cyclic behavior of steel gusset plate connections.pdf

8/19/2019 cyclic behavior of steel gusset plate connections.pdf

1/165

Structural Engineering Report No. 191

University of AlbertaDepartment of Civil &Environmental Engineering

Cyclic Behavior of SteeGusset Plate Connections

by

Jeffrey S. Rabinovitch

and

J.J. Roger Cheng


2/165

Structural Engineering Report No 191

CYCLIC

BEHAVIOR OF

STEEL GUSSET PLATE

CONNECTIONS

by

Jeffrey

S

Rabinoviteh

and

J J Roger Cheng

Department of Civil Engineering

University

of Alberta

Edmonton Alberta


3/165


4/165

connections appear capable absorbing significant amounts ener

proposed strong braced weak gusset concentric braced frame model.


5/165

KNOWLEDGEMENTS

This report is a reprint

o

a thesis by the same name written by the ftrst aut

supervision o the second author

Financial support for the project was provided

by

the Natural Sciences and

Research Council

o

Canada to Dr J J Cheng under grant No 4727 and

Research Fund o the University o Alberta Financial support was provide

author by the Natural Sciences and Engineering Research Council

o

Canad

Walter

Johns Graduate Fellowship

The assistance o the technical staff o the I F Morrison Structural Labo

University

o

Alberta is acknowledged

A special gratitude o thanks goes to Michael Yam for his helpful advice and s


6/165

T LE OF CONTENTS

hapter

INTRODUCTION

General

1 2 Seismic Design Considerations

1 3 Energy Absorption in Braced Frames

1 4 Current Gusset Plate Design

1 5 Objectives and Scope

2 LITERATURE REVIEW

2 1 Introduction

2 2 Early Gusset Plate Research

2 3 Monotonic Gusset Plate Behavior

2 4 Cyclic Gusset Plate Behavior

3 EXPERIMENTAL PROGRAM

3 1 Introduction

3 2 Preliminary Considerations

3 3 Specimen Description

3 4 Test Set up

3 5 Instrumentation


7/165

4 3 2 Specimen A 2

4 3 3 Specimen A 3

4 3 4 Specimen A 4

4 3 5 Specimen A 5

5 DISCUSSION OF TEST RESULTS

5 1 Introduction

5 2 Parameters Affecting Energy Absorption

5 2 1 Gusset Plate Thickness

5 2 2 Plate Edge Stiffeners

5 2 3 Gusset Plate Geometry

5 2 4 Connection Bolt Slip

5 3 Comparison Test Results

5 3 1 Monotonic Tension

5 3 2 Monotonic Compression

5 3 3 Cyclic Load Behavior

5 4 Finite Element Analysis

6 SUMMARY AND DESIGN GUIDELINE

6 1 Summary

6 2 Design Guideline for Gusset Plates Under Cyclic Loads

6 3 Recommendations for Future Research

REFERENCES


8/165

LIST OF TABLES

able

3 Specimen Description

3 2 Slip Resistance of Gusset Plate to Splice Member Connection

4 1 Material Properties

4 2 Ultimate Specimen Capacities

5 1 Comparison ofTest Results Tension

5 2 Comparison of Test Results Compression


9/165

LIST

O

IGUR S

igure

Basic Configurations of Concentric

Braced

Frames

1.2

Hysteresis

Loops

- Concentric Braced

Frame

Wakabayashi, et al. 1974

1.3

Eccentric Braced Frame - Link Segment

1.4

Typical Experimental

Frame

Behavior

under

Cyclic Lateral Load

Popov

and

Engelhardt,

1988

2.1

Whitmore Gusset Plate Prototype

2.2 Whitmore Effective Width

2.3 Thornton Equivalent Column Method

2.4

Block Shear Tear-outModel

2.5

Gusset Plate Studied

by

Astaneh-Asl, et

al.

1981

3.1

Typical Test SpecimenGeometry - Specimen A-I to A-4

3.2

Geometry of Specimen

A-5

3.3 Simulation of Boundary Conditions

3.4 Schematic of Test Set-up

3.5 Test Set-up Reactions

3.6

Tension Reaction Frame

3.7 Test Set-up - Tension Reaction

Frame

Removed

3.8 Typical Strain Gage Locations - Specimen A-I to A-4


10/165

igure

3 13

Test Frame Instrumentation

3 14

LVDT Support Frame

4 1

Typical Material Response

4 2

Load

vs

Axial Displacement Response of the Gusset Plate

Assembly Specimen A I

4 3

Final Specimen Yield Line Pattern Specimen A I

4 4

Load vs Axial Displacement

Response

Specimen A I

4 5

Plate Fracture in

Bolted

Connection Specimen A I

4 6

Out of plane Displacements Specimen A I

4 7 Buckled Long Free Edge Specimen A I

4 8 Strain Distribution Specimen A I

4 9 Load vs Axial Displacement Response of the Gusset Plate

Assembly Specimen

A 2

4 10 Final Specimen Yield

Line

Pattern Specimen

A 2

4 11

Load

vs

Axial Displacement Response Specimen A 2

4 12 Out of plane Displacements Specimen

A 2

4 13 Buckled Plate Free Edges Specimen A 2

4 14 Strain Distribution Specimen A 2

4 15 Load vs

Axial

Displacement Response of the Gusset Plate

Assembly Specimen

A 3


11/165


12/165


13/165

INTRO U TION

General

Concentric braced frames are one

of

the most common lateral load-resistin

steel buildings. n a concentric braced frame the lateral loads applied to the

resisted by a network of inclined bracing members. Depending on the config

braced frame, either tensile

or

compressive loads can be accommodated b

members. These loads are commonly transferred to the beam and column m

frame by gusset plate connections. The gusset plate receives the load from

bracing member and transfers it to the main framing members. The delivery

and

out

of

the gusset plate

will

produce bending, shear, and normal forces

plate. Some common configurations of concentric braced frames are shown

When a structure is subjected to reverse lateral load conditions, the bracing

the gusset plate connections can be subject to both tensile and comp

conditions.

1.2 Seismic Design onsiderations

When a steel structure is required to resist seismic load conditions, the d

concentric braced frame is governed in Canada y the National Building Cod

1990 NRCC, 1990 and the provisions ofCAN/CSA-SI6.1 - Limit States De

Structures CSA, 1989 . The National Building Code outlines the p

determining the minimum lateral seismic force that is to applied to the str


14/165


15/165

test behavior the frame specimen heavy dotted line and an analytical pre

frame response light dotted line .

Eccentric braced frames absorb energy

in

an entirely different manner th

braced frames. The distinguishing characteristic an eccentric braced fra

least one end every brace is connected so that the brace force

is

transmi

another brace or

to

a column through shear and bending

in

a beam segmen

Figure 1.3 . Inelastic activity under severe cyclic loading is restricted pri

links, which are designed and detailed

to

sustain large inelastic deformations

strength. n contrast to concentric braced frames, the braces are des

buckle, regardless the severity lateral loading on the frame. Because b

is

prevented and because the

link

can sustain large deformations without stre

and stable hysteretic loops similar

to

those moment resisting frames

Popov and Engelhardt, 1988 .

