- Email: [email protected]
Numerical study of Slotted-Web–Reduced-Flange moment connection
Journal of Constructional Steel Research 69 (2012) 1–7
Contents lists available at ScienceDirect
Journal of Constructional Steel Research
Numerical study of Slotted-Web–Reduced-Flange moment connection Shervin Maleki a,⁎, Maryam Tabbakhha b a b
Department of Civil Engineering, Sharif University of Technology, Tehran, Iran Ecole Centrale Paris, France
a r t i c l e
i n f o
Article history: Received 5 March 2011 Accepted 6 June 2011 Available online 16 September 2011 Keywords: Moment connections Reduced Beam Section Slotted Beam Web Panel zone Continuity plates Slotted-Web–Reduced-Flange Non-linear analysis
a b s t r a c t Reduced Beam Section (RBS) and Slotted Beam Web (SBW) are two types of seismic resistant moment connections that were introduced after the 1994 Northridge earthquake. These connections have been tested under cyclic loading and have had acceptable performance. In this paper, a new hybrid connection is introduced that is composed of RBS and SBW and is named Slotted-Web–Reduced-Flange (SWRF). Nonlinear finite element analyses are performed on SWRF under cyclic loading. It is shown that the new connection in some cases performs better than its RBS and SBW predecessors. The effects of panel zone strength, continuity plates and slot length are also investigated. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction Steel moment frames were considered an appropriate solution to stability of buildings in seismic regions because of their perceived ductility. After the 1994 Northridge earthquake, when many steel buildings with moment frames were seriously damaged, this notion proved to be wrong. The scientific community realized that more research is necessary to better understand the behavior of these connections under cyclic loading. The “pre-Northridge” moment connections consisted of bolted or welded web shear plates and complete joint penetration (CJP) beam flange welds. These connections proved to be susceptible to brittle failure under seismic loading. Researchers have suggested two solutions to improve ductility of moment connections after the Northridge earthquake. First, increasing the strength of the connections by using stiffeners and appropriate welds in order to prevent premature connection damage, and second, making a beam section weak at a section away from the column face, so that a plastic hinge is forced to occur at this section without damaging the column. The idea of connections with Reduced Beam Section (RBS) and Slotted Beam Web (SBW) is based on the latter concept. RBS connection relies on the selective removal of beam flange material adjacent to the beam-to-column connection, typically from both top and bottom flanges, to reduce the cross sectional area of the beam. This reduction in cross sectional area will reduce the moment capacity at a discrete location in the beam and so plastic hinge occurs in the beam [1,2]. ⁎ Corresponding author. Tel.: + 98 21 66164251. E-mail address: [email protected] (S. Maleki). 0143-974X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcsr.2011.06.003
In SBW connection, slotted beam design allows the beam flanges and the beam web to buckle independently. This circumvents the beam lateral-torsional buckling mode that occurs in non-slotted beams. Therefore, the torsional moment and torsional stresses in the beam flanges and welds at the column flange that result from this buckling mode are omitted. The separation of the beam web and the flange results in a biaxial rather than triaxial stress and strain states in the region of the connection, which increases its fatigue life [3,4]. One of the parameters which can considerably influence the failure mode of the beams with RBS moment connections is the column panel zone (PZ) ductility. Krawinkler [5] and Popov [6] indicated that the beamto-column joints with weak PZs encounter high shear deformation, resulting in brittle fracture within the weld connecting the beam flange to the column face. As a result, in spite of the weak PZ ability to dissipate a large amount of energy, using very weak joints is not recommended. On the contrary, in the presence of strong PZs, the fracture potential is reduced, but the possibility for beam instability rises, especially for RBS connections. Tsai et al. [7] and Jones et al. [8] experimentally illustrated that balanced PZs show appropriate performance. Studies also show desirable performance of SBW connections. Richard et al. [3] conducted successful experiments on these connections. The Slotted Beam Web connection designs reduce the stress concentration factor at the beam-to-column flange interface at the column web from typical values ranging from 4.5 to 5.5 in non-slotted beams down to a typical value of about 1.4 by providing a nearly uniform beam flange/weld stress and strain distribution. The SSDA [4] slotted web design develops the full plastic moment capacity of the beam, moves the plastic hinge region in the beam away from the face of the column, and results in nearly uniform tension and compression stresses, strains across and through the beam flanges from the face of the column
2
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
to the end of the slot. It eliminates the lateral-torsional buckling mode that occurs in non-slotted beams, and enhances ductility by reducing the residual weld stresses. Considering the advantages of RBS and SBW connections, an innovative hybrid connection is examined in this paper in which both beam flanges are reduced in section (like RBS) and the beam web is cut with two slots (like SBW). The connection is named “Slotted Web– Reduced Flange” or SWRF, for short. In this way, the plastic moment forms away from the column face and the web slots reduce stress concentrations and eliminate lateral-torsional buckling modes. A series of nonlinear finite element analyses are performed on this connection using the cyclic load pattern suggested by AISC Seismic Provisions [9] which is similar to the recommendation of SAC committee [10]. The effects of panel zone strength, continuity plates and slot length are also investigated for this new connection. The results are compared with RBS and SBW connections.
