Abstract The “Ultimate Strength” committee of ISSC’2003 carried out a benchmark on the ultimate strength of aluminium stiffened panels。 The purpose of this benchmark was, for a given aluminium stiffened panel, to quantify the variations of the panel’s  ultimate strength when panel’s parameters change。 The studied parameters are weld types, HAZ width, yield stress in the HAZ, initial panel deflection and residual stresses。81941

In this paper the authors focus their discussion on the main outcomes provided by the ISSC benchmark and present additional analysis performed to confirm the previous results。 Finally the authors propose recommendations for future research works。

Author keywords Ultimate strength; Sensitivity analysis; Aluminium stiffened panels; Axial compression; Heat-affected zone

Introduction In San Diego (August 2003) the “Ultimate Strength” committee of ISSC’2003 presented a brief report about their benchmark on the ultimate strength of aluminium stiffened panels [Simonsen et al。, 2003]。 For additional information on the benchmark itself, a comprehensive paper in Marine Structures [Rigo et al。, 2003] is now available。  The  present  paper  provides  additional  and

updated results that are not available in the previous documents

The authors of this paper are the researchers that performed the ISSC benchmark。 They thank all the ISSC’2003 III。1 committee and particularly S。   Estefen,

E。Lehman and B。C。 Simonsen (committee chairman) for their support and advices。

The purpose of the ISSC benchmark was, for a given aluminium stiffened panel, to quantify the variations of the panel’s ultimate strength when panel’s parameters change。 These variations are called the sensitivities of the panel。 The studied parameters are:

Weld types (longitudinal, transversal,  extruded and no extruded components),

The previous results presented in [Rigo et al。, 2003 and Simonsen et al。, 2003] based on a three-spans model are compared to new analysis performed with a ½ +1+ ½ model and with other initial deflection pattern (as recommended by the official discusser of ISSC-III。1 Committee in San Diego, 2003)

The data used in this analysis are taken from Aalberg experiments [Aalberg et al。, 2001]。

Reference panel description

Geometry

A panel with L-shaped stiffeners fabricated from extruded aluminium profiles in alloy AA6082 temper T6, joined by welding, was defined for the  finite element analyses。

In the ISSC benchmark a three-spans model was considered while additional analyses were performed with a standard ½ + 1 + ½ model。 The dimensions of the models are presented in Fig。1 with the XYZ coordinate frame and the U, V, W corresponding displacements。

Boundary conditions

In the present FE computations, the boundary conditions for the stiffened panels were assumed simply supported along the two longitudinal edges (unloaded), which are kept straight (constrained edges)。 The loaded  edges were restrained from rotation and an axial displacement was prescribed on one side (clamping conditions for the ISSC model and symmetric conditions for the Standard model – the Standard model is then less stiff)。 At these two loaded edges, the stiffener cross-section remains plane and the sideways deformation of the stiffeners are not allowed。 In order to simulate stiff transverse frames, the displacements (W) along Z at the location of these transverse plates are not allowed。

Initial imperfections

Fig。1:

(a)ISSC three-spans model

(b)Standard ½ + 1 + ½ model

In the ISSC benchmark plate and stiffener imperfections were considered using the following procedure (Fig。 2): a uniform lateral pressure is applied (on the opposite side of the stiffener––tip of stiffener in tension) on the overall structure。 The pressure has to remain small to stay in the elastic range。 The pressure was calibrated to obtain a linear elastic deflection (W) of 2 mm at the central point of the central panel, i。e。 at mid-span of the central stiffener。 Shape and amplitude of the initial imperfections (for plates and stiffeners) are assumed identical to the deflections induced by the uniform lateral pressure。 The considered initial deflection shape is of a thin-horse mode and is composed of several deflection components including the local buckling mode。