Sei sulla pagina 1di 13

Seismic Performance of Brace-Beam-Column Connections in Concentrically

Braced Frames
K.K. Wijesundara1, G.A. Rassathi2, R. Nascimbene3 and D. Bolognini4
1

European School for Advanced Studies in Reduction of Seismic Risk, IUSS (ROSE
School) Via Ferrata 1, 27100 Pavia, Italy, PH; +390382516927; email: kwijesundara@roseschool.it
2
Department of Civil and Environmental Engineering, University of Cincinnati, 765
Baldwin Hall, Cincinnati, OH 45221-0071, PH; +1 (513) 556-3696, email:
gian.rassati@uc.edu
3
European Centre for Training and Research in Earthquake Engineering (EUCENTRE)
Via Ferrata 1, 27100 Pavia, Italy, PH;
+390382516927, email: roberto.nascimbene@eucentre.it
4
European Centre for Training and Research in Earthquake Engineering (EUCENTRE)
Via Ferrata 1, 27100 Pavia, Italy, PH;
+390382516920, email: davide.bolognini@eucentre.it

ABSTRACT
In typical brace-beam-column connections of concentrically braced frames (CBFs) with
tubular braces, a gusset plate is used to connect the brace to the beam and column. The
slotted tubular brace is welded to the gusset plate and subsequently the gusset plate is
also welded to the beam and column. The beam-to-column connection at the gusset plate
is either welded or bolted at the face of the column flange. Even though a bolted connection is provided at the face of the column flange, stiff gusset plates connected to beam
and column could still provide a fully restrained beam-to-column connection as in the
case of welded connection. Such a fully restrained beam-to-column connections in CBFs
are such that plastic hinges will form in the column upon continued lateral deformation
under severe shaking: hence, higher drift concentrations can be expected in the bottom
storey when the partially restrained or pin-ended column-to-base connections are provided.
For these reasons, this study investigates the local and global seismic performances of
fully restrained brace-to-beam/column connections through numerical analyses. The
global performance is examined using a 4, 8 and 12 storey concentrically braced prototype frames modelled in OpenSees, while the local performances are examined through
the detailed finite element model of a single storey single bay frame located at the
ground floor of the four storey brace frame using the finite element program MIDAS.
Furthermore, this study introduces a partially restrained bolted connection at the corner
of the gusset plate rather than providing it at the face of column flange in order to facilitate beam rotation at the bolted connection upon continued lateral deformation; this is
expected to prevent the formation of plastic hinges in the columns and to distribute the

inelastic drift demand over the full height of the structure. Finally, global performances
of partially and fully restrained connections are compared in terms of storey displacement and inter-storey drift, while local performances of two models are compared in
terms of strain concentrations in gusset plate, beam and the column.
Keywords: Gusset plate connection; Finite element model; Yield mechanism; Equivalent plastic strain; Local buckling
INTRODUCTION
In a gusset plate connection in CBF with the RHS shape brace, the slotted RHS brace is
welded to the gusset plate using 4 fillet weld lines and subsequently, the gusset plate is
also welded to the beam and the column using fillet weld lines. The beam is either
welded or bolted to the face of the column flange. Even though a bolted connection is
provided at the face of the column flange, stiff gusset plate connected to the beam and
the column could still provide a fully restrained beam-to-column connection as in the
case of the welded connection.
Thus, this study discusses effects of different combinations of beam-to-column and
base-to-column connections in the braced bay to the global seismic performance of 4, 8
and 12 storey CBFs designed for 475 years return period of earthquake with the PGA of
0.3g.
Furthermore, to investigate the local performance of a gusset plate connection, three FE
models are developed in MIDAS FEA program. The first FE model is developed without any clearance provided in the gusset plate to facilitate the brace to buckle in out-ofplane direction and it is referred as Model 1.
AISC [2005] provisions require for special concentric braced frames (SCBFs) to provide
2t clearance, where t is the gusset plate thickness, in the gusset plate in order to accommodate the brace to buckle in out-of-plane direction of the frame forming the plastic
hinge in the gusset plate. Such a clearance rule results in a relatively large gusset plates
and the gusset plate connection could be uneconomical. The studies by Roeder et al.
[2004] and Lehman et al. [2008] have indicated that relatively large and thick gusset
plates concentrate the cyclic strain demand in the middle of the brace, thereby reducing
the deformation capacity of the brace. The study by Lehman et al. [2008] proposed to
use elliptical clearance rule in designing the gusset plate connection to accommodate the
brace end rotation, and it results more compact rectangular gusset plate.
In this study, two alternative models of tapered gusset plate connection with 2t clearance
and rectangular gusset plate connection with the clearance to form tri-linear yield
mechanism are designed and analysed. The second and third FE models developed in
this study are provided the linear and tri-linear clearance in the gusset plate and they are
referred as Model 2 and Model 3 respectively. The gusset plate connections in each
model were designed according to the procedure developed based on the performance
based seismic design (PBSD) by Roeder [2004]. This design procedure is developed to
assure the desirable yield mechanisms, to prevent the undesirable failure modes, thereby
achieving the desired seismic performance. Each of the FE model, including the beam,

