Date of Award
Fall 2023
Document Type
Open Access Dissertation
Department
Mechanical Engineering
First Advisor
Austin Downey
Abstract
Fused filament fabrication (FFF) is one of the fastest-growing, most promising, and widely-used Additive manufacturing (AM) technologies, as its capabilities of largely reducing material waste, feasible and flexible designing complicated structure or geometry product. Except for the advantages mentioned above, the unique affordability has made FFF a popular technique for small-scale 3D printing. However, its widescale adaption for industrial applications still faces significant challenges, especially for Industry 4.0, due to the process’s intrinsic uncertainties. During the printing process, even a trivial bonding gap or material void dramatically influences the final product’s structural quality and mechanical properties. Therefore, a defect detection system for the FFF printing process to reduce part-to-part variations and assure product mechanical properties is essential. Extensive defect detection systems utilizing machine learning (ML) for the FFF printing process have been built to detect various faults, including bonding quality problems and infill defects. However, the ultimate goal for defect detection is to guarantee product quality rather than find the defect. In addition, defects’ effects on functional product quality are often more critical than easy-to-spot surface defects. Therefore, effective online defect detection integrated with product structure quality validation for FFF is needed. The proposed system should not only detect visible defects but also infer the defect’s impact on functional qualities caused by variations in the printing process. In the project, a novel real-time ML-based defect detection approach is proposed that is capable of inferring structural quality. The proposed online real-time structural validation approach for FFF can provide a reliable and efficient way to guarantee product quality during manufacturing. The project’s future work will advance the state-of-the-art in real-time model updating while providing advancements in the fields of real-time, data science, and Industry 4.0.
Rights
© 2024, Yanzhou Fu
Recommended Citation
Fu, Y.(2023). Real-Time Product Structural Validation for Fused Filament Fabrication. (Doctoral dissertation). Retrieved from https://scholarcommons.sc.edu/etd/7612