Single crystal elastic constants of additively manufactured components determined by resonant ultrasound spectroscopy

Abstract

Bayesian inference with Sequential Monte Carlo was used to determine the single crystal elastic constants of additively manufactured (AM) cobalt‑nickel-based superalloy specimens from only the resonant frequencies and texture data. This novel framework enables the quantification of the single crystal elastic constants for AM and polycrystalline specimens using only electron backscatter diffraction (EBSD) and Resonant Ultrasound Spectroscopy (RUS), avoiding the expense of bulk single crystal fabrication or synchrotron experiments. A parallelizable and open-source Python package (SMCPy) was used to perform Bayesian inference of the single crystal elastic constants from resonant frequencies of AM specimens. The single crystal elastic constants determined from AM cobalt‑nickel-base superalloy specimens were validated with measurements of the single crystal elastic constants on a bulk single crystal specimen. EBSD texture data was used to determine the single crystal elastic constants from the resonant frequencies of AM specimens, and validated with neutron diffraction data by considering the experimental uncertainty in both the EBSD and neutron diffraction data. The robustness of this framework for varied texture orientations relative to the build direction (BD) was demonstrated for AM specimens printed at 0° and 20° BD-inclinations. • Single crystal elastic constants are determinable with RUS/ EBSD of AM specimens. • Textures both aligned and misaligned to BD are demonstrated. • Single crystal constant variability driven by error in input texture coefficients. • Parallelized SMC enables 10× computation time reduction.