Professional Documents
Culture Documents
Mo.~
?1
?1"
P
P
area of cross-section
area of heat-affected zone
Young's modulus of parent metal
Young's modulus of heat-affected material
tangent modulus
length of member
length of heat-affected zone
nondimensionalized
maximum
bending
strength of member about x axis (=Mx/M0.2x)
----Z~o0.2
knee factor of parent metal in RambergOsgood formula
knee factor of heat-affected material in
Ramberg-Osgood formula
axial force
nondimensional
maximum
compressive
strength of member (Pult/~o.2A)
ultimate compressive strength for nonwelded
member
0141-0296/92/040241-14
1992 Butterworth-Heinemann lad
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
Pu~l
O
Quit
Q*t
Qo.z~
r~, ~
o~
00.2
0o*.2
Oult
o~t
e1
beam (=(Mo.2x/Mcr) u2
, ~ky slenderness (L/r x, L/r>,)
nondintensionahzed
, ~y
slenderness
ratio
Oul t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
'z
"
Stress-strain c u r v e s o f a l u m i n i u m alloys
oI
Clearly the second term on the right hand side of equation (1) will be insignificant at low stress levels, the
curve effectively following the linear elastic line
(o = E . e). Computer programs that use an ultimate
strength approach to simulate buckling behaviour
usually require o as a function of e (i.e. o = f(e)).
Moreoever, since the standard computing technique will
be one of 'trial and error', the resulting nonlinear problem will require iteration techniques, e.g. direct iteration, Newton-Raphson method etc., to ensure the
required convergences. Thus it is computationally
much more convenient to modify the Rand~rg-Osgood
formula into a piecewise form (see Figure I). For the
inelastic and strain-hardening regions, dividing the
curve into approximately 50 pieces and 30 pieces,
respectively is sufficient for the difference in stress
between the original and piecewise form to be less than
0.1%. The tangent modulus Et, which may be determined from equation (2), can also be represented in this
manner. By adopting this approach the authors have
found that computer time is usually reduced by between
10 and 30 times as compared with direct use of equation
(]).
as
(1)
do
de
1 + ((O.O0___2n)~( o "~"-'
E
242
0"0.2
//\0r0. 2 / /
(2)
ae = 1 - [1 - 2-=/"] m
00.2
(3)
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
1.0
0.9
0.8
0.7
N 0.6
,~ 0 . 5
0.4
0,3
0.2
0.1
0
Figure 2
tO
15
20
25
30
Knee rector, tt
35
40
45
50
oe/oo.2
where
m --- 2.30 - 1.7512 TM]
n = knee factor in Ramberg-Osgood formula
The relationship between 0.elao.2 and n is shown in
Figure 2. If the knee factor, n, of the
Ramberg-Osgood formula is used to classify the alloy,
equation (3) may be simplified to
for
5 _< n < 10
0.e =
0.450"0.2
0.680.0. 2
o. = 0.81Oo.2
500
Porent moteriol
E = 7 0 0 0 0 ( N/mm 2)
o'(~2= 2 5 0 (N/ram 2 }
n =25
250
qA
20O
HAZ moterial
E
Z
=. 150
O3
= 70(XX)(N/mm 2)
a~. 2 = 125(N/ram 2)
n = 10
100
5C
Figure 3
Typical strees-strain
5
6
Stroin (%)
curves f o r
IO
II
243
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
Description of B I A X I A L p r o g r a m
,
JjMoterial (~
/Material @
F// o/A
'
Figure 4
1.0
0.9
0.8
0.7 T ~
0.6 ~ 0.5
_ | o m m
0.4
0.3
0.2
0.1
o
\.
..~,,,,.
tp
I
0.2
0.4
o.o
o.8
l.o
,.'z
,!4
1.6
,.~
I.I
l.O
0.9
0.8
0.7
i~, 0.6
0.5
0.4
0.2
0,2i
0.1
0
0
p-ZOOmm-~
_.
