PSI - Issue 17

Ludvík Kunz et al. / Procedia Structural Integrity 17 (2019) 222–229 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

224

3

in Fig. 1. They had spherical shape, the average particle diameter was 15  m and the median of particle diameter was 12  m. As-built specimens were processed by three different heat treatments. The first one consisted from stress relieving at temperature 380 °C for 8 h in argon atmosphere with temperature rise within half an hour and cooling down to room temperature. The second heat treatment was done by heating to temperature of 740 °C during 1.5 h with 2 h dwell at the desired temperature and subsequent cooling down to 520 °C in vacuum followed by cooling to room temperature in argon. The third heat treatment consisted from heating to 900 °C during 1.5 h with 2 h dwell at the temperature in vacuum followed by controlled linear cooling to 500 °C during 2 h and a rapid cooling to room temperature.

specimen a

specimen b

specimen c

Fig. 2. Scheme of specimen building directions.

Fig. 1. Powder used for specimen production.

The characteristic dimension W of CT specimens was 30 mm and the specimen thickness t was 6 mm. Because of inaccurate dimensions of the notch and particularly the notch tip radius the notch was corrected by machining to the required shape. Three different orientations of specimens with regard to the fatigue crack growth plane were built, see Fig. 2. The fatigue crack in specimens denoted as a propagates from the notch in a plane parallel to the build direction and perpendicularly to the layer planes. In specimens denoted as b the crack propagates in the plane parallel to the build direction and the crack growth direction is parallel to the layer planes. Specimens in which the crack propagates in a plane perpendicular to the build direction and the crack growth direction is parallel to the layer plans (growth along the layers) are denoted as c . Because the surface quality of lateral sides of the as-built specimens was not suitable for optical determination of the crack length, the specimens were finely ground and polished. The experimental determination of the growth of fatigue cracks was performed on two fatigue testing machines, namely on a resonant Roell/Amsler HFP 5100 and on an Instron E3000. The length of the crack was monitored optically on both sides of the specimen by means of CCD cameras. The accuracy of the crack length measurement was 0.01 mm. The lowest crack growth rates were reached by load shedding method according to the Standard test method for measurement of fatigue crack growth rates ASTM E647. The results of the determination of the crack growth rate in dependence on the stress intensity factor range  K for cycling at the stress ratio R = 0.1 in material heat treated at 380 °C are presented in Fig. 3. The experimental points for all three orientations lie in one relatively broad scatter band. There is no apparent difference in the fatigue crack growth rate in different directions with respect to the direction of DMLS building. The lowest determined crack growth rate of the order 10 -8 mm/cycle corresponds to the value 3.7 MPam 1/2 . If this value is considered a threshold  K th for all the three orientations, the fatigue crack growth curve da/dN = 3.0 × 10 -8 (  K 3.0 -  K th 3.0 ) which is shown by the curve in Fig. 3 fits well the experimental data in the whole measured interval. The dependence of the crack growth rate on  K for R = 0.1 and material heat treated at 740 °C is shown in Fig. 4. The CT specimens were built on two different DMLS machines with different laser power. In the case of specimens manufactured with 200 W machine all three orientations were investigated whereas for 400 W only orientations a and c were examined. It can be concluded that all data lie in one scatter band. There is no measurable influence of 3. Results