PSI - Issue 76

Daniel Perghem et al. / Procedia Structural Integrity 76 (2026) 107–114

110

curve of each series:

∆ σ = A · N − b

(1)

Hodge-Rosenblatt’s up and down method with small samples and a short stair-case sequence Brownlee et al. (1953) for the fatigue limit estimation was used in the current investigation. The results of the fatigue test at R = 0.1 of the 6 series of 4PB fatigue specimens, are presented in Fig.2, normalized by the endurance limit ( ∆ σ w ) of the machined series. The downskin specimens show the lowest fatigue strength, with

Fig. 2. (a) The results of the fatigue tests conducted at R = 0.1; SEM observations of typical critical anomalies: (b) Vertical series; (c) Horizontal series; (d) Downskin series; (e) Upskin series; (f) Downskin chemically milled series.

about 55% reduction in the fatigue limit region compared to machined specimens. The horizontal specimens exhibit a fatigue limit comparable to the machined series. Surface treatment by chemical milling on the downskin surface enhances the fatigue properties. The vertical series present fatigue behaviour in the finite life region similar to the downskin condition. All series show a comparable slope in the finite life regime, which, however, di ff ers from the one observed for the machined series. Di ff erent types and shapes of anomalies were observed at the failure origin depending on the build orientation. High-quality surfaces, horizontal (Fig. 2c) and upskin (Fig. 2e), typically exhibited grain aggregates or intrusions, whereas rougher surfaces, vertical (Fig. 2b) and downskin (Fig. 2d), were characterized by elongated grooves. Chem ical milling enhanced surface quality by eliminating rough downskin features, leading to failures originating from grain aggregates (Fig. 2f). 3.2.1. Modelling through El-Haddad model In this study to consider the short-crack e ff ect for the endurance limit ∆ σ w the El Haddad et al. (1979) model is utilized: ∆ σ w =∆ σ w , 0 · √ area 0 √ area + √ area 0 with √ area 0 = 1 π · ∆ K th , LC Y · ∆ σ w , 0 2 (2) 3.2. Fatigue strength model