PSI - Issue 77

Jiongyi Yan et al. / Procedia Structural Integrity 77 (2026) 135–142 J. Yan/ Structural Integrity Procedia 00 (2026) 000–000

140

6

Shell elements were selected in the numerical model, justified by their efficiency and accuracy in capturing bending behaviour of composite structures under tensile loading. A displacement-controlled boundary condition was applied to one end of the model (the right end in Fig. 3), reflecting the static tensile tests conducted experimentally, while the opposite end was fixed. The orthotropic elastic properties of the 3D-printed composite were defined based on experimentally characterised data. The longitudinal and transverse Young's moduli were 4.83 GPa and 2.70 GPa (Yan et al., 2023a), respectively, with a Poisson’s ratio of 0.3 (Yan et al., 2023b). To accurately represent the failure behaviour, a damage model was implemented. Damage initiation was governed by the Hashin failure criterion, which provides distinct failure indices for tension, compression, and shear. This criterion is well-established for its ability to predict the onset of damage in the composite. The material orientation was defined along the structure's geometry, with a symmetric fibre orientation specified at the corner to reflect the manufacturing process. Following damage initiation, the stiffness degradation of the composite was controlled by an energy-based damage evolution law. This approach defined the fracture energy required for a crack to propagate after initiation, allowing the model to simulate the progressive hinge-opening and damage propagation rather than immediate, brittle failure. Results of the simulation showed a reasonable match in bending behaviours of the composite between tensile test (Fig. 3a) and numerical models (Fig. 3b). FE models predicted the complex failure sequence observed in the tensile tests, revealing a significant stress concentration up to 68 MPa at the location near the inner side of the corner (Fig. 3c). This concentration first initiated the crack of the composite in the vicinity of the triangular section (Fig. 3d). Subsequently, the model captured the asymmetric propagation of this crack (Fig. 3e). This asymmetry was critical, as it led to an uneven load redistribution within the specimen, causing the observed skewing or bending of the structure prior to failure. The final fracture (Fig. 3f) was predicted to occur following the substantial damage and structural reorientation in the composite.

Loading direction

(a)

(b)

(c)

(d)

(e)

(f)

MPa

50 70 60 40 30 20 10 0

Fig. 3. Mechanical simulations of 3D printed corners. The comparison of deformed and bent specimens between the (a) experiment and (b) simulation. The simulated (c) stress concentration, (d) crack initiation, (e) propagation, and (f) failure due to bridging of cracks.

3.4. Cyclic tensile properties of angular corners For displacement-controlled cyclic tensile, we tested the strongest specimen (30 o corner) and found significant reduction in peak force in the second cycle compared to the first cycle, where the peak force at the target displacement reduced by <5%. This suggests that the yielding has occurred for the specimen when deformed and straightened. After