You are on page 1of 10

Engineering Structures 25 (2003) 681690

www.elsevier.com/locate/engstruct

Towards deformation-based capacity design of RCS beamcolumn


connections
G.J. Parra-Montesinos , X. Liang, J.K. Wight
Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI 48109-2125, USA

Abstract

Hybrid RCS frame systems consisting of reinforced concrete (RC) columns and steel (S) beams have been recognized to possess
several advantages in terms of structural performance and economy compared to pure RC and steel frames. However, their use has
been limited to low or moderate seismic regions due to lack of appropriate design guidelines for RCS frames in high seismic risk
zones. In this article, the implementation of a deformation-based capacity approach for the design of RCS connections in frames
located in high seismic risk zones is described. The main purpose of the design procedure is to control joint distortions and damage,
forcing most of the frame inelastic activity to occur in the beams. The design procedure is validated through the testing of two RCS
beamcolumnslab subassemblies under displacement reversals. Experimental results indicate that the proposed design procedure is
effective in controlling joint deformations and damage, leading to the formation of plastic hinges at the beam ends. The test
specimens showed excellent seismic behavior with significant energy dissipation capacity, which demonstrates the suitability of
RCS frames for use in high seismic risk zones.
2002 Elsevier Science Ltd. All rights reserved.

Keywords: Composite structures; Hybrid structures; RCS frames; Earthquake-resistant structures; Beamcolumn connections; Joint shear defor-
mation; Deformation-based design

1. Introduction have been conducted in the US and Japan to study the


interaction between steel and concrete members in RCS
Hybrid RCS frame systems consisting of reinforced frames, especially in the connections between steel
concrete (RC) columns and steel (S) beams have been beams and RC columns. From testing programs conduc-
increasingly used during the past two decades. These ted in the late 1980s [2,3], design guidelines for RCS
structural systems offer several advantages from the joints in buildings located in low to moderate seismic
structural and economic viewpoints. RC columns have zones were developed by an ASCE task committee [4].
been shown to be significantly more cost-effective than In these guidelines, the proposed strength equations were
steel columns and steel floor systems are substantially calibrated for joint strengths measured at 1% (0.01 rad)
lighter than RC floor systems, reducing building weight joint distortion, which corresponds to approximately the
and foundation costs. In addition, typical construction of same level of story drift. However, these design equa-
RCS frames is performed by first erecting a steel frame tions were not strictly based on joint deformations, but
with light columns, which allows the spreading of con- primarily on strength equations developed for steel and
struction tasks along the height of the frame structure RC beamcolumn connections.
[1]. From a research program conducted at the University
The advantages offered by RCS frame systems have of Michigan during the late 1990s [5], it was concluded
led to increasing attention from researchers and prac- that RCS connections designed following a strength-
titioners. Since the 1980s, several research programs based approach could experience large deformations
under cyclic loading, leading to severe joint damage and
excessive story drifts. In addition, the design equations

Corresponding author. Tel.: +1-734-764-6576; fax: +1-734-764- from the ASCE guidelines could be unconservative
4292. when used for exterior connections. Therefore, a new
E-mail address: gjpm@umich.edu (G.J. Parra-Montesinos). joint model, based on joint shear deformations, was

0141-0296/03/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0141-0296(02)00177-3
682 G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690

