Seismic Design of Rib-Reinforced Steel Moment

Connections based on Equivalent Strut Model

Cheol-Ho Lee, M.ASCE1

Abstract: This paper describes a seismic design procedure for rib-reinforced steel moment connections based on an equivalent strut
Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.

model. It is shown from the results of finite-element analysis that the force transfer mechanism in the rib connections is completely
different from that predicted by the classical beam theory and that a diagonal strip in the rib acts as a strut. By treating the rib as a strut,
an equivalent strut model that could be used as the basis of a practical design procedure is proposed first. A step-by-step design procedure
is then recommended based on the proposed model.
DOI: 10.1061/共ASCE兲0733-9445共2002兲128:9共1121兲
CE Database keywords: Seismic design; Struts; Connections; Finite-element method; Models; Steel.

Introduction et al. 2000; Lee and Uang 2001兲. Especially, it was noted that an
inclined strip in the web of the straight haunch acts as a strut
The 1994 Northridge, California, earthquake caused widespread rather than following the beam theory. Lee and Uang viewed the
brittle fracture in connections of steel moment-resisting frames. A web of a straight haunch as a vertical rib plate and the haunch
variety of improved moment connection details have been pro- flange as a stability element. It was speculated that there exists a
posed after the earthquake. Now there is plenty of evidence that close link between the rib and the straight haunch.
the original welded connection with notch tough weld metal and a
few detail improvements 共the so-called WUF-W connection兲 can Objective and Scope
provide a very ductile response 共Ricles et al. 2000兲. However, the The main objective of this study was to propose a practical design
more popular strategies that have been developed include procedure for rib-reinforced steel moment connections. To this
strengthening the connection or weakening the beams that frame
into the connection 共Bruneau et al. 1998兲. The aim is, based on
the capacity design concept, to shift the plastic hinging away from
the face of the column, thus reducing the possibility of brittle
failure conditions. Fig. 1 shows an example of a rib-reinforced
moment connection 共the test specimen COH-1, Zekioglu et al.
1997兲. In this case, the rib reinforcement was used to supplement
the taper-cut reduced beam section 共RBS兲, i.e., to further limit the
stress in the beam flange welds and to provide increased redun-
dancy for the connection. Rib reinforcement may also be used to
address the situation where the frame design requires an excessive
RBS 共greater than 50% of the beam flange兲 due to short spans, or
larger beam depths 共Iwankiw 1997兲.
Engineers often use rib plates to enhance the seismic perfor-
mance of welded steel moment connections, thinking that the mo-
ment of inertia is increased near the face of the column so that the
tensile stress in the groove weld is reduced. However, such think-
ing is brought into question in this study. Previous studies have
indicated that the classical beam theory cannot provide reliable
force transfer predictions in steel moment connections with a
welded haunch 共Lee and Uang 1997; Lee and Uang 2000; Yu

1
Professor, Dept. of Architectural Engineering, Kyungnam Univ.,
Masan, South Korea, 631-701. E-mail: chlee@kyungnam.ac.kr
Note. Associate Editor: Brad Cross. Discussion open until February 1,
2003. Separate discussions must be submitted for individual papers. To
extend the closing date by one month, a written request must be filed with
the ASCE Managing Editor. The manuscript for this paper was submitted
for review and possible publication on July 12 2001; approved on January
3, 2002. This paper is part of the Journal of Structural Engineering, Vol.
128, No. 9, September 1, 2002. ©ASCE, ISSN 0733-9445/2002/9-
Fig. 1. Rib connection detail 共Zekioglu et al. 1997兲
1121–1129/$8.00⫹$.50 per page.

