?

Thin- Walled Structures 18 (1994) 225-246

© 1994 Elsevier Science Limited
i ,il Printed in Great Britain. All rights reserved
0263-8231/94/$7.00
ELSEVIER

Theoretical and Experimental Investigation of the Collapse

Behaviour of Transversely Stiffened Aluminium Alloy Plate
Girders

K. E. P. Brown
Cleveland Structural Engineers Ltd, Darlington, Co. Durham, UK

H. R. Evans
School of Engineering, University of Wales College of Cardiff, Cardiff, South Glamorgan,
UK
(Received 22 March 1993; revised version received 1 July 1993;
accepted 6 November 1993)

ABSTRACT

This paper reports on part of an ongoing research programme into the

collapse behaviour of lightweight aluminium girders, suitable for transpor-
table bridging systems. It describes the theoretical and experimental inves-
tigation of the behaviour of the transverse web stiffeners of welded
aluminium alloy plate girders loaded predominantly in shear. The theory is
based on a mathematical model that assumes a non-linear shear stress
dtstribution at the boundaries of the adjacent web panels. The experimental
investigation consists of a series of tests in which lO fabricated plate girders
were loaded to failure; careful tests of material properties are also reported.

NOTATION

As Cross-sectional area of the stiffener

d Depth o f web panel
E Young's modulus of web material
Plastic m o m e n t capacity o f the flange
Proposed theoretical force exerted on the stiffener at mid-depth
allowing for the effects of shear lag

225
226 K. E. P. Brown, H. R. Evans

Ps Proposed theoretical force exerted on the stiffener in accordance

with the three modes of response
t Thickness of web plate
Vexp Experimental shear load at failure of the girder
Vf Theoretically predicted shear failure load
Xr Distance between plastic hinges at failure (see Fig. 1)

7 Dimensionless constant for stiffener rigidity

0 Inclination of membrane tension field developed in web plate
O"t Diagonal tensile stress
O'y Yield stress of web material
~cr Shear buckling stress
"Ey Shear yield stress

1 INTRODUCTION

Plate girders have normally been fabricated from steel but other materials
can be used in their production. Following the demands of the aircraft
industry for a lightweight material in the early part of this century, the
production of aluminium was increased. This production increase was
accompanied by a corresponding increase in facilities and production
capacity for high strength aluminium alloys; a high strength aluminium
alloy becomes a very attractive material when a high strength/weight ratio
is required. As all permanent civil engineering structures have to support
applied loads in addition to their own weight throughout their design life,
where heavy loads exist and large sections are required, it would be
uneconomical to use plate girders of aluminium alloy in preference to
those fabricated in steel. However, aluminium alloy plate girders have a
particular use in the aerospace industry, the design of containers, and
military or any transportable bridging.
Earlier tests on aluminium alloy plate girders have indicated that their
mode of failure is similar to that for steel plate girders, i.e. by the forma-
tion of a shear sway collapse mechanism. However, unlike steel plate
girders, the webs of aluminium alloy plate girders may fracture due to the
effects of the heat affected zone adjacent to the perimeter welds.
Previous tests conducted on aluminium alloy plate girders have been
primarily concerned with predicting the ultimate load capacity of web
panels. This paper deals with the analysis of the load on transverse web
stiffeners and the interaction between web panels and stiffeners. The
analysis is based on a mathematical model presented in detail elsewhere
and which can only be outlined in the present paper.l' 2
 Collapse behaviour of lightweight aluminium girders 227

11
C< ~c~ I 1
Iv13~ 1

~r [ ~
x'~C

MO~2

__1 ~cr

Mpf Mpf

-'2- R
2

ltllll

cYfI Xf bl3OE 3
2
II~II
RtA2 2
t,l I t t,t, ~.o° "on++
k"W
Mpf Mpf
Fig. 1. Development of stresses in web panel.
228 K. E. P. Brown, H. R. Evans

