Abstract
Advancing the understanding of material behavior and phenomena related to size effects in small-scale components is critical for settings where limited quantities of material samples can be tested. Established guidelines for sub-sized specimen testing encompass best practices for specimen preparation, testing equipment, test procedures, and data analysis methods. However, prior investigations of specimen size effects in the literature typically involved a relatively small number of tests performed and analyzed. To address this limitation, our team created a large database of 1,050 tensile test records for nuclear structural materials collected from peer-reviewed articles. In this study, we introduced a machine learning-based approach for predicting the tensile properties of sub-sized specimens, and we developed methods for uncertainty quantification of predicted properties. Furthermore, we conducted an experimental validation of the reported critical values for the dimensions and geometry of sub-sized specimens, and we validated existing analytical models for correlating total elongation between sub-sized and standard-sized specimens. Our findings demonstrate the potential of machine learning techniques to enhance understanding of specimen size-dependent material behavior and highlight the need for coordinated efforts in developing large, open-source databases of mechanical testing data to support future research.