Table 1 Details of SWRF specimens. Specimen
Column section
Beam section
Doubler plate thickness (mm)
Vy (kN)
Vr (kN)
Vr / Vy
Weak Balanced Strong
HE200B HE200B HE200B
IPE300 IPE300 IPE300
0 6 10
337.5 517.5 637.5
404.6 404.6 404.6
1.20 0.78 0.63
To determine the required strength (Vr), and the design shear strength of the panel zone (Vy), the recommendation of AISC Seismic Provisions [9] are used. They are computed as follows: 2
Vy = 0:6Fy dc tpz "
2. Analytical study of the SWRF moment connection Vr = βE MP An analytical study is conducted on the slotted web–reduced flange (SWRF) connection to investigate the cyclic non-linear behavior of the connection and to find the moment–rotation curves and critical locations of stresses. Subassemblies with a general configuration as shown in Fig. 1 are considered for analyses. Each subassembly consists of an IPE300 beam and an HE200B column, which are rolled European sections. The column is supported at the base by a hinge, while a vertical roller is used at the other end. The beam flanges are laterally braced at a distance of 150 cm from the column face. This distance satisfies the AISC seismic provisions [9] recommendation. Nonlinear material behavior is considered in all analyses. Nine specimens of SWRF connection are analyzed. The PZ effects are considered by varying the PZ thickness. Three cases of weak, balanced and strong PZs are analyzed. The continuity plates are also added to the 3 PZ cases above to examine their contribution. Four different web slot lengths are also considered. The details of these subassemblies are described in sections to follow.
3bcf tcf 1+ db dc tpz
L + 1 − b 0:95db Lb
!
dc 2
ð1Þ
1 : H
# ð2Þ
in which Fy = yield stress of the PZ material; MP = plastic moment capacity of the beam section; Lb = beam length from the column face to the beam tip; and H = column height. Dimensions dc, db, bcf, tpz and tcf correspond to column depth, beam depth, column flange width, panel zone thickness and column flange thickness, respectively. In Eq. (2), βEMP is the flexural demand imposed at the column face. The suggested value of βE is between 0.85 and 1[11]. Here, βE = 0.85 is used throughout. These are calculated in Table 1 for the three PZ varying specimens. 2.2. Continuity plate variation In addition to the panel zone strength, the effects of continuity plates are investigated on the seismic performance of the SWRF connection. In analytical models with continuity plates, the plate thickness is set equal to the beam flange. 2.3. Details of RBS region
2.1. Panel zone variation The IPE300 beam and the HE200B column are pre-selected so that they can produce a weak column panel zone (PZ) at the intersection. Further, for evaluating the effect of PZ thickness, the column web in some specimens is reinforced with 6 and 10 mm thick doubler plates in order to provide balanced and strong PZs, respectively. The three specimens analyzed are shown in Table 1.
The radius cut shown in Fig. 2 was employed for the RBS region. The RBS connections are detailed according to the recommendations of AISC 358 [12]. The dimensions used are shown in Table 2. 2.4. Details of web slots The beam web slot is designed according to recommendations of reference [4]. Four slot lengths, as shown in Table 3, are used in analyses. In normal cases, the length of slot is 175 mm (LS2) according to the design procedure for this type of connection [4]. However, various lengths for web slots are considered to examine the effects of this parameter. These lengths are shown in Table 3 and correspond to slots extending to the end of shear plate (LS1), to the middle of beam flange reduced section (LS3) and to a distance away from the reduced section (LS4).
R
c c a Fig. 1. Subassembly model of SWRF connection.
b
Fig. 2. RBS connection detail.
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
3
3.1. Elements and meshing
Table 2 RBS region dimensions (mm). Beam
a
b
c
R
tf
tw
bf
d
IPE300
80
200
32
172.25
10.7
7.1
150
300
Table 3 Slot lengths. Specimen
LS 1
LS2
LS3
LS4
Slot length (mm)
110
175
180
460
The subassemblies are modeled using a quadrilateral four-node shell element (element SHELL181 in ANSYS). This element has material plasticity, large deflection, and large strain modeling capabilities. It has six degrees of freedom per node: translations in the x, y, z directions, and rotations about the x, y, z axes. Fig. 4 shows a typical finite element meshing used in this study. As seen in Fig. 4, a more refined mesh is employed for the regions near the RBS and web slots.
3.2. Material modeling In all specimens slot width is 3.2 mm to the end of shear plate and 6.4 mm from the end of the shear plate to the slot termination point (see Fig. 3). All slots are terminated to a hole with 21 mm diameter. The shear plate size is 10 × 110 × 190 mm in all specimens.
Nonlinear material with kinematic hardening is used for steels. The plasticity model was based on the von Mises yielding criterion and its associated flow rule. Stress–strain curve of ASTM A572 steel is shown in Fig. 5. The Young's modulus of 0.21 × 10 6 MPa and Poisson's ratio of 0.3 are used as elastic constants.
3. Finite element analysis 3.3. Loading protocol The ANSYS finite element software [13] is utilized to model the specimens for large deformation nonlinear analysis. The analyses are primarily intended to investigate the overall cyclic behavior of the SWRF subassemblies with an emphasis on the influence of PZ strength, continuity plates and slot length. The details are given below.
