Hancock Et Al 1994 DB Curves Paper

Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

J. Construct.

Steel Research 31 (1994) 169-186


@) 1994 Elsevier Science Limited
Printed in Malta. All rights reserved
0143-974X/94/$7.00
ELSEVIER

Strength Design Curves for Thin-Walled Sections


Undergoing Distortional Buckling

G r e g o r y J. H a n c o c k , a Y o u n g B. K w o n b & E. S t e f a n B e r n a r d a

"Centre for Advanced Structural Engineering, School of Civil and Mining Engineering,
University of Sydney, Sydney, Australia, NSW 2006
bDepartment of Civil Engineering, Yeungnam University, South Korea

(Received 5 January 1994)

ABSTRACT

Certain types of cold-formed sections, notably those composed of high-strength


steel, some decking sections and racking sections may undergo a mode of buck-
ling called distortional in which edge and intermediate stiffeners are inadequate
to prevent lateral movement of the flanges which they support. The existing
design methods in current standards and specifications are often not adequate to
account for distortional buckling. This paper describes proposed strength design
curves for thin-walled sections undergoing distortional buckling.
The design curves are based on testing and the test data base is included in the
paper. The influence of local buckling on distortional buckling is also discussed.

NOTATION

A G r o s s section area
A¢ Effective p a r t o f gross section area
b Plate width
be Effective p a r t of plate width
Fy Yield stress used in design
M~ Ultimate moment capacity
My M o m e n t at first yield
Z G r o s s section m o d u l u s
Z~ Effective section m o d u l u s

O'de Elastic d i s t o r t i o n a l b u c k l i n g stress


O'di Inelastic d i s t o r t i o n a l buckling stress

169
170 G. J. Hancock, Y. B. Kwon, E. S. Bernard

ale Elastic local buckling stress


am Test maximum strength (ultimate load/gross area)
amax Maximum design stress
Oy Test yield stress

1 INTRODUCTION

A mode of failure of thin-walled sections in compression and bending, in


which edge-stiffened flange elements of the sections deform by rotation of
the flange about the flange-web junction, or where the intermediate
stiffener in a flange moves normal to the flange, may occur in sections
composed of high-strength steel. This mode of failure has been called
distortional buckling. The distortional mode of buckling occurs at longer
wavelengths than local buckling and involves membrane displacements of
the edge or intermediate stiffeners forming the section. The distortional
buckling modes are shown for a rack section in compression, a channel
section in compression, a trapezoidal deck in bending and a Z-section in
bending in Figs l(a), (b), (c) and (d) respectively.
Research into the distortional mode of buckling has attracted consider-
able attention in recent years. Distortional buckling in steel storage rack
sections was first discussed by Hancock, 1 where simple design charts for
computing the buckling stress were presented. Further research including
testing by Lau and Hancock 2-4 resulted in a preliminary set of design

,u|-'" i

• !

(a) Rack Section in (b) Channel Section in


Compression Compression

Compression

" " " ~ .... . Compression Ne u_tral Axis


Neutra/ ~_Ax_is__
!

(c) "l'tapezoidalDeck in (d) Z-Sectionin


Bending Bending
Fig. 1. Distortional buckling modes.
Strength design curves for thin-walled sections 171

