PSI - Issue 70

Anil Pradeep Konda et al. / Procedia Structural Integrity 70 (2025) 153–160

156

The I-section used in the model consists of flanges with a thickness of 20.3 mm and a web with a thickness of 12 mm. In the shell-based model, the I-section is modelled by assigning S4R elements (a 4-node linear shell element with reduced integration) to b oth the web and flanges and by applying ‘tie’ constraints to connect the web and flange regions (Fig. 1(b)). For the solid model, the entire I-section geometry, comprising both web and flanges, is extruded as a single monolithic section and meshed using C3D8R elements (8-node linear brick elements), as shown in Fig. 1(c). Since shell elements are generally more effective at transferring in-plane shear than solid elements, a hybrid modelling approach for I-section using S4R elements, while the flanges were assigned C3D8R elements as illustrated in Fig. 1(d). The girder has a total length of 18 meters, with a Young’s modulus of 2 ×10 5 MPa and a Poisson’s ratio of 0.3. The self-weight of the girder is excluded from the analysis to ensure consistency with the damage detection methodology adopted in this study. The analysis was carried out using a 10 mm mesh size, with a concentrated load of 100 kN applied at the mid -span. The resulting deflections are summarized in Table 1.

Table 1: Difference between theoretical and simulated values Steel Plate Girder Theoretical Value at mid-span

Finite Element Simulation

= 3

48 = 25.284

Shell Model Solid Model Hybrid Model

24.844 mm 25.811 mm 25.864 mm

As observed from Table 1, both the monolithic solid I-section and the shell model closely replicate the expected structural behaviour of the I-section. However, the solid model was selected for further analysis, as it allows for explicitly introducing damage in the form of cuts, which is crucial for the subsequent investigation. 4. Castellated Girder The investigators generated the castellated girder geometry by introducing a series of zig-zag cuts along the web of ISMB600, as shown in Figure 2, resulting in a hexagonal opening pattern. The depth of these cuts was 300 mm, leading to an increased overall depth of 900 mm after rejoining the separated segments. The material properties used in this model were identical to those employed in the pilot studies to ensure consistency in comparative analysis.

Fig. 2: Proposed zig-zag cut pattern on ISMB600 section Only the top half of the girder is modelled in this study, considering the geometric symmetry of the castellated girder about its mid-depth. As illustrated in Figure 3, the complete castellated profile is formed by duplicating and inverting the top half, followed by applying a 'tie' constraint between the two halves. Authors intentionally introduce the tie constraint to investigate the behaviour of the castellated girder in scenarios involving potential weld failure. A girder section featuring an increased overall depth of 900 mm was introduced in this study. This section's flange and web thicknesses are 20.3 mm and 12 mm, respectively. The finite element model uses a global mesh size of 10 mm to capture the geometry closely. The model primarily uses C3D8R elements. C3D6 elements (6-node linear triangular prisms) better represent the wedge in the regions created by the castellation geometry, as shown in Fig. 3.