Each subassembly is loaded at the beam free end in displacement control such that the drift angle of the subassembly satisfies the AISC Seismic Provisions loading protocol [9] (see Table 4). All specimens are loaded up to 0.04 rad of drift angle.
Fig. 3. SBW connection details.
Fig. 4. Three-dimensional finite element model.
Fig. 5. Steel stress–strain curve.
4
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
Before loading the beam, buckling mode shapes of the model were computed in a separate buckling analysis by the software. The deformed buckled shape is then analyzed under cyclic loading to take local and lateral buckling effects into account.
Table 4 Loading protocol. 6 cycles 6 cycles 6 cycles 4 cycles 2 cycles 2 cycles 2 cycles 2 cycles
@ @ @ @ @ @ @ @
0.00375 rad 0.005 rad 0.0075 rad 0.01 rad 0.015 rad 0.02 rad 0.03 rad 0.04 rad
4. Results of analysis and discussion In order to verify the validity of the finite element model, the specimen used in the experimental study conducted by Jones et al. [8] was analyzed under cyclic displacement control loading using the
SWRF Weak Pz
SWRF Balanced Pz
SWRF Strong Pz
SWRF Weak PZ+ continuity plate
SWRF Balance PZ+ continuity plate
SWRF Strong PZ+ continuity plate
Fig. 6. Hysteretic response of SWRF connections with and without continuity plates.
LS1=110 mm
LS3=180 mm
LS4=460mm
Fig. 7. Hysteretic response of SWRF with various slot lengths and added continuity plates.
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
5
4.1. Hysteretic responses Beam moment–rotation hysteretic responses of the subassemblies resulting from the finite element analyses are shown in Fig. 6. The beam moment was measured at the column face, and the rotation was computed by dividing the total beam tip deflection by the beam length. The area under this curve indicates the amount of seismic energy dissipation by the connection yielding parts. In Fig. 6 the first row shows the moment–rotation curves for SWRF connections with different panel zone thicknesses when continuity plates are not used; while the second row shows the behavior for the same connections when the continuity plates are present. The effects of slot length in SWRF connection can be examined in Fig. 7. All these subassemblies have weak panel zones and added continuity plates. 4.2. Steel stresses
Fig. 8. von Mises stresses (Pa), SWRF with weak PZ and continuity plates at 4% drift.
ANSYS finite element software. The results agreed well with the experimental results [14]. After this verification, the subassembly models described above are analyzed and the parameters varied. The results of cyclic loading on SWRF connections are presented below and later compared with RBS and SBW connections with the same characteristics.
For the sake of brevity, the stresses at the final step of analysis for the SWRF connection with weak PZ and continuity plates are shown in Fig. 8. The stresses are in Pascal unit. The strength degradation is due to yielding of plate material in different parts of connection. This figure shows the critical (yielding) locations in steel for the SWRF connection. 4.3. Effect of PZ strength In Fig. 6, among the six analyzed specimens, it is observed that the best performing connection is the SWRF with weak panel zone and with continuity plates. It has stable hysteresis loops without strength degradation up to 0.04 rad of drift angle and the greatest energy
RBS Weak PZ
RBS Balanced PZ
RBS Strong PZ
RBS Weak PZ+ continuity plate
RBS Balanced PZ+ continuity plate
RBS Strong PZ + continuity plate
Fig. 9. Hysteretic response of RBS connections with and without continuity plates.
6
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
SBW Weak PZ
SBW Balanced PZ
SBW Strong PZ
SBW Weak PZ+ continuity plate
SBW Balanced PZ+ continuity plate
SBW Strong PZ + continuity plate
Fig. 10. Hysteretic response of SBW connections with and without continuity plates.
dissipation. Also, analysis shows that the smallest von Mises stresses occur in this case (Fig. 8). This result is important because it shows that column web without doubler plate is more suitable for SWRF connections. Balanced PZs and strong PZs experience strength degradation at 0.04 rad drift.
4.4. Effect of continuity plates In Fig. 6, comparing the top row with the bottom row, it is seen that continuity plates increase the energy dissipation in the connection. Analysis also shows that the maximum stresses of the connection are lowered slightly (Fig. 8). Moreover, continuity plates cause the location of maximum von Mises stresses to shift away from the column–beam weld location toward the beam web. Therefore, stress concentration is reduced at the critical weld location.
4.5. Effect of slot length As noted in Table 3, four specimens are examined for determining the effects of slot length on SWRF connection performance. From Fig. 7, it is seen that very long slots have poor performance. It should be noted that the slot length of 175 mm (LS2 of Table 3, standard slot length) was already shown in Fig. 6 (bottom row, left side). Results indicate (see Fig. 7) that slots with longer length than what is recommended do not improve the performance of the connection and cause greater stresses in the specimens. Specimen with 460 mm slot length (LS4) experienced instability because of large deformations. Specimen with the shortest slot length (110 mm) dissipates the greatest amount of energy.
