at Qj/Qc= 1 is the
border of beam or column failure: if Qj/Qc> 1, column
failure is expected; if Qj/Qc< 1, beam failure is expected.
The observed failure modes, which are indicated by
circles and squares, mostly agree with the expectations.
Figure 18 compares the observed and computed
strengths in terms of average shear stresses. The observed strength (Qexp) and computed strength (Qcomputed)
are divided by the gross area of the first-story column
(Ac1). For most of the specimens, the observed and computed shear stresses have good agreement.
(e) A model for pushover analysis
Though Eq.2 is useful to evaluate the test specimens, it
cannot be applied to actual structures. Figure 19 shows
a model for pushover analysis which may be used for
design. The shear wall and boundary columns in the
upper stories are modeled according to Kabeyasawa et
al. (1983). In this model, infinitely rigid beams are assumed at the top and bottom of each floor level. The
axial stiffness of the boundary columns is represented
by truss elements. The first-story columns are modeled
+10%

Qtest
( N / mm2 )
Ac1

-10%

O-1
I-1

I-1t
O-1t

as One-component model (Giberson 1974). Namely,

each column is idealized as a perfectly elastic line element (Fig. 20a) with two non-linear rotational springs at
the two ends. The lower spring is located at a distance
dco from the beam bottom face (Fig. 19b). Figure 20b
shows the characteristics of the lower spring: the yield
strength Mc is computed according to Table 2 and the
yield rotation cy is computed by conventional methods.
To consider the beam failure, another non-linear rotational spring is located at the center of the joint panel
(Fig. 19b). Figure 20c shows the characteristics of the
joint spring: the yield strength Mj is computed according
to Table 2 and the yield rotation jy is discussed by Ichinose et al (2013).
To compute Mc and Mj, N and Qj are required. A preliminary elastic analysis of the structure is required to
obtain the approximated values of these variables.

9. Conclusions
(1) The STM for a prototype frame validates the
specimens constructed in this study. In other words,
the specimens (part of a structural frame) represent
the structural behavior of beam-column joint and
connecting elements appropriately. The model also
implies that the vertical wall reinforcement plays an
important role in a wall with an open frame.
(2) The comparison between STMs for test specimens
and those for usual exterior and knee joints shows
that the flow of the forces of the specimens is dif-

I-1t
O-1 I-1O-1t
Hanai-2 Hanai-1
Hanai-3

Hinge

Qcomputed
Ac1

dco

Column
end speing

Elastic

( N / mm2 )

Fig. 18 Comparison between observed and computed

strengths.

Joint
rotational
spring

Rigid

a) Idialized model of frame

b) Joint details

Fig. 19 Model for pushover analysis.

M

Mj
Mc1

Mc1

O
y

(rad/m)

O
cy

(rad)

(rad)

a) 1F-Column elastic element

b) Column end spring
c) Joint rotational spring
Fig. 20 Rotational springs and elastic element properties.

S. Halim, S. Takahashi, T. Ichinose, M. Teshigawara, T. Kamiya and H. Fukuyama / Journal of Advanced Concrete Technology Vol. 12, 146-157, 2014

ferent from that of the usual joints. All the specimens were stronger than the usual joints and a huge
amount of shear force was transferred to the wall
panel.
(3) In the case of the opening load of O-type joint, the
effective depth of the first-story column is defined
by the embedment length of the beam bottom bars
and shear reinforcement in the joint panel.
(4) The strength related to the beam failure can be approximated as the sum of the beam and secondstory column strengths including the effects of the
wall panel, hoops in the joint, and stirrups in the
beam.
(5) The strength and failure modes predicted by the
proposed procedure (Table 2) are in good agreement with the test results.
References
Giberson, M. F., (1974). Two nonlinear beams with
definition of ductility. Journal of the Structural
Division, ASCE, 95(7), 137-157.
Hanai, N., Goto, Y. and Ichinose, T., (2009).
Anchorage of column re-bars where column depth
greatly reduces and structural wall appears. AIJ J.
Technol. Des., 15(29), 143-146. (in Japanese)
Ichinose, T., Kawai, T., Takahashi, S. and Teshigawara,
M., (2013). Second stiffness of beam-column joint
in soft-first story RC building. Proceedings of the
Japan Concrete Institute, 35(2), 265-270. (in
Japanese)
Kabeyasawa, T., Shiohara, H., Otani, S. and Aoyama, H.,
(1983). Analysis of the full-scale seven-story
reinforced concrete test structure. Journal of the
Faculty of Engineering, University of Tokyo, 36(2),
432-478.
Liang, Q., (2005). Performance-based optimization of

157

structures: Theory and applications. 1sted. New

York: Spon Press.
Wight, J. K. and MacGregor, J. G., (2012). Reinforced
Concrete Mechanics and Design. USA: Prentice
Hall.
Paulay, T., Park, R. and Priestley, M. J. N., (1978).
Reinforced concrete beam-column joints under
seismic actions. ACI Journal, Proceedings, 75(11),
585-593.
Schlaich, J., Schfer, K. and Jennewein, M., (1987).
Toward a consistent design of structural concrete.
PCI journal, 32(3), 77-150.
Shiohara, H., (2001). New model for shear failure of
RC interior beam-column connections. Journal of
Structural Engineering, ASCE, 127(2), 152-160.
Shiohara, H., (2004). Quadruple flexural resistance in
R/C beam-column joints. Proceedings, the 13th
World Conference on Earthquake Engineering,
Vancouver, B. C., Canada, No. 491.
Shiohara, H., (2012) Reinforced concrete beamcolumn joints: An overlooked failure mechanism.
ACI Structural Journal, 109(1), 65-74.
Shobu, H., Teshigawara, M., Nakamura, A., Izumi, N.,
Matsumoto, K., Ichinose, T., Takahashi, S., Takakoshi,
S., Kamiya, T., Fukuyama, H., Suwada, H. and
Kabeyasawa, T., (2012). Effects of vertical
reinforcement and boundary beam on the shear
strength of RC wall located above an open frame.
Proceedings, 15th World Conference on Earthquake
Engineering, Paper ID 4639.
Sritharan, S., (2005). Strut-and-tie analysis of bridge
tee joints subjected to seismic actions. Journal of
Structural Engineering, ASCE, 131(9), 1321-1331.
Yun, Y. M. and Ramirez, J. A., (1996). Strength of
struts and nodes in strut-tie model. Journal of
Structural Engineering, ASCE, 122(1), 20-29.