Table 1: Effect of the HAZ – Reduction of maximum average stress to reference (without HAZ)

Finite element analyses

Two phases were planned。 The aim of Phase A was intended as a calibration assessment。 At the end of this phase, the participants were confident about the quality of their FE models。 Phase A is not discussed in this report。

Then, for Phase B, the participants selected a series of parameters for the sensitivity analyses and each contributor performed different analysis。 These parameters are:

  Welds B + C2 18。1 not considered

  Without HAZ [N/mm²]    160。8 – 173。5 159。6 141。2

with  A  ISSC model and ISSC initial  imperfections

BStandard model and ISSC initial imperfections

CStandard model and Standard initial imperfections

*Interval of values obtained during the ISSC benchmark using different software and meshing features

Phase B2–HAZ width (with Welds A–welded stiffeners)

In Phase B2, the sensitivity analysis concerns the HAZ width (simulations with Welds A)。 Four HAZ widths  (2

1  in plate and 2  in web) were considered (Fig。 3): 2 1

= 25, 50, 75, 100 mm (1 = 2)。 Fig。 7 and Fig。 8 show the results of these analyses, which are compared to the reference case (without HAZ)。

0。000 0。050 0。100 0。150 0。200 0。250 0。300 0。350 0。400

Average strain [%]

Fig。 7: Sensitivity on the HAZ width (2 1 =25–100 mm) (ISSC model and ISSC initial imperfections)

The variation of ultimate strength is not proportional to the HAZ width (2 1)。 The first 25 mm of the HAZ width are the most significant and have the larger effect on the ultimate strength。

For the considered model with Welds A, we observed (Fig。 8):

for 2 1 = 25 mm, the reduction is approximately 9%,

for 2 1 = 50 and 75 mm, the additional reduction of ultimate strength is about 4。5%,

for 2 1 = 100 mm, the additional reduction of ultimate strength becomes smaller, approximately 3。0%。

0 20 40 60 80 100

1,  HAZ Width (mm)

Fig。 8: Sensitivity on the HAZ width (Welds A and A+C1, 21 =25–100 mm)

(ISSC model and ISSC initial imperfections)

Phase B3–initial imperfection

The influence of the amplitude and shape of initial imperfection was studied in this phase (without HAZ)。 The amplitude is considered at the reference point, which is in the centre of the panel。 This deflection may be different from the panel's maximum deflection, as significant local plate deflection may occur。

Shape effect (sensitivity assessment): Two shapes of initial deflection on the standard ½ + 1 + ½ model were considered in the analyses: the initial imperfections defined by the ISSC procedure (uniform lateral pressure) and those defined by the Standard procedure (first buckling mode)。