Issue 78

A. A. Frantsuzov et alii, Fracture and Structural Integrity, 78 (2026) 474-488; DOI: 10.3221/IGF-ESIS.78.27

in etching response, and microstructural banding associated with chemical segregation in rolled pipe material. As a result, some boundaries are discontinuous or poorly contrasted, which directly affects the reliability of manual measurements. When the boundary network is incomplete, the operator tends to select grains with clearly visible contours, which are often larger and easier to identify. This introduces a systematic bias toward overestimation of PAGS. The lineal intercept method partially reduces operator subjectivity because it is based on stereological principles and averages a larger number of boundary intersections. Nevertheless, its reliability still depends on the quality of the revealed boundary network. In its classical form, the method provides mainly an average grain size and does not fully describe the grain-size distribution. Although modern image-analysis software can extract individual chord lengths, the method still inherits the limitations of chemical etching when PAG boundaries are only partially revealed. EBSD-based parent austenite reconstruction provides an alternative approach that does not depend on chemical boundary contrast. Reconstruction algorithms use crystallographic orientation relationships between martensitic variants to restore the parent austenite grain structure. A key advantage of this approach is that each reconstructed grain can be treated as an individual object, allowing additional parameters to be extracted, including grain area, equivalent diameter, shape descriptors, and crystallographic orientation. This makes it possible to obtain not only an average PAGS, but also full grain-size distributions and statistical descriptors of scatter. At the same time, EBSD reconstruction is not free from systematic errors. The results depend on EBSD indexing quality, step size, reconstruction thresholds, and possible local deviations from the assumed orientation relationship. These factors may lead either to artificial subdivision of grains into smaller fragments or to excessive merging of neighboring parent grains. The present results show that EBSD reconstruction at ×500 and ×1000 magnifications give similar PAGS distributions and good spatial agreement between reconstructed PAG maps. This agreement validates the use of the lower magnification for this material. Since ×500 mapping covers a larger area in a shorter acquisition time, it allows a higher number of PAGs to be extracted from a single EBSD map and therefore improves the statistical representativeness of the measurement. In contrast, increasing the magnification to ×3000 does not necessarily improve the reconstruction quality. Although the spatial resolution is higher, the smaller mapped area reduces the number of captured grains, and the reconstruction may subdivide some larger PAGs into smaller fragments. Thus, for this steel, lower-magnification EBSD appears to provide a more practical balance between spatial resolution, acquisition time, and statistical reliability. A direct comparison with true in situ austenite observations reported in the literature shows that all indirect methods may overestimate the average PAG size, even when the same region is carefully aligned before and after transformation. However, the magnitude of this bias differs between methods. EBSD-based reconstruction may be closer to the true austenite structure, whereas picric acid etching can significantly overestimate grain size because of weak twin-boundary detection and local boundary loss in non-uniformly etched regions [19]. This literature trend is consistent with the present observations: when the etched boundary network is incomplete, manual measurements preferentially include more visible and often larger grains, while smaller or weakly revealed grains are underrepresented. he prior austenite grain structure of the investigated C110-grade pipeline steel could be revealed by picric acid etching, but the boundary contrast is not uniform. Fine grain size, weakly revealed boundaries, and chemically banded regions limit the reliability of quantitative manual PAGS measurements. Manual PAGS evaluation showed method- and operator-dependent scatter. The manual diagonal method gave an average grain diameter of approximately 7.2 μ m, while the lineal intercept method gave 6.1 – 8.3 μ m depending on the operator. Standard plate comparison gave a coarse value of approximately 11.2 μ m, reflecting the limited resolution of this approach for fine-grained steels. EBSD-based parent austenite reconstruction provided a more detailed and statistically representative description of the PAG structure. Unlike manual optical methods, it enabled extraction of wider data distributions. EBSD reconstructions obtained at ×500 and ×1000 magnifications showed good agreement in both grain-size distribution and spatial correspondence of reconstructed PAG boundaries. This validates the use of ×500 magnification for quantitative PAGS evaluation in the investigated material. Increasing the magnification did not necessarily improve reconstruction reliability. The smaller field of view reduced statistical representativeness and promoted a shift toward smaller reconstructed grains, partly due to artificial subdivision of larger PAGs. Current research with chosen magnifications and reconstruction parameters should not be applied universally for different microstructures for discussed pipe grade materials. An individual approach to a material state is recommended. However, T C ONCLUSION

486