Engineering Structures 24 (2002) 1105–1117

www.elsevier.com/locate/engstruct

A new design equation for predicting the joint shear strength of

monotonically loaded exterior beam-column joints
P.G. Bakir ab,∗, H.M. Boduroğlu a
a
Istanbul Technical University, Civil Engineering Department, Maslak 80626, Istanbul, Turkey
b
Postal address: Yazmaci Tahir sok, Derya apt. no 11/4, Catalcesme, Suadiye, Istanbul, Turkey

Received 11 June 2001; received in revised form 8 March 2002; accepted 8 March 2002

Abstract

In this study a new design equation for predicting the shear strength of monotonically loaded exterior beam column joints is
proposed. For this purpose, the influence of several key variables on the behaviour of beam-column joints are inspected using
results of parametric studies on an experimental database compiled from a large number of exterior joint tests. The design equation
suggested has three differences from the previously proposed equations. First, the equation proposed considers the influence of
beam longitudinal reinforcement ratio, which was not taken into account in previously suggested design equations. Second, as the
influence of this parameter is taken into account, a more realistic estimate of the influence of joint aspect ratio is obtained. Third,
the influence of stirrups is considered differently for joints with low, medium and high amount of stirrup ratios, in a way, which
was not considered in previously suggested equations. The results showed that the proposed design equation predicts the joint shear
strength of exterior beam column connections accurately with minimal standard deviation and is more reliable than the previously
suggested equations.  2002 Elsevier Science Ltd. All rights reserved.

Keywords: Reinforced concrete; Shear strength; Monotonically loaded exterior beam-column connections

1. Introduction reinforcement detailing, the ratio of beam longitudinal

It is now generally believed that beam-column joints
can be critical regions in reinforced concrete frames
under severe seismic effects. Beam-column joint failures
have been observed in the 1980 El Asnam [1], 1985
Mexico [2], 1986 San Salvador [3], 1989 Lome Prieta
[4] and 1999 Kocaeli earthquakes [5]. During the past
forty years, significant amount of research has been car-
ried out on seismic behaviour of beam-column joints all
over the world. However, compared to cyclically loaded
joints, little information exists in literature for predicting
the shear strength of monotonically loaded exterior
joints.
Remarkable differences exist in the design of joints
for seismic loading or monotonic loading. Parameters
such as column axial load, concrete cylinder strength,
stirrup ratio, stirrup index, joint aspect ratio, beam
reinforcement etc influence the joint shear strength dif-
ferently for interior or exterior and monotonically loaded
or cyclically loaded joints. This investigation was
planned with the objective of adding useful data to the
understanding of the influences of the above-mentioned
parameters on the joint shear strength of monotonically
loaded exterior beam-column joints. Surprisingly, none
of the previously suggested design equations for mono-
tonically loaded exterior joints consider the factors that
influence joint shear strength such as beam longitudinal
reinforcement ratio, joint aspect ratio and concrete cylin-
der strength together.
2. Previously suggested design equations for
exterior beam column joints

In this section, the existing empirical design equations

∗
Corresponding author. Fax: +90-216-386-9742. for predicting the shear strength of the monotonically
E-mail address: gundes@itu.edu.tr (P.G. Bakir). loaded exterior beam-column joints are reviewed.

0141-0296/02/$ - see front matter  2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 1 - 0 2 9 6 ( 0 2 ) 0 0 0 3 8 - X
1106 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117
2.1. The design equation of Sarsam and Phillips
Sarsam and Phillips [6] have proposed the following
equation for the design of monotonically loaded exterior
beam column connections.
冪冉1 ⫹ 0.29A 冊
dc Nu
Vcol ⫽ 5.08fcurc0.33 1.33bcdc
db g total joint shear strength (N); hb is the section depth of
where Vcol is the column shear force at the column-joint
interface (N); fcu is the concrete cube strength (MPa); rc
is the column longitudinal reinforcement ratio
rc ⫽ Aso / bcdc
where Aso is the area of the layer of steel furthest from
the maximum compression face in a column (mm2); Ag
is the gross cross-sectional area of the column at the joint
(mm2); Nu is the axial column load (N);dc is the effective
depth of the layer of steel furthest away from the
maximum compression face in a column (mm); db is the
effective depth of beam tension reinforcement (mm); bc
is the width of column section at the joint (mm).
The shear force resisted by the links is taken as:
Vsd ⫽ 0.87Ajsfyv
Vud ⫽ Vcd ⫹ Vsd
where Ajs is the total area of horizontal link reinforce-
ment crossing the diagonal plane from corner to comer
of the joint between the beam compression and tension
reinforcement (mm2); fyv is the tensile strength of the
link reinforcement (MPa); Vsd is the design link shear
force resistance (N); Vcd is the design shear force resist-
ance of concrete in a joint (N); Vud is the design ultimate
shear capacity of joint (N)
All the joint stirrups are considered to be effective in
increasing the joint shear capacity.
2.2. The design equation of Vollum
Vollum [7] proposed the following equation on the
design of exterior beam column connections:
冉 hb
hc 冊
Vc ⫽ 0.642b 1 ⫹ 0.555(2⫺ ) beffhc冑fc
Vj ⫽ Vc ⫹ (Asjefy⫺abeffhc冑fc)
a is taken as 0.2 for low, medium and high amount of
stirrups. b is 0.9 for U detail beam reinforcement and
1 for L bent down. The joint shear strength should be
limited to:
冉 冊
Vj具0.97beffhc冑fc 1 ⫹ 0.555(2⫺ )
hb
hc
具1.33beffhc冑fc
where Asje is the cross-sectional area of the joint stirrups
within the top five eighths of the beam depth below the
main beam reinforcement (mm2); a is a coefficient that
depends on factors including column load, concrete
strength, stirrup index, and joint aspect ratio, which is
conservatively taken as 0.2 in Vollum’s equation. Vc is
the joint shear strength without stirrups (N); Vj is the
(1)
the beam (mm); hc is the section depth of the column
(mm); fc is the concrete cylinder strength (MPa); beff is
the average of the beam and column widths (mm); fy is
the yield strength of stirrups (MPa).
(2)
2.3. Present design guidelines
The ACI-ASCE Committee 352 [8], and EC8 [9] rec-
ommend the following design equations for the shear
strength of monotonically loaded joints.
Vjd ⫽ 1.058冑fcbeffhc (ACI (9)
⫺ ASCE Committee 352)
Vjd ⫽ 0.525f 2/3
c beffhc (EC8 ductility class DCL) (10)
(3) In order to investigate the reliability of the above
(4) design equations, the authors carried out several para-
metric studies on monotonically loaded exterior beam-
column joints. The parametric studies are explained in
the next section followed by a comparison of the above
design equations with the parametric studies of authors.
3. Parametric studies on joint shear strength
The authors carried out a parametric investigation of
exterior beam-column joint behaviour based on 58 tests
conducted in Europe. The loading in all the tests was
monotonic. The authors realised that there might be
important interactions between variables that at first
were supposed to act independently and this necessitated
viewing the entire population of results as a single para-
meter. For this purpose, monotonically loaded exterior
(5)
beam column joints are investigated through test results
assembled and reconsidered as a unified whole. Examin-
ing a large number of individual series of tests as a single
(6) database has the advantage of observing which variables
have a significant influence on joint shear strength in all
tests and which variables interact with each other. A new
design equation is proposed based on the parametric
studies carried out on the experimental database. Table
1 shows the experimental database used in this study.
The database comprises of the tests of Ortiz [10], Kord-
ina [11], Scott [12], Scott & Hamill [13], Taylor [14],
(7)
and Parker & Bullman [15]. The specimen forms
included in the database are shown in Fig. 1(a). Fig. 1(b)
(8) shows the notations used in this study.
Table 1
Experimental database

