Eccentric
Beam-Column
Connections
Performance and design of joints subjected to seismic lateral load reversals
BY JAMES M. LAFAVE, JOHN F. BONACCI, BURCU BURAK, AND MYOUNGSU SHIN
B eam-column connections are critical regions in
reinforced concrete (RC) moment-resisting frame
structures designed to resist strong earthquakes. Many
experimental and analytical research studies have been
conducted to investigate the behavior of RC beam-column
connections subjected to seismic loading and to establish
guidelines for design. From these studies, several key
design parameters governing the behavior of RC beam-
column connections have been identified, such as the
relative column-versus-beam flexural strengths at the
connection (including slab effects), confinement of the
joint core, joint shear stress, and anchorage of reinforcement
in the connection region. Joint ACI-ASCE Committee 352,
Joints and Connections in Monolithic Concrete Structures,
has integrated the available research results into a state-
of-the-art report entitled, “Recommendations for Design
of Beam-Column Connections in Monolithic Reinforced
Concrete Structures (ACI 352R-02).”1 Many of the research
findings are also reflected in seismic design provisions
found in Chapter 21 of ACI 318-05, Special Provisions for
Seismic Design.2
When an RC beam-column connection is subjected
to lateral earthquake loading, the beam tension T and
compression C forces from bending are transmitted
to the joint at the beam-column interfaces, producing
relatively large joint shear forces. In an eccentric connection,
58 SEPTEMBER 2005 / Concrete international
the column centerline is offset in plan from the beam
centerline (by an eccentricity e, as indicated in Fig. 1),
thereby concentrating the joint shear toward one side of
the joint. Due to this eccentricity, the transmitted forces
can also introduce torsion into the joint, adding to the
shear stresses. ACI 352R-02 addresses eccentric RC
beam-column connections to only a limited extent. The
provisions for eccentric connections are based on
information about their poor joint shear performance in
practice3 and on two research studies,4,5 where a total of
six experimental subassemblies were tested to assess
the effect of eccentricity on joint behavior. The design
approach taken in Section 4.3 of ACI 352R-02 is simply
to place additional restrictions on the permissible joint
shear force in connections where the eccentricity
between the beam and column centerlines e exceeds
1/8 of the width of the column bc . Section 21.5.3 of
ACI 318-05 has adopted a similar approach, with provisions
for assessing the design joint shear strength of eccentric
connections that are in some cases even more restrictive
than those in ACI 352R-02. Codes and guidelines for
seismic design of RC structures elsewhere in the world
(New Zealand and Japan, for example) use comparable
approaches to address eccentric joints.
Section A.1 of ACI 352R-02 further notes that the effect
of eccentric beams on joints is an area in need of additional
 Beam centerline
Flush side T e

C
bb e bc C
Column centerline
T
hc
(a) Offset side (b)

Fig. 1: (a) Plan view; and (b) isometric view of typical eccentric RC beam-column connection (floor slab not shown)

research. This is particularly so in light of how common

this type of connection is in exterior RC building frames,
where beams often frame into a column with flush
outside beam and column faces (as in a building that
exhibited noticeable joint damage in a recent strong
earthquake6). In response to this need, Joint ACI-ASCE
Committee 352 appointed a task group to investigate,
review, and summarize all research currently available on
the subject of eccentric RC beam-column connections
and to propose design recommendations compatible with
ACI 352R-02, as appropriate. This article is the product of
those efforts.