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002 / 1121

J. Struct. Eng. 2002.128:1121-1129.

Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.
Fig. 2. Finite-element mesh for specimen COH-1
end, a simplified equivalent strut model that could be used as the
basis of a design procedure was established first.
Finite Element Analysis and Observations
To gain insight into the behavior of rib connections, the test speci-
men COH-1 was modeled and analyzed using the general-purpose
finite-element analysis program ABAQUS, user’s manual, version
5.8 共1998兲, Hibbit, Karlson & Sorenson, Inc., Pawtucket, R.I. Fig.
2 shows the finite-element mesh in the connection region. The
specimen consisted of a W27X178 beam and W14X455 column
and had a rib length of 229 mm 共9 in.兲, a rib height of 165 mm
共6.5 in.兲, and a rib thickness of 25 mm 共1 in.兲. Both the flanges
and web of the beam and column were modeled with an eight-
node continuum element 共element type C3D8 in ABAQUS兲. The
beam web was directly connected to the column flange in the
model. Steel material properties obtained from tensile coupon
tests were used. Material nonlinearity with the von Mises yielding
criterion was considered in the analysis. In this study comparison
of the cyclic behavior of the model with experimental results was
not tried since this study is concerned with the behavior of the rib
connection, and simple monotonic analysis was conducted to
check the overall validity of the model. The analytically predicted
load versus beam tip deflection relationship was correlated with
the response envelope of the test result in Fig. 3. The correlation
was reasonable. The finite-element model was then used to inves-
tigate the behavior of the rib connection.
Fig. 3. Correlation of analytical and experimental global responses
1122 / JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002
J. Struct. Eng. 2002.128:1121-1129.
Fig. 4. Typical beam span with rib connection supplemented by
radius-cut reduced beam section
Plastic hinging of the beam is often assumed to occur at the rib
tip. However, it is difficult to justify this assumption due to the
light reinforcement nature of the rib. The typical beam span with
the rib connection assumed in this study is shown in Fig. 4, where
the radius-cut RBS is introduced to both confine the plastic hing-
ing of the beam effectively outside the rib region and to push the
occurrence of local buckling away from the rib tip. Since the rib
region is expected to remain essentially elastic under this scheme,
an elastic analysis with a concentrated load applied at the beam
tip was conducted to study the force transfer mechanism of the
connection.
Based on the analysis results, the flexural stress profile at the
column face is presented in Fig. 5. The flexural stress profile from
the beam theory, by treating the beam and ribs as an integral
section, is also presented. It is evident from Fig. 5 that the force
transfer mechanism in the rib connection cannot be reliably pre-
dicted by the beam theory. Note that the beam theory underesti-
mates significantly the stress at the beam flanges, i.e., the stress in
the beam flange groove welds. Based on the finite-element analy-
ses, the percentage of the beam shear transferred by each element
at the face of the column is shown in Fig. 6. The ribs transfer
Fig. 5. Comparison of flexural stress profiles along beam depth
共specimen COH-1兲
Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 6. Shear transfer at column face 共specimen COH-1兲

168% of the beam shear applied and produce ‘‘reverse shear’’ in

the beam web. Again, this phenomenon cannot be explained by
the beam theory. The principal stress distribution shown in Fig. 7
suggests clear diagonal strut action in the rib. This strut action can
be used to explain the reverse shear phenomenon noted above.
Fig. 8. Two types of rib configuration
Analyses were performed to investigate the stress distribution
at the beam–rib interface. In addition to analyzing the test speci-
men COH-1 with a rib length a of 229 mm 共9 in.兲, a rib height b matches the rib diagonal angle. This observation will be used in
of 165 mm 共6.5 in.兲, and a rib thickness t of 25 mm 共1.0 in.兲, reducing the number of unknown forces at the beam–rib interface
additional cases were also included in the parametric study by in the next section.
varying the rib slope and the rib thickness within some practical Fig. 10 compares the deformed shapes of the two types of rib
ranges. Similar parametric study was also conducted for the configuration. It is noted that the load path of the single rib con-
single rib configuration 关see Fig. 8共a兲兴. figuration is more direct because all of the beam web, rib, and
The normal and shear stress distributions along the beam–rib column web plates exist in the co-plane and it does not accom-
interface are presented in Fig. 9, where the stress profile of each pany beam flange bending, which is unavoidable in the dual rib
case has been normalized by the maximum stress and the distance configuration. Accordingly, it is expected that the load transferred
has been normalized by the rib length. Similar distributions were by one rib in the dual rib configuration will be smaller than that
also obtained when the rib thickness was varied 共results not by one rib in the single rib configuration under the same rib size.
shown兲. Fig. 9共a兲 shows that resultant normal force N is located at To confirm this expectation, Q and N values from the finite-
a distance of approximately 0.60a from the face of the column, element analysis for both configurations are compared in Table 2.
where a is the rib length. Fig. 9共b兲 shows that the shear stress
profile is insensitive to the variation of rib length. Stress concen-
tration is also evident at the rib tip. The total shear force at the
beam–rib interface is defined as Q. The resultant angles (␪ R )
from force components N and Q, are compared with the rib diag-
onal angles (␪ G ) in Table 1. As suggested from the principal
stress plot in Fig. 7, the resultant angle of Q and N reasonably

