Experimental tests on pure aluminium shear panels with welded stiffeners.pdf

8/10/2019 Experimental tests on pure aluminium shear panels with welded stiffeners.pdf

http://slidepdf.com/reader/full/experimental-tests-on-pure-aluminium-shear-panels-with-welded-stiffenerspdf 1/11

Engineering Structures 30 (2008) 1734–1744

www.elsevier.com/locate/engstruct

Experimental tests on pure aluminium shear panels with welded stiffeners

G. De Matteisa,∗, F.M. Mazzolanib, S. Panico b

a University of Chieti/Pescara G. d’Annunzio, Faculty of Architecture (PRICOS), Viale Pindaro, 42, 65127 Pescara, Italyb University of Naples Federico II, Department of Structural Engineering (DIST), P.le Tecchio 80, 80125 Napoli, Italy

Received 15 January 2007; received in revised form 23 November 2007; accepted 26 November 2007

Available online 21 February 2008

Abstract

Pure aluminium is an innovative material recently proposed in the field of seismic engineering, due to both low yield strength and high ductility

features. Aimed at evaluating the energy dissipation capacity of stiffened pure aluminium shear panels, an experimental investigation has been

recently developed. In the current paper the results of two cyclic tests on aluminium panels with welded stiffeners subjected to shear loads are

provided. Tested specimens have in-plane dimensions 1500 × 1000 mm, with a thickness of 5 mm and are made of AW 1050A alloy, which is

characterised by a high degree of purity (more than 99.50%), also subjected to heat treatment to improve the material’s mechanical features. Such

a material has been chosen due to its very low yield strength (about 20 MPa after heat treatment), high hardening ratio (about 4) and large ductility

(about 40%). Two different configurations of welded stiffeners have been adopted, in order to investigate the effect of plate local slenderness. On

the whole, the obtained results pointed out a good structural performance of tested panels, measured in terms of strength, stiffness and dissipative

capacity, proving that the proposed system can be usefully employed as a special device for passive seismic protection of new and existing

structures.c 2007 Elsevier Ltd. All rights reserved.

Keywords: Pure aluminium; Shear panels; Shear buckling; Dissipative devices; Ductility; Cyclic behaviour

1. Introduction

A recent approach in the seismic design of structures is

based on the adoption of special seismic protection devices

acting as sacrificial elements during seismic events, thus

limiting the damage of both structural and non-structural

elements. Within metallic yielding-based devices, diagonal

bracings and panel systems can be conveniently used with high

efficiency.

In the first case, the dissipative function is carried out by

either ductile braces [1] or Added Damping And Stiffness(ADAS) elements [2]. Notoriously, in the former category

Buckling Inhibited Braces (BIB) and Unbonded Braces are

included. They are composed of a steel core, as load-carrying

and dissipative element, placed inside a lateral support jacket,

so as to obtain a buckling restrained bracing which is able

∗ Corresponding address: University of Naples Federico II, Department of Structural Engineering (DIST), P.le Tecchio 80, 80125 Napoli, Italy. Tel.: +39081 7682444; fax: +39 081 5934792.

E-mail address: [email protected] (G. De Matteis).

to dissipate energy under both tensile and compression axial

forces, providing a stable hysteretic behaviour without any

pinching and/or degradation of strength and stiffness up to the

system failure. In the latter category, the dissipative system

is installed either between the foundation and the structure

or between two relevant parts of the structure. The system is

differentiated in relation to the shape of the device (for instance

X-shaped, E-shaped, U-shaped, Ω -shaped, honeycomb-shaped

and so on), as well as for the type of dissipative mechanism on

which the system is based (for instance based on flexural, shear,

axial and torsional deformations).When shear panel systems are employed, stiffening,

strengthening and dissipative functions are carried out by either

the basic plate constituting the panel or by the connecting

system between shear panels and the bearing structure [3].

The former solution appears more effective and promising. In

fact, firstly, the adopted plates, when rigidly connected to the

external frame, may easily provide high in-plane strength and

stiffness; secondly, the possibility to have a quite uniform shear

stress distribution throughout the plate ensures a large energy

dissipation capacity due to the large size of yielding fields.

