Machine Learning in Additive Manufacturing: A Review of Learning Techniques and Tasks

Date

2022

Authors

Pike, J.A.
Klett, J.
Kunc, V.
Duty, C.E.

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Due to recent advances, Machine Learning (ML) has gained attention in the Additive Manufacturing (AM) community as a new way to improve parts and processes. The capability of ML to produce insights from large amounts of data by solving tasks such as classification, regression, and clustering provide possibilities to impact every step of the AM process. In the design phase, ML can optimize part design with respect to geometry, material selection, and part count. Prior to printing, process simulations can offer understanding into the how the part will be printed, and energy, time, and cost estimates of a print can be made to assist with resource planning. During printing, AM can benefit from in-situ printing optimization and quality monitoring. Lastly, ML can characterize printed parts from in-situ or ex-situ data. This article describes some of the ML learning techniques and tasks commonly employed in AM and provides examples of their use in previous works.

Description

LCSH Subject Headings

Citation