High carbon steel is prone to macroscopic and semimacroscopic segregation. Segregation is a phenomenon that results in the nonuniformity of solute element distribution during solidification; segregation severely impacts the quality of the casting billet. Based on qualitative analysis (macroscopic quality rating) and relatively simple quantitative analysis (segregation index, mean square error), it is found that the accuracy of existing technologies is limited for determining the nonuniformity of carbon element distribution in large areas of cast high carbon steel billets; this causes difficulties while applying such existing technologies for the evaluation of large samples under actual steelworks usage conditions. Therefore, it is essential to study the nonuniformity of carbon element distribution to precisely evaluate and optimize the quality of high carbon steel. Based on the grayscale image of the casting blank macro-structure of typical high carbon steels (mass fraction of carbon = 0.7%), the standard deviation, differential box-counting, and moment of inertia are introduced to discuss simple methods for the quantitative characterization of the nonuniformity of the major segregation element (C) in large areas. Then, the validity of the characterization results was verified using statistical homogeneity and the near-equilibrium solidification model, and the similarities and differences of the three characterization methods were discussed. The results showed that the standard deviation, differential box-counting, and moment of inertia can effectively reflect the nonuniformity of carbon element distribution. Furthermore, the mean value of the nonuniformity of carbon element distribution in the equiaxed area was 20.85% higher than that in the columnar area. The standard deviation is mainly based on the statistical characteristics of the grayscale value, while differential box-counting and moment of inertia combine the grayscale value statistical and spatial distribution information on a grayscale image of the casting blank macrostructure. In addition, differential box-counting has scale independence, which is less affected by the size and resolution of the grayscale image; the moment of inertia is sensitive to the variation of the nonuniformity of carbon element distribution in the microregion. Finally, the standard deviation is mainly influenced by large segregation points (e.g., area larger than 1 mm(2)), while differential box-counting and the moment of inertia are mainly influenced by medium segregation points (e. g., area of 0.1-1 mm(2)). Therefore, the three characterization methods can be combined to comprehensively evaluate the nonuniformity of carbon element distribution in large areas of cast high carbon steel billets. This research can provide a new theoretical reference for comprehensively evaluating the nonuniformity of carbon element distribution in a large area and for the fine quality evaluation of cast high carbon steel billets.
