Sei sulla pagina 1di 10

XXVII CONGRESSO C.T.A.

VALUTAZIONE DEL MOMENTO ULTIMO DI TRAVI IN LEGA


D’ALLUMINIO A SEZIONE RETTANGOLARE CAVA
MEDIANTE IL METODO DELLO SPESSORE EFFICACE

ULTIMATE MOMENT EVALUATION OF RHS ALUMINIUM


BEAMS BY EFFECTIVE THICKNESS METHOD

Elide Nastri, Vincenzo Piluso

Department of Civil Engineering, University of Salerno, Fisciano (SA), Italy


enastri@unisa.it, v.piluso@unisa.it

ABSTRACT
In this work, a procedure for evaluating the ultimate moment of aluminium alloy RHS beams un-
der non-uniform bending is developed within the framework of the effective thickness approach
to account for the influence of local buckling phenomena. The theoretical approach is presented
and its accuracy is investigated by comparing the results obtained with those of numerical simula-
tions based on an advanced finite element model. To this scope, a sample of 1008 beams whose
ultimate non-dimensional moment has been computed through FEM simulations has been consi-
dered. This sample covers 4 different aluminium alloys, i.e. 6082 T4 and T6, 6061 T6 and 6063
T5. In addition, many approaches for the evaluation of the so-called ξ factor, accounting for the
occurrence of local buckling in the plastic range, have been investigated.

SOMMARIO
In questo lavoro, viene sviluppata una procedura per valutare il momento ultimo di travi RHS in
lega di alluminio soggette a flessione non uniforme, nell'ambito del metodo dello spessore effica-
ce per tenere conto dell'influenza dei fenomeni di instabilità locale. Viene presentato l'approccio
teorico e la sua accuratezza viene esaminata confrontando i risultati ottenuti con quelli di simula-
zioni numeriche basate su un modello FEM avanzato. A tale scopo, è stato preso in considerazio-
ne un campione di 1008 travi il cui momento ultimo è stato calcolato mediante simulazioni FEM.
Questo campione copre 4 diverse leghe di alluminio, ovvero 6082 T4 e T6, 6061 T6 e 6063 T5.
Inoltre, sono stati studiati molti approcci per la valutazione della instabilità locale in campo pla-
stico.

1 INTRODUCTION
One of the main advantages of aluminium profiles is the chance of forming any shape due to the
hot-extrusion moulding process. Due to the higher material expenses, aluminium profiles are of-
BOLOGNA 3-5 OTTOBRE 2019

ten designed to be economic by adopting thinner and weaker plate elements. Moreover, alumi-
nium has a small elastic modulus which is about one-third of that of steel. These aspects make
local buckling the dominant failure mode of aluminium profiles. In this paper, the effective thick-
ness method is adopted to compute the non-dimensional ultimate moment of aluminium alloy
RHS beams under non-uniform bending. Through the effective thickness method, the local buck-
ling of the plate elements is accounted for by using an effective section instead of the true section.
The effective section progressively reduces as far as the strain level increases. A sample of 1008
beams whose ultimate non-dimensional moment has been computed through FEM simulations
has been considered to investigate the accuracy of the theoretical approach. This sample covers 4
different aluminium alloys, i.e. 6082 T4 and T6, 6061 T6, 6063 T5 [1].