4.6. Comparison of SWRF, RBS and SBW connections In order to compare the performance of SWRF connection with that of RBS and SBW connections, similar analyses are conducted on the latter two connections. The moment–rotation curves for the RBS and SBW connections are determined and illustrated in Figs. 9 and 10, respectively (similar to Fig. 6 for the SWRF). The stresses for the weak PZ and added continuity plates (similar to Fig. 8 for the SWRF) are shown in Figs. 11 and 12. Comparing Figs. 6, 9 and 10, shows that RBS connection has the best performance among the cases considered. As shown in Fig. 9, RBS connections experience stable hysteresis curves in all 6 variations; whereas, SBW connections with continuity plates, in balanced and strong PZ cases, and SWRF connections in balanced and strong PZ cases, deteriorate in the final cycle. Although the results are not shown for brevity, a typical preNorthridge connection is also examined as part of this research. This connection is modeled by directly attaching the beam flanges to the column without any slots in the web or cuts in the flanges. The preNorthridge connection experiences nearly double the maximum stresses of the SWRF connection at 0.04 rad of rotation. Therefore, the new connections are all definitely better than the pre-Northridge connection. However, the new connections each have their own merits and they are summarized in the next section. 5. Conclusions Previous analytical and experimental research has indicated many advantages in using RBS and SBW moment connections in steel structures subjected to earthquake loading. This paper investigated the possibility of having both connection details and their advantages in one
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
7
Fig. 12. von Mises stresses (Pa), SBW with weak PZ and continuity plates at 4% drift. Fig. 11. von Mises stresses (Pa), RBS with weak PZ and continuity plates at 4% drift.
new connection called Slotted Web–Reduced-Flange (SWRF) connection. A non-linear finite element model of a typical subassembly was used to examine the hysteretic behavior of the new connection analytically. The effects of panel zone thickness, continuity plate existence and web slot length are also investigated. The results for all three types of connections are compared with each other. Although all connections are better than the non-ductile pre-Northridge connection, but each has its own advantage. The following general conclusions can be drawn. • Results indicate that continuity plates have positive effects on all 3 types of connections and reduce the maximum von Mises stresses. In SWRF connection, continuity plates cause the location of maximum stresses to move away from the beam/column interface, where there exist stress concentrations due to welding. Therefore, using continuity plates is recommended. • In general, in SBW connections the use of doubler plates on the column is not recommended. A weak panel zone works fine in most cases. • The RBS connection overall outperforms the SBW and SWRF connections in terms of energy dissipation and stable hysteresis loops. • In the new SWRF connection, the best performance is achieved by having a weak panel zone (no doubler plate on column web) and continuity plates. In this case, the maximum von Mises stresses are located in the beam web and not flanges. • A short slot length up to the end of the shear plate is recommended for SWRF connections. This can be increased to what is recommended in SBW connections if necessary, but beyond that the connection will be adversely affected. • SWRF and SBW connections both have the advantage of reducing stresses at beam/column interface and preventing lateral torsional
buckling of the beam and reducing the vertical shear forces in the beam flanges. These recommendations are all based on numerical studies and computer simulations. To ascertain, physical testing on SWRF connections is definitely needed.
References [1] Chen SJ, Yeh CH, Chu JM. Ductile steel beam-to-column connections for seismic resistance. ASCE, J Struct Engng 1996;122(11):1292–9. [2] Moore KS, Malley JO, Engelhardt MD. Design of reduced beam section (RBS) moment connections. Steel Tips, Structural Steel Education Council, Moraga, CA; 1996. [3] Richard RM, Allen CJ, Partridge JE. Proprietary slotted beam connection design. Modern Steel Construction 1997:28–33. [4] Seismic Structural Design Associates (SSDA). Beam slot connection design manual. Laguna Niguel, California; 1998. [5] Krawinkler H. Shear in beam–column joints in seismic design of steel frames. Engineering Journal, AISC 1978:82–91 [3 rd Quarter]. [6] Popov EP. Panel zone flexibility in seismic moment joints. In: Chen WF, editor. Joint flexibility in steel frames. London: Elsevier Applied Science; 1987. [7] Tsai KC, Chen WZ. Seismic response of steel reduced beam section to weak panel zone moment connections. In: Mazzolani FM, Tremblay R, editors. Behavior of steel structures in seismic areas. Rotterdam: Balkema; 2002. [8] Jones SL, Fry GT, Engelhardt MD. Experimental evaluation of cyclically loaded reduced beam section moment connections. ASCE, J Struct Engng 2002;128(4):441–51. [9] American Institute of Steel Construction (ANSI/AISC 341–05). Seismic provisions for structural steel buildings. Chicago, IL; 2005. [10] SAC. Seismic design criteria for new moment-resisting steel frame construction. Report no. FEMA 350, SAC Joint Venture, Sacramento, CA; 2000. [11] Englehardt MD. Design of reduced beam section moment connections. Proceedings of North American Steel Construction Conference, AISC, Chicago; 1998. [12] American Institute of Steel Construction (ANSI/AISC 358–05). Prequalified connections for special and intermediate steel moment frames for seismic applications, Chicago; 2006. [13] ANSYS (Release10), ANSYS, Inc., Southpointe 275 Technology Drive, Canonsburg, PA. [14] Maleki S, Tabbakhha M. Analytical study of reduced section moment connections. Proceedings of EASEC 11 conference, Taipei, Taiwan; 2008.
