You are on page 1of 178

Beam-to-Column Web Shear Tab

Connections – Minor Axis Column Loading

By
Andrea Iachetta

Department of Civil Engineering and Applied Mechanics

McGill University

Montreal, Quebec, Canada

August 2019

A thesis submitted to McGill University in partial fulfillment of the requirements of the


Degree of Master’s of Engineering (M.Eng.)

© Andrea Iachetta, 2019


1 Abstract

Single plate shear tab connections are commonly used due to their relatively simple design, ease
of fabrication and safe erection on-site. They are fillet welded to a supporting column or girder,
and then bolted to a supported beam. Extended shear tabs can either frame into the web, a
flexible support, or the flange, a rigid support, of a column. The shear tab connection is assumed
to function as a pin in the analysis of the overall structure. However, testing has shown that it
must withstand shear and moment associated with the eccentricity applied to the shear tab,
relative to the support. Current design procedures have little guidance on how to account for this
added moment to the column, which may not be a concern for shear tab connections framing
into a rigid support (column bending about major axis), but may be more critical when framing
into a column’s minor axis (which is considered as a flexible support condition).

The existing design procedures for shear tab connections were formulated from past research
studies. The 2015 CISC Handbook of Steel Construction lacks a design approach for extended
shear tabs. The 2017 AISC Steel Construction Manual has a conservative design approach only
based on unstiffened connections, despite the frequent use of stiffened extended shear tab
connections. Further research and testing is required to provide recommendations for a more
efficient and complete design procedure.

Finite element models were used to analyse the behaviour and capacity of beam-to-column web
extended shear tab connections. Five validation models were first created to simulate laboratory
tests by D’Aronco (2013) to verify the modelling assumptions. The obtained shear forces,
rotations and displacements proved to be accurate, confirming that the modelling approach
could be used in a subsequent parametric study. The parametric study included the influence of
the supporting column, single- vs. double-sided connections, the effect of the type and
connection pattern of stabilizer plates, the addition of vertical rows of bolts, the shape of the
shear plate, and the effect of the loading protocol.

ii
The results demonstrate that the column size, the addition of a vertical row of bolts and the
shape of the shear tab did not alter the connection capacity. Also, the welds connecting the
stabilizer plates to the column web are not necessary, if the plates are welded to the column
flanges. Introducing a gap between the stabilizer plates and the column flanges can benefit the
column by reducing the moments applied to it. Finally, the loading protocol greatly influences
the connection behaviour; more studies are required to further investigate this aspect.

iii
2 Résumé

Les assemblages par plaque de cisaillement sont communément utilisés en raison de leur
conception relativement simple, de leur facilité de fabrication et de leur montage sauf sur site.
Ils sont soudés sur une colonne ou une poutre de support, puis boulonnés sur une poutre
supportée. Les assemblages par plaque de cisaillement étendues peuvent soit encadrer l’âme,
(soit un support flexible) ou les semelles (un support rigide) d'une colonne. La connexion par
plaque de cisaillement est supposée agir comme une connexion articulée dans l’analyse de la
structure globale. Cependant, des essais ont montré qu’elle devait résister au cisaillement et au
moment associés à l'excentricité de la plaque de cisaillement par rapport au support. Les
procédures de conception actuelles ont peu d’indications sur la manière de considérer ce
moment ajouté à la colonne. Cela peut être correct pour les connexions par plaques de
cisaillement qui s’encadrent dans un support rigide (flexion de la colonne autour de l’axe majeur),
mais peut être plus grave lorsqu’elles s’encadrent dans l’axe mineur d’une colonne (ce qui est
considérée comme un support flexible).

Les procédures de conception existantes pour les connexions par plaques de cisaillement ont été
formulées à partir d’études de recherche antérieures. Le Manuel de Construction en Acier 2015
de l'ICCA manque d'une approche de conception pour les plaques de cisaillement étendues. Le
Manuel de Construction en Acier 2017 de l’AISC a une procédure de conception conservatrice
basée uniquement sur des connexions non-raidies, malgré l'utilisation fréquente de
renforcement pour les plaques de cisaillement étendues. Des recherches et des essais
supplémentaires sont nécessaires pour fournir des recommandations en vue d’une procédure de
conception plus efficace et complète.

Des modèles d'éléments finis ont été utilisés pour analyser le comportement et la capacité des
connexions par plaques de cisaillement étendues de poutre à l’âme de la colonne. Cinq modèles
de validation ont été créés pour simuler les essais de laboratoire exécutés par D’Aronco (2013),

iv
afin de vérifier les hypothèses de modélisation. Les forces de cisaillement, les rotations et les
déplacements obtenus se sont révélés précis, confirmant que la méthode de modélisation
pourrait être utilisée dans une étude paramétrique ultérieure. L’étude paramétrique a porté sur
l’influence de la colonne de support, les assemblages simple face et double face, l’effet du type
et de la méthode de connexion des plaques stabilisatrices, l’ajout de rangées verticales de
boulons, la forme de la plaque de cisaillement et l’effet du protocole de chargement.

Les résultats démontrent que la taille de la colonne, l'ajout d'une rangée verticale de boulons, et
la forme de la plaque de cisaillement n'ont pas modifié la capacité de la connexion. De plus, les
soudures reliant les plaques stabilisatrices à l’âme de la colonne ne sont pas nécessaires.
L'introduction d'un espace entre les plaques stabilisatrices et les semelles de la colonne peut être
bénéfique pour la colonne, en réduisant les moments qui lui sont appliqués. Enfin, le protocole
de chargement influence grandement le comportement de la connexion, car d'autres études sont
nécessaires pour approfondir les connaissances sur ce sujet.

v
3 Acknowledgments

I would like to begin by thanking my supervisor, Professor Colin Rogers, for guiding me
throughout this research project. I am grateful to have had this opportunity. Thank you for your
support, encouragement, patience and feedback.

A special thank you to Mohammad Motallebi for his continuous support and guidance
throughout my Civil Engineering studies. His mentorship and patience greatly aided my research.

I would like to acknowledge ADF Group Inc. and DPHV Structural Consultants for their financial
and technical support which allowed this research to continue. Thank you to Denis D’Aronco,
Martin Frappier and Marie-Eve Lemoine for their insight on the research topic as well as to Marco
D’Aronco for providing past data to aid in the validation of this research project.

Additional thanks goes to NSERC and to Hydro-Québec for their financial support.

A portion of the finite element computations were realized using the Graham supercomputer. I
would like to thank Compute Canada for providing access to this resource.

I would also like to thank Jorge Sayat and Dr. William Cook for their help with all software and
hardware difficulties.

I am also appreciative of the support of fellow students: Andrés González Ureña, Keith Lee, David
Pizzuto, Alina Rudman and Jason Wu.

My deepest appreciation goes out to my family and friends for supporting me throughout my
studies and for providing me with crucial advice.

vi
4 Table of Contents

1 Abstract.....................................................................................................................................ii

2 Résumé ....................................................................................................................................iv

3 Acknowledgments ...................................................................................................................vi

4 Table of Contents.................................................................................................................... vii

5 List of Figures ............................................................................................................................ x

6 List of Tables .......................................................................................................................... xvi

1 Introduction ............................................................................................................................. 1

Overview .......................................................................................................................... 1

Objectives ......................................................................................................................... 7

Methodology and Scope .................................................................................................. 7

Outline .............................................................................................................................. 8

2 Literature Review .................................................................................................................... 9

Overview .......................................................................................................................... 9

Brief Background of Extended Shear Tab Connections ................................................... 9

The Use of Stabilizer Plates ............................................................................................ 19

Finite Element Modelling of Shear Tab Connections ..................................................... 25

Current Design Procedures ............................................................................................ 31

2.5.1 CISC Handbook of Steel Construction 11th Edition (2015) ...................................... 31

2.5.2 AISC Steel Construction Manual 15th Edition (2017) .............................................. 31

3 Validation of Finite Element Models ..................................................................................... 38

Coupon Modelling .......................................................................................................... 38

vii
Developing the True Stress-Strain Curve ....................................................................... 41

Details of D’Aronco’s (2013) Laboratory Tests .............................................................. 45

FE Modelling Procedure ................................................................................................. 53

3.4.1 Part Module ............................................................................................................ 53

3.4.2 Property Module ..................................................................................................... 55

3.4.3 Assembly Module.................................................................................................... 57

3.4.4 Step Module ............................................................................................................ 57

3.4.5 Interaction Module ................................................................................................. 59

3.4.6 Load Module ........................................................................................................... 64

3.4.7 Mesh Module .......................................................................................................... 69

3.4.8 Job Module ............................................................................................................. 71

3.4.9 Imperfection Introduction ...................................................................................... 72

3.4.10 Visualization Module .............................................................................................. 74

Results of Validation Models.......................................................................................... 74

3.5.1 Results of Test #5 .................................................................................................... 74

3.5.2 Results of Test #6 .................................................................................................... 78

3.5.3 Summary of Validation Models .............................................................................. 83

4 Parametric Study ................................................................................................................... 85

Influence of the Column Size ......................................................................................... 87

4.1.1 Shallow Beam .......................................................................................................... 87

4.1.2 Deep Beam .............................................................................................................. 93

Single- vs. Double-Sided Connections ............................................................................ 97

Effect of the Type and Connection Pattern of Stabilizer Plates ................................... 105

viii
4.3.1 Shallow Beam ........................................................................................................ 106

4.3.2 Deep Beam ............................................................................................................ 112

Additional Vertical Rows of Bolts ................................................................................. 117

Effect of the Shape of the Shear Tab ........................................................................... 123

Loading Protocol .......................................................................................................... 128

Recommendations for Design ...................................................................................... 130

5 Conclusions .......................................................................................................................... 134

Summary ...................................................................................................................... 134

Recommendations for Future Work ............................................................................ 135

6 References ........................................................................................................................... 137

Appendix A: Design Calculations................................................................................................. 146

Appendix B: Results of Validation Models .................................................................................. 157

ix
5 List of Figures

Figure 1.1: Extended Shear Tab Connections with Flexible Support a) Beam-to-Girder b) Beam-
to-Column Web ............................................................................................................................... 1

Figure 1.2: Coped Beam Connection .............................................................................................. 2

Figure 1.3: Extended Shear Tab Connections with Rigid Support a) Beam-to-Column Flange b)
Double-Sided Beam-to-Column Web (column faded to facilitate viewing) ................................... 3

Figure 1.4: FBD showing Eccentric Shear Load and Resulting Moment ......................................... 4

Figure 1.5: Moment vs. Rotation Curve of Fully-Restrained (FR), and Partially-Restrained (PR)
Moment Connections and Simple Shear Connections (Faridmehr et al., 2016) ............................ 5

Figure 2.1: Tri-Linear Shear Rotation Curve (Astaneh et al., 1989) .............................................. 12

Figure 2.2: Moment Generation due to Lap Eccentricity (Thornton and Fortney, 2011) ............ 22

Figure 2.3: Extended Shear Tab Configuration Showing Optional Stabilizer Plates (AISC, 2011) 23

Figure 2.4: Stiffened Extended Shear Tab (Fortney and Thornton, 2016) ................................... 24

Figure 2.5: Conventional Shear Tab Configuration (AISC 15th, 2017) ........................................... 32

Figure 2.6: Extended Shear Tab Configurations a) Flexible Support b) Rigid Support (AISC, 2017)
....................................................................................................................................................... 33

Figure 2.7: Possible Failure Modes Considered by the AISC Manual 15th Ed. (2017) ................. 35

Figure 2.8: Girder Web Mechanism (Hertz, 2014) ........................................................................ 36

Figure 3.1: FE Modelling of Coupon: a) Part in Abaqus b) Mesh .................................................. 39

Figure 3.2: Example of Engineering vs. True Stress-Strain Curve for Mild Steel .......................... 40

Figure 3.3: Comparison of FE Model and Experimental Stress vs. Strain Curves of the Beam
Flange of Test #6 ........................................................................................................................... 42

Figure 3.4: True Stress vs. Strain Curve of the Beam Flange of Test #6 ....................................... 43

Figure 3.5: Flexible Support Laboratory Test Setup (D'Aronco, 2013) ......................................... 46
x
Figure 3.6: Dimensions of Extended Shear Tab (Modified Eccentricities).................................... 49

Figure 3.7: General Moment Diagram for Flexible Support Conditions (D'Aronco, 2013) .......... 51

Figure 3.8: General Free-Body Diagram for Flexible Support Conditions (D'Aronco, 2013) ........ 52

Figure 3.9: Exploded View of Meshed FE Model (each colour represents a different part of the
assembly) ...................................................................................................................................... 54

Figure 3.10: Slip Region for the Basic Coulomb Friction Model (SIMULIA, 2018) ........................ 59

Figure 3.11: Difference between Stiffness Scale Factor = 1 and = 0.003 for Test #8 ................... 61

Figure 3.12: Abaqus Results Showing Negligible Difference between Full Restraint and Springs at
Load Cell Locations for FE Models of Test #5 ............................................................................... 64

Figure 3.13: Typical Loading Protocol of Main and Tip Actuators ................................................ 66

Figure 3.14: a) Loading Plate in Abaqus b) Load Application at Main Actuator (D’Aronco, 2013)
....................................................................................................................................................... 67

Figure 3.15: FE Model Assembly showing Coordination System.................................................. 68

Figure 3.16: Beam Connection Part in Abaqus: a) Geometry b) With Partitions and c) Mesh .... 69

Figure 3.17: Bolt Hole Deformation a) Finer Mesh: 5 mm mesh with 40 elements around bolt
holes b) Coarser Mesh: 7 mm mesh with 24 elements around bolt holes................................... 70

Figure 3.18: Imperfection Introduction in FE Models: a) Buckling Region b) Typical Mode Shape
....................................................................................................................................................... 73

Figure 3.19: Test #5 Experimental vs. FE Model Results a) Shear Force vs. Absolute Beam
Rotation b) Shear Force vs. Shear Tab Vertical Displacement ..................................................... 75

Figure 3.20: Test #5 FE Results: a) Yield Initiation at Beam Rotation = 0.0073 rad b) Yield
Propagation at Beam Rotation = 0.0134 rad ................................................................................ 76

Figure 3.21: Deformed Shape of Shear Tab after Test #5 in a) Abaqus b) Laboratory (D'Aronco,
2013) (U3: Out-of-Plane Deformation) ......................................................................................... 77

xi
Figure 3.22: Deformed Shape of Beam after Test #5 in a) Abaqus b) Laboratory (D'Aronco, 2013)
....................................................................................................................................................... 77

Figure 3.23: Test #6 Experimental vs. FE Model Results a) Shear Force vs. Absolute Beam
Rotation b) Shear Force vs. Shear Tab Vertical Displacement ..................................................... 79

Figure 3.24: Test #6 FE Results: a) Yield Initiation at Beam Rotation = 0.0051 rad b) Yield
Propagation at Beam Rotation = 0.0192 rad ................................................................................ 80

Figure 3.25: Deformed Shape of Shear Tab after Test #6 in a) Abaqus b) Laboratory (D’Aronco,
2013) (U3: Out-of-Plane Deformation) ......................................................................................... 80

Figure 3.26: Out-of-Plane Deformation of Shear Tab Test #6 in a) Abaqus b) Laboratory


(D'Aronco, 2013) ........................................................................................................................... 81

Figure 3.27: Beam Yielding and Deformation from Test #6: a) and b) Beam near the Connection
c) and d) Close-up View ................................................................................................................ 82

Figure 3.28: Yielding of Column Flanges of Test #6 in a) Abaqus b) Laboratory (D’Aronco, 2013)
....................................................................................................................................................... 83

Figure 4.1: Shear Force vs. Absolute Rotation Curve of FE Tests #5 and #7 ................................ 87

Figure 4.2: FE Tests #5 and #7: a) Location of Recorded Measurements, b) Shear Force vs.
Vertical Displacement of Shear Tab, c) Out-of-Plane Displacement vs. Absolute Beam Rotation
....................................................................................................................................................... 88

Figure 4.3: Deformed Shape of Shear Tab from FE Models: a) Test #5 and b) Test #7 ................ 90

Figure 4.4: FE Tests #5 and #7 Shear Force vs. Change in Eccentricity ........................................ 91

Figure 4.5: FE Tests #5 and #7: a) Shear Force vs. Moment in Column Web b) Moment vs.
Absolute Beam Rotation ............................................................................................................... 92

Figure 4.6: Shear Force vs. Absolute Rotation Curve of FE Tests #6 and #9 ................................ 93

Figure 4.7: Deformed Shape of Shear Tab from FE Models: a) Test #6 and b) Test #9 ................ 94

Figure 4.8: FE Tests #6 and #9: a) Shear Force vs. Column Web Moment b) Moment vs. Absolute
Beam Rotation c) Shear Force vs. Change in Eccentricity ............................................................ 95
xii
Figure 4.9: Free Body Diagram of Symmetric Double-Sided Connection ..................................... 97

Figure 4.10: a) Double-Sided FE Model Assembly b) Shear Tab with Half of Column Width....... 98

Figure 4.11: FE Models of Single- and Double-Sided Connections: a) Shear Force vs. Absolute
Rotation b) Shear Force vs. Vertical Displacement of the Shear Tab c) Out-of-Plane
Displacement vs. Absolute Beam Rotation................................................................................... 99

Figure 4.12: FE Models of Single- and Double-Sided Connections: a) Shear Force vs. Effective
Eccentricity b) Shear Force vs. Moment in Column c) Moment vs. Absolute Beam Rotation ... 101

Figure 4.13: Single-Sided Connection a) Yield Initiation at Re-Entrant Corners b) Yield


Propagation along Interior Line of Bolts (gray represents yielded regions)............................... 103

Figure 4.14: Double-Sided Connection a) Yield Initiation at Re-Entrant Corners b) Yield


Propagation along Interior Line of Bolts (gray represents yielded regions)............................... 103

Figure 4.15: Single-Sided Connection a) Yield Initiation in Column Web at Out-of-Plane


Displacement of 0.819 mm b) Yield Propagation in Column Web at Out-of-Plane Displacement
of 1.86 mm .................................................................................................................................. 104

Figure 4.16: Double-Sided Connection a) Yield Initiation in Column Web at Beam Rotation of
0.01245 rad b) Yield Propagation (Top View) ............................................................................. 105

Figure 4.17: Stabilizer Plate Types a) Fully Welded (Traditional), b) Reduced Width Welded,
c)Reduced Width with Gaps, d) Further Reduced Width, e) Unstiffened, f) Beam Laterally
Supported ................................................................................................................................... 106

Figure 4.18: Shear Force vs. Absolute Beam Rotation of Shallow Beam Connections with
Different Stabilizer Plate Details ................................................................................................. 107

Figure 4.19: Shallow Beam Connections with Different Stabilizer Plate Details: a) Location of
Recorded Measurements, b) Shear Force vs. Vertical Displacement of Shear Tab, c) Out-of-Plane
Displacement vs. Absolute Beam Rotation................................................................................. 108

Figure 4.20: Out-of-Plane Displacement of a) BC-2-3-10-U-S, b) BC-2-3-10-P-S-b, c) BC-2-3-10-P-


S-b 1/8” gap ................................................................................................................................ 110

xiii
Figure 4.21: Shallow Beam Connections with Different Stabilizer Plate Details: a) Shear Force vs.
Effective Eccentricity, b) Shear Force vs. Moment in Column, c) Moment in Column vs. Absolute
Beam Rotation ............................................................................................................................ 111

Figure 4.22: Shear Force vs. Absolute Beam Rotation of Deep Beam Connections with Different
Stabilizer Plate Details ................................................................................................................ 113

Figure 4.23: Deep Beam Connections with Different Stabilizer Plate Details: a) Shear Force vs.
Vertical Displacement of Shear Tab b) Out-of-Plane Displacement vs. Absolute Beam Rotation
..................................................................................................................................................... 114

Figure 4.24: Column Web Punching a) BC-2-6-16-U-S-a, b) BC-2-6-16-P-S-a (flange removed for
clarity). U1 represents horizontal displacement. ....................................................................... 115

Figure 4.25: Deep Beam Connections with Different Stabilizer Plate Details: a) Shear Force vs.
Effective Eccentricity b) Shear Force vs. Moment in Column c) Moment in Column vs. Absolute
Beam Rotation ............................................................................................................................ 116

Figure 4.26: Comparison of Shear vs. Absolute Beam Rotation of Validation Tests #7 and #8 . 118

Figure 4.27: Bolt Hole Configurations: a) BC-2-3-10-F-S, b) BC-3-3-10-F-S, c) BC-3-3-13-F-S, d) BC-


4-3-13-F-S .................................................................................................................................... 119

Figure 4.28: Shear Force vs. Absolute Rotations for a change from a) 2 to 3 Vertical Rows of
Bolts and b) 3 to 4 Vertical Rows of Bolts ................................................................................... 120

Figure 4.29: Shear Force vs. Effective Eccentricity for a change from a) 2 to 3 Vertical Rows of
Bolts and b) 3 to 4 Vertical Rows of Bolts ................................................................................... 121

Figure 4.30: Additional Vertical Rows of Bolts: a) Shear Force vs. Shear Tab Vertical
Displacement, b) Shear Tab Out-of-Plane Displacement vs. Absolute Beam Rotation, c) Moment
in Column vs. Absolute Beam Rotation ...................................................................................... 122

Figure 4.31: Traditional Shear Tab Configuration with Extensions vs. Rectangular Shear Tab a)
Shallow Beam, b) Deep Beam ..................................................................................................... 124

xiv
Figure 4.32: Shear Force vs. Absolute Beam Rotation for a) Shallow Beam, b) Deep Beam;
Comparison of traditional and rectangular shear tab configurations ........................................ 125

Figure 4.33: Comparison of Traditional and Rectangular Shear Tab Configurations: a) Shear
Force vs. Vertical Displacement b) Out-of-Plane Displacement vs. Absolute Beam Rotation c)
Shear Force vs. Effective Eccentricity d) Moment vs. Absolute Beam Rotation ........................ 126

Figure 4.34: Vertical Displacement of Shear Tab a) Traditional b) Rectangular ......................... 128

Figure 4.35: Shear Force vs. Absolute Beam Rotation of FE Models of Test #5 with Various
Loading Protocols........................................................................................................................ 130

Figure 4.36: Calculated Moment Predictions Induced on the Column; a) Using a Distance
(support to interior bolt line) b) Using e Distance (support to centreline of bolt group) .......... 132

xv
6 List of Tables

Table 3.1: Shear Tab Laboratory Test Specimens (D'Aronco, 2013) ............................................ 47

Table 3.2: Predicted Connection Resistances ............................................................................... 49

Table 3.3: Plastic Region of True Stress-Strain Data of A325 Bolts .............................................. 56

Table 3.4: Predicted and Observed Maximum Shear Force and Failure Mode for the Validation
Models .......................................................................................................................................... 84

Table 4.1: Details of the Parametric Study ................................................................................... 86

xvi
1 Introduction

Overview

Connection design is a vital component to the assembly, behaviour and strength of a steel
building. Beam-to-girder and beam-to-column connections are common in most steel structures
and can be designed in various ways. The former consists of secondary beams framing into main
girders, while the latter consists of beams connected to columns; simple shear connectors are
generally used for both cases. One of the most frequently used connections for the above
mentioned cases in a gravity frame, is the single plate shear tab. This connection is relatively
simple to design, easy to fabricate and to erect on-site. It is typically fillet welded to the
supporting member and then bolted on-site to the supported beam (Figure 1.1).

a) b)

Figure 1.1: Extended Shear Tab Connections with Flexible Support a) Beam-to-Girder b) Beam-
to-Column Web

It is common to require beams and girders to align their tops of steel at the same elevation, for
example in floor construction. This requires a special connection; namely an extended shear tab
or a coped beam. Coped beams are effective for this purpose, however they are costly to
fabricate and reduce the cross-sectional area of the beam at the connection location, therefore

1
reducing its strength. For these reasons, extended shear tabs are an attractive option to
structural engineers. They are fast and easy to fabricate and can be shop welded to the
supporting girders or columns. As well, the secondary beams can simply and safely be lowered
into place and bolted on-site to the supporting member. Extended shear tabs offer this more
efficient erection procedure as compared to equivalent simple shear connections, such as double
angle or end plate connections. Figure 1.1 illustrates the extended shear tab and Figure 1.2 shows
an example of a coped beam connection.

Figure 1.2: Coped Beam Connection

Similar to extended shear tabs in beam-to-girder scenarios, these connections are also ideal for
beam-to-column connections, since they avoid the costly coping procedure and ease the
fabrication and erection of steel frames. Beam-to-column extended shear tabs can either frame
into the flange or the web of a column. They are distinguished from the conventional shear tabs
by certain dimensional limitations. To continue, the supporting member of a shear tab
connection is classified as either rigid or flexible. A support which restricts the free rotation of
the connection is considered to be rigid. This is the case for a beam-to-column flange scenario or
a double-sided beam-to-column web connection (Figure 1.3a and b, respectively). A flexible
support allows the free rotation of the connection. This occurs in single-sided beam-to-column

2
web connections (Figure 1.1b). In both cases, the connection is not capable of developing the full
moment capacity of the supported beam, and therefore carries primarily shear forces and some
moment. The difference lies in their response to this; both types respond by rotating. The
connections with rigid supports rotate about the connection, whereas those with flexible
supports rotate about the connection and the supporting column. Therefore, less rotational
demand is imposed on the shear tab connection with a flexible support since the column moves,
whereas the column remains effectively still with a rigid support. Additionally, the applied shear
force on the beam acts eccentrically to the column centroid. Both these small amounts of
moment, from rotation and from the eccentric load, cause additional moment induced on the
supporting member. The AISC Manual design procedures failed to account for these additional
moments until the 15th Ed. (2017), which states that the weak-axis bending in the column should
be accounted for in design. However only minimal detail on how to design for the eccentric beam
end reaction is provided. The free body diagram in Figure 1.4 shows the eccentricity to the
column centroid and the resulting moment.

a) b)

Figure 1.3: Extended Shear Tab Connections with Rigid Support a) Beam-to-Column Flange b)
Double-Sided Beam-to-Column Web (column faded to facilitate viewing)

3
Figure 1.4: FBD showing Eccentric Shear Load and Resulting Moment

In structural analysis, the extended shear tab is idealized as a pin connection. However, if one
considers the measured moment-rotation response of an actual shear tab connection, this
idealized behaviour is not entirely accurate. Past research has shown that it is not a perfect pin
because the shear plate, by means of its flexural stiffness, usually attracts some forces generated
by the beam end rotation. Moreover, it is presumed that this undesired bending may be amplified
in stiffened extended shear tab connections since the stabilizer plates further stiffen the
connection. See Figure 1.5, where the simple shear connection is compared to partially-
restrained and fully restrained moment connections. Initially, the shear tab will transfer a
moment to its support. However, once the yielding of the plate occurs, as the shear force is
increased, moment transfer is minimized.

