PSI - Issue 37

Mihaela Iordachescu et al. / Procedia Structural Integrity 37 (2022) 203–208 Iordachescu M. et al./ StructuralIntegrity Procedia 00 (2019) 000 – 000

206

4

has been assisted by the un-relieved residual tensile stresses induced during the rolling of the thread and developed during a long service period, once the Zn-layer protection had stopped being effective.

Fig. 4. Fracture surface of the failed threaded bolt: (a) macroscopic sketch; (b) macroscopic image; (c) detail showing the 1 st fatigue cracking initiation site; (d) detail of the ratcheting marks due to multiple fatigue crack initiation.

The sketch and the images contained in Fig. 4 illustrate the complex rupture of the failed bolt. The ratchet marks found on the fracture surface indicate that the rupture was initiated in multiple points located in close but distinct circumferential planes of the thread root (Fig. 4a and Fig. 4b) and propagated in a stable way by fatigue cracking. The first initiated fatigue crack produced a stress redistribution along the thread root able to activate some environmentally induced fatigue initiators, from which new cracks developed. These new cracks initiate on close but distinct planes and grow with the fronts laterally bounded by small shear walls (ratchet marks) that separate the cracking planes (Fig. 4d). The ratchet marks disappear as the crack fronts move away from the root thread and the cracking planes merge to form a single crack which grows at the expense of the resistant ligament of the bolt until it collapses by overloading. 3. Fatigue resistance of the bolt steel In order to determine the fatigue resistance of the bolt steel, three notched cylindrical specimens of diameter D = 4.4 mm were machine cut from the unbroken bolts, smooth notched (Fig. 5a), and fatigue tested. The specimens were subjected to cycles of tensile load at room temperature, with constant amplitude and a frequency of 8 Hz. Maximum load was decreased a number of times along the duration of each test to reduce load amplitude by between 20% and 30% and obtain a maximum number of experimental data from the available material. The maximum load applied in the first fatigue step was well below the yielding load of the notched cross-section. The crack size at the end of each fatigue step was monitored by measuring the stiffness of the cracked specimen as the slope of the load-COD curve when unloaded and reloaded between two consecutive fatigue steps (Fig 5b). A resistive extensometer faced the crack over a short gage length provided the COD measurements. The last fatigue step was prolonged up to fracture in order to expose the cracked cross section, as seen in Fig 5c. The images captured by optical microscopy allowed determining the successive crack fronts and their dimensions a (the crack depth) and 2b when modeled as symmetric arcs of ellipses with the center on the specimen surface (Toribio J. et al., 2009). In accordance with this modelling, the fatigue crack growth rate was assessed at the deepest point of the crack front by applying the Paris – Erdogan law (Paris P. C. and Erdogan F., 1963) to each load cycle at the deepest point of the crack front: da dN = C ∆ K ( ) m (1) where ∆K and da/dN are the stress intensity range and the crack growth per fatigue cycle, in each cycle. For the load