2 THEORETICAL DEVELOPMENT

2.1 General

To appreciate transverse stiffener behaviour fully an understanding of web

panel behaviour is necessary. The mathematical model on which the paper
is based identifies three modes of response experienced by a web panel
during loading.
The first mode takes place before elastic buckling of the web panel
occurs. Here the shear force is sustained by a uniformly distributed shear
stress at the web boundaries, which increases linearly until the buckling
stress is reached.
In the second mode, where the applied shear load is greater than that
load required to produce shear buckling of the panel, the web panel is
assumed to act as a series of diagonal struts whose capacity is governed by
the Euler formula for a pin ended strut. Based on the Euler formula, the
stress distribution at the boundary of the web panel is non-linear and
varies parabolically and inversely with the square of the diagonal strut
length.
For the third, and final, mode, any further increase in the applied shear
load cannot be sustained by an increasing shear stress at the web panel
boundaries since the web has already yielded in shear. The additional
applied load is resisted by the development of a tensile band which induces
the transfer of tensile stress between the web panel and the flanges as well
as the adjacent web panels. The diagonal tensile stress in the web panel is
resolved into its vertical and horizontal components which anchor them-
selves against the flanges, adjacent web panels and stiffeners.
The development of stresses along the web panel for the three modes
identified are shown in Fig. 1.

2.2 Stiffener analysis

The mathematical model for panel behaviour evaluates the forces on the
transverse web stiffener in accordance with the three modes identified
during the loading of the web panel. As each mode occurs the load on the
stiffener is determined.
Up to the end of the first mode, assuming adjacent panels are identical
and without initial imperfections, the resultant in-plane force applied to
the intermediate transverse stiffener, Psi, is zero as the forces acting are
equal and opposite, as shown in Fig. 2.
During the second mode a non-linear shear stress distribution exists at
the web panel boundaries and a resultant compressive force, Ps2, is exerted
 Collapsebehaviouroflightweightaluminiumgirders 229

Psl
stiffener

1 II tl I
1 tl 1 1
1
J web ~1 tl web t
rl 1~4 t
1
1 tl tl t
tl Llll t J_
Z
I-- b -I Psl t- b -4
Fig. 2. The forces on the stiffener during the first mode.

on the stiffener which is a m a x i m u m at the mid-depth of the web panel.

The stresses applied to the web panels and stiffener are then as shown in
Fig. 3. The resultant force on the stiffener is determined by using the
principle of an effective shear depth and integrating the stress distribution
over the depth of the stiffener. The expression for Ps2 thus becomes:

Ps2=zydl W/~r+
"[Y Jfd/2
d~ Zcr
Q~2t" dy - fd~2 "[cr (~2 t. dy

= rydt.~ - 3rcrdt +
zcrdt - dt [ 2 ~ - 3rcr]
V ~y
W ~Y

Ps2
rcr

rl fl ~I
I1 tl
N.A,
VI

~cr ry
~2
Fig. 3. The forces on the stiffener during the second mode.
230 K. E. P. Brown, H. R. Evans

A further increase in the shear force applied to the web panel causes the
formation of a tensile band during the third mode. The web material
eventually yields in tension and the vertical component of the diagonal
tensile band exerts a compressive force on the stiffener by pulling on the
flanges. Under the action of the tensile diagonal stress field plastic hinges
begin to develop in both top and bottom flanges.
The compressive force exerted in addition to the load on the stiffener at
the end of the second mode is:
a t
y.x,-, t (I)
From standard tension field theory the distance between the plastic
hinges at failure, Xf (see Fig. 1 mode 3), is given as:
2
Xr - sinO Vatt

when the inclination of the tension field 0 = 45 °, this reduces to the

following expression:

X,- = 2 ~/M_p~ (2)