0141-0296/$ - see front matter c

2007 Elsevier Ltd. All rights reserved.

doi:10.1016/j.engstruct.2007.11.015

http://www.elsevier.com/locate/engstruct

mailto:[email protected]

http://dx.doi.org/10.1016/j.engstruct.2007.11.015

http://dx.doi.org/10.1016/j.engstruct.2007.11.015

mailto:[email protected]

http://www.elsevier.com/locate/engstruct



G. De Matteis et al. / Engineering Structures 30 (2008) 1734–1744 1735

Table 1

Mechanical features of typical LYS steel and heat treated pure aluminium EN-AW 1050A

Material f 0.2 (N mm−2) f u (N mm−2) εu (%) E (N mm−2) E / f 0.2 α = f u/ f 0.2

LYS steel 86 254 50 210 000 2441 2.95

Heat treated pure aluminium (EN-AW 1050A) 21.3 80 45 70 000 3286 3.76

In order to enhance the dissipative capacity, shear panels

should be conceived in order to increase the shear buckling

threshold and, in the meantime, to reduce the yielding interstory

drift, so to allow the seismic protection of the primary

structure also for reduced lateral loads. To this purpose,

adequate stiffener configurations must be applied, helping shear

panels to provide a pure shear dissipative mechanism with

plastic deformations developing before the occurrence of shear

buckling. Plate buckling may be more easily controlled when

low yield strength metals are adopted. As a consequence,

shear panels made of low yield strength materials may be

characterised by a very stable hysteretic behaviour up to large

deformation levels, with a conspicuous strain-hardening undercyclic loads and with limited strength and stiffness degradation

arising from buckling waves. A first solution is represented by

the so-called LYSW (Low Yield Shear Walls), which in the last

decade has been proposed and adopted mainly in Japan [4].

Also, the use of common aluminium alloys has been proposed

to develop new devices for passive protection of structures

subjected to seismic and wind loading [5–9].

Actually, an alternative and more effective solution for the

fabrication of dissipative shear panels could be based on the

use of pure aluminium owing to lower yielding stress and easier

availability in the world market. Therefore, a wide experimental

campaign to investigate the energy dissipation capacity of aluminium stiffened shear panels has been undertaken at the

University of Naples Federico II in cooperation with the

University of Chieti-Pescara G. d’Annunzio. Two aluminium

alloys have been employed, namely AW 1050A and AW

5154A. The former, also indicated as “pure aluminium”, is

characterised by a very high degree of purity (99.50% of

aluminium), while the latter is an alloy commonly used for the

civil engineering applications. The experimental activity is still

ongoing and six shear panels, four made of AW 1050A and

two made of AW 5154A, have been tested so far [ 10]. This

experimental activity was preceded by a preliminary numerical

study based on both static and dynamic inelastic analyses

carried out in order to evaluate the seismic performance of steel frames equipped with pure aluminium shear panels [11].

The obtained results clearly showed that these panels provide

a useful stiffening effect, a significant energy dissipation

contribution and also a remarkable increase of global strength,

on the whole increasing the performance of the structural

system. These effects allowed a significant global economical

advantage, measured in terms of saving of steel weight of the

basic structure when compared to the employment of other

technical solutions, i.e. the ones based on the use of simple

moment resisting steel frames. The economical analysis also

considered the major cost related to the employed aluminium

for shear panels, which was duly taken into account.

In this paper, the results of two cyclic tests carried out on the

AW 1050A aluminium panels are presented. In the following,

for each test the obtained cyclic response of the specimen

is described and the corresponding experimental behaviour is

interpreted by means of synthetic numerical parameters, which

are used to evaluate the supplied structural performance in

terms of dissipated energy, secant stiffness and plastic strength.

2. The experimental program

2.1. The adopted material

According to the purpose of this study, the selected material

is the aluminium alloy AW 1050A, which is characterised

by a limited conventional yield strength (about 20 MPa) and

large ductility (larger than 40%). It is important to observe

that the shear panel specimens, after the fabrication, have been

subjected to heat treatment, which aids a further improvement

of the mechanical features and is also useful to reduce the

residual stresses produced by welding during the fabrication

process. In particular, a number of specimens have been

submitted to a cycle of heat treatments characterised by

different phases with constant temperature, each one having a

duration of four hours [9].