Figure 1.4, reprinted from Popov and Engelhardt 1988 , provides a crude

between the hysteresis behavior

a moment resisting frame MRF . a conc

frame CBF , and an eccentric braced frame EBF . The observed difference

behavior

is

recognized

by

the National Building Code

Canada NRCC,

favorable energy absorption behavior

ductile moment resisting frames

force reduction factor R =4.0. Eccentric braced frames are designed usin


16/165

can be confmed to an element designed and detailed to sustain large ine

without significant loss

strength. The current research attempts to

frame model to be examined considers a concentric braced frame wher

is designed not to buckle. The energy of the system s designed to

gusset plate connection. The gusset plate

s

designed to yield n tensi

compression. n compression, a stable post-buckling behavior s des

plate

an improved energy absorption behavior

s

to be obtained. It

s

gusset plate element under severe cyclic loads that s the focus

this in

4 Current Gusset Plate Design

Design specifications n North America currently provide very little

design

steel gusset plates. Generally, only design philosophy s

specific formulas for evaluating the dimension and thickness

a guss

the design

gusset plates often draws upon the experience

the struc

When gusset plates are designed to resist strictly monotonic tensio

CAN/CSA-S6-88 - Design

Highway Bridges CSA, 1988 states onl

shall be

ample thickness to resist shear, direct load, and flexure, act

r

critical section. A simple design equation

s

provided to determine

resistance

the gross area of the gusset plate.

n

addition, a provis

avoid local buckling

the unsupported edge

a gusset plate. CAN


17/165

their gravity axes intersect at one point, otherwise, provisions shall be made

and shearing stresses due to the eccentricity. Design strength for gusset pl

considered for tension loading conditions. It is stated that the design strength

lower value obtained according to the limit states of yielding, fracture of the

element, and block shear rupture of the connection. Provisions relating specif

design

of

gusset plates under seismic conditions are provided in the Seismic Pr

Structural Steel Buildings AISC, 1992 . A distinction

is

made between gusse

are connected to brace members that buckle in-plane or out-of-plane of the g

When the brace member buckles out-of-plane it is required that the brace term

gusset a minimum

of

two times the gusset thickness from the theoretical line

which is unrestrained y the column or beam joints. This detail is provided to

formation of a hinge line in the gusset plate.

Under seismic loading conditions, CAN/CSA-SI6.I-M89 - Limit States Desi

Structures CSA, 1989 provides detailed provisions for the design of braced

required factored resistance for the design of bracing connections is prov

resistance of the bolted gusset plate connection

is

based on the ultimate tens

u to block shear considerations.

addition, it

is

suggested that gusset pla

detailed to avoid brittle failures due to rotation

of

the brace when

it

buckles. T

design recommendation, and its origins, will be considered in detail in Chapter 5


18/165

element undergoing large inelastic deformations no design guidelines x

5 Objectives andScope

order for the proposed strong brace weak gusset concentric br

effective under seismic loading conditions the gusset plate must be cap

large inelastic deformations without significant loss o load. There

research project was initiated to investigate the behavior o steel gusset p

loads. The main objectives o the project are s follows:

Observe the general behavior of gusset plates under cyclic loading

2. Provide experimental data for the various design parameters

3. Improve the compressive behavior o gusset plates under cyclic c

4. Determine the ultimate gusset plate capacity under tension

is

a

compressive inelastic deformations and vice versa

5. Determine the feasibility

o

having the gusset plate s the

absorption element in a concentric braced frame

6. Establish preliminary design rules possible

7. Identify areas requiring further investigation

Because o the complexity o the problem the research program develop

purposes was primarily experimental

in

nature. The test results are com

the current design practices and the results o tests performed on gu


19/165


20/165

a Diagonal Bracing

b X-bracing


21/165

8 9

3

' : . . ~ , 1 . '

: : i :

.

4

-50


22/165

Detail

UnkSegment


23/165

l

~

LATERAL OI PLACEMENT lN)

bl

,

S T - - - ~ · _ · _ - _ · _ - . - - - - - - r - - - - O - O - S - H

so

0

0

·

-


24/165

2 LITERATURE REVIEW

2.1

ntroduction

The current design method for gusset plates

is

mainly the result

of

practice, and the engineer s intuition. Recent research has attempt

knowledge

of

the behavior

of

gusset plates

in

an effort to provide

approach. However, the focus has been only on the behavior unde

tension or compression. Research into the behavior of steel gusset

loads

is

severely lacking.

n

this chapter, past work done on gusset plate connections is revie

examines gusset plate research

from an

historical perspective, while S

recent research into the ultimate load behavior

of

gusset plates und

tensile and compressive loads. The limited investigations of the behav

under cyclic loads

is

considered

in

Section 2.4.

2.2

arly

Gusset Plate Research

Early research focused on determining the general elastic stress dis

plates. One

of

the early gusset plate studies that was to prove signific

by

Whitmore 1952). Whitmore tested a one-quarter scale model

connection from the lower chord of

a Warren Truss. A schematic dra

plate prototype

is

shown

in

Figure 2.1. Experimental tests were perfo


25/165

diagonals does not accurately reflect the stress condition in gusset plates. B

observations, Whitmore found that the maximum tensile and compressive st

be approximated quite closely by assuming the force in each diagonal to b

distributed over an area obtained

by

multiplying the plate thickness by an eff

normal to the axis

o

the diagonal. This effective length

is

obtained by draw

from the outside bolts o the first row, to intersect with a line perpendicular to

through the bottom row o bolts. This concept compares quite well to test res

since been used as one o the primary tools

in

gusset plate design.

An

est

gusset plate yield load can be determined by multiplying the yield stress by th

t the effective width section. The method o determining the effectiv

illustrated in Figure 2.2.

Subsequent investigations attempted to confirm Whitmore s fmdings. Ir

investigated the primary stress in the double gusset plates o a Pratt truss. T

o

the maximum stresses were similar to Whitmore s. However, his method

o

the maximum stress was slightly different than Whitmore s. A further study

(1958) o a gusset connection o a Pratt truss confirmed lrvan s [mdings. Fi

studies were first performed by Davis (1967) and Vasarhelyi (1971) to de

elastic stress distribution in gusset plates. Davis performed s study

on

the

model used by Whitmore and confirmed his results. Vasarhelyi performed test

[mite element analyses on a scale model o a Warren truss. He found that th

stress determined by various simplified analytical methods are only slightly d


26/165

problem of ultimate gusset plate capacity under monotonic loads,

compressive.

Tho rn ton 1984) presented an approach to the design of vertical br

based on satisfying the dual requirement of equilibrium and yield, with s

given to stiffness to preclude buckling and failure. Thornton considered

a typical bracing connection.

To

determine the gusset plate ultima

tension, he considered the tear-out of the gusset plate. The capacity is re

shear requirements of the 1978 AISC Specification. The tear-out capac

net section with hole size taken as bolt diameter plus 1/16 inch. n comp

considered gusset plate buckling by establishing the capacity

of

an e

section. This method considers an imaginary fixed-fIXed column strip

coefficient, k=O.65) of unit width below the Whiunore section. The len

strip may be taken as the largest

of l LZ

and

L3

as defmed in Figure

used to determine an equivalent slenderness ratio. Alternately, Thornto

shorter length, such as the average

of

L1 LZ and L3 may give a

approximation of the buckling strength. Thornton originally presented h

allowable stress method. From an ultimate strength perspective, the com

resistance

of

the gusset plate can be evaluated according to the colu

CAN/CSA S16 1 M89

standard CSA, 1989) using the Whiunore

Whiunore, 95Z as the column area and the equivalent slenderness fro

column strip. Thornton states that this approach is conservative becau


27/165

were performed. with three tests each at two different plate thicknesses and t

bracing member angles. For the type

of

bracing connection examined, the p

mode

of

the gusset plate was a t ear across the boltom bolt holes

of

the splic

It was also determined that the type and location

of

the gusset

p ~

boundari

with the load transfer mechanism into the plate. have important secondary eff

buckling and associated out-of-plane bending.

