Page 70 - ESAM-1-1

Engineering Science in
Additive Manufacturing ML in additive manufacturing

Figure 1. Scope of existing reviews on research at the intersection of machine learning (ML) and additive manufacturing (AM), with core ML, AM
lifecycle, AM applications, AM software “SW,” and AM hardware “HW” as the main themes. Core ML themes in AM include data, learning techniques,
and algorithms.

printing” AND “artificial intelligence (AI)” OR “machine are divided into four categories. Applications of AI and
learning” OR “deep learning” OR “data driven”) have been ML support surrogate modeling across the process-
published since 2018 to summarize the applications of ML structure-property linkage. The resulting data-driven
in AM (Appendix A). Figure 2 shows the temporal and predictive models cover the lack of first-principles-driven
geographical distribution of review publications at the domain-based models at each phase of the process chain.
intersection of ML and AM. Exponential increase in such In-process monitoring and quality control have been
reviews highlights the research growth of ML applications a major outcome for the applications of AI and ML in
in the AM field. The most recent reviews have focused AM. These models support process repeatability through
on explainable and physics-informed learning and parameter optimization and defect detection. Applications
31
5,32
applications to enhance fabrication of polymers, metals, of AI and ML in AM have synergy with advanced
34
33
alloys, composites, biomaterials, functionally graded manufacturing themes such as digitization and industry
37
35
36
materials, metamaterials as well as improve technologies 4.0 through enhanced productivity. By eliminating
39
38
such as digital light processing, selective laser sintering, typical AM challenges (e.g., lack of process repeatability),
24
40
robotic arc AM, and MEX. AM technologies also enable productivity can be improved through minimized
42
41
repair and remanufacturing of existing parts through repetitions and destructive evaluations. AM provides
freedoms in material handling and deposition. 41,43,44 potential to enhance sustainable manufacturing practices.
Several review papers have also provided broader focus on As a result, applications of AI and ML can optimize energy
the overall integration of ML in AM. 45 consumption, and minimize failure and waste while
Table 1 highlights some of the most prominent results supporting repair and remanufacturing applications in
from the applications of AI and ML in AM. The results AM thereby improving sustainability.
Volume 1 Issue 1 (2025) 3 doi: 10.36922/ESAM025040004