SUMMARY OF AVAILABLE RESEARCH

Eleven research studies on eccentric RC beam-column
connections subjected to reverse-cyclic lateral loading
were found in the literature, comprising nearly 40 test
specimens (experimental subassemblies). The tests
included 22 interior (cruciform) connections4,5,7-14 and 15
corner connections.15,16 A brief summary of many of these
testing programs can be found elsewhere.17 Five of the
interior connections were edge connections with a floor
slab and transverse beam on one side only (see Fig. 2).9,10,14
The rest of the interior connections did not have floor slabs
or transverse beams.4,5,7,8,11-13 Of the corner connections,
four out of five in one study had spread-ended (tapered
width) beams to minimize eccentricity at the joint,15 while
all 10 in the other study were so eccentric that the beam
often framed into a perpendicular girder rather than
directly into the column.16 Modest axial column
compression loads were applied to some specimens
prior to testing.4,5,8,10-12,14
It was noted in most of the studies that eccentric
connections had somewhat lower joint shear strengths
and an earlier onset of strength degradation (less
distortion capacity) than similar concentric connections.
Some researchers proposed addressing this by using a
modified (reduced) effective joint width in design for
eccentric connections.5,8,9,11,15 Eccentric connections were
Fig. 2: Eccentric RC beam-column edge connection test with floor
slab and transverse beam
almost always observed to experience greater stiffness
degradation in the joint than comparable concentric
connections due to extensive joint cracking at the flush
face of the column and beams. This damage was typically
noted in conjunction with larger joint shear deformations
and/or larger joint hoop strains near the flush side of the
joint.4,5,8,13,15 However, differences in joint hoop strains
between the flush exterior side of a joint and the offset
interior side were much less when a floor slab and
transverse beam were present.9,10 The small differences
seen in such cases were actually quite similar to those
noted in concentric edge connections with floor slabs
and transverse beams.17

Concrete international / SEPTEMBER 2005 59

TABLE 1:
EFFECTIVE JOINT WIDTHS FOR ECCENTRIC RC BEAM-COLUMN CONNECTIONS
bb , bc , Vj,m , bj,est , bj,est / bj,est / b′j,est / bj,est /
Specimen in. in. e/bc kips in. bj,318 bj,352 bj,318 (bb+bc)/2
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Joh, Goto,
JX0-B5 5.9 11.8 0.25 66.0 8.03 1.36 1.05 1.09 0.91
and Shibata4
1 10 14 0.14 146 13.5 1.35 1.12 1.08 1.12

Raffaelle and 2 7 14 0.25 94.6 9.02 1.29 0.99 1.03 0.86

Wight5 3 7.5 14 0.23 106 8.54 1.14 0.89 0.91 0.79

4 7.5 14 0.23 92.7 10.4 1.39 1.09 1.11 0.97

Teng and S3 7.9 15.7 0.25 161 15.9 2.02 1.64 1.61 1.34
Zhou8 S6 7.9 15.7 0.25 87.9 12.5 1.56 1.36 1.25 1.04

Shin and 1* 11 18 0.19 145 14.1 1.29 1.09 1.03 0.97

LaFave9 2* 7 18 0.31 146 13.0 1.85 1.45 1.48 1.04

Burak and 2-S* 10 21 0.26 194 15.4 1.54 1.27 1.23 0.99
Wight10 3-S* 10 21 0.26 186 17.1 1.71 1.42 1.37 1.10

UM-60 7.9 17.7 0.13 175 20.7 1.59 1.81 1.27 1.62
Goto and Joh11
UM-125 7.9 17.7 0.28 148 17.3 2.20 1.79 1.76 1.35
Kamimura,
Takimoto, and NN.2 7.1 13.8 0.16 94.2 11.0 1.17 1.13 0.93 1.06
Tanaka12

Kusuhara JE-55 7.1 12.6 0.17 104 12.6 1.53 1.35 1.22 1.28
et al.13 JE-55S 7.1 12.6 0.17 105 12.7 1.54 1.36 1.23 1.29

Average 1.53 1.30 1.23 1.11

Standard deviation 0.30 0.28 0.24 0.22

1.14 0.89 0.91 0.79
Range
to 2.02 to 1.81 to 1.76 to 1.62
Note: 1 in. = 25.4 mm; 1 kip = 4.45 kN
*
Indicates specimens with floor slab and transverse beam on one side of joint