Vat't
The term, at, in expression (1), is determined by applying the Von-Mises-
Hencky yield criterion, which may be expressed as:

aa + a~0 + 90) - a0a/0 + 90i + 3r~ = a;

to the stress condition shown in Fig. 4. Based on the stress distribution
shown the components of stress are evaluated as:
ao = r,, sin 20 + a~

a(o+oo) = 0
750 ~- 0

Thus, the tensile stress, ay, takes the form:

at = a>-rysin20

and when 0 = 45 ~ and assuming ry = a y / V ~ , the expression for a t becomes:

at = 0.4226 ay (3)
Thus, from eqns (1)-(3) the force that is exerted on the stiffener at failure
of the girder, P~, may now be expressed as:
P~,~ = Ps2 + X/0"8452 Mpf. ayTtt
 Collapse behaviour of lightweight aluminium girders 231

(r-rot}

7 .0.~qo f

z ~t %,qo

t z'-rcr)
Fig. 4. The post-buckling behavour of a web panel as postulated by standard tension field
theory.

Using available results 3 from previously conducted tests on aluminium

alloy plate girders, experimental information was obtained about the load
exerted on the stiffener during the loading of a girder. When the theory for the
new mathematical model was applied to the specific test the predicted load on
the stiffener was found to be in excess of that measured experimentally. The
reason for this is that the total force applied is not entirely taken by the stif-
fener alone, but instead a portion of the adjacent web panels also contributes
to sustaining the load. The following derivation uses the phenomenon of
shear lag to explain the load distribution and to determine what effective area
of web contributes to sustaining the load in association with the stiffener.

2.3 Analysis of the load on the stiffener

F r o m the theory of elasticity 4 it can be shown that if a load is continu-

ously distributed along a length, or in this case, along the depth (d) of a
rectangular beam of narrow cross-section, then the stress function may be
expressed in the form of a polynomial.
The equation for the stress function is:
O4~b 2~4q~ 64~b
~X 4 ~- 6X
- 2- 6y 2 ~- 6y
-- 4 -0

The expression above may be satisfied by expressing ~b, in the form:

=
sin m rex
232 K. E. P. Brown, H. R. Evans

For the case of stiffeners under shear action the equations are written in
terms of shear stress. The distribution of shear stress over the depth of the
stiffener is defined by:
mrcx
Zxy : "~oCOS - ~ -

where m = 1 for the simplest wave pattern.

The stress model in Fig. 5 represents an anti-symmetrical stress system
and applying the relevant theory of elasticity the stress in the x-direction is
calculated. Based on that stress the strain in the x-direction at mid-depth
of the stiffener can be written as:
" 7zb
7r cos h 7rb _ _2. 2 COS h 2 ~-~
+~-~ c o s h
d 2d b rob
dzot rc sin h 2---d (4)

rCsinh ~b 2 ~b r ~ b C ° S h ~
d 2d b cos h ~-~ + 2d sin h ~-~

The strain at mid-depth of the stiffener can also be written in terms of

the theoretical applied load and the modelled stress condition. Hence,
based on the forces exerted on the stiffener shown in Fig. 6 the strain at
mid-depth of the stiffener equates to:

- P~-2t] rocos dx - Ps-2tz0 • (5)

Web Pa r~l..,.,.,~

- ~ cc6Tr~
1

-Y " ~ t - - ]

Fig. 5. Anti-symmetrical stress system model for a web panel.

 Collapse behaviour of lightweight aluminium girders 233

where E is the m o d u l u s of elasticity; and As is the cross-sectional area of

the stiffener.
Equating expressions (4) and (5), the expression for the m a x i m u m shear
z0 becomes:
Ps
Z0 --
2td
- - + ~As

where 7 is written as:

_ 2.2cos h 2 ~zb
b 2d
n rrb 2 rob rtb ~z h2 r~b
sin h 2 2--d- b cos h 2-d" sin h ~-~ - ~ cos 2--d

Considering the force(s) on the stiffener, as shown in Fig. 6, the force at

mid-depth becomes:
d/2 gX
Pc = P s - 2t f
,/o
z o c o s - - d . d X = Ps - 2t z-2°d
g

i rcYAs ]
PC = es "
~z)'A~l

where Pc is the theoretical force at mid-depth of the stiffener.