Investigator Specimen Detail H (mm) L (mm) hb/hc bc/bb Beam Column rein. fc(MPa) Column axial SI (MPa)–0.5 Vjpredicted/Vjactual Failure
rein.ratio ratio load modes

Ortiz BCJ 1 L bar 2000 1050 1.33 1.00 1.1 2.19 34 0 0 0.68 is
BCJ 2 L bar 2000 1100 1.33 1.00 1.1 2.19 38 0 0.16 0.77 is
BCJ 3 L bar 2000 1100 1.33 1.00 1.1 2.92 33 0 0 0.64 is
BCJ 4 L bar 2000 1100 1.33 1.00 1.1 3.65 34 0 0.33 0.78 is
BCJ 5 L bar 2000 1100 1.33 1.00 1.1 3.65 38 300 0 0.72 Is
BCJ 6 L bar 2000 1100 1.33 1.00 1.1 3.65 35 300 0 0.68 is
BCJ 7 L bar 2000 1100 1.33 1.00 1.1 3.65 35 300 0.76 0.68 b
Kordina RE 2 L bar 3000 1000 2.00 1.00 0.9 2.41 25 240 0 0.66 is
RE 3 L bar 3000 1000 1.50 1.00 1.8 2.41 40 400 0.26 0.96 is
RE 4 L bar 3000 1000 1.50 1.00 1.2 2.41 32 51 0.19 0.83 is
RE 6 L bar 3000 1000 1.50 1.00 1.2 2.41 32 213 0.38 0.91 is
RE 7 L bar 3000 975 1.40 1.00 1.3 1.61 26 650 0.43 0.87 is
RE 8 U bar 3000 975 1.40 1.00 1.3 1.61 28 525 0.42 0.90 is
RE 9 U bar 3000 975 1.40 1.00 1.3 1.61 28 770 0.41 0.86 is
RE 10 U bar 3000 975 1.56 1.00 1.2 1.61 24 551 0.45 0.94 is
Taylor P1/41/24 L bar 1290 470 1.43 1.40 2.4 4.10 33 240 0.3 0.97 is
P2/41/24 L bar 1290 470 1.43 1.40 2.4 4.10 29 240 0.3 0.94 is
P2/41/24A L bar 1290 470 1.43 1.40 2.4 4.10 47 240 0.26 0.92 is
A3/41/24A L bar 1290 470 1.43 1.40 2.4 4.10 27 240 0.3 0.88 is
D3/41/24 L bar 1290 470 1.43 1.40 2.4 4.10 53 60 0.24 0.89 is
B3/41/24 L bar 1290 470 1.43 1.40 2.4 4.10 22 240 0.75 0.92 is
C3/41/24BY U bar 1290 470 1.43 1.40 2.4 4.10 32 240 0.31 1.04 is
C3/41/13Y U bar 1290 470 1.43 1.40 1.4 4.10 28 240 0.33 0.95 is
Scott C1AL L bar 1700 750 1.40 1.36 1.1 4.29 33 50 0.188 0.87 is
C4 L bar 1700 750 1.40 1.36 2.1 4.29 41 275 0.203 0.89 is
C4A L bar 1700 750 1.40 1.36 2.1 4.29 44 275 0.196 0.86 is
C4AL L bar 1700 750 1.40 1.36 2.1 4.29 36 50 0.218 0.86 is
C7 L bar 1700 750 2.00 1.36 1.4 4.29 35 275 0.22 0.90 is
C3L U bar 1700 750 1.40 1.36 2.1 4.29 35 50 0.22 1.03 is
C6 U bar 1700 750 1.40 1.36 2.1 4.29 40 275 0.21 1.05 is
C6L U bar 1700 750 1.40 1.36 2.1 4.29 46 50 0.19 0.94 is
P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117