Hardash and Bjorhovde 1985) continued the study

of

gusset plate capacity

loads. In o rd er to develop an ultimate strength approach t o the design

of

g

test results from the University

of

Arizona, the.University

of

Illinois, and the

Alberta were incorporated into the evaluation. For

all

42 gusset plate specim

te ar across the last r ow

of

bolts was observed, regardless

of

the strength para

size,

or

plate material. It was concluded that the governing block shear m

incorporating the tensile ultimate strength. F

u

on the net area between the

bolts. and a uniform effective shear stress, Feff. acting on the gross area alon

bolt holes Figure 2.4). The set

of

equations developed to give the nom

resistance, R

n

•

of

a gusset plate loaded in tension are as follows:

R

n

FuSnett

1.15F

e

ff

Lt

Feff

=

1-Cl F

y +

CIF

u

CI =0.95 - 0.047 L L in inches)

It was discovered that the shear stress distribution is

not

uniform, but rather


28/165

design procedures for steel gusset plates

in

diagonally braced fra

considered both the tensile and compressive behavior

gusset pla

included the nonlinear behavior the fasteners and the frame to which

attached. Finite element results demonstrated that the additional stiffne

gusset plate cause the beam to develop end moments equivalent to a h

beam. To determine the tensile capacity, Richard developed a block sh

it is assumed that the plate will fail along the gross section of the bo

diagonal connections. The strength the gusset plate

in

tension may

combining the tensile stress resultant acting at the end the bolt patt

stress resultants acting along the sides

the bolts. the case b

with less than six rows

bolts, it

is

suggested that it may be appropr

shear area, as opposed to the gross shear area,

in

the block shear model.

were also generated to predict the gusset-to-frame fastener force dis

determined that fastener forces do not act in pure shear as is commonly

design procedures.

compression, the ftnite element analysis was

elastic behavior.

Full-scale diagonal bracing members were tested by Cheng and H

University Alberta. The compressive behavior and elastic buckl

examined. The gusset plate parameters considered

in

the investi

thickness, geometric conftguration, boundary condition, and out-of-plan

concentrically loaded tests failed

in

plate buckling. The eccentrically


29/165

Additional analysis of the test results was performed by Cheng, et

al

199

element analysis by the program ANSYS gave reasonable predictions

of

buckling strength of the gusset plate test specimens. Furthermore, the sp

connection length and the thickness of the splice member were found to affe

buckling strength of the specimens. Increasing the length of the spl

connection, or the thickness of the splice member, results in an increase in

buckling strength.

An experimental program was undertaken by Gross 1990 to determine

1

t

of the members framing into the connection

on

the behavior and stren

connection, 2 the effect

of

connection eccenuicity on gusset plate capac

distribution

of

forces to the framing members, and 3 the difference

in

p

between a connection made to the column flange and one made to the colum

behavior of three nearly full-scale braced steel subassemblies were studied exp

It was determined that computing gusset plate buckling using the equival

method with a value of k=O 5 appears to be conservative. Additionally, gus

capacity is predicted very closely using the block shear approach. t was fou

eccentric connection, which had a compact gusset plate, had a higher buckli

than larger concentric gusset plates, and that the gusset plates produced a

condition for the beam to column connection.

Additional gusset plate research was performed at the University of Alberta by


30/165

moving out-of-plane. The paper presents the results o the inelastic s

Cheng and Yam and reviews the test results o the elastic specimens t

Hu 1987 . It was determined that the plate thickness and plate size

strength o the gusset plate significantly. The failure mode for the

sway buckling o the gusset plate connection, while the failure mode fo

local buckling o the free edges o the plate. The out-of-plane restrain

on the inelastic buckling strength o the specimens while the elastic bu

greatly affected by the restraint. The test results show that the Whitm

concept overestimates the strength o the specimens that failed in e

underestimates the inelastic buckling strength. addition, the T

column method produced a large margin o safety for the gusset p

failed in inelastic buckling.

Cheng and Yam are considering the effect o the framing members

eccentric loading conditions

on

the compressive behavior

o

gusset pla

the research

is

currently

in

progress at the University o Alberta.

4 Cyclic Gusset Plate Behavior

Research into the cyclic behavior o steel gusset plates in concentrical

severely lacking. One of the most relevant studies

in

this area

is

a ser

investigations conducted at the University o Michigan. Astaneh-

conducted an experimental study to investigate the cyclic behavior

o


31/165

simulate the effects of strong earthquakes. Although the study focused o

member behavior, gusset plate behavior was also monitored. Observations

plane specimen tests revealed that no major plastification occurred in the guss

they behaved mostly elastic until the end

of

the test. Observations from the

specimen tests showed that during post-buckling deformations the rotation

o

hinges in the gusset plate is about an axis normal to the axis

of

the brace mem

the study, i t was discovered that the portion of the gusset plate connected t

must be allowed to rotate freely about the axis of the hinge. Any restric

freedom would cause early fractures in the gusset plate.

s

such, it was con

the tests that an optimum free length of 2t, where t is the thickness of the guss

necessary for the free space in the gusset to prevent its premature fracture.

shows the type of gusset plate employed in this study and the recommended fr

gusset plate required t allow free rotation about the plastic hinge.

A further research paper by Astaneh-AsI,

et

al. 1985 focuses on the o

specimen tests. The results of both the welded and bolted specimens are inc

report.

t

was observed that three plastic hinges formed in all test specimens.

formed at midspan

of

the brace and one in each end gusset plate. The formatio

hinges in the end gusset plates was due to the relatively small strength and stif

gusset plate in out-of-plane bending, compared to that

of

the double-angle me

combined effect

of

bending and

x l

load in the post-buckling region caused


32/165

he cyclic behavior of gusset plate connections in V-braced fram

Astaneh 1992 . o improve the behavior of V-braced frames, this

shear inelasticity of the gusset plate connection be utilized as a reliab

of

ductility and energy dissipation. The experimental part

of

the re

subjecting three gusset plate connections to cyclic loading. t wa

specimen with the largest eccentricity of point of intersection of mem

most desirable manner, while the behavior of the typical concentric co

braced frames was relatively brittle and undesirable.

n

the specim

eccentricity, the governing failure mode was the shear yielding

of

the

was a very ductile and stable energy dissipating mechanism. Add

design approach is proposed. The main component of the design ph

the shear yielding of the gusset plate the weakest link in the system. It

gusset plate will yield before the buckling of the bracing member an

shear deformation of the gusset will result in a more ductile bracing sys

Research into the behavior of

connections for seismic-resistant eccentr

EBFs is described by Engelhardt and Popov 1989 . Although b

EBFs may be subject to a significantly different stress environment tha

in concentrically braced frames, the observed gusset plate behavior is

was observed that large compressive stresses were produced along the

plate nearest the

l nk

region. n order to preclude gusset plate buckl


33/165

inelastic compressive behavior he current study

o

the cyclic behavior o

plates draws upon the experienced gained in these previous investigations


34/165

russ utline

o

o

o

onnection

o o


35/165

Whitmore Eff

Width

Figure 2 2 Whitmore Effective Width


36/165

Whitm

W

Imaginary Fixed Fixed

Column Strip:

k

65

Equivalent slenderness kL


37/165

U1tmate Tear-out Resistance, R

n

:

Hardash and Bjorhovde, 1985


38/165

Figure 2.5. Gusset Plate Studied y Astaneh-Asl, t al.