developed for interior and exterior RCS connections. used to mobilize the concrete regions between the beam
This model allows the estimation of the force vs. joint flanges; (2) two-part U-shaped stirrups passing through
shear deformation response of RCS joints, as well as the holes in the steel web panel; (3) a light steel column
strains in the concrete panel and steel reinforcement. welded to the beam flanges, which is used for erection
In this article, design equations based on joint shear purposes and to mobilize the concrete regions outside
deformation levels corresponding to moderate damage the beam flanges; and (4) closed stirrups just above and
are proposed for use in RCS beamcolumn connections below the steel beam to resist the outward thrust pro-
in zones of high seismicity. The effectiveness of the duced by the transfer of shear forces through bearing of
design procedure is evaluated through the design and the steel columns on the surrounding concrete [3].
reversed-cyclic load testing of two RCS beamcolumn For the case when steel transverse beams frame into
slab subassemblies. The behavior of the test specimens is the column (Fig. 2a), a joint design consisting of the use
evaluated in terms of loaddisplacement response, joint of steel band plates wrapping around the column regions
deformations and beam rotations. The ability of the just above and below the steel beam is proposed (Fig.
design equations to limit joint deformations, and thus 2b). These band plates are connected to the beam flanges
damage, is evaluated, as well as the contributions from through stiffeners and, if properly designed, can provide
different mechanisms to total story drift. sufficient confinement to the joint region to eliminate the
need for steel transverse reinforcement [7]. In this arti-
cle, the proposed deformation-based capacity design pro-
2. Recommended details for RCS connections cedure is applied to RCS connections with either stan-
dard details or steel band plates.
A large number of details have been proposed for RCS
connections in the US and Japan. This makes the appli-
cability of RCS construction difficult since design rec- 3. Deformation and failure modes in RCS
ommendations need to be available for each joint detail. connections
Because one of the targets of a design code should be
clearness and simplicity, two simple joint details that In RC and steel beamcolumn connections, the domi-
have been shown to behave well under large displace- nant type of deformation is given by panel shear defor-
ment reversals [3,6,7] are proposed for use in RCS con- mations (Fig. 3a). However, in RCS connections, bear-
struction located in high seismic risk zones. These two ing of the steel beam flanges on the surrounding concrete
details are shown in Figs. 1 and 2. In these two proposed can lead to concrete crushing and the opening of a gap
designs, the steel beam passes continuously through the that allows a rigid body rotation (RBR) of the steel beam
column, representing what is referred to as a through- inside the joint (Fig. 3b). Therefore, the total distortion
beam RCS connection. in RCS connections is given by panel shear deformation
The joint details shown in Fig. 1b correspond to the and RBR of the steel beam inside the joint. Conse-
standard design for RCS connections and can be used quently, two main failure modes have been identified in
when no transverse beams frame into the joint (Fig. 1a). RCS connections: (1) panel shear failure, and (2) bearing
These standard details consist of: (1) face bearing plates failure [4]. A panel shear failure is characterized by sig-

Fig. 1. RCS joint with standard details.


G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690 683

Fig. 2. RCS joint with steel band plates.

Fig. 3. RCS joint distortions.

nificant diagonal cracking of the concrete panels and steel web panel, (2) inner diagonal concrete strut, and
yielding of the steel web panel in the joint region. A (3) outer diagonal concrete strut. The state of plane strain
bearing failure is characterized by crushing of the con- in the connection is defined through the development of
crete regions adjacent to the steel beam flanges. Kanno a ratio, ktc, between the principal tensile and compression
[6] reported that a panel shear failure generally triggers strains. For the same level of joint shear deformation,
a bearing failure, but a bearing failure does not always this ratio is assumed larger for exterior joints compared
cause a shear failure. Therefore, even though the bearing to interior joints. From the state of plane strain in the
strength of an RCS connection could be significantly connection, and using appropriate constitutive relations
higher than its shear strength, attention must be given to for steel and concrete, the force vs. shear distortion
the RBR of the steel beam inside the joint. response of RCS joints can be estimated. In addition, the
From testing of exterior RCS connections under large strains in the joint transverse reinforcement and diagonal
displacement reversals, it was concluded that the concrete struts can be estimated from the joint model.
maximum shear distortion that RCS connections can sus- Detailed information on this model can be found else-
tain without severe damage and loss of stiffness is where [5,8].
approximately 0.75% [7]. Therefore, if only minor to The bearing behavior of RCS connections was exten-
moderate joint damage is to be allowed under a major sively studied by Kanno [6]. Because the bearing
earthquake, it seems reasonable to design the connec- response of RCS joints is significantly less stable than
tions for a target joint shear distortion of 0.5%. the joint shear mechanisms [9], joint bearing failures
In order to study the relation between joint shear force must be prevented in the design of RCS connections.
and shear distortion in RCS connections, Parra-Montes- However, as mentioned earlier, panel shear failures gen-
inos and Wight [5] developed a model based on the state erally cause subsequent bearing failures. To account for
of plane strain in the hybrid connection. This model the effect of RBR of the steel beam inside the joint
assumes that at any given level of joint shear defor- region, even though the connection is designed to pos-
mation, the strength is given by three mechanisms: (1) sess sufficient bearing strength, a simple relation
684 G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690