4
Figure 1.5: Moment vs. Rotation Curve of Fully-Restrained (FR), and Partially-Restrained (PR)
Moment Connections and Simple Shear Connections (Faridmehr et al., 2016)

To continue, a theoretically ideal shear tab connection will transfer shear to the supporting
member without transferring any moments. This allows the connection and the supporting
member to be designed for shear loads alone. In order to achieve this ideal behaviour, shear tab
connections should have low rotational stiffness and high rotational ductility. The ductility of the
connection comes from the distortion of bolt holes in the shear tab and/or beam web and the
yielding of the plate.

To design shear tabs and their connected members, in practice, North American engineers refer
to two design procedures, which are based on past research, and therefore, are largely limited
to the scope of these completed studies. First, the design procedure provided in the Handbook
of Steel Construction 11th edition of the Canadian Institute of Steel Construction (CISC) (CISC,
2015) includes only conventional configurations. Therefore, it does not address extended shear
tabs framing into the minor axis of a column. The procedure is based on research by Astaneh et
al. (1989) and is restricted to shear tabs with one vertical row of 2 to 7 bolts, among other
dimensional limitations. This is further discussed in Chapter 2 Literature Review. Second, the 15th

5
edition of the Steel Construction Manual of the American Institute of Steel Construction (AISC)
(AISC, 2017) has independent design procedures for conventional and for extended shear tab
connections. These differ by their dimensional limitations. The extended shear tab design
procedure allows for multiple vertical rows of bolts, and has no limit on the distance from the
weld line to the bolt line. Initially, the AISC introduced a design procedure for unstiffened
extended shear tab connections in 2005, which has since been modified by Muir and Hewitt
(2009) based on the findings of full-scale laboratory testing and finite element modelling. The
AISC Manual 15th Ed. (2017) gives some recommendations for member design in stiffened
extended connections, but there remain many aspects of the design procedure to be studied.

In buildings with a moment frame along the major axis of the columns, stiffeners (continuity
plates) are required in the column webs. In these cases, it is common to use these stiffener plates
to stabilize the shear tabs in the minor axis of the column. Stabilizer plates above and below the
shear tab may also be used in column webs when no moment frame acts in the opposite
direction, to shorten the unsupported length of the shear tab plates. The stabilizer plates are
shown in yellow in Figure 1.1b. Recommendations, based on engineering judgment, have been
made with regards to the load path and connection behaviour when using stabilizer plates in
conjunction with shear tabs (Fortney and Thornton, 2016). Concerns have arisen regarding the
additional stiffness they create. Sherman and Ghorbanpoor (2002), Goodrich (2005) and
Thornton and Fortney (2011) have investigated the advantages and disadvantages of using
stabilizer plates and different weld patterns. In current practice, there are no design standards
or procedures to determine stabilizer plate thickness, their location or the location/size of welds
to connect them. In fact, the stabilizer plates have been removed from the most recent AISC
Manual 15th Ed. (2017) design procedure, whereas they were listed as optional in the previous
versions (AISC, 2011 and 2005). These previous versions do not mention the column design, but
the recent manual gives minor recommendations to the column design for a stiffened connection
to a column web (AISC Manual 15th Ed., 2017, p.10-127). Therefore, steel fabricators have the
responsibility to size these stabilizer plates by following limited guidelines. A standardized design

6
procedure would generate consistency among fabrication facilities and ensure a safe and optimal
design of the extended shear tab connections.

Despite the frequent use of extended shear tab connections with flexible supports, limited
research on the connection and member design has been conducted. Past research focusses
primarily on unstiffened extended shear tabs under gravity-induced shear force. Further research
and testing is required to provide a more efficient design procedure for stiffened extended shear
tab connections, as these are common when connecting beams to a supporting column web.

Objectives

The research contained in this thesis focusses on extended shear tab connections framing into a
flexible supporting member; specifically, beam-to-column web connections. The goal is to
provide recommendations for practicing engineers regarding the connection and member
design. The objective includes investigating the eccentricity and resulting moment induced on
the column when subjecting the beam to a shear load. As well, two recommendations by Fortney
and Thornton (2016) will be evaluated: the design using a modified eccentricity, and the use of
reduced width stabilizer plates. These objectives will be executed through finite element
modelling.

Methodology and Scope

The methodology consists of the following:

- Develop finite element (FE) models in Abaqus of previous laboratory tests on beam-to-
column (minor axis) shear tab connections performed by D’Aronco (2013).
- Perform feasibility and sensitivity studies to validate the modelling assumptions and
procedure.

7
- Build a matrix of beam-to-column web shear tab configurations which have not been
previously tested, and determine which ones will be investigated through FE modelling.
- Perform a parametric study of these configurations. The parametric models will be made
according to the validated modelling assumptions and procedure previously determined.
- Provide recommendations on the connection and member design of extended beam-to-
column web shear tab connections to be considered by practicing structural engineers.

A number of parameters can be investigated in extended shear tab connections. The variables of
interest in this research consist of the influence of the supporting column, single- and double-
sided connections, the type and connection pattern of stabilizer plates, the addition of vertical
rows of bolts, the shape of the shear tab, and the effect of the loading protocol.

Outline

This study is organized as follows:

Chapter 2 Provides a summary of past research on shear tab connections. It is focussed on


extended shear tab connections, the use of stabilizer plates, finite element
modelling techniques and the current design procedures in both the CISC
Handbook of Steel Construction (2015) and the AISC Manual of Steel Construction
(2017).

Chapter 3 Contains the validation of the finite element modelling techniques and
assumptions. It also includes an in-depth review of the laboratory testing by
D’Aronco (2013), the results of which will be used to validate the finite element
models for this research project.

Chapter 4 Describes and details the parametric study. Provides recommendations for shear
tab connection and member design.

Chapter 5 Provides a summary and concluding remarks on the findings from this study.

8
2 Literature Review

Overview

This chapter contains a brief review of research related to the study of extended shear tab
connections framing into a flexible support. The use of stabilizer plates is also discussed. Past
finite element analyses of shear tab connections are presented. The design procedures provided
in the AISC 15th edition Manual of Steel Construction (2017) and in the CISC 11th edition Handbook
of Steel Construction (2015) are explained in detail.

Brief Background of Extended Shear Tab Connections

Research on shear tab connections dates back to 1965 when new steel members were introduced
into practice which therefore needed new connection configurations (White, 1965). A history of
the development and research from this date is presented in the literature review of Behaviour
of single plate shear connections with rigid and flexible supports by Creech (2005). Please refer
to this thesis for an in-depth review of past research. The focus here will be on extended shear
tab connections and particularly those framing into flexible supporting members.

Richard et al. (1980) sought to better understand the load and stress distributions of single-plate
shear connections to determine the adequacy of the accepted design procedure at the time. This
procedure assumed that each bolt carries an equal portion of the shear load and that the
connection acts as a simple support. There was concern regarding the assumption that the
connection acts as a perfect pin, which allows free rotation to occur between the supported
beam end and the supporting member. Upon testing, it was shown that the connection does
carry a moment, which can be transferred to the beam and the supporting member. This moment
depends on several factors, including:

 number, size and configuration of bolt pattern,

9
 plate thickness and/or beam web thickness,

 beam span to depth ratio,

 distribution of the loading, and

 relative flexibility of the supporting member.

Richard et al. (1980) developed a design procedure to ensure that shear tab connections have
adequate rotational ductility. The rotational ductility is of importance since the connection is
expected to act as a simple support (pinned connection), so as to transfer the end shear reaction
of the beam to the connection without developing large moments. This is achieved by
connections having sufficient shear strength to transfer loads and ductility to rotate the beam
end without failure (Astaneh, 1989). The full-scale tests by Richard et al. (1980) demonstrated
that connection ductility came from:

 bolt deformations in shear,

 distortions around bolt holes (whether in the plate and/or beam),

 plane bending of the plate and/or beam web, and

 bolt slippage.

Richard et al. (1980) proposed a design procedure which included a limit on the shear plate
thickness to guarantee it would yield prior to other failure modes.

The first appearance of research on flexible support, single plate connections came from Stiemer
et al. (1986). They conducted research to evaluate the behaviour under shear loading of
connections with flexible supports. The full-scale tests consisted of shear tabs connecting beams
to girder webs. Concerns had been raised on the potential for deformation in the supporting
member (girder or column) due to the eccentric loading. The research led to two conclusions:
the supporting girder behaves flexibly when the ultimate capacity of the connection is reached,
and a design formula based on three-ratio interactions is recommended for connections of this

10
configuration. The three-ratio interaction is a combination of the Vmax/Vult ratio (maximum shear
force to ultimate shear force resistance), Mt/Tult ratio at the centre of the girder web (applied
torsional moment to ultimate torsional moment resistance) and the Mb/Mr ratio at the mid-span
of the girder (applied bending moment to ultimate bending moment resistance). It was
determined that each of the three ratios did not adequately predict the failure mode on their
own, rather, combining the three ratios gave the most accurate results.

To continue, Astaneh-Asl (1989) was concerned with the proportions of shear and moment
applied to shear tab connections in past laboratory tests. Astaneh-Asl recognized the importance
of applying realistic shear and moment to the connection to observe representative shear
stresses, as would be found in a building. He therefore proposed a modified beam-line curve (to
that of Richard et al. (1980)) to predict more realistic end moments and end rotations. This Tri-
Linear Shear-Rotation Curve (Figure 2.1 solid line abcd) would provide a more realistic loading
path for shear tab connection tests to follow. The other dotted lines are typical plots of end shear
vs. end rotation of simply supported, uniformly loaded beams. The beams were assumed to have
elastic perfectly plastic stress-strain properties. These curves represent the theoretical shear-
rotation demand on the simple connections. However, as the midspan moment on the beam
approaches the plastic moment, the rotation approaches infinity. In reality, structural members
experience strain hardening and have limits on the maximum deflection, therefore a connection
should never see an infinite rotation and therefore is not required to have infinite rotational
ductility (Astaneh-Asl, 1989). Consequently, the more realistic curve represented in Figure 2.1 by
the solid line abcd is suggested for simple beams.

11
Figure 2.1: Tri-Linear Shear Rotation Curve (Astaneh et al., 1989)

The first segment of the curve ab represents the elastic behaviour of the beam until the beam
starts to soften at point b. Here the moment at midspan of the beam reaches the yield moment,
My, and the rotation is equal to 0.02 rad. The next segment, bc, represents the inelastic behaviour
of the beam until the shear yielding capacity of the connection is reached at point c. Here the
moment reaches the plastic moment, Mp, and the rotation is equal to 0.03 rad. In the final
segment of cd, the beam experiences strain hardening until the ultimate moment is reached at
point d. Here, the rotation is said to be 0.1 rad, which is considered equivalent to infinite rotation.

Using this loading curve, a series of full-scale tests were performed on shear tabs consisting of a
single vertical row of bolts (Astaneh et al., 1989). The test results allowed the researchers to
develop the conventional shear tab design procedure in the AISC Manual (1993). The current
design procedure in the CISC Handbook (2015) is also based on this research. The design
procedure considers several variables: the stiffness of the supporting member, number of bolts,
material grade, plate thickness and spacing requirements (CISC, 2015). The design checks are set

12
to ensure that shear yielding of the shear tab plate occurs prior to brittle failure types. More
details on the design procedure will be discussed in Section 2.5.

Astaneh et al. (1989) also recommended equations to calculate the eccentricities of both the bolt
line and the weld line to the inflection point (where moment is zero).

n: number of bolts

a: distance between bolt line and weld line

eb: distance between the bolt line and the point of inflection

ew: distance between the weld line and the point of inflection

For the bolt line to the inflection point:

At rigid supports: 𝑒𝑏 = (𝑛 − 1) − 𝑎, (𝑖𝑛) (2-1)

At flexible supports: 𝑒𝑏 = (𝑛 − 1) − 𝑎 > 𝑎, (𝑖𝑛) (2-2)

For the weld line to the inflection point:

At rigid supports: 𝑒𝑤 = (𝑛 − 1), (𝑖𝑛) (2-3)

At flexible supports: 𝑒𝑤 = (𝑛 − 1) > 𝑎, (𝑖𝑛) (2-4)

A few years later, Astaneh et al. (1993) made a small change to the equations by adding absolute
values to better predict the eccentricities, these are shown in Equations 2-5 to 2-8.

For the bolt line to the inflection point:

At rigid supports: 𝑒𝑏 = |𝑛 − 1 − 𝑎|, (𝑖𝑛) (2-5)

At flexible supports: 𝑒𝑏 = |𝑛 − 1 − 𝑎| > 𝑎, (𝑖𝑛) (2-6)

For the weld line to the inflection point:

At rigid supports: 𝑒𝑤 = (𝑛 − 1), (𝑖𝑛) (2-7)

At flexible supports: 𝑒𝑤 = 𝑛, (𝑖𝑛) (2-8)

13
During the course of this research, Astaneh et al. (1993) also identified all possible failure modes
of the shear tab connections; below they are listed in order from ductile (more desirable) to
brittle failure modes.

 Yielding of gross area of plate

 Bearing yielding of plate and beam web bolt holes

 Fracture of edge distance of bolts

 Shear fracture of net area of plate

 Fracture of bolts

 Fracture of welds

Sherman and Ghorbanpoor (2002) published research on extended shear tabs (shear tabs with
an a distance greater than 76 mm (3”)). The goal was to determine the applicability of the
conventional shear tab design procedure suggested by Astaneh et al. (1989) to extended shear
tab connections. The variables of interest were the stiffness of the supported beam, size of the
supporting member, weld configuration and the type of bolt holes. It was found that the
conventional shear tab design procedure did not adequately predict the failure modes of
extended shear tabs, and therefore the researchers proposed a new design procedure. This
included an equation to address the distortion of the column web for unstiffened beam-to-
column web connections. However, it was decided only to permit the use of stiffened extended
shear tabs, and therefore this equation would not be used. The proposed design procedure was
based on 31 full-scale tests of both beam-to-column and beam-to-girder connections. The design
procedure in the 2005 AISC Manual was modified according to the recommended procedure by
Sherman and Ghorbanpoor.

For clarification, it should be noted that after publishing the AISC LRFD 3rd edition Manual (AISC,
2001), the LRFD and ASD Manuals were combined. Therefore, the next version published was the
13th edition AISC Manual (2005).

14
Muir and Thornton (2011) conducted research on both conventional and extended shear tabs in
order to address the bolt shear resistance values in the 13 th edition AISC Manual (2005). Their
results were used to modify the design procedure, which was subsequently included in the 14th
edition AISC Manual (2011). Research showed that the design procedure in the 13 th edition AISC
Manual (2005) relied on bolt shear values from Table J3.2 from the AISC 360-05 Specification
(2005), which are listed as 20% lower than theoretical values (Muir and Thornton, 2011). This
reduction accounts for the uneven force distribution among the bolts and permits a secondary
factor of safety to all bolted connections. By relying on this reduction in bolt shear, the shear tab
design procedure allowed engineers to neglect the eccentricity in the bolt group. However, in the
AISC 360-10 Specification (2010), the bolt shear strength values were increased, and therefore
the shear tab design procedure needed to be revised.

When a connection is made to a rigid support, the centre of rotation is approximately the same
as the centre of the bolt group (Muir and Thornton, 2011). For flexible supports, the centre of
rotation is less important since the end rotation can be accommodated through movement of
the flexible support. In other words, the rotational demand on the shear tab connection is lower
when connected to a flexible supporting member. To continue, beam end rotations are due to a
combination of plate flexural yielding, bolt deformation, bolt plowing and support rotation. The
latter mechanism only exists for connections with a flexible support. Conventional shear tabs
have a relatively short distance between the bolts and the welds, therefore allowing minimal
area for plate yielding to occur, and consequently allowing next to no beam end rotation. Bolt
deformation and bolt plowing are dependent on the movement of the bolts with respect to the
plate (Muir and Thornton, 2011).

Furthermore, Marosi (2011) performed full-scale tests on 16 beam-to-column specimens with


single and double vertical rows of bolts. The connections were designed based on the AISC
Manual of Steel Construction 13th edition (2005) and the CISC Handbook of Steel Construction
10th edition (2010) to compare the resistances calculated by both standards. He determined that

15
the CISC procedure yielded highly conservative values. This can be attributed the fact that Table
3-41: Shear Tab Beam Connections is based on the outdated bolt resistance factor (φb = 0.67)
from the CSA S16-01 Standard (2001). This value has since changed to φb = 0.8 in the CSA S16-09
Standard (2009). In addition, the design procedure in the CISC Handbook (2010) is only valid for
shear tabs with a single vertical row of up to 7 bolts. Marosi’s tests did not meet this requirement.
Based on the test results, Marosi proposed a modified design procedure based on that in the
AISC Manual (2005) for shear tabs of single and double vertical rows of bolts.

D’Aronco (2013) performed tests on 10 full-scale beam-to-column shear tab connections with
two and three vertical rows of bolts. Of the ten tests, two were weld retrofits, two were bolted
connections to a rigid support and the remaining six were bolted connections to a flexible
support. These final six specimens were shear tabs that connected the beam to the web of the
column. The purpose of this research was to evaluate the accuracy of the design procedure
recommended by Marosi (2011). It was found that the procedure was adequate for predicting
connection resistances for both rigid and flexible supports, as well as for multiple vertical rows
of bolts (D’Aronco, 2013). An in-depth review of D’Aronco’s research is presented in Section 3.3.

Hertz (2014) performed 12 full-scale tests on beam-to-girder and beam-to-column extended


shear tabs. Rotational restraint of the beam was not provided in order to simulate a steel grating
floor, which is assumed to provide no rotational stiffness. This floor setup is typical in industrial
buildings and was assumed for the majority of past research. The tested shear tab thickness of
9.5 mm (⅜”) was used to ensure yielding failure in the plate, and therefore to evaluate the
accuracy of the AISC Manual’s (2010) extended shear tab design procedure. Hertz suggests to
calculate the weld thickness using an equation which considers the probable material properties
rather than the ⅝ tp, which is found in the AISC Manual. Hertz’s recommendation is supported by
two tests, which resulted in weld tearing due to higher yield stresses of the shear plate (Hertz,
2014).

16
The addition of a concrete slab on top of the beam can have varied effects on the results of the
bare steel connection. In a building, there would typically be a concrete slab above the steel
beam, acting as a floor. However, since the slab is not directly connected to the shear tab, rather
it is connected to the beam, the reliability of the stiffening and strengthening of the slab is
unknown. It may only transfer moment but not shear. The added weight and stiffness may
prevent lateral torsional buckling of the beam and potentially out-of-plane deformations of the
shear plate. This added stiffness can also decrease the rotational demand on the shear tab
(Motallebi, 2018).

An experimental-numerical study was conducted to address various shortcomings of the current


design procedures for extended shear tab connections in Canada and USA (Motallebi, 2018). The
laboratory testing provided data to validate numerical finite element models. One phase of the
research comprised single- and double-sided shear tab configurations. The second phase of
testing extended to connections subjected to gravity-induced shear forces along with axial forces.
The third phase consisted of an expanded range of applied axial forces and connection
configurations. Information gathered from these phases led to a deeper understanding of the
inelastic response of extended shear tabs. Analysis of the validated FE models enabled Motallebi
to study the load path and behaviour of each connection. Local imperfections were included in
the FE models to account for potential local instabilities.

Based on the observations from the three phases of testing, a parametric FE study was then
performed by Motallebi (2018) to investigate the influence of various parameters. Some of the
conclusions drawn from this research are discussed here:

 The governing failure mode for stiffened full-depth configurations (both for beam-to-
girder and beam-to-column connections) is plate buckling.
 The governing failure modes for stiffened partial-depth configurations are shear plate
yielding and twisting.

17
 The inflection point was observed close to the centre of the bolt group, therefore much
smaller than assumed in the AISC design method (2017). (Note here that the CISC design
procedure (2015) is not mentioned since extended shear tab configurations fall outside
the range of connections applicable to the Canadian design procedure). The AISC (2017)
recommends designing for an eccentricity from the weld line to the centre of the bolt
group. Consequently, bolt shear fracture is not a governing failure mode since the group
is designed for a much higher moment than it will actually experience. On the other
hand, this difference in eccentricity may not always be conservative since a flexural
demand on the supporting member is induced.
 The stiffened portion of full-depth extended beam-to-girder shear tabs was subjected to
vertical axial and horizontal shear forces simultaneously. This is not taken into account
in the current design procedures. The demands on the connections depend on the minor
axis stiffness of the supporting member and on the eccentricity.
 Yielding initiated at the re-entrant corner of both the single- and double-sided full-depth
extended beam-to-girder connections.
o The propagation of yielding in the single-sided configurations tended toward the
stiffened portions of the shear tab. Out-of-plane deformations of the girder web
increased as the shear force increased.
o The propagation of yielding in the double-sided configurations spread to the
interior bolt line until the entire net section yielded. Then, the connection
experienced a significant reduction in stiffness and the yielding spread to the
stiffened portions of the shear plate. In this case, out-of-plane deformations of
the girder web are restricted by the connections on either side of the web.
 Motallebi (2018) suggests determining the bolt shear strength based on a bolt group
eccentricity of between the location of inflection point and the centre of the bolt group.
Connection eccentricity is a function of the bolt pattern depth, therefore this assumption
needs further justification.

18
 The local buckling failure mode of the shear plate should be considered. The use of shear
plates that satisfy the CSA S16 compactness ratio for stiffeners (200 / √Fy) results in a
stable shear tab connection behaviour.
 The suggestion by Fortney and Thornton (2016) to consider the inflection point at the
toe of the stiffener predicted the ultimate connection strength more accurately
compared to the AISC Manual 15th Ed. (2017) method. However, the suggestion
overestimates the resistance of the connection to gross cross sectional yielding. This is
due to the large rotations observed in testing which are beyond acceptable limits for
serviceability of the beam.

The Use of Stabilizer Plates

Note: Sherman and Ghorbanpoor (2002) refer to stabilizer plates as stiffeners or continuity plates
(both are used here). Stabilizer plates may also be called continuity plates because they are
typically installed as continuity plates for an orthogonal moment frame and simultaneously used
as stabilizer plates for the shear tab connection.

Sherman and Ghorbanpoor (2002) conducted full-scale tests on extended shear tabs. Both
stiffened and unstiffened specimens were evaluated. Various weld configurations were tested
including one or a combination of: vertical weld plate to web, horizontal weld plate to top
stiffener, horizontal weld plate to bottom stiffener and horizontal weld stiffener to inside column
flanges. It was concluded that unstiffened tabs could be used for connections with a very low
beam reaction. Otherwise, stiffeners should be used for the connection to have enough strength
and ductility, as well as to avoid web mechanism in the column and flexural buckling. Web
mechanism is a failure mode which has substantial distortions in the column web. Additionally,
using only one stiffener at the top of the plate yields similar resistances to using top and bottom
plates, but experiences large displacements due to column web mechanism at the unstiffened
base of the shear tab (Sherman and Ghorbanpoor, 2002). Welding the stiffening plates to the

19
column web had no significant impact on the connection capacity. As well, increasing the
thickness of the stiffener plates had minimal effect on the behaviour of the connection. The
results showed that an additional limit state, the yield line mechanism, for unstiffened shear tabs
was necessary to account for the web mechanism failure. Motallebi (2018) determined that large
axial and shear forces accumulate at the location where the shear tab extends beyond the flanges
of the supporting member. This is the cause for the out-of-plane deformation of the girder or
column web resulting in a girder or column web mechanism.

Goodrich (2005) provides a design procedure for the continuity plates: design the continuity
plates for flexure according to Section F of the AISC Specification 360-05 (2005), for tensile
strength according to Section D of AISC Specification 360-05 (2005) and for compressive strength
according to Section E of the AISC Specification 360-05 (2005). Commonly, the size of the
continuity plates is taken equal to the thickness of the flanges of the strong-axis beams; these
are typically heavy members since they make up part of the rigid frame, and therefore the
continuity plates are normally at least ¾”. However there is a minimum thickness requirement
the continuity plates should meet depending on the load. It can be shown that this limit is 1.5
times the thickness of the shear tab that is designed using Table 10-9 of the AISC Manual (2005).
As well, the thickness of the continuity plate to supporting member welds is equal to 75% of the
thickness of the continuity plates and has a minimum value of 1.5 times the thickness of the shear
tab-to-continuity plate welds. The length and width of the continuity plates are determined
based on the geometry of the column to which they are welded.

As previously mentioned, Sherman and Ghorbanpoor (2002) recommend only using stiffened
shear tabs to avoid excessive lateral twist. Thornton and Fortney (2011) suggest that this
conclusion may be invalid. The authors analysed the past results and suggested that the excessive
distortion only occurred at loads much beyond the capacity of the connection as calculated with
the AISC design procedure. For this reason, Thornton and Fortney recommend using stiffened
shear tabs only when one or both of two conditions are not satisfied:

20
1) the lateral-torsional buckling strength of the plate is insufficient and/or

2) the torsional moment strength due to lap eccentricity is insufficient.