39/165

3 EXPERIMENTAL PROGRAM

3.1

ntroduction

The purpose of the experimental program was to investigate the general

gusset plate connections

in

a braced steel frame under cyclic loading con

bracing system subjected to cyclic loads may

fail

either in the bracing membe

the gusset plate which connects the bracing member to the beam and co

proposed strong brace, weak gusset model to be examined in this investigatio

a concentric braced frame where the brace member

is

designed not to buckle

case, the bracing member acts

as

a restraining member to provide rotational

the gusset plate. The assumption was made in designing the present experimen

that the gusset plate would fail prior to the bracing member and that th

restraint provided by the bracing member

is

effectively infmite.

3.2

reliminary

Considerations

The experimental program was designed to represent the conditions of actual

connections. Thus, full-scale single gusset plate connections of a diago

member a t the joint of a beam and column were used. The variables of t

connection include plate thickness, plate size and configuration, plate boundary

angle of bracing, type of connection welded or bolted), cross-section of braci

length of splice member. order to simplify the problem, the test parameters


40/165

designed to best represent the conditions

of

actual gusset plate conne

common bracing angles in practice range from 30 to 60 degrees.

As

bracing angle was used in this program. It was

chosen to connect the b

gusset plate through the use

of

a splice member. A bolted connection w

the splice member to the gusset plate

and

the gusset plate itself was dir

beam and column members in the braced frame assembly. All gusset

concentrically through the brace member.

3.3 Specimen Description

A series

of

five specimens was tested in the experimental program.

thickness

of

the test specimens are listed in Table

3.1

and a typical gusse

shown in Figure 3.1. All specimens were fabricated from CSA

structural quality steel. Two different thicknesses

of

gusset plate, 9.32

were used in the test program. The test specimens were rectangular

exception

of

Specimen A-5, which was designed to allow the free for

hinge under compressive buckling deformations. Based on the reco

previous research investigation Astaneh-Asl, et al., 1985 , a plastic hin

of

2t, where t is the thickness

of

the test specimen, was provided for

splice member connection. The modified geometry

of

Specimen

A

Figure 3.2.

order to investigate the effect

of

the stiffness

of

the free edge


41/165

energy absorbing capacity of these specimens.

nall cases, two Tee-sections WTl25x22.5) and two

nun thick plates w

splice member to ensure that

the

gusset plate

failed

prior to the splice assemb

nun

specimens were connected to the splice member with five rows of 7

ASTM A325 high strength bolts designed for bearing. Due to the loads antic

connection, all 9.32 nun thick specimens were expected to experience bo

gusset plate to splice member connection during the cyclic loading history. T

thick specimens utilized 7/8 diameter ASTM A490 bolts in order to achieve

connection.

All

bolts were pretensioned using turn-of-nut tightening as

CAN/CSA-S16.1-M89, Clause 23.5 CSA, 1989). The calculated slip-resi

gusset plate to splice member connections are recorded

in

Table 3.2.

All

spe

designed to be loaded at

45

degrees

by

the bracing member and all spe

directly welded to the beam

and

column members. n the design of the weld

the force distribution between the gusset plate and the beam and column m

based on a model proposed

by

Williams

and Richard 1986).

n

connectin

plate specimens to the frame members, a mm ftilet weld was used for t

specimens and an 8 mm ftllet weld was used for the 6.18 mm specimens.

3.4 st Set-up

A gusset plate connection under cyclic loading conditions might deform out-


42/165

allowed to move out-of-plane along with the frame assembly. To simp

the simulation Figure 3.3 b

w s

used. To further simplify the tes

that would normally exist in the framing members under lateral loadin

neglected in the testing program.

The test set-up is shown schematically n Figure 3.4. The cyclic test loa

the MTS 6000 testing machine. Two W31Oxl29 steel sections were us

column members in the test frame assembly. The frame assembly wa

WWF400x202 distributing beam

in

order

to

transfer the specimen loads

up reactions. The diagonal bracing member W250x67 was res

displacements in the plane

the test specimen

by

the restraint provi

bracing roller assemblies. The roller assemblies were fixed to the f

testing machine and restrained the brace member at the contact point

3.4. The test frame, with the specimen in place, was then sandwiched

rollers at each end of the set-up to allow it to sway laterally out-of

compressive and tensile loading Figure 3.5 . The compressive reaction

directly to the strong floor

the test lab, while the tensile reactions

tension reaction frame s shown in Figure 3.6. To prevent a sudden kic

frame a guided rod mechanism was affixed to each end

the distrib

movement the guided rod mechanism was monitored throughout th

unrestrained sway

the test

fr me

Figure 3.7 is a photograph

the a

up, with the tension reaction frames removed.


43/165

the gusset plate. The location

the strain gages was decided on the basis

gusset plate test experience. addition, a pair rosette strain gages wer

either side the test specimens at the possible location the maximum no

The strain gage locations were the same for Specimens A-I through A-4, wh

locations were modified for Specimen

A 5

due to the unique geometry

tha

The locations

the specimen strain gages are shown in Figures 3.8 and 3.9.

LVDTs were used to monitor both the out-of-plane and in-plane displacem

gusset plate specimens and the test frame assembly. The out-of-plane buckl

the specimen plates were monitored by three sets of LVDTs which recorded th

shapes

the two free edges

the gusset plates and the center line of the l

Figure 3.10 and 3.11 .

the in-plane direction, a pair

LVDTs were use

the axial displacement the gusset plate itself. An additional pair LVDTs

.to monitor the axial displacement of the splice member and to observe the beh

proposed fixed boundary condition. By monitoring the axial displacement

gusset plate and the splice member, the deformation and bolt slip

in

the splice

gusset plate connection could be isolated. addition, an LVDT was used to

axial displacement the brace member relative

to

the distributing beam to c

the brace member

to

splice member connection was performing adequately. Th

the axial displacement LVDTs are shown

in

Figure 3.12. All axial displacem

were referenced

to

the distributing beam.


44/165

lateral and longitudinal twist that might develop in the test frame. F

deflection

the distributing beam was recorded using a cable transdu

the test frame instrumentation is detailed in Figure 3.13. ll in-plan

out-of-plane LVOTs were placed on a support frame affixed

to

the

the point load application s shown

Figure 3.14. The rema

affixed to the frame of the MTS testing machine.

The cyclic load applied

to

the test specimens was monitored by the int

MTS 6000. Furthermore load cells were incorporated into the tens

ends

the test set-up. Under tension loading the MTS load was com

the loads recorded at the tension reactions. The MTS load was

str

the loads recorded at the reactions. Therefore the statics

the sy

and all loads are accounted for. The specimen load s recorded by th

machine was used the test results. Load cells were not inc

compression reactions due

to

stability considerations.

The electronic readings generated from the strain gages LVOTs load

testing machine were continuously monitored by the Fluke data acqu

gage readings were monitored manually at regular intervals through

addition a whitewash coating was applied to ll specimens

order to

yielding process.


45/165

properties Fy=300 MPa). The procedure for calculating the Whitmore yield c

gusset plate

is

outlined

Section 2.2. During each cycle, the specimen w

loaded

n

tension to the desired

maximum

cycle load level. The specimen

unloaded from tension and cycled. through zero

to

the same load level

n

co

wo

cycles were conducted at each elastic load level. The first cycle was

start-stop fashion with the testing stopped at regular intervals for elect

acquisition. The second cycle at each load level was tested under continuo

conditions; the loading was only stopped at the peak tensile load, zero load, an

compressive load during each cycle so that qualitative specimen observation

gage readings could

be

recorded.