Fig. 9. Typical normal and shear stress distributions along beam–rib

Fig. 7. Principal stress distribution in rib interface

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002 / 1123

J. Struct. Eng. 2002.128:1121-1129.

 Table 1. Comparison of Rib Diagonal and Resultant Angles Table 2. Comparison of Q and N Values in Single and Dual Rib
Configurations
Single Rib Dual Rib
Rib length 共mm兲 Q N Qa Na
a⫽178 共1:1 slope兲 1.0 1.0 0.99b 0.98
a⫽229 共1.5:1 slope兲 1.0 1.0 1.09 1.10
a⫽305 共2:1 slope兲 1.0 1.0 1.15 1.15
a
Sum of two rib forces.
b
Relative values.

␪ R ⫽tan⫺1(N/Q) 共deg兲 ers the strut action in the rib, is proposed for practical design
␪ G ⫽tan⫺1(b/a) purposes. First, equivalent strut area A e is defined as
Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.

共deg兲 Single rib Dual rib

␩ 共 ab⫺c 2 兲 t
42.9 39.4 41.8 A e⫽ (1)
35.8 35.2 36.1 冑共 a⫺c 兲 2 ⫹ 共 b⫺c 兲 2
28.4 30.9 32.9 where a⫽rib length, b⫽rib height, c⫽trimmed length, t⫽rib
thickness, and ␩⫽equivalent strut area factor 共refer to Fig. 11兲.
The overall features of the proposed model are shown in Fig. 12,
It is observed that the load transfer by one rib in the single rib where V pd is the design beam shear which can readily be calcu-
configuration is about two times that by one rib in the dual rib lated by using Eq. 共16兲 in the next section. In the model, the rib is
configuration. Additional analyses also confirmed this observa- replaced by an equivalent strut 共or a simple truss element兲. The
tion. All these observations noted above were incorporated in es- angle and location of the strut incorporate the findings from the
tablishing an equivalent strut model in the following section. finite-element analysis results. The equivalent strut force corre-
sponds to the resultant of Q and N at the beam–rib interface in the
actual connection. The key remaining problem is to determine
Equivalent Strut Model equivalent strut area factor ␩ in a way which will give the strut
force acting in the actual rib. Theoretically, the number of degrees
Based on the observations from the finite-element analysis results of freedom along the beam–rib interface is infinite. It seems
in the previous section, an equivalent strut model, which consid- hardly feasible to determine the equivalent strut area factor by a
purely analytical method due to both the continuum problem na-
ture and the complexity of the boundary conditions around the rib
connection. A calibration approach is adopted in this study.
Considering the results summarized in Table 1, normal force N
can be computed with reasonable accuracy by using Eq. 共2兲.

N⫽ 冉冊
b
a
Q (2)

Then, horizontal shear force Q remains as the only unknown. The

relationship between Q and ␩ is established first by applying the
horizontal deformation compatibility condition at the strut tip
共point A in Fig. 12兲.

Horizontal Displacement Component of Equivalent

Strut at Point A: d x (strut)
By referring to Figs. 11 and 12, the axial stiffness of equivalent
strut k e is computed as Eq. 共3兲

Fig. 10. Comparison of deformed shapes Fig. 11. Definition of rib cross section width

1124 / JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002

J. Struct. Eng. 2002.128:1121-1129.

 A eE ␩ 共 ab⫺c 2 兲 tE Note that Eq. 共2兲 was used in Eq. 共6兲. The beam moment produces
k e⫽ ⫽ (3) a compressive strain in the beam top flange
Le 共 0.60兲 冑共 a 2 ⫹b 2 兲 冑共 a⫺c 兲 2 ⫹ 共 b⫺c 兲 2
where
M共x兲 db
␧ 共 x 兲 flange⫽ ⫻
L e ⫽ 共 0.60兲 冑共 a ⫹b 兲
2 2 (7)
EI b 2
(4)
E⫽Young’s modulus The horizontal component of the beam deformation at point A is
The horizontal displacement component of the equivalent strut at obtained by integrating the axial strain of the beam top flange
point A is thus obtained by dividing Q with k e as follows: within the equivalent strut region as follows:
冑Q 2 ⫹N 2 cos共 ␪ G 兲 Q

冕
d x 共 strut兲 ⫽ ⫽ 0.60a
ke ke
d x 共 beam兲 ⫽ ␧ 共 x 兲 flange dx
Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.