Contents lists available at ScienceDirect
Journal of Constructional Steel Research
Numerical study of Slotted-Web–Reduced-Flange moment connection Shervin Maleki a,⁎, Maryam Tabbakhha b a b
Department of Civil Engineering, Sharif University of Technology, Tehran, Iran Ecole Centrale Paris, France
a r t i c l e
i n f o
Article history: Received 5 March 2011 Accepted 6 June 2011 Available online 16 September 2011 Keywords: Moment connections Reduced Beam Section Slotted Beam Web Panel zone Continuity plates Slotted-Web–Reduced-Flange Non-linear analysis
a b s t r a c t Reduced Beam Section (RBS) and Slotted Beam Web (SBW) are two types of seismic resistant moment connections that were introduced after the 1994 Northridge earthquake. These connections have been tested under cyclic loading and have had acceptable performance. In this paper, a new hybrid connection is introduced that is composed of RBS and SBW and is named Slotted-Web–Reduced-Flange (SWRF). Nonlinear finite element analyses are performed on SWRF under cyclic loading. It is shown that the new connection in some cases performs better than its RBS and SBW predecessors. The effects of panel zone strength, continuity plates and slot length are also investigated. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction Steel moment frames were considered an appropriate solution to stability of buildings in seismic regions because of their perceived ductility. After the 1994 Northridge earthquake, when many steel buildings with moment frames were seriously damaged, this notion proved to be wrong. The scientific community realized that more research is necessary to better understand the behavior of these connections under cyclic loading. The “pre-Northridge” moment connections consisted of bolted or welded web shear plates and complete joint penetration (CJP) beam flange welds. These connections proved to be susceptible to brittle failure under seismic loading. Researchers have suggested two solutions to improve ductility of moment connections after the Northridge earthquake. First, increasing the strength of the connections by using stiffeners and appropriate welds in order to prevent premature connection damage, and second, making a beam section weak at a section away from the column face, so that a plastic hinge is forced to occur at this section without damaging the column. The idea of connections with Reduced Beam Section (RBS) and Slotted Beam Web (SBW) is based on the latter concept. RBS connection relies on the selective removal of beam flange material adjacent to the beam-to-column connection, typically from both top and bottom flanges, to reduce the cross sectional area of the beam. This reduction in cross sectional area will reduce the moment capacity at a discrete location in the beam and so plastic hinge occurs in the beam [1,2]. ⁎ Corresponding author. Tel.: + 98 21 66164251. E-mail address: [email protected] (S. Maleki). 0143-974X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcsr.2011.06.003
In SBW connection, slotted beam design allows the beam flanges and the beam web to buckle independently. This circumvents the beam lateral-torsional buckling mode that occurs in non-slotted beams. Therefore, the torsional moment and torsional stresses in the beam flanges and welds at the column flange that result from this buckling mode are omitted. The separation of the beam web and the flange results in a biaxial rather than triaxial stress and strain states in the region of the connection, which increases its fatigue life [3,4]. One of the parameters which can considerably influence the failure mode of the beams with RBS moment connections is the column panel zone (PZ) ductility. Krawinkler [5] and Popov [6] indicated that the beamto-column joints with weak PZs encounter high shear deformation, resulting in brittle fracture within the weld connecting the beam flange to the column face. As a result, in spite of the weak PZ ability to dissipate a large amount of energy, using very weak joints is not recommended. On the contrary, in the presence of strong PZs, the fracture potential is reduced, but the possibility for beam instability rises, especially for RBS connections. Tsai et al. [7] and Jones et al. [8] experimentally illustrated that balanced PZs show appropriate performance. Studies also show desirable performance of SBW connections. Richard et al. [3] conducted successful experiments on these connections. The Slotted Beam Web connection designs reduce the stress concentration factor at the beam-to-column flange interface at the column web from typical values ranging from 4.5 to 5.5 in non-slotted beams down to a typical value of about 1.4 by providing a nearly uniform beam flange/weld stress and strain distribution. The SSDA [4] slotted web design develops the full plastic moment capacity of the beam, moves the plastic hinge region in the beam away from the face of the column, and results in nearly uniform tension and compression stresses, strains across and through the beam flanges from the face of the column
2
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
to the end of the slot. It eliminates the lateral-torsional buckling mode that occurs in non-slotted beams, and enhances ductility by reducing the residual weld stresses. Considering the advantages of RBS and SBW connections, an innovative hybrid connection is examined in this paper in which both beam flanges are reduced in section (like RBS) and the beam web is cut with two slots (like SBW). The connection is named “Slotted Web– Reduced Flange” or SWRF, for short. In this way, the plastic moment forms away from the column face and the web slots reduce stress concentrations and eliminate lateral-torsional buckling modes. A series of nonlinear finite element analyses are performed on this connection using the cyclic load pattern suggested by AISC Seismic Provisions [9] which is similar to the recommendation of SAC committee [10]. The effects of panel zone strength, continuity plates and slot length are also investigated for this new connection. The results are compared with RBS and SBW connections.