T~T,s.,~
~,x
Figure 6 Comparison of INSTAF and Hong 1 results (longitudinally welded columns)
244
Experimental verification
Strength of a/uminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
R= 2.3rnm
~5.4mrn
Cross-sectional properties
R = 3.2m~
f
= 1415.1
mmZ 1
/
2"x
rx
1.7996 x l0 s mm4~t~
/
35.7
mm J
IO2.2 turn
4.7 mm
---m
50.9 mm
~1
I 24.5mm_1
I-
-i
12.4 rnrn
._L
~" Fillet welds
Approximate cross-sectional
properties
A = 2548.1
mmZ /
Zx
rx
45.2
mm J
B e a m test results
using I N S T A F
b
Figure 7
Q0z, -
M0.2x
0.3
or
Q0.2~ -
Mo.2x
0.8
2Q
~1~
II II II
/
Roller support
II
t~ LOodspreOder
Beam
5 0 mm
~400
mm
50 mm
==
Span = I O 0 0 m m or 2 0 0 0 r a m
I
Figure 8
srnl
1--
245
i~
L 7"L
LTL = 1200 mm
LTL = 2200 mm
N - 1000-P- 1
N-IOOO-P-2
N -2000-P- 1
N-2000-P-2
W- 1000-L/2-1
W-1000-L12-2
W-1000-L/2-3
W- 1000-L/2-4
W-2000-IJ2-1
W-2000-L/2-2
W-20OO-L/2-3
W-2000-L/2-4
W-2000-L/4-1
~-i
Lwp
N-lOOO-P-3
LTL = 1200 mm
LTL = 2200 mm
(without welds)
~.. I"-
L wp
1
-(at mid-span)
6 0 0 mm
200
6 0 0 mm
~
50
i....i
--~ ~,--Lwp
200
W-2OOO-L/4-3
"~ F.--Lwp
(symmetrically at quarter-span)
150ram
150ram
r-!
I-1
W-10OO-E-1
W-IOOO-E-2
W-1000-E-3
25
W-IOOO-F-1
W-IOOO-F-2
W-2OOO-F-1
W-2OOO-F-2
W-2000-F-3
wW- 1000-F-4
1200
2200
W- 1000-F-5
(along the whole specimen)
Specimen designation convention is as shown below:
~
Condition of
welding
W-lOOO-L/2-1 ~ _
Span of beam
during test
Location of
welded plate
Sequence number
for specimen
Table 2 Parent metal properties, n = In 211n(oo.21oo.1). All tensile coupons were obtained from beam N-IOOO-P-1
E
O0.1
GO. 2
(]rult
(Et
Specimen
Location
(N/ram2)
(N/ram2)
(N/ram2)
(N/ram2)
(%)
~=t/00.2
N- 1
N-2
N-3
N-4
N-5
N-6
N-7
Flange
Flange
Flange
Web
Web
Web
Web
68500
69800
68300
72000
72900
70200
74600
363
376
358
373
370
366
352
370
385
366
360
378
374
359
432
431
420
436
439
434
415
10.4
11.6
10,4
10.2
9.0
9.1
10.6
36.3
29.3
31.4
37.3
32.4
32.1
35.1
1.168
1.119
1.173
1.147
1.161
1,165
1.156
Maximum
Minimum
Mean
74600
68300
70900
376
352
365.4
385
359
373.1
439
415
429.6
11.6
9.1
10.3
37.3
29.3
33.4
1.173
1.119
1.155
Standard
deviation
2175.2
7.8
8.2
8.1
0.73
2.15
2.20
1.89
7.11
Coefficient
of variation
(%)
246
Eng. S t r u c t .