2 EFFECTIVE THICKNESS APPROACH


The reference elastic buckling stress of an isolated plate element is given by:
𝜋𝜋 2 𝐸𝐸
𝜎𝜎𝑐𝑐𝑐𝑐 .0 = 𝑘𝑘𝑜𝑜𝑜𝑜 (1)
12(1 − 𝜈𝜈𝑒𝑒2 )(𝑏𝑏⁄𝑡𝑡 )2
where 𝐸𝐸 is the elastic modulus, 𝜈𝜈𝑒𝑒 is the Poisson’s ratio in the elastic range, 𝑏𝑏 is the plate width, 𝑡𝑡
is the plate thickness and 𝑘𝑘𝑜𝑜𝑜𝑜 is the buckling factor accounting for the edge restraining conditions
and the stress distribution along the loaded edges.
In the case of the plate elements constituting the section of a structural member, the occurrence of
elastic buckling is also affected by the interaction between the plate elements constituting the
member section and by the longitudinal stress gradient. These effects can be taken into account by
modifying Eq. (1) using two factors: 𝜁𝜁 and 𝛼𝛼𝐿𝐿𝐿𝐿 . The factor 𝜁𝜁 accounts for the interaction between
the plate elements composing the section. The factor 𝛼𝛼𝐿𝐿𝐿𝐿 accounts for the influence of the longi-
tudinal stress gradient occurring in structural members under non-uniform bending.
Therefore, including the effects of interactive buckling and longitudinal stress gradient, the elastic
buckling stress can be expressed as:
𝜋𝜋 2 𝐸𝐸
𝜎𝜎𝑐𝑐𝑐𝑐 = 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝑜𝑜 (2)
12(1 − 𝜈𝜈𝑒𝑒2 )(𝑏𝑏⁄𝑡𝑡 )2
Buckling in the plastic range can be accounted for using a correction factor which depends on the
non-linear behaviour of the material. By denoting with 𝜉𝜉 such correction factor, the buckling
stress in the plastic range 𝜎𝜎𝑐𝑐𝑐𝑐 .𝑝𝑝 is given by:
𝜋𝜋 2 𝐸𝐸
𝜎𝜎𝑐𝑐𝑐𝑐 .𝑝𝑝 = 𝜉𝜉 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝜓𝜓 (3)
12(1 − 𝜈𝜈 2 )(𝑏𝑏⁄𝑡𝑡 )2
where also the Poisson’s ratio 𝜈𝜈 must be evaluated accounting for the non-linear behaviour of the
material.
Regarding the 𝜉𝜉 factor, many different expressions have been proposed in the technical literature
for its evaluation. They will be presented in the following discussion.
Concerning the Poisson’s ratio in the yield region, Gerard and Wildhorn (1952) [2] have studied
the problem with reference to several aluminium alloys and have shown that it is seriously af-
fected by the anisotropy of the material. In case of materials exhibiting the same properties in the
two directions orthogonal to the loading direction, they proposed the following relation:
𝐸𝐸𝑠𝑠
𝜈𝜈 = 𝜈𝜈𝑝𝑝 − �𝜈𝜈 − 𝜈𝜈𝑒𝑒 � (4)
𝐸𝐸 𝑝𝑝
where νe = 0.30 is the Poisson’s ratio in the elastic range, νp = 0.50 is the Poisson’s ratio for
perfectly plastic material and Es is the secant modulus. It can be easily recognised that Eq. (4)
provides 𝜈𝜈 = 𝜈𝜈𝑒𝑒 for 𝐸𝐸𝑠𝑠 = 𝐸𝐸. However, in the case of perfectly plastic material, the condition
ν = νp is reached only when the strain assumes infinite value (ε → ∞) so that 𝐸𝐸𝑠𝑠 → 0. Therefore,
XXVII CONGRESSO C.T.A.

an alternative relation can be proposed:


𝐸𝐸𝑡𝑡
𝜈𝜈 = 𝜈𝜈𝑝𝑝 − �𝜈𝜈 − 𝜈𝜈𝑒𝑒 � (5)
𝐸𝐸 𝑝𝑝
which assures ν = νe in the elastic range (𝐸𝐸𝑡𝑡 = 𝐸𝐸) and the condition ν = νp when a plastic pla-
teau is reached (𝐸𝐸𝑡𝑡 = 0).
Eq. (3) can be written as follows:
𝜋𝜋 2 𝐸𝐸
𝐸𝐸𝑠𝑠 𝜀𝜀 = 𝜉𝜉 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝑜𝑜 (6)
12(1 − 𝜈𝜈 2 )(𝑏𝑏⁄𝑡𝑡 )2
where 𝜀𝜀 is the strain corresponding to buckling. As a consequence, the effective ratio 𝑏𝑏⁄𝑡𝑡 can be
defined as:
𝑏𝑏 𝜉𝜉𝐸𝐸 𝜋𝜋 2 1
� � =� 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝑜𝑜 (7)
𝑡𝑡 𝑒𝑒𝑒𝑒𝑒𝑒 𝐸𝐸𝑠𝑠 12(1 − 𝜈𝜈 2 ) 𝜀𝜀
The effective thickness can be derived by noting that:
𝑡𝑡 𝐸𝐸𝑠𝑠 1 12(1 − 𝜈𝜈 2 )
� � =� 𝜀𝜀 (8)
𝑏𝑏 𝑒𝑒𝑒𝑒𝑒𝑒 𝜉𝜉𝐸𝐸 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝑜𝑜 𝜋𝜋 2
and by introducing a parameter 𝜇𝜇, which accounts for the non-linear behaviour of the material, as
follows:
𝜉𝜉𝐸𝐸 1
𝜇𝜇 = (9)
𝐸𝐸𝑠𝑠 1 − 𝜈𝜈 2
the following relation is obtained:
𝑡𝑡𝑒𝑒𝑒𝑒𝑒𝑒 𝑏𝑏 12 1 1
= � 2 𝜀𝜀 (10)
𝑡𝑡 𝑡𝑡 𝜋𝜋 𝜇𝜇 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝑜𝑜
Remembering that:
𝑡𝑡𝑒𝑒𝑒𝑒𝑒𝑒
𝜆𝜆̅𝑝𝑝 = (11)
𝑡𝑡
it means that the normalised slenderness in the non-linear range (elastic-plastic range) can be defined
as:
𝑏𝑏 12 1 1
𝜆𝜆̅𝑝𝑝 = � 2 𝜀𝜀 (12)
𝑡𝑡 𝜋𝜋 𝜇𝜇 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 𝑘𝑘𝑜𝑜𝑜𝑜
In order to use the buckling curves of EN1999-1-1 with the normalised slenderness corrected to
account for the non-linearity depending on the strain level, it has to be remembered that:
𝛽𝛽
𝜆𝜆̅𝑝𝑝 = 0.03143 (13)
𝜖𝜖0
where 𝜖𝜖0 = �250⁄𝑓𝑓0 , therefore:
𝛽𝛽 𝜆𝜆̅𝑝𝑝 1 3 𝜀𝜀 2 𝑏𝑏
= = � (14)
𝜖𝜖0 0.03143 0.03143 𝜋𝜋 2 𝜇𝜇 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿 � 𝑘𝑘𝑜𝑜𝑜𝑜 𝑡𝑡
Taking into account that, in EN1999-1-1, the factor accounting for the stress distribution along
the loaded edge is:
2
𝜂𝜂 = (15)
𝑘𝑘
� 𝑜𝑜𝑜𝑜
Eq. (14) provides:
BOLOGNA 3-5 OTTOBRE 2019