Table 1 Details of SWRF specimens. Specimen
Column section
Beam section
Doubler plate thickness (mm)
Vy (kN)
Vr (kN)
Vr / Vy
Weak Balanced Strong
HE200B HE200B HE200B
IPE300 IPE300 IPE300
0 6 10
337.5 517.5 637.5
404.6 404.6 404.6
1.20 0.78 0.63
To determine the required strength (Vr), and the design shear strength of the panel zone (Vy), the recommendation of AISC Seismic Provisions [9] are used. They are computed as follows: 2
Vy = 0:6Fy dc tpz "
2. Analytical study of the SWRF moment connection Vr = βE MP An analytical study is conducted on the slotted web–reduced flange (SWRF) connection to investigate the cyclic non-linear behavior of the connection and to find the moment–rotation curves and critical locations of stresses. Subassemblies with a general configuration as shown in Fig. 1 are considered for analyses. Each subassembly consists of an IPE300 beam and an HE200B column, which are rolled European sections. The column is supported at the base by a hinge, while a vertical roller is used at the other end. The beam flanges are laterally braced at a distance of 150 cm from the column face. This distance satisfies the AISC seismic provisions [9] recommendation. Nonlinear material behavior is considered in all analyses. Nine specimens of SWRF connection are analyzed. The PZ effects are considered by varying the PZ thickness. Three cases of weak, balanced and strong PZs are analyzed. The continuity plates are also added to the 3 PZ cases above to examine their contribution. Four different web slot lengths are also considered. The details of these subassemblies are described in sections to follow.
3bcf tcf 1+ db dc tpz
L + 1 − b 0:95db Lb
!
dc 2
ð1Þ
1 : H
# ð2Þ
in which Fy = yield stress of the PZ material; MP = plastic moment capacity of the beam section; Lb = beam length from the column face to the beam tip; and H = column height. Dimensions dc, db, bcf, tpz and tcf correspond to column depth, beam depth, column flange width, panel zone thickness and column flange thickness, respectively. In Eq. (2), βEMP is the flexural demand imposed at the column face. The suggested value of βE is between 0.85 and 1[11]. Here, βE = 0.85 is used throughout. These are calculated in Table 1 for the three PZ varying specimens. 2.2. Continuity plate variation In addition to the panel zone strength, the effects of continuity plates are investigated on the seismic performance of the SWRF connection. In analytical models with continuity plates, the plate thickness is set equal to the beam flange. 2.3. Details of RBS region
2.1. Panel zone variation The IPE300 beam and the HE200B column are pre-selected so that they can produce a weak column panel zone (PZ) at the intersection. Further, for evaluating the effect of PZ thickness, the column web in some specimens is reinforced with 6 and 10 mm thick doubler plates in order to provide balanced and strong PZs, respectively. The three specimens analyzed are shown in Table 1.
The radius cut shown in Fig. 2 was employed for the RBS region. The RBS connections are detailed according to the recommendations of AISC 358 [12]. The dimensions used are shown in Table 2. 2.4. Details of web slots The beam web slot is designed according to recommendations of reference [4]. Four slot lengths, as shown in Table 3, are used in analyses. In normal cases, the length of slot is 175 mm (LS2) according to the design procedure for this type of connection [4]. However, various lengths for web slots are considered to examine the effects of this parameter. These lengths are shown in Table 3 and correspond to slots extending to the end of shear plate (LS1), to the middle of beam flange reduced section (LS3) and to a distance away from the reduced section (LS4).
R
c c a Fig. 1. Subassembly model of SWRF connection.
b
Fig. 2. RBS connection detail.
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
3
3.1. Elements and meshing
Table 2 RBS region dimensions (mm). Beam
a
b
c
R
tf
tw
bf
d
IPE300
80
200
32
172.25
10.7
7.1
150
300
Table 3 Slot lengths. Specimen
LS 1
LS2
LS3
LS4
Slot length (mm)
110
175
180
460
The subassemblies are modeled using a quadrilateral four-node shell element (element SHELL181 in ANSYS). This element has material plasticity, large deflection, and large strain modeling capabilities. It has six degrees of freedom per node: translations in the x, y, z directions, and rotations about the x, y, z axes. Fig. 4 shows a typical finite element meshing used in this study. As seen in Fig. 4, a more refined mesh is employed for the regions near the RBS and web slots.
3.2. Material modeling In all specimens slot width is 3.2 mm to the end of shear plate and 6.4 mm from the end of the shear plate to the slot termination point (see Fig. 3). All slots are terminated to a hole with 21 mm diameter. The shear plate size is 10 × 110 × 190 mm in all specimens.
Nonlinear material with kinematic hardening is used for steels. The plasticity model was based on the von Mises yielding criterion and its associated flow rule. Stress–strain curve of ASTM A572 steel is shown in Fig. 5. The Young's modulus of 0.21 × 10 6 MPa and Poisson's ratio of 0.3 are used as elastic constants.
3. Finite element analysis 3.3. Loading protocol The ANSYS finite element software [13] is utilized to model the specimens for large deformation nonlinear analysis. The analyses are primarily intended to investigate the overall cyclic behavior of the SWRF subassemblies with an emphasis on the influence of PZ strength, continuity plates and slot length. The details are given below.