C9 U bar 1700 750 2.00 1.36 1.4 4.29 36 275 0.22 0.93 is
(continued on next page)
1107
Table 1 (continued) 1108

Investigator Specimen Detail H (mm) L (mm) hb/hc bc/bb Beam Column rein. fc(MPa) Column axial SI (MPa)–0.5 Vjpredicted/Vjactual Failure
rein.ratio ratio load modes

Scott & C4ALN0 L bar 1700 750 1.40 1.36 2.1 4.29 42 50 0 0.88 p
Hamil
C4ALN1 L bar 1700 750 1.40 1.36 2.1 4.29 46 50 0.229 0.85 js
C4ALN3 L bar 1700 750 1.40 1.36 2.1 4.29 42 50 0.478 0.78 js
C4ALN5 L bar 1700 750 1.40 1.36 2.1 4.29 50 50 0.718 0.85 js
C4ALH0 L bar 1700 750 1.40 1.36 2.1 4.29 104 100 0 0.86 p
C6LN0 U bar 1700 750 1.40 1.36 2.1 4.29 51 50 0 0.92 js
C6LN1 U bar 1700 750 1.40 1.36 2.1 4.29 51 100 0.19 0.96 js
C4ALH1 L bar 1700 750 1.40 1.36 2.1 4.29 95.2 100 0.159 0.93 b
C4ALH3 L bar 1700 750 1.40 1.36 2.1 4.29 105.6 100 0.302 0.97 b
C4ALH5 L bar 1700 750 1.40 1.36 2.1 4.29 98.4 100 0.469 1.00 b
C6LN3 U bar 1700 750 1.40 1.36 2.1 4.29 49 50 0.44 0.92 js
C6LN5 U bar 1700 750 1.40 1.36 2.1 4.29 37 50 0.765 0.74 js
C6LH0 U bar 1700 750 1.40 1.3 2.1 4.29 101 100 0 0.72 js
C6LH1 U bar 1700 750 1.40 1.36 2.1 4.29 102 100 0.153 0.98 js
C6LH3 U bar 1700 750 1.40 1.36 2.1 4.29 97 100 0.472 0.93 js
Parker 4a L bar 2000 850 1.67 1.20 0.9 1.09 39 0 0 - c
4b L bar 2000 850 1.67 1.20 0.9 1.09 39 300 0 1.05 js
4c L bar 2000 850 1.67 1.20 0.9 1.09 37 600 0 0.83 js
4d L bar 2000 850 1.67 1.20 0.9 4.38 39 0 0 0.97 js
4e L bar 2000 850 1.67 1.20 0.9 4.38 40 300 0 0.92 js
4f L bar 2000 850 1.67 1.20 0.9 4.38 38 600 0 0.78 js
5a L bar 2000 850 1.67 1.20 0.9 2.67 42 0 0.404 - c
5b L bar 2000 850 1.67 1.20 0.9 2.67 43 300 0.4 1.08 js
5d L bar 2000 850 1.67 1.20 1.4 2.67 43 0 0.6 - c
5e L bar 2000 850 1.67 1.20 1.4 2.67 45 300 0.589 - c
5f L bar 2000 850 1.67 1.20 1.4 2.67 43 600 0.6 0.86 js
Average 0.88
Standard devi- 0.10
ation
P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117