two gusset plate connections at each end of the brace and two columns as shown in Figure 1, is half bay width of the braced bay at the ground floor of the four storey CBF
building.

5Bays@7m

5Bays@7m

Figure 1. Selection of half bay single storey model.


The results of the inelastic analyses show that the gusset plate connections can significantly influence the cyclic performance of the CBFs and the tri-linear clearance rule
could optimize the design of the gusset plate connection.
EFFECTS OF DIFFERENT COMBINATIONS OF BEAM-TO-COLUMN AND
COLUMN-TO-BASE CONNECTIONS TO THE GLOBAL PERFORMANCE OF
CBFS.
4, 8 and 12 storey braced frames with inverted V configuration with the middle column
(IVMC) as shown in Figure 1 were selected for further investigation on the effects of
beam-to-column and column-to-base connections in the braced bay to the seismic performance of the CBFs. Basically 3 cases were considered in this section as:
(a) Partially restrained beam-to-column and column-to-base connections.
(b) Fully restrained beam-to-column and partially restrained column-to-base connections.
(c) Fully restrained beam-to-column and column-to-base connections.
In these combinations, fully restrained beam-to-column connections can be considered
as welded connections while partially restrained beam-to-column connections can be
considered as bolted connections provided at ends of the gusset plates. Fully restrained
column-to-base connections are embedded base plate connections and partially restrained base-to-column connections are typical end base plate connections.
Design and Modelling of the CBFs. Type 1 elastic spectrum as specified in EC8 [2005]
which is corresponding to ground motion of magnitude 5.5 or above and class-A (hard
soil), was selected for designing of the 4, 8 and 12 storey frames. The reference peak
ground acceleration, having a probability of exceedance equal to 10% in 50 years, is
0.3g. The design gravity loads were as follows: Gravity dead (Gk) and live (Qk) loads of
6.4 and 3.0 kPa at each floor level. For simplicity, accidental torsion was omitted in the
seismic design and total lateral load in each direction was resisted by two IVMC-braced
frames. Lateral force method as specified in EC8 [2005] for the structure in regular in

plan and elevation, was used as the seismic design method with reference to q factor of 4.
Braces were sized assuming that storey shears at all floor levels were entirely resisted by
axial forces in braces. Brace forces due to gravity loads were neglected in this calculation. All braces were made of hollow square section (HSS) shapes and their nominal
yield stress Fy =355 MPa. HSS shapes belong to the required cross sectional class proposed by EC8 [2005] and EC3 [2005]. A pair of braces at each storey level was selected
such that the design storey shear is higher than the shear resistance produced by the pair
of braces assuming that the post buckling strength is 0.25 times the initial buckling load.
All the columns in the braced bay were designed to resist the maximum expected axial
load in combining of non seismic actions and seismic actions following the brace buckling and yielding.
The brace model used to represent the nonlinear behavior of a brace in this study consists of two inelastic beam-column elements available in OpenSees plateform. The
force-displacement relation of the element is transferred to the global reference system
considering of nonlinear geometry of large displacements in accordance with the corotational theory (Crisfield [1991], Filippou and Fenves [2004]) whereas, the small deformation theory is used for computation of local stresses and strains of the inelastic beamcolumn element. The middle node of each brace model was placed by amount of Ls/200
(where Ls is length of the brace) in the direction perpendicular to the plane of the frame
in order to consider the buckling of axial loaded braces. Beams and columns were also
modeled using nonlinear beam-column elements. The gusset end restrained condition of
a brace was modeled using additional two nonlinear beam-column elements at each end
of the brace. The width of the fiber section of the element was the Whitmores width or
the effective width of the gusset plate while the thickness of the section was the thickness of the gusset plate. Fully restrained beam-to-column connections were modeled as
continuous members while the partially restrained connections were modeled using six
springs: five elastic springs to constraint the three translational degrees of freedoms and
two rotational degrees of freedoms about the two vertical axes on the plane of the frame
and one nonlinear elasto-plastic spring to constraint the rotational degree of freedom
about the axis perpendicular to the plane of the frame at each of the connection. The rotational stiffness of the spring was calibrated in the experimental results from Astaneh et
al. [2001]. The material response of the brace steel was represented by Menegotto-Pinto
material model with kinematics and isotropic hardening. The strain hardening ratio was
assumed to be 0.8% and values of Youngs modulus and yield stress of the steel were
taken as 210 GPa and 355 MPa respectively.
Results and Discussion. Figure 2 compares the average drift profiles resulted in inelastic time history analysis (ITHA) of 7 real accelerograms which are scaled to matched
with designed spectrum for three cases. This comparison illustrates that the combination
of partially restrained base-to-column and fully restrained beam-to-column connections
results more drift concentration at the 1st storey, compared to other two cases. Even
though fully restrained beam-to-column connections are provided, the partially restrained connections at the base fail to make the system similar to the dual system. Fur-

