Application of machine learning techniques to control surface quality of as-printed wire arc additive manufactured components

Nowadays, the wire arc additive manufacturing is making its path toward providing benefits to aerospace, defense, and oil and gas sectors, ascribed to the process capacity to fabricate components with minimum waste of material and lead time. However, the main challenges associated with the WAAM that have hindered the wide-spread application of the technology include the irregular and random quality of the WAAM fabricated surfaces. The mission of this project is to control aforementioned irregularities in fabrication by implementing machine learning-based algorithms and modify the process parameters to achieve a defect-free part with high surface quality. The anticipated trained machine learning method in this project will foster the progress toward completion of an autonomous in-situ defect recognition and correction (AIDRAC) system, which is the primary goal of the intern.

Faculty Supervisor:

Ali Nasiri

Student:

Salar Salahi

Partner:

Springboard Atlantic

Discipline:

Engineering - mechanical

Sector:

Professional, scientific and technical services

University:

Memorial University of Newfoundland