curves based on the Johnston parabola s for distortional buckling. These


design charts and curves have recently been incorporated in the Australian
Steel Storage Rack Standard (Standards Association of Australia6). Sim-
plified formulae for computing the elastic distortional buckling stress of
sections with edge-stiffened elements were also provided by Lau and
Hancock. 7 Charnvarnichborikarn and Polyzois 8 have modified the for-
mulae for elastic buckling to predict the distortional buckling stress of
Z-section columns. They have used the same design curves as proposed by
Lau and Hancock. 2 Their method accurately predicted the distortional
buckling strength of Z-section columns in compression. In neither the
method of Lau and Hancock, nor in that of Charnvarnichborikarn and
Polyzois, is allowance made for the postbuckling reserve of strength in the
distortional mode.
Early work on the distortional buckling mode in edge-stiffened elements
and intermediate stiffened elements was performed by Desmond, Pek6z
and Winter. 9'1° In these two papers the buckling mode was given the
name 'stiffener buckling mode' in cases where the stiffener was not
adequate to prevent its deformation in a plane normal to the element
which it supported. This is the same mode of buckling referred to as
distortional buckling in this paper. The design methods proposed for edge
and intermediate stiffened elements by Desmond, Pek6z and Winter have
been incorporated in the AISI Specification. 11 They account for the
inability of the stiffener to prevent distortional buckling by reducing the
local buckling coefficient for the plate element supported by the stiffener
below 4.0 and then including this reduced buckling coefficient in the
Winter effective width formula used to compute the strength of the plate
element. The design methods do not account for the restraint to distor-
tional buckling provided by the web of the section. The use of the Winter
effective width formula allows for postbuckling in the distortional mode.
However, as discussed in Kwon and Hancock, 12 the design method in the
AISI Specification is unconservative for distortional buckling of channel
sections composed of high-strength steel of yield stress 550 MPa, and so
alternative design methods are necessary to design against distortional
buckling in this case.
An alternative approach has been to account for the distortional mode
of buckling of edge and intermediate stiffeners as a compressed strut on
an elastic foundation where the elastic foundation is represented by a
spring which depends upon the bending stiffness of adjacent parts of plane
elements and on the boundary conditions of the element. This procedure
has been adopted in Eurocode 3 Part 1.3.13 The design strength of the
stiffener is based upon the conventional compression member design
curve in the Eurocode with ~ = 0-13. The method accounts for the elastic
172 G. J. Hancock, Y. B. Kwon, E. S. Bernard

restraint of all elements in the section including the web by incorporation


of their flexibility in the elastic spring restraint. However, the method does
not allow for any postbuckling reserve in the distortional buckling mode.
A detailed discussion of this method applied to channel sections is given
in Buhagiar, Chapman and Dowling. 14
A related problem is that of the buckling of the unsupported compres-
sion flanges of panel beams (decks with vertical upstands used to provide
interlocking between adjacent panels). The distortional buckling mode
involves rotation of the web-flange element about the axis of the web and
tension flange. A design procedure for this type of distortional buckling
has been provided by Serrette and Pek6z. ~5 The method is based on the
elastic buckling stress of the compression flange computed using formulae
for the flexural-torsional buckling of the web-flange element elastically
restrained by the tension flange. The strength design curve is the Johnston
parabola as used by Lau and Hancock 2 and Charnvarnichborikarn and
Polyzois. s The method does not allow for postbuckling in the distortional
mode, although allowance is made for the interaction of local and
distortional buckling.
The purpose of this paper is to compare two different sets of design
curves which have been proposed for distortional buckling and which are
applicable for sections buckling in the distortional mode with the test
results performed on a range of sections at the University of Sydney. The
design methods are applicable to sections subjected to compression or
bending, and can account for sections which buckle in the distortional
mode interacting with the local mode.

2 TEST RESULTS

A detailed test programme on a range of thin-walled cold-formed steel


sections has been performed at the University of Sydney to determine the
strength of sections failing in the pure distortional buckling mode, or
the distortional buckling mode interacting with local buckling (mixed
mode). The results of the tests are described for hat, channel and storage
rack sections with yield stress in the range 200-480 M P a in Lau and
Hancock, 2-4 for channel sections with and without intermediate web
stiffeners with yield stress in the range 585-640MPa by Kwon and
Hancock 12,16 for trapezoidal decking sections with intermediate stiffeners
and yield stress 650 MPa by Bernard, Bridge and Hancock 17-19, and for
Z-sections in bending with yield stress 450 MPa by Minnett 2° and O'Dey. 21
The tests by Lau and Hancock 2-4 were performed on sections for which
the distortional buckling stress was greater than half the yield stress and so
Strength design curves for thin-walled sections 173

no substantial postbuckling reserve was observed. However, the tests by


Kwon and Hancock 12'16 and by Bernard, Bridge and Hancock 17 19 were
performed on sections for which the yield stress was significantly higher
than the distortional buckling stress and so a substantial postbuckling
reserve of strength occurred. The tests by Minnett 2° and O'Dey 21 were
performed for sections for which the distortional buckling stress was
approximately half the yield stress and so failure occurred at approximate-
ly the distortional buckling stress.
The complete set of test results for the channel (CH), rack (RA, RL) and
hat (HA) sections shown in Fig. 2 is given in Lau and Hancock 2'4 and is
summarised in Table 1, where only those sections which underwent
distortional buckling in the tests have been included. The elastic distor-
tional buckling stress (ade) in Table 1 is based on the spline finite strip
analysis accounting for the fixed ends as described in Section 3 below. The
yield stress (try) is based on compression coupons. The last number in the
specimen nomenclature (e.g. 700 in CH17-700) is the specimen length in
mm.
The complete set of test results for the high yield strength channel
(CH1, CH2) sections shown in Fig. 3 is given in Kwon and Hancock 12'16
and is summarised in Table 2. The elastic distortional buckling stress (true)
in Table 2 is based on the spline finite strip analysis accounting for the

bf
1
~t
b bw] ] Front Rear
Front
Flange J

(a) SimpleLippedChannel(CH) (b) RackColumnUpright(RA)


bf
k
T
b Front Rear
/Flange ~ Flange

(c) RackColumnUprightwith (d) Hat (HA)