The inelastic loading sequence for all specimens began with a set o yield leve

100 o the expected nominal tensile

yield

load. The loading procedure in t

range varied slightly between the specimens tested.

n

general, subsequent inel

were conducted under increasing levels

o

specimen axial deformation. The

axial deformation was monitored using the LVDTs attached to the splice me

that the bolt slip in the connection, any, would be included the determina

cycle axial deformation level.

During each cycle, a specimen

was

initially loaded

in

tension to a predetermin

axial deformation. The specimen

was

unloaded from tension and then de

compression to the same axial deformation level, as referenced to the


46/165

made to a nominal yield level defonnation.

As

such the cycle axia

was increased

in

roughly evenly spaced intervals over the course

of

th

he

level of axial defonnation was increased

in

subsequent cycles unt

of the specimen was achieved. Failure in tension was signified by a de

load carrying capacity under increasing MTS machine stroke. The tes

loading the specimen in compression until

an

excessive level

of

defonnation was obtained. The test loading was limited by the lev

defonnation that was able to be accommodated by the test set up.

All inelastic cycles were tested under continuous loading conditions wi

the testing of Specimen A I. It was originally thought that the effect of

stop start loading on the specimen behavior could be investigated in the

such the loading scheme for Specimen A I involved a set

of

three c

defonnation level in the inelastic range. The fIrst cycle was loaded in a

with the testing stopped at regular intervals for electronic data acquisiti

cycles of the set were loaded continuously. However when the axial d

held constant for successive inelastic cycles there is inadvertently an ob

load carrying capacity due to inelastic straining and residual defonnatio

no significant infonnation could be obtained from this cycle loading sc

inelastic loading cycles

of

increasing axial defonnation with no cy

employed in the testing of all remaining specimens.


47/165

behavior

o

steel structures

is

relatively unaffected

by

moderate variations

i

rate within the range o loading rates employed this investigation Davis,

Therefore, the loading rates chosen reflects values that enabled careful an

observation o the specimen throughout the testing process.


48/165

Table 3.1. Specimen Description

Test

Plate Size Thickness Speci

Specimen

mmxmm

mm

Test Par

A I

550 x 450 9.32

Basic

A 550x 450 6.18

Basic

A 3 550x

450

9.32 Stiffened F

A

550 x 450

6.18 Stiffened F

A 5

550x

450

9.32 Free Formatio

Hinge Fa


49/165

Table 3 2 Slip Resistance Gusset Plate to

Splice Member Connection

Slip Resistance kN)

Test Connection

5 Probability of Slip

Specimen

Details

Class

A Surface

Class B S

Clean

mill

scale

Blast cl

9 32mm

10

ASTM

A325

918 151

Specimens

8

n

Bolts

6.18

mm

10

ASTM A490 1096 181

Specimens

8 n

Bolts

As

per

CAN/CSA SI6 I M89

Clause

13.12


50/165

45

5

33

55

igur

3 1 Typical Test Specimen Geometry


51/165


52/165


53/165

Distributing

Beam

WWF400x202

•

if

Brace Member

o 0

Splice Member1 :

W250X67

WT125x22.5

0 0

o 0

o 0

Gusset Plate

Lateral Brace

o

0

/ Specimen

Location

Fram

/

/

Asse

W31O

t

r r l


54/165


55/165


56/165


57/165

Legend

Strain

k Rosett

50

50

r

5

~ f i

50

5

25

25

Figure 3 8 Typical Strain Gage Locations

Specimens A I to A 4


58/165


59/165

:

:

Figure 3 10 Out or plane LVDT Locations

Specimens A I to A 4


60/165

Figure 3 11 Out of plane LVDT Locations for Specime


61/165

o

o

o

o

o

o

V

on both sides

of splice mem er


62/165


63/165


64/165

4 EXPERIMENTAL RESULTS

4 Introduction

The results

the experimental program are presented

in

this chapte

properties

the test specimen are discussed in Section 4.2. The resu

plate tests are presented in Section 4.3. The

full

test behavior each

presented. This chapter will focus on the physical behavior

the spec

cyclic loading

as

well

as

consider the load versus defonnation respo

configuration, and the observed strain distribution the specimens. Fur

the test results in relations

to

the test parameters will be covered in Chap

4 2 Material Properties

Table 4 1 lists the values the elastic modulus, the static and dynamic

the dynamic ultimate strength

as

detennined

from

tensile tests on coupo

material used

in

the test specimens. Four coupons were fabricated from

thicknesses utilized. A 2 mm gage length was used for all material tes

from each plate were tested in the direction parallel to the sheet steel rol

while the remaining two coupons were tested

in

the perpendicular direct

values from each plate direction and the average values from the four co

reported. As expected, higher strengths were observed from the coupo

to the rolled direction

the steel plate. However, the response

the


65/165

4 3 Test Results

The ultimate tensile and compressive loads attained during testing are record

4.2. The ultimate load

is

defmed as the maximum load level reached by

throughout its cyclic loading history. The test results each specimen are p

tum in Section 4.3.1 through Section 4.3.5. The discussion of each specime

focus on the physical response the specimen during testing and will then c

load versus deformation response, the buckled configuration, and the obse

distribution

within

the specimen. Additional test data, including the gusset

versus deformation response and the load versus out-of-plane displacement

frame, is provided in Appendix

4 3 Specimen ·

Specimen A-I

is

a basic, unreinforced gusset plate specimen

9.32

rom

thick

7/8 diameter ASTM A325 high strength bolts designed for bearing were used

the splice member to the gusset plate.

The load versus axial displacement response Specimen A-I throughout

history is shown in Figure 4.2. The area under the load-deformation curve is a

the energy absorbed by the gusset plate throughout the cyclic loading. Plotted

average axial displacement the gusset plate assembly. The displacement wa

by the LVDTs attached to the splice member and referenced to the distributing


66/165

By examining the tension portion

o

the loading

in

Figure 4.2, it is

specimen response under tension

is

relatively stable; as the axial de

increased the load carrying capacity o the specimen

is

maintained

compression, the overall plate buckling

o

the specimen is accompa

increase in axial defonnation and a significant drop in load carrying ca

the cyclic specimen loading was not continued very far into the post-bu

still be observed that there is a deterioration

o

the compressive load c

the cyclic axial defonnation level and the specimen out-of-plane defonn

photograph

o

the failed specimen and the final yield line pattern is show

Figure 4.4 a) shows the specimen load versus defonnation response

o

assembly during the elastic portion o loading, Cycle 1-7. sexpec

were observed on the specimen during the elastic cycles. During the

Cycle 5,

t

a load o approximately 1000 kN, boIt slip occurred in t

splice member to gusset plate bolted connection. Due to the slack cre

system, the observed load dropped during the slip process. Upon comp

bolt slip, the specimen once again began

to

pick up load.

s

the testing progressed, the splice member to gusset plate conn

slippage twice during every test cycle, once during tension loading, and

the specimen was cycled through into compression. The connection


67/165

Figure 4.4 b) shows the load versus deformation response the specimen

inelastic loading cycles up to the onset plate buckling. Cycle 8-13. The

signs yielding were recorded on the specimen along the welded boundary

gusset plate and the frame assembly. Yielding was indicated by the flak

whitewash coating on the specimen. Yielding

in

this region appeared to be

restraint provided by the welded boundary s the specimen began to deform o

the cycle axial deformation level increased. yield lines developed on the g

about the sides the splice member at the mid-height the connection. Th

the yielding plate material

in

the connection under both tensile and comp

bearing deformations. Yield lines about the base the splice member r

beginning a plate buckle stretching from the bottom comer one free edge

beneath the splice member. to the bottom corner the other free edge.