between the normalized joint shear force and beam RBR horizontal shear strength of the steel web panel, Vwh, can
was proposed by Parra-Montesinos and Wight [5], and be estimated as follows:
is shown in Fig. 4. In this figure, the normalized joint
fy
shear force is given by the ratio between joint shear Vwh kw
3
twhc (1)
force, Vj, and ultimate joint shear strength, Vuj. As shown
in Fig. 4, excessive beam RBR can occur when the con-
nection is subjected to shear forces approximately equal where kw 0.9 and 0.8 for interior and exterior joints,
respectively, fy and tw are the yield strength and thick-
to its ultimate shear strength.
ness of the steel web panel, respectively, and hc is the
column depth. The horizontal shear strength of the inner
4. Deformation-based capacity design of RCS and outer diagonal concrete struts, Vih and Voh, respect-
connections ively, can be estimated from Eqs. (2) and (3) as follows:
Vih ki fc (0.0048 fc 1.13) hc(bftw) (2)
The design of RCS connections involves the checking
of bearing and panel shear strengths. As mentioned earl- Voh kofc (0.0048 fc 1.13)hcbo (3)
ier, because of the more stable behavior observed in where ki and ko are strength factors, and bf and bo are
joints with panel shear failure compared to those exhibit- the width of the beam flange and outer concrete panel,
ing bearing failure, the bearing strength of an RCS con- respectively. Procedures to determine the width of the
nection should be greater than or equal to its joint shear outer concrete panel, bo, can be found elsewhere [5]. In
strength. Bearing strength of RCS connections can be Eqs. (2) and (3), fc must be expressed in MPa. Values
checked following the procedure given in the ASCE of ki and ko for joint shear distortions of 0.5 and 1.2%
guidelines [4], which assumes a maximum concrete are listed in Table 1.
bearing strength of 2fc, where fc is the concrete com- Once the design shear strength of the connection is
pressive strength. In the ASCE guidelines, the bearing estimated for a maximum joint shear distortion of 0.5%,
stresses are assumed to act over an area given by the the RBR of the steel beam inside the joint can be esti-
effective joint width and 30% of the column depth. mated from Fig. 4, based on the ratio between the design
The shear strength of the connections can be determ- shear strength and the ultimate shear strength of the con-
ined based on the joint model developed by Parra-Mon- nection.
tesinos and Wight [5]. From this model, design equations The design of the composite joint should be performed
are proposed in this article for a target joint shear distor- so that the maximum shear force transferred to the con-
tion of 0.5% in order to limit joint damage and excessive nection does not exceed the design shear strength of the
story drifts. Equations are also proposed for estimating joint. The maximum horizontal joint shear force, (Vj)max,
the ultimate shear strength of the connections, calibrated can be approximated as follows:
for a joint shear deformation of 1.2%, which would cor-
respond to significant damage in the connection.
As was mentioned earlier, the shear strength of the
(Mu)beam
(Vu)col (Vj)max (4)
dbeamtf
connection is given by the contributions from the steel
web panel, and the inner and outer concrete panels. The assuming a lever arm between the joint horizontal shear
resistant forces equal to the steel beam depth, dbeam,


minus the beam flange thickness, tf. In Eq. (4),
(Mu)beam is the summation of the ultimate moment

Table 1
k factors recommended for RCS joints

Target shear k Factor Joint detail


deformation (%)

Standard Steel band plates

Interior Exterior Interior Exterior

0.5 ki 0.25 0.17 0.32 0.24


ko 0.14 0.10 0.17 0.12
1.2 ki 0.32 0.21 0.40 0.29
ko 0.17 0.11 0.22 0.15
Fig. 4. Normalized load vs. beam RBR response.
G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690 685