Additionallip Stiffeners(RL)
Fig. 2. Lau and Hancock test sections.
174 G. J. Hancock, Y. B. Kwon, E. S. Bernard

TABLE 1
Lau and Hancock Tests (Refs 2-4)

Test ,~,,° ~. ~ ,~,,,/,~y x/~y/a~°


section (MPa) (MPa) (MPa)

CH 17-700 307 295 406 0-725 1.15


CH17-1100 308 260 406 0-640 1.15
CH 17-1370 308 255 406 0-629 1-15
CH17-1640 300 248 406 0.610 1.16
CH17-1900 295 250 406 0-615 1.17
CH20-700 489 204 217 0.941 0.67
CH24-800 629 403 479 0.841 0.87
CH24-1100 562 375 479 0,783 0-92
RA 17-800 404 321 406 0.791 1.00
RA 17-1300 349 304 406 0.749 1.08
RA 17-1500 343 302 406 0.744 1.09
RA17-1700 324 292 406 0.719 1.12
RA17-1900 316 289 406 0.711 1.13
RA20-800 601 208 217 0.958 0.60
RA20-1300 472 206 217 0.949 0.68
RA24-800 662 413 479 0.862 0.85
RA24-1100 578 382 479 0.797 0-91
RL 17-800 322 307 406 0-756 1.12
R L 17-1300 320 288 406 0-709 1.13
R L 17-1500 320 287 406 0.707 I. 13
RL17-1700 321 280 406 0.689 1.12
RL 17-1900 315 262 406 0.645 1.14
RL20-800 571 217 217 1.00 0-62
RL24-800 808 410 479 0.856 0.77
RL24-1100 703 394 479 0.823 0-83
HA 17-800 278 262 406 0.645 1.21.
HA 17-1300 270 247 406 0.608 1.23
HA20-800 426 200 217 0.922 0.71
HA24-800 464 341 479 0.712 1.02

fixed ends as described in Section 3 below. T h e yield stress (ay) is based o n


the 0"2% p r o o f stress of tensile c o u p o n s . The s e c o n d last n u m b e r in the
specimen n o m e n c l a t u r e (e.g. 5 in C H 1 - 5 - 8 0 0 ) is the overall lip d e p t h d in
m m , a n d the last n u m b e r (e.g. 800) is the specimen length in m m .
The c o m p l e t e set o f results for the d e c k i n g sections (IST, TS3) s h o w n in
Fig. 4 is given in Bernard, Bridge a n d H a n c o c k 17 19 a n d is s u m m a r i s e d in
T a b l e 3, where o n l y those I S T sections (V-stiffeners) which u n d e r w e n t
d i s t o r t i o n a l b u c k l i n g first or d i s t o r t i o n a l b u c k l i n g i m m e d i a t e l y following
local b u c k l i n g are included. T h e TS3 sections (fiat hat stiffeners) which
Strength design curves for thin-walled sections 175

~ . - r (internalcorner1 - - [
=l radius) //

Et
uw .I
(a) Simple Lipped Channel(CH1)

K r (internalcorner ~-
= = t radius)
f

bw l .

(b) StiffenedLipped Channel(CH2)

Fig. 3. Kwon and Hancock test sections.

TABLE 2
Kwon and Hancock Tests (Refs 12 and 16)

Test cr.o am ~, ~r~,l~, ,f2,1~,o


section (MPa) (MPa) (MPa)