stiffener
~ f

web web

"
!
I
g
Fig. 6. F o r c e s e x e r t e d o n the stiffener.
234 K. E. P. Brown, H. R. Evans

3 EXPERIMENTAL P R O G R A M M E A N D G I R D E R DETAILS

This series of tests on aluminium alloy plate girders was conducted to

investigate the collapse behaviour and the loads imposed upon both
single- and double-sided transverse web stiffeners. The girders were of two
distinct sizes, having the basic layout shown in Fig. 7; values of the speci-
fic dimensions and material properties for the different girders are given in
Table 1.
The girders were loaded predominantly in shear and loaded to failure
while carefully monitoring the performance to test the assumptions made
in the mathematical model.

3.1 Testing procedure and instrumentation

Each girder was simply supported at its ends on roller supports and
subjected to a continuously increasing vertical point load applied
centrally. All the girders were instrumented with electrical resistance strain
gauges to monitor the developing strains. Where possible, strain gauges
were placed at the same position on opposite faces of the material to
calculate bending and membrane stresses. A continuous record was also
kept of the load/deflection response. In addition to studying panel and
stiffener behaviour the distortion of the compression flange was carefully
monitored.

3.2 Material properties

The material used to fabricate the girders was 7020 aluminium alloy. To
determine the strength of the virgin material, extensive tensile tests were
conducted on web, flange and stiffener material; the values obtained are
summarised in Table 1.
The susceptibility of aluminium alloy to the heat of welding, thus
producing a reduced strength in the material adjacent to the weld, is well
known and documented. Within the heat affected zone (HAZ) the
strength of the material varies from a minimum near the weld to the full
strength of the material at a certain distance away from the weld. The
determination of the extent of the HAZ has been approached by various
methods, each focusing on specific parameters that may govern the extent
of the HAZ. In the present study, for a single fillet weld, zones of low
hardness were quite well defined as shown in Fig. 8. However, as the
complexity of the weld increased by introducing several heat sources and
several heat paths, the extent of the HAZ was not so clearly defined, as
shown in Fig. 9, where a cruciform specimen was studied. The results
 Collapse behaviour o f lightweight aluminium girders 235

~, ,, q

g- g-

.~ ×

I I ~j
::~:--I
t

L_
-- = - ~ - z : l

r-=±-----I

I la,

tl N II --t--I
 tO

TABLE 1
Dimensions and Material Properties of Test Girders

Girder test panel Web details Flange details

d b t ¢70.2 tfcomp bfcomp iftens bftens 0"0.2f
(ram) (ram) (ram) (N/mm 2) (mm) (mm) (ram) (ram) (N/mm 2)
Test 1 P3 895 443 3.23 334.2 12-75 175-2 12.75 176.1 338-7
ATG7-1d P4 895 443 3-23 334.2 12.75 175.2 12.75 176.1 338-7
Test 2 P1 895 442 3-23 334-2 12-75 175.2 12.75 176.1 338.7
P2 895 442 3.23 334.2 12.75 175.2 12.75 176.1 338.7
Test 3 P3 896 442 3.23 334.2 12.74 175-3 12.72 175.8 338.7
A7G7-2d P4 896 442 3.23 334.2 12.74 175.3 12.72 175-8 338-7
Test 4 P1 895 445 3-23 334.2 12.74 175.3 12.72 175.8 338-7
P2 895 433 3.23 334.2 12-74 175.3 12-72 175.8 338.7
A7G7-1s Test 5 P1 895 442 3.23 334.2 12.75 175-9 12-76 176.1 338-7
P2 895 442 3.23 334.2 12.75 175-9 12.76 176.1 338.7
A7G8-2s Test 6 P1 895 441 3-23 334.2 12.70 175.4 12.76 175.6 338.7
P2 895 445 3.23 334.2 12.70 175.4 12.76 175.6 338-7
A7G9-2s Test 7 P1 604 309 3.23 334.2 9.60 130-1 9-62 130.2 341.9
P2 604 311 3-23 334-2 9.60 130-1 9.62 130-2 341.9
A7G9-1s Test 8 P1 605 308 3.23 334.2 9-61 130.4 9.60 130.5 341.9
P2 605 310 3.23 334.2 9.61 130.4 9.60 130.5 341.9
A7G9-1d Test 9 P1 603 310 3.23 334.2 9.60 130.6 9-62 131.1 341.9
P2 603 310 3.23 334.2 9.60 130.6 9.62 131.1 341.9
A7G9-2d Test 10 PI 601 311 3.23 334.2 9.62 130.2 9.62 130.4 341-9
P2 601 309 3.23 334.2 9-62 130.2 9-62 130.4 341.9
 web
h Outlineof darker~d arm
J re.re(lied by etch

