Machine learning based prediction of mechanical properties of WC-Co cemented carbides from magnetic data only

• DOI: 10.1016/j.ijrmhm.2024.106665
• Publication Date: 2024-03-21T08:28:35Z
• Authors: Hubert Brueckl; Leoni Breth; Johann Fischbacher; Thomas Schrefl; Saskia Kuehrer; Julia Pachlhofer; Maria Schwarz; Thomas Weirather; Christoph Czettl
• Keywords: Cemented carbide

Based on experimental data and extensive experience, magnetic coercivity saturation moment are traditionally used to estimate the microstructure quality of cemented carbides, especially in manufacturing industry. This work demonstrates that predictions structural mechanical properties manufactured WC-Co elements can be derived principle from alone using an artificial neural network (ANN). A collection pellet samples with a wide variety powder compositions processing parameters was produced cover range characteristic features for ANN training. The total field distribution, extracted first-order-reversal-curves, serves as input ANN. Microstructural such mean grain size hardness fracture toughness purely measurements high accuracy, while transverse rupture strength shows large errors cannot predicted.