CH 1-5-800 46" 1 147'8 590 0"251 3-58


CH1-6-800 62-5 147.3 590 0"250 3"07
CH 1-7-800 76" 1 149-5 590 0"253 2"78
CH 1-7-600 81'2 155-7 590 0-264 2-70
CH1-7-400 83"8 160.0 590 0-271 2"65
CH2-7-800 87"0 198-5 590 0"336 2"60
CH2-7-1000 78"3 193.2 590 0-327 2-75
CH2-8-1000 94'9 192"1 590 0"326 2-49
CH2-10-1000 119"8 202"2 590 0'343 2"22
CH2-12-1000 144-6 206-2 590 0"349 2-02
CH2-14-1000 145'4 214.4 590 0"363 2'01

underwent local buckling well before distortional buckling are marked


with an asterisk (*). The elastic distortional buckling stress (ode) was
computed using a semi-analytical finite strip analysis as described in
Section 3 below. The yield stress (o-y) is based on the 0.2% proof stress of
tensile coupons. The second numeral in the nomenclature for the IST
specimens (e.g. 7 in IST 47A) is the approximate stiffener height S in mm.
176 G. J. Hancock, Y. B. Kwon, E. S. Bernard

b 1

I bd21
I~
i
' b4

(a) IST decking section (V stiffener)


b3
t- bl -I

_ ' b5

(b) TS3 decking section

Fig. 4. Bernard, Bridge and Hancock test sections.

TABLE 3

Bernard, Bridge and Hancock Tests (Ref 17-19)

Test ~rde Mu(test) My Mu/My x~y/tr~e


section (MPa) (kN m) (kN m)

IST 43A 141 3"85 8.92 0"431 2"15


IST 44A 190 4.00 8"88 0"450 1"85
IST 44B 130 3-68 8"95 0'411 2"24
IST 45B 176 3"88 8.90 0"436 1'93
IST 46A 282 4'59 9"26 0"496 1-52
IST 47A 273 4'56 9.01 0'506 1-55
IST 47B 244 4'57 9-06 0"504 1"64
IST 48B 296 4.54 9.14 0"497 1'49
TS3 A1 80 2-96 9'13 0"324 2.86
TS3 A2 104 2.93 9-17 0"320 2"50
TS3 A3 168 2"96 9"01 0'329 1'97
TS3 A4* 260 3'52 8"90 0"396 1"58
TS3 A5* 317 4'10 8"95 0"458 1"44
TS3 A6* 314 3'87 8"90 0'435 1"44
TS3 B4* 231 3"63 8.91 0-407 1.68
TS3 B5* 301 4"01 8.81 0"455 1'47

* Local buckling well before distortional buckling.


Strength design curves for thin-walled sections 177

TABLE 4
Minnett and O'Dey Tests (Refs 20 and 21)

Test ,0, o= ,, Om/~, x/~-~/a.e


section (MPa) (MPa) (MPa)

Z20015 N 18 283 302 450 0"67 1"26


Z20015 W 18 262 295 450 0-66 1'31
Z20015 W 15 224 267 450 0"59 1"42
Z20015 W 12 182 245 450 0"54 1-57

The last numeral in the nomenclature for the TS3 specimens (e.g. 6 in
TS3 A6) is the approximate stiffener height S in mm.
The complete set of test results for the Z-sections (Z20015) in bending is
given in Minnett 2° and O'Dey 21 and is summarised in Table 4. The elastic
distortional buckling stress (ado) was computed using a semi-analytical
finite strip method as described in Section 3 below. The yield stress (ay) is
based on the 0.2% proof stress of tensile coupons. The last number in the
nomenclature for the Z20015 specimens (e.g. 15 in Z20015 W 15) is the
overall lip depth in mm.

3 CALCULATION OF THE ELASTIC DISTORTIONAL


B U C K L I N G STRESS

A graph of buckling stress versus buckle half-wavelength for a rack section


column in compression is shown in Fig. 5. The half-wavelength of the
distortional buckling mode at point B is 350mm whereas the local
buckling mode at point A is 65 mm. Overall buckling, including flexural
and flexural-torsional buckling where the section remains undistorted,
occurs at longer wavelengths. Local buckling involves only plate flexure
with the line junctions between intermediate plates remaining straight.
Sections in bending may undergo distortional buckling of two forms, as
shown for the channel section in bending in Fig. 6. At point B the mode of
buckling is distortional of the same form as that shown at point B in Fig.
5, except that only the compression flange and web participate in the
buckling mode for the section in bending. At point C in Fig. 6, distortional
buckling occurs where the web bends with a lateral movement of the
flange. This type of distortional buckling occurs at much longer half-
wavelengths of the order of 2000-10 000 mm. This paper is confined to
those forms of distortional buckling of the type shown at point B in Figs 5
and 6. The longer wavelength distortional buckling, which involves
178 G. J. Hancock, Y. B. Kwon, E. S. Bernard

I I ~ I I III I t I I I I Ifll I I I I 1 IT

7O0
f
600
500 Local Distor tional
mode mode
40O
300
200
100

0 I L
10 100 1000 10000
Buckle Half-wavelength (ram)
Fig. 5. R a c k section c o l u m n : b u c k l i n g stress versus half-wavelength.