During the compressive loading of Cycle 13. overall plate buckling occurred

free edge

the gusset plate buckled. allowing the specimen to deform signil

of-plane. The load carrying capacity the specimen dropped signillcantly a

plate buckling occurred. a result, the maximum load 1682 kN reached d

11 is the compressive capacity this specimen.

Once overall plate buckling h d occurred, the remaining loading cycles for Spe

involved only an increase in the axial deformation level for the tension loading

each cycle. Since the specimen had already buckled

in

compression, it was


68/165

Cycle 17). The specimen underwent significant tensile axial defonna

failure was reached at a load

1794 kN.

The gusset plate specimen was examined after the completion the test.

that the specimen failed

in

tension due a fracture

the plate material

holes in the bottom row the connection. Figure 4.5

is

a photograph

material. 11ris tensile failure mode was characteristic

all the specim

remaining bolt holes in the connection had been defonned in both tension

due to the bearing of the bolts under load, with the greatest defonnation

fIrst and last row the connection.

The linear variable displacement transducers recorded the out-of-plane d

the specimen free edges and the center line deflection. Figure

displacements that existed

in

the specimen after overall plate buckling o

13. It

is

observed that the deflected shape

the specimen center line

that a fIxed-fIxed but guided column. The long free edge has buckled

shape indicates that the restraint provided to the top the specimen by t

appears to be localized, and does not provide f ity

to

the free edges

Figure 4.7

is

a photograph

the long free edge

the specimen a

buckling occurred.

plotting the deflected shape of the specimen center line, an interpolati


69/165

the splice member was assumed to be rougWy equal to the out-of-plane disp

the top

the gusset plate free edges. Since the splice member is relatively rigi

to the gusset plate, this assumption appears to be valid.

Specimen strain gage readings were recorded throughout the loading history.

only strain levels obtained from the elastic portion loading are relevant in an

the strain distribution in the specimen. Once local yielding occurs and the stres

redistribute in the plate, the strain readings are difficult to interpret. Strain

were investigated at the cycle maximum loads at both the 50 and

iClO

Whitmore load levels, 530 kN and 1060 kN respectively. From this anal

observed that the

p k

strain levels in the specimen were recorded at the roset

beneath the splice member. Figure 4.8 shows both the tensile and compressive

distributions at the 530 kN and 1060 kN levels. At the 530 kN load level th

significant difference between the strain levels and pattern distribution b

tension and compression stress states.

This

indicates that the elastic loading

the specimen is generally the same under both tension and compression loading

When the specimen load is increased to the I060 kN level, strain levels ar

twice those recorded at the 530 kN load level. The linearity strain s

assumption that the material response is generally elastic at this load level.

strain readings at the -1060 kN load level are analyzed, the strain levels ar


70/165

For the purpose of comparison, the maximum strain levels

in

the specim

Whitmore critical stress theory, should be approximately ±1500 m

Whitmore section at the 100% nominal yield load level (1060 kN). The

runs through the bottom row of bolts

in

the splice member connection,

rosette gage locations.

4 3 Specimen A·

Specimen A-2 is a basic, unreinforced gusset plate specimen of 6.18 rom

7/8 diameter ASTM A490 high strength bolts were used to connect the

the gusset plate in order to achieve a slip-resistant connection. No bolt

to occur during the testing of this specimen.

Figure 4.9 shows the load versus the axial displacement response of

assembly throughout the loading of Specimen A-2. By examining the t

the loading, it is observed that the specimen response under tensio

compression, the buckling behavior of the specimen is revealed. The occ

plate buckling

is

well defmed

by

the compressive peak load plotted in

observed during testing, the compressive capacity immediately drops

post-buckling load capacity level. The figure confirms that the post-buc

relatively stable; the post-buckling load capacity is seen to deteriorate v

axial

plate deformations increase significantly. A photograph of the fail


71/165

long free edge of the plate. Yielding was likely due to the restraint prov

welded boundary condition.

Figure 4.11 b) shows the load versus deformation response of the specimen

inelastic loading cycles

up

to the onset of plate buckling, Cycle 7-8.

th

deformation level increased, a significant increase in compressive plate y

observed about the base of the splice member. The fIrst signs of tension y

observed during Cycle 8 ±2.0l1

y

), when yield lines were noted about the

splice member as the material in the connection region was beginning to yield

loading was cycled through to compression, overall plate buckling occurred

1128

le

Immediately upon buckling, the load carried by the specimen dro

le

The plate buckle extended from the midpoint of both free edges down to

the splice member. Both plate free edges buckled in order to enable the

undergo signifIcant out-of-plane deformation.

Figure 4.11 c) shows the specimen load versus deformation response for the r

the test loading.

the test proceeded under sequentially increasing levels of

plate deformation, the tensile yield lines about the splice member connectio

h maximum tensile loads reached continued to increase as the cycle axial

level increased.

compression, only minor deterioration was observed i

buckling behavior of the specimen. The compressive loads attained only decrea


72/165

Figure 4.10).

The tensile failure of the specimen occurred during Cycle 16 ±16.0d

carrying capacity reached 1340 le then quickly began to drop as the

axial deformation continued to increase. The test was concluded by load

in

compression until an excessive level

of

deformation was achieved. Eve

deformation level that was imposed during this [mal cycle, the compress

still 630

le

only 100 le less than the initial post-buckling capacity. Th

shape of the specimen Figure 4.10) reveals the complex pattern of plate

associated plastic hinge locations. It

is

believed that the complex pa

allowed the specimen to redistribute the compressive load through a ser

load paths, resulting in a relatively stable post-buckling capacity.

The failure

of

the specimen

in

tension was a fracture

of

the plate mate

bolts in the bottom row

of

the connection. Only the bottom row

of

bolt h

be significantly deformed. Since no major connection slip occu

deformations observed were due

to

overall connection material yielding,

deformations.

addition, tensile yielding caused the plate to spread out

upwards at the top

of

the splice member connection region.

The out-of-plane deflected shape

of

the plate free edges and the center l


73/165

plane deformations increased, the fIxity

of

the free edges decreased as pl

developed at the bottom comers of the free edges· at the location

of

the weld

to the frame assembly. Photographs

of

the free edges (Figure 4.13) show th

shapes, including the location

of

the plastic hinge lines that developed as th

buckled.

Strain gage values were investigated at the maximum cycle loads at both th

100 cycle levels. For Specimen A-2, this corresponds to 353 kN an

respectively.

general for Specimen A-2, the strains are highest at the roset

at the base

of the splice member, and the edge strains are higher along the lon

The recorded plate strain distributions at both the 353 kN and 707 kN load

shown in Figure 4.14. At the +707 kN load level, the distribution

of

strains

unchanged from the +353 kN level. addition, strains have approximate

indicating an elastic linearity

of strains

in

tension. Under compressive loading

have more than doubled at the rosette locations when the loading increased fro

kN to the -707 kN level. The strains at the rosette locations are approaching

levels, which

is

in agreement with the whitewash flaking observed on the speci

load level.

4.3.3 Specimen A·3

Specimen A-3

is

a 9.32 mm thick specimen with reinforced plate free edges.


74/165

Specimen A-I

in

that

as

the axial deformation level is cyclically inc

carrying capacity of the specimen

is

stable.

n

compression, the enha

Specimen A-3 is revealed. Not only is the ultimate compressive capaci

that of Specimen A I but the onset of overall plate buckling was not a

sudden increase

in

axial plate deformation nor a significant decrease

capacity. Since Figure 4.2 and Figure 4.15 are both plotted to the

increased energy absorption capacity

of

the reinforced plate, Specimen A

to the basic plate, Specimen A-I,

is

evident. A photograph of the failed

fmal yield line pattern

is

shown

in

Figure 4.16.