capacity of the composite beams in one plane framing the concrete deck, the slab bars did not pass through the
into the connection and (Vu)col is the column shear force column. Shear studs with a diameter of 19 mm and
when the beams reach their ultimate moment strength. spaced at 200 mm were used along the length of the
Appropriate material overstrength and strain hardening beam, including the plastic hinge region, to assure a full
effects, as given in AISC [10], should be considered composite action between the steel section and the con-
when determining the beam ultimate moment strength. crete slab. The size of the beam was chosen such that a
beam yielding mechanism would dominate the behavior
of the specimen at large drift levels with only moderate
5. Behavior of RCS beamcolumnslab damage in the connection region, as will be shown later.
subassemblies with joints designed following a The specimens were subjected to 20 lateral displacement
deformation-based capacity approach cycles with story drifts ranging from 0.5% to a
maximum of 5%. The lateral displacement history is
To evaluate the effectiveness of the proposed design shown in Fig. 6.
procedure for RCS connections, two RCS beamcol-
umnslab subassemblies were designed, constructed and
tested under lateral displacement cycles at the Structures 6. Material properties
Laboratory in the University of Michigan. The two test
specimens represented interior connections with the two For Specimens 1 and 2, concrete with compressive
joint configurations shown in Figs. 1 and 2. The test strengths of 38.6 and 40.7 MPa, respectively, was used
setup used for this testing program is shown in Fig. 5. in the columns and joint regions. For the concrete deck
Each test specimen consisted of a 2240 mm long RC
column with a 380 380 mm2 cross-section. The longi-
tudinal column reinforcement consisted of 12#19 bars,
representing a steel ratio of 2.4%. A W6 16 steel sec-
tion was embedded in the RC column and welded to the
beam flanges as shown in Fig. 1b. A column axial load
equivalent to approximately 4% of the column axial
capacity was applied to the test specimens.
The steel beam used in the test specimens consisted
of a 4480 mm long W12 16 steel shape with a 1220
mm wide and 90 mm thick concrete slab poured on a
steel deck supported by the steel beam. As shown in Fig.
5, the ends of the RC column and steel beam were
pinned, assuming inflection points at column mid-height
and beam mid-span, respectively. The slab was
reinforced with 4#10 deformed bars in the longitudinal
direction and #10 bars at 300 mm in the transverse direc-
tion. Because the RC column was constructed prior to Fig. 6. Lateral displacement pattern.

Fig. 5. Test setup.


686 G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690

a compression strength of 44.2 and 31.0 MPa was posite W12 16 steel beam-concrete slab using actual
determined on the testing day for Specimens 1 and 2, material strengths and assuming a 10% increase in steel
respectively. Dual Grade A36/A572-Grade 50 steel was strength due to strain hardening, a maximum shear force
used for the wide flange beams. The actual yield strength transferred into the connection of 1100 kN (lateral force
of the dual grade steel was approximately 320 MPa. of 160 kN) was estimated, which is similar to the design
Grade 60 steel was used for the longitudinal and trans- strength of the connection. Even though the expected
verse reinforcement in the RC column and slab. joint shear deformation for a lateral force of 160 kN,
0.75%, is larger than the proposed limit of 0.5%, it has
been suggested as the maximum deformation for which
7. Design of test specimens no significant damage and stiffness deterioration is
expected in RCS joints [7]. With respect to joint bearing
In the following, a brief description of the design pro- strength of Specimen 1, it was approximately equal to
cedure used for the test specimens is given. For the the ultimate joint shear strength, based on the ASCE
determination of joint and beam strengths, actual guidelines [4]. The RC column in Specimen 1 was
material properties were used. designed to remain elastic during the test and sufficient
transverse reinforcement just above and below the steel
7.1. Specimen 1 beam was used to resist the outward thrust caused by
bearing of the steel column on the surrounding concrete,
Specimen 1 had a standard joint detail, as shown in as suggested by Deierlein et al. [3]. The size of the steel
Fig. 1b, with no transverse beams framing into the con- column and face bearing plates was checked following
nection. The joint transverse reinforcement used between the ASCE design guidelines [4].
the beam flanges consisted of three two-part #13 U-
shaped stirrups passing through holes in the web of the
7.2. Specimen 2
steel beam. This amount of transverse reinforcement rep-
resented a stirrup volume to joint volume ratio of 1.1%,
which is larger than the minimum of 0.9% recommended Specimen 2 represented the recommended design for
by Parra-Montesinos and Wight [7]. The estimated lat- the case of transverse beams framing into the connec-
eral load vs. joint shear deformation envelope for Speci- tion, as shown in Fig. 2. For this specimen, the joint
men 1 is shown in Fig. 7. Also, the specimen lateral detail consisted of the use of 13 mm thick steel band
loads corresponding to design and ultimate strength of plates wrapping around the column regions just above
the connection are shown in the figure. The design and and below the steel beam. These band plates, which
ultimate strength of the hybrid joint are given by the extended 25% of the steel beam depth above and below
summation of the shear strength obtained from Eqs. (1) the steel beam, were connected to the beam flanges
through (3) and using the strength factors listed in Table through 13 mm thick stiffeners located at the middle of
1. The size of the steel beam to be used in the specimen the beam flanges. The ratio between the volume of each
was dictated by the design lateral strength of approxi- band and the joint volume was 3.3%. A light steel col-
mately 150 kN obtained for Specimen 1. For a com- umn welded to the beam flanges was embedded in the
RC column, as for Specimen 1. The RC column in
Specimen 2 was identical to that of Specimen 1 except
for the connection region. Because no joint transverse
reinforcement is used when transverse beams frame into
the connection, special attention must be given to the
unsupported length of the longitudinal column bars. To
avoid buckling of these bars, small ties, anchored in the
column core, were used in between the beam flanges
(Fig. 2b). Fig. 7 shows the predicted lateral load vs. joint
shear distortion envelope curve for this specimen. The
design and ultimate lateral strengths are also shown in
this figure. It should be noted that the design lateral
strength of Specimen 2 (170 kN) is very similar to the
maximum lateral force expected in Specimen 1 (160
kN), and thus the use of a W12 16 steel beam was
suitable to assure a strong columnweak beam response
with limited joint damage in the specimen. In Specimen
Fig. 7. Predicted lateral load vs. joint shear deformation response for 2, bearing was not expected to be a problem because of
test specimens. the use of the steel band plates, which provide excellent
G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690 687