9OO I ! I I I !III i ! I !iI I r i II I'

800 Top flange in c o m p r e s s i o n • I


b o t t o m flange in tension
Distortional
700

r~
=~
m

i
~
8
600

5oo

300
IA • i I1!
m

.~_
r£1 buckle
I ] buckl[ - ibuck~e B (Flixural torsional)
100 I

o I I I II Illl I I i L I ilhJ i i i illi i L i i l iiIJ


10 100 1000 10000 100030
Buckle Half-wavelength (mm)
Fig. 6. C h a n n e l section: b u c k l i n g stress versus h a l f - w a v e l e n g t h for m a j o r axis bending.

transverse bending of the web, is discussed in more detail in two papers on


purlins (Rousch, Rasmussen and Hancock 22 and Trahair23).
The design method proposed in Section 4 below requires computation
of the elastic distortional buckling stress. The distortional buckling stress
Strength design curves for thin-walled sections 179

can be calculated using either a computer program based on the finite


strip or finite element methods, or analytical formulae specifically deter-
mined for distortional buckling. For the sections described in this paper,
the distortional buckling stress has been computed using the semi-analyti-
cal and spline finite strip methods of buckling analysis (Plank and
Wittrick, 24 Lau and Hancock, 25 Hancock1). The semi-analytical method
assumes simply supported end boundary conditions and is applicable for
longer sections where multiple buckle half-waves occur in the section
length. It has been used for the trapezoidal decks of Bernard, Bridge and
Hancock, and the Z-sections of Minnett and O'Dey. The spline finite strip
method is applicable to sections where the end boundary conditions may
have an influence such as for sections compressed between rigid end
platens. It has been used for the rack sections of Lau and Hancock and the
channel sections of Kwon and Hancock. The results of buckling analyses
for the CH1 channel section of Kwon and Hancock are shown in Fig. 7,
where BFINST denotes the semi-analytical finite strip analysis and
BFPLATE denotes the spline finite strip analysis. The BFINST analysis
gives the local (ale) and distortional (O'de) buckling stresses at given
half-wavelengths, whereas the BFPLATE analysis gives the actual buck-
ling stress of a given length of section between ends that may be fixed. The
abscissa in Fig. 7 for the BFINST analysis indicates one half-wave,
whereas for the BFPLATE analysis the abscissa indicates the real column

550
I I I I I II I I I I I I I IIII I I I I I II
500 m L ~ l,,~.al

450 D

400 B

350 m

30o
J
m

250 D

200 m

m
15o B

lOO D

50
i 1222221 il _ ._ c,)l

o
o 50 100 500 10G0 5000 10000
Buckle Half-wavelength or Column Length (ram)

Fig. 7. Finite strip buckling analyses.


180 G. J. Hancock, Y. B. Kwon, E. S. Bernard

length of the section. The different marks indicate local (L), distortional (D)
and overall buckling (FT), respectively. It can be observed in Fig. 7 that
where the sections buckle in multiple distortional half-waves, the change in
the buckling stress in the distortional mode with increasing section length
reduces as the effect of the end conditions reduces, and eventually
approaches the value produced by the semi-analytical finite strip method.
Simplified design charts have been presented for rack sections in
compression by Hancock.1 For rack, channel and hat sections in compres-
sion, simplified formulae for computing the distortional buckling stress of
sections with edge-stiffened elements were presented by Lau and Han-
cock. 7 For Z-sections in compression, formulae for the elastic distortional
buckling stress have been given by Charnvarnichborikarn and Polyzois. 8
Formulae for the elastic distortional buckling stress of deck sections in
bending are given in Serrette and Pek6z. xs Distortional buckling stresses
can be computed approximately using the elastically supported strut
model in Eurocode 3 Part 1.313 and Buhagiar, Dowling and Chapman. 14

4 DESIGN FORMULAE

4.1 Effective section approach for distortional buckling

The Winter f o r m u l a 26 is commonly used in design procedures to deter-


mine an estimate of the ultimate load-carrying capacity of plates in
compression. In its usual form, it is expressed as

be= a ~ l e ( l _ 0 . 2 2 a/~l~'] (1)


b ~Fy ~Fy}
where be is the effective part of the plate width b, ale is the elastic local
buckling stress and Fy is the yield stress.
Kwon and Hancock 12'16 proposed a modification of this formula to
permit application to the distortional mode of buckling for columns
undergoing essentially uniform compression. This was done in two stages.
At first, the elastic local buckling stress (the) was replaced by the elastic
distortional buckling stress (ade) to arrive at
h_
-° = 1 2 ~<0.673 (2a)
b

b-=~/ Fy 1-fiZZ 2>_-0.673 (2b)


