About this Abstract |
| Meeting |
Materials in Nuclear Energy Systems (MiNES) 2025
|
| Symposium
|
Materials in Nuclear Energy Systems (MiNES) 2025
|
| Presentation Title |
A Statistically Robust Area Fraction Measurement Tool for Digital Micrographs |
| Author(s) |
Mitchell L. Mika, Mary Sevart, Paul McIntyre, Assel Aitkaliyeva |
| On-Site Speaker (Planned) |
Mitchell L. Mika |
| Abstract Scope |
Quantifying the volume fraction of a constituent within a sample is often a key step in nuclear materials characterization, such as the need for quantifying the fission gas porosity in an irradiated fuel sample. A common way of quantifying volume fraction is to approximate it using area fraction measurements, which could be done manually via point counting or using automated image segmentation algorithms. While image segmentation algorithms greatly increase the volume of samples that can be analyzed compared to manual point counting, their sensitivity to imaging conditions and artifacts requires some level of validation against manual methods to ensure measurements are accurate. Little work has focused on these manual methods. Common methods such as adaptations of ASTM E562 are sensitive to periodicity, gradients, and heterogeneity within the micrograph. Manual segmentation of the image avoids these issues but can be extremely time intensive and prone to user errors. In this work, we present an efficient point-counting algorithm for quantifying the area fraction of a constituent based on statistical sampling principles. This algorithm allows users to specify a confidence interval and desired margin of error of their measurement, providing a more complete understanding of the measurement uncertainty. The algorithm is then implemented into an open-source program to facilitate the measurement process. This presentation will also highlight how these 2D area fraction measurements relate to the 3D volume fractions. |
| Proceedings Inclusion? |
Undecided |