一方向凝固材のミクロ偏析/組織における介在物分布の評価手法の開発

抄録

The formation of inclusions during solidification in steelmaking process is a critical issue for the optimal processing and the quality of steel products. Therefore, it is required to clarify the mechanism on the inclusion formation for its adequate control. In the present work, the evaluation method of inclusion distribution via the combination of inclusion positions analysis and image analysis of dendrite structure with machine learning is proposed. Image analysis using a conditional deep convolutional generative adversarial network enabled the detection of domain boundaries and the directions of secondary dendrite arms in the cross-sectional structure of unidirectionally solidified specimens. In addition, by combining this with the analysis of inclusion position, a correlation was confirmed between micro segregation behavior and the formation behavior of TiN inclusions.