Strength design curves for thin-walled sections 181

where

2 = _ [-:-: (2c)
~/ Ode
In this form, it was assumed that all plate elements forming the cross-
section were reduced to effective widths in the same proportions. This was
equivalent to

q=~F'~'-dy
A e
( 1-0"22/-~e~/) /Fy
O'de
(3)

where Ae is the effective part of the gross section area.


Kwon and Hancock found that this approach yielded unconserva-
tive estimates under certain circumstances and thus produced their 'De-
sign Proposal 2' in which the exponent of the (trde/Fy) term in eqn (2b) was
changed from 0.5 to 0.6 and 0.22 coefficient increased to 0-25,
thus:

be = 1 2 ~<0.561 (4a)

be (aaeX~°'6(1 (O'deX~0"6'~
k -0-25 / 2>_-0.561 (4b)

where

2=/~ (4c)
~/ One
As for eqn (2), this reduction of the effective widths was performed for all
the elements of the section and thus the reduction was equivalent to
reducing the gross area (A) to an effective area (Ae).
If a similar approach is applied to members in flexure, then, by direct
analogy with eqn (4b), the effective section modulus (Ze) is calculated
as

Ze (O'de~0"6(1 025(0"de~0"6~ O'u (5)


182 G. J. Hancock, Y. B. Kwon, E. S. Bernard

leading to the ultimate moment capacity

M u = F y Z e = o'uZ (6)

where Z is the section modulus of the gross section.


A comparison of eqns (3), (5) and (4b) with the test results is shown
in Fig. 8, where the ratios M u / M y and a ~ / a y are the measured values
given in Tables 1-4. The measured face yield stress (ay) has been used
in the calculation of My. In general, eqn (4b) provides a reasonable
mean fit to the test results of sections in compression, and eqn (5)
provides a reasonable lower bound to the sections in bending, although
the trapezoidal deck sections with flat hat stiffeners which underwent
local then distortional buckling (solid circles) are predicted unconser-
vatively.

4.2 Maximum stress formulae for distortional buckling

A similar design formula to that uscd by Chajcs et al. 27 for inelastic


flexural-torsional buckling was adopted for determining the inelastic
distortional buckling stress by Lau and Hancock. 2 The proposed formula

1.5 I I I
0 Lau and Hancock
e-
Q • Minnettand O'Dey
Q K w o n and Hancock C H I
+ K w o n and Hancock C H 2
¢:h
o Bernard Bridge and Hancock - V Stiffeners
E
O = Yield • Bernard, Bridge and Hancock - Flat Hat Stiffeners
.~ 1.0 OR ~1 Buckling First)
v o~,, • Bernard, Bridge and Hancock - Flat Hat Stiffeners
"o'~ (Distortional Buckling First)

~ , ~ ~Eqn. 3
; s 40.5
Eqns. 4b, 5 / v~

0 I I I
0 1 2 3
DimensionlessSlenderness FyvF ~ , ~y/Oae
Fig. 8. Comparison of test results with effectivesection design curves.
Strength desion curves for thin-walled sections 183

is based on the Johnston 5 parabola and is given by

Gdi ~ Gde
_<Fy
O'de~ 2 (7b)

where trde and trdi are the elastic and inelastic distortional buckling
stresses, respectively.
In Kwon and Hancock, 16 an extension of this design curve was
proposed for determining the strength of slender sections which are
formed from thin steel of high yield strength and which may buckle in the
distortional or mixed local-distortional modes in the elastic range of
material properties. The proposed formulae for the maximum stress (O'max)
(called 'Design Proposal 1' in Kwon and Hancock 16) are given by

---- (1--&) Fy
~,>~-~ (8a)

trmax=Fy 0"055 -3"6 +0"237 ]-~<trde~ 2

where trde is the elastic distortional or mixed mode buckling stress.


Equation (8a) is the same as (7a) except that amax has been substituted for
trdi. Equation (8b) is a parabolic fit to the Kwon and Hancock 12'~6 test
results. A comparison of eqns (8a) and (8b) with the test results in Tables
1-4 is shown in Fig. 9. In general, eqns (8a) and (8b) provide a reasonable
mean fit to the test results.

