PSI - Issue 42

Niklas Ladwig et al. / Procedia Structural Integrity 42 (2022) 647–654 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

653

7

4.2. Traction vectors As mentioned before, the traction at the initial crack tip can be calculated according to the previously given equations (5) by applying the direct identification method [8]. That method does not require integrability of the traction-separation law, but it assumes that the process zone (the region in which energy is released) can be approximated to shrink to a point. That assumption is expected to be fulfilled in case of cantilever-like specimens and rather stiff adhesives and it can easily be checked by plotting the measured components of COD. In case that the process zone can be assumed to be concentrated in a single point, the measured COD paths 3 ⁡( 1 ) must be a straight line passing the origin up to the point of crack propagation. These paths in COD are given (grey lines) in figure 5. It can be seen that the assumption is fulfilled and the traction is then calculated by the direct method. Tractions vectors are given as red arrows in figure 5.

Figure 5 - traction vector and COD paths for Henkel Teroson® EP4510GB and SikaPower®-498

5. Conclusion Multiple tests at a constant mixed-mode ratio between mode I and mode III loading were performed until crack propagation. Two different adhesive systems were investigated, one linear elastic brittle and one elastoplastic. Fracture envelopes with a high resolution in mode-mixity for these adhesive systems were determined, which confirmed the applicability of the new test method for different kinds of adhesive joints. An increase in fracture energy close to single mode III loading was observed for both adhesives, but a physical explanation for that observation is not given so far. However, it seems that possible effects of plasticity may not play the major role as the observed effect is of considerable magnitude also in case of an adhesive, which is not expected to show relevant plastic behavior. In addition, traction vectors at the crack tip were calculated based on a direct identification method, which has been recently proposed in previous work. The method relies on some requirements such as strictly monotonic loading of the specimen and small dimensions of the fracture process zone. The validity of these assumptions could be proven and confirmed experimentally by interpreting the crack tip opening displacement. As a consequence, the test setup seems to be appropriate to directly extract information about constitutive laws for cohesive zone modeling, which might be used afterwards in numerical simulations.