15 PDF
15 PDF
15 PDF
1 (2016)
PAGES 15-42
ABSTRACT
Keywords: Special moment frame; ductility; strength; welded flange plate (WFP)
connection.
1. INTRODUCTION
The steel moment-resisting frames suffered a verity of brittle fractures in their beam-to-
column connections during the 1994 Northridge earthquake. This initiated widespread
research to find out the causes [1-3] and to propose changes to design procedures. Several
methods have been presented so far in order to increase the ductility of pre-Northridge
connections. These modifications can be implemented either by reinforcing the connection
or weakening the beam section. The purpose of each method is to create a stronger
E-mail address of the corresponding author: amirahmad1356@yahoo.com (A.A. Hedayat)
16 A.A. Hedayat, H. Saffari and E. Jazebi
connection in comparison with the beam. Using these methods lead to move the plastic
hinges in a region away from the face of the column and avoid the brittle fractures. In the
first method some additional elements are added to the connection at the column face level
[4-6], while in the second method, weakening the beam section can be done either by cutting
a portion of the beam flange, (reduced beam section connections, [7]) or the beam web
(RBW connections,[8-11]).One of the most common used methods after Northridge
earthquake to enhance the ductility of rigid connections was using plates in two forms of
cover plate and flange plate (Fig. 1). The latter is named by FEMA350 (2000) [12] as
welded flange plate (WFP) connection where the width and the length of the top flange plate
are as same as those of the bottom flange plate.
Prior to the Northridge earthquake, a number of investigators conducted cyclic loading
tests on steel moment connections reinforced with cover plates [13]. These early tests
showed that very high levels of cyclic ductility could be achieved by reinforced connections.
However, these earlier tests were very limited in number and were conducted on relatively
small size of beams. Popov and Jokerst[14] tested two WFP connections which were
composed of W30x99 beams and W14x283 columns of ASTM A572 Grade 50 steel. Two
connections were reinforced by two same size and trapezoidal shaped flange plates. Both
specimens achieved 0.034 radian plastic rotations. Whittaker and Gilani [15] tested three
WFP connections. Specimens were composed of W21 and W24 beams and I-shaped
columns of Grade 50 steel with two same size-rectangular shaped flange plates. They
investigated the effect of contribution of the column panel zone (PZ) to the plastic rotation
capacity of the joint. In these tests, in the case of having weak PZ, the large distortions in the
PZ led to the column flange fracture and caused minimum contribution of the beam into the
joint plastic rotation capacity. The joint plastic rotation capacity varied between 0.028 and
0.045 radian. All specimens (except the specimen of weak PZ) experienced excessive
strength degradation. Noel and Uang [16] tested three double-sided and one single-sided
WFP connections. The beams ranged from W24x94 through W18x86 of Grade 50 steel with
two same size rectangular or trapezoidal shaped flange plates. Maximum beam plastic
rotation ranged between 0.013 and 0.046 radian. A total of 12 large scale specimens were
constructed with cover plated connections and subjected to cyclic loading tests by
Engelhardt [4]. Out of the 12 specimens tested by Engelhardt et al. [4], 10 specimens
showed excellent performance, achieving high levels of plastic rotation. Tests done by
Engelhardt et al. [4] indicated that using very long cover plates could result in a large
amplification of moment from the hinge location to the face of the column. This research
also showed that using very thick cover plates result in a very high restraint at the face of the
column, thereby promoting more brittle behavior.SAC Joint Venture (SAC 2000) [17] tested
five WFP and five cover plate reinforced connections. All specimens were fabricated from
W30x99 Grade 50 beams and Grade 50 rectangular or trapezoidal shaped reinforcing plates.
The key design variables were loading history, PZ strength and stiffness (which several
investigation on the behavior of panel zone have been done [18-20]), and flange plate
thickness and geometry. The results revealed substantially better behavior in these types of
connections as opposed to unreinforced connections. None of the ten connections failed in a
brittle manner and the maximum beam plastic hinge rotations ranged between 2.3% and
3.9% radians. However, specimens experienced excessive strength degradation at large
rotations.
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 17
WFP connection is also the most common connection in Iran while its configuration
and fabricating processes is different from what has been mentioned in FEMA350 (2000)
[12] and this may lead to the achievement of a different behavior. Fig. 2 shows this
connection where the bottom beam flange has been strengthened by using a rectangular
flange plate which is wider than the beam flange. Whilst, the top beam flange has been
strengthened by using a flange plate which is narrower than the beam flange. In this
connection, the bottom flange plate is shop welded to the column flange which is
consequently used as a seat plate during the erection process of the structure. W hilst the
top flange plate is field welded to the column flange. These forms of plates prevent
overhead welding which finally leads to the achievement of a stronger weld. Ghobadi et
al. [21] [22] and Gholami et al. [23] tested few WFP connections where the connection
configuration was similar to what is common in Iran (i.e. similar to Fig. 2). In all tests the
section used for beams was either IPE270 (beam depth=270 mm) or a built up section with
the overall depth equal to 380 mm. The key design parameters in these tests were: (1) type
of weld for connecting the bottom beam flange plate to the column flange, (2) T -stiffeners
for retrofitting purposes, (3) flange plate length, and (4) plate-to-beam flange fillet weld
geometry. Based on the experimental tests done by Gholami et al. [23] and Ghobadi et al.
[10], the WFP connections used in Iran can reach to an acceptable level of ductility. It
should be noted that in these tests, the maximum beam depth was only 380 mm. However,
such beams are generally referred as shallow beams. Shallow beam connections often
show a high plastic rotational capacity even in the case of using pre-Northridge
connections (FEMA355D, 2000) [24]. Because of this, generally all connection
modifications are developed by using deep beams (e.g. W36x150, W30x99 and W24x68
beams). In addition, for all specimens tested by Gholami et al. [23] and Ghobadi et al.