The relatively flexible joints observed with eccentric
RC beam-column connections were found to contribute
about 1/6 to 1/4 of the overall story displacements at
low to moderate subassembly drifts (approximately 1 to
2%). The contribution increased to about 1/3 to 1/2 of
the overall subassembly drift at larger displacements
(when significant joint damage had occurred).4,5,8,9,12,14,15
Finally, the presence of a floor slab was observed to
add to the total possible joint shear demand in an
eccentric connection, but it also appears that the slab
(along with the transverse beam, when present) effectively
reduced the connection eccentricity and aided in the
resistance mechanisms.9,10 Both eccentric and concentric
edge connections typically exhibited greater effective
slab widths than those commonly prescribed for use
in design.17
EFFECTIVE JOINT WIDTHS
To better evaluate the joint shear capacity of eccentric
RC beam-column connections, a detailed examination was
conducted of all 16 eccentric connection subassemblies
found in the literature that were reported to have failed

60 SEPTEMBER 2005 / Concrete international

due to joint shear.4,5,8-13 They are listed in numbered
Columns 1 to 4 of Table 1 along with their beam width
bb, column width bc, and normalized eccentricity e/bc
values (see also Fig. 1). These specimens were all
interior connections (most of which had one beam face
flush with the exterior face of the column) and had well-
distributed and well-detailed joint hoop reinforcement.
Four of the specimens had a floor slab and transverse
beam on one side of the joint, as noted below the table.
During testing, a few of the specimens underwent beam
plastic hinging in conjunction with joint shear failure. The
joint shear strength for each specimen was considered to
be equal to the maximum shear force applied to the joint
during the test (Vj,m) and is given in Column 5 of Table 1.
The method used to determine the maximum joint shear
force from the maximum applied column shear force
during testing is provided elsewhere.17
The estimated effective joint width bj,est, shown in
Column 6 for each specimen, was calculated from the
experimentally-determined maximum joint shear
force as
(1)
where ƒ′c is the actual measured concrete compressive
strength (psi) and hc is the column depth (in.), with bj,est
in inches, Vj,m in lb, and √ƒ′c in psi. The joint shear stress
factor specified by ACI 352R-02 and ACI 318-05 (γn = 12) is
used to place eccentric connections on an equal basis for
comparison with similar concentric connections. (If Eq. (1)
were used with Vj,m in N, ƒ′c in MPa, and hc and bj,est in mm,
then γn would be 1.0.) For the 16 specimens, ƒ′c ranged
from about 2800 to 5600 psi (19 to 39 MPa), and the
column cross-sectional aspect ratio (hc /bc) ranged from
0.50 to 1.00.
Table 1 also contains ratios of the estimated effective
joint width bj,est to the effective joint widths computed
following ACI 318-05 (bj,318) and ACI 352R-02 (bj,352). Per
ACI 318-05, bj,318 = bb + 2x, where x is the smaller distance
between the beam and column edges. Most of the
connections tabulated are one-sided (flush) eccentric
connections where the bj,318 value is by definition simply
equal to the beam width (bb). Per ACI 352R-02,
bj,352 = bb + Σmhc /2, where m is 0.3 when e is greater than
bc /8 and m is 0.5 otherwise. The ACI 318-05 effective joint
width definition could be conservative (simply representing
the width of the eccentric beams in most of the tabulated
cases). Such a narrowly defined joint region might be
considered as effectively confined on at least two
opposite vertical faces, a case where γn is typically taken
as 15 in design (1.25 if N and mm units are used). The b ′j,est
values used in the table were computed using Eq. (1) with
γn = 15. Finally, the estimated experimental effective joint
width of each specimen was compared to the basic