Note: C is column failure; b is beam failure; js is jointshear failure; p is connection zone reinforcement pullout failure.
 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117
Fig. 1. (a) Typical specimen shape in the experimental database. (b)
Typlical elevation and notations used for exterior beam colum joints.
(c) The strut and truss mechanisms.
It is commonly accepted that beam column joints
resist shear by the strut and truss mechanisms as sug-
gested first by Paulay [16] as shown in Fig. 1(c). The
strut mechanism accounts for the contribution of the con-
crete, whereas the truss mechanism represents the contri-
bution of stirrups to joint shear strength. In this study,
the resistance of the concrete to the joint shear will be
determined first and then the influence of the stirrups
will be added.
4. Influence of concrete cylinder strength
The relation between maximum joint shear stress at
the instant of joint failure and the uni-axial compressive
stress of concrete fc is shown in Fig. 2. The authors car-
ried out a regression analysis and found that concrete
1109
Fig. 2. The influence of concrete cylinder strength on joint shear
strength.
cylinder strength is related to the joint shear strength by
the following equation.
nj⬇0.9155冑fc (11)
where fc is the uniaxial concrete cylinder strength with-
out any factors of safety and vj is in MPa.
C4ALHO and C4ALNO of Scott have been investi-
gated in order to find the possible influences of the inter-
action of other parameters with the concrete cylinder
strength. These specimens were deliberately chosen so
as to eliminate other factors such as stirrups, column
axial load and joint aspect ratio. Because they had no
stirrups, they had similar joint aspect ratios and they
belonged to the same group of tests.
The analysis confirmed the above relationship
between the concrete cylinder strength and the joint
shear strength. Hamill has also suggested that the joint
shear strength is proportional to the square root of the
concrete cylinder strength.
5. Determining the joint shear strength
The joint shear is calculated using the following pro-
cedure:
Mb ⫽ P(L ⫹ d1) (12)
where P is the failure load (N); L is the distance from
the point of application of the load to the face of the
column (mm); d1 is the cover (mm).
A value is assumed for the strain in the beam tensile
reinforcement. The force in the beam tensile reinforce-
ment and the moment produced by it are calculated and
if this is equal to the moment calculated in Eq. (2), the
procedure is stopped. If not, the strain assumed is
increased in small increments up until, the moment is
equal to the moment given in Eq. (2). The reinforcement
was assumed to have an elastic modulus of 200 GPa.
The joint shear strength is calculated as below:
Vj ⫽ Tb⫺Vcol (13)
where Vj is the joint shear force (N); Tb is the tensile
1110 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117
force in the beam longitudinal reinforcement (N); Vcol is
the shear force in the upper column (N).
The normalised joint shear strength is determined as:
Vj
nj ⫽
beffhc冑fc
(14)
where beff is the average of the beam and the column
width; fc is the concrete cylinder strength; hc is the height
of the column.
The unit of v j is (MPa)0.5.
6. Influence of column reinforcement ratio
The authors made a parametric study on the specimens
in the experimental database which failed by joint shear,
which had low amount of stirrups and which were
detailed by L bars bent down detail. Parker and
Bullmann specimens were deliberately excluded from
this parametric study because they have high joint aspect
ratios and low beam reinforcement ratio. Furthermore
due to low column longitudinal reinforcement ratio and
low column axial stress, some of them have failed due
to column failure. The Kordina specimens were also
excluded as they are provided by inclined reinforcement
in their joints which significantly increased their joint
shear strength. The parametric study in Fig. 3 clearly
shows that joint shear strength is independent of the col-
umn longitudinal reinforcement ratio in joints that fail
by joint shear failure.
The authors also analysed the joints of Parker and
Bullman. The analysis of specimens 4a and 4d of Parker
and Bullman showed that these two specimens were
identical except for their column longitudinal reinforce-
ment ratios. The provision of 75% less column reinforce-
ment than specimen 4d, did not only decrease the joint
shear strength of specimen 4a, but also caused the speci-
men 4a to fail by column failure instead of joint shear
failure. It can be concluded that, the joints that have low
column longitudinal reinforcement ratios and column
axial stresses are more likely to fail by column hinging
Fig. 3. The influence of the column longitudinal reinforcement ratio
on the normalised joint shear strength.
and consequently have lower joint shear strengths with
respect to joints that fail by joint shear failure.
7. Influence of beam longitudinal reinforcement
ratio
The authors having investigated the specimens of
Parker and Bullman decided that the possible low
strength of Parker and Bullman specimens apart from
detailing and high joint aspect ratio as well as low radius
of bend could be the very low beam longitudinal
reinforcement ratio. The relation between the beam
longitudinal reinforcement ratio and the normalised joint
shear strength is demonstrated in Fig. 4. Fig. 4 shows
that the ratio of the beam longitudinal reinforcement is
related to the normalised joint shear strength by the fol-
lowing equation:
nj⬇ 冉 冊
Asb
b bd
0.4289
(15)
where Asb is the total area of beam reinforcement; bb is
the breadth of the beam; d is the depth of the beam.
In order to understand the possible influence of the
ratio of beam longitudinal reinforcement on the joint
shear strength, the authors investigated the specimens
C4AL and C1AL of Scott, which were nearly identical
except the beam reinforcement ratios they had. The joint
shear strength of C1AL that had a beam longitudinal
reinforcement ratio of 1.1 was nearly 30% lower than
C4AL that had a beam reinforcement ratio of 2.1. The
analysis of these two specimens showed that the beam
longitudinal reinforcement is related to the normalised
joint shear strength by the above equation.
8. Influence of beam reinforcement detailing
The beam reinforcement detailing affects both the nor-
malised joint shear strength and the failure modes of
joints. The authors’ analysis of two identical specimens
of Kordina with different beam reinforcement detailing
Fig. 4. The influence of the beam longitudinal reinforcement ratio on
the normalised joint shear strength.
 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117
showed that the joint shear strengths of monotonically
loaded exterior beam column connections decrease by
15% if detailed by U bars rather than L bars. This is
shown in Fig. 5. Inspection of Scott and Hamill speci-
mens C4ALH3 (L bars bent down detail) and C6ALH3
(U bars detail) which were identical except for their
beam reinforcement detailing shows that providing L
bars bent down detail reinforcement can change the fail-
ure modes of monotonically loaded exterior beam col-
umn connections from joint shear to beam failure.
Further inspection of Kordina specimens with that of