In order to check the first condition, the shear tab connection is represented as a double-coped
beam end and therefore uses the same lateral-torsional stability check. This resemblance has
been proven through research by Cheng et al. (1984).

The lateral torsional buckling strength of the shear plate is (AISC 14 th Eq. 10-6):

1500𝜋𝐿𝑡 3
𝜙𝑅𝑛 = 𝜙 (2-9)
𝑎2

Where φ = 0.9, t is the thickness of the plate, a is the distance from the support to the interior
line of bolts and L is the depth of the plate. The buckling strength is compared to the shear force
applied to the connection to determine if it is adequate without stabilizer plates.

The second check to determine the need for stabilizer plates accounts for the torsional moment
strength due to lap eccentricity. The lap eccentricity is the offset between the beam and the shear
plate along their longitudinal axis:
𝑡𝑝 +𝑡𝑤
𝑒= (2-10)
2

This lap eccentricity produces a torsional moment, which is resisted by the torsional strength of
the shear tab as well as the local torsional strength of the beam due to the floor slab or roof deck.
See Figure 2.2.

21
Figure 2.2: Moment Generation due to Lap Eccentricity (Thornton and Fortney, 2011)

The available torsional moment strength due to lap eccentricity is as shown in Eq. 2-11 (AISC 14th
Ed., 2011, Eq. 10-7):
2
𝐿𝑡𝑝 2 (𝑡 +𝑡 )𝑏
2𝑅𝑢
𝑅 𝑤 𝑝 𝑓
𝜙𝑀𝑡 = [𝜙𝑣 (0.6𝐹𝑦𝑝 ) − 𝐿𝑡𝑢 ] ( )+ 2 (2-11)
𝑝 2 𝜙𝑏 𝐹𝑦𝑏 𝐿𝑠 𝑡𝑤

Where Ls is the span length of the beam, Ru is the required strength, φb = 0.9 and φv = 1.0. This
accounts for the lateral shear strength of the plate as well as the lateral bending strength of the
beam in the connection region. This second portion should only be included when there is a slab
present, since it occurs as bending of the beam flange comes into contact with the slab (AISC 14th
Ed., 2011). The result is compared to the required torsional moment strength calculated as the
shear force times the lap eccentricity from Equation 2-10.

The 14th edition of the AISC Manual (2011) provides a sketch of a flexible support, extended shear
tab connection which has “stabilizer plates, if required”. It is assumed that these stabilizer plates
provide a bracing force to the connection (Figure 2.3). Upon further research, it was suggested
by Fortney and Thornton (2016) to avoid the use of stabilizer plates unless absolutely necessary.
The two checks explained above to determine the need for stabilizer plates are still valid in the
15th edition of the AISC Manual (2017), however the sketch no longer includes these optional
plates. The reason for removal was the misinterpretation of engineers assuming the plates were

22
there to provide stability to the beam end. To continue, if either of these limits are not met, best
practice would suggest altering the shear plate thickness or connection geometry to try making
the connection adequate without the use of stabilizer plates. However if adequate geometric
changes are not possible, stabilizer plates may be used so long as they are designed
appropriately. The intent for the stabilizer plates is only to laterally stabilize the shear tab and
not to alter the load path of the applied shear force. Therefore avoiding them would be ideal to
prevent additional costs and added stiffness to the connection.

Figure 2.3: Extended Shear Tab Configuration Showing Optional Stabilizer Plates (AISC, 2011)

In the case where stabilizer plates are used, to minimize their effects on the behaviour of the
connection, they should provide negligible stiffness to the connection (Fortney and Thornton,
2016). This would result in minimally inducing rotational demands on the supporting member,
and consequently the AISC extended shear tab design procedure would remain valid since no
changes are made to the load path. If additional rotational demands are induced, these tend not
to be accounted for in frame analysis. The challenge in achieving this desired result is that
stiffness attracts loads. Consequently, experimental and numerical studies are required to justify
these claims.

23
Therefore, two design methods were analysed. 1) The bolt group is designed for a moment M =
R a as previously done. Here a is the distance from the face of the column web to the first line of
bolts (Figure 2.4) and R is the reaction force. In this case, through static analysis, the force in the
stabilizer plates, Vs is equal to zero (Fortney and Thornton, 2016). This reflects a case where no
load is transferred between the shear plate and the stabilizer plate and therefore no additional
moment is transferred to the column (or girder) web. However, engineering judgement
presumes that some load is transferred, regardless of what the static analysis illustrates.

2) The bolt group is designed for a moment M = R g (reduced). Here g is the distance from the
toe of the stabilizer plate to the first line of bolts (Figure 2.4). In this case, Vs is no longer equal
to zero. Therefore, a force is transferred and a moment generated in the minor axis of the column
(or girder). This rotational demand imparted on the column is equal to Vs*a’*L, where L is the
vertical separation between stabilizer plates (Fortney and Thornton, 2016). As well, the moment
demand on the shear plate is reduced since the design considers a smaller eccentricity. This
second design method is similar to the design of rigid extended shear tabs which frame into the
flange of a column rather than into the web. Finally, Fortney and Thornton (2016) suggest
designing the stabilizer plates for moment, shear and shear rupture.

Figure 2.4: Stiffened Extended Shear Tab (Fortney and Thornton, 2016)

24
FEM analysis was performed on extended beam-to-column flange connections to examine their
failure modes, in particular twisting of the plate, and consequently to study the need for stabilizer
plates (Suleiman et al., 2017). This research simulated connections with a beam having full lateral
support from a roof diaphragm or floor slab. Large lateral displacements of the shear tab were
recorded at ultimate load, but the lateral displacement remained small under service level loads.
The authors suggest a more economical solution to reduce twisting would be to increase the
thickness of the shear plate prior to adding stabilizer plates (Suleiman et al, 2017).

Finite Element Modelling of Shear Tab Connections

In this section, a review of past finite element analyses on shear tab connections is presented.
Both finite element modelling assumptions and results will be discussed. The details from these
past models were taken into consideration when modelling for this research project.

Ashakul (2004) performed forty-five finite element analyses on beam-to-column flange shear tab
connections. The research goal was to evaluate the shear tab design procedure in the 2001 AISC
LRFD 3rd Edition Manual (AISC, 2001). In order to validate the numerical models, 8 of the 45
models simulated laboratory experiments by Astaneh et al. (1989) and Sarkar and Wallace (1992).
Four more models were created to verify certain assumptions used in the models: effect of beam
size/length and effect of loading and bolt strength on the connection behaviour. The finite
element models included the effects of geometric nonlinearities. All other relevant modelling
assumptions used in this research project are explained here.

The material properties of the welds and bolts were assumed to be elastic perfectly plastic. The
plate and beam had full stress-strain relationships including strain hardening. Poisson’s ratio and
Young’s modulus were taken as 0.3 and 29000 ksi, respectively, for all tests.

Ashakul (2004) used the symmetry technique for his models. Half of the beam span was modelled
with a shear release boundary condition to simulate the mid-length of the beam. Ashakul

25
modeled areas expected to see high stresses with C3D20R elements. These are solid continuum
20 node hexahedral second order elements. Reduced integration was used to capture the stress
in critical areas. In other areas of the model where stress was not expected to be significant,
similar brick elements with less nodes were used (C3D8I). Such incompatible elements (I) add
additional internal degrees of freedom to the first-order elements to improve their bending
behaviour. Finally, in areas of more complex geometry, C3D15 elements were used. This is a solid
15 node triangular prism second order element and was used in areas such as the innermost
elements of bolts. Two mesh sizes, 0.25” and 0.125”, were applied to the shear plate to compare
their results. The shear rotation curve of the 0.125” mesh matched the Von Mises’ yield criterion
and therefore was considered as an adequate mesh size.

Ashakul (2004) did not include the column in his models, rather the backside of the welds were
restrained in a way to simulate this connection to the column. Tie constraints were used between
the welds and the shear plate. For the contact between the bolts and bolt holes, small sliding was
used. The interactions between the bolts and the bolt holes were simulated using the GAP
element in Abaqus, since there is an initial gap around the bolt shaft for the 1/16” oversize of the
bolt holes. Gap elements allow for contact and separation between two nodes. The contact
property used was “Hard” contact in the normal direction.

Boundary conditions in the model included fixing the welds to simulate the connection to the
column and restraints along the entire beam length to prevent lateral torsional buckling.

Observation of shear vs. rotation of the plate at the bolt line, moment vs. beam end rotation at
the weld line, shear vs. beam rotation and shear vs. distance to point of inflection from the weld
line curves led to the following conclusions:

 the lack of effect on the bolt group by changing the a distance,


 plates of a higher grade of steel impeded the bolt group capacity,

26
 the bolt group moves altogether as opposed to moving about the instantaneous centre
of rotation and,
 the position of the connection with respect to the beam’s neutral axis (centre of rotation)
impacted the bolt group strength (Ashakul, 2004).

Rahman et al. (2007) conducted an FE investigation on shear tab connections with 3 and 5 bolts
placed in a single vertical row to validate previous experiments by Sherman and Ghorbanpoor
(2002). Material properties were measured and therefore included strain hardening effects. By
using the program ANSYS, fully-integrated cubic elements (C3D8) were used for the column and
shear tab and C3D10 elements for the bolts. A pretension force was incrementally applied until
the bolt came into contact with the shear tab and the beam. Consequently, stresses were
generated and then converted into strain, causing the bolts to become locked. Once the
pretension step is complete, external loads can be applied.

To continue, Rahman et al. (2007) suggested a coefficient of friction (steel to steel) of 0.3. The
researchers analyzed several models with different coefficients of friction ranging between 0.2
and 0.6, as suggested by the AISC LRFD Manual (1993) and the optimal was found to be 0.3. The
detailed finite element models proved to correlate well with the laboratory tests.

Marshall (2011) simulated laboratory tests performed by Marosi (2011) and by D’Aronco (2013)
using Abaqus software. In order to accurately input material properties of the beam flanges,
webs and shear plates, finite element models were first made for the coupon tests. An extensive
study into the relationship between the true stress-strain and the engineering stress-strain
curves was performed. Equations will be discussed in Section 3.2 to convert the engineering data
into true values to be inputted into Abaqus. To do this, an equation suggested by Holzer et al.
(1975) was used. The same procedure was followed by other researchers such as Motallebi
(2018), Goldstein Apt (2015), Hertz (2014), etc.

27
The FE models of the coupons were based on average measured dimensions since there were
more than one coupon per specimen. Using the coupon data, the material properties were
assumed to be isotropic. It is important to note that in reality the material properties change with
direction and location. This particularly affects steel which has undergone significant work
hardening; this may be the case for the plates used to fabricate the shear tabs, which may
undergo cold-rolling to obtain the desired thickness.

Coupon testing was not performed for all members of Marosi’s (2011) and D’Aronco’s (2013)
assemblies; only for beams and shear plates. Therefore, Marshall used elastic properties for the
columns.

Interactions and ties were specified at all contact locations. Surface-to-surface tie constraints and
contact interactions were used with finite sliding. The coefficient of friction used in the model
was 0.3. These assumptions are identical to past FEM studies described in the literature, e.g.
Ashakul (2004), Rahman et al., (2007), Hijaj and Mahamid (2016) and Motallebi (2018).

In the Load Module of Abaqus, pressure loading proved to yield more accurate results as
compared to displacement loading. However, this technique can result in modeling instabilities
once large deformations occur. Therefore, Marshall used pressure loading up until approximately
83% of the run time and then switched to displacement loading. The boundary conditions on the
column were a fixed base and a displacement of the column corresponding to the test data as
recorded with an LVDT. These horizontal displacements ranged from 1-5 mm.

For the mesh, several element types were evaluated for their efficiency. C3D20 elements were
used in the shear tab, beam near connection, loading plate for main actuator and column. C3D8I
elements were used in the beam portion furthest from the connection and in the loading plate
for the tip actuator. C3D15 elements were used for the portions of the model which are not

28
rectangular (welds and k areas of W-shapes). Finally, C3D6 elements were used for the curved
portion of the beam web furthest from the column. It was noted that the C3D8 first order
elements yielded better results than the second order elements since the analysis is not governed
by bending. First order hexahedral elements can experience shear locking, which is the
development of “parasitic shears” which do not exist in reality. This phenomenon of shear locking
is mostly encountered in analyses with significant bending stresses. Incompatible modes can help
avoid “parasitic shears” and this was the idea behind using the C3D8I elements. For this reason,
the C3D8 elements proved to be adequate for the shear tab models. To better understand the
behaviour of the element types, an element with a reduced number of integration points tends
to be softer (more flexible) and is prone to hourglassing. Hourglassing is when there is significant
distortion at the centre of an element due to large deformations; this can cause problems in
calculating the stiffness. On the contrary, increasing the number of integration points makes the
element stiffer, subjecting it to the potential for shear locking (SIMULIA User Assistance, 2018).

Marshall (2011) assumed the maximum recorded shear force was the shear plateau value rather
than the actual maximum in Abaqus. It was noted by the author that future versions of Abaqus
have the potential to model fracture by including the option for a ductile damage model. Marshall
suggested this inclusion would provide results even closer to the experimental observations.

Abou-zidan and Liu (2015) performed a numerical study on unstiffened extended shear tabs to
investigate beam-to-column web connection behaviour. Each model was subject to a gravity-
induced shear load. Parameters were altered to investigate their effects on the connection
capacity and distribution of stresses. These parameters included the number of horizontal and
vertical rows of bolts, thickness of the shear plate, the a distance, lateral restraint of the beam
and the web slenderness of the supporting column. The connections were designed and detailed
to fail in bolt shear rather than plate yielding. It should be noted that elastic perfectly plastic
material properties were defined for the column, beam and shear tab.

29
The results show that as the shear load increases but remains small, the inflection point moves
toward the bolt line until it reaches a relatively constant value. Once the beam web begins to
yield, the inflection point moves back toward the support. This occurs because of the decrease
in capacity of the beam, therefore redistributing the forces to the supporting column.

Furthermore, the results showed that in general, having more bolts in a single row leads to higher
eccentricities and therefore higher moment. However, it also means higher capacity. In the case
of two vertical rows of bolts, larger rotations are reached. They also showed that the plate
thickness had minimal effect on the bolt eccentricity (distance between the inflection point and
the centre of the bolt group). The connection capacity decreases approximately linearly with an
increase in a distance. This decrease is due to large rotations and deformation of the shear tab.
It was found that providing lateral restraint at the connection location was as effective as
providing lateral restraint along the entire beam length. Finally, increasing the web slenderness
of the column resulted in increasing the eccentricity of the bolt group and consequently reducing
the connection shear capacity. This was particularly noticed at slenderness ratios above 30 (h/w
> 30). Overall, Abou-zidan and Liu (2015) concluded that the AISC design method (2011) provides
better accuracy in predicting the strength of connections with a higher number of bolts in a single
row. Therefore, they suggest a modification to the eccentricity of connections with fewer than 6
bolts in a single row.

Suleiman et al. (2017) performed FEM analysis using Abaqus 2010 software on extended beam-
to-column flange connections to examine their failure modes. In contrast to the modelling
approach used by previous researchers, Suleiman et al. applied the bolt pretension using a
temperature change on the bolt shank. Various monitoring techniques were used to observe the
behaviour of the connection to determine the controlling failure mode. These included
monitoring stresses in the bolt shank to detect bolt shear failure, principle stresses around bolt
lines to detect net shear and bending interaction, stresses around bolt holes to detect bolt
bearing and finally angle of twist of the shear plate.

30
Current Design Procedures

In this section, a presentation is provided of the current design procedures for shear tab
connections in Canada and USA.

2.5.1 CISC Handbook of Steel Construction 11th Edition (2015)

The Canadian Institute of Steel Construction (CISC, 2015) contains limited information on the
design of shear tab connections. The information is based on research by Astaneh et al. (1989)
and assembled into tabular form. The design procedure addresses only conventional shear tab
connections; these include shear tabs with 2-7 bolts in a single vertical row. The procedure
excludes any configuration with multiple vertical rows of bolts. There are also dimensional
limitations for the bolts and bolt holes, as well as specified steel and electrode grades. Essentially,
the CISC provides a table giving plate thickness and weld size required to achieve a given factored
load resistance based on the number of bolts and whether the support is flexible or rigid.
Therefore, many questions remain unanswered for designing shear tabs according to Canadian
standards. Consequently, Canadian engineers tend to look to the American Institute of Steel
Construction (AISC, 2017) for further guidance.

2.5.2 AISC Steel Construction Manual 15th Edition (2017)

The American Institute of Steel Construction (AISC) provides a more detailed design procedure
as compared to the CISC. It differentiates between conventional and extended shear tab
connections, each with a unique design method. The conventional configuration resembles that
in the CISC’s Handbook of Steel Construction and is described as having:

 Single vertical row of bolts (2-12 bolts),

 Distance a must be less than or equal to 3 ½” (a: horizontal distance from the weld line
to the bolt line closest to the support) (Figure 2.5),

31
 Standard holes or short-slotted holes transverse to the direction of the supported
member reaction are permitted,

 Vertical edge distance must satisfy AISC Specification 360-16 (2016) Table J3.4
requirements. Horizontal edge distance should be greater than or equal to 2d for both
the plate and the beam web, where d is the bolt diameter,

 Either the plate thickness or the beam thickness must satisfy the maximum requirement
given in Table 10-9 (AISC, 2017). This requirement follows from previous editions of the
AISC. The thickness is limited to one-half the diameter of the bolts + 1/16” (Muir and
Thornton, 2011). This allows bolt plowing to occur (bolts can cause elongation of the bolt
holes due to local buckling of the beam web or plate).

Figure 2.5: Conventional Shear Tab Configuration (AISC 15th, 2017)

The design checks for the conventional configuration are listed below:

1. Bolt shear is checked in accordance with AISC Specification 360-16 (2016) Section J3.6
assuming the eccentricity, e, shown in Table 10-9 (AISC, 2017). The bolt shear capacity is
computed based on the reaction force, R, and the corresponding bending moment, M =
R e (AISC, 2017). (e: horizontal distance from the weld line to the centroid of the bolt
group).
2. Plate bearing and tearout are checked in accordance with AISC Specification 360-16
(2016) Section J3.10 assuming the reaction is applied concentrically.

32
3. Plate buckling will not control for the conventional configuration. The reason is the small
distance between the bolts and welds therefore leaving only a minor area for plate
yielding (Muir and Thornton, 2011).

If the dimensional limitations of the conventional configuration are not satisfied, then the shear
tab is considered an extended shear tab (Figure 2.6). Listed here are the dimensional limitations
for extended shear tabs:

 Number of bolts n is not limited,

 Distance a is not limited,

 The use of holes must satisfy AISC Specification 360-16 (2016) Section J3.2 requirements,

 Horizontal and vertical edge distances must satisfy AISC Specification 360-16 (2016) Table
J3.4 requirements.

a) b)

Figure 2.6: Extended Shear Tab Configurations a) Flexible Support b) Rigid Support (AISC, 2017)

33
The design procedure for single-plate extended shear tabs was developed by Muir and Hewitt
(2009). Their suggested equations were based on past research. It was noted that the moment
distribution throughout the connection was not well understood. Therefore, Muir and Hewitt
developed an equation for the maximum shear plate thickness. The purpose of this equation is
to limit the plate strength, to guarantee it acts as a fuse since it is the most ductile element in the
connection (2009). Furthermore, the limit ensures forces in the connection will redistribute,
allowing the connection to have enough rotational ductility. In other words, the internal and
external forces are in equilibrium while the external forces are below the critical load causing
failure. This confirms that the limit states are satisfied, therefore enough rotational ductility is
provided to redistribute the force (Muir and Hewitt, 2009).

Extended configuration shear tab connections follow six design checks. The design checks apply
to unstiffened shear tab connections. Due to a lack of any other methodology, practicing
engineers use these guidelines for stiffened connections as well. Some information was provided
in the literature by Fortney and Thornton (2016) on the use of stabilizer plates (which create a
stiffened connection), however these recommendations have yet to be validated through
experimental or numerical studies. The design checks are listed below and Figure 2.7 illustrates
the possible failure modes.

1. Bolt group must satisfy bolt shear, bearing and tearout with the eccentricity, e, defined
as the distance from the support to the centroid of the bolt group or alternative
eccentricities when justified.

2. Maximum shear plate thickness such that the plate moment strength does not exceed
the moment strength of the bolt group in shear. This limit is placed to ensure that plowing
of the bolts occurs rather than bolt shear failure since a ductile failure mode is desirable
over a more brittle failure mode.

3. The plate is checked for the limit states of shear yielding, shear rupture, block shear
rupture and flexural rupture. The beam web is checked against the same limit states, as
applicable.

34
4. The Von Mises stress criteria is checked to evaluate the limit states of shear yielding, shear
buckling and yielding due to flexure of the plate.
5. The plate is checked for the limit state of buckling using the double-coped beam
procedure.
6. Lateral supports for the beam must be provided at the connection location.

Beyond the design checks, there are further suggestions. The weld size is recommended to be ⅝
times the thickness of the plate. This limit sets the minimum weld size in order to ensure that
shear yielding of the plate occurs prior to failing the welds. Shear plate and beam web materials
are designed to be either Fy = 250 MPa or 345 MPa (36 ksi or 50 ksi). As well, if stabilizer plates
are used in a beam-to-column web connection, the weak axis moment in the column should be
accounted for in design.

Figure 2.7: Possible Failure Modes Considered by the AISC Manual 15th Ed. (2017)

Motallebi (2018) found the governing failure mode for partial- and full-depth extended beam-to-
girder shear tabs to be girder web mechanism (or punching shear failure) and its interaction with
shear plate buckling of the shear tab. This mechanism was also observed by Hertz (2014) and is
illustrated in Figure 2.8. This failure mode can be explained by the development of horizontal

35
compressive stresses at the base of the shear tab as well as vertical tensile stresses below the
top flange of the supporting girder. Consequently, rotation of the top flange of the girder occurs
(Hertz, 2014). This failure mode is not considered in the CISC or the AISC design procedures. In
addition, the above mentioned design method was developed for extended beam-to-column
shear tabs and not for extended beam-to-girder shear tabs (AISC, 2017).

Figure 2.8: Girder Web Mechanism (Hertz, 2014)

It is clear there is a need for a standardized design procedure to account for the girder web
mechanism in extended shear tab configurations. It has been shown throughout the past
research on extended shear tabs that flexural stresses are induced into the minor axis of the
supporting girder member. However, the current design procedures lack information on the
magnitude and design method to account for these stresses. Two failure modes of the supporting
member’s web have been identified: bending failure or punching shear failure. It has been
suggested to address the bending failure with yield line analysis (Goldstein Apt, 2015 & Sherman
and Ghorbanpoor, 2002). Upon bending failure, the yield line analysis considers the formation of
plastic hinges along yield lines. The analysis proportions the shear tabs to ensure yielding occurs
prior to shear fracture in the web. The second failure mechanism is punching shear failure of the
supporting member’s web. This occurs when the applied load exceeds the shear resistance of the
shear tab.

36
For unstiffened extended shear tabs, the AISC design procedure allows the supporting column to
be designed for an axial force without eccentricity (AISC, 2017). This is due to the design of the
bolt group for an eccentricity from the weld line to the centre of the bolt group. However, it has
been seen that some moment is induced into the supporting column. Therefore another option
is to design the bolt group for a smaller eccentricity, for example, if 5% of the column’s minor
axis moment resistance is assumed to transfer through the extended shear tab. In this case, the
additional moments on the column must be accounted for in the member design (AISC, 2017).
This method would decrease the demand on the bolt group, therefore potentially having a more
economical connection.

Another consideration neglected in the Canadian and the American standards is the case where
axial loads are applied simultaneously with the gravity-induced shear force. Some adjustments
to the design procedure are suggested for combined shear and axial loads in the Steel Connection
Handbook Section 2.5.3 (Tamboli, 2016) and the AISC Design Example IIA-19B (2017). However,
minimal laboratory testing and numerical modelling have been performed to validate these
adjustments. Motallebi (2018) performed laboratory and finite element analyses on stiffened
extended shear tab connections under combined axial and shear forces. The results showed that
a compressive axial load allowed the connection to withstand higher shear forces, but increased
the out-of-plane deformation of the shear plate which could result in plate buckling. A tensile
axial load accelerated plate yielding since the interior bolt line suffered higher force demands,
however the out-of-plane deformation was decreased. Thomas (2014) also conducted laboratory
tests of shear tab connections under combined axial and shear loads. The resulting shear
resistance of the connections were lower than the same setups under shear load alone.
Compressive axial loads added to the out-of-plane deformation, similar to Motallebi’s (2018)
results.

37
3 Validation of Finite Element Models

This research project comprises finite element (FE) modelling (FEM) of steel shear tab
connections using the software Abaqus 6.18-1 (2018). An extensive review and study into the
parameters and assumptions involved is necessary to validate the numerical models. This chapter
covers the FE modelling of coupons, development of the true stress-strain curves, a review of
D’Aronco’s (2013) laboratory specimens, the FE modelling procedure and the FE simulation
results of D’Aronco’s (2013) Tests #5-9. Through comparison of the FEM and laboratory results,
a numerical model capable of simulating the response of the extended shear tab connections
under gravity-induced shear loading was created. The validated model was then used for the
parametric study, discussed in Chapter 4.