thermore, it is noticed that as increasing of number of stories, the drift concentration at


the 1st storey is increased.
In the case of fully restrained column-to-base and beam-to-column connections, the significant drift reduction in the 1st storey is observed in all frames. However, better drift
distributions are resulted for medium-rise frame building since the additional stiffness
makes the system similar to the dual system. More importantly, it is noticed that the drift
amplifications at the upper storeys due to the distribution of slenderness over the building height have been mitigated markedly even without providing over strength braces.
4

Case (a)
Case (b)

Case (a)
Case (b)
Case (c)

Case (c)

Storey No.

5
4
3
2

1
0
0.00

0
0.0

0.5

1.0

1.5

2.0

0.50

1.00

1.50

2.00

2.50

Inter-storey drift (%)

Inter-storey drift (%)


12

Case (a)
Case (b)
Case (c)

11
10
9
8

Storey No.

Storey No.

7
6
5
4
3
2
1
0
0.00

0.50

1.00

1.50

2.00

2.50

Inter-storey drift (%)

Figure 2. Comparison of drift profiles obtained in 3 case studies for 4, 8 and 12


store frame

DESIGN OF GUSSET PLATE CONNECTION.


In this study, the gusset plate connections were designed according to the performance
based seismic design (PBSD) as proposed by Roeder et al. [2004]. This design procedure was developed to assure the desirable yield mechanisms, to prevent the undesirable
failure modes and, thereby achieving the desired seismic performance. The formation of
yield mechanism provides the inelastic deformation of the structure without significant
strength reduction. The formation of a failure mode could lead to the fracture, the loss of
strength or deformation capacity. A single failure mode of the connection will produce a
considerable reduction in strength or deformation capacity of the system whereas the
multiple failure modes generally result to complete failure of the connection. Figure 3
illustrates possible yield mechanisms and failure modes which could be developed in
CBFs under different performance levels.
As shown in Figure 3, primary or favourable yield mechanisms in CBFs are expected to
be the inelastic shortening due to the post-buckling behaviour of the brace and the tensile yielding of the brace. Secondary or less favourable yield mechanisms are expected
to be yielding of the gusset plate under the combine actions of axial, moment and shear
forces, and the yielding of the beam and the column adjacent to the re-entrant corners of
the gusset plate. The possible failure modes of the brace in CBF are fracture of the brace
at the mid region due to the excessive local deformation and the net section fracture at
the slotted brace end. Severe weld tearing, the block shear, the net section fracture and
the buckling just beyond the brace end are failure modes of the gusset plate.

(1)

Yielding of Beams
and Columns at Guest
Plate Edge (4)

Inelastic Shortening Due to


Post-Buckling Behaviour of
Brace

Tensile Yielding of
Brace (2)

Yielding of Gusset Plate (3)

Severe Weld
Tearing

(9)
Block
Shear (7)

Net Section
Fracture of
Brace (11)

Buckling of
free edge (6)

Fracture of Tearing of
Brace(10)
Net Section Fracture of
Gusset Plate

Buckling of Gusset Plate (5)

(8)