𝛽𝛽 𝑏𝑏 𝜀𝜀
= 17.54 𝜂𝜂 � (16)
𝜖𝜖0 𝑡𝑡 𝜇𝜇 𝜁𝜁 𝛼𝛼𝐿𝐿𝐿𝐿
which is the final expression to compute the slenderness parameter of the plate element to be
used, in combination with the buckling curves of EN1999-1-1 [1], to compute the effective thick-
ness in the non-linear range as a function of the strain level 𝜀𝜀. In fact, according to EN1999-1-1
[1], the reduction factor accounting for local buckling is computed as:
𝛽𝛽 1
𝜌𝜌𝑐𝑐 = 1 𝑖𝑖𝑖𝑖 ≤ �𝐶𝐶 + �𝐶𝐶12 − 𝐶𝐶2 (3 + 𝜓𝜓)� (17)
𝜖𝜖0 2 1
𝐶𝐶1 𝐶𝐶2 (3 + 𝜓𝜓) 𝛽𝛽 1
𝜌𝜌𝑐𝑐 = − 𝑖𝑖𝑖𝑖 > �𝐶𝐶 + �𝐶𝐶12 − 𝐶𝐶2 (3 + 𝜓𝜓)� (18)
𝛽𝛽 ⁄𝜖𝜖0 4 (𝛽𝛽 ⁄𝜖𝜖0 )2 𝜖𝜖0 2 1
The parameter 𝜓𝜓 accounts for the strain distribution along the loaded edge of the plate. It is given
by the ratio between the maximum compression strain at one end of the plate and the strain at the
second end of the plate element. In the case of uniform compression, it results 𝜓𝜓 = 1 while 𝜓𝜓 < 0
when the second end of the plate element is subject to tension.
Regarding the 𝜉𝜉 factor needed in Eq. (9), as already stated, a variety of different formulations
have been proposed in the technical literature:
• tangent modulus theory: 𝜉𝜉 = 𝐸𝐸𝑡𝑡 ⁄𝐸𝐸 (19)
• secant modulus theory: 𝜉𝜉 = 𝐸𝐸𝑠𝑠 ⁄𝐸𝐸 (20)
• Pearson (1950) [3], Bleich (1952) [4] and Vol’Mir
𝜉𝜉 = �𝐸𝐸𝑡𝑡 ⁄𝐸𝐸 (21)
(1965) [5]:
• Radhakrishnan (1956) [6]: 𝜉𝜉 = �𝐸𝐸𝑡𝑡 ⁄𝐸𝐸𝑠𝑠 (22)
• Gerard (1957) [7]: 𝜉𝜉 = �𝐸𝐸𝑡𝑡 𝐸𝐸𝑠𝑠 ⁄𝐸𝐸 2 (23)
• Weingarten et al. (1960) [8]: 𝜉𝜉 = 𝐸𝐸𝑠𝑠 ⁄𝐸𝐸 �𝐸𝐸𝑡𝑡 ⁄𝐸𝐸 (24)
𝐸𝐸𝑠𝑠 1 2 1 3 𝐸𝐸𝑡𝑡
• Stowell (1948) [9], Bijlaard (1949)[10]: 𝜉𝜉 = � + � + � (25)
𝐸𝐸 3 3 4 4 𝐸𝐸𝑠𝑠

𝐸𝐸𝑠𝑠 1 1 1 3 𝐸𝐸𝑡𝑡
• Li and Reid (1992) [11]: 𝜉𝜉 = � + � + � (26)
𝐸𝐸 2 2 4 4 𝐸𝐸𝑠𝑠
Regarding the factor accounting for the influence of the longitudinal stress gradient, the following
relations can be adopted:
• for a flat internal compression cross-section part
1 1.70
𝛼𝛼𝐿𝐿𝐿𝐿 = 1 +
4 𝐿𝐿𝑠𝑠 0.75 (27)
� � − 0.20
𝑏𝑏
• in case of flat outstand compression elements:
1 1.70
𝛼𝛼𝐿𝐿𝐿𝐿 = 1 +
0.425 𝐿𝐿𝑠𝑠 0.95 (28)
� � − 0.60
𝑏𝑏
where 𝐿𝐿𝑠𝑠 is the shear length and b is the plate width in compression. The shear length is defined
as the distance between the point of zero moment and the section where the maximum bending
moment occurs.
The correction factor 𝜁𝜁 for interactive buckling can be evaluated taking into account that it repre-
sents the ratio between the buckling factor 𝑘𝑘 accounting for interactive buckling and the buckling
factor 𝑘𝑘0 evaluated for the isolated plate element, i.e. 𝜁𝜁 = 𝑘𝑘 ⁄𝑘𝑘0 .
XXVII CONGRESSO C.T.A.

• in case of plate elements, acting as flange, connected to webs on both edges (curve 1 of Fig.
1)
0.45 𝑏𝑏2 ⁄𝑏𝑏1
𝜁𝜁 = 1.75 − − 0.02275(𝑏𝑏2 ⁄𝑏𝑏1 )3 ≥ 1 (29)
0.15 + 𝑏𝑏2 ⁄𝑏𝑏1
which is derived from the expression of 𝑘𝑘 given by BS5950-5 considering that, in this case,
𝑘𝑘0 = 4.
• in case of plate elements, acting as flange, connected to the web on one edge and lip on an-
other edge (curve 2 of Fig. 1)
0.35 𝑏𝑏2 ⁄𝑏𝑏1
𝜁𝜁 = 1.35 − − 0.005(𝑏𝑏2 ⁄𝑏𝑏1 )3 ≥ 1 (30)
0.60 + 𝑏𝑏2 ⁄𝑏𝑏1
which is derived from the expression of 𝑘𝑘 given by BS5950-5 considering that, in this case,
𝑘𝑘0 = 4.
• in case of unstiffened elements, acting as a flange (Fig. 2)
1.882 𝑏𝑏2 ⁄𝑏𝑏1
𝜁𝜁 = 3.00 − − 0.0059(𝑏𝑏2 ⁄𝑏𝑏1 )2 ≥ 1 (31)
2.0 + 𝑏𝑏2 ⁄𝑏𝑏1
which is derived from the expression of 𝑘𝑘 given by BS5950-5 considering that, in this case,
𝑘𝑘0 = 0.425.