Coupon Modelling

To validate the material properties used for the shear tab FE models, additional models of the
steel test coupons were created in Abaqus. During D’Aronco’s (2013) laboratory testing, each
beam had 7 coupons: 2 from the top flange, 2 from the bottom flange and 3 from the web. For
each beam, the 4 coupons from the flanges were averaged to give the material properties used
for both bottom and top beam flanges in the Abaqus model. The 3 web coupons were averaged
to give the material properties of the beam webs. Three coupons were tested for each different
shear plate; the results of these tests were again averaged to give the material properties used
in Abaqus.

38
a) b)

Figure 3.1: FE Modelling of Coupon: a) Part in Abaqus b) Mesh

To numerically reproduce the coupons tested in the laboratory, a coupon was drawn with its
measured dimensions in AutoCAD 2019 and then imported into Abaqus (Figure 3.1). The material
properties inputted were Young’s Modulus, Poisson’s ratio and the true stress-strain curve. The
engineering stress-strain curve is based on the assumption that the cross-sectional area of the
coupon remains constant throughout the entire test. Whereas, in reality, necking occurs, and
therefore the area decreases as the loading (or displacement) continues. After the ultimate
strength is reached, this decrease in area occurs faster than the load decreases. Therefore, the
engineering stress-strain curve shows a decrease in strength beyond this point, whereas the true
stress-strain relationship continues to increase until failure (Figure 3.2). The development of the
true stress-strain curve is presented in Section 3.2.

39
Figure 3.2: Example of Engineering vs. True Stress-Strain Curve for Mild Steel

The numerical model of the coupon was then held fixed in all six degrees of freedom (3
translational and 3 rotational directions) at one end. An axial displacement in tension was applied
to all nodes at the opposite end. The displacement of the FE model followed the movement
implemented in the laboratory tests. Furthermore, the coupon had a global seed size of 5 mm
and local seed size of 3 mm at its centre, where necking was expected. Wedge Sweep C3D6 and
Hex Structured C3D8 elements were used for the curves and rectangular portions of the model,
respectively.

The stress and strain were required from the results of the finite element coupon models. Neither
of these were extracted explicitly. Rather, the stress was calculated as the force divided by the
initial cross-sectional area. This force came from a free body cut at the fixed end, which gave the
average force with respect to time. The strain was calculated by the change in length at the
central section of the coupon. The displacement of two points, 100 mm along the length in either
direction of the centre were recorded, i.e. a total gauge length of 200 mm. The strain was then

40
equal to the difference in their respective movements along the axis of the coupon length,
divided by the initial separation of the points (200 mm), ε = Δ L / L. By extracting the data as
mentioned, the engineering stress-strain curve was acquired.

Therefore, to compare the results from Abaqus to the laboratory coupon tests, the numerical
stress-strain curve was compared to the experimental results. This procedure was completed to
validate the use of Equations 3-1 to 3-4 presented in Section 3.2. The results proved that the
equations accurately represented the material properties, and could therefore be used to
calculate the average stress-strain curve for each specimen.

Developing the True Stress-Strain Curve

To obtain the true stress-strain curve for each material, the average yield strength (Fy), ultimate
strength (Fu) and strain at yield, at strain hardening and at ultimate were extracted from the
coupon tests (D’Aronco, 2013). The yield strength was determined by dividing the recorded
applied force by the initial cross-sectional area. The maximum stress was taken at the ultimate
strength. The yield point can be ambiguous to locate and therefore an offset method was used
to define it. This was done by fitting a line to the experimental engineering stress-strain curve in
the elastic region. First, results of some of the coupon tests were adjusted to eliminate the noise
caused by slippage. Then, by offsetting the fitted linear line overtop the experimental stress-
strain curve by 0.002, the intersection of the two curves was taken as the strain at yielding (εy).
The offset of 0.002 is a typical value representing 0.2% of the strain (Craig, 2011). The εsh is
typically between 5-15 times the strain at yielding. However this range tended to be much higher
than observed in the coupon testing of the shear tabs. This was due to the shear plates being
cold-rolled and therefore having a small, if any, yield plateau. Therefore, the εsh was determined
by selecting the strain on the curve at which strain hardening began. Of course, there is some
error in selecting this point manually, but the results matched the experimental data quite
accurately. Figure 3.3 illustrates the comparison for the beam flange material of D’Aronco’s
(2013) Test #6.

41
Figure 3.3: Comparison of FE Model and Experimental Stress vs. Strain Curves of the Beam
Flange of Test #6

With this data, curve fitting was performed using the suggested equations (Equations 3-1 and 3-
2) by the SAP Documentation under “Technical Notes - Material Stress Strain Curves - Simple
Structural Steel Parametric Stress-Strain Curve” (SAP 2000 Documentation, 2008). Professor
Robert Tremblay (Polytechnique Montréal) modified the exponential portion of the equation as
well as the constants A and C (Marshall, 2011).
𝐹
𝜎 = 𝐹𝑦 ∗ [1 + [(𝐹𝑢 − 1) ∗ (1 − 𝐶 ∗ (1 − 𝑟) ∗ 𝑒 (𝐴∗(𝜀−𝜀𝑠ℎ )) )]] (3-1)
𝑦

Where,
𝑟 = (𝜀 − 𝜀𝑠ℎ )/(𝜀𝑢 − 𝜀𝑠ℎ ) (3-2)

These equations are valid for the region of the stress-strain curve between the strain hardening
strain (εsh) and the ultimate strain (εu). In between the yield strain (εy) and εsh, the stress was set
equal to Fy. Beyond the ultimate point, no curve was developed, since this would require complex
equations to account for the progression of damage. The constant A influences the rate of strain

42
hardening (slope) and the constant C influences the start of the curve. It was determined by trial
and error that the most accurate results were obtained with A = -25 and C = 1 (Marshall, 2011).

The stress and strain values were then transformed into true data points as follows. (T: true, E:
engineering)

𝜀𝑇 = ln⁡(1 + 𝜀𝐸 ) (3-3)
𝜎𝑇 = 𝜎𝐸 ⁡(1 + 𝜀𝐸 ) (3-4)

The plastic region begins after the elastic region, therefore the yield strain (εy) was subtracted
from all true strain values for input into Abaqus. Figure 3.4 shows an example of the true stress-
strain curve from a W610x140 beam flange used in D’Aronco’s Test #6 (2013).

Figure 3.4: True Stress vs. Strain Curve of the Beam Flange of Test #6

43
Finally, since the true material properties of the columns were not known, expected properties
were assumed. The same method described above was used to generate the true stress-strain
curve for the columns based on a yield strength of RyFy and an ultimate strength of RuFu. The
strain at strain hardening and at ultimate were assumed based on a 0.2% offset to the yield
plateau. As well, a Young’s Modulus of 200000 MPa was used for the column material.

44
Details of D’Aronco’s (2013) Laboratory Tests

The validation of the finite element modelling procedure used in Abaqus was based on the results
of full-scale shear tab connection testing by D’Aronco (2013). These experiments were performed
at McGill University in 2011, and consisted of ten full-scale setups of shear tab connections. Six
of the specimens had flexible supports, four rigid supports and the final two were weld retrofit
connections. The objective of the research was to broaden the understanding of the behaviour
of shear tab connections with more than one vertical line of bolts. In particular; the rotational
ductility and shear resistance were monitored. In this study, five of the six flexible test setups
were modelled for validation of the modelling assumptions. Only the flexible tests were modelled
since this research focusses on the behaviour of beam-to-column web shear tab connections; i.e.
flexible supports. Some details of the laboratory experiments performed by D’Aronco are
presented; for further information please refer to the thesis Behaviour of Double and Triple
Vertical Rows of Bolts Shear Tab Connections and Weld Retrofits (D’Aronco, 2013).

The flexible specimens consisted of the shear tab welded to the web of the column, and then
bolted to the supported beam (Figure 3.5). The beams were laterally braced to prevent lateral-
torsional buckling and to force the failure to occur at the connection. To best simulate a flexible
support, full-scale 3.7 m long columns were used. The end conditions of the supporting columns
played a vital role in the behaviour of the connection. They simulated pins, while allowing
movement in plane. This was achieved by using a roller at the base and blocking the out-of-plane
movement. Similarly, at the top, a cap plate was placed with heavy angles to prevent out-of-
plane translations. Teflon pads between the column flanges and the angles facilitated movement
in-plane. The cap plate was bolted to a backup HSS to support any torque generated on the cap
plate due to possible movement of the test column. The HSS was firmly braced to the strong
floor. Rotations were allowed at both top and bottom since the columns were designed to
simulate a real building. The column length extended from inflection point to inflection point
(between adjacent floors), and therefore the columns were assumed to carry zero moment at
these inflection points.

45
Figure 3.5: Flexible Support Laboratory Test Setup (D'Aronco, 2013)

The shear tabs were shop welded in a flat position using a semi-automatic flux-core arc welding
wire feeding process (FCAW) and a gas shield. Due to widespread usage of this dual-shielding
welding procedure in the heavy structural steel industry, this welding process was selected for
this research project; it provides efficient, consistent and reliable results. It is a prequalified
procedure upon the requirements of the American Welding Society (AWS) and Canadian Welding
Bureau (CWB) (D’Aronco, 2013).

The materials used for testing were as follows:

 ASTM A572-Gr.50 (50 ksi) for shear tabs

46
 A992-Gr.50 for test beams

 A992-Gr.50 for test columns

 Grade E70 electrodes for welding

 A325 bolts with threads excluded from the shear plane

The majority of previous research was conducted using ASTM-A36 (36 ksi) steel for the shear
tabs. A higher grade of steel was used in D’Aronco’s tests to determine the ductility and rotation
it is capable of providing compared to the previous findings. Additionally, it is important to note
that the stress-strain curves for the tested shear tabs do not follow the typical stress-strain curve
of steel. They have no yield plateau, suggesting that the steel is of high strength and has been
worked in its preparation for fabrication. The beams were designed to withstand the applied
shear forces, allowing the failure to occur in the shear tab. The beams were also designed
indirectly to withstand the flexural moment generated due to the eccentricity of the loading.
Table 3.1 provides details of the laboratory specimens which are the focus of the FEM research
in this thesis.

Table 3.1: Shear Tab Laboratory Test Specimens (D'Aronco, 2013)

Vertical Horizontal Shear Tab


Test Support Column Bolt
Test Beam Rows of Rows of Thickness
ID Condition Size Size
Bolts Bolts (mm)
5 Flexible W360x122 W310x60 2 3 3/4" 10
6 Flexible W360x122 W610x140 2 6 7/8" 16
7 Flexible W360x314 W310x60 2 3 3/4" 10
8 Flexible W360x314 W310x74 3 3 3/4" 13
9 Flexible W360x314 W610x140 2 6 7/8" 16

The bolt spacing and gauge were 3” for all tests. Horizontal and vertical edge distances were 1.5”
and the first bolt hole was 3” below the top of the beam. All holes were standard and sized
according to Table J3.3 of the AISC Manual 14th Ed. (2011).

47
The laboratory tests from D’Aronco’s research were designed according to the standard design
procedures at the time of testing, i.e. the AISC Manual 14th Ed. (2011). Since the shear tab
connections did not comply with the requirements of the CISC Handbook of Steel Construction
(2010), the results could not be compared to tabulated design values; rather they were used to
provide recommendations to update the design procedure. As previously mentioned, the CISC
Handbook did not provide a valid design procedure since it only considers shear tabs with one
vertical row of bolts. Therefore the design procedure in the AISC Manual 14th Ed. (2011) was used
for preliminary design and sizing of the specimens. In the AISC Manual, all tested shear tabs fell
under the extended configuration category. Preliminary design calculations were performed
assuming a 10% increase in yield strength (Fy) and ultimate strength (Fu) since coupon testing
could only be performed after this step. It is standard practice to use RyFy and RyFu, where Ry =
1.1. Once the coupon testing was completed, and true values could be obtained, the material
properties were adjusted in the design procedure and compared once again.

The connection eccentricity was taken as the distance between the centreline of the bolt group
and the toe of the stabilizer plate (same as the edge of the column flange in this case) (Figure
3.6). This resembles the eccentricity of connections with rigid supports; for example if the shear
tab was welded to the column flange rather than to the column web (flexible support). This
eccentricity was also recommended by Fortney and Thornton (2016). Contrarily, the AISC Manual
14th Ed. (2011) design method recommends taking the eccentricity as the distance between the
centroid of the bolt group and the weld line at the column web. To examine the difference, all
D’Aronco’s shear tab connections framing into a flexible support were stiffened by horizontal
plates at the top and bottom of the shear tab. Therefore, the remaining eccentricity of the node
was transferred as a couple to the top and bottom column stiffener plates (D’Aronco, 2013). In
other words, due to the inclusion of the stiffener plates, the eccentricity of the connection with
respect to the column was assumed to be reduced. A revised version of the AISC Manual has
been made available since the laboratory tests performed by D’Aronco. Therefore, the new AISC
Manual 15th Ed. (2017) was used to calculate the predicted resistances of each test specimen in
this thesis. Table 3.2 gives these predictions as well as those predicted using the modified

48
eccentricity suggested by Fortney and Thornton (2016). Both design calculations were performed
using measured material properties; the details are provided in Appendix A: Design Calculations.

Figure 3.6: Dimensions of Extended Shear Tab (Modified Eccentricities)

Table 3.2: Predicted Connection Resistances

Test # AISC 15th Ed. (2017) Predicted Predicted Fortney & Predicted Predicted
Recommendation Resistance Failure Thornton (2016) Resistance Failure
(kN) Mode Recommendation (kN) Mode
e (mm) a (mm) e (mm) a (mm)
5 211.1 173 230 Bolt Shear 88.9 50.8 450 Bolt Shear
292 Von Mises 533 Von Mises
6 223.1 185 993 Bolt Shear 101.6 63.5 1626 Bolt Shear
1536 Von Mises 2032 Von Mises
7 277.1 239 181 Bolt Shear 88.9 50.8 450 Bolt Shear
224 Von Mises 533 Von Mises
8 315.2 239 287 Bolt Shear 127 50.8 579 Bolt Shear
268 Von Mises 639 Von Mises
9 289.1 251 803 Bolt Shear 101.6 63.5 1626 Bolt Shear
1295 Von Mises 2032 Von Mises

49
Displacements were applied to the test setup by both a main actuator near the connection and
a tip actuator at the tip of the beam. The role of the main actuator was to apply force to the
connection, while the tip actuator displaced the beam end to control the beam rotation. The
target rotation at the probable ultimate connection capacity was 0.02 radians. As previously
mentioned, this probable ultimate resistance was calculated using a 10% increase to the material
properties (Fy and Fu), since true values were not known at the time of testing. At this point, when
the target rotation and ultimate strength were reached, the speed of the displacement was
increased; the ratio between the main actuator displacement and the tip actuator displacement
was kept constant. In general, the initial rate of the main actuator was 0.1125 mm/min for the
flexible support condition tests, and the rate of the tip actuator was adjusted to maintain the
target rotation ratio.

Tests #5 and #7 differed by column size; Test #7 had a much heavier column (W360x314) which
was 7 times stiffer in terms of minor axis flexure as compared to the W360x122 column used for
Test #5. The laboratory results of Test #5 showed shear deformations in the bolt holes of the
shear tab; greater deformations were observed in the vertical line of bolt holes closest to the
face of the column. Test #7 demonstrated shear deformations and shear rupture through the
first line of bolts. It is known that shear yielding took place prior to flexural yielding from evidence
from the strain gauges.

Test #8 had three vertical rows of three bolts. It also suffered shear deformations around the bolt
holes of the shear tab. Shear yielding along the vertical line of bolt holes closest to the column
also occurred and led to shear fracture through the net area of the plate.

The same two columns used in Tests #5 and #7 were used for Tests #6 and #9; Test #9 having the
heavier and stiffer column. The relative rigidity of the supporting column explains why Test #6
did not reach the relative rotation target of 0.02 rad at the ultimate capacity. Yet, this difference
in stiffness had minimal influence on the shear capacity of the connection. The failure mode of

50
both Tests #6 and #9 was shear rupture through the net area of the shear tab. Similar to other
tests, shear deformations around the bolt holes closest to the column were more pronounced.
Also in Test #6, minor axis bending in the supporting column was observed; it exceeded the
bending capacity of the column. Therefore Mp was reached and surpassed, causing the column
to yield in flexure. Permanent damage could be observed in the bolt at the top of the connection
closest to the column. Additionally, weld fracture occurred at the top shear tab to stiffener weld.
It was assumed that the Test #6 would have reached higher rotation values should the weld not
have fractured.

Overall, in the five connection tests with flexible supports, shear yielding of the shear tab
occurred prior to flexural yielding. This was due to the bolts allowing rotation early in the test. To
assess the bending moments induced on the columns during the extended shear tab connection
tests, the following moment diagram (Figure 3.7) and free body diagram (Figure 3.8) are
presented (D’Aronco, 2013).

Figure 3.7: General Moment Diagram for Flexible Support Conditions (D'Aronco, 2013)

51
Figure 3.8: General Free-Body Diagram for Flexible Support Conditions (D'Aronco, 2013)

The negative moment at the face of the column can be calculated by equilibrating the forces and
moments at point g (location of tip actuator). Similarly, the moment in the beam can be
calculated by summing the moments at point f. Knowing these two moments, the inflection point
(ecalc) can be determined. The moments induced in the supporting column are not accounted for
in the design of the columns. It may be acceptable to ignore these moments in core columns
since they have shear tab connections on both sides, therefore counteracting the imposed
moments. However, this is not the case for perimeter columns having a shear tab connection
acting only on one side of the column.

It was noticed that as the loading increased from the calculated nominal shear capacity of the
connection to the ultimate shear load, the point of inflection in the beam moment diagram
shifted along the beam span towards the bolt group. With the shallow beams, W310, the actual
eccentricity increased with the increase in the applied shear load at the connection, whereas the

52
opposite occurred in the deeper beams, W610 (D’Aronco, 2013). The actual eccentricity
discussed here is the distance between the point of zero moment and the centreline of the bolt
group.

FE Modelling Procedure

To validate the finite element modelling techniques, preliminary models of full-scale Tests #5-9
performed by D’Aronco (2013) were analysed. As previously mentioned, these tests consisted of
extended shear tabs welded to a column web (flexible support) and bolted to a supported beam.
A detailed discussion of each of the Modules in Abaqus 6.18-1 is presented in Sections 3.4.1 to
3.4.10 to justify the finite element modelling techniques and assumptions.

3.4.1 Part Module

All the parts were initially drawn and assembled in AutoCAD to ease the procedure in Abaqus.
Each component was then imported as an individual deformable 3D part into Abaqus (Figure 3.9).

53
Figure 3.9: Exploded View of Meshed FE Model (each colour represents a different part of the
assembly)

Parts were needed for the column top, column bottom, shear tab, bolt, loading plate, beam and
beam near the connection. The beams and columns were both modelled in several parts since a
finer mesh was required near the connection where failure was expected. The parts further from
the connection, which remained elastic, had a coarser mesh to reduce the computational
demand of the models. To match the laboratory tests, the beams and shear plates were drawn
using average measured dimensions since this data was available. All other parts not measured
at the time of testing were drawn with dimensions from the CISC Handbook (2015).

54
Shop drawings and technical drawings aided in the dimensioning of the test setup. The shear tab
was modelled as one part with the stiffener plates on a portion of the column extending 100 mm
from the top and bottom of the stiffener plates. This additional 100 mm was to prevent sudden
changes in mesh size to occur right at the connection, potentially leading to inaccurate results. A
coarse mesh can result in convergence issues or unreliable results in areas of high deformation,
particularly due to the calculation of stresses and strains based on displacement gradients. For
the same reason, the beam near the connection part extended approximately 50 mm beyond the
edge of the loading plate at the location of the main actuator. There was no consideration of
weld tearing in the FE models since many uncertainties are introduced when modelling damage
initiation and propagation. As well, the weld material properties were unknown.

All bolt holes were 1/16” larger than the bolt diameter as specified in the AISC Manual 15th Ed.
(2017) Table J3.3. The bolt shaft was drawn exactly the length of the beam web plus the thickness
of the shear tab; this allowed the bolt head and nut to be in contact with the beam web and shear
plate as they were when snug-tightened in the laboratory tests.

The AutoCAD drawings were imported into Abaqus as homogenous, deformable, solid parts.
They were partitioned to simplify the geometry and ease the meshing. This avoids having
distorted elements and gives freedom to specify different mesh sizes in various locations of the
same part.

3.4.2 Property Module

In this Module, the different material properties were defined. Similar to the dimensions of each
part, the measured properties were used whenever possible and probable properties were used
otherwise. Coupon tests were performed for the beams and shear plates, but material properties
for the columns were not measured.

55
Abaqus requires the true stress-strain data to adequately capture the steel necking, since finite
element modelling considers large deformations and displacements. Therefore, as mentioned in
Section 3.2, the true stress-strain curves were derived using Equations 3-3 and 3-4.

In Abaqus, the required elastic properties are Young’s Modulus (E) and Poisson’s ratio (taken as
0.3). The plastic properties required are the data points along the true stress-strain curve,
beginning at the yield point. The plastic region of the true stress-strain data of an A325 bolt is
shown in Table 3.3 and was taken from the literature (Kulak et al., 2001).

Table 3.3: Plastic Region of True Stress-Strain Data of A325 Bolts

Yield Stress (MPa) Plastic Strain


637 0
680.6 0.002084
720.2 0.004872
745.8 0.007645
757.8 0.011089
766.7 0.016569
795.9 0.023389
839.1 0.034867
863.4 0.046207
896.6 0.080005

The loading plates located between the beam and the actuators transferred the load evenly,
without causing localized damage. Elastic properties were used since they were not the focus of
the test. The stiffeners on the beam at the location of the main actuator were also assumed to
have elastic properties. Finally, the stabilizer plates above and below the shear plates had the
same true stress-strain curve as the shear plate of the same thickness. This is based on the
knowledge that the stiffener plates were fabricated from the same steel plate as the shear tabs.

56
3.4.3 Assembly Module

All the parts were then brought into the Assembly Module as dependent part instances.
Dependent part instances were used since they are pointers to the original part. Therefore they
cannot be meshed as an instance, but rather must be meshed as a part. For example there was
only one bolt part but several instances of this part in the Assembly. On the contrary,
independent part instances are a copy of the geometry of the original part. In this case, each
instance is meshed rather than the part itself. This consumes more memory resources than using
dependent part instances (Simulia User Assistance, 2018).

All parts were initially in contact except for the bolts. Since there was no way of determining the
exact contact points of the bolts within the bolt holes of the beam and shear plate, the bolt shank
was initially presumed to be at the centre of the bolt hole in the numerical models.

3.4.4 Step Module

All Abaqus models have an “Initial” default step which cannot be edited. Boundary conditions are
typically initiated in this step. In the full-scale tests, some force was applied to the bolts when
they were snug-tightened. To simulate this initial pretension force prior to the initiation of
loading, a Static, General Procedure Type step “Preload” was created after the “Initial” step. A
general step means that the starting conditions of the step are the ending conditions of the
previous general step. The step times of all general steps add up to the total time of the analysis.
Therefore, the step time of the “Preload” step was 0.1; which was small in comparison to the
subsequent step in which the loads were applied. Nonlinear geometry was turned on and
remained activated for the entirety of the analysis. This accounted for the nonlinear effects from
large displacements and deformations. If the model was expected to undergo small
displacements, nonlinear effects may have been ignored; however, in the case of the shear tab
connections, which were loaded to failure, large deformations were expected and as such,
nonlinear effects had to be included. To continue with the definition of the “Preload” step, a
small initial increment size of 0.001 was used to avoid convergence problems. Also, automatic
57
incrementation was used to allow the increment sizes to vary between specified limits. Finally,
the matrix storage was set to unsymmetric. Abaqus can automatically determine which type of
matrix storage to use or it can be explicitly defined by the user. The stiffness matrix is used
regardless if the analysis is linear (solved by Stiffness Method) or nonlinear (solved by Newton’s
Method). The unsymmetric matrix storage is suggested when the friction coefficient is greater
than 0.2 which was the case for these steel shear tab models (coefficient of 0.3).

Second, the “Loading” step was defined. All parameters were identical to the “Preload” step
except for the step time. Since the actuator displacements were applied in this step, a larger step
size of 1.0 was used. Therefore the total time of the analysis was 1.1.

In the Step Module, Field Output Requests and History Output Requests were specified. Field
outputs are infrequent requests for a large portion of the model. Complete sets of basic variables
(ex: all the stress components) can be requested rather than specific components of basic
variables. History outputs are frequent output requests for small portions of the model. Contrary
to field output requests, history output requests can include individual variables (ex: stress in U1
direction).

In the laboratory tests, instrumentation was used to record the displacement, rotation, force and
strain of specific locations on the test setup. Therefore, sets were created in the Assembly
Module at the instrumentation locations to then output history requests based on these defined
sets. In general, these included: shear tab vertical displacement, shear tab out-of-plane
displacement, beam lateral displacement, beam vertical displacement, bolt vertical
displacement, rotation of beam near connection, rotation of beam near the tip, rotation of the
column, load cells on the column, displacement of main actuator, and displacement of tip
actuator.

58
3.4.5 Interaction Module

The Interaction Module defines the contact properties and locations of the assembly which are
in contact or which can come into contact upon deformation. In the shear tab models, there were
two interaction properties: steel-to-steel contact, and frictionless rollers at the points of loading.
The friction between steel and steel was defined as the “Friction” property. It had both tangential
and normal behaviours. The tangential behaviour was defined as follows:

 Friction formulation: Penalty

 Directionality: Isotropic

 Friction coefficient: 0.3

The friction formulation “Penalty” uses the basic Coulomb friction model. This relates the
maximum allowable frictional shear stress to the contact pressure between the two bodies in
contact (SIMULIA, 2018). Therefore, the surfaces carry shear stresses up until a certain limit when
they begin to slide with respect to one another (Figure 3.10). Isotropic means that the coefficient
of friction is the same in all directions. The chosen coefficient of friction of 0.3 was suggested for
steel-to-steel contact in the literature (see Chapter 2 Literature Review).

