PSI - Issue 68

Louka Eleftheria-Sotiria et al. / Procedia Structural Integrity 68 (2025) 894–900 Louka Eleftheria-Sotiria et al. / Structural Integrity Procedia 00 (2025) 000–000

897

4

4. Finite element modelling To evaluate the effect of corrosion damage on the mechanical behavior, a 3D finite element model of the gauge length of the tensile specimen was modelled in the ANSYS software. The inputs of the model are the geometry of the specimen and the stress-strain curve of the reference material up to the maximum stress. Symmetry was considered in the width and length directions but not in the thickness direction due to one-sided corrosion damage. Elasto-plastic analysis was conducted with a displacement load step corresponding to 0.4% of nominal strain. Stress was calculated from the reaction force in the loading direction. The output of the model is the stress-strain curve, i.e., the tensile mechanical behaviour of the corroded specimen. 4.1. Modelling of corrosion damage To model corrosion damage, experimental measurements of mean and maximum depths of attack and surface coverage were used. The corrosion damage was modelled using ellipsoidal volumes “cutting” the outer surface of the specimen. The position of the ellipsoid in the width and length of the specimen is chosen completely randomly. The size and the distance of the ellipsoid from the surface (thickness direction) is selected such that when it cuts the surface the area covered, and the depth are equal to prescribed values. In this way, the total area covered, and the distribution of depth can be controlled. The depth of dents is random between a minimum and maximum value. The number of dents ranges from 10 to 30 depending on the coverage. Typical full mesh and outer surface mesh are shown in Fig. 2.

Fig. 2. Images from the finite element modelling of the corroded specimen: (a) full mesh; (b) outer surface mesh with corroded products.

4.2. Failure criterion Failure of the specimen was assumed when the equivalent strain exceeds the maximum strain of the reference material through the width of the specimen. The corresponding nominal strain is assumed to be the failure strain of the corroded specimen. The following figures (Fig. 3 & Fig. 4) show typical results of the finite element modelling with respect to equivalent strain distribution for the T3 temper and for different corrosion exposure times, namely 2 h (Fig. 3) and 24 h (Fig. 4) pre-exposure to exfoliation corrosion solution. Gray area corresponds to equivalent tensile strain greater than the maximum strain of the reference material.