부산대학교 도서관

Additive Manufacturing (AM) for Industry 4.0 requires a number of networking, integrated control and cloud technologies to enhance the connectivity and performance. There is now, however, a dearth of widely available solutions for Cloud-based AM. Regrettably, the repeatability and monitoring of quality in the production process are not sufficiently dependable to be used in mass production. Thus, quality monitoring can be used as an important tool in AM for defect detection to minimize material and time waste during printing. Therefore, this study is aimed to provide a cyber physical system-based AM framework of evaluating the reliability of the automatically printed components by including sensor to capture images and machine learning approach in an industry 4.0 environment. Images of semi-finished parts are taken while the extruder's vibration goes into the above threshold level vibration. The proposed system incorporates an accelerometer and camera module connected with raspberry pi attached with 3D Printer. Azure machine learning studio, connected to the Azure IoT hub, where a machine learning method, convent, is proposed to classify the parts into the ‘good’ or ‘defective’ category. Thus, experimental runs are reduced in Fused Deposition Modeling (FDM) AM printing through parametric optimization using Taguchi Design studies approach. Finally, the developed model has been validated using variance analysis (ANOVA) and signal-to-noise ratio (S/N), in an intelligent environment with Industry 4.0 for defect-free production. This Framework have the potential to successfully implementation of cloud based closed loop quality monitoring system for FDM based Additive manufacturing process in industry 4.0 environment.