3.07
1 9 9 2 , V o l . 14, N o 4
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
Table 3 Heat-affected properties, n = In 2/Io(o~.21~.~ ). (see Reference 6). All tensile coupons were obtained from top flange of beam
W-10OO-P-1
E*
~xil.1
o~.z
oust
E~
Specimen
(Ntrnm=l
iN/ram 2)
(NJRr~IR 2 )
(NJnlll~ 2 )
(%)
17
oru~t/ord.2
W-I
W-2
W-3
W-4
W-5
72900
72700
70700
69000
736OO
224
232
232
221
229
237
248
245
235
243
388
355
353
331
362
13.8
12.8
12.4
10.5
11.7
12.3
10.4
12.7
11.3
11.7
1.637
1.431
1.441
1.408
1.490
Maximum
Minimum
Mean
73600
69000
71780
231
221
227.6
248
235
241.6
388
331
357.8
13.8
10. 5
12.2
12.7
10.4
11.7
1.637
1.408
1.481
Standard
deviation
1691.6
4.4
4.9
18.3
1.10
1.94
2.02
Coefficient
of
variation
2.36
5.12
9.05
(%)
1.4
Q02,
are
1.2
I.C
Q0.2. = 68.1 kN
25.5 kN
.gl-- ----o--
~O.E
\
a..,. ,N=~,,.~
0.6
W - I000 - L/2- I
IV-IOOO-L/2-2
0.4
w- ,ooo-L/z-4
I000- L/2-3
.....
~Meaa
....
0.2
Minimum
~'~*~
0
I
'
|
2
I
:3
I
4
/
5
= 21.6 kN
I
6
!
8
I
7
6c/SPan x 100%
Figure 9
Compmison between test results and predicted loaddeflection curves at mid-span (specimen W-IOOO-L/Z
1.4
1.2
From Tables 2 and 3, it is clear that the mechanical properties for the parent and heat-affected material show a
certain variability. To take this effect into account, three
theoretical curves are shown in Figures 9 and 10, these
three curves being obtained by inputting the maximum,
mean and minimum values of 00.2 and o~.2, respectively
into the INSTAl: program. The effect of residual
stresses is neglected in the computer simulation.
1.0
~o.a
6
.~f
0.6
o W-2000-L/A-
~
j
0.2
76 rnm
~i
6.3mm
. W-2000-L/&-2
W - 2000- L/A-3
0.4
~,,
aeemde~g~tion
/~/~/
.....
Minimum
~ M e o n
. . . .
Maximum
~
~
I
Ac/spon x 100%
76 mm
4.3 mm
247
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
Table 4
Comparison between test results and theoretical results obtained using INSTAF program
Experimental
load
Specimen
designation
~c(max)
(mm)
Ao(max)
(mm)
Ornax
(mm)
N-10OO-P-2
N-IOO0-P-3
51.7
63.1
36.6
45.3
68.2
69.7
W-1000-4/2-1
W-1000-4/2-2
W-1000-4/2-3
W-1000-4/2-4
31.7
62.3
48.1
60.1
22.2
62.7
34.4
45.1
W-1000-E-1
W-1000-E-2
W-1000-E-3
61.1
65.9
64.4
W-IOO0-F-1
W-1000-F-3
W-1000-F-4
W-IO00-F-5
N-2OOO-P- 1
N-2000-P-2
030
(kN)
Mean of
experimental
Theoretical
load
Oso
(kN)
030
(kN)
Aso
(kN)
03o
(kN)
Oeo
(kN)
Difference
(%)
62,7
61.9
62.3
58.2
6.6
LB
LB
58.8
54.8
64.2
57.2
57.2
54.8
60.3
51.7
56.0
52.5
6.2
BF
LB
44.3
46.6
45.2
64.0
64.4
64.4
57.2
55.2
55.6
56.0
57.6
-2.9
LB
LB
LB
>51.9
>59.0
56.7
>54.0
>38.1
>41.0
>38.5
>38.5
>102.6
>104.2
>98.7
>97.9
88.5
87.7
89.3
89.3
88.7
80.3
9.5
NB
NB
NB
NB
126.7
128.7
79.9
82.2
23.0
23.6
20.0
20.5
20.3
20.0
1.5
LB
LB
W-2000-4/2-1
W-2000-4/2-2
W-2000-4/2-3
W-2000-4/2-4
208.5
155.2
105.3
>105.0
135.7
103.8
69.8
>67.8
20.7
23.6
23.7
>20.9
17.0
19.6
21.2
18.2
19.1
18.6
2.6
BF
BF
BF
BF
W-2000-4/4-1
W-2000-4/4-2
W-2000-4/4-3
>128.7
> 127.4
> 122.7
>83.0
> 81.9
> 78.2
>22.9
> 24.7
> 23.3
19.9
21.7
20,7
20.8
19.6
5.8
NB
NB
NB
W-2000-F-1
W-2000-F-2
W-2000-F-3
>153.4
> 144.6
>146.8
>98.5
>95.6
>95.9
>39.3
>38.6
>39.6
31,6
31.5
31,8
31.6
28.5
9,8
NB
NB
NB
Remark
LB
LB
030 corresponding applied lateral patch load when mid-span of beam deflects 3 0 mm
Qso corresponding applied lateral patch load w h e n mid-span of beam deflects 6 0 mm
Qmax maximum applied lateral patch load measured in test
Ac(max) maximum mid-span deflection of beam measured in test
Ao(max) maximum quarter-span deflection of beam measured in test
LB beam failed in local buckling
BF beam fractured during test
N8 no local buckling or fracture up to OmB~
248
Eng. Struct.