共 0.60兲 冑共 a 2 ⫹b 2 兲 冑共 a⫺c 兲 2 ⫹ 共 b⫺c 兲 2

0

⫽ ⫻Q
冕
(5)
␩ 共 ab⫺c 2 兲 tE db 0.60a
⫽ M 共 x 兲 dx
2EI b 0

Horizontal Displacement Component of Beam Flange

共 0.21a 2 ⫹0.15aL ⬘ 兲 d b
at Point A: d x (beam) ⫽ ⫻V pd
EI b
From Fig. 12, the beam moment within the equivalent strut region
is
M 共 x 兲 ⫽V pd 共 x⫹0.40a⫹L ⬘ /2兲 ⫺2Nx⫺Qd b
⫺ 冉 0.18abd b ⫹0.30ad 2b
EI b
冊 ⫻Q (8)

⫽V pd 共 x⫹0.40a⫹L ⬘ /2兲 ⫺2 冉冊 b
a
Qx⫺Qd b (6)
The relationship between Q and ␩ is obtained by equating Eqs.
共5兲 and 共8兲 for deformation compatibility as follows:

冉 共0.21a⫹0.15L ⬘ 兲 ad b
Ib 冊 ⫻V pd
冉冊
Q⫽ (9)
1 共 0.60兲 冑共 a 2 ⫹b 2 兲 冑共 a⫺c 兲 2 ⫹ 共 b⫺c 兲 2 共 0.18b⫹0.30d b 兲共 ad b 兲
⫹
␩ 共 ab⫺c 2 兲 t Ib

To calibrate the equivalent strut area factor, horizontal shear force
values Q obtained from the finite-element analysis within some
practical range of the rib slope and thickness were inserted into
Eq. 共9兲. The results are summarized in Table 3. Although the
factors vary more or less depending on the rib slope and thick-
ness, they are fairly stable. The average value is close to 1.50.
With an equivalent strut area factor of 1.50, Table 4 shows the
degree of accuracy of Q and N values predicted by using Eqs. 共2兲
and 共9兲. Using a larger equivalent strut area factor 共say, 1.80兲
produces only slightly larger predictions. Thus, it is recom-
mended, for practical design purposes, that equivalent strut area
factor ␩ be taken as

␩⫽1.50 (10)
Knowing the two interface forces, the moment at the column face
is

M f ⫽V pd 共 a⫹L ⬘ /2兲 ⫺Qd b ⫺2N 共 0.60a 兲 (11)

Flexural stress f b f in the beam flange groove can be calculated by

elementary mechanics as follows:

Table 3. Equivalent Strut Area Factors

Rib Thickness 共mm兲
Rib length 共mm兲 0.6t(⫽15) 1.0t(⫽25) 1.4t(⫽35)
a⫽178 共1:1 slope兲 1.65 1.44 1.31
a⫽229 共1.5:1 slope兲 1.72 1.52 1.31
Fig. 12. Interaction model for deformation compatibility a⫽305 共2:1 slope兲 1.85 1.65 1.48

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002 / 1125

J. Struct. Eng. 2002.128:1121-1129.

 Table 4. Predictions of Q and N Values with ␩⫽1.50
Q N
Rib Thickness 共mm兲 Rib Thickness 共mm兲
Rib length 共mm兲 0.6t(⫽15) 1.0t(⫽25) 1.4t(⫽35) 0.6t(⫽15) 1.0t(⫽25) 1.4t(⫽35)
a⫽178 mm 共1:1 slope兲 0.93 a
1.03 1.09 0.99 0.96 1.01
a⫽229 mm 共1.5:1 slope兲 0.91 0.99 1.08 0.91 1.02 1.11
a⫽305 mm 共2:1 slope兲 0.86 0.94 1.01 0.76 1.14 0.92
a
Predicted values were normalized by the corresponding finite-element analysis results.