Figure 3.10: Slip Region for the Basic Coulomb Friction Model (SIMULIA, 2018)

59
The normal behaviour was defined as follows:

 Pressure-Overclosure: “Hard” Contact

 Constraint enforcement method: Penalty (Standard)

 Allow separation after contact

 Contact stiffness behaviour: Linear

 Stiffness value: Use default

 Stiffness scale factor: 1

 Clearance at which contact pressure is zero: 0

The “Hard” contact relationship means that two surfaces come into contact when the clearance
distance between them reduces to zero, and the surfaces separate when the contact pressure
reduces to zero. By using this pressure-overclosure relationship, the enforcement method is
typically “Penalty.” With this method, the contact force is proportional to the penetration
distance. Linear contact stiffness behaviour was chosen for the friction contact property since
this keeps the penalty stiffness constant; therefore the contact pressure and overclosure
relationship is linear. The default stiffness scale factor is 1, however it was suggested that a value
of 0.001 to 0.003 could be used for steel connection models (Daneshvar, 2013). A comparison of
models with the default stiffness scale factor and a reduced factor of 0.003 showed little
difference in the results (Figure 3.11). Finally, separation was allowed after contact, because the
loading may have caused some surfaces to deform away from others, or the bolts may have lost
contact to the bolt holes as bolt hole elongation occurred.

60
Figure 3.11: Difference between Stiffness Scale Factor = 1 and = 0.003 for Test #8

The second contact property was the “Rollers” property which was used at the locations where
the actuators apply displacement to the beam (one near the connection and one near the tip). In
this case, the tangential behaviour was frictionless since the rollers were free to move. The
normal behaviour was identical to the “Friction” contact property.

To continue, all contact pairs were found and created in Abaqus; they have various parameters
which could be altered. These include:

 Type: interaction or tie,

 Sliding: finite or small,

 Discretization: surface-to-surface,

 Property: friction or rollers as defined previously,

 Adjust interactions to remove overclosure,

 Surface smoothing.
61
To begin explaining interactions, master and slave surfaces must be defined. In general, the slave
surface is that which follows the movement of the master surface. Therefore it should be the
instance which is expected to see higher deformations. Following this logic, the slave surface
should have a smaller mesh density. If the two surfaces are not significantly different in mesh
size, the master surface should be the one with a coarser mesh or which is stiffer and therefore
is expected to experience less deformation.

If two elements cannot move with respect to one another, this is defined as a tie constraint. One
example of this is the two portions of the beam which were tied together since they were one
member in reality, and therefore the parts could not move relative to one another. All other
interactions represented elements in contact, with either friction between them or the rollers at
the locations of the actuators.

To continue, a finite sliding formulation was selected since this is the general formulation which
allows any motion of the surfaces. Small sliding allows large motions but minimal sliding between
the surfaces.

The discretization method was surface-to-surface rather than node-to-surface. The former
provides a more accurate stress distribution which is less sensitive to master and slave surface
designations. It considers the shape of both the master and the slave surface and enforces
contact conditions as an average nearby the slave nodes. In the node-to-surface contact
discretization, contact conditions are enforced only at individual slave nodes. Node-to-surface
discretization should be used when contact surface normals are not approximately normal to one
another. The node-to-surface method constrains the slave nodes not to penetrate into the
master surface but the nodes of the master surface can penetrate into the nodes of the slave
surface (SIMULIA, 2018).

62
Next, interactions can be adjusted to remove overclosure. This moves the slave surface nodes to
be precisely in contact with the master surface without creating any initial strain in the model.
Abaqus requires all tied surfaces to use this feature but it can be toggled on or off for contact
interactions. Models with and without the adjustment to remove overclosure feature were
compared and the results showed no differences. The reason for this is the assembly was drawn
with all interaction locations in contact except for the bolts within the bolt holes. The bolts would
not be adjusted to remove overclosure since their separation distance is larger than the
tolerance. Therefore, no interactions were adjusted to remove overclosure.

Surface smoothing is important when curved surfaces are in contact in the assembly. In finite
element analysis, curved geometric surfaces are approximated as a connected group of element
faces. These faceted surfaces differ from the true surface geometry and can consequently lead
to contact stress inaccuracies. By applying surface smoothing, the noise created in the contact
pressure is reduced as compared to using the true geometry with no smoothing.

To continue, in the Interaction Module, contact controls can be specified. These include
automatic stabilization which helps to automatically control rigid body motion in static problems,
before contact closure and friction restrain such motion. This feature should be used when it is
clear that contact will be established but cannot necessarily be done in the modelling. Abaqus
activates viscous damping for relative motions of the contact pair at all slave nodes. If no contact
controls are specified, Abaqus uses a default contact control which is adequate for most analyses.
Models with and without defining an automatic stabilization contact control were analysed and
the results showed that the model without the defined controls yielded the most accurate
results. Therefore, default contact controls were used for all the shear tab models.

Finally, in the initial runs of the validation models, the load cell locations were assumed to have
some flexibility since, in the laboratory tests, they were connected to a backup frame and then
bolted to the strong floor (see Figure 3.5). These load cells were placed just offset from the

63
column top and bottom. To accurately represent this flexibility, a SAP 2000 19 model of the
backup frame was created. This included the steel sections of the backup HSS column and
diagonals fixed to the ground. By applying a displacement to the SAP 2000 model, the resulting
force was measured to then calculate the stiffness at both load cell locations. Validation Test #5
was modelled using these stiffness values implemented as springs in Abaqus. The results proved
to differ slightly from the results using full restraints at the load cell locations in the U1 direction.
Since the differences between the models were deemed negligible, the springs were not used in
further models. Should they have been included, additional variables and uncertainties would be
introduced into the finite element models. Therefore using a full restraint simplified the models,
while maintaining sufficient accuracy and reliability. The force vs. absolute beam rotation curves
of Test #5 with and without springs are presented in Figure 3.12 for comparison.

Figure 3.12: Abaqus Results Showing Negligible Difference between Full Restraint and Springs at
Load Cell Locations for FE Models of Test #5

3.4.6 Load Module

First, the pretension in the bolts was applied in the “Preload” step and then modified in the
“Loading” step. The mechanical bolt load was applied to the cross-section of the bolt shaft where
64
the beam web and the shear tab came into contact. Bolt loads can be applied by two methods:
apply force or adjust length. The apply force technique requires a force magnitude to be explicitly
stated and the adjust length method requires a length change. In the laboratory, the bolts were
snug-tightened to the full capacity of a worker using a spud wrench with no further
pretensioning. Therefore, the apply force method was used with a magnitude of F = Ab * 50 MPa.
Therefore, for A325 ¾” bolts: F = 285 mm2 * 50 MPa = 14250 N. This calculation was performed
in accordance with past researchers: Motallebi (2018) and Daneshvar (2013). Higher values of
bolt-pretension were used by other researchers, such as Kline et al. (1995). Therefore, models
with higher pretension, up to 50 kN, were tried and the results proved to be very similar. Rather
than propagating this force in the “Loading” step, the bolt load was fixed at the current length
after the “Preload” step, restricting the bolt from any further changes.

In previous versions of Abaqus, one bolt load was needed for each bolt. In Abaqus 2018, there is
an option to “pretension section at part level.” This allows for one bolt load to be specified and
applied to all instances of this part. To verify this new feature, an analysis was performed on a
model with a bolt load created at each bolt and an identical model using the “pretension section
at part level” option applied only to one bolt. Both yielded identical results.

To continue, the full-scale testing was performed by applying a displacement with respect to time
at both the main and tip actuators. Therefore, these were implemented in the model as boundary
conditions in the “Loading” step. First, amplitudes were defined as tabular displacements with
respect to the normalized time. Figure 3.13 shows the loading protocol from Test #5; similar
protocols were used for all other tests by D’Aronco (2013). To create these displacements at the
points of loading, the amplitude was applied to the centreline of the loading plate (perpendicular
to the beam) (Figure 3.14a). Note that the tip actuator displacement was monitored and
controlled during the laboratory tests to follow the beam end rotation. During the testing in 2011,
the tip actuator was adjusted based on the relative rotation of the beam, however it should have
been controlled based on the absolute rotation. Therefore, other loading protocols were

65
investigated in the FE models to determine realistic moment-shear interactions acting on the
connection. This is discussed further in Section 4.6.

Main
Actuator

Tip
Actuator

Figure 3.13: Typical Loading Protocol of Main and Tip Actuators

Other boundary conditions were necessary for the stability of the model. On both loading plates,
the two corner points of the line of load application were restricted in all three translational
degrees of freedom U1, U2 and U3 (x, y and z, respectively as seen in Figure 3.15). This was
applied in the “Initial” step, propagated in the “Preload” step and modified in the “Loading” step
to have U1 = U3 = 0 and allowing U2 to translate with the movement of the actuator. This
boundary condition ensured that the load was always applied normal to the beam flange. In the
laboratory, this was achieved using steel rollers and a half cylinder to connect the actuator to the
beam (Figure 3.14). This technique to apply the actuator loads was consistent with methods used
by other researchers: Mirzaei (2014) and Motallebi (2018). Next, the beam had lateral restraints
to prevent lateral torsional buckling during the test and to force the failure to occur in the shear
tab connection. To simulate these lateral restraints, a boundary condition was created along the

66
thickness of the flange (top and/or bottom depending on the setup). Out-of-plane movement
was restricted, therefore the U3 degree of freedom was set equal to zero in all three steps of the
analysis.

a) b)

U2

Figure 3.14: a) Loading Plate in Abaqus b) Load Application at Main Actuator (D’Aronco, 2013)

Additionally, the column was designed to be continuous at the floor level (at the beam) as it
would be designed in a typical building. This means the length of the column was designed to
extend from midway of one storey to midway of the next storey; therefore at the points of
inflection (points of zero moment). Consequently, the end fixity of the column must allow
movement in-plane and restrict movement out-of-plane, while allowing rotation at the top and
bottom. Therefore in the model (see Figure 3.15 for global coordination system used):

 The entire cross-section at the top of the column was restrained in the U3 (z) direction to
prevent out-of-plane movement. This simulated the steel angles with Teflon pads used in
the laboratory. Since rotations were permitted, this method adequately simulated the
inflection point of the column since no moments were generated at the column top.

 At the location of the top and bottom load cells, the column experienced some resistance
and was therefore restrained in the U1 (x) direction. Several other boundary conditions

67
were investigated at the load cell locations. These areas were not fully restrained; there
was some flexibility based on the backup frame. Therefore, a spring was used to simulate
the load cells, backup HSS and angles fixing it to the ground. As explained in Section 3.4.5,
this proved to have minimal influence on the results and consequently was not
implemented into the final shear tab models.

 The bottom portion of the column was considered to extend to the base of the half
cylinder, below the column, as this is the actual point of rotation of the column. Therefore,
the column base cross-section was coupled to a reference point at its centre. The
reference point was then restrained in U2 and U3 (y and z). Consequently, the column
was free to rotate about its base. This method adequately simulated the inflection point
of the column since no moments were generated at the column base.

Figure 3.15: FE Model Assembly showing Coordination System

The purpose of these boundary conditions was to allow the column to rotate at the top and
bottom and to shorten at the top. Therefore, the boundary conditions simulated flexural hinges
at both ends of the column. Brick elements were used, which have no rotational degrees of
freedom at their nodes. Therefore, no restraints were required in the UR1, UR2 and UR3
directions. Should coupling be used at the top and bottom of the column, a restraint at either the

68
top or bottom reference point would have been necessary in UR2 for the model to be stable.
Both techniques proved to give equal results.

3.4.7 Mesh Module

The mesh plays a vital role in finite element modelling. The number of elements and
consequently the size of the model are determined based on the seed size. The seeds are
propagated along all edges. Therefore, to create a consistent and smooth mesh, partitions were
required to divide each part into simple geometric shapes (example in Figure 3.16).

a) b)

c)

Figure 3.16: Beam Connection Part in Abaqus: a) Geometry b) With Partitions and c) Mesh

69
Since dependent part instances were used in the assembly, the meshes were specified at the part
level. In general, a finer mesh is desired at locations where deformations and stresses are
expected to be higher. A denser mesh yields more accurate results, however it is computationally
expensive and requires large amounts of memory. Therefore, the ideal mesh size is the largest
mesh which provides acceptably accurate results. To determine this mesh size, several models
were compared to conduct a mesh refinement analysis. Some models with a mesh size of 7 mm
on the shear plate and 24 elements around the circumference of the bolt holes could not
converge, as they had trouble bringing the bolts into contact with the beam and the shear plate.
Other models, with the same element sizes, converged but showed significant discrepancies in
the results as compared to models with a finer mesh size of 5 mm and 40 elements around the
circumference of the bolt holes. For example, the bolt hole deformation could be captured more
accurately with a mesh size of 3-5 mm as compared to a mesh of 7 mm (Figure 3.17). As well, a
mesh size greater than 5 mm proved to lack accuracy in monitoring the out-of-plane deformation
of the shear plate.

a) b)

Figure 3.17: Bolt Hole Deformation a) Finer Mesh: 5 mm mesh with 40 elements around bolt
holes b) Coarser Mesh: 7 mm mesh with 24 elements around bolt holes

70
For all the parts, the element shape and meshing technique used was either Hex Structured
(C3D8) or Wedge Sweep (C3D6). The hex elements were used for all rectangular shapes and the
wedges were used for triangular shapes, i.e. plate corners and k areas of the W-sections. The
element names describe the element type. The C represents the continuum stress/displacement
family which is for all solid and fluid elements. Next, 3D indicates that it is a three-dimensional
element. The second number is the number of nodes in the element; 8 for hex and 6 for wedge.
The element types used in the FE models were inspired by past researchers: Motallebi (2018) and
Rahman et al. (2007). To continue, reduced integration lowers the number of integration points
to create the stiffness matrix and can therefore yield inaccurate results. Full integration elements
were used for all models in this research.

The mesh size used for the majority of the tests in this study is as follows (Figure 3.9):

 Column bottom: 10 mm

 Column top: 10 mm

 Loading plates: 10 mm

 Bolts: 1.5 mm shaft, 3 mm head

 Beam: 40 mm

 Beam near connection: 20 mm

 6 seeds per quarter of bolt hole edge = 24 elements around bolt holes

 Shear tab: 5 mm

 10 seeds per quarter of bolt hole edge = 40 elements around bolt holes

3.4.8 Job Module

A job was created for each analysis. A job runs a model based on all the parameters defined in
each Module in Abaqus, as described in Sections 3.4.1 to 3.4.7. Any warnings and/or errors can
be monitored as the job is running.
71
3.4.9 Imperfection Introduction

In FE simulation, it is easy to draw all dimensions with perfect accuracy and straightness, however
imperfections are nearly impossible to avoid in reality. Therefore, the introduction of
imperfections into the numerical models was performed to yield more realistic results. The
introduction of imperfections was not performed on all models; it was performed when deemed
necessary. To implement the imperfections, appropriate buckling modes were scaled and applied
to the model prior to analysis.

To implement this in Abaqus, a copy of the entire model was made for the buckling analysis.
Here, both the “Preload” and “Loading” steps were suppressed. To replace these steps, a new
Linear Perturbation – Buckle step (“Buckle”) was created. The subspace eigensolver was selected
since this is a faster iteration method when less than 20 eigenmodes are required as compared
to the Lanczos eigensolver (SIMULIA, 2018). Eight eigenvalues were requested, each giving a
buckling mode.

To perform the buckling analysis, two new loads were created in the “Buckle” step. Both were
General - Surface Traction type loads. The surface where buckling was expected to occur is
highlighted in Figure 3.18a. One load was created in the U1 direction (-1,0,0) and the other in the
U2 direction (0,-1,0), both with magnitudes of 1.0. The traction was defined per unit of
undeformed area and allows the force to change with rotation. At this point, a new job “Buckle”
was created and submitted. The resulting deformed model gave the eight mode shapes, one of
which is presented in Figure 3.18b. Buckling depends primarily on the gap between the shear tab
and the beam, the depth and the length of the shear tab.

72
a) b)

Figure 3.18: Imperfection Introduction in FE Models: a) Buckling Region b) Typical Mode Shape

The next step was to apply the scaled mode shape deformation in the initial step of the
displacement analysis to represent the imperfections. The scale factor for local imperfections is
= d/150 for the web (where d = depth of shear tab in the column) and = bf/150 for the flanges.
Global imperfections for out-of-straightness are = L/1000 (CISC, 2016, CSA G40.20-13, 2013 and
ASTM, 2003). In the literature, these tolerances have been verified and considered appropriate
(Elkady, 2016). These are maximum values and therefore represent a worst case scenario; most
fabrication facilities can easily meet these tolerances. For example, for Test #5, the maximum
scale factor = 283/150 = 1.89. The following lines were added to the input file to include
imperfections:
**
**Imperfection Introduction
*Imperfection, file=Buckle, step=1
2, -1.89
**
The file is Buckle since this was the ODB output file where the displacements of the buckled shape
were recorded. Step = 1 since the buckling step was the first (after the “Initial” step) in the
“Buckle” model. On the third line, the first number is the mode shape selected and the second is
the scale factor. The mode shape which most accurately represents the deformation of the model

73
should be chosen. Combinations of several mode shapes can also be specified. The scale factor
shall be positive or negative depending on the direction the shear tab is expected to buckle.

A new job was then created with this input file. The job was submitted and the results were
analysed to determine the importance of including imperfections in the models.

3.4.10 Visualization Module

The history outputs previously created were used to write XY Data in the Visualization Module in
Abaqus. Further results could be obtained using the field output requests. Free body cuts were
used to obtain nodal forces (NFORC) based on element stresses. The free body cuts provided
average force and moment magnitudes of all the nodes and elements selected for the cut. The
numerical outputs were then used to plot force vs. rotation curves, displacements and stresses.
These results will be discussed in Section 3.5 for the validation models and in Chapter 4 for the
parametric study.

Results of Validation Models

The validation models described in Section 3.3 were modelled in Abaqus and the numerical
results were compared to those obtained in the laboratory. The primary model verification
criteria were the shear force and beam rotation. Some discrepancies in the initial increments of
the loading step were due to the uncertainty of the bolt shank placement within the bolt hole, as
discussed in Section 3.4.3.

3.5.1 Results of Test #5

Test #5 comprised a W310x60 beam connected to a W360x122 column by means of 3 rows of 2


bolts. The bolts were ¾” and the shear plate was 10 mm thick. In the laboratory test, shear
deformation occurred around the bolt holes of the shear plate and the observed failure mode
74
was shear rupture through the interior line of bolts, closest to the column, illustrated in Figure
3.21b (D’Aronco, 2013).

Test #5 was modelled according to the assumptions described in Section 3.4. The connection
shear force, absolute beam rotation and vertical displacement of the shear tab were analysed
and compared to the experimental results in Figure 3.19.

a)

b)

Figure 3.19: Test #5 Experimental vs. FE Model Results a) Shear Force vs. Absolute Beam
Rotation b) Shear Force vs. Shear Tab Vertical Displacement
75
The progression of yielding up to failure in the Abaqus model matched the experimental results.
Shear rupture through the interior line of bolts was the primary failure mode. Yielding initiated
at the re-entrant corners and due to bearing on the exterior bolt line. The yielding then
propagated from the re-entrant corners to the interior line of bolts and to the stabilizer plates
simultaneously (Figure 3.20). The coloured contours in the figure represent the Von Mises stress
distribution; gray areas have yielded.

a) b)

Figure 3.20: Test #5 FE Results: a) Yield Initiation at Beam Rotation = 0.0073 rad b) Yield
Propagation at Beam Rotation = 0.0134 rad

The final deformed shape of the model was then compared to the deformed shape captured in
the laboratory (Figure 3.21). The shear tab suffered out-of-plane deformations at the base and
bearing damage around its bolt holes. Elongation of the bolt holes along the interior bolt line led
to net section failure. Primary failure occurred at the interior line of bolts because it experienced
the highest rotation. There was no visible deformation of the shear tab within the limits of the
stabilizer plates, suggesting that they provide adequate stability and stiffness to the connection.

76
a) b)

Figure 3.21: Deformed Shape of Shear Tab after Test #5 in a) Abaqus b) Laboratory (D'Aronco,
2013) (U3: Out-of-Plane Deformation)

The beam also experienced yielding (Figure 3.22). It initiated in the bolt line furthest from the
column and propagated through the web up to the beam stiffeners.

a)

Figure 3.22: Deformed Shape of Beam after Test #5 in a) Abaqus b) Laboratory (D'Aronco, 2013)

77
b)

Figure 3.22 Cont.: Deformed Shape of Beam after Test #5 in a) Abaqus b) Laboratory (D'Aronco,
2013)

3.5.2 Results of Test #6

Test #6 consists of a W610x140 beam connected to a W360x122 column by means of 6 rows of


2 bolts. The bolts were 7/8” and the shear plate was 16 mm thick. In the laboratory tests, shear
deformation occurred around the bolt holes of the shear plate and the observed failure mode
was again shear rupture through the first line of bolts. The entire weld along the length of the
shear tab to top stabilizer plate fractured. It was suggested that this fracture was caused by the
extensive deformations to the column and stabilizer plates (D’Aronco, 2013). The minor axis
bending capacity of the column was exceeded, therefore causing flexural yielding near the
bottom stabilizer plate.

Test #6 was modelled according to the assumptions described in Section 3.4. The connection
shear force, absolute beam rotation and shear tab vertical displacement were analysed and
compared to the experimental results (Figure 3.23).

78
a)

b)

Figure 3.23: Test #6 Experimental vs. FE Model Results a) Shear Force vs. Absolute Beam
Rotation b) Shear Force vs. Shear Tab Vertical Displacement

The observed failure mode in the Abaqus model matched the experimental results. Yielding
initiated at the re-entrant corners and then propagated to the stiffened regions (toward the

79
stabilizer plates), before yielding through the interior line of bolts (Figure 3.24). Therefore, shear
rupture through the interior line of bolts was the primary failure mode.

a) b)

Figure 3.24: Test #6 FE Results: a) Yield Initiation at Beam Rotation = 0.0051 rad b) Yield
Propagation at Beam Rotation = 0.0192 rad

a) b)

Figure 3.25: Deformed Shape of Shear Tab after Test #6 in a) Abaqus b) Laboratory (D’Aronco,
2013) (U3: Out-of-Plane Deformation)

80
Figure 3.25 shows the deformed shape of Test #6. Similar to Test #5, the shear tab suffered out-
of-plane deformations at its base and top (Figure 3.26) as well as bolt hole elongation in the
interior bolt line.

a) b)

Figure 3.26: Out-of-Plane Deformation of Shear Tab Test #6 in a) Abaqus b) Laboratory


(D'Aronco, 2013)

Due to the deeper beam section and increased number of bolts, the connection was much
stronger than Test #5, as expected. The high shear forces caused the beam and column to suffer
some damage, as well as the shear tab. The beam web yielded and bolt holes were deformed as
seen in Figure 3.27. Figure 3.27c and d show beam deformation at the bottom row of bolts.
Unfortunately no photos were taken of the beam without the bolts, therefore the bolt hole
deformations cannot be directly compared to the experimental results.

81
Figure 3.27: Beam Yielding and Deformation from Test #6: a) and b) Beam near the Connection
c) and d) Close-up View

Again, due to the high forces, the column also suffered yielding. The column flanges at the bottom
stabilizer plate yielded in both the laboratory and the numerical tests as can be seen in Figure
3.28.

82
Figure 3.28: Yielding of Column Flanges of Test #6 in a) Abaqus b) Laboratory (D’Aronco, 2013)

3.5.3 Summary of Validation Models

Finite element models of five full-scale extended shear tab connections from beam-to-column
web were created to replicate the laboratory specimens and loading of the tests conducted by
D’Aronco (2013). Tests #7-9 were analysed and compared to the laboratory results similar to
Sections 3.5.1 and 3.5.2. These results are presented in Appendix B: Results of Validation Models.
All tests provided accurate representations of the laboratory test results, therefore proving the
modelling assumptions and techniques to be reputable and adequate for use in the parametric
study.

83
Table 3.4 contains the results of a comparison of the maximum shear force obtained from
laboratory results and Abaqus modelling results. The predicted resistances based on measured
material properties calculated using the AISC Manual 15th Ed. (2017) design procedure with
eccentricities modified according to Fortney and Thornton (2016) are also included. These
significantly under-predict the actual forces reached during the tests. The discrepancies are due
to the laboratory tests determining the ultimate resistances of each connection, however these
are typically achieved at high rotations which are unlikely to occur, or to be acceptable, in real
applications. Finally, the test-to-predicted ratios range from 0.99 – 1.07 (between numerical and
experimental results). Due to the numerous uncertainties between laboratory tests and finite
element modelling assumptions, these results are acceptable and can be used as validation for
the modelling parameters described in Section 3.4.

Table 3.4: Predicted and Observed Maximum Shear Force and Failure Mode for the Validation
Models

Test Predicted Predicted Observed Observed Maximum Observed Test-to-


# Resistance Failure Maximum Failure Connection Failure Predicted
Based on Mode Connection Mode Shear Mode Ratios
Measured Shear Capacity (Abaqus
Material Laboratory Abaqus vs. Lab)
Properties (kN) (kN)**
(kN)*
5 450 BG 589 SR 617 SR 1.05
6 1626 BG 2115 SR 2156 SR 1.02
7 450 BG 597 SR 606 SR 1.02
8 579 BG 707 SR 757 SR 1.07
9 1626 BG 2141 SR 2122 SR 0.99
BG: Bolt Group Shear Failure
BS: Block Shear Failure
PFY: Plate Flexural Yielding
SR: Shear Rupture through Plate Net Area
* Calculated according to the AISC Manual 15th Ed. (2017)
** Recorded using the loading protocol from the laboratory tests. FE models could have reached
higher loads under larger displacements, but these would have been inaccurate since fracture
was not considered in the FE models.