In Table 1, the mechanical features of the above heat treated

aluminium alloy are shown and are also compared with the onesrelated to a typical low yield strength (LYS) steel. It is worth

noticing that due to the processed heat treatment, the aluminium

alloy AW 1050A is very suitable for the application under

consideration. In fact, it is characterised by a very high value

of the ratio between Young’s modulus ( E ) and the conventional

yield strength at 0.2% offset strain ( f 02), which means a high

aptitude to avoid buckling [12].

2.2. Geometry of tested panels

The experimental tests have been carried out on panel speci-

mens measuring 1500 by 1000 mm, with a thickness t =

5 mm

[10,11,13]. Tested specimens are endowed with longitudinal

and transversal open rectangular-shaped stiffeners having a

depth of 60 mm and obtained by the same sheeting used for

the base shear plate and welded to the base shear plate. In order

to reduce the sheeting shape distortion produced by shrinkage,

the welds has been subdivided into segments and the ribs have

been placed on both sides of the sheeting, therefore balancing

the residual strain produced during the welding process.

In the present paper, two different types of panel

configuration have been considered, which are characterised by

a different geometry of the applied ribs (see Fig. 1). For panel

type B, the ribs are placed on both sides of the plate according

to square fields with side length b = 500 mm. Contrarily, the



1736 G. De Matteis et al. / Engineering Structures 30 (2008) 1734–1744

Fig. 1. Geometrical configuration of tested shear panels.

panel configuration type F is stiffened with ribs alternatively

placed on the two sides of the plate in order to obtain squarefields with b = 250 mm and to balance the out-of-plane

deflection induced by the welding process. As a consequence,

the internal fields of the tested shear panels are characterised by

different slenderness ratios, namely b/t = 100 for panel type

B and b/t = 50 for panel type F.

The external shear load has been applied by means of an

articulated steel frame composed of very rigid members and

equipped with lateral out-of-plane braces [9] (see Fig. 2).

3. Test results

The main results of the experimental tests are provided

in terms of system reaction force versus the applied lateral

displacement. In the following, for each specimen such a

relationship has been normalized considering the average shear

stress acting on the horizontal panel side, assuming a nominal

cross-section A = 5000 mm2 (the panel width B is 1000 mm),

and the equivalent shear strain γ , which has been evaluated

as the ratio between the applied displacement and the panel

nominal depth H = 1500 mm. It is worth noting that the

considered shear strain includes only the part related to the

panel shear deformation, since the slips occurring in the panel

connections as well as the displacements of the reaction frame

have been deducted from the global measured displacement.

The cyclic response of tested specimens has also beeninterpreted by considering three synthetic numerical parameters

(see Fig. 3), which characterise the behaviour of the system

in terms of maximum hardening ratio (τ max/τ 02 with τ 02 = f 02/

√ 3), secant shear stiffness (Gsec) and equivalent viscous

damping factor (ζ eq). Therefore, the experimental data have

been processed for each panel by drawing the τ max/τ 02–∆γ ,

Gsec–∆γ and ζ eq–∆γ curves. Moreover, in order to evaluate

the degradation of strength, energy dissipation capacity and

secant shear stiffness due to low cycle fatigue effects,

τ max/τ 02, Gsec and ζ eq have also been represented in the form

of bar diagrams as a function of the cycle number.

The obtained hysteretic curves for tested specimens togetherwith the applied displacement history are shown in Fig. 4. In

the same figure, some pictures illustrating the main collapse

modes of tested specimens are also provided. In order to better

emphasise the progressive behaviour of the system during the

loading process, the cyclic response of the tested panels has

Fig. 2. Testing lay-out.




Fig. 3. Definition of dissipated energy ( E d ), equivalent viscous damping ratio

(ζ eq) and secant shear stiffness (Gsec).

been divided into four different significant phases. In Figs. 5

and 6, the relevant phenomena are depicted for each phase

and the corresponding panel hysteretic cycles are shown. The

above phases have been subdivided according to the followingcriteria:

– phase 1: Negligible buckling phenomena;– phase 2: Occurrence of local buckling;– phase 3: Occurrence of global buckling;– phase 4: Collapse of the specimen.