effective joint width definition currently used by
ACI 352R-02 for concentric connections, namely
bj = (bb + bc)/2.
The ACI 318-05 approach (see Column 7 of the table)
systematically underestimates the joint shear strength of
eccentric connections (by an average of more than 50%),
especially in cases with floor slabs and transverse beams.
The ACI 352R-02 effective joint width definition (tabulated
in Column 8) was found to be about 30% conservative
on average. A comparison of Column 9 to Column 7 in the
table indicates that the computed b ′j,est values were closer
to the bj,318 values than were the bj,est values. Finally, the
simple effective joint width (see Column 10 in the table)
equal to the average of the beam and column widths
matched the estimated experimental value fairly well on
average and had the smallest standard deviation of all
the approaches examined. This effective joint width
definition worked especially well for cases with floor
slabs and transverse beams. While it was not conservative
for all cases without slabs, in only one specimen did it
give a nominal joint shear strength prediction slightly
less than 85% of the experimentally determined value.
RECOMMENDATIONS
Based on the preceding information, it is recommended
that an effective joint width (bj) equal to the average of the
beam and column widths [(bb + bc )/2] can be used to estimate
the joint shear strength of eccentric RC beam-column
connections for design. This recommendation should be
considered in conjunction with the other design provisions
of ACI 352R-02 such as using a design yield stress multiplier
of at least 1.25 and including the contribution to joint
shear forces (“demand”) of slab reinforcing steel within
an appropriate effective tension flange width.
Acknowledgments
The authors would like to thank fellow members of Joint ACI-ASCE
Committee 352, Joints and Connections in Monolithic Concrete
Structures, for their constructive comments and suggestions
regarding the subject of this article.
References
1. Joint ACI-ASCE Committee 352, “Recommendations for Design
of Beam-Column Connections in Monolithic Reinforced Concrete
Structures (ACI 352R-02),” American Concrete Institute, Farmington
Hills, MI, 2002, 37 pp.
2. ACI Committee 318, “Building Code Requirements for Structural
Concrete (ACI 318-05) and Commentary (ACI 318R-05),” American
Concrete Institute, Farmington Hills, MI, 2005, 430 pp.
3. Ohno, K., and Shibata, T., “On the Damage to the Hakodate
College by the Tokachioki Earthquake, 1968,” Proceedings of the U.S.-
Japan Seminar of Earthquake Engineering with Emphasis on the Safety
of School Buildings, Sendai, 1970, pp. 129-144.
Concrete international / SEPTEMBER 2005 61
 4. Joh, O.; Goto, Y.; and Shibata, T., “Behavior of Reinforced
Concrete Beam-Column Joints with Eccentricity,” Design of Beam-Column
Joints for Seismic Resistance, SP-123, James O. Jirsa, ed., American
Concrete Institute, Farmington Hills, MI, 1991, pp. 317-357.
5. Raffaelle, G.S., and Wight, J.K., “Reinforced Concrete Eccentric
Beam-Column Connections Subjected to Earthquake-Type Loading,”
ACI Structural Journal, V. 92, No. 1, Jan.-Feb. 1995, pp. 45-55.
6. Hirosawa, M.; Akiyama, T.; Kondo, T.; and Zhou, J., “Damages
to Beam-to-Column Joint Panels of RC Buildings Caused by the 1995
Hyogo-Ken Nanbu Earthquake and the Analysis,” Proceedings of the
12th World Conference on Earthquake Engineering, Paper No. 1321,
Auckland, New Zealand, 2000.
7. Lawrance, G.M.; Beattie, G.J.; and Jacks, D.H., “The Cyclic Load
Performance of an Eccentric Beam-Column Joint (Report 91-25126),”
Central Laboratories, Lower Hutt, New Zealand, 1991, 81 pp.
8. Teng, S., and Zhou, H., “Eccentric Reinforced Concrete
Beam-Column Joints Subjected to Cyclic Loading,” ACI Structural
Journal, V. 100, No. 2, Mar.-Apr. 2003, pp. 139-148.
9. Shin, M., and LaFave, J.M., “Seismic Performance of Reinforced
Concrete Eccentric Beam-Column Connections with Floor Slabs,”