M f V pd 共 a⫹L ⬘ /2兲 ⫺Qd b ⫺1.2Na 2M pd

Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.

f bf⫽ ⫽ (12) V pd ⫽ ⫹V G (16)

S S 共 L⫺2e 兲
where S⫽elastic section modulus of the beam. The flexural where V G ⫽beam shear due to the gravity load between the beam
stresses based on Eq. 共12兲 are compared in Fig. 13 with the flex- plastic hinges.
ural stress profiles obtained from both the beam theory and finite-
element analysis. The flexural stress obtained from Eq. 共12兲 is
Strong Column – Weak Beam Condition
satisfactory.
By applying equilibrium equations to the free body in Fig. 14, the
strong column–weak beam condition can be checked as follows:
Recommended Design Procedure 兺M*
pc
⬎1.0 (17)
The recommended step-by-step design procedure is presented 兺M*
pb

冉 冊
below. H C ⫺d b ⫺2b
兺 M *pb ⫽ 兺 关 M pd ⫹V pd共 e⫹d c /2兲兴 HC
(18)
Preliminary Rib Sizing
Consider a portion of the moment frame in Fig. 4. Based on the 兺 M *pc ⫽ 兺 Z c共 F yc ⫺ f a 兲 (19)
available experimental and analytical database, suggested depth b
and angle ␪ G b ⫽tan⫺1(b/a)c are chosen as where Z c ⫽plastic section modulus of the column, F yc ⫽nominal
yield strength of the column, and f a ⫽column axial stress.
b⬇ 共 41 – 15 兲 d b (13)
␪ G ⬇30° – 40° (14) Rib Design
Determine interaction forces Q and N by employing Eqs. 共2兲, 共9兲,
Beam Design Moment and Shear and 共10兲. Consider the free body of the rib shown in Fig. 15.
Based on the von Mises yield criterion, the minimum rib thick-
For design purposes, the plastic hinge is assumed to develop at ness can be computed as follows:
the narrowest section in the RBS 共see Fig. 4兲. Following the AISC
Seismic Provisions 共Seismic 1997兲, the design beam plastic mo- N
f v⫽ (20)
ment is bt
M pd ⫽1.1Z RBSF ye (15) Q
f n⫽ (21)
where Z RBS and F ye ⫽plastic section modulus at the narrowest bt
section in the RBS and expected yield strength of the beam, re-
spectively. The corresponding beam shear is then

Fig. 13. Comparison of flexural stress predictions Fig. 14. Free body for strong column–weak beam condition

1126 / JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002

J. Struct. Eng. 2002.128:1121-1129.

 共1兲 Fillet weld 共two sided兲 between the rib and beam:

冑冉 Q
0.707S⫻2a 冊 冉
2
⫹
N
0.707S⫻2a 冊 2
⭐␾ 共 0.6F EXX 兲 (25)

where ␾⫽0.75 and F EXX ⫽tensile strength of the welding elec-

trode. Solving for fillet weld size S gives
冑Q 2 ⫹N 2
S⭓ (26)
0.636共 a 兲共 F EXX 兲
共2兲 Fillet weld 共two sided兲 between the rib and column flange:

冑冉 Q
冊 冉
2
N
冊 2
Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.

Fig. 15. Free body of rib ⫹ ⭐␾ 共 0.6F EXX 兲 (27)

0.707S⫻2b 0.707S⫻2b
Solving for fillet weld size S gives
f 2n ⫹3 f 2v ⭐ 共 ␾F y 兲 2 (22)
where ␾⫽0.90. Substituting Eqs. 共20兲 and 共21兲 into Eq. 共22兲 and 冑Q 2 ⫹N 2
S⭓ (28)
solving for t gives 0.636共 b 兲共 F EXX 兲
冑Q 2 ⫹3N 2 For the dual rib configuration, use 50% of Q and N values in
t⭓ (23) Eqs. 共26兲–共28兲. Complete joint penetration groove weld may be
b 共 ␾F y 兲
used to connect the rib to the column flange to avoid the clearance
For the dual rib configuration, use 50% of Q and N values in Eqs. problem in the field welding.
共20兲–共23兲.
A designer can use the flexural stress formula given in Eq. 共12兲
to ensure that the beam flange groove weld stress level is within Design of Continuity and Doubler Plates
some limit. For example,
The requirement in Chap. K of the LRFD Specification 共LRFD
f b f ⭐F ye (24) 1993兲 should be followed to design continuity and doubler plates.
The requirement in Eq. 共24兲 will limit the maximum moment Based on Eq. 共11兲, the force developed in the beam flange, P u f ,
carried by the beam at the column face to the level of actual first can be calculated as follows:
yield moment resistance at beam end 共Iwankiw 1997兲.
Mf
Puf⫽
0.95d b
Rib Weld Design
The rib can be fillet welded to both the column and beam flanges V pd 共 a⫹L ⬘ /2兲 ⫺Qd b ⫺1.2Na
⫽ (29)
to transmit normal force N and shear force Q 共see Fig. 15兲. 0.95d b

