Vol.11, No1, 2011, pp. 9-14
UDC 620.178.2 : 669.15-194.2

PREDICTION OF FRACTURE TOUGHNESS TEMPERATURE DEPENDENCE APPLYING NEURAL NETWORK

I. Dlouhý1,2, H. Hadraba1, Z. Chlup2, T. Šmida3

1) Institute of Physics of Materials, Academy of Sciences of the Czech Republic, Brno,

2) Institute of Material Science and Engng, Faculty of Mech. Engng., University of Technology, Brno

3) IBOK, a.s., Bratislava, Slovak Republic

Abstract

Reference temperature localizing the fracture toughness temperature diagram on temperature axis is predicted based on tensile test data. The regularisation neural network is developed to solve the correlation of these properties. Three-point bend specimens are applied to determine fracture toughness. The fracture toughness transition dependence is quantified by means of the master curve concept enabling to represent it by using one parameter, i.e. the reference temperature. Tensile samples with circumferential notch are also examined. In total 29 data sets from low-alloy steels are applied for the analysis. A good correlation of the predicted and experimentally determined values of the reference temperature is found.

Keywords: brittle to ductile transition, fracture toughness, artificial neural network, steels

full article (295 kB)