Abstract: In this study, novel reconstruction methods, including grain graph and variant graph, were established to reconstruct parent austenite on the basis of electron backscatter diffraction (EBSD) data. The evaluation indicators included boundary identification and variant distribution. Moreover, an innovative variant pair analysis method was proposed. The results indicated that the Kurdjumov-Sachs orientation relationship was the most appropriate because it had the smallest refinement error and deviation. In addition, the variant graph reconstruction was more effective in reducing mis-indexing areas than the grain graph, exhibiting a robust capacity to accurately identify austenite grain boundaries. Additionally, the variant graph reconstruction induced the transformation of variants, variant pairs, close-packed plane (CP) groups, and Bain groups. Moreover, various reconstructed datasets (calc-grain data and EBSD data) affected the distribution of variants. The austenite grains reconstructed from the calc-grain data featured two or more variants clustered within the same region due to the preprocessing (calculating, filtering, and smoothing) of the EBSD data. These variations did not impede the microstructural analysis when consistent original data and reconstruction methods were used. The reconstruction of parent austenite grains holds promise for providing a fresh perspective and a deeper understanding of strengthening and toughening mechanisms in the future.

Key words: Martensitic steel, Parent austenite grain, Reconstruction, Electron backscattering diffraction, Variant