2 mm
. . . . . . . . . . ,, . _ ~ _ . ./,--~ . . . . . . . . ", . _ . - I - . T
flange " '
, .// /
t .......

V i c k e r s Hardness N~mber
1~0.
I
I
145. I
I

140. m P',
"q I
ts "'-, ,T,, ,,,, - I,,,
#DI-
135. I II ,I. / N
',~, i J

130. m lJa~
Vl I
t !
12~. i

120. i
I

115. .~.
i
110.

105

l°°eo' ~o ~o ~o ~o 40 ~o ~o io LIO L20 L30 ~40 L~O L60 LTO LSO LSO
Distance Prom centre oP weld (mm)

Fig. 8. V a r i a t i o n in hardness across the specimen.

--,.I
238 K. E, P. Brown, H. R. Evans

I
,g
T
P~
i
.J
/
b "1

i+
= '1 "0

o-~ ~

i"

i~ + ~ d ~ + ~ o

in o IN o ii~ o ii~
 Collapse behaviour of lightweight aluminium girders 239

obtained gave no direct relationship between strength and hardness so

some of the test pieces were machined into standard tensile test specimens.
These material tests provided information to reinforce the work of other
authors 5 as to the extent of the HAZ. In particular, these results high-
lighted the importance of ensuring good quality welding and that for each
particular weld arrangement the extent and severity of the HAZ could be
most accurately determined experimentally. For those tensile test pieces
formed from the cruciform test samples fracture was observed to occur in
the weld material and not where expected, in the zones of minimum
hardness.

4 RESULTS

The experimental study produced a large amount of information which

has been fully discussed elsewhere2; only the main points will be briefly
described here.
In developing the theory for the load on a stiffener, tension field action,
as found in steel plate girders, is assumed to occur. However, the idea that
the tension field forms a clearly defined band is observed not to be entirely
true. This is because that, after buckling of a web panel occurs, the web
panel still has some capacity to sustain additional compressive forces due
to its shear resistance; however, the major additional load resistance is
provided by the developing diagonal tensile membrane field.
In Fig. 10 the development of the principal strains with the applied load
is shown for a position on the web panel near the idealised tensile zone. In
evidence are the substantial compressive strains that continue to increase as
the loading increases although the tensile strains are predominant. This
indicates that the basic assumption that beyond the buckling load of the
web panel all the compressive resistance is exhausted and the diagonal
tensile band is the only form of load carrying mechanism is not entirely true.
The development of strain shown in Fig. 10 is characteristic of all the
tests conducted although the relative magnitude of the tensile and
compressive strains is dependent on the position of the strain gauges upon
the web panel.
One objective of the tests was to validate the basic assumption of the
proposed mathematical model, viz. that a non-linear shear stress distribu-
tion is developed at the web panel boundaries. This provides the basis for
the analysis of the loads on a stiffener. In some of the tests conducted the
web panel was heavily instrumented adjacent to the compressive flange to
study the shear stress distribution. Typical measured values from such a
test are shown in Fig. 1 1.
 4~
GO0. AppL~.edLoo.d (kN)

551

500.

450

400.

350

300.

