PSI - Issue 57

Philipp Ulrich Haselbach et al. / Procedia Structural Integrity 57 (2024) 169–178 P. U. Haselbach and P. Berring / Structural Integrity Procedia 00 (2023) 000–000

175

7

the crack growth parameters for the adhesively bonded joints in the study presented. During the experiments, the structural adhesive showed a relatively brittle behaviour and the crack paths and crack path transitions mainly grew inside the adhesive or were observed to be associated with micro-fracturing along laminate surface and showed only little cracking kinks into the Biax material for Mode II loading of the fully bonded joints. Similar behaviour for the same adhesive was observed by Samborsky, D. et al. (2012). Due to the relatively brittle material characteristic, it is decided to model the crack with the VCCT.

3. Results and discussion

3.1. Numerical simulations and analyses

Several numerical simulations were conducted for di ff erent load scenarios, assuming multi- / dual-axis excitement during fatigue testing. The most serve load case is a combination of suction to pressure side (flapwise) loading and simultaneously being loaded in edgewise direction (leading towards trailing edge). Here, with the use of the VCCT, the E ff ective Energy Release Ratio (EFEBRRTR) at nodes between the fully debonded area and the adhesive low (30%) bonded surfaces is investigated, as well as the individual SERR for Mode I, Mode II and Mode III in comparison to the critical SERR based on the fracture toughness as material property determined by material characterisation. This specific load case leads to SERR and an E ff ective Energy Release Ratio exceeding the fracture toughness of the adhesive bondline and thus, leading to crack growth exceeding the initial debonded area towards the blade root. Moreover, the result clearly shows that the L-shaped form of the shear web (see Figure 2) a ff ects the stress distribution in the bondline by introduction stress concentrations, where the vertical shear web is connected to the spar cap region as shown in Figure 4. The initial debonded area grew only towards the direction of the blade root. No critical SEER nor E ff ective Energy Release Ratio were predicted to develop from the initial debonded area towards the blade tip in any of the simulated load cases. Several di ff erent aspect ratios between fully debonded regions and areas with low and medium bonds were inves tigated. In Table 4 the SEER and E ff ective Energy Release Ratio are given for four di ff erent dual- / multi-axial load sequences, where crack growth at a moderate speed is expected to occur. This crack growth behaviour is regarded as being favourable, because crack growth seems to be likely to occur but is also expected to progress at moderate rate and thus, not leading to a sudden blade failure. For the four load cases, the shear web / spar cap bond is experiencing di ff erent loads caused by combinations of Suction towards Pressure side (STP), Leading Towards Trailing edge (LTT), Pressure Towards Suction side (PTS) and / or Trailing Towards Leading Edge (TTL) load combinations. Based on the evaluation of di ff erent load scenarios, it was decided to manufacture the blade with an initial debonding between r = 1.45mand r = 2.55 m, enclosed by low bonded and medium bonded areas to both sides, each with a length of 0.45 m in length as depicted in Figure 5, which refers to the here presented SERR and e ff ective Energy Release Ratio.

Fig. 4. E ff ective Energy Release Ratio (left hand figure) and SEER for mode II (right hand figure) both exceeding the fracture toughness of the adhesive bondline and thus, leading to crack growth beyond the initial debonded area. The result clearly shows that the L-shaped form of the shear web a ff ects the stress distribution in the bondline by introduction stress concentrations, where the vertical shear web is connected to the spar cap region.