[21]a single shear tab was used to transfer shear force from the beam to the column.
While, in Iran, in order to transfer the shear forces, normally the rib plates are used. All
these differences finally may lead to the achievement of different behavior from that has
been mentioned by the FEMA350 [12] or Gholami et al. [23] and Ghobadi et al. [21].
Hence, the present study was aimed to do a parametric study on the behavior of WFP
connections by using deep beams considering all mentioned details used in Iran to reach a
better understanding of the behavior of such connections. The key design variables were:
flange plate thickness, methods to transfer shear force from the beam to the column,
transverse welding at the noise of flange plates, end form of the top flange plate, equality
and inequality of the top and the bottom plate length and rib plate form. Most of these
parameters were not investigated by previous researchers. In addition, experimental tests
done by SAC2000 [17] and Whittaker and Gilani [15]showed that in the case of using deep
beams, the WFP connection experiences remarkable strength degradation such that at 4
percent total rotation the connection may not reach to 80% of the beam plastic moment
capacity. However, such significant strength degradation was not reported by Gholami et al.
[23] and Ghobadi et al. [21] where the shallow beams were used. Hence, the last aim of this
study was to investigate the efficiency of doubler plates to resolve such problem for deep
beams.
18 A.A. Hedayat, H. Saffari and E. Jazebi
(a) (b)
In order to verify the accuracy of finite element modeling, at first, specimens RC06 of
reference (SAC2000) [17] which its configuration is based on the FEMA350 (2000) [12](i.e.
rectangular flange plates) was remodeled using ANSYS software [25]. The used materials
for both beam and column were made of A572 Grade 50 steel (SAC 2000)[17]. For this
specimen the length of the column and the beam were 3450 mm and 3400 mm respectively.
The thickness of the continuity plates and the web doubler plate are 25 mm and 10 mm
respectively. The high fracture toughness weld metal used was E70. Fig. 3 shows the finite
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 19
element mesh of this specimen. In this model, all connection’s parts were modeled using
shell elements (SHELL181 in ANSYS program). Interaction between plates (e.g. interaction
between the flanges of beam and the top or the bottom flange plates) was modeled using
contact elements.SHELL181 is a multi-layer four-node shell element which has the ability to
model plasticity, large deflection, and large strain phenomena. In this case each element was
separated into five layers across the thickness. The number of layers was selected based on
the finite element study carried by Gilton and Uang [26]. In order to determine the
appropriate mesh density, a study on mesh sensitivity was carried based on the
recommendations given by ANSYS software. Results were then compared with the
experimental results presented by SAC (2000) [17].
To perform material nonlinearity analyses, plasticity behavior was based on the von-
Mises yielding criteria and the associated flow rule. Isotropic hardening was assumed for the
monotonic analysis, whereas kinematic hardening was assumed for the cyclic analysis as
used by Mao et al. [27] and Ricles et al. [28]. For base metals, both bilinear and multi linear
material responses were investigated based on the material properties given by SAC (2000)
[17]. However, the analytical results showed that the use of multi linear response agree well
with the experimental results. For weld metals a multi-linear material response based on the
material property given by Mao et al. [27]and Ricles et al. [28] was used.
Nonlinear geometric analyses were performed through a small strain, large displacement
formulation. The monotonic analyses were conducted by applying a monotonic vertical
displacement load to the beam tip until more than 4 percent total rotation at column web
center was achieved and the load history recommended by the AISC (2010a) [29] was
utilized for cyclic analyses. When loads are applied only in the vertical direction, then the
out of plane deformations (normal to the web) may not occur. Therefore, in order to ensure
that buckling would occur due to instability of model, the imperfect model subjected to
cyclic or monotonic loading was used. In this study, in order to determine the imperfect
model, first the buckling mode shapes were computed in a separate buckling analysis and
then were implemented to perturb the original perfect geometry of the model.
In order to verify the accuracy of the model, the experimental results of RC06 from SAC
(2000) [17] were compared to the analytical results obtained from the finite element models
in terms of applied load and rotation, plastic equivalent strain distribution and buckling
modes. Figs 4 and 5 show these comparisons. As Fig. 4 shows the applied load-rotation
curve obtained from the finite element analysis is in good agreement with the experimental
one. Fig. 5 shows the plastic equivalent strain distribution at 5.5 percent rotation. In this
figure high concentration of the plastic equivalent strain at the beam at the end of the flange
plates is quite evident and is similar to what was achieved during the experimental test.
Furthermore, at the same location the beam flange experienced a wide local buckling which
is completely clear in the experimental test (Fig. 5).
It should be noted that the fracture prediction is the most questionable part of a finite
element study. Because it is inherently a complicated phenomenon and is dependent on many
parameters such as weld and base metal properties, weld defects, notch effects, weld quality
and weld toughness. In this study it was assumed that the qualified welders and fabricators are
employed and high fracture toughness weld metals are used (as these should be, based on the
AISC [30], Specification for Structural Steel Buildings).Considering the fact that the locations
of high level of strains have significant probability of premature fracture and by comparing the
20 A.A. Hedayat, H. Saffari and E. Jazebi
finite element results obtained from the specimen RC06 with the experimental results
presented by SAC (2000) [17] a failure criterion was assumed as follows:
The connection fracture occurs when the Von-Mises strains at the whole width of the
beam flange (at the end of flange plate length) or at the entire width of the flange plates (at
the column face level) exceed the strain associated with the ultimate strength of the
materials, based on the material properties reported by SAC (2000) [17].A similar failure
criterion also was used by other researchers (e.g. Saffari, et al. [31];Berman, et al. [32]).
