Smart Monitoring Program for Selective Laser Melting 3D Printing Technology

3D Printing (3DP) technologies are in demand nowadays and their application is rapidly increasing in modern days. 3DP is a state-of-the-art technology that is convenient, and user-friendly for manufacturing components and materials. Recent studies present smart monitoring programs to increase the mobility of the 3DP. These improve the 3DP technology in general, however, it does not resolve faults and errors during printing. To cope with it, we proposed a smart monitoring program and prototyped a diagnosis toolbox for fault detection on selective laser melting 3D printing technology. The values from the monitoring system are utilized in the diagnostic toolbox equipped with a machine learning (ML) paradigm to produce forecast values. The output from the ML scheme is subjected to anomaly detection through threshold classification. To examine the appropriate machine learning scheme used in the diagnostic toolbox, the ML schemes are evaluated with root mean square error (RMSE), mean average error (MAE), and r-square (R
²
). The experiment results showed that the temporal neural network (TCN) outperformed gated recurrent networks and long-short-term memory schemes.