About this Abstract |
Meeting |
MS&T22: Materials Science & Technology
|
Symposium
|
Energy Materials for Sustainable Development
|
Presentation Title |
The Development of a Machine Learning Guided Process for the Additive Manufacturing of Thermoelectric Materials |
Author(s) |
Connor Headley, Roberto J. Herrera del Valle, Ji Ma, Prasanna Balachandran, Vijayabarathi Ponnambalam, Saniya LeBlanc, Dylan Kirsch, Joshua Martin |
On-Site Speaker (Planned) |
Connor Headley |
Abstract Scope |
The implementation of additive manufacturing promises to create thermoelectric devices with increased efficiency and lowered production costs. However, the optimal additive manufacturing processing parameters for any thermoelectric material are currently unknown, and the development of an additive manufacturing process for a new material is traditionally an arduous task that requires numerous rounds of experimental trial-and-error. Through the integration of machine learning techniques alongside well-curated additive manufacturing experimentation, we quickly draw vital connections between processing parameters, melt pool geometries, and defects while significantly reducing experimental burden. We rapidly developed process parameters for laser powder bed fusion that created highly dense, geometrically complex bismuth telluride parts through additive manufacturing. A system of high throughput sample fabrication and characterization was devised to determine the relationship between processing conditions and resulting thermoelectric properties. These connections have allowed us to intentionally vary the character of these samples from n-type to p-type through processing parameter modifications. |