Note that the connection failure can also be predicted using rupture index, RI (e.g. Chao et
al. [33]; Prinzet al. [34]). This index is defined as ratio between the plastic equivalent strain
index and the ductile fracture strain. However, this method may be better when the fracture
initiates from the beam web rather than the beam flange. Note that none of the mentioned
methods are intended to predict exact rotation capacities for connections; rather, they
provide a tool for comparing various models.
200
monotonic
150 cyclic
100
50
Actuator force kips
0
50-
100-
150-
200-
6- 5- 4- 3- 2- 1- 0 1 2 3 4 5 6
Story drift angle , perecent radian
(a) Experimental test conducted by SAC (2000) (b) Finite element model
Figure 4. Applied force-rotation curve of specimen RC06 obtained from
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 21
(
a) (
b)
(a) Finite element model (b) Experimental test done by SAC (2000)
Figure 5. Plastic equivalent strain distribution of specimen RC06 obtained from
Proposed design procedures to modify pre and post Northridge connections are usually based
on three beam sections: W36x150, W30x99 and W24x68. Therefore, all parametric studies
were done on specimens SPE07 (beam: W36x150; column: W14x257), SPE05 (beam:
W30x99; column: W14x176) and SPE03 (beam: W24x68; column: W14x120). These
specimen sizes were chosen since they may be good representatives of the conventional
pre/post Northridge specimen sizes, large, medium and small (Lee et al.) [35]and were also
tested in phase 1 of SAC Steel Projects (SAC 1996) [36].The beam and column lengths of
these specimens were similar to those that were used for specimen RC06. However, the
configuration of these specimens was similar to the one that is shown in Fig. 2.
All specimen were designed based on the Iranian National Building code[37]and
AISC[29,30]. To design the flange plates, the maximum applied moment at the column face
(Mcf) was determined using equation 1. In this equation Zb and fye are the beam plastic
section modulus and the expected yield stress of the material respectively. Sc is equal to the
length of the flange plate plus ¼ of the beam overall depth and Vexp is the expected shear
force at the plastic hinge location. Cpr is a factor to account for the peak connection strength,
including strain hardening, local restraint and additional reinforcement. In FEMA350 [12],
the Cpr factor is given by equation (fy+fu)/2fy, where fuis the specified minimum tensile stress
of the material. FEMA350 [12] proposes the use of value 1.2 for any case of modified
connections except where otherwise noted in the individual connection design procedures.
Similar to SAC [17], in the present study the value used for parameter Cpr was 1.1.
the noise of the flange plate (width at the column face: 337 mm; width at the noise of the
flange plate: 245 mm).
Fig. 6 compares the moment-rotation curves of the two specimens. As this figure shows
by reducing the steel grade, initial rotational stiffness, strength and ductility of the
connection increased. Both specimens could achieve the minimum required ductility. At
four percent radian total rotation, the connection strength of both specimens was greater than
the minimum required strength (i.e. M/MP was greater than 0.8). However, the specimen of
the higher steel grade showed lower strength and experienced faster strength degradation
than the one of the lower steel grade. Therefore, investigation of the behavior of specimens
of the higher steel grade (i.e. A572-Gr50) has been regarded more critical. Furthermore,
using high grade steel is more desirable for designers. Sine it is the most economical way to
increase the flexural capacity of beams without change in the beam section. Also in the case
of having deep beams, using of high steel grade helps to reduce the beam overall depth. It
finally leads to increase in the connection ductility and to reduce in the total height of the
building. Hence, hereafter all parametric studies have been done using the high grade steel
(A572-Gr50).
Table 1: Details of WFP specimens SPE05 with different beam material properties
Specimen Material property Flange plate dimension (mm) Shear tab Continuity
SPE05 Beam column Flange plate top bottom (mm) plate (mm)
NO.1 A36 A572.Gr.50 A36 (337&245)x380x25 337x380x25 660x147x13 2x353x188x25
NO.2 A572.Gr.50 A572.Gr.50 A572.Gr.50 (337&245)x380x25 337x380x25 660x147x13 2x353x188x25
1.4
1.2
1.0
0.8
M/Mp
0.2
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Total rotation(%)
Figure 6. Moment-rotation curves of WFP specimen SPE05 for different beam material
properties
3.2 Effect of the common methods to transfer the beam-column shear on the connection
behavior
In this section the effects of the common methods to transfer the shear forces form the beam
to the column flange on the connection behavior were investigated. The shear forces can be
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 23
transferred through: (1) Shear tab; (2) The top and the bottom rib plates; (3) The beam web
which is directly welded to the column flange and (4) The top and the bottom flange plates.
For initial investigation, specimen RC06 (with rectangular flange plates) was re-designed by
considering the four methods mentioned above. In the first case the length, the width and the
thickness of the shear tab plate were660 mm, 145 mm and 13 mm respectively. Fig. 7 shows
the moment-rotation curve of this specimen. As this figure shows this specimen could
achieve 5.5 percent total rotation while its strength at four percent total rotation was greater
than one. It indicates the acceptable performance level of this specimen. In the second case,
the shear force was transferred by totally four rib plates (two rib plates at the top and two of
them at the bottom) with the dimensions of 15x100x200 mm which were connected to the
beam flange and the column flange through the fillet welds. In this case, the connection only
achieved 2 percent radian total rotation which is less than the minimum required total
rotation (see Fig. 7). In the third case (i.e. direct welding of the beam web to the column
flange), the connection achieved same strength and rotational capacity as those that were
obtained for specimen RC06 in the first case. The worst behavior was achieved when shear
force was transferred through the flange plates (forth case). In this case, the maximum total
rotation was around 1.5 percent radian and the connection fracture occurred when it was in
the elastic region (Fig. 7).
