PSI - Issue 19

Marc J.W. Kanters et al. / Procedia Structural Integrity 19 (2019) 698–710 Marc Kanters et al./ Structural Integrity Procedia 00 (2019) 000–000

708

11

Figure 11: Local stresses in the failure location of the demonstrator part loaded in with a minimum load of -1500N (a) and a maximum load of 1500N (b).

Note that on top of the geometrical asymmetry, boundary conditions also play a role in the local stress state. Here, for example, the critical location with maximum stress is nearby the bolt and washer that are used to mount the demonstrator part to the fixture (not shown in Figure 11, but present in the FEA model). Hence, the critical location experiences also a small tensile load due to the bolt pretension, also influencing � . The local stresses and stress ratio, � , depend on the load level in tension and the load level in compression, defined by the maximum force applied and � . As highlighted in Figure 12b, the local magnitude of � clearly deviates from the macroscopically applied � . For example, when applying � = −1 the value of � varies from 0.08 to -0.53 within a maximum applied force range of 0 - 3 kN. And, as Figure 7a already shows, the variations in lifetime can be easily in the order of a factor 5 to 100. Note that where the linear elastic computation makes the local stresses scale with applied load, the slope of the Basquin curve from part testing should be equal to that of the standard specimens. However, the dependence of � on maximum force will affect the slope of the fatigue curves at constant � . This highlights the necessity of considering the local stress ratio, instead of assuming a (constant) ratio equal to the macroscopically applied ratio. The Digimat software can do this via three routes: i) Manually determine the local magnitudes of principal stresses in the maximum and minimum load to manually compute � and let the software compute lifetime using that value, ii) Use two load cases with opposite directions, define the increments of interest that correspond to a macroscopic load and � , after which the software computes � and the resulting lifetime, or iii) Use a load case that simulates one load cycle, and allows the software to determine the minimum and maximum stress levels, the value of � , and the lifetime.

Figure 12: (a) Model of the demonstrator part. (b) R-value as function of applied load for the demonstrator part, where markers represent the local stress ratio, � , and the dashed lines the macroscopically applied force ratio, � .