Ortiz or Scott shows that the normalised joint shear
strengths of Kordina specimens are significantly higher
than that of other researchers’ specimens. The authors
are of the opinion that there are two important reasons
for this. First, the specimens of Kordina except for RE4
are detailed by inclined reinforcement, which signifi-
cantly increases the normalised joint shear strength.
Second, the concrete cylinder strengths of Kordina’s
specimens have been underestimated because they are
based on the minimum rather than the average cube
strengths. The consequence of this is that the normalised
joint shear strengths may have been overestimated.
9. Influence of the vertical anchorage length and
the radius of bend
The analysis of BCJ3 of Ortiz showed that the joint
shear strength of BCJ3 was considerably increased com-
pared to other specimens of Ortiz with low amount of
stirrups. The authors are of the opinion that this increase
is due to higher vertical anchorage length and higher
radius of bend of beam reinforcement of Ortiz specimen
BCJ3. Thus, it can be concluded that there is evidence
from Ortiz tests that the shear strength of joints can be
increased if the vertical anchorage length is higher than
26db (db is the diameter of beam reinforcement) and the
radius of bend is higher than 8db.
Fig. 5. The normalised joint shear strength of identical specimens
with different beam reinforcement detailing.
1111
10. Influence of joint aspect ratio
The authors have carried out a parametric study to
investigate the influence of the joint aspect ratio on the
normalized joint shear strength. The specimens which
fail by failure modes other than joint shear failure, the
U bar specimens, the specimens with inclined bars, all
of Parker and Bullmann specimens, all specimens with
medium and high amount of stirrups are excluded from
this study. In order to minimise the possible interaction
of parameters such as the concrete cylinder strength and
the beam longitudinal reinforcement ratio, the joint shear
strength is normalised by the square root of the concrete
cylinder strength and the 0.4289 power of the beam
longitudinal reinforcement ratio. Fig. 6 shows the
relation between the joint aspect ratio and the normalised
joint shear strength. It is evident that the joint aspect
ratio is related to the normalised shear strength as shown
in Eq. (16).
Vj
beffhc冑fc
⬇ 冉冊
hb
hc
⫺0.61
(16)
where hb is the cross-sectional height of the beam; hc is
the cross-sectional height of the column.
In order to investigate the possible interactions of dif-
ferent parameters, the authors also inspected C7 and
C4AL of Scott, which were the only test data available
that investigates the joint aspect ratio. The notable differ-
ence between C7 and C4AL was that they had got differ-
ent beam reinforcement ratios as well as different joint
aspect ratios. Analysis of the test data confirmed the
reliability of Eq. (16).
11. Influence of Joint Stirrups
Without the influence of transverse reinforcement, the
authors’ design equation takes the following form:
0.71bg 冉 冊 冉
100Asb
b bd
0.4289
bc ⫹ bb
冊hc冑fc
冉冊
Vc ⫽ (17)
hb 0.61 2
hc
Fig. 6. The influence of the joint aspect ratio on the joint shear
strength normalised by the concrete sylinder strength and the beam
reinforcement ratio.
1112 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117
where b ⫽ 0.85 for joints detailed by U bars and b ⫽
1 for joints detailed by L bars. and g ⫽ 1.37 for inclined
bars in the joint and g ⫽ 1 for others.
The constant 0.71 is a capacity reduction factor
determined empirically. The authors have plotted the
joint shear strengths of the specimens in the experi-
mental database and their stirrup ratios in Fig. 7. It is
very apparent from the figure that the relation between
the stirrup ratio and the normalised joint shear strength
is tri-linear. Up to stirrup ratios of 0.003, the joints are
named as joints with low amount of stirrups. Joints with
joint stirrup ratios between 0.003 and 0.0055 are named
as joints with medium amount of stirrups. All the stirrups
yield within this region. When the stirrup ratio is higher
than 0.0055, not all the stirrups yield as evident from
the analysis of specimen BCJ7 of Ortiz and C4ALN3 of
Scott and Hamill and the contribution of stirrups to joint
shear strength substantially reduces under the yield
capacity of stirrups.
The stirrups that are considered effective in this study
are those that are situated above beam compressive
chord and below the top beam reinforcement. This is
because, the reported stirrup strains of Ortiz specimen
BCJ7 and BCJ4 both showed that the stirrup strains sub-
stantially reduce if positioned between the beam tensile
reinforcement and the beam compressive chord. Fig. 1
shows this.
It is commonly assumed in the literature [16] that joint
shear strength is given by the addition of concrete resist-
ance to shear, symbolising the resistance of the concrete
strut mechanism and stirrup yield capacity, symbolising
the resistance of the truss mechanism:
Vj ⫽ Vc ⫹ Asjefy (18)
where Vc is the joint shear strength of the concrete
(without stirrups); Asje is the area of the stirrups; fy is
the stirrup yield strength.
As mentioned above, the analysis of tests showed that
when the joints have a stirrup ratio higher than 0.0055,
not all the stirrups yield within the joint. The authors are
of the opinion that Eq. (18) should be corrected as:
Fig. 7. The influence of stirrup ratio on the joint shear strength.
Vj aAsjefy
⫽ vc ⫹ (19)
beffhc冑fc behc冑fc
Therefore the final equation can be formulated as:
冉 冊 冉 冊
Asb 0.4289 bc ⫹ bb
0.71bg 100∗ hc冑fc
bbdb 2
冉冊
Vj ⫽ (20)
hb 0.61
hc
⫹ aAsjefy
where a ⫽ 0.664 for joints with low amount of stirrups;
a ⫽ 0.6 for joints with medium amount of stirrups;
a ⫽ 0.37 for joints with high amount of stirrups
The stirrups not only affect the normalised joint shear
strength but also the failure modes of joints. The authors’
inspection of joints C4ALNO (no stirrup in joint) and
C4ALN1 (single stirrup in joint) of Scott and Hamill has
shown that provision of a single stirrup in joints which
have hc/db ratios less than 10 changes the failure mode
from connection zone reinforcement pull out to joint
shear failure. Furthermore, it is apparent from the datab-
ase that anchorage failures are not anticipated in joints
which have medium or high amount of stirrups.
12. Influence of column axial stress
The authors have depicted the relation between the
normalised joint shear strength and the column axial
stress in Fig. 8. There is considerable scatter in the
experimental data. The results show that the column
axial load certainly does not influence the joint shear
strength of monotonically loaded exterior beam column
connections. Vollum reaches a similar conclusion.
Nevertheless, the authors’ investigation of the speci-
mens of Parker and Bullman showed that column axial
load influences the behaviour of joints by changing their
failure modes. The authors’ comparison of Parker and
Bullmann specimen 4a with specimens 4b, 4c showed
that the former failed by column failure while the latter
two specimens all failed by joint shear failure. The
Fig. 8. The influence of the column axial stress on the normalised
joint shear strength.
 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117 1113