1992,
Vol.
14, No 4
the
commonly
used aluntinium
alloys,
Parent
E - - 70000 N/ram 2
00.2 = 2 5 0 N / r a m 2
n = 25
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
..J
I.I
-I
8 0 mm
Eulercurve
I I Ilmm
O.S
0.8
~Y/L~/L
= 0.0 (Nonwelded)
0.7
~\~/o.I
IQ.O.E ~ 0 . 2 ~
O.~
0.4
.0.3
~
0.2
---~.....~
0.2
o.,
7 mm
200 mm
Oo o15 ,Io
Figure 14
4'0
xx
;5
5.'0
(in-plane buckling)
I.I I
1.0
I
Figure 12 Cross-sectionused for parametric studies (BIAXIAL)
\ /.
Elastic criticx=lcurve
~
~L~ =0 (Nonwetded)
\Y"/L* = ,50 mm o, mid-height
" ~ / L
~= L (Wholly HAZ moteriol)
0.9
0.8
0.7
i~. 0.6
05
04
03
02
HAZ
O,I
0
E* = 70000 N/mm 2
(7~.2 = 125 N/ram 2
Figure 15
n* = 10
I
I000i
c~
L~'L
0,0
90C
C
O
E
F
G
0.1 At botheflds
0.1
0.2
0.3
1.0
OB
70C
60C
Goi
soc
40C
3O
20(:
IOC
0
ell
o!2 o13
~~OCoB
0.4
'
I. 0
I.Jl
Pult/Pult
Figure 13
I
0.5
Column
I .tO
curves
for transversely
L * = 5 0 m m at mid-height, f l e x u r a l - t o r s i o n a l
80C
I
1.5
welded columns
buckling)
249
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
Curve
I000
900
800
700
600
500
40C
50C
200
I00
0
A
C
ED
F
L~/L
B A
0.0 (Nonwelded)
o i At i~thI~l=
0.1
0.2
0.3
i.o (WhollyHAZmoterlol)
D C
011
I
0.2
0,3
0!4
1.5
0.6
0.7
0.8
09
A
1.0
Quit/Quit
Figure 16 Effect of location and extent of H A Z on bending strength
ot b e a m s (in-plane bending)
I.I1.0.
o.9o8
\x\
0.7
L'~=0 ( Nonwelded)
L
= L ( Wh011yHAZ rnoterlol)
,~ 06
0.5
0.4
0.3
0.2
Ol
00
Or,5
Figure 17
mid-span,
1.10
1!5
S t u d y of transversely w e l d e d b e a m s L * = 50 m m at
Iflexural-torsional
buckling)
oo 1:
,f
Table 5
Lc
(mm)
954
1590
2226
250
tL
~o
\\\\
0.9
o.e
0.7
,~o.6
o.
