PSI - Issue 42

Sarim Waseem et al. / Procedia Structural Integrity 42 (2022) 1692–1699

1697

6

Waseem et al. / Structural Integrity Procedia 00 (2019) 000–000

No Overload 30% Overload 50% Overload 70% Overload 100% Overload

Fig. 2. Notched specimen crack pattern (Left); Crack growth vs cycles (Right)

overload demonstration is limited to 400 cycles. Due to the relatively slow evolution of the phase field per cycle, we were able to be less conservative with the incrementation without the usual stability problems of staggered schemes and hence allowing the easy computation of a higher number of cycles. However, this is not a sustainable methodology to reduce computation times while retaining accuracy and a better solution would be using Quasi-Newton monolithic solution schemes to simulate a realistic number of cycles within a reasonable computation time (see e.g. Navidtehrani et al. (2021), Kristensen and Mart´ınez-Pan˜eda (2020)).

3.2. Three Point Notched Bending Test with Asymmetric Perforations

Next, the crack propagation path prediction is studied through a 3-point bending test with stop holes to study the performance of the framework in complex geometries. To avoid the damage at the load and boundary condition points, the specimen is dissected into regions (see Fig. 3), with the phase field UMAT only applicable in the region where the crack is expected to grow. The mesh is made dense in the region where the crack is expected to propagate.

UMATREGION

Fig. 3. 3 point bending test with asymmetric perforations (Left); Mesh used (Right)

First, a simulation using a simple phase field fracture model without fatigue e ff ects is conducted to make a compar ison in the crack pattern. It is found that in the simple static case, the crack is drawn into the stop hole, in contrast to the Ingra ff ea and Grigoriu (1990) experiments (see Fig. 4). Molna´r and Gravouil (2017) shows that the crack is drawn into the stop hole for a similar case, noting that di ff use crack approaches are perhaps unsuitable for modeling such problems with crack paths narrowly avoiding stop holes. However, the phase field fatigue methodology introduced in the current work is capable of capturing the crack path experimentally obtained by Ingra ff ea and Grigoriu (1990) as shown in Fig. 4. There is a deviation in the phase field solution and the experimental results near the end of the