PSI - Issue 77

Magdalena Mieloszyk et al. / Procedia Structural Integrity 77 (2026) 256–263 M.Mieloszyk & S.Bhadra / Structural Integrity Procedia 00 (2026) 000–000

260

5

Fig. 3. THz images of 2-layer GFRP structure: (a) B-scan, (b) C-scan of top surface, (c) C-scan of mid surface, (d) C-scan of bottom surface.

fibre optics with FBG sensors results in the epoxy resin flow disturbances. In the final structure, they are covered by an additional layer of polymer in the form of a resin pocket. The sensors were embedded between the 1st and the 2nd layers, counting from the PTFE plate. During the THz measurements, the sample was placed analogously. Therefore, in the B-scan (Figure 2(a)), the bottom surface is flat (PTFE plate) and the top surface is rough (porous ply). Also, the embedded FO influence is observable in the structure as a local discontinuity and an increase in sample thickness. C-scan (Figure 2(b)) was determined for the plane where the FOs were embedded. They can be easily determined as darker, straight parallel lines. An example of the internal structure of the GFRP sample manufactured using the mFDM method is shown in Fig ure 3. As it was presented in the scheme (Figure 1(b)), GFRP sample structure has a characteristic fibre reinforcement pattern. Due to the high roughness of the sample, it is hard to determine the fibre reinforcement pattern on the upper surface (Figure 3(b)). However, it is well visible in C-scan for the bottom surface (Figure 3(d)). Glass fibre layers are separated by a polymer layer, and it is well visible in C-scan for the mid-surface (Figure 3(c)). A comparison of the structures of AM samples after manufacturing (denoted as A) and exposition to elevated temperatures (denoted as B) is presented in Figure 4. The thermal treatment was in the form of slow heating from 10 ◦ C to 50 ◦ C with stabilisation every 5 ◦ C. The cross-sections of the samples with marked embedded FO are presented in Figure 4(a). In the structure of 4-layer samples is also possible to determine polymer layers between consecutive fibre layers. They are analogous to the mid-surface presented in Figure 3. The heat treatment process influenced the polymer layers, so the polymer filled the emptiness between fibres. It is observed as a reduced thickness and width of the sample. Embedded FO is not as visible as in the samples manufactured using the infusion method. Because no additional load was added during the process, all glass fibres 3D-printed over the FO are moved up. It is visible in two C-scans for the intact sample, for the top surface (Figure 4(b)) and the FO-plane (Figure 4(c)). As the elevated temperature influenced the polymer, it resulted in a higher regularity of the sample structure. The FO is less visible in the image for the FO plane (Figure 4(c)) in the sample after heating than in the intact state.

3.2. Scanning Electron Microscopy

SEM of the additively manufactured GFRP exposed to elevated temperature revealed the sequence of thermally activated damage. The first micrograph (Figure 5(a)) shows fibre-matrix debonding at the ply edge, with a clear interfacial gap and localised tearing of the polymer. Such decohesion is consistent with softening of the matrix material (PLA) near the glass transition temperature Tg ( ∼ 65 ◦ C) and di ff erential thermal expansion between glass fibres and