Neural networks model for prediction of the hardness of steels cooled from the austenitizing temperature

Journal Title: Archives of Materials Science and Engineering - Year 2016, Vol 82, Issue 2

Abstract

Purpose: The paper presents the new neural networks model making it possible to estimate the hardness of continuously-cooled steel from the austenitizing temperature.Design/methodology/approach: The method proposed in the paper employs two applications of the neural networks of: classification and regression. Classification and consists in determining the value of dichotomous variables describing the occurrence of: ferrite, pearlite, bainite and martensite in the microstructure of a steel. The values of dichotomous variables have been used for calculating steel hardness. The other task is regression, which aims at calculating the steel hardness.Findings: The presented neural networks model can be used only in the range of concentrations of alloying elements shown in this paper.Practical implications: The model worked out makes it possible to calculate hardness for the steel with a known chemical composition. This model deliver important information for the rational selection of steel for those parts of the machines that are subjected to the heat treatment. The presented model make it possible the analysis of the interaction of the chemical composition on the hardness curves of the steel cooled from the austenitizing temperature.Originality/value: The paper presents the method for calculating hardness of the structural and engineering steels, depending on their chemical composition, austenitizing temperature and cooling rate.

Authors and Affiliations

J. Trzaska

Keywords

Related Articles

Warm stretch-formability of 0.2%C- 1.5%Si-(1.5-5.0)%Mn TRIP-aided steels

Purpose: Warm stretch-formability of 0.2%C-1.5%Si-(1.5-5.0)%Mn transformation-induced plasticity (TRIP)-aided sheet steels with annealed martensite matrix was investigated for automotive applications. Additio...

Chromous capacities of ferrochromium and matte smelting slags

Purpose: of this paper is to extend the slag capacity concept to sparingly soluble cationic species encountered for example in matte smelting and ferrochromium processing in terms of chromous capacities. Design/methodolo...

Chasing shadows: technology and socioeconomic barriers versus climate targets for iron and steel industry

Purpose: To analyse the potential of various scenarios for reduction of carbon footprint of iron and steel sector and to reveal plausible pathways for modernisation. Design/methodology/approach: Several scenarios hav...

CAD/CAM silicone auricular prosthesis with thermoformed stiffening insert

Purpose: Epitheses (facial prostheses) for large facial tissue defects manufacturedfrom silicones exhibit unsatisfactory rigidity and its stiffening is required, which createstechnological problems. Moreover, facial epit...

Mathematical model for determining the expenditure of cooling and lubricating fluid reaching directly the grinding zone

Purpose: The purpose of this article is to discuss the method of determining the mathematical model used for calculating the amount of emulsion reaching directly the grinding zone during the hob...

Download PDF file
  • EP ID EP216249
  • DOI 10.5604/01.3001.0009.7105
  • Views 74
  • Downloads 0

How To Cite

J. Trzaska (2016). Neural networks model for prediction of the hardness of steels cooled from the austenitizing temperature. Archives of Materials Science and Engineering, 82(2), 62-69. https://europub.co.uk/articles/-A-216249