confinement to the concrete regions just above and specimens up to story drifts approximately equal to
below the steel beam flanges. 1.0%. Little pinching of the loaddisplacement hyster-
esis loops due to cracking of the joint region can be
noticed for story drifts up to 2.0%. However, because
8. Experimental results beam plastic hinging dominated the specimen response
for larger amplitude cycles, full hysteresis loops with
8.1. Concrete cracking and beam yielding pattern excellent energy dissipation capacity were obtained at
large drift levels. In addition, little loss of stiffness dur-
Diagonal cracking in the joint region of Specimens 1 ing repeated cycles at the same drift level was observed.
and 2 started at approximately 0.5% story drift. At this At approximately 3.0% drift, both specimens reached
drift level, hairline flexural cracks that spanned over the their maximum lateral strength (approximately 160 kN),
full slab width were observed with an average spacing which was governed by the plastic moment strength of
approximately equal to the depth of the composite beam. the composite steelconcrete beams. At this displace-
Column flexural-bearing cracks, which originated from ment level, a drop in the load was observed in Specimens
the corners of the steel beam flanges, were also noticed 1 and 2, primarily caused by buckling of the steel beam
at low drift levels. At 1.5% drift, new diagonal cracks bottom flanges. However, the response of the specimens
had formed in the connection region and the slab flexural was stable up to story drift levels larger than 4.0%.
cracks became wider. In addition, yielding of the steel To better illustrate the amount of deformations experi-
beam flanges near the column faces had started to take enced by the composite steelconcrete beams, a typical
place. At 3.0% story drift, the joint region was crossed moment vs. rotation response for a beam plastic hinge
by several diagonal cracks, but joint damage was moder- region of Specimen 1 is shown in Fig. 10. For both
ate (Fig. 8). At this displacement level, significant yield- specimens, beam rotations were measured over a length
ing of the steel beams had already occurred, and at equal to 75% of the composite beam depth from the col-
approximately 4.0% drift, buckling of the bottom steel umn face. As shown in this figure, beam rotations in
beam flanges was observed, limiting the amount of shear excess of 1.0% were measured at story drifts of approxi-
force transferred into the connection. For the cycles per- mately 2.5%. For larger drift levels, beam rotations of
formed at 4.0 and 5.0% drift, no new cracks developed up to 3.0% were measured in the plastic hinge region,
in the joint regions of Specimens 1 and 2, and thus they leading to substantial energy dissipation capacity in the
remained with only moderate damage at the end of the test specimens. It should be mentioned that even though
tests. Slight crushing of the slab concrete, due to large shear studs were located in the beam plastic hinge
beam rotations, was observed during these displace- region, no adverse effect was noticed during the tests.
ment levels. The energy dissipation capacity of the test specimens
was evaluated by comparing the energy dissipated, ED,
8.2. Load vs. displacement response during the first and second cycle at drift levels ranging
from 2 to 5% to the energy dissipated by an equivalent
The lateral load vs. story drift response measured in elasto-plastic system, EEP. The energy dissipated for the
the test specimens is shown in Fig. 9a and b. In these equivalent elasto-plastic system was calculated based on
plots, the story drift has been adjusted to account for the initial stiffness of the specimen and the peak positive
displacements that occurred in the test setup during the and negative loads during the first cycle performed at
testing of the specimens. As shown in Fig. 9a and b, a each drift level. Fig. 11 shows this energy ratio for
nearly linear elastic response was observed in the two Specimen 2. As can be observed, Specimen 2 possessed
excellent energy dissipation capacity as indicated by the
energy ratios ranging from approximately 40 to 55% and
from 27 to 45% for the first and second cycle at each
drift level, respectively. It is worth mentioning that the
behavior shown in Fig. 11 for Specimen 2 is representa-
tive of the energy ratios exhibited by Specimen 1.