Fig. 1. Correction factor for interactive buckling for stiffened compression elements of
beams

Fig. 2. Correction factor for interactive buckling for unstiffened compression elements
of beams
• in case of channel flanges acting as bending elements (Fig. 3)
BOLOGNA 3-5 OTTOBRE 2019

2(3 + 𝜓𝜓)
𝜁𝜁 = ≥1
𝑏𝑏2 ⁄𝑏𝑏1 (32)
2.0 +
1 + 𝑏𝑏2 ⁄𝑏𝑏1

Fig. 3. Correction factor for interactive buckling for channel flanges as bending elements

3 RESULTS OBTAINED AND THEIR ACCURACY


The main results of the application of the effective thickness method to the prediction of the ulti-
mate flexural resistance of RHS aluminium beams subjected to non-uniform bending are herein
reported. A computer program has been developed for the evaluation of the moment versus curva-
ture relation, accounting for the occurrence of local buckling either in the elastic or in the plastic
range by means of the effective thickness approach using the formulations derived in the previous
Sections. In the case of RHS beams subjected to bending, all the plate elements constituting the
member section behave like flat internal elements. Two elements constitute the beam flanges; one
of them is subjected to tension and another is subjected to uniform compression. The remaining
plate elements constitute the webs and subjected to a linear strain distribution. Therefore, both the
flange in compression and the compressed part of the web have to be modelled with an effective
thickness whose value is dependent on the strain level. As a consequence, the effective cross-
section reduces because of the reduction of the effective thickness of the compressed parts as far
as the curvature increases. The moment versus curvature relation is characterised by an increasing
non-linear branch up to the complete development of local buckling which is followed, because of
local buckling, by a softening branch. The ultimate moment is the maximum value 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 attained
by the moment-curvature relation which can be non-dimensionalised by considering the ratio
𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 ⁄𝑀𝑀0 , where 𝑀𝑀0 is the bending moment corresponding to the conventional elastic limit stress
𝑓𝑓0 .
The accuracy of the proposed effective thickness approach has been investigated by comparing
the non-dimensional ultimate moment obtained with such approach with those provided by the FE
numerical simulations reported in previous papers [9]-[10], based on an advanced FE model
which was preliminarily validated by comparison with the available experimental results.
In addition, the different formulations available in the technical literature for computing the 𝜉𝜉 fac-
tor, accounting for the occurrence of buckling in the plastic range, have been investigated. Obvi-
ously, the results confirmed that the tangent modulus theory, i.e. Eq. (19), provides the lowest
values of the flexural resistance; conversely, the secant modulus theory, i.e. Eq. (20), provides the
highest values.
Even though the secant modulus theory provides the highest values of the ratio 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 ⁄𝑀𝑀0 , the
obtained results are on the safe side providing values which are, on average, from 6% to 10%,
depending on the alloy, less than the values of 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 ⁄𝑀𝑀0 obtained from the FE model. An excep-
XXVII CONGRESSO C.T.A.

tion is made for 6061 T4 and 6082 T4 alloys, i.e. low yielding high hardening alloys. In fact, for
such alloys, the obtained results are on the unsafe providing an average overestimation of the ul-
timate flexural resistance equal to about 6% (Fig. 4).