1.4
1.2
1.0
0.8
M/Mp
RIBS2 CASE
0.4
BEM WEB3 CASE
0.2
T&B FALANGE PLATE4 CASE
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Total rotation(%)
Figure 7. The moment-rotation curves in effect of common methods for beam- column shear
transfer in specimen RC06
Direct welding of the beam web to the column flange (case 3) is an acceptable method to
increase the connection ductility (FEMA355D) [24]. In this method shear force is properly
transferred to the column flange through the beam web (not through the flange plates) and
the beam web participates in the moment transferring. Hence, in this case the proper
behavior of the connection is completely expected. In the fourth case all shear forces are
transferred to the column flange through the flange plates. In this case the flange plates are
24 A.A. Hedayat, H. Saffari and E. Jazebi
overstressed (due to the combination of the normal and the shear stresses) and consequently
the early fracture of the flange plates happens. When case 1 and 2 are compared, it might be
concluded that the use of the shear tab is a better method to transfer the shear forces. The
question is if this is the case for any connection configuration.
Initial investigations showed that the connection ductility can also be a function of the
connection length (the length of the flange plates, Lp) and the beam overall depth (db).
Hence, by defining a non-dimensional parameter, LP/db, the behavior of WFP connections
was investigated. Tables 2 and 3 summarize the results of these specimens. In these tables θu
and θd are the maximum total rotation achieved under upward loading and downward
loading respectively and θis the lesser of the two values.
With reference to Tables 2 and 3, when LP/db is greater than 0.6, generally using the rib
plates was better than the use of the shear plate, whereas in LP/db range which is less than
0.6, using the rib plates led to the brittle fracture at the face of the column in small rotation.
The reason for this behavior might be explained as follows:
It is well known that in the welded connections at a region near the column face the
classic beam theory is invalid (SAC 2000 [17]; Cheol [38]). In this region, shear stresses are
at a maximum level near the column flanges, especially at the beam flange or at the flange
plate area. Generally, the combination of the shear stresses and the normal stresses at the
beam flanges near the column face causes high tri-axially stress at the face of the column
and consequently it promotes the brittle fracture of the welded connections. Based on this
discussion, the distribution of the Von-Mises strains (which is a combination of the normal
and the shear strains) for WFP connections were compared in two main cases, LP/db less
than 0.6 and greater than 0.6.
For instance, Figs. 8.a and 8.b compare the normalized von-Mises strain (von-Mises
strain divided by the yield strain) distributions at the center line of the top and the bottom
flange plates along the beam length for specimens SPE05 and RC06 respectively. These
curves are drawn for two different methods used to transfer the shear forces (i.e. using shear
tab and rib plates) when the ratio of the LP/db is 0.5 (LP=380mm, db =750 mm). As these
figures show, for the top flange plates, there is no significant difference between the von-
Mises strain distribution curves when the shear forces are transferred either through the
shear tab or the rib plates. However, it was not the case for the bottom flange plates. There
can be seen a common behavior for both bottom flange plates regardless of the method used
to transfer the shear forces. For both methods:
By moving from the end of the flange plate towards the column face the von-Mises
strains gradually increased which is due to the increase in the applied moments.
There was a sudden increase in the von-Mises strains at the vicinity of the column face
which was due to the presence of the shear strains and the associated secondary normal
strains at the flange plates due to the improper transformation of the shear forces through
the beam web. However, for both SPE05 and RC06 specimens, this increase was much
more when the rib plates were used to transfer the shear forces.
Figs 9.a and 9.b compare the normalized von-Mises strain distributions at the center line
of the top and the bottom flange plates along the beam length for specimens RC06 and
SPE05 respectively when the LP/db ratio is 1.0 (LP=750 mm, db =750 mm). Similar to the
previous case, for the top flange plates, there is no significant difference between their von-
Mises strain distribution curves when the shear forces are transferred either through the
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 25
shear tab or the rib plates. As Figs. 9.a and 9.b show, for the bottom flange plates, by
moving from the end of the flange plate towards the column face, the Von-Mises strains
gradually increased. Again a sudden increase can be seen in the von-Mises strains at the
vicinity of the column face which are due to the presence of the shear strains and the
associated secondary normal strains at the flange plates. However, in contrast to the previous
case, for both SPE05 and RC06 specimens, this increase was much less when the rib plates
were used to transfer the shear forces.
These might be the reasons of facing of different connection behaviors when LP/db is
more or less than 0.6. Similar behaviors were also observed for WFP connections SPE07.