Figure 3. Yield mechanisms and the failure modes from Roeder et al [2004].
Primary yield mechanisms should have smaller resistances than secondary mechanisms
and all failure modes. This requirement makes the connection strength and stiffness able
to ensure the immediate occupancy performance level. Once this connection strength
and stiffness are provided, the connection deformation capacity becomes the dominant
concern with life safety and collapse prevention performance levels. According to the
performance based design strategy, the brace buckling and the incipient of the brace
yielding are expected in the immediate occupancy whereas the brace yielding and the
incipient of yielding of the gusset plate are expected in the life safety performance level.
Due to significant out-of-plane buckling of the brace in the life safety performance level,
the gusset plate could subject to significant rotation and subsequently, yielding of the
gusset plate initiates under the combined actions of the moment and the axial load. In the
collapse prevention performance level, the excessive out-of-plane deformation is expected. Due to the excessive out-of-plane deformation, incipient of the brace fracture
and significant yielding in the gusset plate are expected but improperly detailed connection could lead to either the fracture of the net section of the gusset plate or weld tearing
of the gusset plate connection.
DESCRIPTION OF PARAMETRIC STUDY
Four different types of brace-beam-column connections as shown in Figure 4 were designed using aforementioned procedure. Model 1 and 3 were provided the clearance in
the gusset plate to facilitate for the brace end rotation when the brace buckles in out-ofplane direction in forming tri-linear yield lines whereas Model 2 and 4 were provided
the linear clearance of two times the gusset plate thickness (2t) as specified in AISC
[2005] between the brace end and the line which connects the re-entrant corners of the
gusset plate as shown in Figure 5. The difference between Model 1 and 3 is the angle
between the vertical line and the line which connects the brace end and the re-entrant
corner of the gusset plate. This angle is 15 degree for Model 1 and 30 degree for Model
3. Furthermore, authors have indicated the clearance provided in Model 3 in Figure 5(b).

Model 1

Model 2

Model 3

Model 4

Figure 4. Expected yield line formed in the gusset plate of different models.
When the 2t linear clearance rule was used to design the gusset plate in rectangular
shape, relatively large and thick gusset plate was resulted compared to other models.

Relatively long length beyond the brace end in the gusset plate required thick gusset
plate to prevent the buckling of the gusset plate under the direct compression load. In
experimental studies by Lehman et al. [2008] and numerical studies by Yoo et al. [2008]
have clearly indicated that relatively large and thick gusset plate concentrates the cyclic
strain demand in the middle of the brace, thereby reducing the deformation capacity of
the brace. Hence, this numerical study is limited to Model 1, 2 and 3 in order to investigate alternative proposals to improve the performance of the system.
In Model 3, to form a tri-linear yield line in the gusset plate as proposed in this study,
the angle between the line which connects the brace end and the re-entrant corner, and
the vertical line should be equal to 30 degree as shown in Figure 5(b). The reason to select 30 degree angle is to use the maximum possible effective width of the gusset plate
in resisting the direct axial tensile load developed in the brace when rectangular gusset
plate is used. If the angle is reduced more that 30 degree as in the case of Model 1, the
higher strain concentration is observed just beyond the brace end when full tensile
strength or the excessive rotation is developed in the brace due to the reduced effective
width.
2t

2t

(a) Model 2
(b) Model 3
Figure 5. Connection details of the models.
Furthermore, tri-linear clearance rule results more compact gusset plate in rectangular
shape with leg length of the gusset plate along the column and the beam similar to
Model 2 of tapered connection. The controlling yield mechanism of the design of Model
3 is the yielding of the effective width or Whitmores section of the gusset plate under
the direct tension.
FINITE ELEMENT MODEL
MIDAS FE models were developed to simulate behaviours of the gusset plate connections in the CBF belonged to three different categories. Each model represents the subassemblage of the half bay width of the braced bay at the ground floor of four storey
CBF. The bay width and the storey height of the braced frame are 7m and 3.5m respec-

