PhD Internal Presentation - Amir Zarei - Department of Mechanical Engineering-FR

Event date(s): August 27, 2026
Time(s): 01:00 PM - 04:00 PM
Category: Fredericton
Location: Fredericton

Abstract

The lack of consistency in alloy’s microstructure and properties has led to a degree of conservatism in adopting additive manufacturing (AM) in critical sectors including marine, aerospace, and nuclear industries. This highlights the significance of modeling approaches in predicting the behavior of different alloys produced using AM. In this regard, recent promising accomplishments of machine learning (ML) methods in a wide variety of tasks exemplified by computer vision and natural language processing (NLP) have rendered these tools as an alternative for conventional modeling frameworks already used in AM. This work proposes an intelligent algorithm for predicting the melt pool geometry and microstructure characteristics of 316L stainless steels fabricated by laser powder bed fusion (LPBF). To fulfill this goal, a benchmark dataset is introduced, which is experiment-oriented and alloy-specific. This dataset contains the state-of-the-art melt pool geometry reported in the literature as well as some in-house experimental results conducted for this study. The proposed algorithm was validated on this dataset and its performance was compared with the ML frameworks deployed recently in AM. Results demonstrated the superiority of the proposed methodology with acceptable degree of accuracy. This approach enhances the repeatability and consistency of LPBF and facilitates process parameter optimization to develop desired properties.

Building: n/a

Room Number: Microsoft Teams meeting Join: https://teams.microsoft.com/meet/26608630230791?p=xUEkNZVjVrSTAbpCOi