A Study on Monitoring and Fault Diagnosis of Fused Deposition Modeling Process Based on Data-driven Approach

Abstract

Additive manufacturing (AM) is a layer by layer manufacturing process that can fabricate a threedimensional part directly from a computer aided design (CAD) model. In particular, a fused deposition modeling (FDM) process is the most widely used AM technique for fabrication of thermoplastic parts. They can be used for making functional prototypes with advantages of low cost, minimal wastage and ease of material change. Despite its recent popularity, FDM still faces many technical challenges for insufficiency of process reliability and controllability and product quality. Therefore, to overcome such disadvantages, the monitoring and fault diagnosis on FDM process is of much significance. In this study, the monitoring on quality of parts and components fabricated by the FDM process is conducted by analyzing mass data which are obtained from various sensors such as accelerometers, acoustic emission sensors and temperature sensors. After extracting critical features from the measured process signals, they are related with quality evaluation indices in the model by using data-driven modeling techniques. Finally, the developed model is validated via a series of experiments.