Dresden 2020 – wissenschaftliches Programm
Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
MM: Fachverband Metall- und Materialphysik
MM 64: Mechanical Properties: Plasticity, fracture, fatigue, wear - II
MM 64.1: Vortrag
Donnerstag, 19. März 2020, 17:30–17:45, IFW B
Can machine learning be used to extract single phase properties based on nanoindentation mapping? A case study in steels — •Robin Jentner1, Kinshuk Srivastava2, Bastian Philippi2, Christoph Kirchlechner1, and Gerhard Dehm1 — 1Max-Planck-Institut für Eisenforschung GmbH, 40237 Düsseldorf, Germany — 2AG der Dillinger Hüttenwerke, 66763 Dillingen, Germany
Advanced high strength steels exhibit an intricate microstructure comprising of several different phases and interfaces that determines the mechanical behavior of the bulk material. In order to examine the mechanical behavior of each individual phase, micro pillar compression tests have been used. However, micro pillar compression is tedious and time consuming. In this work, we explore the capability of high throughput nanoindentation combined with k-means clustering to determine the mechanical properties of each single phase.
We have trained first the k-means clustering method with a two-phase laboratory sample which was finally applied to a HSLA bainitic steel consisting of three different phases. The clustering revealed two to three clusters in the first case and at least five clusters for the bainitic steel. To unravel the origin of the additional clusters we have performed correlative microscopy and found that indents close to grain or phase boundaries are responsible for them. Based on the obtained results we conclude that analyzing the mechanical properties of complex bulk materials by k-means clustering provide a suitable microstructural characterization method, which will be discussed in the talk.