Table 5. Summary of Example Design Calculations

Relevant
Design step Calculations equations
Preliminary rib sizing 1
Assume b⫽( 4 )d b ⫽150 mm, ␪ G ⫽35°; t⫽15, and c⫽50 mm; 共13兲,共14兲
a⫽b/tan(35°)⬇220 mm.

Beam design moment and shear M pd ⫽1.1Z RBSF ye ⫽(1.1)(2,269,000)(314)⫽783,713 kN mm, 共15兲,共16兲
L ⬘ ⫽L⫺2a⫽4,560 mm, e⫽a⫹X/2⫽415 mm, L⫺2e⫽4170 mm; V pd ⫽417 kN.

Strong column–weak beam condition 兺M* pb ⫽共2,043,904 kN mm兲/共780,126 kN mm兲⫽2.62⬎1.0

pc / 兺 M * 共17兲–共19兲

Rib design 共1兲 Interaction forces; Q⫽637 kN, N⫽(150/220)(637)⫽434 kN; 共2兲,共9兲,共10兲
共2兲 Flexural stress level; f b f ⫽211 MPa⬍F ye ⫽314 MPa; 共24兲
共3兲 Rib thickness check; assuming dual rib configuration; 共23兲
Q共per rib兲⫽637/2⫽319 kN, N 共per rib兲⫽434/2⫽217 kN,
t⫽15 mm ⭓ 冑Q 2 ⫹3N 2 /(b␾F y )⫽11.2 mm.

Rib weld design 共1兲 Fillet weld 共two-sided兲 between the rib and beam 共26兲
S⭓ 冑Q 2 ⫹N 2 /(0.636aF EXX )⫽5.6 mm; use 6 mm fillet weld;
共2兲 Use CJP groove weld to connect the rib to column flange.

Design of continuity and doubler plates Beam flange force P u f ⫽957 kN; no doubler plates were needed in this case and the 共29兲
continuity plates were provided according to the SAC 共2000兲 recommendations.

