Volume 57, Issue 5

Metallographic and electron microscopic methods were used to understand the microstructural changes occurring during the additive manufacturing of components made of X40CrMoV5-1. In this hardening steel, tempering effects, which can locally be very different, occur as early as during additive manufacturing. For the determination of the micro- and macroscopic microstructural inhomogeneities, this work focused on the metallographic characterization based on macro- and micrographs as well as on hardness measurements. In addition, local inhomogeneities in the chemical composition and the microstructure were characterized by ESMA and EBSD. The examinations were accompanied by the characterization of the powder.

In this work, the effect of friction stir processing (FSP) on microstructure and mechanical properties of an Al-5Mg-0.2Sc-0.1Zr alloy was investigated in detail. The results showed that the as-cast alloy exhibited equiaxed grain structure with a large number of primary Al 3 (Sc,Zr) particles inside. FSP imposed a significant grain-refining effect on the alloy, changing the average grain size of 50.5 μm to 2.4 μm, which is attributed to the occurrence of dynamic recrystallization and generation of nanoscaled Al 3 (Sc,Zr) precipitates. These Al 3 (Sc,Zr) dispersoids intense impeded the movement of dislocation and grain boundaries and hindered the growth of recrystallized grains, resulting in a remarkable grain-refining strengthening, substructure strengthening and precipitation strengthening effect. On the other hand, the residual primary Al 3 (Sc,Zr) particles gave rise to a loss in ductility.

One of the most important but yet partly neglected steps in quantitative microstructural analysis is microscopic image segmentation. Despite major advances in the currently available segmentation methods and first of all, machine learning segmentation, the metal industry in many cases still relies on segmentation by thresholding. Based on microstructural images of low-alloy steels, this work will show to what extent machine learning segmentation can accomplish a significant quality improvement compared to the traditional thresholding method.

Swivel arms in a turning gear application failed in service due to fatigue cracking. 2 out of 52 components were affected. Turning gears are mainly used to spin heavy-duty gas turbine engines rotors after shut-down to prevent distortion of the still-hot rotor. The design of the swivel arms, that was not robust enough to tolerate deviations in fabrication, was identified as the root cause of the failures.