0.4
L~ = 0 (Nonwelded)
~30 mm
60 mm
*= L ( Wholly HAZ moteriol)
o.3
02
O.o
I
0.5
I
1.0
I
t.5
Figure 18
)~x
(column)
30
50
70
Eccentric load -
No H A Z
L* = 30mm
WhollyHAZ
P.It(kN)
Pu*~l (kN)
P~it2(kN)
301.7
282.1
254.7
191.3
190.7
180.1
158.7
140.3
126.7
Figure 19(a)
Pu*kl
Pu*it2
P.=t
Pu~t
0.63
0.68
0.71
0.53
0.50
0.50
No H A Z
L* = 30 mm
Wholly HAZ
Pult(kN)
Pu~tl (kN)
P~t2(kN)
298.5(278.7) + 183.5
280.9(236.3)
180.5
253.5(158.5)
173.9
155.5(139.5)
138.1(118.7)
125.7(100.1)
Pu~tl
P~tl
Pult
P~t
0.61
0.64
0.69
0.52
0.49
0.50
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
Table 6 Effect of partially HAZ affected cross-section on member strength
Degree of softening
on cross-section
Cases
|_
135mm'---~
r-
-~ ~5mm
80ram
J_ 1444m I
r'-
"-1
(kN m)
Qp (kN)
(Plastic
theory)
OplO*,,
43.55
16.75
46.43
0.938
0.3
38.97
12.35
34.21
1.139
0.5
40.07
14.14
39.17
1.023
0.5
34.71
11.84
32.81
1.058
0.5
32.11
11.28
31.26
1.027
1.0
31.55
10.39
28.75
1.097
1.0
31.25
10.39
28.75
1.097
Mp
A/A *
O,~,(kN)
(INSTAF)
0.0
l Sm~m
144.4 mm
. . . .
U...-I-
-~-- 70mm-~--2.4mm
J'=~--135 mm
iP"/J.fJJJ.f JJJiJJ~l
,c,
9mm
I
144. 4 mm
mm-~-
~--70
Il'MIl
9.4 mI
144.4
(e)
mm
"~
144.4 mm
T~f
(g)
0
~..
---~.
f f/il//~
~.~
Ill
.5.7n,~
I #3 --~ii--t
- - ~ - 7 0 m m - - ~ , 5.7ram
"
v#'JJ
144.4 mm
mm
f/JJ/J/JI
J Ar~r
rJ ~ / L / / J
Arl
/ / ill
k ,444mm I
251
0~"
Column
section
'L*I
I- i
~T
I,--76 m r r ~ ._L
L*
= 3 0 mm
I,,,-76 m m q,I L
'mmt/ "I4';m"
/
l l[4'3mm
J--,
Lc
Beom
section
152rnm
J.
L * : 30mm
_L ";J ~.
~k.,
I_
r-
~l
954 mm
I_
r-
954mm
U
--q
Figure 19 Welded f r a m e e x a m p l e
(a)
Condition
Quit~Quit
Qp (kN)
82.80
1.0
69.64
45.54
0.55
43.13
60.86
0.73
60.80
61.90
0.75
69.64
44.50
o. 54
43.13
Quit (kN)
(D
Z~
(b)
~ (_b
,,
,,
,,
"
0
(c)
,, @
,,
,d,
(D
"
;"
0
(e)
252
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
(3)
(4)
(5)
(6)
Acknowledgements
This work forms part of a project funded by RARDE;
the authors are grateful for assistance from Mr. D. Webber and Dr. P. S. Bulson.
References
Conclusions
Computer programs have been prepared that permit the
in-plane and out-of-plane response of aluminium
members containing transverse welds to be studied.
From the parametric studies undertaken the main findings are
(1) For end-welded columns, it is unsafe to neglect the
softening effect even if the dimensions of the HAZ
are small.
(2) The maximum reduction in column strength will
occur when the HAZ is located at the mid-height of
the column. For the column curves, the behaviour
of centrally-welded columns is quite similar to that
6
7
8
9
10
253
Strength of aluminium members containing local transverse wells: Y. F. Lai and D. A. Nethercot
11 Gilson, S. and Cescotto, S. 'Experimental research on the buckling
of aluminium alloy columns with unsymmetrical cross-section',
University of Liege, Belgium (undated)
12 Gilson, S. and Cescotto, S. 'Experimental research on the buckling
of aluminium alloy columns with unsymmetrical cross-section complementary tests', University of Liege, Belgium (undated)
254