|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Integrative Materials Design III: Performance and Sustainability
||Data Science and Machine Learning Opportunities in Additive Manufacturing
||Elizabeth Holm, Brian DeCost, Anna Smith, Andrew Kitahara
|On-Site Speaker (Planned)
Additive manufacturing (AM) promises to rewrite manufacturing workflows, offering an unprecedented ability to build complex, low-volume parts economically. However, this advantage is lost if a large number of additional parts must be produced for qualification purposes. One tactic for overcoming this challenge is to apply data science and machine learning to determine successful process parameters using data from many different builds. In this talk, we will focus on data science approaches to materials science aspects of the AM process. For example, computer vision and machine learning can characterize the complex and multifaceted properties of feedstock powders that directly impact the success of powderbed AM builds. Similar techniques can identify both normal and anomalous build microstructures and associate them with process parameters for feedback control loops. Together, these methods promise to capitalize on the unique data streams intrinsic to the AM process.
||Planned: Supplemental Proceedings volume