Issue 35

A. Nikitin et alii, Frattura ed Integrità Strutturale, 35 (2016) 213-222; DOI: 10.3221/IGF-ESIS.35.25 219 (a) (b) Figure 7 : Surface (a) and subsurface ‘fish-eye’ (b) crack initiations in extruded VT3-1 titanium alloy under torsion loading. The next difference between torsion and pull-push ‘fish-eye’ is the nature of smooth area. In the case of axial loading, the formation of such area is governed by crack growth rate, while in the case of torsion ‘fish-eye’ a second factor can be stated. Indeed the smooth area of torsion crack is limited by an ellipse, that ‘touches’ the specimen surface. More rough fracture surface is starting to form when an internal crack reaches the specimen surface. Probably, there are several factors acting together and leading to fracture roughness modification. One can say: (1) the crack growth rate increasing when the crack reaches the specimen surface; (2) presence of environment (gasses) into the crack, when it connects to the surface. Anyway, a smooth area of torsion ‘fish-eye’ exists till an internal crack turns to a surface crack. The next pronounceable difference between surface and subsurface initiation is less expressed branching of internal crack. On Fig.7a several clear traces can be observed. These traces are formed due to crack propagation in series of parallel planes, orientated at 45° with regard to the specimen longitudinal axis. In the case of subsurface crack initiation, growth of several cracks in 45° planes is also possible, but this is well limited. It is interesting to point out, that in the case of high strength steel, Fig.1a [10], a fracture surface with internal crack initiation does not show a significant branching pattern (no 45° ‘wings’ ). D ISCUSSIONS irst of all, the comparison of SN-curves for push-pull and torsion loadings (Fig.4 and 5) shows a more important slope of SN-curve in the case of torsion loading. This tendency keeps being the same as for forged, as well for extruded titanium alloy in spite of small difference of SN-curves for these alloys under push-pull fatigue. This means that VT3-1 titanium is more sensitive to shear stress than to normal one. This is typical for ductile metals. However, long crack propagation is observed in planes of maximum normal stress which is typical for more brittle material. Therefore, at the very first stage of fatigue crack initiation, when a crack length is about the same order than the grain size the fatigue behavior of titanium is similar to ductile material and fatigue damage accumulation is due to sliding process. In the case of two-phase    titanium alloy a higher capacity to accommodate plastic sliding has a hexagonal alpha-phase. Thus, the fatigue resistance of titanium alloy to torsion loading may be related to features of alpha-platelets. In the case of surface torsion crack it can be observed macroscopically at surface in one of two maximum shear stress planes: along or perpendicularly to specimen’s axis. Fig.8b shows an example of first torsion crack growth stage along an axis. When torsion crack reaches a length of several micrometers, the crack growth turns into a plane of maximum normal stress, Fig. 8b. Typical size of alpha-platelets is very small for the studied titanium alloys and micro-plasticity of alpha- phase is not enough to accumulate fatigue damage at later stage of crack growth that turns fatigue behavior of material to the brittle-mode failure (governed by the normal stress cracking mechanisms). Another reason that can limit the stage of crack growth in shear plane is quite high deformation rate in the case of ultrasonic loading. But anyway, a transition from maximum shear to normal stress plane is typical for torsion cracking at different loading frequencies. In reference [18] fully reversed torsion fatigue tests were carried out on titanium alloy Ti-6Al-4V in HCF regime at a loading frequency of F