PSI - Issue 54

Rahul Iyer Kumar et al. / Procedia Structural Integrity 54 (2024) 164–171 Iyer Kumar, De Waele / Structural Integrity Procedia 00 (2023) 000–000

169

6

which positive peel strains are distributed is defined as the process zone obtained from the experiments. The calculated process zone length and the length obtained from the DIC images for the specimens tested are given in Table 1.

Table 1: The process zone length calculated from the Kanninen-Penado model compared with the length obtained from the DIC images

Specimen number

Calculated process zone ( mm )

DIC process zone ( mm )

Di ff erence

01 04 07 08

21.27 20.83 21.14 21.06

20.96 21.18 20.92 20.67

1.46% -1.68% 1.04% 1.85%

The absolute di ff erence between the calculated process zone length and the one obtained from DIC images is very small, varying from 1 . 00%to 1 . 90%. This suggests that the model describes the stress state near the crack tip well.

3.1.2. Mode I loading The Kanninen-Penado model assumes pure mode I loading condition at the crack tip and the DCB test specimen is experimentally loaded under Mode I condition. However, the location of the initial crack at the adhesive-adherend interface and the thickness of the adhesive introduce asymmetry which could potentially lead to a mixed-mode (mode I + mode II) loading condition at the crack tip. If that would be the case, the SERR should be calculated for a mixed mode condition rather than the mode I condition. The DIC images are once again used to determine the loading conditions by evaluating peel and shear strains at the crack tip region. As seen from Figure 5b and 5c, the peel strain at the crack tip, which corresponds to the mode I loading condition, is at least one or two orders of magnitude greater than the corresponding shear strain. The shear strain around the crack tip has typical values in the range 0 . 008% − 0 . 1% while the corresponding peel strain values vary between 1 . 00% − 1 . 50%. This demonstrates that at the crack tip, mode I loading is the dominant loading mode despite the asymmetric condition introduced in the DCB specimen due to the adhesive thickness and the location of the initial crack tip. DCB specimens with thick adhesive bondline, which were manufactured under shipyard conditions and had an initial crack located at the adhesive-adherend interface, were subjected to fatigue loading. The specimens consisted of two steel plates bonded together with a methyl-methacrylate adhesive that had a nominal bondline thickness of 8 mm. Currently, there are no standards that describe a method to determine the SERR or FCGR curve of specimens with thick adhesive bondlines. This research proposes a method to determine a FCGR curve by taking inspiration from existing literature and a standard developed for testing DCB specimens made of fibre-reinforced polymer matrix composites. The Kanninen-Penado model initially developed by Kanninen based on the Winkler elastic foundation, and ex tended by Penado to determine the energy release rate of the DCB specimen is used to determine the FCGR curve. The Kanninen-Penado model was originally developed and validated for a symmetric DCB specimen under mode I loading conditions. Notwithstanding these conditions, the model is shown to be valid for an asymmetric DCB speci men by analysing strains near the crack tip obtained by the DIC technique. The experimentally determined process zone length is compared to the value calculated using the Kanninen-Penado model, and an excellent correspondence is observed. It is shown that the shear strain near the crack tip is typically one or two orders of magnitude lower than the corresponding peel strain. This shows that the loading condition at the crack tip is mode I dominated. Combined with the process zone length comparison it is concluded that the Kanninen-Penado model is valid for an asymmetrical DCB specimen with thick adhesive bondlines. 4. Conclusion