Table 2: Effect of LP/db ratio on the connection’s strength and ductility when the shear tab is
used
M/MP at 4% M/MP at
Specimen LP (cm) LP/db θu (%) θd (%) θ (%)
total rotation failure time
38 0.51 5.5 5.5 5.5 1.01 0.93
48 0.64 5.3 5.3 5.3 1.05 0.97
RC06
60 0.8 5.2 5.2 5.2 1.12 1.04
75 1 2.7 2.7 2.7 1.19 1.04
32 0.43 5.5 5.5 5.5 1.02 0.92
38 0.51 5.5 5.5 5.5 1.04 0.95
48 0.64 6 6 6 1.09 0.97
SPE05
60 0.8 6 3.8 3.8 - 1.16
70 0.93 6 3.7 3.7 - 1.41
75 1 3.5 3.7 3.5 - 1.41
40 0.44 4.7 4.7 4.7 1.11 0.91
50 0.55 5.2 5.1 5.1 1.27 1.1
SPE07 70 0.76 4.7 5.7 4.7 1.29 1.16
80 0.87 2.9 1.9 1.9 - 1.37
90 1 2.7 2.7 2.7 - 1.34
Table 3: Effect of LP/db ratio on the connection’s strength and ductility when the rib plates are
used
M/MP at 4% M/MP at
Specimen LP (cm) LP/db θu (%) θd (%) θ (%)
total rotation failure time
38 0.51 1.9 1.9 1.9 - 1.17
48 0.64 5.2 5.2 5.2 1.03 0.96
RC06
60 0.8 4.6 4.6 4.6 1.08 1.03
75 1 4.9 4.9 4.9 1.1 1.38
38 0.51 1.9 1.9 1.9 - -
SPE05 48 0.64 5.8 5.8 5.8 1.07 0.95
60 0.8 6 6 6 1.14 1.04
26 A.A. Hedayat, H. Saffari and E. Jazebi
4
Normalized von mises strain
0
0 50 100 150 200 250 300 350
Distance from the end of beam flange groove welds at column face
level(mm)
5
4
Normalized von mises strain
0
0 50 100 150 200 250 300 350
Distance from the end of beam flange groove welds at column face level(mm)
Figure 8. Normalized von-Mises strain distribution along the center line of the top and bottom
flange plates when LP/db=0.5 for the WFP specimens: (a) SPE05; (b) RC06
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 27
6
(a)
5
Normalized von mises strain
0
0 100 200 300 400 500 600
Distance from the end of beam flange groove welds at column face
level(mm)
4
(b) WFP specimens
RC06
3 Top flange plate with ribs
Normalized von mises strain
0
0 100 200 300 400 500 600 700
Distance from the end of beam flange groove welds at column face level(mm)
Figure 9. Normalized von-Mises strain distribution along the center line of the top and bottom
flange plates when LP/db=1 for the WFP specimens: (a) SPE05; (b) RC06
3.3 Effect of the transverse welding at the noise of the flange plates on the connection
behavior
In order to evaluate the effect of the transverse welding on the connection behavior,
specimen RC06 was designed and modeled with and without transverse welds. Note that in
this study fillet welds were not explicitly modeled but their effects were considered using
couple command in the ANSYS program. In this case the degree of freedom of the nodes at
28 A.A. Hedayat, H. Saffari and E. Jazebi
the intersection of the noise of the flange plates and the beam flange plate were coupled
together.
Fig. 10 compares the normalized Von-Mises strain distribution across the beam flange
and the flange plate at the various levels in the presence and the absence of the transverse
welds for the WFP specimen RC06 at 6% total rotation. As Fig. 10.a shows the presence of
the transverse welds was ineffective to reduce the normalized strains across the flange plate
at the column face level. At this location the maximum von-Mises strain was very low and
was around 1.1 times of the yield strain of the flange plate material. However, as Fig. 10.b
shows the presence of the transverse weld caused a significant reduction in the normalized
strains across the beam flange width at the noise of the flange plate. It was due to the
composite function of the beam flange and the flange plate to resist the von-Misesstrains.By
moving from the noise of the flange plate towards the plastic hinge length, this positive
effective reduced such that at 40 mm away from the noise of the flange plate the presence of
the transverse welds were somehow ineffective (Fig. 10.c). By moving more inside the
plastic hinge length, the presence of the transverse welds became detrimental (Figs. 10.d and
10.e) such that at a distance equal to 100 mm away from the noise of the flange plate the
presence of the transverse welds causes10 percent increase in the beam flange strains in
compare to the specimen without beam transverse welds. It finally caused that the specimen
with the transverse welds to show a lower ductility than the one without the transverse welds
(Fig. 11).However, their strengths are approximately the same. Based on these results it
canbe concluded that to avoid the strain concentration at the plastic hinge area and to
increase the connection ductility, only longitudinal welds to be used to connect the top
flange plates to the beam flange. However, using the transverse welds can be effective in
reducing the size of the fillet welds and consequently in reducing the thermal stresses at this
location.
1.2
Normalized von-Mises strain
1.0
0.8
0.6
with transverse welds
0.4
with out transverse welds
0.2
a) At the flange plate at face of column
0.0
0.5- 0.3- 0.1- 0.1 0.3 0.5
Normalized distance from the flange plate center
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 29
35
with transverse welds
Normalized Von Mises Strain
60
Normalized von-Mises strain
50
40
30
80
60
40
80
Normalized von-Mises strain
60
40
e) At the beam flange at 180
mm away from the noise of
20 with transverse welds
flange plate
with out transverse welds
0
0.5- 0.4- 0.3- 0.2- 0.1- 0.0 0.1 0.2 0.3 0.4 0.5
Normalized distance from the beam flange center
Figure 10. Normalized von-Mises strain distribution across the beam flange and the flange plate
at various distances away from the column face
1.4
1.2
1.0
M/Mp
0.8
0.6
Without Transverse Welding
0.4
With Transverse Welding
0.2
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Total rotation (%)
Figure 11. Moment-rotation curves of WFP specimens RC06 with and without transverse welds
3.4 Effect of the end form of the top reinforcing plate on the connection ductility
As mentioned before, in Iran, in order to avoid overhead welding, the common used form of
the top flangeplate is a combination of the trapezoidal and the rectangular form (Fig. 2).
Hereafter for simplicity, this form of the top flange plate is named as TR form. In FEMA350
(2000) [12], the top flange plate has a rectangular form and its width is larger than the beam
flange width. In this section the effect of the different forms of the top flange plate on the
ductility of specimen SPE05 was assessed. Fig. 12 compares the Von-Mises strain
distribution for these flange plates and their details are summarized in Table 4.