84
4 Parametric Study

The validated FE modelling parameters described in Chapter 3 were used to model new
connection configurations which were not tested in the laboratory. The benefits of using finite
element analysis rather than full-scale testing are the reduction in cost and time, as well as the
ability to manipulate parameters with ease. For example the same numerical model can be run
using different loading protocols, whereas a full-scale model would need to be fabricated,
constructed and tested for each new run, thus requiring new materials for each test.

Upon research, analysis of the validation models, and consultation with industry professionals,
the extent of the parametric study was determined. It should be noted that there are infinite
configuration possibilities due to the numerous parameters which can be altered. These
parameters include, but are not limited to: column size, beam size, bolt size, number of bolts,
thickness of shear tab, a distance, and stabilizer plate details, etc. In order to draw some
conclusions and to provide tangible recommendations to practicing engineers, six main
parameters were investigated in this study. The scope includes the influence of the supporting
column, single- vs. double-sided connections, the effect of the type and connection pattern of
stabilizer plates, the addition of vertical rows of bolts, the shape of the shear plate, and the effect
of the loading protocol. The variables investigated as part of the parametric study are presented
in Table 4.1. The following identification names are used throughout this thesis to describe the
model configurations: BC-#1-#2-#3-L1-L2:

 BC: beam-to-column connection,


 #1: number of vertical rows of bolts,
 #2: number of horizontal rows of bolts,
 #3: thickness of the shear tab (mm),
 L1: type of connection:
o F: fully restrained (welded); such is the case for the stabilizer plates from
D’Aronco’s (2013) tests (Figure 4.17a),

85
o P: partial restraint; reduced width stabilizer plates which extend up to the k area
of the columns and therefore do not touch the column web (Figure 4.17b and c),
o R: rectangular shear tab (Figure 4.31),
o U: unstiffened (no stabilizer plates) (Figure 4.17e),
o P-Reduced: reduced width Type II stabilizer plates (Figure 4.17d),
 L2: (S) single- or (D) double-sided connection.

Table 4.1: Details of the Parametric Study

Section Parameter Details


W360x122
4.1 Column Size
W360x314
Single- vs. Double- Single
4.2
Sided Double
Type F
Type P
Type P with gaps
4.3 Stabilizer Plates
Type P-Reduced
Type P-Reduced with gaps
Type U
Vertical Rows of 2 to 3 rows bolts
4.4
Bolts 3 to 4 rows bolts
Traditional
4.5 Shear Tab Shape
Rectangular
2 Actuators
4.6 Loading Protocol
Uniformly Distributed

A survey of commonly used column and beam sizes in recent building construction in North
America was reviewed to determine the range of sizes to consider in the parametric study. The
beam and column combinations used in D’Aronco’s (2013) laboratory tests fell within the limits
of those in the survey and therefore were used in the parametric study.

86
Influence of the Column Size

An investigation into the effect of the column size (minor axis bending) was performed. Two
column sizes, W360x122 and W360x314, were studied in both shallow (3 rows of bolts) and deep
beam (6 rows of bolts) scenarios. The heavier column was nearly 7 times stiffer in the minor axis
than the small column (Ismall = 61.5x106 mm4 and Ilarge = 426x106 mm4). The larger column also
had a greater a distance since its width was larger, however the modified a distances (from the
toe of the stabilizer plates to the interior bolt line) were identical. Laboratory Tests #5 and #6
(D’Aronco, 2013) used the same smaller column (W360x122), and Tests #7 and #9 (D’Aronco,
2013) used the same larger column (W360x314).

4.1.1 Shallow Beam

Finite element models with identical parameters other than the column size were analysed. The
behaviour and connection capacity are presented in Figure 4.1. It is clear that the column rotation
was small compared to the beam rotation. Therefore the relative rotation between the beam
and the column, i.e. the connection rotation, is nearly identical to the absolute beam rotation.

Figure 4.1: Shear Force vs. Absolute Rotation Curve of FE Tests #5 and #7

87
In all the laboratory tests (#5 to #9), shear rupture through the net area of the plate was
observed. This was consistent with the failure mode observed in the FE models (Figure 4.3). The
numerical results suggest that the column had no influence on the connection capacity, according
to the shear force vs. absolute beam rotation curve (Figure 4.1). However, differences in the
deformed shape were observed. Test #7 suffered large out-of-plane displacements and
consequently, smaller vertical displacement of the shear tab (Figure 4.2).

a)

b)

Figure 4.2: FE Tests #5 and #7: a) Location of Recorded Measurements, b) Shear Force vs.
Vertical Displacement of Shear Tab, c) Out-of-Plane Displacement vs. Absolute Beam Rotation

88
c)

Figure 4.2 Cont.: FE Tests #5 and #7: a) Location of Recorded Measurements, b) Shear Force vs.
Vertical Displacement of Shear Tab, c) Out-of-Plane Displacement vs. Absolute Beam Rotation

The difference in deformed shape can be explained by analysing the stiffened region of the
connection, i.e. the region between the stabilizer plates. The large column had a distance from
the weld line to the edge of the flange of 188 mm, while the equivalent small column distance
was 122 mm; a difference of 66 mm (2.6”). Therefore, the stabilizer plate of the large column,
which was made of the same thickness plate as the small column, was wider and consequently
more flexible. This resulted in the top stabilizer plate of Test #7 deforming more than in Test #5.
Due to this deformation, the shear tab of Test #7 was compressed, causing it to buckle out-of-
plane (Figure 4.3b). Contrarily, Test #5 with the smaller column suffered more bolt hole
deformation and vertical displacement since its stabilizer plates were stiffer (Figure 4.3a).

89
a)

b)

Figure 4.3: Deformed Shape of Shear Tab from FE Models: a) Test #5 and b) Test #7

Next, the change in eccentricity with respect to the centreline of the bolt group was monitored.
The eccentricity was calculated from the FE models in two parts. First, the moment and shear at
the toe of the column flanges was monitored (location of M3 in Figure 4.5b). The eccentricity at
that location equals the moment divided by the shear. The distance from the weld line to the M3
location was then added to give the total connection eccentricity. According to the AISC Manual
15th Ed. (2017) design procedure, the eccentricity extends from the weld line to the centreline of
the bolt group. However, it can be seen in Figure 4.4 that the eccentricity fluctuates as a function

90
of the shear force applied. Note that, in Figure 4.4, the change in eccentricity represents the
movement of the inflection point with respect to the centreline of the bolt group.

Figure 4.4: FE Tests #5 and #7 Shear Force vs. Change in Eccentricity

The column size had little influence on the change in eccentricity. Nonetheless, when considering
the additional distance to the column’s web of Test #7, the resulting moment induced in the
column of Test #7 was 35 % higher than in Test #5 (Figure 4.5a). The moment in the column was
calculated from the FE models as the moment from the toe of the column flanges (location of M3
in Figure 4.5b) plus the shear at that location multiplied by the distance between the column
centreline and the M3 location. The column’s minor axis bending capacity was not reached in
either of the FE models. To continue, it can be deduced from Figure 4.5b that the majority of the
moment in the column is due to the eccentric shear load.

91
a)

b)

Figure 4.5: FE Tests #5 and #7: a) Shear Force vs. Moment in Column Web b) Moment vs.
Absolute Beam Rotation

92
4.1.2 Deep Beam

Again, two FE models with identical parameters, apart from the column size, were analysed.
These connections differ from the shallow beam models since they have 2 vertical rows of 6 bolts
and a shear tab thickness of 16 mm. The behaviour and connection capacity are presented in
Figure 4.6.

b = 257 b = 401
mm mm

Figure 4.6: Shear Force vs. Absolute Rotation Curve of FE Tests #6 and #9

Similar to Tests #5 and #7, the connection rotation is nearly identical to the absolute beam
rotation for Test #9. However, the column of Test #6 had a larger rotation, therefore reducing
the rotational demand on the connection. Both the experimental test and the numerical model
showed yielding in the column, consequently resulting in higher rotations (Figure 3.28).

Additionally, the deformed shapes are presented in Figure 4.7. Again, the shear tab with the
larger column suffered higher out-of-plane displacements due to the increased flexibility of the
wider stabilizer plates.

93
a)

b)

Figure 4.7: Deformed Shape of Shear Tab from FE Models: a) Test #6 and b) Test #9

To continue, Figure 4.8a illustrates the moment induced on the column’s web. As previously
mentioned, the column of Test #6 exceeded its plastic moment capacity, and therefore suffered
yielding and deformation. The column of Test #9 experienced bending moments 30% higher than
Test #6, but its minor axis bending capacity is 1231 kNm, two times higher than the moment to
which it was subjected. The change in eccentricity with respect to the centreline of the bolt group

94
is presented in Figure 4.8c. Minor differences existed between the shallow beam Tests #5 and #7
(Figure 4.4) which differs from the results of Tests #6 and #9. The larger column (Test #9) had
significantly higher change in eccentricity. This was mainly due to two reasons: the higher shear
forces reached with the deep beam connection and the larger lever arm. The distance between
the stabilizer plates in the deep beam scenario was significantly greater than in the shallow beam
connection, and therefore the resulting moments in the column web were greater. From Figure
4.8c, it can also be observed that the changes in eccentricity tended toward the same path as the
shear plate yielded.

a)

Figure 4.8: FE Tests #6 and #9: a) Shear Force vs. Column Web Moment b) Moment vs. Absolute
Beam Rotation c) Shear Force vs. Change in Eccentricity

95
b)

M web

M3

c)

Figure 4.8 Cont.: FE Tests #6 and #9: a) Shear Force vs. Column Web Moment b) Moment vs.
Absolute Beam Rotation c) Shear Force vs. Change in Eccentricity

96
Single- vs. Double-Sided Connections

Along the edges of a building, single-sided shear tab connections can be used to connect a beam
to a column web or flange. On the contrary, double-sided connections are typically found at
interior columns, where they are used to connect the column to beams on either side. Due to the
large eccentricity in an extended shear tab connection, a moment is introduced into the column.
In a double-sided scenario, both sides introduce an equal and opposite moment counteracting
one another, assuming the vertical beam loads are the same. Therefore the column is only
expected to play a significant role in the connection capacity and ductility of single-sided
connections. Note that a non-equilibrium state is possible if the beam size and/or gravity forces
and/or span length on either side of the column are different. For the purpose of this research,
the double-sided connections are assumed to be symmetric (Figure 4.9).

Figure 4.9: Free Body Diagram of Symmetric Double-Sided Connection

To compare both the single- and double-sided connections, models similar to Test #5 of
D’Aronco’s (2013) laboratory experiments was used. The single-sided connection consisted of a
W310x60 beam framing into the web of a W360x122 column. There were no stabilizer plates on
the backside of the column. To simulate the double-sided shear tab connection, rather than

97
modelling the entire setup, with a beam on either side of the column, one side was modelled
with boundary conditions to represent symmetry. The double-sided model included the column
sliced in half, lengthwise therefore leaving only half of the column web thickness and flanges
(Figure 4.10b). The backside of the column was then fixed in all degrees of freedom since it was
assumed that the symmetric connection on the backside will counteract any displacements or
moments induced into the column. By following the same logic, since the column was expected
to have minimal impact on the connection response, it was deemed unnecessary to model the
full column height. Therefore, the column height extended to the edges of the stabilizer plates
atop and below the shear tab. The FE model assembly is illustrated in Figure 4.10a.

a)

b)

Figure 4.10: a) Double-Sided FE Model Assembly b) Shear Tab with Half of Column Width

The connection shear force, absolute rotation, vertical displacement of the shear tab, and out-
of-plane displacement of the shear tab and the column web are presented in Figure 4.11. Both

98
the single- and double-sided connections are compared. Note that the out-of-plane
displacements of the shear tab and the column web are not in the same direction; they are
perpendicular to one another.

a)

b)

Figure 4.11: FE Models of Single- and Double-Sided Connections: a) Shear Force vs. Absolute
Rotation b) Shear Force vs. Vertical Displacement of the Shear Tab c) Out-of-Plane Displacement
vs. Absolute Beam Rotation

99
c)

Figure 4.11 Cont.: FE Models of Single- and Double-Sided Connections: a) Shear Force vs.
Absolute Rotation b) Shear Force vs. Vertical Displacement of the Shear Tab c) Out-of-Plane
Displacement vs. Absolute Beam Rotation

The shear force vs. absolute beam rotation curves resulted in a nearly bilinear response. The
double-sided connection exhibited a slightly stiffer response, however the overall behaviour and
strength were very similar between the two models. Further investigation proved that the
double-sided connection experienced slightly higher rotation as compared to the equivalent
single-sided connection. This difference was due to the influence of the column. The column of
the single-sided connection rotated about its minor axis and therefore reduced the rotational
demand on the connection itself. However, the column’s rotation remained small in comparison
to the beam (Figure 4.11a).

The shear tab out-of-plane displacement vs. beam rotation plot demonstrated a loss in stiffness
as the slope changed, at relatively low rotation (Figure 4.11c). This loss was linked to the elastic
buckling of the shear plate. Similarly, the column web deformation increased significantly beyond

100
this bifurcation point for the single-sided connection. Finally, the column web of the double-sided
connection was not expected to deform because the counteracting moments on the backside are
presumed to restrict the out-of-plane deformation. The double-sided model showed no
deformation in the column web, as predicted.

Figure 4.12a illustrates the movement of the point of inflection. The double-sided connection
had a larger eccentricity since it had a higher stiffness. The bending moment in both single- and
double-sided connections followed the same response (Figure 4.12b and c). However, the
moment induced in the column of the single-sided connection developed after that of the
double-sided connection. This suggests that the flexibility of the column slowed the development
of moments. Overall, the maximum moment reached the same value of 119 kNm for both
models.

a)

Figure 4.12: FE Models of Single- and Double-Sided Connections: a) Shear Force vs. Effective
Eccentricity b) Shear Force vs. Moment in Column c) Moment vs. Absolute Beam Rotation

101
b)

c)

Figure 4.12 Cont.: FE Models of Single- and Double-Sided Connections: a) Shear Force vs.
Effective Eccentricity b) Shear Force vs. Moment in Column c) Moment vs. Absolute Beam
Rotation

102
Furthermore, the progression of yielding for both the single- and double-sided connections was
identical. Yielding initiated at the re-entrant corners of the shear tab and propagated to the
interior line of bolts and then finally toward the stiffened regions of the connection (stabilizer
plates). This can be seen in Figure 4.13 and Figure 4.14. This failure mode was also observed in
Motallebi’s (2018) finite element research on single- and double-sided connections.

a) b)

Figure 4.13: Single-Sided Connection a) Yield Initiation at Re-Entrant Corners b) Yield


Propagation along Interior Line of Bolts (gray represents yielded regions)

a) b)

Figure 4.14: Double-Sided Connection a) Yield Initiation at Re-Entrant Corners b) Yield


Propagation along Interior Line of Bolts (gray represents yielded regions)

103
Furthermore, column yielding initiated after net section failure through the interior line of bolts
of the shear tab. For the single-sided connection, since the column web was free to displace out-
of-plane, this displacement caused the column web to yield (Figure 4.15).

a) b)

Figure 4.15: Single-Sided Connection a) Yield Initiation in Column Web at Out-of-Plane


Displacement of 0.819 mm b) Yield Propagation in Column Web at Out-of-Plane Displacement of
1.86 mm

For the double-sided connection, yielding in the column web initiated as the shear tab’s net
section yielded. The column web was restrained from moving out-of-plane, and so the shear
force caused yielding in the column web (Figure 4.16). Therefore, the column did not contribute
to the net section failure mechanism. This differed from the behaviour of the single-sided
connection which had the column yield due to out-of-plane displacements.

104
a) b)

Figure 4.16: Double-Sided Connection a) Yield Initiation in Column Web at Beam Rotation of
0.01245 rad b) Yield Propagation (Top View)

Effect of the Type and Connection Pattern of Stabilizer Plates

Researchers Fortney and Thornton (2016) made a recommendation to use reduced width
stabilizer plates which do not extend to the column web. Currently, no physical or numerical tests
have been performed on these reduced width stabilizer plates. Therefore, to verify the adequacy
of this type of plate, FE models of the connection types shown in Figure 4.17 were created. Note,
regardless of the stabilization provided directly to the shear tab plate, all models were laterally
supported along the length of the beam. These types and connection patterns were investigated
for both a shallow and a deep beam connection.

105
Figure 4.17: Stabilizer Plate Types a) Fully Welded (Traditional), b) Reduced Width Welded,
c)Reduced Width with Gaps, d) Further Reduced Width, e) Unstiffened, f) Beam Laterally
Supported

4.3.1 Shallow Beam

The models presented herein connect a W310x60 beam to a W360x314 column by means of 2
vertical rows of 3 bolts with a 10 mm thick shear tab. To begin, the behaviours of the connections
are compared in Figure 4.18. The plots were separated into three categories: welded, gap and
unstiffened (beam laterally braced for all). The results showed negligible difference between the
welded configurations; including both the fully welded stabilizer plates (to web and flanges
(Figure 4.17a)) and those welded only to the column flanges (Figure 4.17b and d). This suggested
that the weld to the column web was not necessary to attain the same shear capacity and
rotation. Secondly, the connections with a gap between the stabilizer plates and the column
flanges (Figure 4.17c) resulted in lower shear capacity as compared to the welded cases. This was
expected since the connection behaved as if it were unstiffened up until the point when the
stabilizer plates came into contact with the column. Varying the gap size, from 1/64” to 1/8” on
either side of the stabilizer plates, did not affect the connection capacity. The width of the

106
stabilizer plates was further reduced by 2*k of the column (Figure 4.17d); a model with welds to
the column flanges and models with gaps between the stabilizer plates and the column flanges
were investigated with this reduced width. Finally, the unstiffened model resulted in the lowest
shear capacity. Note that the column rotations for each connection were very small (less than 2%
of the absolute beam rotations) and therefore are not shown in Figure 4.18.

Figure 4.18: Shear Force vs. Absolute Beam Rotation of Shallow Beam Connections with
Different Stabilizer Plate Details

It should also be noted that the factored design capacity calculated according to the AISC Manual
15th Ed. (2017) was greatly exceeded by all the connection configurations. By performing the
factored design calculations using to the modified eccentricity (toe of the stabilizer plate to the
bolt group centreline) recommended by Fortney and Thornton (2016), the prediction was much
closer to the actual capacity of the shear tab connections with stabilizer plates. Note, all design
calculations were performed with probable material properties.

107
Furthermore, the vertical displacement was highest in the unstiffened connection since it was
free to deform (Figure 4.19b). The connections with the gaps acted similar to the unstiffened
connection until bearing between the column and the stabilizer plates initiated. At this point, the
stabilizer plates became engaged and resisted some of the vertical and out-of-plane
displacement.

a)
Location of FBC 3

Vertical
Displacement
Out-of-Plane
Displacement

b)

Figure 4.19: Shallow Beam Connections with Different Stabilizer Plate Details: a) Location of
Recorded Measurements, b) Shear Force vs. Vertical Displacement of Shear Tab, c) Out-of-Plane
Displacement vs. Absolute Beam Rotation

108
c)

Figure 4.19 Cont.: Shallow Beam Connections with Different Stabilizer Plate Details: a) Location
of Recorded Measurements, b) Shear Force vs. Vertical Displacement of Shear Tab, c) Out-of-
Plane Displacement vs. Absolute Beam Rotation

From Figure 4.19c, it is clear that the unstiffened connection buckled out-of-plane (Figure 4.20a).
This behaviour was restricted by the stabilizer plates in the other models (Figure 4.20b and c). As
well, in the models with the gaps, a change in the slope of the out-of-plane displacement curves
occurred when the stabilizer plates came into contact with the column. This represented the
point when the stabilizer plates added stiffness to the connection and reduced the out-of-plane
displacement.

109
a) b)

c)

Figure 4.20: Out-of-Plane Displacement of a) BC-2-3-10-U-S, b) BC-2-3-10-P-S-b, c) BC-2-3-10-P-


S-b 1/8” gap

The movement of the point of inflection confirmed that the moment induced in the column was
larger in the welded models compared to the models with gaps and the unstiffened model (Figure
4.21a). Logically it would seem that the eccentricity of the welded connection would be smaller
since it can be assumed to be taken from the toe of its stabilizer plate, however the opposite was
true. Since the welded configurations added stiffness to the connection, they attracted more
forces and consequently transferred higher moments to the column (Figure 4.21b). Therefore,
the addition of gaps between the stabilizer plates and the column flanges proved to benefit the
supporting column by reducing the moments applied to it.

110
a)

b)

Figure 4.21: Shallow Beam Connections with Different Stabilizer Plate Details: a) Shear Force vs.
Effective Eccentricity, b) Shear Force vs. Moment in Column, c) Moment in Column vs. Absolute
Beam Rotation

111
c)

Figure 4.21 Cont.: Shallow Beam Connections with Different Stabilizer Plate Details: a) Shear
Force vs. Effective Eccentricity, b) Shear Force vs. Moment in Column, c) Moment in Column vs.
Absolute Beam Rotation

4.3.2 Deep Beam

The models presented herein connect a W610x140 beam to a W360x122 column by means of 2
vertical rows of 6 bolts with a 16 mm thick shear tab. To begin, the behaviours of the connections
are compared in Figure 4.22. Similar to the shallow beam scenario, the results showed negligible
difference between the welded configurations; these include both the fully welded stabilizer
plates (Figure 4.17a) and those welded only to the column flanges (Figure 4.17b). This again
suggested that the weld to the column web was not necessary to attain the same shear capacity
and rotation.

112
Figure 4.22: Shear Force vs. Absolute Beam Rotation of Deep Beam Connections with Different
Stabilizer Plate Details

The connection capacity of the unstiffened model (Figure 4.17e) was lower compared to the
welded models (Figure 4.17a, b). This was due to the increased flexibility of the connection having
no stabilizer plates. However, the model with a gap between the stabilizer plates and the column
flanges (Figure 4.17c) resulted in a similar shear capacity to the unstiffened connection model.
These results were different from the shallow beam connection results, in which the models with
flange gaps had significantly higher capacity compared to the unstiffened model. To explain this
difference, the out-of-plane displacement plots were analysed (Figure 4.23b).

113
a)

b)

Figure 4.23: Deep Beam Connections with Different Stabilizer Plate Details: a) Shear Force vs.
Vertical Displacement of Shear Tab b) Out-of-Plane Displacement vs. Absolute Beam Rotation

It was expected that the unstiffened connection would buckle out-of-plane, similar to its
response in the shallow beam connection model (Figure 4.23c); however, this was not observed
for the deep beam scenario. The small column combined with the deep beam resulted in high

114
shear capacities but caused damage to the column. The shear tab punched through the column’s
web rather than buckling out-of-plane, without substantially reducing the shear capacity of the
connection. This behaviour was observed in all four models presented in this section, with more
pronunciation in the unstiffened and gap between stabilizer plates and column flanges models.

a)

b)

Figure 4.24: Column Web Punching a) BC-2-6-16-U-S-a, b) BC-2-6-16-P-S-a (flange removed for
clarity). U1 represents horizontal displacement.

115
To continue, the movement of the inflection point was monitored (Figure 4.25a). The welded
configurations had higher eccentricities due to their increased stiffness. This consequently
resulted in higher moments in the column web (Figure 4.25b, c). As well, this proved that the
eccentricity prediction by the AISC Manual (2017) was conservative for unstiffened connections.

a)

b)

Figure 4.25: Deep Beam Connections with Different Stabilizer Plate Details: a) Shear Force vs.
Effective Eccentricity b) Shear Force vs. Moment in Column c) Moment in Column vs. Absolute
Beam Rotation

116
c)

Figure 4.25 Cont.: Deep Beam Connections with Different Stabilizer Plate Details: a) Shear Force
vs. Effective Eccentricity b) Shear Force vs. Moment in Column c) Moment in Column vs.
Absolute Beam Rotation

Additional Vertical Rows of Bolts

In practice, some building designs require a shallow floor beam, therefore if additional bolts are
needed, this can only be realised by adding more vertical rows. This may be necessary to achieve
an adequate connection resistance according to the AISC Manual 15 th Ed. (2017) design
procedure. However, the efficacy of adding vertical rows of bolts is questionable. Since the AISC
Manual (2017) requires the bolt group to be designed for a conservative eccentricity extending
from the support to the centreline of the bolt group (e), a large moment is assumed to act on the
bolts. This often results in the primary failure mode being wrongly predicted as bolt shear.
Therefore, additional bolts are required to increase the capacity of the bolt group. In reality, bolt
shear is rarely observed as the governing failure mode since the eccentricity is not as high as the
prediction from the AISC Manual 15th Ed. (2017).

117
A comparison between the results of validation Tests #7 and #8 (D’Aronco, 2013) is presented in
Figure 4.26. Test #7 had two vertical rows of three bolts, and Test #8 had three vertical and
horizontal rows of bolts. Both connections show little difference in their behaviour, but Test #8
increased the connection resistance by 20% compared to Test #7 with the same a distance.
However, the shear tab thickness was also larger, therefore the full increase in strength may not
be attributed to these additional three bolts. Further investigation into the addition of vertical
rows of bolts is necessary.

Figure 4.26: Comparison of Shear vs. Absolute Beam Rotation of Validation Tests #7 and #8

To verify if additional vertical rows of bolts are effective, four models were created; BC-3-3-10-F-
S which was identical to BC-2-3-10-F-S (Test #5) with an added third vertical row of bolts and BC-
4-3-13-F-S which was identical to BC-3-3-13-F-S (Test #8) with an added fourth vertical row of
bolts (Figure 4.27).