4.3 Influence of local buckling on distortional buckling

Local buckling can occur simultaneously with distortional buckling, or at a


higher or lower load. The tests by Kwon and Hancock 12' 16 were designed
to determine whether adverse interaction occurred if local and distortional
buckling were simultaneous or nearly simultaneous. No adverse inter-
action was found between local and distortional buckling for the channel
sections tested, so the distortional buckling strength can be assessed in-
dependently of whether local buckling is occurring simultaneously.
184 G. J. Hancock, Y. B. Kwon, E. S. Bernard

1.5 I I I
0 Lau and Hancock
• Minnett and O'Dey
[] Kwon and tlancock CH1
+ Kwon and Hancock CH2
o BernardBridge and Hancock - V Stiffeners
Yield • Bernard, Bridge and Hancock - Flat Hat Stiffeners
e 1.0 i

•,. ~0a~ ~ (Local Buckling First)


.)~0 ~ • Bernard, Bridge and Hancock - Flat Hat Stiffeners
J ~b~.~ ~ k (Distortional Buckling First)
/ °"L%•
Eqn.(8a) ~ o , Eqn. (8b)
0.5

Elastic Bucklingj ~ - ~ .... m

I I t
0 1 2 3 4
DimensionlessSlenderness ~/Fy~de, ~/Oy/Ode
Fig. 9. Comparison of test results with maximum stress design curves.

The tests of trapezoidal decks by Bernard, Bridge and Hancock 17 19


included sections which underwent local buckling before and after distor-
tional buckling. For the sections with V-stiffeners shown by the open circles
in Figs 8 and 9, distortional buckling occurring before, or in the approxi-
mate vicinity of, local buckling could be predicted using the design formulae
proposed for distortional buckling alone. However, sections which under-
went local buckling well before distortional buckling needed to be designed
using conventional methods to account for local buckling as reported in
Bernard, Bridge and Hancock.l 7,1 s For the sections with fiat hat stiffeners
shown in Figs 8 and 9 by the solid triangles when distortional buckling
occurred first, the design proposals in Sections 4.1 and 4.2 adequately
predict the strength of the sections. However, sections which underwent
local buckling well before distortional buckling (i.e. ate < 0"85 ade), as shown
by the solid circles in Figs 8 and 9, could not be predicted conservatively by
the formulae in Section 4.1, and a modified effective section method has
been proposed by Bernard, Bridge and Hancock. 19
The modified effective section method involves computing the effective
section using conventional methods based on local buckling alone, and
then using the maximum stress for distortional buckling as the limiting
stress rather than the yield stress. The method allows for postbuckling in
the distortional buckling mode.
Strenffth desiffn curves for thin-walled sections 185

5 CONCLUSIONS

Two sets of design curves for cold-formed sections undergoing distortional


buckling have been proposed and compared with the test results. The first
set is based on an effective section approach whilst the second provides a
prediction of the maximum stress in the distortional buckling mode
including the postbuckling reserve of slender sections. Both approaches
were found to produce a reasonable estimate of the results for sections
which underwent distortional buckling before or at the same time as local
buckling. For sections which undergo local buckling before distortional
buckling, a separate method accounting for local buckling should be
performed.

ACKNOWLEDGEMENTS

Technical and material support throughout the experimental programme


was generously provided by Lysaght Brownbuilt Industries Pty Limited,
Colby Engineering Pty Limited and BHP Research, Melbourne Labora-
tories.

REFERENCES

1. Hancock, G. J., Distortional buckling of steel storage rack columns. J. Struc.


Enginff ASCE, 111 (12) (1985) 2770-83.
2. Lau, S. C. W. & Hancock, G. J., Distortional buckling tests of cold-formed
channel sections. 9th International Specialty Conference on Cold-Formed Steel
Structures, St Louis, MO, 1988, pp. 45-73.
3. Lau, S. C. W. & Hancock, G. J., Strength tests and design methods for cold-
formed channel columns undergoing distortional buckling. Research Report
No. R579, School of Civil and Mining Engineering, University of Sydney,
1988.
4. Lau, S. C. W. & Hancock, G. J., Inelastic buckling of channel columns in the
distortional mode. Thin-Walled Structures, l0 (1990) 59-84.
5. Johnston, B. G., Guide to Stability Design Criteria for Metal Structures, 3rd
edn. John Wiley, New York, 1976.
6. Standards Association of Australia, Steel Storage Rackin#, AS4084-1993.
7. Lau, S. C. W. & Hancock, G. J., Distortional buckling formulas for channel
columns. J. Struct. Enffinff ASCE, 113 (5) (1987) 1063-78.
8. Charnvarnichborikarn, P. & Polyzois, D., Distortional buckling of cold-
formed steel Z-section columns, l lth International Specialty Conference on
Cold-Formed Steel Structures, St Louis, MO, 1992, pp. 353-78.
9. Desmond, T. P., Pek6z, T. & Winter, G., Edge stiffeners for thin-walled
members. J. Struct. Enffinff ASCE, 107 (2) (1981) 329-53.
186 G. J. Hancock, Y. B. Kwon, E. S. Bernard