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 31
Forms 1 and 2 are the same as those are presented in the FEMA350(2000) [12] and the
Iranian steel code respectively. Comparing the von-Mises strain distributions of these two
forms shows that the maximum strain developed for the FEMA350 (2000) [12] form is at
the beginning of the flange plate (i.e. at column face level) and is 11.3εy where εy is the yield
strain of the flange plate material. By using the TR form, the maximum Von-Mises strain
decreased which is due to its higher thickness in compare to the one of the FEMA350 form.
In this case, the maximum strain concentration occurred at the intersection of the rectangular
and the trapezoidal parts (named as SC point).However, these differences between the first
and the second forms did not affect the connection ductility (see Table 4) since in these
connections the failure was due to the beam flange fracture at the plastic hinge area (beyond
the noise of the flange plates) rather than at the flange plates. Fig. 13 compares the
normalized Von-Mises strains across the beam flange width at the plastic hinge area at 6%
total rotation. This figure clearly shows the high level of strains at this area in compare to the
column face level. Also this figure indicates that the change of the top flange plate form did
not have any effect on the strain distribution across the beam flange at the plastic hinge area.
FORM1 FORM2
SC point
32 A.A. Hedayat, H. Saffari and E. Jazebi
FORM3 FORM4
FORM5
FORM6
Figure 12. Von-Mises strain distribution for different forms of the top flange plate
100
Normalized von-Mises strain
80
60
40
1Form 2Form
20
3Form 6Form
0
0.5- 0.4- 0.3- 0.2- 0.1- 0.0 0.1 0.2 0.3 0.4 0.5
It should be noted that the strain concentration at the SC point can be detrimental if a
crack exist at this point which can easily be created during the fabrication of the top flange
plates of the TR form. In this case the stress intensity factor (SIF) may easily exceed the
fracture toughness of the flange plate metal and finally leads to a brittle fracture at this point.
Note that in the finite element modeling the cracks were not considered and the SIFs were
not calculated.
To avoid such brittle failure, in the forms 3 through 5, it was tried to remove the SC
point. As Fig. 12 shows by rounding the sharp corner at the SC point, the maximum strainof
forms 3 and 4 reduced to 6.7εy and 7.2εy respectively. While the fifth form was ineffective to
reduce the maximum Von-Mises strains and also achieved the lowest ductility among the
other forms (see Table 4). Note that the width of this flange plate at its tip was
approximately half of the beam flange width and it caused a higher strain concentration at
the plastic hinge areawhich promoted the beam flange fracture at this region. The sixth form
of the top flange plate is the same as the first form except that its width is less than the beam
flange width and consequently its thickness is larger than the one of the first form. This form
of the flange plate also avoidsthe overhead welding of the rectangular flange plate to the
beam flange. For this form, the connection ductility was 5.3 percent total rotation. This form
caused highest von-Mises strain concentration across the beam flange at the plastic hinge
area in compare to the other forms (Fig. 13).
As the results presented in this sectionshowed, generally there is no remarkable
difference between the connection’s ductility and strength of the WFP connections of
different form of the top flange plate. However, when the connection fabrication follows the
common practice in Iran, among the investigated forms, the third one might be the best
form. Since it reduced the strain concentration at SC point and avoids the overhead welding.
But when the connection fabrication is similar to the American code of practice, still the first
form is the most desirable one since the flange plate thickness is in a minimum level which
results in lower cost and thermal stresses at the column face level.
3.5 Effect of equality and inequality of the top and the bottom plate length on the connection
ductility
The lengths of the top and the bottom flange plates are a function of the required weld length
to connect these plates to the beam flange plates. Generally, in the case of using a flange
plate of the TR form, the initial part (the trapezoidal part) is not welded to the beam flange
and it causes the weld length of the top flange plate to be less than the bottom one. In this
case, in order to avoid increasing the fillet weld sizes, in many designs the top plate length is
considered longer than the bottom one while for the WFP connections presented in
FEMA350 [12], these plates have same length. In this section the effect of the inequality of
the top and the bottom flange plate lengths on the connection ductility was investigated.
Note that in this section due to the asymmetry shape of the connection; all specimens were
analyzed under both upward and downward loading. Fig. 14 shows the moment-rotation
curves of specimen SPE05 for three different cases where the length of the top or the bottom
flange plates are either 48 cm or 38 cm. Note that in all cases the shear force was transferred
to the column flange through the shear tab.
As this figure shows, under upward loading, when the top flange plate is longer, the
connection achieved a higher rotational capacity and experienced lower strength degradation
34 A.A. Hedayat, H. Saffari and E. Jazebi
when it is compared with the one with the top and bottom flange plates of the same length.
However, under downward loading, when the top flange plate is longer, the relevant
moment-rotation curve doesn’t show a better rotation and mostly indicates much more
strength reduction when it is compared to the one of the top and bottom flange plates of the
same length.
When the top flange plate is shorter than the bottom flange plate, under downward
loading, this specimen experienced same rotational capacity and strengthas that was
obtained for specimen of the equal flange plate lengths. However, under upward loading,
when the top flange plate is shorter, the relevant moment-rotation curve doesn’t show a
better rotation and mostly indicates much more strength reduction when it is compared to the
one of the top and bottom flange plates of the same length.The same results have been
observed in assessing the specimen SPE07. Therefore, it might be concluded that for the
both upward and downward loading, the best connection performance may happen when the
length of the both top and bottom flange plates are equal.
