About this Abstract |
| Meeting |
MS&T26: Materials Science & Technology
|
| Symposium
|
Computational Materials for Qualification and Certification
|
| Presentation Title |
Probabilistic Fatigue Modeling of Powder Bed Fusion – Laser Beam Ti-6Al-4V with Model- and Measurement-Based Uncertainty |
| Author(s) |
Justin Patrick Miner, Pavel Shevchenko, Viktor Nikitin, Sneha Prabha Narra |
| On-Site Speaker (Planned) |
Sneha Prabha Narra |
| Abstract Scope |
Process-induced porosity remains a barrier to adopting Powder Bed Fusion–Laser Beam (PBF-LB) components in fracture-critical applications. X-ray micro computed tomography (X-μCT) can characterize porosity, but its reliability is limited by measurement uncertainty. In this work, we quantify X-μCT porosity measurement uncertainty using high-resolution synchrotron X-μCT. A Ti-6Al-4V specimen with intentional porosity was fabricated to systematically evaluate measurement uncertainty across the varying degrees of porosity. Results show that detectability and bias can be modeled as functions of pore size, shape, and location, for constant imaging conditions and segmentation methods. These models outperform approaches in nondestructive evaluation standards. We then apply these uncertainty models to the AMBench fatigue dataset to estimate the true pore distribution. Using extreme value statistics with model uncertainty, we assess pore criticality and propagate it through an empirical fatigue model. This work demonstrates that incorporating measurement uncertainty enables more robust fatigue strength estimations at a given risk level. |