Fig. 4. Secant modulus theory Fig. 5. Gerard formulation


The use of the Gerard formulation provides safe side results also in the case of low-yielding high-
hardening alloys having a Ramberg-Osgood exponent of the stress-strain curve n=8, but for all
the other alloys the underestimation of the ultimate flexural resistance seems excessive. These
results are depicted in Fig. 5 where it can be observed that, in the case of 6063 T5, 6082 T6 and
6061 T6 alloys, the underestimation of the ultimate flexural resistance ranges, on average, from
16% to 20%.
Better results are obtained by means of the Li and Reid formulation, i.e. Eq. (26). Such results are
depicted in Fig. 6 and in Fig. 7. Regarding the formulation for computing the Poisson’s ratio in
the plastic range, Fig. 6 is referred to Eq. (5) while Fig. 7 is referred to Eq. (4). However, the
comparison between the obtained results shows that the influence of the formulation for
computing the Poisson’s ratio in the plastic range is negligible.
Similar results are obtained by using the Stowell and Bijlaard formulation, i.e. Eq. (25). In par-
ticular, such results are depicted in Fig. 8 with reference to Eq. (5) for the Poisson’s ratio in the
plastic range and in Fig. 9 with reference to Eq. (4).

Fig. 6. Li and Reid formulation Fig. 7. Li and Reid formulation

The results obtained have pointed out that the 𝜉𝜉 factor, to account for the occurrence of buckling
in the plastic range, plays a role of paramount importance on the accuracy of the predicted values
of the ultimate flexural resistance. In addition, the best formulation to be adopted for the 𝜉𝜉 factor
is dependent on the hardening behaviour of the alloy, i.e. on the exponent n of the Ramberg-
BOLOGNA 3-5 OTTOBRE 2019

Osgood stress-strain law. Therefore, the accuracy of the effective thickness approach for predict-
ing the ultimate flexural resistance of aluminium beams subjected to local buckling under bending
actions can be improved looking for new formulations of the 𝜉𝜉 factor where the influence of n is
explicitly considered.

Fig. 8. Stowell and Bijlaard formulation Fig. 9. Stowell and Bijlaard formulation

Two new possibilities have been investigated aiming to the codification of the effective thickness
approach, in Annex L of forthcoming revision of EN1999-1-1, for the prediction of the ultimate
resistance of members explicitly considering the influence of local buckling either in the elastic or
in the plastic range.
These alternative formulations have been conceived with the aim of obtaining a formula provid-
ing results close to those of the Gerard formulation for small values of the Ramberg-Osgood ex-
ponent n of the stress-strain law and close to the secant modulus theory of high values of n.
The first proposal is:
n − 8 𝐸𝐸𝑠𝑠 8 𝐸𝐸𝑡𝑡 𝐸𝐸𝑠𝑠
𝜉𝜉 = + � (33)
n 𝐸𝐸 n 𝐸𝐸 2
Such proposal assures that in the elastic range, being 𝐸𝐸𝑡𝑡 = 𝐸𝐸𝑠𝑠 = E, 𝜉𝜉 = 1. The results obtained
with such proposal are depicted in Fig. 10 with reference to Eq. (5) for the Poisson’s ratio in the
plastic range and in Fig. 11 with reference to Eq. (4).

Fig. 10. The first proposal for EN1999-1-1 Fig. 11. The first proposal for EN1999-1-1
Annex L Annex L
XXVII CONGRESSO C.T.A.

In particular, the use of Eq. (33) and of Eq.(5) provides the minimum value of the standard devia-
tion of the ratio between the values of 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 obtained by the effective thickness approach and
those obtained by the ABAQUS finite element model.
The second proposal is:
n − 5 𝐸𝐸𝑠𝑠 5 𝐸𝐸𝑡𝑡 𝐸𝐸𝑠𝑠
𝜉𝜉 = + � (34)
n 𝐸𝐸 n 𝐸𝐸 2
Obviously, also such proposal assures that in the elastic range, being 𝐸𝐸𝑡𝑡 = 𝐸𝐸𝑠𝑠 = E, 𝜉𝜉 = 1.
The results obtained with such proposal are depicted in Fig. 12 with reference to Eq. (5) for the
Poisson’s ratio in the plastic range and in Fig. 13 with reference to Eq. (4).

Fig. 12. The second proposal for EN1999-1-1 Fig. 13. The second proposal for EN1999-1-1
Annex L Annex L

In particular, the use of Eq. (34) and of Eq.(4) provides a mean value of the ratio between the val-
ues of 𝑀𝑀𝑚𝑚𝑚𝑚𝑚𝑚 obtained by the effective thickness approach and those obtained by the ABAQUS
finite element model equal to 1.014, which is the closest to 1.0 among all the adopted formula-
tions.