1.6
1.2
0.8
M/Mp
0.4
0.0
38BFP48 TFP
0.4-
48BFP38 TFP
0.8-
48BFP48 TFP
1.2-
1.6-
6.0- 4.0- 2.0- 0.0 2.0 4.0 6.0
Total rotation (%)
Figure 14. Moment-rotation curves of WFP specimens SPE05 in case of using different length
for the top and the bottom flange plates
(Hrib/LP=1). While in the case of using long flange plates (LP/db>0.6), the rib plates must
have a minimum Hrib/LP equal to 0.6. Fig. 16 compares the moment-rotation curves of the
WFP specimens RC06 with the shear and the flexural rib plates. This figure clearly shows a
delay in the onset of the beam flange and web local buckling (i.e. the onset of the strength
degradation) when the horizontal length of the rib plates were increased. It finally led to an
increase in both the connection’s strength and ductility.
Table 5: Strength and ductility of WFP specimens RC06 with various rib plates
Rib M/Mp at
LP vrib Number
Specimen plate LP/db Hrib(mm) trib(mm) Hrib/LP θ(%) failure
(mm) (mm) of ribs
type time
1 380 0.51 100 200 15 0.26 4 1.92 1.17
2 380 0.51 200 200 15 0.53 4 2.6 1.2
3 380 0.51 300 200 15 0.79 4 2.8 1.25
4 380 0.51 380 200 15 1 4 3.08 1.11
RC06
5 380 0.51 380 200 15 1 4 3.08 1.11
1 480 0.64 100 200 15 0.21 4 5.15 0.96
2 480 0.64 200 200 15 0.42 4 5.38 0.97
3 480 0.64 380 200 15 0.79 4 6 1.08
Figure 15. Details of the various types of the used rib plates
36 A.A. Hedayat, H. Saffari and E. Jazebi
1.4
1.2
1.0
0.8
M/Mp
0.6
)1TYPE (
0.4
)2TYPE (
0.2
)3TYPE (
0.0
0.0 1.0 2.0 3.0 4.0 5.0 6.0
Total rotation(%)
Figure 16. Moment-rotation curves of the WFP specimens RC06 (LP=48 cm)with the shear and
the flexural rib plates
4. SUMMARY OF RESULTS
Evaluation of the WFP connections presented in the previous section revealed that these
connections generally can achieve adequate rotational capacity (i.e. total rotation greater
than 0.04 rad) provided that:
1. The length of the top and the bottom flange plates is same.
2. Depends on the ratio of the Lp/db, the appropriate method is used to transfer the shear
forces.
3. Depends on the fabrication method, the appropriate form of the top flange plate is
used (i.e. forms 1 and 3).
4. Flexural rib plates are used when the shear forces are transferred through the rib
plates.
As mentioned before, the connection strength can be evaluated using M/MP ratio. Based
on the seismic provision (AISC) [29], from the strength point of view, a connection is
acceptable if the measured flexural resistance of the connection determined at the column
face, is at least 0.80MP of the connected beam at a story drift angle of 0.04 rad. Generally
connections show higher strength degradation when they are subjected to the cyclic
loading[9]. Examination of the RC06 cyclic and monotonic analyses (Fig. 4.a), it was found
that the ratio of the connection strength under the cyclic loading to the one under the
monotonic loading at 4 percent total rotation is equals to 0.71. Hence, it might be concluded
that the connection strength obtained from the monotonic analysis can be turned to the cyclic
one by multiplying to a reduction factor equal to 0.71. Note that in the previous section, all
the presented results were obtained based on the monotonic loading. This indicates that all
the connection strengths presented in the previous section must be modified by multiplying
to a reduction factor of 0.71. With respect to this issue, the strength at four percent total
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 37
For the WFP connections the beam web buckling is generally followed by the beam
flange buckling and it finally leads to the beam flange fracture at the plastic hinge region.
Hence, the beam flange fracture at the plastic hinge region might be controlled if the beam
web is strengthened using a doubler plate. This strengthening method is shown in Fig. 17,
where the thickness of the doubler plate is 5 mm.
The maximum height of the doubler plate (Hdp) was determined using Hdb=db-2k formula,
where db and k are the beam overall depth and the distance from the outer face of the beam
flange to the web toe of the fillet respectively. The length of the doubler plate (Ldp) was
defined as a function of parameter b (Ldp=100 mm +b) and was located in such a way that
100mm of that, was located before the noise of the flange plate (Fig. 17).
Fig. 18 compares the moment-rotation curves of specimens RC06 (LP=380mm) under
monotonic loading for four cases: without doubler plate and with doubler plate when the
parameter b is equal to 100 mm, 200 mm and 300 mm which are corresponded to the b/db
ratio equal to 0.13, 0.27 and 0.4 respectively (db is the beam overall depth). As Fig. 18
shows the presence of the doubler plate was effective to increase the maximum connection
38 A.A. Hedayat, H. Saffari and E. Jazebi
flexural capacity. It also caused a remarkable increase in the connection strength at four
percent total rotation such that at this point the ratio of the M/MP increased from 1.04 to
1.21when b/dbwas equal to 0.27.By increasing the doubler plate length (or b/db ratio) the
connection strength increased. However, it caused a reduction in the connection ductility.
This investigation was also carried out for specimens SPE03 and SPE07 where the beam
overall depth of these specimens was 600 and 912 mm respectively. Table 6 summarizes the
results of these specimens. In this table, M/MP is the connection strength at four percent total
rotation under cyclic loading and θ is the total rotation of connection at failure time. The
results of this table clearly show the efficiency of the doubler plate to enhance the strength
of all WFP specimens. However, as this table shows excessive increase in the horizontal
length of the doubler plate was detrimental and caused a significant reduction in the
connection ductility. Based on these limited results it might be concluded that the optimum
value for b/db ratio to achieve both adequate connection strength and ductility is 0.13.
