Machine Learning Trends in Design for Additive Manufacturing

Additive Manufacturing is becoming a widespread manufacturing system in several industrial fields such as automotive, aerospace, biomedical, etc. Design for Additive Manufacturing represents the branch of research that considers the technological constraints from the early stages of design, arriving at a geometrical model to be exported in G-code. The limitations of additive manufacturing are related to the complexity of the process, the high costs, the processing time, and the difficulties of ensuring adequate geometric and dimensional tolerances. A data-driven approach can be a solution to improve the Design for Additive Manufacturing. Artificial Intelligence and Machine Learning methods are employed in the literature to shorten the time for assessing the optimal combination of parameters and supporting decision-making. The current state of the art shows three macro-areas to apply Machine Learning methods in Design for Additive Manufacturing. These applications concern Geometrical Design Level, Process Configuration Level, and Process Monitoring Level. This paper aims to identify and classify the Machine Learning methods and algorithms most used in Design for Additive Manufacturing practices, analyzing parameters, results, processes, and materials involved.