Identifications of various defect types during a fused deposition modeling process based on deep learning technology

This project is to purpose to use computer vision to identify the various error types during the operation of 3D printers to boost their throughput and enhance their application in the manufacturing industry. Nevertheless, due to the lack of precision and controllability inside the printers, engineers cannot achieve a reliable printing process and acceptable quality of final 3D printing products. In this research, a monitoring system equipped with computer vision is proposed to address this challenge. The computer vision program can monitor the printing process and easily detect the error type and inform the operators immediately when a failure occurs. As a result, this program will significantly reduce waste and improve productivity for 3D printers.

Faculty Supervisor:

Yu Zou

Student:

Xingchen Liu

Partner:

Mech Solutions Ltd

Discipline:

Engineering - mechanical

Sector:

Manufacturing

University:

University of Toronto