|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||General Poster Session
||O-7: Bayesian Inference Based Uncertainty Quantification and Propagation Analysis of a Polycrystal Plasticity Finite Element Model Used for High Cycle Fatigue Analysis of Ti-6Al-4V
||Ritwik Bandyopadhyay, Kartik Kapoor, Barron J. Bichon, Michael D. Sangid
|On-Site Speaker (Planned)
Polycrystal plasticity (CP) models, integrated within finite element framework, are used for the microstructure based predictive modelling in Integrated Computational Materials Engineering. Such models are complex, non-linear, and often involve different length scales. Therefore, it is essential to characterize these models based on uncertainties associated with the predicted quantities (for instance stress and strain). Here, a Bayesian inference method based uncertainty quantification and propagation analysis is performed on a CP model, which is used for the high cycle fatigue analysis of Ti-6Al-4V alloy. First, global sensitivity analysis is performed to find the set of the most influential parameters. Next, Bayesian inference method along with Markov chain Monte Carlo algorithm is used to obtain the posterior distributions of the most influential parameters. Finally, quantified uncertainties are sampled based on Monte Carlo method and propagated through the model to obtain the uncertainty associated with the predicted value of the macroscopic stress.
||Planned: Supplemental Proceedings volume