PSI - Issue 79

Abubakr E.S. Musa et al. / Procedia Structural Integrity 79 (2026) 206–216

213

Fig. 5. A typical example of the generated nodes with different thicknesses. The output of this code includes the nodes numbers and their respective thickness values. These data were then used to define a discrete field in ABAQUS, replacing the uniform thickness assigned in the previously prepared base FE model. This process ensures the accurate representation of the thickness reduction imperfection in the FE model, providing a realistic simulation of corrosion damage for subsequent analysis. 4. Results and discussion The FE simulations were conducted for all the specimens tested, accounting for both the unintended initial imperfections, represented by equivalent geometric imperfections, and the intended imperfections, simulating corrosion damage. The numerical results of the failure loads obtained from these simulations are presented in Table 3, compared with the corresponding experimental results. These data are also used to generate Fig. 6 to enhance the comparison further. The comparison reveals that the proposed FE model exhibits a reasonable level of accuracy in predicting the failure loads, with the experimental-to-FE load ratio ranging from 0.86 to 1.16. Table 3. Comparison between the experimental results and FE model predictions. No. Specimen Designation Thickness variation ∆ t ∆ t/t P Exp (kN) P Exp (ave.) (kN) P FE (kN) P Exp /P FE

1 2 3 4 5 6 7 8

T1-1 T1-2 T2-1 T2-2 T3-1 T3-2 T4-1 T4-2

0 0 1 1

327 318 259 283 252 255 229 271

0

322.5

322.47

1.00

0.33

271

315.74

0.86

1.5 1.5

0.5

253.5

264.21

0.96

2 2

0.67

250

214.85

1.16