Analysis of acoustic emission signals at austempering of steels using neural networks

• Published in: Metals and materials international 2017, 23(3), , pp.426-433
• Main Authors: Łazarska, Malgorzata, Wozniak, Tadeusz Z., Ranachowski, Zbigniew, Trafarski, Andrzej, Domek, Grzegorz
• Format: Article
• Language: English
• Subjects: Austempering; Bearing steels; Characterization and Evaluation of Materials; Chemistry and Materials Science; Engineering Thermodynamics; Heat and Mass Transfer; Machines; Magnetic Materials; Magnetism; Manufacturing; Martensitic transformations; Materials Science; Metallic Materials; Neural networks; Processes; Solid Mechanics; Tool steels; 재료공학
• ISSN: 1598-9623; 2005-4149
• DOI: 10.1007/s12540-017-6347-z

Bearing steel 100CrMnSi6-4 and tool steel C105U were used to carry out this research with the steels being austempered to obtain a martensitic-bainitic structure. During the process quite a large number of acoustic emissions (AE) were observed. These signals were then analysed using neural networks resulting in the identification of three groups of events of: high, medium and low energy and in addition their spectral characteristics were plotted. The results were presented in the form of diagrams of AE incidence as a function of time. It was demonstrated that complex transformations of austenite into martensite and bainite occurred when austempering bearing steel at 160 °C and tool steel at 130 °C respectively. The selected temperatures of isothermal quenching of the tested steels were within the area near to MS temperature, which affected the complex course of phase transition. The high activity of AE is a typical occurrence for martensitic transformation and this is the transformation mechanism that induces the generation of AE signals of higher energy in the first stage of transition. In the second stage of transformation, the initially nucleated martensite accelerates the occurrence of the next bainitic transformation.