118
a) b)

c) d)

Figure 4.27: Bolt Hole Configurations: a) BC-2-3-10-F-S, b) BC-3-3-10-F-S, c) BC-3-3-13-F-S, d) BC-


4-3-13-F-S

First, the shear force vs. absolute rotation plots are compared in Figure 4.28. Both plots show
little benefit to adding a vertical row of bolts when considering the shear capacity. Note, the
factored design capacities were calculated using probable material properties.

119
a)

b)

Figure 4.28: Shear Force vs. Absolute Rotations for a change from a) 2 to 3 Vertical Rows of
Bolts and b) 3 to 4 Vertical Rows of Bolts

The movement of the effective eccentricity shows some offset in the models changing from two
to three vertical rows of bolts and no difference when shifting from three to four vertical rows of
120
bolts (Figure 4.29). This difference was attributed to the slightly lower stiffness of the BC-3-3-10-
F-S model, allowing the shear tab to displace vertically at lower forces as compared to the BC-2-
3-10-F-S model.

a)

b)

Figure 4.29: Shear Force vs. Effective Eccentricity for a change from a) 2 to 3 Vertical Rows of
Bolts and b) 3 to 4 Vertical Rows of Bolts

121
Other properties, including the vertical displacement, out-of-plane displacement, and the
moment in the column web, were compared to demonstrate the negligible effect of additional
vertical rows of bolts (Figure 4.30).

a)

b)

Figure 4.30: Additional Vertical Rows of Bolts: a) Shear Force vs. Shear Tab Vertical
Displacement, b) Shear Tab Out-of-Plane Displacement vs. Absolute Beam Rotation, c) Moment
in Column vs. Absolute Beam Rotation

122
c)

Figure 4.30 Cont.: Additional Vertical Rows of Bolts: a) Shear Force vs. Shear Tab Vertical
Displacement, b) Shear Tab Out-of-Plane Displacement vs. Absolute Beam Rotation, c) Moment
in Column vs. Absolute Beam Rotation

Achieving a higher connection capacity by adding a vertical row of bolts proved to be inefficient.
Alternatively, having a better prediction of the connection’s shear resistance would not require
additional bolts. As previously explained, the bolt group must currently be designed for a large
moment according to the AISC Manual design procedure (2017). This overestimates the actual
force to which it is subjected. If the eccentricity were better predicted for stiffened connections,
the shear tab’s capacity could be calculated more accurately, and the additional vertical rows of
bolts would likely be unnecessary.

Effect of the Shape of the Shear Tab

Traditionally, shear tabs were designed with extensions to align the stabilizer plates with the
beam’s flanges (Figure 4.31). These extensions may still be required if a moment connection
exists in the major axis of the column. Otherwise, the shear tab is typically designed as a
rectangular plate to ease the fabrication procedure (Figure 4.31). It is presumed that no real

123
advantage is achieved by using the extensions. To verify this assumption, models with and
without extensions were considered in both a shallow and a deep beam connection.

a)

b)

Figure 4.31: Traditional Shear Tab Configuration with Extensions vs. Rectangular Shear Tab a)
Shallow Beam, b) Deep Beam

The shear force and absolute rotation curves are shown in Figure 4.32. They demonstrate nearly
identical results. Note, the factored design calculations were performed using probable material
properties. Investigation into the vertical displacement, out-of-plane displacement, effective
eccentricity and moment further confirm the assumption that the two shapes of stabilizer plates
studied result in similar behaviours and identical capacities (Figure 4.33).

124
a)

b)

Figure 4.32: Shear Force vs. Absolute Beam Rotation for a) Shallow Beam, b) Deep Beam;
Comparison of traditional and rectangular shear tab configurations

125
a)

b)

c)

Figure 4.33: Comparison of Traditional and Rectangular Shear Tab Configurations: a) Shear
Force vs. Vertical Displacement b) Out-of-Plane Displacement vs. Absolute Beam Rotation c)
Shear Force vs. Effective Eccentricity d) Moment vs. Absolute Beam Rotation

126
d)

Figure 4.33 Cont.: Comparison of Traditional and Rectangular Shear Tab Configurations: a)
Shear Force vs. Vertical Displacement b) Out-of-Plane Displacement vs. Absolute Beam Rotation
c) Shear Force vs. Effective Eccentricity d) Moment vs. Absolute Beam Rotation

The results are similar and show that the shape of the shear tab did not influence the connection
behaviour. The only difference was noticed in the out-of-plane displacement of the shear tab. In
the rectangular plates, the ductility of the connection was provided by vertical movement and
deformation in the stabilizer plates. In the traditional shear tab models, the stabilizer plates and
stiffened region of the shear tab experienced minimal displacement, therefore the ductility came
mainly from bolt hole deformations. Figure 4.34 illustrates these differences. In addition, the out-
of-plane displacements in the traditional shear tab connections were higher. This change was
due to the vertical displacements. Since the stabilizer plates did not deform as significantly as
they did in the rectangular connections, the shear tab was compressed and therefore
experienced more out-of-plane buckling. Despite this difference, the assumption that the two
shapes of stabilizer plates studied result in similar behaviours and identical capacities is correct.

127
a) b)

Figure 4.34: Vertical Displacement of Shear Tab a) Traditional b) Rectangular

Loading Protocol

The loading protocol has a substantial effect on the connection response. This was proven by
Motallebi (2018) in a study on beam-to-girder extended shear tab connections. He investigated
the effect of the beam size, L/d ratio, and girder length and boundary conditions. Different ratios
of main to tip actuator displacements applied to a specimen will result in different combinations
of shear, moment, and rotation to the connection. Astaneh (1988) developed a test setup which
included a test beam and two actuators. The purpose of the two actuators was to achieve the
same beam end rotation and shear on the connection as would be applied under a uniform
gravity load. Therefore, the testing protocol used by D’Aronco (2013) in the laboratory comprised
two actuators to simulate these displacements. The actuator at the beam tip was controlled
based on the relative rotation of the beam. Ideally, it should have been controlled based on the
absolute beam rotation to remove the effect of the column. Therefore, as the testing program
continued at McGill University, future students performed their laboratory testing by controlling
the tip actuator based on the absolute beam rotation. For this reason, the loading protocol used
in D’Aronco’s laboratory tests was used with caution in the FE models.

128
The general behaviour of the loading protocol used in the laboratory was achieved by targeting
an absolute beam rotation of 0.02 rad at the calculated probable ultimate resistance of the shear
tab (D’Aronco’s loading protocol was based on relative rotation as mentioned previously). The
ratio between the tip and the main actuator rates was adjusted continually until the target
rotation was reached. Then, the ratio of the actuator displacement rates was held constant for
the remainder of the test.

To determine the connection response in a real building and to verify the ultimate capacities
reached in the laboratory and in the FE models, a uniform gravity load was applied to the FE
models. This was done using surface traction in Abaqus. Surface traction applies a vertical
uniformly distributed gravity load, regardless of the beam deflection and deformation. The force
in the connection can be limited by the resistance of the beam at mid-span. By modelling the
assembly with a gravity load (surface traction), the beam may fail first, and as such, the true
potential resistance of the connection would not be observed. In the laboratory tests and in the
previous FE models, this was not problem because the failure of the beam at mid-span was not
simulated (two displacement controlled actuators were used). Thus, two beam sizes were
investigated: W310x60 and W310x97. The selected beam sizes were calculated to withstand a
bending moment according to the maximum shear capacity of the connection, predicted by the
AISC Manual (2017) design procedure and by the AISC with modified eccentricity according to
Fortney and Thornton (2016), respectively. A third model with the W310x60 beam was modelled
with the beam having only elastic material properties. This model was assumed to reach the full
capacity of the connection since it did not depend on plastic hinging of the beam. The results of
the three surface traction models are compared to the results using D’Aronco’s (2013) two
displacement controlled actuators in Figure 4.35. Note that a beam length to depth ratio, L/d, of
17 was used in the surface traction models, as suggested by Astaneh (1989) and Motallebi (2018).

129
Figure 4.35: Shear Force vs. Absolute Beam Rotation of FE Models of Test #5 with Various
Loading Protocols

From Figure 4.35, the difference between the heavier W310x97 and the lighter W310x60 beam
is evident. The heavier beam reached values much closer to those observed by D’Aronco (2013).
The lighter beam with elastic material properties also provided a good estimate of the ultimate
connection capacity. Therefore, it is clear that the connection response was highly dependent on
the other parameters of the assembly. Further studies on the influence of the loading protocol
are required.

Recommendations for Design

The information obtained from Sections 4.1 to 4.6 was used to provide recommendations for the
design of shear tab connections. The recommendations are as follows.

- Where stabilizer plates are used, but not required because of an orthogonal moment
frame, they can be designed with a reduced width, therefore eliminating the weld to the

130
column web. With the stabilizer plates welded to the column flanges, two reduced
widths were studied: one extending to the k-area of the column (stabilizer plate reduced
by k, Figure 4.17b), and a second model reducing the stabilizer plate width by 2*k of the
column (Figure 4.17d). Should larger width reductions be used, their behaviour should
be verified. The studied width reductions resulted in negligible change to the behaviour
and capacity of the connection, compared to the traditional fully welded stabilizer plates
(Figure 4.17a). Thus, the connection capacity is significantly higher than the AISC Manual
15th Ed. (2017) design capacity and exceeds the design capacity with modified
eccentricity according to Fortney and Thornton (2016) as well.
- Reduced width stabilizer plates with gaps between them and the column flanges (Figure
4.17c) reach the AISC Manual 15th Ed. (2017) design capacity and that with modified
eccentricity according to Fortney and Thornton (2016), only at high rotations. They
behave as unstiffened connections until bearing, between the stabilizer plates and the
column, increase their capacity. Therefore, it may not be adequate to design them
according to the modified eccentricity.
- Column web punching shall be considered in the design of the supporting column for
unstiffened connections and those with gaps between the stabilizer plates and the
column flanges.
- Beam-to-column web stiffened extended shear tab connections were found to behave
similar to beam-to-column flange extended shear tab connections when considering the
modified eccentricity, from the toe of the stabilizer plates. This further reinforces the
suggestion by Fortney and Thornton (2016) stating that the eccentricity shall be taken
from the toe of the stabilizer plates to the centreline of the bolt group. The stabilizer
plates adequately stiffened the shear plate in the region between the column flanges.
- Using the modified eccentricity according to Fortney and Thronton (2016), the design
capacity for stiffened shear tab connections was much closer to the actual connection
capacity reached in the laboratory tests and in the FE models. This may eliminate the
need for additional bolts to satisfy the required design load of a given connection. It was
proven that the addition of vertical rows of bolts was not an efficient way of increasing

131
the capacity of the connection, thus by designing with the modified eccentricity, the
actual connection capacity could be predicted more accurately.
- The moment induced on the column should be considered in the design of the column.
The AISC Manual 15th Ed. (2017) has a new section compared to previous manuals which
accounts for the weak-axis bending in the column due to the eccentric beam end
reaction. The recommendations are still preliminary. Estimates of the moment are
presented in Figure 4.36a and b; calculations were performed using the maximum shear
recorded from the FE models multiplied by the a distance and the e distance,
respectively. Therefore, the moment induced in the column’s minor axis should be
calculated as the beam reaction (shear force) multiplied by an eccentricity, whether a or
e. Using the eccentricity, a, the moment was fairly well predicted but some moments
were underestimated, whereas the eccentricity, e, gave a conservative estimate of the
moment induced on the column in all cases.

a)

Figure 4.36: Calculated Moment Predictions Induced on the Column; a) Using a Distance
(support to interior bolt line) b) Using e Distance (support to centreline of bolt group)

132
b)

Figure 4.36 Cont.: Calculated Moment Predictions Induced on the Column; a) Using a Distance
(support to interior bolt line) b) Using e Distance (support to centreline of bolt group)

133
5 Conclusions

Summary

Finite element models were used to analyse the behaviour and capacity of beam-to-column web
extended shear tab connections. The FE modelling assumptions were first calibrated to
laboratory data from full-scale tests performed by D’Aronco (2013). Then, a parametric study was
conducted to investigate the influence of the supporting column, single- vs. double-sided
connections, the effect of the type and connection pattern of stabilizer plates, the addition of
vertical rows of bolts, the shape of the shear plate, and the effect of the loading protocol.

As a result of this study, the behaviour of extended shear tab connections with a flexible support
is better understood. The parametric study demonstrated that the column size and the shape of
the shear tab did not alter the connection capacity. In a deep beam scenario, when high shear
forces are reached, the column should be carefully selected to account for the high moments
induced on it.

A comparison between single- and double-sided connections demonstrated that the supporting
column in the single-sided connection rotated with the beam, therefore reducing the rotational
demand on the shear tab. Otherwise, little differences among the two models were observed.
However, the column yielding was induced by two different means. The column of the single-
sided connection yielded due to out-of-plane displacements in the column web, while the column
yielding in the double-sided configuration was due to the applied shear force.

To increase the capacity of the connection, adding a vertical row of bolts was proven ineffective.
According to the AISC Manual 15th Ed. (2017), more vertical rows of bolts should increase the
connection capacity. To address this current issue in the design procedure, a better prediction of

134
the connection’s shear resistance is needed. This can be achieved by using a more accurate
eccentricity value, thus enabling engineers to predict the capacity and failure mode of stiffened
shear tab connections more correctly.

Also, the welds connecting the stabilizer plates to the column web are not necessary if the plates
are welded to the column flanges. These reduced width stabilizer plates proved to give the same
results as the traditional, fully welded stabilizer plates (within the width reductions studied; k of
the column and 2*k). Therefore, a reduction in costs and labour is possible by using reduced
width stabilizer plates. Additionally, introducing a gap between the stabilizer plates and the
column flanges can benefit the column by reducing the moments applied to it. This connection
type behaves initially as an unstiffened connection and therefore reaches high capacities only at
high rotations.

The results of this study reinforce the recommendation made by Fortney and Thornton (2016),
to use a modified eccentricity, extending from the toe of the stabilizer plates to the centreline of
the bolt group for stiffened extended shear tab connections. However, the moment induced in
the column should be accounted for in the design of the supporting column.

Recommendations for Future Work

Some of the unknown parameters of extended shear tab connections were investigated in this
thesis. However, there remain numerous aspects yet to be studied. Mainly, with respect to the
reduced width stabilizer plates, the thickness and width limit of the plates should be explored
further. Additionally, all models presented in this thesis included either top and bottom stabilizer
plates or no stabilizer plates. The effect of having either top or bottom plates should be
considered, to potentially reduce the moment induced on the column.

135
Also, in many buildings, a concrete slab is installed atop the steel beam. This may affect the
rotational ductility and shear capacity of the connection. However, the shear connectivity of the
concrete to the beam is of concern; composite action can only be relied on if they are fully shear
connected. It is common practice in Canada not to use shear studs in residential buildings,
therefore the slab would add a dead load which stiffens the beam, but could not be relied upon
to transfer forces. The models considered in this research comprised of beams with full lateral
support, but no consideration of the concrete slab was included. Consequently, the shear tab
was indirectly supported laterally by the beam. Should the building use steel grating for example,
this lateral stability would not be present; its effect should be studied.

Additionally, the effect of adding a vertical row of bolts in a stiffened extended shear tab
connection proved to provide no benefit to its capacity. An unstiffened connection may not result
in the same conclusion, due to the potential of experiencing a different failure mode. Further
investigation of unstiffened connections with various numbers of vertical rows of bolts is
necessary.

Another aspect to investigate would be the influence of the welds. The weld material properties
would need to be known to determine if they play a role in the transfer of loads. Damage
initiation and propagation in the FE models may also be factors to include in future models for
yielding more accurate results.

136
6 References

ABAQUS 6.18-1 (2018). [Computer software]. Dassault Systemes Simulia Corp., Providence, RI,
USA.

AISC 360-05 (2005). Specification for Structural Steel Buildings. American Institute of Steel
Construction, Chicago, IL, 2005.

AISC 360-10 (2010). Specification for Structural Steel Buildings. American Institute of Steel
Construction, Chicago, IL, 2010.

AISC 360-16 (2016). Specification for Structural Steel Buildings. American Institute of Steel
Construction, Chicago, IL, 2016.

AISC (2017). Design Examples: Companion to the AISC Steel Construction Manual. American
Institute of Steel Construction, Chicago, IL. Version 15.0.

AISC (2005). Steel Construction Manual, 13th edition. American Institute of Steel Construction,
Chicago, IL.

AISC (2011). Steel Construction Manual, 14th edition. American Institute of Steel Construction,
Chicago, IL.

137
AISC (2017). Steel Construction Manual, 15th edition. American Institute of Steel Construction,
Chicago, IL.

AISC (2010). Code of Standard Practice for Steel Buildings and Bridges. American Institute for
Steel Construction, Chicago, IL.

Al Hijaj, M., Mahamid, M. (2016). Behaviour of Skewed Extended Shear Tab Connections Part I:
Connection to Supporting Web. Journal of Constructional Steel Research 128, pp. 305-320. Civil
and Materials Engineering Department, University of Illinois, Chicago, US.

Ashakul, A. (2004). Finite Element Analysis of Single Plate Shear Connections. Ph.D Dissertation,
Virginia Polytechnic Institute and State University. Blacksburg.

Astaneh, A. (1989). Demand and Supply of Ductility in Steel Shear Connections. Journal of
Constructional Steel Research, 14, pp. 1-19.

Astaneh, A., Call, S.M., McMullin, K.M. (1989). Design of Single Plate Shear Connections. American
Institute of Steel Construction – Engineering Journal, First Quarter, Volume 26, 21-32.

Astaneh, A., McMullin, K. M., and Call, S. M. (1988). Design of Single-Plate Framing Connections.
Report No. UCB/SEMM-88/12. Department of Civil Engineering, University of California, Berkeley,
CA.

138
Astaneh-Asl, A., McMullin, K.M., and Call, S.M. (1993). Behaviour and Design of Steel Single Plate
Shear Connections. Journal of Structural Engineering, ASCE, Volume 119, No. 8, pp. 2421-2440.

ASTM (2003) Standard Specification for General Requirements for Rolled Structural Steel Bars,
Plates, Shapes, and Sheet Piling. American Society for Testing and Materials, West
Conshohocken, PA.

AutoCAD (2019). [Computer software]. Autodesk. San Rafael, CA, USA.

Caccavale, S.E. (1975). Ductility of Single Plate Framing Connections. MSc Thesis, University of
Arizona. Tucson.

Carter, C.J., Muir, L.S. and Dowswell, B. (2016). Establishing and Developing the Weak-Axis
Strength of Plates Subjected to Applied Loads. American Institute of Steel Construction –
Engineering Journal, Vol. 53, No.3, pp. 147-157.

Cheng, J. J., Yura. J. A., and Johnson, C.P. (1984). Design and Behavior of Coped Beams. Ferguson
Structural Engineering Laboratory Report No. 84-1. Dept. of Civil Engineering, University of Texas
at Austin: Austin, TX, USA.

CISC (2010). Handbook of Steel Construction, 10th Edition. Canadian Institute of Steel
Construction: Markham, ON, Canada.

139
CISC (2015). Handbook of Steel Construction, 11th Edition. Canadian Institute of Steel
Construction: Markham, ON, Canada.

Craig, Roy R. Jr. (2011). Mechanics of Materials. John Wiley & Sons, 3rd ed. USA.

Creech, D. D. (2005). Behaviour of Single Plate Shear Connections with Rigid and Flexible Supports.
M.Sc. Thesis, North Carolina State University, USA.

CSA-G40.20-13/G40.21-13 (2013). General Requirements for Rolled or Welded Structural Quality


Steel. Canadian Standards Association, Toronto, ON.

Daneshvar, H. (2013). One-Sided Steel Shear Connections in Column Removal Scenario. Ph.D.
Thesis, University of Alberta, Edmonton, AB, Canada.

D’Aronco, M. (2014). Behaviour of Double and Triple Vertical Rows of Bolts Shear Tab Connections
and Weld Retrofits. MAsc Thesis, Dept. of Civil Engineering, École Polytechnique, Montreal, QC,
Canada.

Dassault Systemes, 2018. SIMULIA User Assistance.

Dowswell, B. and Whyte, R. (2014). Local Stability of Double-Coped Beams. American Institute of
Steel Construction – Engineering Journal, 1st Quarter, 43-51.

140
Dowswell, B. (2016). Stability of Rectangular Connection Elements. American Institute of Steel
Construction – Engineering Journal, 4th Quarter 2016, Vol. 53, No. 4, 171-202.

Elkady, A. (2016). Collapse Risk Assessment of Steel Moment Resisting Frames Designed with
Deep Wide-Flange Columns in Seismic Regions. Ph.D Dissertation, McGill University, Montreal,
QC, Canada.

El-Khoriby, S. et al. (2016). Modelling and Behaviour of Beam-to-Column Connections under Axial
Force and Cyclic Bending. Journal of Constructional Research 129, pp. 171-184.

Faridmehr, I., Tahir, M., Lahmer, T. (2016). Classification System for Semi-Rigid Beam-to-Column
Connections. Latin American Journal of Solids and Structures 13. pp. 2152-2175.

Fortney, P.J. and Thornton, W.A. (2016). Analysis and Design of Stabilizer Plates in Single-Plate
Shear Connections. American Institute of Steel Construction – Engineering Journal, Vol. 53, No.1,
pp. 1-28.

Goldstein Apt, N. (2015). Testing of Extended Shear Tab and Coped Beam-to-Girder Connections
Subject to Shear Loading. M.Sc. Thesis, McGill University, Montreal, QC, Canada.

Goodrich, W. S. (2005). Behavior of Extended Shear Tabs in Stiffened Beam-to-Column Web


Connections. MSc Thesis, Vanderbilt University. Nashville, TN.

141
Hertz, J. (2014). Testing of Extended Shear Tab Configurations Subjected to Shear. MEng Thesis,
Dept. of Civil Engineering and Applied Mechanics, McGill University, Montreal, QC, Canada.

Holzer, S.M., Melosh, R.J., Barker, R.M., Somers, A.E. (1975). Singer: A Computer Code for General
Analysis of Two-Dimensional Concrete Structures. Air Force Weapons Laboratory, Dept. of Civil
Engineering. Virginia Polytechnic Institute and State University, Blacksburg, Virginia. Volume I.

Kline, D., Rojiani, K., and Murray, T. M. (1995). Performance of Snug-Tight Bolts in Moment End-
Plate Connections. Virginia Tech Structural Engineering Report No. CE/VPI-ST-89/04, Blacksburg,
VA.

Kong, Z., Kim, S.E. (2017). Numerical estimation for initial stiffness and ultimate moment of T-stub
connections. Journal of Constructional Research 141, pp. 118-131.

Kulak, G.L., Fisher, J.W., Struik, H.A. (2001). Guide to Design Criteria for Bolted and Riveted Joints.
2nd Edition American Institute of Steel Construction, AISC. Chicago, IL.

Lipson, S. L. (1968). Single Angle and Single Plate Framing Connections. Canadian Structural
Engineering Conference Proceedings, 1968, 139-162.

Marosi, M. (2011). Behaviour of Single and Double Row Bolted Shear Tab Connections and Weld
Retrofits. MEng Thesis, Dept. of Civil Engineering and Applied Mechanics, McGill University:
Montreal, QC, Canada.

142
Marshall, J. (2011). Finite Element (FE) Modeling of a Welded Retrofit Shear Tab Connection.
MEng Thesis, Dept. of Civil Engineering and Applied Mechanics, McGill University: Montreal, QC,
Canada.

Muir, L.S., Thornton, W.A. (2004). A Direct Method for obtaining the Plate Buckling Coefficient for
Double-Coped Beams. A Technical Note, Engineering Journal, AISC 41, pp. 133-134.

Muir, L.S., Hewitt, C.M. (2009). Design of Unstiffened Extended Single-Plate Shear Connections.
American Institute of Steel Construction – Engineering Journal, 46 (2).

Muir, L.S. and Thornton, W.A. (2011). The Development of a New Design Procedure for
Conventional Single-Plate Shear Connections. American Institute of Steel Construction –
Engineering Journal, Vol. 48, No.2, pp. 141-152.

Muir, L.S. (2017). A Shear Connection Extends its Reach. American Institute of Steel Construction
- Modern Steel Construction Magazine. December edition.

Motallebi, M. (2018). Behaviour of Extended Shear Tab Connections under Combined Axial and
Shear Forces. Ph.D Dissertation, McGill University, Montreal, QC, Canada.

Rahman, A., et al. (2005). 3D FE Model of Extended Shear Tab Steel Connections. University of
Wisconsin-Milwaukee, Dept. of Civil Engineering and Mechanics. Milwaukee, WI.

143
Rahman, A., Mahamid, M., Amro, A., Ghorbanpoor, A. (2007). The Analyses of Extended Shear
Tab Steel Connections Part I: the Unstiffened Connections. American Institute of Steel
Construction – Engineering Journal, 44 (2).

Richard, R., Gillett, P., Kriegh, J., and Lewis, A. (1980). The Analysis and Design of Single Plate
Framing Connections. American Institute of Steel Construction – Engineering Journal, 2nd
Quarter, 38-52.

Ricles, J.M. (1980). The Behaviour and Analysis of Double Row Bolted Shear Web Connections.
M.Sc. Thesis, University of Texas. Austin, TX.

SAP 2000 Documentation. (2008). Technical Note - Material Stress-Strain Curves. Berkeley,
California: Computers and Structures, Inc.

SAP 2000 (2019). [Computer software]. Computers and Structures, Inc. (CSI), CA, USA.

Sherman, D. R., Ghorbanpoor, A. (2002). Design of Extended Shear Tabs. University of Wisconsin-
Milwaukee, Milwaukee, WI.

Stiemer, S.F., Wong, H.H., and Ho, A. (1986). Ultimate Capacity of Single Plate Connectors.
Proceedings of the Pacific Structural Steel Conference, 117-132. Auckland, NZ.

