PSI - Issue 41

Venanzio Giannella et al. / Procedia Structural Integrity 41 (2022) 298–304 V. Giannella / Structural Integrity Procedia 00 (2022) 000–000

301

4

To this aim, a crack propagation simulation was performed for each FEM model (i.e. for various fillet radius dimensions) so as to derive a distribution of K values along the crack depth for each fillet radius. These crack propagation simulations were developed thanks to the commercial code FRANC3D; through this latter it was possible to simulate the propagation of a crack having user defined size and shape, initiated in the most stressed location, and propagating through the structure according to geometry, loads and material properties. Particularly, an initial crack having a circular cross-section and radius r = 0.3 mm was assumed, with an initial orientation defined as being perpendicular to the local maximum principal stress, see Figure 3. Then, the crack propagation simulation was performed automatically and incrementally, by considering the Maximum Tangential Stress (MTS) as the criterion to calculate the crack deflection angle along the propagation (Erdogan et al., 1963). Further examples and details on how these crack-growth simulations work can be found in literature (Giannella et al., 2017, 2021a-c; Citarella et al., 2018). According to the MTS criterion, the crack resulted to propagate perpendicularly to the maximum stress acting along the tangential direction (i.e. radially), see Figure 3. This confirmed also further similar investigations available in literature (Citarella et al., 2016; Shlyannikov et al., 2021; Giannella et al., 2019). The simulations were stopped once reached critical crack sizes (i.e. at nearly 10 mm of depth) and can therefore be considered as up to failure. Consequently, K values were extracted at a fixed position along the crack size and were compared for various fillet radii, see Figure 3 (only K I values were reported, being the only non-null K factors). It is worth noting that K I values resulted to be very different for the initial crack size (crack depth = 0.3 mm) with the expectable trend of having much higher K I for the lower fillet radii. This well agrees with the same “notch effect” already highlighted by the stress distribution reported in Figure 2. However, as long as the crack radially propagates through the component, stresses result to be lower and this in turn affects the K I values that turns out to have an opposite trend. As a consequence, the influence of the fillet radius seems to play a minor role on the crack propagation life than that expected in terms of crack nucleation life.

Figure 3. Crack-growth propagation in the aircraft compressor stage perpendicular to the hoop stress acting in the tangential direction; crack shape evolution during the growth; K values along crack depth calculated for various fillet radius sizes.