10. Desmond, T. P., PekSz, T. & Winter, G., Intermediate stiffeners for thin-
walled members. J. Struct. Enging ASCE, 107 (4) (1981) 627-48.
11. Specification for the Design of Cold-Formed Steel Structural Members. Ameri-
can Iron and Steel Institute, Washington, DC, 1986.
12. Kwon, Y. B. & Hancock, G. J., Design of channels against distortional
buckling. 11th International Specialty Conference on Cold-Formed Steel Struc-
tures, St Louis, MO, 1992, pp. 323-52.
13. Eurocode 3, Cold Formed Thin Gauge Members and Sheeting,
CEN/TC250/SC3-PT 1A, 1992, 3(1.3).
14. Buhagiar, D., Chapman, J. C. & Dowling, P. J., Design of C-sections against
deformational lip buckling, l lth International Specialty Conference on Cold-
Formed Steel Structures, St Louis, MO, 1992, pp. 75-94.
15. Serrette, R. & PekSz, T., Local and distortional buckling of thin-walled
beams. 1I th International Specialty Conference on Cold-Formed Steel Struc-
tures, St Louis, MO, 1992, pp. 63-74.
16. Kwon, Y. B. & Hancock, G. J., Strength tests of cold-formed channel sections
undergoing local and distortional buckling. J. Struct. Enging ASCE, 117 (2)
(1992) 1786-803.
17. Bernard, E. S., Bridge, R. Q. & Hancock, G. J., Tests of profiled steel decks
with V-stiffeners. J. Struct. Enging ACSE, 119 (1993) 2277-93.
18. Bernard, E. S., Bridge, R. Q. & Hancock, G. J., Tests of profiled steel decks
with V-stiffeners. 11th International Specialty Conference on Cold-Formed
Steel Structures, St Louis, MO, 1992, pp. 17-43.
19. Bernard, E. S., Bridge, R. Q. & Hancock, G. J., Intermediate stiffeners in
cold-formed profiled steel decks. Part 2: 'Flat hat' shaped stiffeners. Research
Report No. R658, School of Civil and Mining Engineering, University of
Sydney, 1992.
20. Minnett, A. J., Stiffener buckling of cold-formed steel purlins. Bachelor of
Engineering Honours thesis, School of Civil and Mining Engineering, Univer-
sity of Sydney, 1985.
21. O'Dey, N. A., Stiffener buckling of cold-formed steel purlins. Bachelor of
Engineering Honours thesis, School of Civil and Mining Engineering, Univer-
sity of Sydney, no date.
22. Rousch, C. J., Rasmussen, K. J. R. & Hancock, G. J., Bending strengths of
cold-formed channel and Z sections restrained by sheeting. Proceedings,
International Workshop on Cold-Formed Steel Structures, Sydney, 1993.
23. Trahair, N. S., Lateral buckling strengths of unsheeted cold-formed beams.
Proceedings, International Workshop on Cold-Formed Steel Structures, Sydney,
1993.
24. Plank, R. J. & Wittrick, W. H., Buckling under combined loading of thin
fiat-walled structures by a complex finite strip method. Int. J. Num. Methods
Engng, $ (2) (1974) 323-39.
25. Lau, S. C. W. & Hancock, G. J., Buckling of thin flat-walled structures by a
spline finite strip method. Thin-Walled Structures, 4 (4) (1986) 269-94.
26. Winter, G., Thin-walled structures: theoretical solutions and test results.
Preliminary Publication of the 8th Congress, IABSE, 1968, pp. 101-12.
27. Chajes, A., Fang, P. J. & Winter, G., Torsional flexural buckling, elastic and
inelastic, of cold formed thin walled columns. Research Bulletin No. 66-1,
School of Civil and Mining Engineering, Cornell University, Ithaca, NY, 1966.

You might also like