8.3. Load vs. joint distortions

As mentioned earlier, distortions in RCS connections


consist of shear deformations and RBR of the steel beam
inside the joint region. In this section, the ability of the
design procedure to limit joint distortions is evaluated
through the measurements of shear distortions and beam
Fig. 8. Cracking pattern in Specimen 1 at 3.0% story drift. RBR obtained during the tests. Fig. 12a and b show the
688 G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690

Fig. 9. Load vs. displacement response.

lateral load vs. joint shear deformation and beam RBR


response, respectively, obtained for Specimen 2 with
steel band plates. As Fig. 12a indicates, joint shear defor-
mations in Specimen 2 were effectively limited to a
maximum of 0.5% (0.005 rad), the target of the design
procedure. Beam RBR of approximately 0.6% were
measured during the two cycles performed at 5.0% story
drift. From these two joint deformations, a total joint
distortion of approximately 1.1% was measured at 5.0%
drift, which roughly corresponds to 20% of the total drift
of the subassembly. The response of Specimen 1, with
standard details, was similar to that of Specimen 2. For
this specimen, joint shear distortions were kept below
0.5% up to story drifts of 3.0%. At 5.0% story drift, a
maximum total joint distortion of 1.5% was measured,
which consisted of 0.7% of joint shear distortion and
0.8% of beam RBR.
Fig. 10. Typical beam moment vs. rotation response.

8.4. Story drift components

In RCS beamcolumn subassemblies, four major


components contribute to total story drift: (1) beam
rotations, (2) column rotations, (3) joint shear defor-
mations, and (4) beam RBR (beam and column shear
deformations generally represent a negligible percentage
of the subassembly story drift). In Specimens 1 and 2,
the contribution from each of these mechanisms to speci-
men story drift was determined. Fig. 13 shows the story
drift contribution from each mechanism for Specimen 1,
which is also representative of the contributions
observed in Specimen 2. Displacement contributions
from the test setup were previously subtracted from the
total specimen drift. For all levels of story drift, the cal-
culated lateral displacement was within 5% of the
actual drift. As shown in Fig. 13, for the 1.02.0% story
drift levels, the beams contributed to approximately 50%
Fig. 11. Energy ratio for Specimen 2. of the specimen drift. Joint shear distortions and beam
RBR represented about 20% each, and column rotations
contributed approximately 10% of the total story drift.
However, for larger displacement cycles during which
G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690 689

Fig. 12. Joint distortion in Specimen 2.