tively. The sub-assemblage included a brace, top beam, two columns and two gusset
plates. The brace was 203x203x12.5 HSS shape while the beam and columns were wideflange section of W360x237 and W360x101 respectively.
All elements in the sub-assembly were also developed in four node shell elements available in MIDAS FEA programme. As it is discussed before, shell elements are capable of
taking into account for the in-plane deformation (membrane) and the out-of-plane deformation (bending and shear). The Von Mises constitutive model was adopted with bilinear isotropic hardening to represent the material non linearity. Parameters for the
constitutive model were adopted in the coupon test results from the experimental program by Walpole [1995]. The nominal yield strength is 350 MPa for the grade 350 steel
but yield stress specified in the numerical model is the stress that obtained from 0.2%
offset method using the coupon testing of the bracing member and subsequently, the
hardening function is assumed to be bilinear with 1% hardening till the strain is 0.035
and beyond that the platue is specified with the stress is equal to 490 MPa. Furthermore,
it is highlighted that the isotropic hardening is considered with Von-Mises plasticity
model.
Since large in-plane and out-of-plane deformations were also expected in these analyses,
geometric nonlinearity was included together with material nonlinearity. The initial imperfection ratio of Ls/700 at the middle of the brace in the direction of out-of-plane was
prescribed in each model. Figure 6 shows typical FE model of the subassembly and refined meshes at the top and bottom gusset plate connections. In all models, approximate
mesh size of 25x25mm was used in the mid region of the brace and gusset plates where
significant inelastic deformation was expected and the rest was modelled with the coarse
mesh. The weld lines were modelled with shared nodes along lines of welds.

(b)

(a)

(c)

Figure 6 (a) Half bay single storey model (b) enlarged view of top gusset plate connection (c) enlarged view of bottom gusset plate.

Welded beam-to-column connections were also modelled with sheared nodes along weld
lines. The out-of-plane deformation of the top flange of the beam was restrained at 300
mm interval in order to provide the restrained to the top flange of the beam. All translational degrees of freedoms at base nodes of each model were restrained. The displacement history was applied to the column flange as shown in Figure 6(a).
RESULTS AND DISCUSSION
Free space in gusset plate. The strain distributions at gusset plates and mid regions of
braces of Model 1, 2 and 3 at ductility levels (/y) of 11 in compression are presented in
Figure 7. It is important to note that comparisons are made for surface strains those
demonstrate the greatest magnitude of the local strain states. In Model 1, higher strain
concentrations are developed just beyond the brace end and re-entrant corners of the
gusset plate at the ductility level, compared to Model 2 and 3.

Model 1

Model 2

Model 3
Figure 7. Equivalent Von Mises strain distributions at the deformation of /y=11
in compression.
Furthermore, significant yielding in the beam and columns adjacent to re-entrant corners
of the gusset plate are observed without local buckling in compression flanges in the
beam and the column.
In Model 2, significant yielding is developed beyond the brace end in the gusset plate
under the combined actions of axial, moment and shear forces at considered compres-

10

sion ductility level but, there is no sign of high strain concentrations as observed in
Model 1. It is also observed that linear yield line is formed in the gusset plate when the
excessive out-of-plane deformation is occurred in the brace. However, high strain concentrations at corners of the gusset plate are observed. No local buckling is formed in the
gusset plate beyond brace ends, free edges of the gusset plate, column and beam flanges.
Similar to Model 2, significant yielding is observed in Model 3 just beyond the brace
end in the gusset plate under the combined actions of axial, moment and shear forces at
the considered ductility level. Furthermore, tri-linear yield line is formed in the gusset
plate when excessively large out-of-plane buckling occurred in the brace without concentrating the strain just beyond the brace end in the gusset plate. Again similar to
Model 2, high strain concentrations at corners of the gusset plate are observed and no
local buckling is formed in the gusset plate beyond brace ends, free edges of the gusset
plate, column and beam flanges.
Depending on the free space provided in the gusset plate, the deformation capacity of the
gusset plate connection could be limited. In order to better understand yield mechanisms
and failure modes of Model 1, 2 and 3, the distributions of peak equivalent strains at the
mid regions of braces, along weld lines and regions just beyond brace ends in gusset
plates at the end of each compression cycle are plotted against corresponding drift ratios
of half bay single storey frame models.
0.08

0.3

Initiation of crack

0.2
0.1

Model 1
Model 2
Model 3

0
0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Equivalent Strain

Equivalent Strain

0.4

0.06 Initiation of crack


0.04
Model 1
Model 2
Model 3

0.02
0
0.0 0.5

1.0 1.5

3.0 3.5

Drift (%)

Drift (%)

(a)

2.0 2.5

(b)

Figure 8. Strain distribution (a) in mid region of the brace (b) along the weld lines
which connect the gusset plate to beam/column.
Figure 8(a) illustrates peak equivalent strain distributions in Model 1, 2 and 3 at locally
bucked mid regions of braces against storey drift levels. All three models show very
closer strain distributions and hence, it is evident that there is no significant effect of the
free space provided in the gusset plate to the strain distribution in the mid region of the
brace. Furthermore, Figure 8(a) shows that the strain level at the mid region of the brace
is increased dramatically after the formation of local buckling. Figure 8(b) illustrates that
considerably increased peak equivalent strain distributions along the weld lines are observed in Model 1 and Model 3 compared to Mosel 2. These peak strain values are
mostly resulted due to the concentration of strain at the corners of the gusset plate.

