Assessment of Fracture Mechanisms and Temperature of Ductile-Brittle Transition of Metals Obtained on the Basis of 3D Printing Using Neural Network Modeling and Fractal Analysis

Abstract

In this work, a study of the structure, impact toughness and mechanisms of destruction of steels 08Mn2Si and 07Cr25Ni13 was carried out using optical and electron microscopy. It was found that with a decrease in temperature, the level of impact toughness of the studied steels significantly decreases, so at a temperature of t = -100°C, the impact toughness of 08Mn2Si and 07Cr25Ni13 steel decreases by 83% and 45%, respectively, relative to the temperature t = +20°C. For steel 07Cr25Ni13, in the entire considered range, only the tough component is present in the fracture and there is no temperature of ductile-brittle transition, whereas for steel 08Mn2Si, already at temperatures below t = -41°C (ductile-brittle transition of steel 08Mn2Si), the brittle component begins to dominate and significantly the likelihood of its fragile destruction increases. On the basis of the obtained electronic photographs of the kinks, an artificial neural network (ANN) for the classification of kinks was trained and tested for intermediate test temperatures. When recognizing the viscous component in the fractures of 08Mn2Si steel obtained using the 3D-printing technology by electric arc surfacing, the recognition error does not exceed 7% with the use of ANN.