4. Analysis of the obtained results

The obtained shear force–shear deformation response

diagrams are rather complicate, since they incorporate a number

of different phenomena occurring during the tests. On the other

hand, it has to be observed that the hysteretic response of thetested panels appears to be quite dissipative, even though, in the

initial part of the loading process, the behaviour is conditioned

by some slipping phenomena due to initial imperfections and

local buckling of some plate portions. In particular, the buckling

phenomena developing in the lower portions of the panels are

due to normal stresses equilibrating the flexural moment at the

base of the plate, producing a ‘pinching’ effect on the shape

of the shear stress–shear strain curve. These phenomena are

evident at displacement amplitude equal to ± 8 mm (±0.0053

rad) for panel type F and ±5 mm (±0.0033 rad) for panel type

B. For both the panel types, the first behavioural phase develops

up to shear deformation of ±

0.01 rad. In such a range local

shear buckling waves are reduced and involve only the upper

and bottom corners of the panel (see Figs. 5 and 6—phase 1).

Local shear buckling of the single plate portions developed for

an applied deformation amplitude equal to ±30.0 mm (±0.02

rad) for panel type F and ±23.0 mm (±0.015 rad) for panel

type B (see Figs. 5 and 6—phase 2). As a consequence, out-

of-plane displacements throughout diagonals in compression

and the consequent activation of diagonal tension fields were

evident. For increasing out-of-plane displacements of the local

buckling waves, strength peaks became clearly visible in the

cyclic curves. This is due to a sort of stiffening effect produced

by corrugated shape of buckled panel portions. This effect is

also reflected in the fatter shape of hysteretic cycles, evidencing

an enhanced dissipative capacity of the specimen (see Figs. 5

and 6—phase 3).

It is important to observe that at this deformation stage, the

shear stress on the panel was decidedly larger than the one

corresponding to the conventional elastic limit (τ 02 = f 02/√

3),

emphasising that significant plastic deformations were ongoing.

Due to the stable post-buckling behaviour, a significant increaseof the global stiffness and dissipative capacity of the system

was evident, especially for panel type F. The phenomenon

increased as long as it involved more plate portions. As a

consequence, the dissipative capacity of the system gradually

increased as long as the deformation amplitude increased. A

better performance was noticed when the maximum strength

was attained, which practically coincides with the pure shear

strength of the system. Buckling phenomena of the ribs

developed when a displacement larger than ±47.0 mm (±0.031

rad) and ±55.0 mm (±0.037 rad) was imposed for panel type

B and for panel type F, respectively. This behavioural phase

was characterised by a strength degradation followed by a

significant increase of stiffness. Due to rib buckling, the shearstress–shear strain diagram showed a sort of ‘double bulge’

effect. Finally, a noticeable strength degradation occurred for

displacements larger than ±53.0 mm (±0.035 rad) for panel

type F and larger than ±80.0 mm (±0.053 rad) for panel

type B. Global buckling phenomena developed up to the

complete collapse of the system, which was due to failure of

surrounding connections, by tearing of the aluminium plate in

correspondence of the panel-to-frame bolted connection holes

for shear deformation equal±0.05 rad and±0.095 rad for panel

type F and type B, respectively (see Figs. 5 and 6—phase 4).

5. Comparison of results

By comparing the cyclic shear stress–shear strain curves of

the two examined panel configurations, it is clearly evident that

the number of cycles performed by the panel type B is higher

than panel type F, despite the larger b/t ratio of the former

(b/t = 50 versus b/t = 100). Such a result can be ascribed

to two main reasons: (a) in the panel type F a premature

tension failure of perimeter panel-column connections occurred

due to both excessive slipping in the cover plates and over-

dimensioned diameters of joining holes carried out in the

fabrication process; (b) in some parts of the panels, owing to the

fact that the ribs were alternatively placed on each side of the

plate and not continuously welded, the buckling waves passedfrom one side to the other side of a stiffener, causing a reduction

of the exerted lateral restraining action (see details of Fig. 6—

phase 2). For this reason, the effective b/t field ratio resulted

larger than the nominal one. On the contrary, for panel type B,

thanks to the bilateral restraining action exerted by the ribs, the

shear buckling was confined to the single plate portion up to the

complete failure of the panel, which occurred for tension failure

of connections preceded by the tearing of central points of each

plate portion enclosed by the ribs, forming a sort of X-shaped

failure lines in each field (see details of Fig. 5—phase 4).