buildings located in zones of high seismicity is proposed


and experimentally evaluated. Two simple joint details
for RCS construction are also proposed for connection
regions either with or without transverse beams. For
beams framing into the joint region from only one direc-
tion, the use of standard RCS joints with two-part U-
shaped stirrups, face bearing plates and steel columns
embedded in the RC column is proposed. For the con-
nections with transverse beams, U-shaped stirrups are
eliminated and steel band plates that wrap around the
column regions just above and below the steel beam are
proposed to provide confinement to the connection
region.
A joint deformation-based capacity approach was
used to design two strong columnweak beam RCS
beamcolumnslab subassemblies, one with standard
Fig. 13. Components of total story drift (Specimen 1). joint details and one with steel band plates. The lateral
load testing of these two specimens demonstrated the
significant inelastic activity occurred in the beams, a effectiveness of this design procedure for controlling
reduction in the percentage contribution from joint joint deformations and damage, and thus leading to the
deformations and column rotations was observed due to occurrence of large inelastic deformations in the com-
the increase in the contribution from beam rotations to posite steelconcrete beams. In addition, the two RCS
total story drift. For a story drift of 5.0%, beam rotations specimens tested in this research program exhibited
contributed up to 65% of the total drift, while joint shear excellent loaddisplacement response and high energy
distortions and beam RBR represented approximately dissipation capacity when subjected to large displace-
15% each and column rotations represented less than 5% ment reversals. Beam rotations exceeding 3% were mea-
of the total specimen drift. It should be mentioned that sured in the test specimens, which represented the main
these contributions contrast with those obtained for two contribution to the total story drift of the subassemblies.
RCS beamcolumn subassemblies tested by Parra-Mon-
tesinos and Wight [8] that exhibited a joint shear failure.
For those particular specimens the contribution from
joint distortions represented approximately 65% of the
total drift, while beam and column rotations, represented Acknowledgements
only 30 and 5%, respectively.

The research study described herein was sponsored by


9. Summary and conclusions the National Science Foundation under the grants CMS-
9520295 and CMS-9710204. The opinions expressed in
In this article, a deformation-based capacity design this paper are those of the writers and do not necessarily
procedure for beamcolumn connections in RCS frame reflect the views of the sponsors.
690 G.J. Parra-Montesinos et al. / Engineering Structures 25 (2003) 681690

References [6] Kanno R. Strength, deformation, and seismic resistance of joints


between steel beams and reinforced concrete columns. PhD the-
sis, Cornell University, Ithaca, NY; 1993.
[1] Griffis L. Some design considerations for composite-frame struc-
[7] Parra-Montesinos G, Wight JK. Seismic response of exterior RC
tures. AISC Eng J 1986;23(2):5964.
column-to-steel beam connections. ASCE J Struct Eng
[2] Sheikh TM, Yura JA, Jirsa JO. Moment connections between
2000;126(10):111321.
steel beams and concrete columns. PMFSEL Report No. 87-4.
[8] Parra-Montesinos G, Wight JK. Seismic behavior, strength and
Austin, TX: University of Texas at Austin, 1987.
retrofit of exterior RC column-to-steel beam connections. Report
[3] Deierlein GG, Yura JA, Jirsa JO. Design of moment connections UMCEE 00-09. Ann Arbor, MI: Department of Civil and
for composite framed structures. PMFSEL Report No. 88-1. Aus- Environmental Engineering, University of Michigan, 2000.
tin, TX: University of Texas at Austin, 1988. [9] Kanno R, Deierlein GG. Seismic behavior of composite (RCS)
[4] ASCE Task Committee on Design Criteria for Composite Struc- beam-column joint subassemblies. In: Composite Construction in
tures in Steel and Concrete. Guidelines for design of joints Steel and Concrete III. Proceedings of an Engineering Foundation
between steel beams and reinforced concrete columns. ASCE J Conference, Irsee, Germany. 1996. p. 23649.
Struct Eng 1994;120(8):233057. [10] AISC. Seismic provisions for structural steel buildings. Chicago,
[5] Parra-Montesinos G, Wight JK. Modeling shear behavior of IL: American Institute of Steel Construction, 1997.
hybrid RCS beamcolumn connections. ASCE J Struct Eng
2001;127(1):311.

You might also like