Figure 4.17 a) shows the specimen load versus deformation response

of

assembly during the elastic portion

of

loading, Cycle 1-6.

As

expecte

were observed

on

the specimen during the elastic cycles. Figure 4.17 b

versus deformation response during the inelastic loading cycles up to th

buckling, Cycle 7-14. Connection bolt slip

in

Specimen A-3 first occ

tension loading

of

Cycle 7 at a load

of

1290

kN As

in the testing

of

Sp

splice member to gusset plate connection experiences slippage twic

subsequent test cycle. The bolt

slip

experienced during each cycle of loa

the load versus axial deformation response of Figure 4.15. The saw-tooth

slip on the load versus deformation curve

is

a function of the discrete na

observations.


75/165

tensile yielding was first noticed on the specimen along the sides the splice

the plate material in the connection region began

to

yield. Yielding about the

splice member connection was likely due to both tension and compressive lo

testing progressed, plate yielding spread out from the regions previously indica

During the compressive loading portion Cycle 14, a maximum compres

1990 kN was reached, accompanied by the first signs

yielding on the outsid

plate edge stiffeners. 1bis was a slight decrease from the maximum comp

reached during the previous cycle. The slight decrease in maximum load and

yielding

in

the stiffeners indicated that plate buckling or local plate cripp

specimen had begun. The maximum load of 2004 kN reached during Cycle

compressive capacity of this specimen.

The test proceeded under sequentially increasing levels

cyclic axial plate

until tensile failure of the specimen was observed. Figure 4.l7 c) shows the sp

versus deformation response for the remainder the test loading. The fa

specimen in tension occurred during Cycle 15 just prior to reaching the d

axial deformation level. The maximum tensile load reached prior to failure

1bis was slightly less than the ultimate specimen tensile capacity 1884 k

during the previous cycle. The test was concluded by loading the s

compression until

an

excessive level compressive deformation was achiev

this

fmal

loading a

peak

compressive load

1910 kN was reached prior t


76/165

to spread outwards and to bulge upwards at the top o the connection.

the high compressive loads realized the compressive bolt hole deforma

severe resulting in some piling-up of material on the compressive side o

holes.

Figure 4.18 shows the out-of-plane deflected shape

o

the stiffened plat

center line deflection

o

the specimen after the onset

o

overall plate b

14. The deflected shape of the center line again resembles that

o

a

guided column. However the presence of the edge stiffeners restricts

from experiencing any signillcant out-of-plane deformation t this load le

severe compressive axial deformations were imposed during the fmal

significant deformations become visible along the stiffened edges. As seen

the stiffener displacements are predominately confIned to rotational defo

stiffener to frame weld locations while the rest of the stiffener remains relat

Gusset plate strain distributions for Specimen

A

were investigated at th

Specimen A-I. Figure 4.20 shows the strain distributions at both the 530 k

load levels for Specimen A-3. The general distribution

o

strains re

unchanged between the tensile and compressive load states. The peak s

are recorded

t

the rosette locations beneath the splice plate while the e

strains are higher along the long edge side

o

the specimen. Along each


77/165

stiffeners onto the specimens, the plate stresses and strains are distributed ov

area

of

material. addition, at the -1060 leN load level, no significant

indicated by the strain readings in Specimen A-3.

is

observation sup

conclusion that the addition of stiffeners to Specimen A-3 delays the onset of yi

out-of-plane buckling deformation in

the gusset plate.

4 3 4 Specimen A 4

Specimen A-4 is a 6.18

rnm

thick specimen with reinforced plate free edges.

6

8

rnm

stiffeners, of the same material as the specimen, were welded alon

length

of

each plate free edge. Ten 7/8 diameter ASTM A490 high strength

used to connect the splice member to the gusset plate

in

order to achieve a sli

connection. No bolt slip was expected

to

occur during the testing

of

this specim

Figure 4.21 shows the load versus axial displacement response

of

Specimen

tensile portion of the cyclic loading curve shows the stable response of the speci

tension loading. compression, the effect of the edge stiffeners is observed.

the compressive capacity of Specimen A-4

is

not significantly higher than Spec

the presence

of

the plate edge stiffeners greatly affected the post-buckling behav

specimen. The attainment of the specimen compressive capacity was not follo

sudden drop in load carrying capacity, but instead, a steady deterioration is obse

comparing Figure 4.9 and Figure 4.21, the increased energy absorption capac


78/165

observed during Cycle 5, at the 100 nominal Whitmore yield load lev

lines were noted running parallel to the long free edge at the base

of

the

onset of yielding was supported by the presence of residual plate strains a

of the cycle.

The testing proceeded with cycles

of

increasing axial plate deformation.

shows the load versus deformation response of the specimen during the

up to the onset of plate buckling, Cycle 7-10. During Cycle 9, at an a

level of

5d

y

, the fIrst signs of tensile yield lines were observed on the

sides

of

the splice member connection.

n

compression, yield lines inc

base

of

the splice member as a localized buckle began to form at the base

plate. The onset

of

yielding was also observed on the outsides

of

the pla

at the bottom corner of each edge.

During Cycle 10 (±3.0d

y

) the compressive capacity

of

the specimen w

compressive load of 1149

l

was attained, accompanied by an incre

activity about the base of the splice member and on the plate edge stiffener

compressive capacity had been reached, the plate appeared to have only b

overall plate buckling was being restrained

by

the plate edge stiffeners. U

A-2, the buckling

of

the

spe imen

was not immediately followed by a sig

the specimen load carrying capacity.


79/165

steady pace as the testing cycles continued. There was an increase

in

yielding

the specimen as the plate continued to buckle under constraint

from

the

stiffeners. Yield lines developed along the long edge of the gusset plate just

edge stiffener. Eventually the buckled shape became more defined s t r e t ~ h i n

bottom comers the splice member out towards the comers of the plate ed

lines on the plate stiffeners increased significantly under cycles increasing

defonnation. Ultimately the plate edge stiffeners themselves defonned both in

out-of-plane allowing the outside edges of the specimen plate to buckle.

The tensile capacity of the specimen was reached during Cycle 17 ± 2 I . O ~

maximum load

1265

leN

was reached. However the tensile failure of th

occurred during Cycle 18 when the tensile load carrying capacity began to

increasing axial plate defonnation. During Cycle 18 a

new

region

yi

observed along the gusset plate to frame assembly boundary due to the cyclic o

defonnations

in

the specimen. The test

was

concluded

by

loading the sp

compression. The test was stopped when the compressive capacity

the spe

deteriorated

to

680

leN

The tensile failure mode of the specimen was a fracture of the plate material b

bolts in the bottom row the connection. There was no sign bolt

connection region. Tensile yielding defonnations were the most severe in the b


80/165

connection region after the completion

of

testing.

Figure 4.26 shows the out-of-plane deflected shape

of

the stiffened pla

center line deflection of the specimen during the compressive loading

Initially, the deflected shape

of

the center line resembles the buckled shap

but guided column. The deflected shapes of the reinforced edges revea

themselves had not buckled at this load cycle level.

the co

deformations increased, the deflected shape of the specimen center line b

that of a fixed-pinned but guided column.

the stiffened edges began t

plate buckling was facilitated. This created a plastic hinge region beneat

splice member, thereby reducing the fixity provided by the relatively rigid

Photographs of the [mal deformed shape of the plate edge stiffeners Fi

that the stiffeners have buckled.