Figure 17. Strengthening of the beam web at the plastic hinge area using a doubler plate
1.4
1.2
1.0
0.8 b/db=0.27
M/Mp
0.6 b/db=0.13
0.4 b/db=0.4
0.2 with out doubler plate
0.0
0 1 2 3 4 5 6
Total rotation
Figure 18. Moment-rotation curves of specimens RC06 with doubler plates of different size
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 39
Table 6: Results of WFP specimens with different sizes of the doubler plate
Specimen Hdb(mm) Ldb(mm) t db(mm) b/db M/Mp θ(%)
- - - - 0.71 5.5
600 200 5 0.13 0.82 6
RC06
600 300 5 0.27 0.86 6
600 400 5 0.4 - 2.7
- - - - 0.81 5.46
SPE07 750 200 5 0.11 0.92 6
750 300 5 0.22 - 3.07
- - - - 0.79 4.92
SPE03 400 200 5 0.16 0.82 6
400 300 5 0.32 - 3
6. CONCLUSIONS
This study was aimed to investigate the effective parameters on the strength and ductility of
the WFP connections with deep beams using the finite element method. The investigated
parameters included, the material property of the beam, flange plate thickness, flange plate
length, the transverse welding of the top flange plates, the form of the top flange plate, the
equality or inequality of the top and bottom flange plate lengths and form of the rib plates.
The conclusions drawn from this finite element study can be summarized as follows:
1. Decrease in the yield stress of the beam material delayed the initiation of the beam flange
fracture at the plastic hinge region. This finally led to increase in the connection’s
strength, ductility and initial rotational stiffness. It also caused a reduction in the strength
degradation of the connection after its yield point.
2. Similar to the other welded connections (e.g. pre or post-Northridge connections), the
classic beam theory was invalid for the WFP connections. In other word, the flange plates
considerably participated to transfer the shear force from the beam to the column.
Consequently, It makes importance the selection of an appropriate method for
transferring the shear forces from the beam to the column in order to reduce the tri-
axially stresses at the flange plates at face of the column and to prevent the early fracture
at this region.
3. The appropriate method to transfer the shear force from the beam to the column was a
function of the LP/db ratio. When LP/db is more than o.6, generally using the rib plates
was better than the use of the shear plate, whereas in LP/db range which is less than 0.6,
using the shear tab was the appropriate method.
4. Transverse welds which connect the noise of the top flange plate to the beam flange had no
effect on the strength of the WFP connections. These welds caused a strain reduction across
the beam flange at the noise of the top flange plate. However, this positive effect caused a
strain concentration at the plastic hinge area and finally led to a slightly reduction in the
connection ductility in compared to the one without transverse welds. Hence, it is
recommended to use only the longitudinal welds for connecting the top flange plates to the
beam flange. However, using the transverse welds can be effective to reduce the size of
40 A.A. Hedayat, H. Saffari and E. Jazebi
fillet welds and consequently to decrease the thermal stresses at this area.
5. Generally the form of the top flange plate had no remarkable effect on the strength and
ductility of the WFP connections. However, among all six forms, using the first, the
second and the third forms led to the achievement of higher connection ductility with less
strain concentration at the top flange plate at the column face level. Despite of the higher
fabrication cost of the third form in compare to the second form, using the third form is
recommended. Because in this case by rounding the intersection of the rectangular and
the trapezoidal parts of the top flange plate, a remarkable reduction was achieved in the
strain concentration at the SC point and consequently this reduces the probability of the
fracture initiation at this critical point. In addition, using the fifth form of top flange plate
led to the achievement of the minimum connection rotational capacity. However, it had
no effect on the connection strength.
6. The best connection performance can be achieved when the length of the both top and
bottom flange plates are equal.
7. By increasing the horizontal length of the rib plates the connection strength and ductility
increased. Hence, in the case of using rib plates to transfer the shear forces from the beam
to the column, the highest connection performance might be achieved when the flexural
rib plates are used.
8. Considering all the recommendations presented in the above paragraphs, was
significantly effective to increase the ductility of the WFP connections to achieve the
minimum required ductility. However, the effects of these recommendations on the
connection’s strength enhancement were not pronounced.
Adding doubler plate at the beam web was effective to delay the beam web buckling and
to increase the connection strength. The horizontal length of the doubler plate played a
significant role on the connection ductility. Excessive increase in this parameter significantly
reduced the connection ductility. Based on these limited results, the highest connection
performance might be achieved when the b/db ratio of the doubler plate is limited to 0.13.
REFERENCES
1. Ramirez CM, Lignos DG, Miranda E, Kolios D. Fragility functions for pre-Northridge
welded steel moment-resisting beam-to-column connections, Engineering Structures,
45(2012) 574-84.
2. Mahin ST. Lessons from damage to steel buildings during the Northridge earthquake,
Engineering Structures, 20(1998) 261-70.
3. Miller DK. Lessons learned from the Northridge earthquake, Engineering Structures,
20(1998) 249-60.
4. Engelhardt MD, Sabol TA. Reinforcing of steel moment connections with cover plates:
benefits and modifications, Engineering Structures, 20(1998) 510-20.
5. SAC. Experimental investigations of beam-column sub assemblages, Technical report
SAC-96-01, parts 1 and 2. Sacramento, CA, SAC Joint Venture, (1996) 739-46.
INVESTIGATION OF THE EFFECTIVE PARAMETERS ON THE STRENGTH ... 41