11

From the comparison of Figure 8(a) and 8(b), the immediately occupancy performance
level could be achieved in all models, but the achievement of life safety and collapse
prevention performance level in Model 1 is raised serious doubt due to the fact that high
concentration of strain is developed just beyond the brace end and corners of the gusset
plate in the medium level of drift. These high concentrations of strain could lead to either net section fracture of the gusset plate or incipient weld tearing due to excessive
brace rotation resulting in the out-of-plane buckling, well before the initiation of brace
fracture. Even in Model 2 and 3, collapse prevention performance level could not be
achieved because of possible incipient weld tearing before the incipient brace fracture.
CONCLUSION

From the results of this study, following recommendations for the design of the gusset
plate connection and conclusions could be made as:
Providing a free space in the gusset plate to form tri-linear yield line results more compact gusset plate. Furthermore, it shows that there is no significant difference in the distribution of strains and stresses in the middle of the brace, gusset plate and along the
weld lines compared to the current 2t linear clearance model with tapered gusset plate.
According to the tri-linear model, it is recommended to insert the brace into the gusset
plate till the minimum angle between the line which connects the brace end and the reentrant corner of the gusset plate and the weld line which connect the gusset plate to the
beam or the column, should be 30 degree as shown in Figure 5.2 in order to prevent the
highly concentrated stress and strain demands to the weld lines and the gusset plate. Furthermore, it is recommended that when the tapered gusset plate is used, the minimum
angle of 30 degree between the weld line which connects the brace to the gusset plate
and the free edge of the gusset plate should be provided.
REFERENCES

AISC [2005] Seismic provisions for structural steel buildings. American Institute of
Steel Construction, Inc., Chicago, Illinois, USA.
A Astaneh-Asl. [1998] Seismic Behavior and Design of Gusset Plates for Braced
Frames, Steel Yips, Structural Steel Education Council, Moraga,CA.
CEN [2005] ENV 1993-1-1, Eurocode 3: Design of steel structures Part 1.1: General
rules and rules for buildings, European Committee for Standardisation, Brussels.
CEN [2005] EN 1998-1, Eurocode 8: Design provisions for earthquake resistance of
structures, Part 1: General rules, seismic actions and rules for buildings, European
Committee for Standardisation, Brussels.
Elghazouli A.Y. [2008] Seismic design of steel framed structures to Eurocode 8, Proceedings of the 14th World Conference on Earthquake Engineering, Beijing, China, 12-17 October.

12

Lehman, D.E., Roeder, C.W., Herman, D., Johnson, S. and Kotulka, B. [2008] Improved Seismic Performance of Gusset Plate Connections, Journal of Structural
Engineering, ASCE, Vol. 134,No. 6 pp. 890-901.
MIDAS, Nonlinear and Detail FE Analysis System for Civil Structures, FEA Analysis
and Algorithm Manual.
OpenSees: Open System for Earthquake Engineering Simulation, Pacific Earthquake
Engineering Research Center, University of California, Berkeley, CA.
Roeder, C. W., Lehman, D. E., and Yoo, J. H. [2004]. Performance based seismic design of braced-frame connections. 7th Pacific Structural Steel Conference.
Uriz P. Filippou F.C. and Mahin S.A. [2008] Model for cyclic inelastic buckling for
steel member, Journal of Structural Engineering ASCE, Vol. 134, No. 4, pp. 616628.
Yoo, J.H., Lehman, D.E., Roeder, C.W. [2008] Influence of connection design parameters on the seismic performance of braced frames, Journal of Constructional Steel
Research, Vol. 64, pp. 608-622.
Yoo, J.H., Roeder, C.W., Lehman, D.E., [2008] Analytical Performance Simulation of
Special Concentrically Braced Frames, Journal of Structural Engineering, ASCE,
Vol. 134, No. 6, pp. 881-889.
Yoo, J.H., Roeder, C.W., and Lehman,D.E. [2008] Simulated Behavior of Multi-Story
X-Braced Frames, Journal of Engineering Structures, (In Press).

13

Potrebbero piacerti anche