experimental data indicates that high column axial

stresses and high column longitudinal reinforcement
ratios are necessary in order to avoid column failures.

13. Critique of previously suggested design

equations for exterior beam column joints

In this section, the previously suggested empirical

equations are discussed and compared with the design
equation of authors. Because the influence of beam
longitudinal reinforcement ratio is not taken into
account, previous researchers’ estimates of the influence
of the joint aspect ratio on the joint shear strength are
flawed. None of the previously suggested design equa-
tions take account of the influence of the stirrups realisti-
cally. Using the authors’ new proposal, a more realistic
estimate of joint shear strength was obtained.
13.1. Remark on Sarsam and Phillips equation
The authors’ as well as Vollum’s analysis showed that
the column axial load does not influence the joint shear
strength. Contrary to the experimental evidence, Sarsam
and Phillips equation suggests that column axial load
influences the joint shear strength.
Furthermore, all stirrups are considered as effective in
Sarsam and Phillips equation while the tests of Ortiz
showed that the effective stirrups are those that are situ-
ated above the beam compressive chord and below the
beam reinforcement. Sarsam and Phillips equation pre-
dicts that all the transverse reinforcement yield. How-
ever, inspection of BCJ7 of Ortiz demonstrates clearly
that not all the stirrups yield in some joints.
Sarsam and Phillips equation predicts that column
longitudinal reinforcement ratio affects the joint shear
strength. However, the parametric study of authors’ in
Fig. 6 showed that there is no apparent relationship
between the column longitudinal reinforcement ratio and
the normalised shear strength of joints that fail by joint
shear failure.
The equation predicts that there is a linear relationship
between the concrete cube strength and the normalised
joint shear strength. The authors’ parametric study in
Fig. 3 showed that the normalised joint shear strength is
proportional to the square root of the concrete cylinder
strength. The equation of Sarsam and Phillips overesti-
mates the influence of the concrete cylinder strength on
the normalised joint shear strength.
Fig. 9. The influence of the beam reinforcement ratio on the predicted
normalised joint shear strength of the Vollum equation.
ratio on the joint shear strength. Because the effect of
beam longitudinal reinforcement ratio is disregarded, the
equation incorrectly predicts a linear relationship
between the joint aspect ratio and the joint shear
strength. If the above parameters are taken into account,
joint shear strength is predicted as proportional with the
0.61 power of the joint aspect ratio as given by the sug-
gested equation in this paper.
Vollum has not given the basis of the upper limits
chosen. It is not clear what the basis of the constant 0.97
is. Moreover, the author suggests that the other limit
1.33√fcbeffhc is chosen because it is the highest shear
strength of the database used by the author. So the accu-
racy of the above constraint is limited by the experi-
mental database used. The authors are of the opinion that
more experiments are needed to propose upper limits for
joint shear strength.
The authors plotted Vjpredicted/Vjactual values for both
Vollum and the authors’ equation against beam
reinforcement ratio, joint aspect ratio, stirrup ratio and
the stirrup index in Fig. 9–16. The figures clearly show
that the authors’ model is an improvement on the design
equation of Vollum. The suggested design equation is
more conservative than Vollum equation under varying
beam reinforcement ratio, joint aspect ratio, stirrup ratio
and the stirrup index.

13.2. Remark on the design equation of Vollum

As mentioned in the above paragraphs, beam

reinforcement ratio significantly influences the joint
shear strength of exterior joints. Vollum’s equation neg- Fig. 10. The influence of the beam reinforcement ratio on the pre-
lects the influence of beam longitudinal reinforcement dicted joint shear strength of author’s equation.
1114 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117

Fig. 11. The influence of the joint aspect ratio on the normalised joint Fig. 15. The influence of the stirrup index on the predicted joint shear
shear strength predicted by Vollum. strength of Vollum’s equation.

Fig. 12. The influence of the joint aspect ratio on the predicted shear Fig. 16. The influence of the stirrup index on the predicted joint shear
strength of the author’s equation. strength of the author’s equation/

14. Present design guidelines

Both the ACI-ASCE Committee 352 [8], and EC8 [9]

methods calculate the joint shear strength on the assump-
tion that the tensile reinforcement yield and for both
equations factors of safety are included. The above
methods are considered to be inadequate by the authors
because they neglect the influence of the stirrups, beam
longitudinal reinforcement ratio, and joint aspect ratio
on the joint shear strength.
Fig. 13. The influence of the stirrup ratio on the predicted joint shear
strength of the Vollum equation.
14.1. Comparison of the parametric studies with the
established principles of joint mechanics

In order to investigate the reliability of the parametric

Fig. 14. The influence of the stirrup ratio on the predicted joint shear
strength of the author’s equation.
studies in Fig. 2–9, the authors investigated the estab-
lished equations on the basic mechanics of reinforced
concrete beam-column joints. This has been also pre-
viously discussed by Paulay [17] and by Bonacci & Pan-
tazopoulou for interior joints [18] who have also taken
into account the joint deformations. Both of the authors
use the average stresses for equilibrium as shown in Fig.
18. The typical loading system considered in analysis of
exterior beam-column joints is shown in Fig. 17. Fig. 18
depicts the equilibrium of vertical and horizontal forces.
Fig. 18 shows that equilibrium of forces in the horizontal
direction require the average transverse compressive
stress in the joint sx defined as:
 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117 1115

I2 ⫽ sxsy ⫹ sysz ⫹ sxsz⫺t2av (24b)