In order to allow a clear comparison between the two

tested panel configurations, the experimental data have been




Fig. 4. Hysteretic curves, applied shear deformation histories and collapse mechanisms of tested AW 1050A specimens: (a) panel type B, (b) panel type F.




Fig. 5. Sequence of behavioural mechanisms for tested shear panel AW 1050A type B and corresponding hysteretic cycles.

processed by drawing for each panel theτ max/τ 02–∆γ, Gsec–∆γ

and ζ eq–∆γ diagrams (see Figs. 7a, 8a). Moreover, in order to

evaluate the degradation of strength, energy dissipation capac-

ity and secant shear stiffness due to low cycle fatigue effects,




Fig. 6. Sequence of behavioural mechanisms for tested shear panel AW 1050A type F and corresponding hysteretic cycles.

τ max/τ 02,, Gsec and ζ eq have been also drawn in the form of bar

diagrams as a function of the cycle number (see Figs. 7b, 8b).

In the same Figures, the characteristic zones related to the

phases illustrated in Figs. 5 and 6 are also shown. Finally in




Fig. 7. Cyclic performance of shear panel AW 1050A type B: (a) interpolation curves; (b) histograms.

Fig. 9, a direct comparison of the performance of tested pan-

els is provided in terms of interpolation curves of the above

selected synthetic parameters.

By observing the above diagrams, it is worth noting that theexamined F and B panels present an average equivalent viscous

damping equal to 26% and 23%, respectively. The maximum

value of factor ζ eq (even larger than 30%) is got in the initial

and in the final loading phases. On the contrary, in both the

cases, there is a strong reduction of such an equivalent viscous

damping ratio for intermediate deformation amplitude, namely

∆γ ranging from 0.013 rad to 0.054 rad for panel type F and

0.005 rad to 0.045 rad for panel type B. This evidences the

fact that the dissipative behaviour of pure aluminium shear

panels can be more profitable for very small deformation

amplitude (i.e. at serviceability limit state of the structure)

and for very large deformation amplitude (i.e. in the collapse

condition of the structure). On the contrary, the dissipative

capacity of tested shear panels appears to be slightly reduced

for an intermediate deformation range, the equivalent viscous

damping ratio assuming for both panel types a minimum value

of about 13% for ∆γ = 0.03 rad.

It is also interesting to observe the trend of Gsec–∆γ curves,

which shows that the global stiffness of tested specimens

rapidly decreases as long as the applied deformation amplitude

increases. In particular, even for very small values of ∆γ , the

obtained stiffness of the specimen is remarkably lower than the

nominal shear stiffness of the specimen (about 26 000 MPa).

This aspect emphasises that the actual susceptibility of the

specimen to buckle is significantly larger than the one that could

be predicted on the basis of the elastic material features E

and f 02.




Fig. 8. Cyclic performance of shear panel AW 1050A type F: (a) interpolation curves; (b) histograms.

Finally, the τ max/τ 02–∆γ curves allow the determination of

the maximum strength exploitable by the system for a given

deformation amplitude. In both the cases, the maximum value

of τ max/τ 02 is greater than 3, assuming a value comparableto the material hardening ratio, thus evidencing the actual

possibility of tested shear panels to exploit the full shear

strength. This outcome confirms the important issue that

the experienced buckling phenomena do not limit the shear

capacity of the system.

6. Conclusions

In this paper the results of two experimental tests on AW

1050A heat treated pure aluminium shear panels have been

provided. The examination of the obtained results emphasised

that the proposed system is able to exhibit a structural behaviour

characterised by large hysteretic cycles and large ductility with

a shear plastic collapse mechanism. Therefore, on the whole it

can be profitably employed as passive control device in framed

structures.An additional significant outcome of the study is that

the sheeting shape distortion produced by weld shrinkage

and the heat effects due to welding process may influence

the hysteretic response of the system, conditioning the

occurrence of premature buckling phenomena, even though the

experienced buckling phenomena do not limit the maximum

shear capacity of the system. On the other hand, the tests

presented in this paper have to be intended as a part of a

wider experimental program aimed at assessing the effect of

buckling phenomena on aluminium shear panels with different

stiffener configurations. In fact, the aforementioned drawbacks

may be overcome by avoiding welded stiffeners, but employing




Fig. 9. Comparison between tested panels (interpolation curves).

steel channel ribs bolted to the basic aluminium sheeting by

means of friction high-strength steel bolts as already shown

in [11].