ACI Structural Journal, V. 101, No. 3, May-June 2004, pp. 403-412.
10. Burak, B., and Wight, J.K., “Seismic Behavior of Eccentric
RC Beam-Column-Slab Connections Under Sequential Loading in
Two Principal Directions,” Innovations in Design with Emphasis on
Seismic, Wind and Environmental Loading; Quality Control and
Innovation in Materials/Hot-Weather Concreting, SP-209, V.M.
Malhotra, ed., American Concrete Institute, Farmington Hills, MI,
2002, pp. 863-880.
11. Goto, Y., and Joh, O., “Shear Resistance of RC Interior
Eccentric Beam-Column Joints,” Proceedings of the 13th World
Conference on Earthquake Engineering, Paper No. 649, Vancouver,
BC, Canada, 2004, 13 pp.
12. Kamimura, T.; Takimoto, H.; and Tanaka, S., “Mechanical
Behavior of Reinforced Concrete Beam-Column Assemblages with
Eccentricity,” Proceedings of the 13th World Conference on Earth-
quake Engineering, Paper No. 4, Vancouver, BC, Canada, 2004, 10 pp.
13. Kusuhara, F.; Azukawa, K.; Shiohara, H.; and Otani, S., “Tests
of Reinforced Concrete Interior Beam-Column Joint Subassemblage
with Eccentric Beams,” Proceedings of the 13th World Conference on
Earthquake Engineering, Paper No. 185, Vancouver, BC, Canada, 2004,
14 pp.
14. Burak, B., and Wight, J.K., “Experimental Investigation of
Eccentric Reinforced Concrete Beam-Column-Slab Connections Under
Earthquake Loading,” Proceedings of the 13th World Conference on
Earthquake Engineering, Paper No. 2150, Vancouver, BC, Canada, 2004,
14 pp.
15. Chen, C.C., and Chen, G.K., “Cyclic Behavior of Reinforced
Concrete Eccentric Beam-Column Corner Joints Connecting
Spread-Ended Beams,” ACI Structural Journal, V. 96, No. 3, May-
June 1999, pp. 443-449.
16. Vollum, R.L., and Newman, J.B., “Towards the Design of
Reinforced Concrete Eccentric Beam-Column Joints,” Magazine of
Concrete Research, V. 51, No. 6, Dec. 1999, pp. 397-407.
17. Shin, M., and LaFave, J.M., “Reinforced Concrete Edge
62 SEPTEMBER 2005 / Concrete international
Beam-Column-Slab Connections Subjected to Earthquake Loading,”
Magazine of Concrete Research, V. 56, No. 5, June 2004, pp. 273-291.
Received and reviewed under Institute publication policies.
ACI member James M. LaFave, PE, is an
Associate Professor of Civil Engineering
at the University of Illinois at Urbana-
Champaign. He is the Chair of Joint
ACI-ASCE Committee 352, Joints and
Connections in Monolithic Concrete
Structures, and is a member of ACI
Committees 439, Steel Reinforcement,
and E 802, Teaching Methods and Educational
Materials. His research interests include earthquake-resistant
design of reinforced concrete structures and durability of
structural concrete.
John F. Bonacci, FACI, is an Associate
Professor of Civil Engineering at the
University of Toronto, Ontario, Canada.
He is a member and past Chair of Joint
ACI-ASCE Committee 352, Joints and
Connections in Monolithic Concrete
Structures, and is a member of ACI
Committees 374, Performance-Based Seismic
Design of Concrete Buildings; 318-D, Flexure
and Axial Loads: Beams, Slabs, and Columns; and Joint ACI-ASCE
Committee 445, Shear and Torsion.
ACI member Burcu Burak is a PhD candidate
in civil engineering at the University of
Michigan, Ann Arbor, MI. She is an
associate member of Joint ACI-ASCE
Committee 352, Joints and Connections
in Monolithic Concrete Structures. Her
research interests include earthquake-
resistant design, analysis, and rehabilitation
of reinforced concrete and fiber-reinforced
composite structures.
ACI member Myoungsu Shin is an Assistant
Professor of Industrial and Engineering
Technology at Morehead State University,
Morehead, KY. He received his PhD in civil
engineering from the University of Illinois at
Urbana-Champaign. He is an associate
member of Joint ACI-ASCE Committee 352,
Joints and Connections in Monolithic Concrete
Structures. His research interests include
earthquake-resistant design of reinforced concrete structures.