144
Suleiman, M.F., Shahrooz, B.M., Bill, H.L., Fortney, P.J., Thornton, W.A. (2017). 3-D Finite Element
Modeling of Extended Single Plate Shear Connections: Predicting the Mode of Failure.
International Journal of Steel Structures, 17(2) 525-534.

Tamboli, A.R. (2016). Handbook of Structural Steel Connection Design and Details. Third edition,
McGraw-Hill, New York, NY.

Thornton, W.A. and Fortney, P.J. (2011). On the Need for Stiffeners and the Effect of Lap
Eccentricity on Extended Shear Tabs. American Institute of Steel Construction – Engineering
Journal, 48 (2), pp.117-125.

White, R.N. (1965). Framing Connections for Square and Rectangular Structural Tubing. American
Institute of Steel Construction – Engineering Journal, Vol. 2, No. 3, Third Quarter, pp. 94-102.

145
Appendix A: Design Calculations

Design calculations performed according to AISC Manual 15th Edition (2017) – Extended
Configuration with eccentricity modified according to Fortney and Thornton (2016).

146
TEST #5 D'Aronco
Beam W310x60
Column W360x122

Plate and Beam Properties


Plate Thickness, tp 9.23 mm 0.363 in
Thickness of Beam Web, tw 7.76 mm 0.306 in
Measured Yield Stress of Plate, Fy,plate 478 MPa 69 ksi
Measured Tensile Strength of Plate, Fu,plate 548 MPa 79 ksi
Measured Yield Stress of Beam, Fy,beam 399 MPa 58 ksi
Measured Tensile Strength of Beam, Fu,beam 502 MPa 73 ksi

Bolt Properties
Bolt Grade A325
Threads in Shear Plane? NO
Diameter of Bolts, db 19.05 mm 0.75 in
Diameter of Holes, dh 20.64 mm 0.81 in
Number of Rows of Bolts, n 3 3
Number of Columns of Bolts, m 2 2

Eccentricity (Modified Method)


a Distance 50.8 mm 2 in
Eccentricity, e 88.9 mm 3.5 in
Eccentric Loading Coefficient, C 3.37 3.37 * AISC Table 7-6 to 7-13
ICR Coefficient, C' 401.3 mm 15.8 in * AISC Table 7-6 to 7-13

Design Check 1
Shear Strength, φrn 100 kN 22.5 kips * AISC Table 7-1
Predicted Shear Strength, φrn 133 kN 30.0 kips *φ=1
Predicted Shear Strength, φrn (eccentric) 450 kN 101.1 kips Bolt Shear (One Bolt)
Predicted Bearing Strength, rn 211 MPa/bolt 47.4 kips/bolt * φ = 1
Predicted Bearing Strength, φrn (eccentric) 710 kN 159.6 kips Bolt Bearing on Shear Plate
Available Strength 87.8 kips/in * AISC Table 7-4
Predicted Bearing Strength, φrn 119 kN 26.8 kips
Predicted Bearing Strength, φrn (eccentric) 402 kN 90.4 kips Bolt Bearing on Beam Web

Design Check 2
Mmax 59.6 kNm 527 kip-in * AISC eq. 10-4
Measured Maximum Thickness of Plate, tmax 14.3 mm 0.564 in * AISC eq. 10-3

147
Design Check 3 (Plate)
Gross area subject to shear, Agv 2110 mm2 3.27 in2
Predicted Shear Yielding Resistance, Vsy 605 kN 136 kips * AISC Eq. J4-3
Net Area of Plate, ANV 1537 mm2 2.32 in2 * AISC B4.3b
Predicted Shear Rupture Resistance, VsR 505 kN 110 kip * AISC Eq. J4-4
Shear Area, Agv 1758 mm2 2.73 in2
Net Area in Shear, Anv 1281 mm2 1.93 in2
Net Area in Tension, Ant 747 mm2 1.16 in2
Predicted Block Shear Resistance, Rn 614 kN 138 kip * AISC Section J4.3

Design Check 3 (Beam)


Gross area subject to shear, Agv 2351 mm2 3.64 in2
Probable Shear Yielding Resistance, Vsy 562 kN 126 kips * AISC Eq. J4-3
Net Area of Beam, ANV 1871 mm2 2.90 in2
Predicted Shear Rupture Resistance, VsR 563 kN 127 kip * AISC Eq. J4-4

Design Check 4
Probable Shear Capacity, Vpc 605 kN 136 kip
Plastic Section Modulus, Zpl 120585 mm3 7.36 in3
Probable Moment Capacity, Mpc 57.6 kNm 510 kip-in
Probable Shear and Flexural Yielding Resistance, Vr 533 kN 120 kip * AISC eq. 10-5

Design Check 5
Lateral-Torsional Buckling of Plate
Lateral-Torsional Buckling Modification Factor, Cb 1.84 1.84 * AISC eq. 9-15
Lbd/t^2 136
Elastic Section Modulus, Sx 80390 mm3 4.91 in3
Plastic Section Modulus, Zx 120585 mm3 7.36 in3
Yield Moment Capacity, My (measured) 38.4 kNm 340 kip-in
Plastic Moment Capacity, Mp (measured) 57.6 kNm 510 kip-in
Predicted Nominal Flexural Strength, Mn (measured) 57.6 kNm 510 kip-in * AISC Spec. eq. F11-2
Flexural Rupture of Plate
Znet 90948 mm3 5.55 in3 * AISC Table 15-3
Mn 49.8 kNm 441 kip-in * AISC eq. 9-4

Eccentric Weld to Column Web


Length of Weld, L 245 mm 9.64 in
Horizontal Eccentricity of Load, ex 89 mm 3.5 in
a 0.363 0.363
Eccentric Loading Coefficient, C 2.82 2.82 * AISC Table 8-4
Probable Strength of Weld Group, φRn 578 kN 130 kip

Strength of Column Web at Weld


Minimum Column Web Thickness, tmin 4.5 mm 0.179 in * AISC Eq. 9-2

148
TEST #6 D'Aronco
Beam W610x140
Column W360x122

Plate and Beam Properties


Plate Thickness, tp 15.71 mm 0.619 in
Thickness of Beam Web, tw 13.75 mm 0.541 in
Measured Yield Stress of Plate, Fy,plate 497 MPa 72 ksi
Measured Tensile Strength of Plate, Fu,plate 574 MPa 83 ksi
Measured Yield Stress of Beam, Fy,beam 409 MPa 59 ksi
Measured Tensile Strength of Beam, Fu,beam 510 MPa 74 ksi

Bolt Properties
Bolt Grade A325
Threads in Shear Plane? NO
Diameter of Bolts, db 22.23 mm 0.88 in
Diameter of Holes, dh 23.81 mm 0.94 in
Number of Rows of Bolts, n 6 6
Number of Columns of Bolts, m 2 2

Eccentricity (Modified Method)


a Distance 63.5 mm 2.5 in
Eccentricity, e 101.6 mm 4 in
Eccentric Loading Coefficient, C 8.93 8.93 * AISC Table 7-6 to 7-13
ICR Coefficient, C' 1376.7 mm 54.2 in * AISC Table 7-6 to 7-13

Design Check 1
Shear Strength, φrn 137 kN 30.7 kips * AISC Table 7-1
Predicted Shear Strength, φrn 182 kN 40.9 kips *φ=1
Predicted Shear Strength, φrn (eccentric) 1626 kN 365.5 kips Bolt Shear (One Bolt)
Predicted Bearing Strength, rn 354 MPa/bolt 79.7 kips/bolt * φ = 1
Predicted Bearing Strength, φrn (eccentric) 3166 kN 711.7 kips Bolt Bearing on Shear Plate
Available Strength 102.0 kips/in * AISC Table 7-4
Predicted Bearing Strength, φrn 246 kN 55.2 kips
Predicted Bearing Strength, φrn (eccentric) 2193 kN 493.1 kips Bolt Bearing on Beam Web

Design Check 2
Mmax 278.4 kNm 2463 kip-in * AISC eq. 10-4
Measured Maximum Thickness of Plate, tmax 16.1 mm 0.633 in * AISC eq. 10-3

149
Design Check 3 (Plate)
Gross area subject to shear, Agv 7183 mm2 11.13 in2
Predicted Shear Yielding Resistance, Vsy 2141 kN 481 kips * AISC Eq. J4-3
Net Area of Plate, ANV 4788 mm2 7.42 in2 * AISC B4.3b
Predicted Shear Rupture Resistance, VsR 1650 kN 371 kip * AISC Eq. J4-4
Shear Area, Agv 6584 mm2 10.21 in2
Net Area in Shear, Anv 4521 mm2 6.80 in2
Net Area in Tension, Ant 1197 mm2 1.86 in2
Predicted Block Shear Resistance, Rn 1856 kN 417 kip * AISC Section J4.3

Design Check 3 (Beam)


Gross area subject to shear, Agv 8484 mm2 13.15 in2
Probable Shear Yielding Resistance, Vsy 2080 kN 468 kips * AISC Eq. J4-3
Net Area of Beam, ANV 6519 mm2 10.10 in2
Predicted Shear Rupture Resistance, VsR 1995 kN 449 kip * AISC Eq. J4-4

Design Check 4
Probable Shear Capacity, Vpc 2141 kN 481 kip
Plastic Section Modulus, Zpl 820973 mm3 50.10 in3
Probable Moment Capacity, Mpc 407.9 kNm 3611 kip-in
Probable Shear and Flexural Yielding Resistance, Vr 2032 kN 457 kip * AISC eq. 10-5

Design Check 5
Lateral-Torsional Buckling of Plate
Lateral-Torsional Buckling Modification Factor, Cb 1.84 1.84 * AISC eq. 9-15
Lbd/t^2 118
Elastic Section Modulus, Sx 547315 mm3 33.40 in3
Plastic Section Modulus, Zx 820973 mm3 50.10 in3
Yield Moment Capacity, My (measured) 272.0 kNm 2407 kip-in
Plastic Moment Capacity, Mp (measured) 407.9 kNm 3611 kip-in
Predicted Nominal Flexural Strength, Mn (measured) 407.9 kNm 3611 kip-in * AISC Spec. eq. F11-2
Flexural Rupture of Plate
Znet 553883 mm3 33.8 in3 * AISC Table 15-3
Mn 318.1 kNm 2815 kip-in * AISC eq. 9-4

Eccentric Weld to Column Web


Length of Weld, L 548 mm 21.57 in
Horizontal Eccentricity of Load, ex 102 mm 4 in
a 0.185 0.185
Eccentric Loading Coefficient, C 3.56 3.56 * AISC Table 8-4
Probable Strength of Weld Group, φRn 2991 kN 672 kip

Strength of Column Web at Weld


Minimum Column Web Thickness, tmin 8.3 mm 0.328 in * AISC Eq. 9-2

150
TEST #7 D'Aronco
Beam W310x60
Column W360x314

Plate and Beam Properties


Plate Thickness, tp 9.23 mm 0.363 in
Thickness of Beam Web, tw 7.76 mm 0.306 in
Measured Yield Stress of Plate, Fy,plate 478 MPa 69 ksi
Measured Tensile Strength of Plate, Fu,plate 548 MPa 79 ksi
Measured Yield Stress of Beam, Fy,beam 399 MPa 58 ksi
Measured Tensile Strength of Beam, Fu,beam 502 MPa 73 ksi

Bolt Properties
Bolt Grade A325
Threads in Shear Plane? NO
Diameter of Bolts, db 19.05 mm 0.75 in
Diameter of Holes, dh 20.64 mm 0.81 in
Number of Rows of Bolts, n 3 3
Number of Columns of Bolts, m 2 2

Eccentricity (Modified Method)


a Distance 50.8 mm 2 in
Eccentricity, e 88.9 mm 3.5 in
Eccentric Loading Coefficient, C 3.37 3.37 * AISC Table 7-6 to 7-13
ICR Coefficient, C' 401.3 mm 15.8 in * AISC Table 7-6 to 7-13

Design Check 1
Shear Strength, φrn 100 kN 22.5 kips * AISC Table 7-1
Predicted Shear Strength, φrn 133 kN 30.0 kips *φ=1
Predicted Shear Strength, φrn (eccentric) 450 kN 101.1 kips Bolt Shear (One Bolt)
Predicted Bearing Strength, rn 211 MPa/bolt 47.4 kips/bolt * φ = 1
Predicted Bearing Strength, φrn (eccentric) 710 kN 159.6 kips Bolt Bearing on Shear Plate
Available Strength 87.8 kips/in * AISC Table 7-4
Predicted Bearing Strength, φrn 119 kN 26.8 kips
Predicted Bearing Strength, φrn (eccentric) 402 kN 90.4 kips Bolt Bearing on Beam Web

Design Check 2
Mmax 59.6 kNm 527 kip-in * AISC eq. 10-4
Measured Maximum Thickness of Plate, tmax 14.3 mm 0.564 in * AISC eq. 10-3

151
Design Check 3 (Plate)
Gross area subject to shear, Agv 2110 mm2 3.27 in2
Predicted Shear Yielding Resistance, Vsy 605 kN 136 kips * AISC Eq. J4-3
Net Area of Plate, ANV 1495 mm2 2.32 in2 * AISC B4.3b
Predicted Shear Rupture Resistance, VsR 491 kN 110 kip * AISC Eq. J4-4
Shear Area, Agv 1758 mm2 2.73 in2
Net Area in Shear, Anv 1281 mm2 1.93 in2
Net Area in Tension, Ant 747 mm2 1.16 in2
Predicted Block Shear Resistance, Rn 614 kN 138 kip * AISC Section J4.3

Design Check 3 (Beam)


Gross area subject to shear, Agv 2351 mm2 3.64 in2
Probable Shear Yielding Resistance, Vsy 562 kN 126 kips * AISC Eq. J4-3
Net Area of Beam, ANV 1871 mm2 2.90 in2
Predicted Shear Rupture Resistance, VsR 563 kN 127 kip * AISC Eq. J4-4

Design Check 4
Probable Shear Capacity, Vpc 605 kN 136 kip
Plastic Section Modulus, Zpl 120585 mm3 7.36 in3
Probable Moment Capacity, Mpc 57.6 kNm 510 kip-in
Probable Shear and Flexural Yielding Resistance, Vr 533 kN 120 kip * AISC eq. 10-5

Design Check 5
Lateral-Torsional Buckling of Plate
Lateral-Torsional Buckling Modification Factor, Cb 1.84 1.84 * AISC eq. 9-15
Lbd/t^2 136
Elastic Section Modulus, Sx 80390 mm3 4.91 in3
Plastic Section Modulus, Zx 120585 mm3 7.36 in3
Yield Moment Capacity, My (measured) 38.4 kNm 340 kip-in
Plastic Moment Capacity, Mp (measured) 57.6 kNm 510 kip-in
Predicted Nominal Flexural Strength, Mn (measured) 57.6 kNm 510 kip-in * AISC Spec. eq. F11-2
Flexural Rupture of Plate
Znet 90948 mm3 5.55 in3 * AISC Table 15-3
Mn 49.8 kNm 441 kip-in * AISC eq. 9-4

Eccentric Weld to Column Web


Length of Weld, L 245 mm 9.64 in
Horizontal Eccentricity of Load, ex 89 mm 3.5 in
a 0.363 0.363
Eccentric Loading Coefficient, C 2.82 2.82 * AISC Table 8-4
Probable Strength of Weld Group, φRn 578 kN 130 kip

Strength of Column Web at Weld


Minimum Column Web Thickness, tmin 4.5 mm 0.179 in * AISC Eq. 9-2

152
TEST #8 D'Aronco
Beam W310x74
Column W360x314

Plate and Beam Properties


Plate Thickness, tp 12.47 mm 0.491 in
Thickness of Beam Web, tw 9.71 mm 0.382 in
Measured Yield Stress of Plate, Fy,plate 424 MPa 61 ksi
Measured Tensile Strength of Plate, Fu,plate 492 MPa 71 ksi
Measured Yield Stress of Beam, Fy,beam 389 MPa 56 ksi
Measured Tensile Strength of Beam, Fu,beam 487 MPa 71 ksi

Bolt Properties
Bolt Grade A325
Threads in Shear Plane? NO
Diameter of Bolts, db 19.05 mm 0.75 in
Diameter of Holes, dh 20.64 mm 0.81 in
Number of Rows of Bolts, n 3 3
Number of Columns of Bolts, m 3 3

Eccentricity (Modified Method)


a Distance 50.8 mm 2 in
Eccentricity, e 127 mm 5 in
Eccentric Loading Coefficient, C 4.34 4.34 * AISC Table 7-6 to 7-13
ICR Coefficient, C' 711.2 mm 28 in * AISC Table 7-6 to 7-13

Design Check 1
Shear Strength, φrn 100 kN 22.5 kips * AISC Table 7-1
Predicted Shear Strength, φrn 133 kN 30.0 kips *φ=1
Predicted Shear Strength, φrn (eccentric) 579 kN 130.2 kips Bolt Shear (One Bolt)
Predicted Bearing Strength, rn 255 MPa/bolt 57.4 kips/bolt * φ = 1
Predicted Bearing Strength, φrn (eccentric) 1109 kN 249.2 kips Bolt Bearing on Shear Plate
Available Strength 87.8 kips/in * AISC Table 7-4
Predicted Bearing Strength, φrn 149 kN 33.6 kips
Predicted Bearing Strength, φrn (eccentric) 648 kN 145.7 kips Bolt Bearing on Beam Web

Design Check 2
Mmax 105.6 kNm 935 kip-in * AISC eq. 10-4
Measured Maximum Thickness of Plate, tmax 28.6 mm 1.127 in * AISC eq. 10-3

153
Design Check 3 (Plate)
Gross area subject to shear, Agv 2376 mm2 3.68 in2
Predicted Shear Yielding Resistance, Vsy 604 kN 136 kips * AISC Eq. J4-3
Net Area of Plate, ANV 2019 mm2 3.13 in2 * AISC B4.3b
Predicted Shear Rupture Resistance, VsR 596 kN 134 kip * AISC Eq. J4-4
Shear Area, Agv 2376 mm2 3.68 in2
Net Area in Shear, Anv 1730 mm2 2.61 in2
Net Area in Tension, Ant 1683 mm2 2.61 in2
Predicted Block Shear Resistance, Rn 910 kN 205 kip * AISC Section J4.3

Design Check 3 (Beam)


Gross area subject to shear, Agv 3010 mm2 4.67 in2
Probable Shear Yielding Resistance, Vsy 703 kN 158 kips * AISC Eq. J4-3
Net Area of Beam, ANV 2409 mm2 3.73 in2
Predicted Shear Rupture Resistance, VsR 703 kN 158 kip * AISC Eq. J4-4

Design Check 4
Probable Shear Capacity, Vpc 725 kN 163 kip
Plastic Section Modulus, Zpl 162914 mm3 9.94 in3
Probable Moment Capacity, Mpc 69.0 kNm 611 kip-in
Probable Shear and Flexural Yielding Resistance, Vr 639 kN 144 kip * AISC eq. 10-5

Design Check 5
Lateral-Torsional Buckling of Plate
Lateral-Torsional Buckling Modification Factor, Cb 1.84 1.84 * AISC eq. 9-15
Lbd/t^2 75
Elastic Section Modulus, Sx 108609 mm3 6.63 in3
Plastic Section Modulus, Zx 162914 mm3 9.94 in3
Yield Moment Capacity, My (measured) 46.0 kNm 407 kip-in
Plastic Moment Capacity, Mp (measured) 69.0 kNm 611 kip-in
Predicted Nominal Flexural Strength, Mn (measured) 69.0 kNm 611 kip-in * AISC Spec. eq. F11-2
Flexural Rupture of Plate
Znet 121264 mm3 7.40 in3 * AISC Table 15-3
Mn 59.6 kNm 528 kip-in * AISC eq. 9-4

Eccentric Weld to Column Web


Length of Weld, L 246 mm 9.68 in
Horizontal Eccentricity of Load, ex 127 mm 5 in
a 0.516 0.516
Eccentric Loading Coefficient, C 2.25 2.25 * AISC Table 8-4
Probable Strength of Weld Group, φRn 618 kN 139 kip

Strength of Column Web at Weld


Minimum Column Web Thickness, tmin 6.1 mm 0.239 in * AISC Eq. 9-2

154
TEST #9 D'Aronco
Beam W610x140
Column W360x314

Plate and Beam Properties


Plate Thickness, tp 15.71 mm 0.619 in
Thickness of Beam Web, tw 12.81 mm 0.504 in
Measured Yield Stress of Plate, Fy,plate 497 MPa 72 ksi
Measured Tensile Strength of Plate, Fu,plate 574 MPa 83 ksi
Measured Yield Stress of Beam, Fy,beam 390 MPa 57 ksi
Measured Tensile Strength of Beam, Fu,beam 500 MPa 73 ksi

Bolt Properties
Bolt Grade A325
Threads in Shear Plane? NO
Diameter of Bolts, db 22.23 mm 0.88 in
Diameter of Holes, dh 23.81 mm 0.94 in
Number of Rows of Bolts, n 6 6
Number of Columns of Bolts, m 2 2

Eccentricity (Modified Method)


a Distance 63.5 mm 2.5 in
Eccentricity, e 101.6 mm 4 in
Eccentric Loading Coefficient, C 8.93 8.93 * AISC Table 7-6 to 7-13
ICR Coefficient, C' 1376.7 mm 54.2 in * AISC Table 7-6 to 7-13

Design Check 1
Shear Strength, φrn 137 kN 30.7 kips * AISC Table 7-1
Predicted Shear Strength, φrn 182 kN 40.9 kips *φ=1
Predicted Shear Strength, φrn (eccentric) 1626 kN 365.5 kips Bolt Shear (One Bolt)
Predicted Bearing Strength, rn 354 MPa/bolt 79.7 kips/bolt * φ = 1
Predicted Bearing Strength, φrn (eccentric) 3166 kN 711.7 kips Bolt Bearing on Shear Plate
Available Strength 102.0 kips/in * AISC Table 7-4
Predicted Bearing Strength, φrn 229 kN 51.4 kips
Predicted Bearing Strength, φrn (eccentric) 2043 kN 459.4 kips Bolt Bearing on Beam Web

Design Check 2
Mmax 278.4 kNm 2463 kip-in * AISC eq. 10-4
Measured Maximum Thickness of Plate, tmax 16.1 mm 0.633 in * AISC eq. 10-3

155
Design Check 3 (Plate)
Gross area subject to shear, Agv 7183 mm2 11.13 in2
Predicted Shear Yielding Resistance, Vsy 2141 kN 481 kips * AISC Eq. J4-3
Net Area of Plate, ANV 4788 mm2 7.42 in2 * AISC B4.3b
Predicted Shear Rupture Resistance, VsR 1650 kN 371 kip * AISC Eq. J4-4
Shear Area, Agv 6584 mm2 10.21 in2
Net Area in Shear, Anv 4521 mm2 6.80 in2
Net Area in Tension, Ant 1197 mm2 1.86 in2
Predicted Block Shear Resistance, Rn 1856 kN 417 kip * AISC Section J4.3

Design Check 3 (Beam)


Gross area subject to shear, Agv 7904 mm2 12.25 in2
Probable Shear Yielding Resistance, Vsy 1848 kN 415 kips * AISC Eq. J4-3
Net Area of Beam, ANV 6074 mm2 9.41 in2
Predicted Shear Rupture Resistance, VsR 1822 kN 410 kip * AISC Eq. J4-4

Design Check 4
Probable Shear Capacity, Vpc 2141 kN 481 kip
Plastic Section Modulus, Zpl 820973 mm3 50.10 in3
Probable Moment Capacity, Mpc 407.9 kNm 3611 kip-in
Probable Shear and Flexural Yielding Resistance, Vr 2032 kN 457 kip * AISC eq. 10-5

Design Check 5
Lateral-Torsional Buckling of Plate
Lateral-Torsional Buckling Modification Factor, Cb 1.84 1.84 * AISC eq. 9-15
Lbd/t^2 118
Elastic Section Modulus, Sx 547315 mm3 33.40 in3
Plastic Section Modulus, Zx 820973 mm3 50.10 in3
Yield Moment Capacity, My (measured) 272.0 kNm 2407 kip-in
Plastic Moment Capacity, Mp (measured) 407.9 kNm 3611 kip-in
Predicted Nominal Flexural Strength, Mn (measured) 407.9 kNm 3611 kip-in * AISC Spec. eq. F11-2
Flexural Rupture of Plate
Znet 553883 mm3 33.80 in3 * AISC Table 15-3
Mn 318.1 kNm 2815 kip-in * AISC eq. 9-4

Eccentric Weld to Column Web


Length of Weld, L 547 mm 21.53 in
Horizontal Eccentricity of Load, ex 102 mm 4 in
a 0.186 0.186
Eccentric Loading Coefficient, C 3.56 3.56 * AISC Table 8-4
Probable Strength of Weld Group, φRn 2985 kN 671 kip

Strength of Column Web at Weld


Minimum Column Web Thickness, tmin 8.3 mm 0.328 in * AISC Eq. 9-2

156
Appendix B: Results of Validation Models

The FE results of the validation models from Tests #7, #8 and #9 are compared to the
experimental results here; their results are similar to Tests #5 and #6.

157
Test #7

Figure B.1: Experimental vs. FE Abaqus Model Results for Test #7

158
Figure B.2: Deformed Shape of Shear Tab from Test #7

Figure B.3: Deformed Shape of Beam from Test #7

159
Test #8

Figure B.4: Experimental vs. FE Abaqus Model Results for Test #8

160
Figure B.5: Deformed Shape of Shear Tab from Test #8

161
Test #9

Figure B.6: Experimental vs. FE Abaqus Model Results for Test #9

Figure B.7: Deformed Shape of Shear Tab from Test #9

162

You might also like