I3 ⫽ sxsysz⫺s t 2
z av (24c)
The tensile stress in the concrete is negligible and
therefore s1 ⫽ 0, which consequently gives:
t2av
sy ⫽ (25)
sx
From the Mohr’s circle,
Fig. 18. Stress equilibrium.
2tav
tan2q ⫽ (26)
sx⫺sy
If Eq. (25) is substituted into Eq. (26), the following
quadratic equation ensues:

t2av ⫹ 冉 1
tanq 冊
⫺tanq sxtav⫺s2x ⫽ 0 (27)

which gives:
sx
tav ⫽ ⫺ (28)
tanq
Using Eq. (25), we have:
tav
sy ⫽ ⫺ (29)
tanq
Fig. 17. Joint geometry.
Collins and Mitchell [19] suggest the following equation
for the maximum stress in concrete panels:
Asb Asje
sx ⫽ ⫺ fs⫺ f (21) fc
beffhb beffhb w f2max ⫽ 具f (30)
0.8 ⫹ 170e1 c
where fs is the average stress in the beam reinforcement;
The principal compressive stress is given by:
fw is the average stress in the transverse reinforcement.
Consequently, the average normal concrete stress in
the y direction sy can be expressed as : s2 ⫽ 2 冉冉 e2
⫺0.002
⫺ 冊冉e2
⫺0.002 冊冊
2
f2max (31)
Ascol N
sy ⫽ ⫺ f ⫺ (22) s2 is also given from Mohr’s circle as:
beffhc scol beffhc
where fscol is the average stress in the column reinforce-
ment; N is the column axial load
s2 ⫽ sx ⫹ sy ⫽ ⫺tav tanq ⫹ 冉 1
tanq 冊 (32)

Defining the average joint shear stress in the joint as Thus the average joint shear stress can be expresses as:
tav, the maximum principal stress associated with the
stress tensor is given as; s2

冉 冊
tav ⫽ ⫺ (33)
1

冤 冥
tanq ⫹
sx tav 0 tanq
s ⫽ tav sy 0 (23) Eq. (30)–(33) show very clearly that as the principal
0 0 sz tensile strain increases, the average joint shear stress
decreases. Thus it is necessary to express the principal
tensile strain in terms of the strains in the x and y direc-
where sz is the confining stress provided by stirrups in tions in order to investigate the factors that influence the
the z direction. joint shear strength. From Mohr’s circle, it is known
s3 ⫽ I1s2 ⫹ I2s⫺I3 ⫽ 0 (24a) that:
In order to determine the principal stresses, Eq. (24a) g
tan2q ⫽ (34)
has to be solved; where I1 ⫽ sx ⫹ sy ⫹ sz ex⫺ey
1116 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117

冢冢
From Mohr’s circle, the principal tensile strain will be:

冪冉 冊 冉冊
(ex ⫹ ey) (ex⫺ey) 2
g 2 1 hb 1

冉 冉 冊冊
e1 ⫽ ⫹ ⫹ (35) e1 ⫽ tav (43)
2 2 2 hb 2 hc Asbm Asje
Es 1⫺ ⫹
hc beffhb beffhb
If Eq. (34) is substituted into Eq. (35) and appropriate

冉冊
trigonometric transformations are carried out, Eq. (36)

冣 冣
given by Bonacci and Pantazopoulou is obtained. hb 2
N

冉 冊
1 hc
ex⫺eytan2q ⫺ ⫹
e1 ⫽ (36) Ascol Ascol
1⫺tan2q beffhc
The next step will be to express the strains in the x
and y directions in terms of the stresses. The above equation shows that the principal tensile

冉
strain is increased by the joint aspect ratio and column
Asbfs Asjefw Asb longitudinal reinforcement ratio and the axial load on
sx ⫽ ⫺tavtanq ⫽ ⫺ ⫺ ⫽ m (37)
beffhb beffhb beffhb the column whereas it is decreased by increasing beam

冊
longitudinal reinforcement ratio and the stirrup ratio. The
Asje shear stress in the joint is dependent on the principal
⫹ f
beffhb w tensile strain as evident from Equations 30 and 33. It is
therefore evident from Eq. (28)(29) and (43) that the
where
joint shear strength increases as the beam longitudinal
m ⫽ fs / fw (38) reinforcement ratio and the transverse reinforcement
ratio increases. Eq. (29) shows that the joint shear
The strain in the x direction can therefore be strength increases as the column load and the column
expressed as: longitudinal reinforcement increases but Eq. (43) shows
fw tavtanq that as the longitudinal column reinforcement and the

冉 冊
ex ⫽ ⫽ (39) column load increases, the principal tensile stresses
Es Asbm Asje
Es ⫹ increase which consequently decreases the normalised
beffhb beffhb joint shear strength. Therefore the increase in the joint
The strain in the y direction can similarly be shear strength due to Eq. (29) is offset by the increase
expressed as: in the principal tensile strain. The above conclusions are
totally in accordance with the predictions of the

ey ⫽
fscol
Es
⫽
tav
冉
⫺
N beffhc
冊
tanq beffhc AscolEs
(40)
authors’ equation.

If Eq. (39) and (40) are substituted into Eq. (36): 15. Conclusions

冢 冢
The purpose of this investigation was to study the
1 1 effect of the parameters influencing the behaviour of
e1 ⫽ t tanq (41) beam to column connections and to determine if the
Es(1⫺tan2q) av Asb Asje
m⫹ present design guidelines are unconservative. From the
beffhb beffhb
analysis of the tests and results of the parametric studies,
the following design recommendations can be made.

