| About this Abstract |
| Meeting |
2021 TMS Annual Meeting & Exhibition
|
| Symposium
|
Recycling and Sustainability for Emerging Technologies and Strategic Materials
|
| Presentation Title |
Uncertainty Analysis and Reduction for Environmental Impact Modeling of Emerging Manufacturing Technologies |
| Author(s) |
Jiankan Liao, Daniel Ross Cooper |
| On-Site Speaker (Planned) |
Jiankan Liao |
| Abstract Scope |
Reliable environmental impact models of emerging manufacturing technology are needed in order to compare new and traditional processes and to identify research priorities. We use a data-driven approach to tailor generalized mechanistic process impact models to a specific machine with the model uncertainties quantified and then reduced using Bayesian inference and the principles of optimal experimental design. We demonstrate the approach by modeling the primary energy requirements in metal laser powder bed fusion. Forward error propagation is used to quantify the overall uncertainty in the results before a Sobol indices analysis is used to reduce the dimensionality of the problem. Bayesian inference is used to reduce the uncertainty in unobservable parameters (e.g., the adiabatic efficiency) using experimental build time data from a specific machine; thus, reducing the overall uncertainty. The updated model is used to guide future research priorities for reducing environmental impacts. |
| Proceedings Inclusion? |
Planned: |