About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Verification, Calibration, and Validation Approaches in Modeling the Mechanical Performance of Metallic Materials
|
| Presentation Title |
A Bayesian Framework for the Calibration of Constitutive Models Across Yield and Creep Regimes Across Alloy Classes |
| Author(s) |
Anjana Anu Talapatra, Jobin K Joy, Andrea Rovinelli, Laurent Capolungo |
| On-Site Speaker (Planned) |
Anjana Anu Talapatra |
| Abstract Scope |
The development of microstructure-informed constitutive models is essential for designing and optimizing engineering alloys, but their high-dimensional parameter spaces make manual calibration impractical. We introduce an automated workflow for calibrating a temperature-dependent, high-fidelity crystal plasticity model using high-throughput sampling, global sensitivity analysis, surrogate modeling, and Bayesian inference. Sensitivity analysis identifies key parameters for yield and creep, reducing problem dimensionality. Heteroscedastic Gaussian process surrogates emulate model responses and are embedded in a Markov chain Monte Carlo sampler for Bayesian inversion using experimental data upto 800C. Applied to Grade 91 (ferritic-martensitic) and 347H (austenitic) steels, the workflow yields robust posterior distributions for yield and creep across stress–temperature conditions. For Grade 91, it achieves accuracy within ±5% for yield stress and within an order of magnitude for creep—while reducing calibration time from weeks to hours. The approach yields reproducible, uncertainty-aware calibrations and provides a transferable template for other complex alloy systems. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Modeling and Simulation, Iron and Steel, |