~o

0_~
"4000 13500 !3000 22500 12000 21500 "~I000 2500 0 ~oo ¢ooo Csoo ~ooo ~soo ~ooo ~50o -~ooo
Micro-Stroin Compression Tension Micro-Strain

Fig. 10. Variation of principal strains.

Shear S~ress (N/am2)
240

220

200.

180.

e~
160. e~

140.

120.

! 00.

80.

60.
O~

40l

20.

C e n t r e Pos~ S~iPFener

Fig. 11. Typical development of shear stress on top edge of web panel.
242 K. E. P. Brown, H. R. Evans

The shear stress distribution along the top edge of the web panel closely
resembles that postulated by the proposed mathematical model by exhi-
biting a distinct plateau. Although the instrumented sub-panel of the web
was not isolated, the rigidity of the intermediate transverse web stiffener
was sufficient to allow the forces developed in the sub-panel to anchor
against the stiffener: the shear stresses in the web were then able to
increase up to the yield stress value. The measured value of the web shear
stress at the central stiffener was approximately five times the theoretical
buckling stress of the unstiffened panel; this was mainly due to the
increased resistance to local web buckling provided by the intermediate
stiffener. However, the minimum measured value of shear stress was only
slightly higher than the theoretically calculated shear buckling stress.
The influence of the intermediate stiffener is of great importance to the
behaviour of the web panels. When stiffener failure occurs at an early
stage as load is applied to the girder, the two web panels adjacent to the
failing stiffener effectively become one which then exhibits the behaviour
of a single panel. The behaviour of the web in such a case is illustrated in
Fig. 12 which shows the orientation of the principal strains on the web
panel at loads approaching failure. As the applied load increased the out
of plane deflection of the stiffener allowed the two previously distinct
buckles in the adjacent web panels to merge into one, passing through the
stiffener and running between the load bearing stiffener at mid-span and
the end post. For this 'new' panel, i.e. effectively one half of the girder, the
inclination of the panel diagonal was approxinlately 45 ~. From tension
field theory the predicted angle of inclination of the tensile band at failure
is two-thirds the panel diagonal, i.e. approximately 30. The experimental
value for the orientation of strains at mid-depth of the web panel at failure
was measured at 33, i.e. very close to the predicted value from tension
field theory.
One of the primary objectives of the test programme was to establish
the actual loading imposed upon the stiffener during the loading of a
girder.
In the tests conducted the strains were measured at positions across the
laces of the stiffener at mid-depth as loading proceeded. From the values
obtained, the value of the strain at the web panel were linearly inter-
polated. The strains at the web panel were then converted into loads by
multiplying by the Young's modulus of the stiffener material and the stif-
fener's cross-sectional area. The predicted theoretical loads on the stiffener
are compared to those values obtained from typical experimental results in
Fig. 13. A good correlation is noted and this was observed in all tests
where the transverse web stiffener remained effective during the loading of
the girder.
 Collapse behaviour of lightweight aluminium girders 243

4SOd PU3

t.~

\ /
7 "x;I

%
e-,

~SOd as4ua]

e-.,

4SOd pu:l
e,.

- r"~

.,-.

~m t~

,../ \, / \ ,, ~ "~
0

"o.
.or

l
,z \ IS0d aJ4ua3
244 K. E. P. Brown, H. R. Evans

800
(Theo)
700

600 (Exp)
z
500

4OO-

300 -

200

100

I 110 l ,
S 15 20

Load on the sfiffeners (kN)

Fig. 13. Typical comparison betweenmeasuredand predicted stiffenerloads.

Thus, the proposed mathematical model for determining the load on a

stiffener allows stiffeners to be designed to guard against any premature
failure of the girder.
In the present series of tests three girders had transverse web stiffeners
of suitable proportions so as not to fail before the overall collapse of the
plate girder occurred. In these cases the theoretical failure load for the
girder, calculated by the method proposed in this paper as the basis for the
derivation of the load on a stiffener, slightly overestimated the girder
capacity by about 8% on average.
In Table 2 the results of tests conducted by the authors and by several
others who undertook research on aluminium alloy plate girders at the
University of Wales College of Cardiff are presented.
Table 2 gives a comparison of predicted and measured failure loads for
a number of transversely stiffened girders where failure was due to a web
panel rather than to a stiffener failure mode. The overall level of agree-
ment noted between the measured and predicted failure loads shows the
validity of the proposed theoretical model.