Each subassembly is loaded at the beam free end in displacement control such that the drift angle of the subassembly satisfies the AISC Seismic Provisions loading protocol [9] (see Table 4). All specimens are loaded up to 0.04 rad of drift angle.
Fig. 3. SBW connection details.
Fig. 4. Three-dimensional finite element model.
Fig. 5. Steel stress–strain curve.
4
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
Before loading the beam, buckling mode shapes of the model were computed in a separate buckling analysis by the software. The deformed buckled shape is then analyzed under cyclic loading to take local and lateral buckling effects into account.
Table 4 Loading protocol. 6 cycles 6 cycles 6 cycles 4 cycles 2 cycles 2 cycles 2 cycles 2 cycles
@ @ @ @ @ @ @ @
0.00375 rad 0.005 rad 0.0075 rad 0.01 rad 0.015 rad 0.02 rad 0.03 rad 0.04 rad
4. Results of analysis and discussion In order to verify the validity of the finite element model, the specimen used in the experimental study conducted by Jones et al. [8] was analyzed under cyclic displacement control loading using the
SWRF Weak Pz
SWRF Balanced Pz
SWRF Strong Pz
SWRF Weak PZ+ continuity plate
SWRF Balance PZ+ continuity plate
SWRF Strong PZ+ continuity plate
Fig. 6. Hysteretic response of SWRF connections with and without continuity plates.
LS1=110 mm
LS3=180 mm
LS4=460mm
Fig. 7. Hysteretic response of SWRF with various slot lengths and added continuity plates.
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
5
4.1. Hysteretic responses Beam moment–rotation hysteretic responses of the subassemblies resulting from the finite element analyses are shown in Fig. 6. The beam moment was measured at the column face, and the rotation was computed by dividing the total beam tip deflection by the beam length. The area under this curve indicates the amount of seismic energy dissipation by the connection yielding parts. In Fig. 6 the first row shows the moment–rotation curves for SWRF connections with different panel zone thicknesses when continuity plates are not used; while the second row shows the behavior for the same connections when the continuity plates are present. The effects of slot length in SWRF connection can be examined in Fig. 7. All these subassemblies have weak panel zones and added continuity plates. 4.2. Steel stresses
Fig. 8. von Mises stresses (Pa), SWRF with weak PZ and continuity plates at 4% drift.
ANSYS finite element software. The results agreed well with the experimental results [14]. After this verification, the subassembly models described above are analyzed and the parameters varied. The results of cyclic loading on SWRF connections are presented below and later compared with RBS and SBW connections with the same characteristics.
For the sake of brevity, the stresses at the final step of analysis for the SWRF connection with weak PZ and continuity plates are shown in Fig. 8. The stresses are in Pascal unit. The strength degradation is due to yielding of plate material in different parts of connection. This figure shows the critical (yielding) locations in steel for the SWRF connection. 4.3. Effect of PZ strength In Fig. 6, among the six analyzed specimens, it is observed that the best performing connection is the SWRF with weak panel zone and with continuity plates. It has stable hysteresis loops without strength degradation up to 0.04 rad of drift angle and the greatest energy
RBS Weak PZ
RBS Balanced PZ
RBS Strong PZ
RBS Weak PZ+ continuity plate
RBS Balanced PZ+ continuity plate
RBS Strong PZ + continuity plate
Fig. 9. Hysteretic response of RBS connections with and without continuity plates.
6
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
SBW Weak PZ
SBW Balanced PZ
SBW Strong PZ
SBW Weak PZ+ continuity plate
SBW Balanced PZ+ continuity plate
SBW Strong PZ + continuity plate
Fig. 10. Hysteretic response of SBW connections with and without continuity plates.
dissipation. Also, analysis shows that the smallest von Mises stresses occur in this case (Fig. 8). This result is important because it shows that column web without doubler plate is more suitable for SWRF connections. Balanced PZs and strong PZs experience strength degradation at 0.04 rad drift.
4.4. Effect of continuity plates In Fig. 6, comparing the top row with the bottom row, it is seen that continuity plates increase the energy dissipation in the connection. Analysis also shows that the maximum stresses of the connection are lowered slightly (Fig. 8). Moreover, continuity plates cause the location of maximum von Mises stresses to shift away from the column–beam weld location toward the beam web. Therefore, stress concentration is reduced at the critical weld location.
4.5. Effect of slot length As noted in Table 3, four specimens are examined for determining the effects of slot length on SWRF connection performance. From Fig. 7, it is seen that very long slots have poor performance. It should be noted that the slot length of 175 mm (LS2 of Table 3, standard slot length) was already shown in Fig. 6 (bottom row, left side). Results indicate (see Fig. 7) that slots with longer length than what is recommended do not improve the performance of the connection and cause greater stresses in the specimens. Specimen with 460 mm slot length (LS4) experienced instability because of large deformations. Specimen with the shortest slot length (110 mm) dissipates the greatest amount of energy.