⫺
1
Ascol
beffhc
⫹
冣 冣
Ntan2q

beffhc
Ascol
beffhc
1. All the experimental evidence points to the fact that
the U bar details should not be used in monotonically
loaded exterior beam column joints. As mentioned in
the upper paragraphs, providing L bars bent down
detail beam reinforcement can change the failure
It is evident from the inspection of experiments that modes of monotonically loaded exterior beam-column
cracks extend throughout the diagonal of the joint. So joints from joint shear to beam failure. Furthermore,
the angle of principal stresses can be expressed as: the joint shear strength of joints is increased by 15%
hb if detailed by L bars bent down detail beam reinforce-
tanq ⫽ (42) ment.
hc
2. Experimental evidence shows that anchorage failures
If Eq. (42) is substituted into Eq. (41), are not anticipated in joints with stirrups in monoton-
 P.G. Bakir, H.M. Boduroğlu / Engineering Structures 24 (2002) 1105–1117
ically loaded exterior beam column joints. In joints
without stirrups on the other hand, hc/db values should
be larger than 10 in order to avoid anchorage failures.
3. Column axial load has no influence on ultimate shear
capacity of the joint but high column axial load and
high column longitudinal reinforcement ratios are
necessary in the joint to avoid column failures.
4. Transverse reinforcement in the joint improves the
joint shear capacity but not in the same ratio as indi-
cated by the addition rule Vc+Vs. The authors plotted
the stirrup ratio against the joint shear strength. The
results showed that the diagram is trilinear. Based on
this parametric study, the authors classified the stir-
rups into three in increasing the joint shear strength.
Up to stirrup ratios of 0.003, the joints are named as
joints with low amount of stirrups. Joints with joint
stirrup ratios between 0.003 and 0.0055 are named as
joints with medium amount of stirrups. All the stir-
rups yield within this region. When the stirrup ratio
is higher than 0.0055, not all the stirrups yield as evi-
dent from the analysis of specimen BCJ7 of Ortiz and
C4ALN3 of Scott and Hamill and the contribution of
stirrups to joint shear strength substantially reduces
under the yield capacity of stirrups. Parametric studies
further carried out on the experimental database in
Table 1 showed that the addition rule Vc+Vs, should
be corrected as (Vc ⫹ aVs)beffhc冑fc where α is 0.664
for low amount of stirrups, 0.6 for medium amount
of stirrups, 0.37 for high amount of stirrups.
5. Increasing the beam longitudinal reinforcement ratio
increases the joint shear strength. Because the influ-
ence of beam longitudinal reinforcement ratio is taken
into account, the proposed equation predicts that the
joint shear strength is proportional to (hb/hc)⫺0.61.
6. The present guidelines, design equations and code
recommendations for predicting the shear strength of
monotonically loaded exterior beam column joints are
unconservative. The suggested design equation gives
more conservative and reliable results for predicting
the joint shear strength under varying beam longitudi-
nal reinforcement ratio, joint aspect ratio, stirrup ratio
as well as stirrup index.
7. The analysis of BCJ3 of Ortiz shows that the joint
shear strength will further increase if the vertical
anchorage length is higher than 26db and the radius
of bend of beam bars are higher than 8db. The authors
are of the opinion that these figures can be used as
lower limits for the radius of bend and vertical
anchorage length.
8. The authors applied their equation on the experi-
mental database in Table 1. The results showed that
the average Vjpredicted Vjtest values for the authors’ equ-
1117
ation applied on all the experiments in the experi-
mental database is 0.88 and the standard deviation is
0.1. The results show that the equation suggested
gives realistic and conservative estimates of the joint
shear strength.
9. The extremely good results of the proposed design
equation on the experimental database confirmed and
supported the value of the parametric studies.
References
[1] Bertero VV, Shadh H. Algeria earthquake, October 10, 1980.
Oakland, CA: Earthquake Engineering Research Institute, 1983.
[2] Mitchell D. Structural damage due to the 1985 Mexico earth-
quake. In: Proceedings of the 5th Canadian Conference on Earth-
quake Engineering. Rotterdam: A.A. Balkema; 1987. p. 87–111.
[3] Seismology Committee, Structural Engineers Association of Cali-
fornia. Recommended lateral force requirement and tentative
commentary, 1998.
[4] EERI Reconnaissance Team. Lome Prieta earthquake, October
17, 1989. Preliminary reconnaissance report. Oakland, CA: EERI.
[5] Bakir PG, Boduroglu MH. Earthquake risk and hazard mitigation
in Turkey. Earthquake Spectra, accepted for publication.
[6] Sarsam KF, Phillips ME. The shear design of insitu reinforced
beam-column joints subjected to monotonic loading. Magazine
of Concrete Research 1985;37(130):16–28.
[7] Vollum RL. Design and analysis of exterior beam column con-
nections. PhD thesis, Imperial College of Science Technology
and Medicine-University of London, 1998.
[8] Recommendations for design of beam-column joints for design
of beam column joints in monolithic reinforced concrete struc-
tures, 1991 edition. Reported by ACI-ASCE Committee 352,
ACI.
[9] Eurocode 8: design provisions for earthquake resistance of struc-
tures. London: BSI, 1995.
[10] Ortiz R. Strut and tie modelling of reinforced concrete short
beams and beam column joints. PhD thesis, University of
Westminster, 1993.
[11] Kordina K. Bewehrungsfuhrung in Ecken und Rahmenendknoten,
Deutscher Ausschuss für Stahlbeton, Heft 354, 1984.
[12] Scott RH. The effects of detailing on RC beam column connec-
tion behaviour. The Structural Engineer 1992;70(18):318–24.
[13] Scott RH, Hamill SJ. Connection zone strain in reinforced con-
crete beam column connections, In: Proceedings of the 11th Inter-
national Conference on Experimental Mechanics, Oxford, UK,
1998. pp. 65-69.
[14] Taylor HPJ. The behaviour of in situ concrete beam column
joints. Technical report 42.492, Cement and Concrete Associ-
ation, May 1974.
[15] Parker DE, Bullman PJM. Shear strength within reinforced con-
crete beam-column joints. The Structural Engineer
1997;75(4):53–7.
[16] Park R, Paulay T. Reinforced concrete structures. New York:
John Wiley and Sons, 1975.
[17] Paulay T. Critique of the special provisions for seismic design
of the building code requirements for reinforced concrete (ACI
318-83). ACI Journal 1986; 83(2): 274, 283.
[18] Pantazopoulou S, Bonacci J. Consideration of questions about
beam-column joints. ACI Structural Journal 1992;89(1):27–37.
[19] Collins MP, Mitchell D. Prestressed concrete structures. Engle-
wood Cliffs, NJ: Prentice-Hall, 1991.