About this Abstract |
| Meeting |
2020 TMS Annual Meeting & Exhibition
|
| Symposium
|
Additive Manufacturing Fatigue and Fracture IV: Toward Confident Use in Critical Applications
|
| Presentation Title |
Additive Manufacturing-enhanced Durability Prediction Supported by a Machine-learning Based Material Model |
| Author(s) |
Nicolas Lammens, Matthias Schulz, Stefan Straesser, Hunor Erdelyi |
| On-Site Speaker (Planned) |
Nicolas Lammens |
| Abstract Scope |
Predictable fatigue performance of AM components is a significant challenge because of the large range of printing parameters selectable and the occurrence of artefacts such as key-hole and lack of fusion pores which cannot always be resolved.
A solution is the use of durability software accounting for these typical artefacts. Essential in such a simulation, is the availability of a material model that can account for the impact of multiple AM artefacts. Such a material model is difficult to develop given the many complex interactions between parameters, and the cost of test campaigns.
As part of the Flemish FATAM project, an AM-enhanced durability prediction software was developed supported by a machine-learning based material model. The flexibility and limited a-priori assumptions offered by the machine learning allows the material model to adapt as more test data becomes available and detect cross-sensitivities which would be hard to predict using conventional material modelling. |
| Proceedings Inclusion? |
Planned: Supplemental Proceedings volume |