Issue 35

A. Nikitin et alii, Frattura ed Integrità Strutturale, 35 (2016) 213-222; DOI: 10.3221/IGF-ESIS.35.25 216 Experimental method Torsion fatigue tests were performed on specimens, designed according to ultrasonic concept applied to torsion [3, 14] and taking into an account dynamic elastic properties of materials at 20 kHz (Tab.2). Specimens made of forged VT3-1 were cut from the rim part of compressor disk along an axis of disk symmetry. Specimens made of extruded VT3-1 were machined from bar so that specimen's longitudinal axis is the same than extrusion direction. Geometry of ultrasonic torsion specimens is presented on Fig.3a. Working section of specimen was polished by emery papers from grade 600 to 1000. Fatigue tests were performed continuously (without pulse - pause) by using a self-designed ultrasonic torsion system [14], (Fig.3b) up to fatigue failure or run out limit of 10 9 cycles. All the tests were performed at room temperature with permanent compressed dry-air cooling. An infrared camera was used for monitoring surface temperature of specimen during fatigue tests. Result shows that there is now significant self-heating effect in VT3-1 titanium alloy under torsion loads [15]. (a) (b) Figure 3 : (a) Geometry of ultrasonic torsion specimen and (b) ultrasonic torsion testing system. The calibration of the testing system was performed with strain gauge and Vishay conditioning device with a large bandwidth (up to 100 kHz). Calibration shows a perfect linear relation between applied tension and measured deformation. The fatigue crack under torsion is detected by drop of resonance frequency. This crack detection is done automatically with a high-performance computer feet-back controller. After each test the crack existence is verified by optical microscope. When the crack is detected, a specimen was subjected to self-designed method of specimen opening. As was discussed above, titanium alloy is ductile material and an important plastic deformation may leads to destruction of fracture surface during direct opening. In order to minimize these risks a 'life section' of specimens is reduced by electro-erosion cut. A fill is placed beside a fatigue crack so that surface crack tips are placed on the same line with a wire (in present case it is inclined 45° with respect the specimen longitudinal axis). After reducing the ‘life section’, the specimen is cooled by liquid nitrogen and subjected to sharp shock, so that provide fatigue crack opening. After opening, all the cracked specimens were analysed by using scanning electron microscopy (SEM). An additional attention has been paid to the crack initiation mechanisms. R ESULTS SN-curves for forged and extruded Ti-alloys he results of fatigue tests on both forged (Fig.4a) and extruded (Fig.5a) titanium alloy shows that fatigue failure under pure torsion loading may occur well beyond 10 6 cycles. In spite of limited fatigue data on forged VT3-1 it is possible to plot a curve fitting the results. This curve will have an important slope in the VHCF range. Decreasing in fatigue strength with increasing number of loading cycles is more pronounced for torsion mode compared to results of axial tension-compressing tests [16] that is presented on Fig.4b. In order to compare the results of torsion and axial tests, the following equation was used to recalculate the shear stress amplitude into equivalent (Von Mises) normal stress T