| About this Abstract |
| Meeting |
Additive Manufacturing Benchmarks 2022 (AM-Bench 2022)
|
| Symposium
|
Additive Manufacturing Benchmarks 2022 (AM-Bench 2022)
|
| Presentation Title |
Fast and Effective Sensitivity and Uncertainty Quantification for Metal-based Additive Manufacturing |
| Author(s) |
David Restrepo, Juan Sebastian Rincon Tabares, Matthew Balcer, Mauricio Aristizabal , Arturo Montoya, Harry Millwater |
| On-Site Speaker (Planned) |
David Restrepo |
| Abstract Scope |
Two novel technologies are presented in support of Qualification and Certification process for AM. The first technology corresponds to the hypercomplex finite element method that allows one to calculate highly-accurate arbitrary-order sensitivities of thermal and material responses with respect to initial conditions, loadings, material properties, or shape. This technology allows one to quantify the relative importance of the build parameters. Moreover, these sensitivities provide the basis for a fast uncertainty quantification (UQ) method that uses a Taylor series expansion to approximate the probability of distributions of any output. As a result, one can utilize all the full fidelity of a finite element model and obtain UQ information without the need on relying on Monte Carlo sampling of simplified or surrogate models, which limits accuracy, or relying on polynomial chaos or design of experiment approaches that limit the number of parameters considered. |
| Proceedings Inclusion? |
Undecided |