5 CONCLUSION

The tests described in the paper have shown that the measured shear stress
distribution at the periphery of the web panel adjacent to the compression
flange clearly agrees with the distribution proposed by the mathematical
model.
 Collapse behaviour of lightweight aluminiurn girders 245

TABLE 2
Comparison of Predicted and Measured Collapse Loads for Aluminium Plate Girders

Investigator Girder Failure loads Vexp/Vf

Measured Predicted
Vexp (kN) Vf (kN)
Rockey & Evans 3 AG4 52.3 52.5 1.00
AG1 56.3 56.3 1.00
AG5 69.3 61.7 1.12
AG3 33.9 31-1 1.09
AG2 41.4 35.9 1.15
Hamoodi 3 AGS1 47.0 35.0 1.34
AGS2-T1 62.0 47.7 1.30
AGS2-T2 63.5 47.7 1.33
AGS3-T2 59.0 47.8 1-23
AGS4-T1 60.5 47.7 1.27
AGS5-T 1 62.0 61-5 1.01
AGS5-T2 68.5 61-9 1.11
AGCS1-T1, T2 58.8 49.4 1.19
AGCS2-TI, T2 61-8 49.3 1.26
AGCS3-T1, T2 58.7 49.3 1.19
Burt 5 A7G 1 185.3 211.3 0.88
A7G2 234.2 222-0 1.05
A7G3 285.9 257.6 1.11
A7G4 454-9 323-1 1.41
A7G5 447.9 392.0 1.14
A7G6 192.3 187.2 1.03
A7G7 298.9 255.0 1.17
A7G8 360.7 276.7 1-30
A7G9 236.2 233.5 1.06
A7G 10 236.2 239-9 0.99
A7GI 1 354.2 258.9 1-37
Brown 2 A7G7-1 s 338.1 358.5 0.94
A7G9-2s 267.5 305.6 0.88
A7G9- I s 289.2 306.6 0.94

It has also been s h o w n that the p r e d i c t i o n o f the load on the stiffener

b a s e d o n identifying the three m o d e s o f response experienced by the web
panel d u r i n g the l o a d i n g o f the girder p r o v i d e s a g o o d c o r r e l a t i o n with
e x p e r i m e n t a l results.
Finally, it has been established that the m a t h e m a t i c a l m o d e l p r o v i d e s a
s a t i s f a c t o r y m e t h o d for d e t e r m i n i n g the c a p a c i t y o f transversely stiffened
a l u m i n i u m alloy plate girders l o a d e d p r e d o m i n a n t l y in shear.
246 K. E. P. Brown, H. R. Evans

ACKNOWLEDGEMENTS

The authors wish to acknowledge the contribution made by Dr O. Vilnay

during the early part of the study described in the paper, and to thank Mr
D. Webber of the Defence Research agency for his continued support and
interest.

REFERENCES

1. Vilnay, O., The behaviour of a web plate loaded in shear. Thin-Walled

Structures, 10 (1990) 161-74.
2. Brown, K. E. P., The post-buckling and collapse behaviour of aluminium
plate girders. PhD thesis, University of Wales College of Cardiff, UK, 1990.
3. Hamoodi, M. H., The behaviour of reinforced aluminium alloy web plates in
a shear loading. MSc thesis, University of Wales, College of Cardiff, UK,
Dec. 1983.
4. Timoshenko, S. & Goodier, J. N., Theo O, o f ElasticiO" (2nd edn). McGraw-
Hill, New York, 1951.
5. Burr, C. A., The ultimate strength of aluminium plate girders. PhD thesis,
University of Wales College of Cardiff, UK, 1987.