4. CONCLUSIONS
The issues concerning the codification of the effective thickness approach to explicitly account
for the influence of local buckling phenomena either in the elastic or in the plastic range have
been presented. The resulting method for predicting the ultimate flexural resistance of aluminium
alloy members has been outlined within the framework of EN1999-1-1 whose buckling curves for
flat internal elements and for flat outstand elements are exploited.
Software for evaluating the moment-curvature relation accounting explicitly the influence of local
buckling has been developed. As a result the ultimate flexural resistance of aluminium RHS
beams subjected to local buckling under non-uniform compression has been evaluated and the
results obtained have been compared with those resulting from an advanced finite element model
in ABAQUS environment. Such comparison was aimed at the investigation of the accuracy of the
effective thickness approach for predicting the ultimate flexural resistance.
The results have pointed out the paramount importance of the formulation adopted to account for
local buckling in the plastic range. The formulations already available in the technical literature
are not always sufficiently accurate because of the influence of the hardening behaviour of the
material expressed by the exponent n of the Ramberg-Osgood stress-strain law. Therefore, the
accuracy of alternative formulations, explicitly accounting for the exponent of the Ramberg-
Osgood law, has been investigated.
BOLOGNA 3-5 OTTOBRE 2019

The results obtained have shown that the use of such formulations allows the improvement of the
accuracy of the effective thickness approach, providing satisfactory results for all the considered
alloys, both low yielding high hardening alloys and high yielding low hardening alloys.

REFERENCES
[1] Eurocode 9: Design of aluminium structures, Part 1-1, 2007.
[2] Gerard G., Wildhorn S. (1952): “A Study of Poisson’s Ratio in the Yield Region”, National
Advisory Committee for Aeronautics, NACA Technical Note 2561, Washington, January,
1952.
[3] Pearson C. E. (1950). Bifurcation criterion and plastic buckling of plates and columns. J.
Aerospace Sci. 7,417-424.
[4] Bleich F. (1952) “Buckling Strength of Metal Structures”, McGraw-Hill.
[5] Vol’mir A. S., Kil’dibekov I. G., Investigation of the buckling of bars under impact, Dokl.
Akad. Nauk SSSR, 1966, Volume 167, Number 4, 775–77.
[6] Radhakrishnan S., (1956) “Plastic Buckling of Circular Cylinders”, Journal of the Areonau-
tical Sciences.
[7] Gerard G., Becker H. (1957): “Handbook of Structural Stability: Part I – Buckling of Flat
Plates”, National Advisory Committee for Aeronautics, NACA Technical Note 3781,
Washington, July, 1957.
[8] Weingarten V.I. , E.J. Morgan, Siede P. “Elastic stability of thin-walled cylindrical and con-
ical shells under combined internal pressure and axial compression”, May 1965 AIAA
Journal
[9] Stowell E.Z. (1951): “Compressive Strength of Flanges”, National Advisory Committee for
Aeronautics, NACA Report 1029, Washington, 1951.
[10] Bijlaard P.P. (1949): “Theory and Tests on the Plastic Stability of Plates and Shells”. Jour-
nal of the Aeronautical Science, Vol. 16, n. 9, pp. 529-541.
[11] Li S., Reid S.R., (1992) “The Plastic Buckling of Axially Compressed Square Tubes”, J.
Appl. Mech. Jun 1992, 59(2): 276-282 (7 pages)
[12] Castaldo, P., Nastri, E., Piluso, V. "Ultimate behaviour of RHS temper T6 aluminium alloy
beams subjected to non-uniform bending: Parametric analysis", (2017) Thin-Walled Struc-
tures, 115, pp. 129-141.
[13] Castaldo, P., Nastri, E., Piluso, V. "FEM simulations and rotation capacity evaluation for
RHS temper T4 aluminium alloy beams" (2017) Composites Part B: Engineering, 115, pp.
124-137.

KEYWORDS
Effective thickness approach, RHS, aluminium alloys, non-dimensional ultimate moment.

Potrebbero piacerti anche