4.6. Comparison of SWRF, RBS and SBW connections In order to compare the performance of SWRF connection with that of RBS and SBW connections, similar analyses are conducted on the latter two connections. The moment–rotation curves for the RBS and SBW connections are determined and illustrated in Figs. 9 and 10, respectively (similar to Fig. 6 for the SWRF). The stresses for the weak PZ and added continuity plates (similar to Fig. 8 for the SWRF) are shown in Figs. 11 and 12. Comparing Figs. 6, 9 and 10, shows that RBS connection has the best performance among the cases considered. As shown in Fig. 9, RBS connections experience stable hysteresis curves in all 6 variations; whereas, SBW connections with continuity plates, in balanced and strong PZ cases, and SWRF connections in balanced and strong PZ cases, deteriorate in the final cycle. Although the results are not shown for brevity, a typical preNorthridge connection is also examined as part of this research. This connection is modeled by directly attaching the beam flanges to the column without any slots in the web or cuts in the flanges. The preNorthridge connection experiences nearly double the maximum stresses of the SWRF connection at 0.04 rad of rotation. Therefore, the new connections are all definitely better than the pre-Northridge connection. However, the new connections each have their own merits and they are summarized in the next section. 5. Conclusions Previous analytical and experimental research has indicated many advantages in using RBS and SBW moment connections in steel structures subjected to earthquake loading. This paper investigated the possibility of having both connection details and their advantages in one
S. Maleki, M. Tabbakhha / Journal of Constructional Steel Research 69 (2012) 1–7
7
Fig. 12. von Mises stresses (Pa), SBW with weak PZ and continuity plates at 4% drift. Fig. 11. von Mises stresses (Pa), RBS with weak PZ and continuity plates at 4% drift.
new connection called Slotted Web–Reduced-Flange (SWRF) connection. A non-linear finite element model of a typical subassembly was used to examine the hysteretic behavior of the new connection analytically. The effects of panel zone thickness, continuity plate existence and web slot length are also investigated. The results for all three types of connections are compared with each other. Although all connections are better than the non-ductile pre-Northridge connection, but each has its own advantage. The following general conclusions can be drawn. • Results indicate that continuity plates have positive effects on all 3 types of connections and reduce the maximum von Mises stresses. In SWRF connection, continuity plates cause the location of maximum stresses to move away from the beam/column interface, where there exist stress concentrations due to welding. Therefore, using continuity plates is recommended. • In general, in SBW connections the use of doubler plates on the column is not recommended. A weak panel zone works fine in most cases. • The RBS connection overall outperforms the SBW and SWRF connections in terms of energy dissipation and stable hysteresis loops. • In the new SWRF connection, the best performance is achieved by having a weak panel zone (no doubler plate on column web) and continuity plates. In this case, the maximum von Mises stresses are located in the beam web and not flanges. • A short slot length up to the end of the shear plate is recommended for SWRF connections. This can be increased to what is recommended in SBW connections if necessary, but beyond that the connection will be adversely affected. • SWRF and SBW connections both have the advantage of reducing stresses at beam/column interface and preventing lateral torsional
buckling of the beam and reducing the vertical shear forces in the beam flanges. These recommendations are all based on numerical studies and computer simulations. To ascertain, physical testing on SWRF connections is definitely needed.
References [1] Chen SJ, Yeh CH, Chu JM. Ductile steel beam-to-column connections for seismic resistance. ASCE, J Struct Engng 1996;122(11):1292–9. [2] Moore KS, Malley JO, Engelhardt MD. Design of reduced beam section (RBS) moment connections. Steel Tips, Structural Steel Education Council, Moraga, CA; 1996. [3] Richard RM, Allen CJ, Partridge JE. Proprietary slotted beam connection design. Modern Steel Construction 1997:28–33. [4] Seismic Structural Design Associates (SSDA). Beam slot connection design manual. Laguna Niguel, California; 1998. [5] Krawinkler H. Shear in beam–column joints in seismic design of steel frames. Engineering Journal, AISC 1978:82–91 [3 rd Quarter]. [6] Popov EP. Panel zone flexibility in seismic moment joints. In: Chen WF, editor. Joint flexibility in steel frames. London: Elsevier Applied Science; 1987. [7] Tsai KC, Chen WZ. Seismic response of steel reduced beam section to weak panel zone moment connections. In: Mazzolani FM, Tremblay R, editors. Behavior of steel structures in seismic areas. Rotterdam: Balkema; 2002. [8] Jones SL, Fry GT, Engelhardt MD. Experimental evaluation of cyclically loaded reduced beam section moment connections. ASCE, J Struct Engng 2002;128(4):441–51. [9] American Institute of Steel Construction (ANSI/AISC 341–05). Seismic provisions for structural steel buildings. Chicago, IL; 2005. [10] SAC. Seismic design criteria for new moment-resisting steel frame construction. Report no. FEMA 350, SAC Joint Venture, Sacramento, CA; 2000. [11] Englehardt MD. Design of reduced beam section moment connections. Proceedings of North American Steel Construction Conference, AISC, Chicago; 1998. [12] American Institute of Steel Construction (ANSI/AISC 358–05). Prequalified connections for special and intermediate steel moment frames for seismic applications, Chicago; 2006. [13] ANSYS (Release10), ANSYS, Inc., Southpointe 275 Technology Drive, Canonsburg, PA. [14] Maleki S, Tabbakhha M. Analytical study of reduced section moment connections. Proceedings of EASEC 11 conference, Taipei, Taiwan; 2008.