Acknowledgments

This study has been developed in the framework of both

research projects “Innovative steel structures for the seismic

protection of new and existing buildings: design criteria

and methodologies” (MIUR) and “Development of innovative

approaches for the design of steel and composite structures”

(RELUIS).

References

[1] Clark W, Kasai K, Aiken ID, Kimura I. Evaluation of design

methodologies for structures incorporating steel unbonded braces for

energy dissipation. In: Proceedings of the 12th world conference on

earthquake engineering. Paper No. 2240, CD-ROM. 2000.

[2] Aiken ID, Whittaker AS. Development and application of passive energy

dissipation techniques in the USA. In: Proc. of the international Post-

Smirt conference seminar on isolation. Energy Dissipation and Control

of Vibration of Structures. 1993.

[3] Pinelli JP, Moor C, Craig JI, Goodno BJ. Testing of energy dissipating

cladding connections. Earthquake Engineering and Structural Dynamics

1996;25:129–47.

[4] Nakagawa S, Kihara H, Torii S, Nakata Y, Matsuoka Y, Fyjisawa K,et al. Hysteretic behaviour of low yield strength steel panel shear walls:

experimental investigation. In: Proc. of the 11th WCEE. Elsevier; 1996.

CD-ROM, Paper No. 171.

[5] Rai DC, Wallace BJ. Aluminium shear-link for enhanced seismic

performance. Journal of Earthquake Engineering and Structural Dynamics

1998;27:315–42.

[6] Rai DC, Wallace BJ. Aluminium shear-link for seismic energy dissipation.

In: Proc. 12th world conf. on earthquake engineering. Paper No. 0279 on

CD-ROM. 2000.

[7] Foti D, Diaferio M. Shear panels for seismic protection of buildings. In:

Structural dynamics-EURODYN ’99. 1999.

[8] Rai DC. Inelastic cyclic buckling of aluminium shear panels. Journal of

Engineering Mechanics 2002;128(11):1233–7.

[9] De Matteis G, Mazzolani FM, Panico S. Pure aluminium shear panels




as passive control system for seismic protection of steel moment resisting

frames. In: Proceedings of the 4th STESSA conference (behaviour of steel

structures in seismic areas). 2003. p. 599–607.

[10] De Matteis G, Panico S, Mazzolani FM. Experimental tests on stiffened

aluminium shear panels. In: Proc. XI international conference on metal

structures. 2006.

[11] De Matteis G, Mazzolani FM, Panico S. Pure aluminium shear

panels as dissipative devices in moment-resisting steel frames.

Journal of Earthquake Engineering and Structural Dynamics 2007;

36:841–59. Published online 2 January 2007 in Wiley InterScience

www.interscience.wiley.com . doi:10.1002/eqe.656.

[12] De Matteis G, Formisano A, Panico S, Mazzolani FM. Numerical

and experimental analysis of pure aluminium shear panels. Comput

Struct 2007, in press, available on line at http://www.sciencedirect.com,

doi:10.1016/j.compstruc.2007.05.027 .

[13] De Matteis G, Panico S, Formisano A, Mazzolani FM. Experimental tests

on pure aluminium shear panels under cyclic loading. In: Proc. IZIIS

40EE 21C – earthquake engineering in 21st Century. 2005.

http://www.interscience.wiley.com/


http://dx.doi.org/http://dx.doi.org/doi:10.1002/eqe.656


http://www.sciencedirect.com/


http://dx.doi.org/10.1016/j.compstruc.2007.05.027






Experimental tests on pure aluminium shear panels with welded stiffeners.pdf

Documents

Transcript of Experimental tests on pure aluminium shear panels with welded stiffeners.pdf