About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Fatigue and Fracture
|
| Presentation Title |
Probabilistic Safe Life Modeling of Laser Powder Bed Fusion Additively Manufactured Ti-6Al-4V Using Extreme Value Statistics with Uncertainty
|
| Author(s) |
Justin Miner, Austin Ngo, Brett Ley, Anthony Rollett, John Lewandowski, Sneha Prabha Narra |
| On-Site Speaker (Planned) |
Sneha Prabha Narra |
| Abstract Scope |
Extreme value statistics (EVS) is frequently used in modeling fatigue behavior of additively manufactured (AM) parts by estimating the largest pore sizes that can drive failure. We extend the traditional EVS framework to incorporate model-based uncertainty, enabling a representation of the distribution of critical defect sizes across different regions of a single part. This enhanced EVS model is combined with Ti-6Al-4V fatigue crack growth data, four-point bend fatigue data, and X-ray computed tomography data to probabilistically calibrate safe life models using maximum likelihood estimation. The defect size in the model is further modified to account for stress gradients inherent in bending tests. This approach allows for accurate risk assessment at stress levels with low failure probabilities that would otherwise require extensive testing to characterize. Overall, this work presents a data-driven framework for probabilistic assessment of fatigue strength for porosity-driven fatigue failure in AM parts. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Additive Manufacturing, Modeling and Simulation, Titanium |