Gusset plate strain distributions for Specimen A-4 were investigated a

levels as Specimen A-2. Figure 4.27 shows the tensile and compressive str

at both the 353 kN and 707 kN level for Specimen A-4.

n

general, the

highest

at

the rosette locations beneath the splice member. The strai

generally the same along both stiffened edges of the specimen, with slight

observed

at

the top

comer

of

the long free edge. When the strain di

tensile and compressive loading

s

compared, the compressive loading sta


81/165

edges and within the interior of the plate, but are at similar levels at the rosett

By providing stiffeners to the plate edges, the stresses

and

strains in the edge

the plate are distributed over a greater area material, thereby reducing the s

in

those locations.

4 3 5 Specimen A 5

Specimen A-5 is a 9.32 mrn thick specimen. The plate geometry was m

accommodate the free formation of a plastic hinge under compressiv

deformations. Ten 7/8 diameter ASTM A325 high strength bolts designed

were used to connect the splice member to the gusset plate.

The load versus axial displacement response of Specimen

A 5

is plotted

in

F

Although the specimen geometry

was

significantly altered, the tensile re

Specimen A-5

is

similar to the other test specimens. contrast, the d

compressive behavior

is

evident. The compressive capacity

Specimen A-5

fraction of the compressive capacity of the other specimens the same plate

(Table 4.2). By comparing the load deformation response Specimen A-5

response of Specimens A-I and A-3, the decreased energy absorption capac

specimen geometry is evident. A photograph of the [mal yield line pattern

specimen is shown

in

Figure 4.29.


82/165

compression revealed the onset of yield lines along the specimen plate

welded boundary.

addition, strain gage readings also indicated

occurring in the plate in the region beneath the splice member. Ther

yielding had begun prior to reaching the expected nominal yield level. H

actual measured material yield level was significantly higher than the no

MPa, the 100 or nominal Whitmore yiel level cycles were actual

elastic range. Therefore, compressive yielding

h

begun well below

level based on the Whitmore critical stress theory.

Cycle 6 was loaded to the same

xi l

deformation level

s

Cycle 5. O

t

the maximum compressive load

o

905 kN revealed that the plat

visibly deformed out-of-plane and the strain gage readings in the 'pl

indicated significant plate bending. During compressive loading o Cyc

load o only 860 kN was reached. Specimen observations revealed t

buckle extending from the bottom comer

o

each free edge. The

themselves buckled

in

a region near the frame boundary, allowing the s

out-of-plane.

The testing was continued with cycles o increasing axial plate deformat

post-buckling behavior

o the gusset plate and to reach the tensil

specimen. Figure 4.30(b) shows the specimen load versus deformation


83/165

plot Figure 4.30 b).

As

the testing progressed, the maximum tensile loads reached increased as the

deformation level increased. The fIrst signs tension yielding on the gusset

observed about the sides

the splice member as the bolt holes yielded un

bearing deformations. As the cycle axial deformation levels increased, the te

lines about the connection region began extending outwards towards the free ed

specimen. Under compression, the capacity the buckled specimen deteriora

but steadily from the peak capacity attained during Cycle 5. As the co

deformations increased, yield lines indicated the presence a secondary pl

extending from the base

the splice member out towards the middle

the

edges. n addition, the increase in cyclic out-of-plane deformations resulted

the weld along the gusset plate to frame weld boundary, at the extreme outsi

the long welded side. Initially, a fracture length about I cm was visible. A p

the weld fracture under tension loading is shown in Figure 4.3

The tensile failure the specimen occurred during Cycle 14. However,

capacity the specimen was reached during the previous cycle when a maximu

1887 kN was attained. The test was concluded by loading the specimen in co

until an excessive level compressive deformation was achieved. As

deformations increased, the length the weld fracture grew as the plate was b


84/165

Figure 4.32 shows the out-of-plane deflected shape of the plate free ed

line deflection

the specimen recorded during Cycle 7 once

deformations

had

become significant The deflected shape the cen

that

a fixed-pinned but guided column. A hinge

is

evident on the gu

the base

the splice member.

At

this load level, the deformed shape

resembles the buckled shape a fixed-pinned but guided column, w

shape

the short free edge

is

closer to the buckled shape

fixed

column. The difference in deformation between the free edges is

increased

ftxity

provided to the top the short free edge

by

the close

splice member to that edge, as shown

in

the specimen geometry (Figure

is

a photograph the deformed plate free edges at the completion

observed that further deformation the specimen caused hinges

to

form

plate edges along the frame member boundary.

An

analysis

the specimen strain distribution during the assumed elastic

confirms the poor response of Specimen A-5. Figure 4.34 illustrates the

Specimen A-5 at both the 530 k N and +1060/-907 k N load level.

level, plate strains are highest under the splice member at the rosette loc

distribution

is

similar along both free edges with slightly higher strain

middle

the edges.

In

compression, the magnitude and distribution

s

unchanged from the +530 k N load level. However, the strain readings


85/165

rosette region.

n

addition strain readings at the middle the free edges n

hinge region and at the rosette locations are at or beyond yield levels. Gag

appear distorted due the effects significant plate bending.

-

J


86/165

T

e

M

a

e

a

P

e

P

a

e

E

a

c

Y

e

d

e

U

m

e

T

c

M

o

u

S

a

c

D

m

c

S

e

h

m

M

P

M

P

M

P

M

P

6

.

1

8

-

a

2

4

4

5

6

1

b

1

4

4

5

A

a

2

0

6

0

0

0

4

4

5

9

3

a

2

4

4

5

9

3

b

1

4

4

5

A

a

2

4

4

5


87/165

J


88/165

6

0

0

t

,

5

f

_

r

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

_

-

S

e

M

P

4

-

3

-

2

-

1

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

-

.

-

-

.

-

-

-

-

-

-

-

o

i

I

I

i

I

o

5

1

1

2

2

M

i

c

o

a

n

m

m

F

i

g

u

r

e

4

1

T

c

M

a

e

a

R

p

J

J


89/165

2

1

Z

J

I

-

0

J

I

r

1

f

)

1

2

5

3

1

3

5

7

9

1

1

1

R

a

D

e

o

m

a

o

o

G

u

P

a

e

R

m

b

y

m

m

F

g

2

L

v

A

a

D

s

p

a

m

e

n

R

p

C

e

G

P

a

e

m

b

y

S

m

e

n

A


90/165


91/165

11i 1

o

-1000

-2l l00

Cycle

1-7

Bolt

Slip

r \

J

J ;

>J

I

-5 -3 -1

1 3 5

?

•

11 13 15

Fbcial

Displacement

or

Gusset

Plate Assembly mm)

a

2l l00

Cycle

S :13

1000

:l IZI00

-2l l00

-5 -3

-1

1 3 5 ?

•

11 13 15

Fbcial

Displacement

or

Gusset

Plate

Assembly

mm )

b)

2l l00

Cycle :13-:17

1000

0


92/165


93/165

7

C

y

e

P

o

t

o

C

m

p

S

o

U

n

o

n

5

•

·

·

C

e

e

L

n

L

E

S

E

·

·

·

-

_

_

_

_

-

_

_

_

_

_

1

3

C

l

J

u

r

u

Q

1

o

_

T

o

u

a

e

5

o

1

2

3

4

5

L

o

m

m

)

F

g

6

O

o

p

a

D

s

p

a

m

e

cyclic behavior of steel gusset plate connections.pdf

Documents