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002 / 1127

J. Struct. Eng. 2002.128:1121-1129.

Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.
Fig. 16. Connection details of example design
Design Example and Analytical Verification
By assuming some realistic design conditions, the example design
was conducted by following the proposed design procedure. In
this example, the connection design by the weakening strategy
alone is not feasible, because a short span length 共assumed as 5 m
long兲 requires an excessive RBS 共greater than 50% of the beam
flange兲. The rib was designed by assuming a 30% radius cut of
the beam flange. Design conditions were as follows.
Frame Data
An exterior moment connection was considered with the follow-
ing data: story height H c ⫽3600 共mm兲; bay width L⫽5,000
共mm兲; beam H-600X200X11X17 共SS400 steel, F y ⫽236 MPa,
F ye ⫽314 MPa兲; column H-400X408X21X21 共SM490 steel, F yc
⫽325 MPa兲; rib plate 共SM490 steel, F y ⫽325 MPa兲; a filler metal
with a specified CVN value of 26.7 J 共20 ft lb兲 at ⫺28.9°C
共⫺20°F兲 is to be used to connect the beam flanges and rib plates
to the column (F EXX ⫽492 MPa兲; a uniformly distributed gravity
load for the beam w g ⫽19.6 kN/m; and column axial stress f a
⫽69 MPa.
Member Section Properties
H-600X200X11X17 beam; d b ⫽600 共mm兲 I b ⫽776,000,000
(mm4) S⫽2,590,000 (mm3) Z b ⫽2,979,000 (mm3), RBS proper-
ties 共30% cut of the beam flange assumed兲; X⫽390 共mm兲
Y ⫽30 共mm兲 R⫽649 共mm兲 Z RBS⫽2,269,000 (mm3),
H-400X408X21X21 column; d c ⫽400 共mm兲; and Z c ⫽3,992,000
(mm3).
Detailed calculations are not presented here due to space limi-
tations. The results for each step are summarized in Table 5. The
connection detail drawings from this example design are pre-
sented in Fig. 16. ABAQUS analysis was conducted for the exte-
rior beam–column subassembly. The equivalent plastic strain
共PEEQ兲 distribution, shown in Fig. 17, indicates that the proposed
design procedure works as intended; the rib in combination with
1128 / JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002
J. Struct. Eng. 2002.128:1121-1129.
Fig. 17. Equivalent plastic strain distribution at 4% story drift ratio
response
the RBS effectively pushed the plastic hinging of the beam out-
side the rib region. In the finite-element model, a circular hole
was included near the rib tip to reduce stress concentration; the
finite-element analysis results showed that drilling a circular hole
decreases the magnitude of the horizontal shear, vertical normal,
and von Mises stresses at the rib tip by about 17, 2, and 4%,
respectively.
Conclusions
The main conclusions on the seismic design of steel moment
connections reinforced with ribs are summarized as follows.
1. The force transfer mechanism in the rib connections is com-
pletely different from that predicted by the classical beam
theory. The flexural stress prediction from the beam theory,
treating the beam and ribs as an integral section, underesti-
mates significantly the stress in the beam flange groove
welds. The diagonal strip in the rib acts as a strut, and strut
action tends to produce reverse shear in the beam web.
2. Idealizing the rib as a strut, an equivalent strut model that
can be used to determine the interaction forces at the inter-
face between the beam and rib was proposed. With an
equivalent strut area factor of 1.50, the proposed model pre-
dicted the interaction forces satisfactorily. Based on the pro-
posed model, a step-by-step design procedure was also rec-
ommended.
Acknowledgments
Financial support was provided by the Korea Earthquake Engi-
neering Research Center 共KEERC兲 under Grant No. 2000G0206.
The writer would like to express his special gratitude to Professor
C.-M. Uang at UCSD for support to this study.
References
Bruneau, M., Uang, C.-M., and Whittaker, A. 共1998兲. ‘‘Ductile design of
steel structures.’’ Design of ductile moment-resisting frames,
MaGraw-Hill, New York, 273–380.
 Iwankiw, N. 共1997兲. ‘‘Ultimate strength consideration for seismic design
of the reduced beam section 共internal plastic hinge兲.’’ Eng. J. 34共1兲,
3–16.
Lee, C.-H., and Uang, C.-M. 共1997兲. ‘‘Analytical modeling of dual panel
zone in haunch repaired steel MRFs.’’ J. Struct. Eng., 123共1兲, 20–29.
Lee, C.-H., and Uang, C.-M. 共2000兲. ‘‘Analytical modeling of seismic
steel moment connections with welded straight haunch.’’ Proc., SEE-
BUS 2000, Kyoto, Japan, 31–140.
Lee, C.-H., and Uang, C.-M. 共2001兲. ‘‘Analytical modeling and seismic
design of steel moment connections with welded straight haunch.’’ J.
Struct. Eng., 127共9兲, 1028 –1035.
Load and resistance factor design specification. 共1993兲. 2nd ed., AISC,
Chicago.
Downloaded from ascelibrary.org by UTEP LIBRARY-SERIALS on 08/20/14. Copyright ASCE. For personal use only; all rights reserved.
Ricles, J. M., Mao, L., Kaufmann, E. J., Lu, L., and Fisher, J. W. 共2000兲.
‘‘Development and evaluation of improved details for ductile welded
unreinforced flange connections.’’ Rep. No. SAC BD 00-24, SAC Joint
Venture, Sacramento, Calif.
SAC 共2000兲. ‘‘Seismic design criteria for new moment-resisting steel
frame construction.’’ Rep. No. FEMA 350, SAC Joint Venture, Sacra-
mento, Calif.
Seismic provisions for structural steel buildings. 共1997兲. 2nd Ed., AISC,
Chicago.
Yu, Q.-S., Uang, C.-M., and Gross, J. 共2000兲. ‘‘Seismic rehabilitation
design of steel moment connection with welded haunch.’’ J. Struct.
Eng., 126共1兲, 69–78.
Zekioglu, A., Mozaffarian, H., Chang, K. L., and Uang, C.-M. 共1997兲.
‘‘Designing after Northridge.’’ Modern Steel Constr., 37共3兲, 36 – 42.

JOURNAL OF STRUCTURAL ENGINEERING / SEPTEMBER 2002 / 1129

J. Struct. Eng. 2002.128:1121-1129.