Additive manufacturing processes are prone to production errors. Specifically, the unique physical conditions of Laser-Directed Energy Deposition (DED-L) lead to unexpected process anomalies resulting in subpar part quality. The resulting costs and lack of reproducibility are two major barriers hindering a broader adoption of this innovative technology. Combining sensor data with data from relevant steps before and after the production process can lead to an increased understanding of when and why these process anomalies occur. In the present study, an IoT-based data mining framework is presented to assess the stability of processing Ti6Al4V on an industrial-grade DED-L machine. The framework employs an edge-cloud computing methodology to collect data efficiently and securely from various steps in the part lifecycle. During manufacturing, multiple sensors are employed to monitor the essential process characteristics in situ. Mechanical properties of the 160 printed specimens were obtained using appropriate destructive testing. All data are stored on a central database and can be accessed via the web for data analytics. The results prove the successful implementation of the proposed IoT framework but also indicate a lack of process stability during manufacturing. The occurring part errors can only be partially correlated with anomalies in the in situ sensor data.     «

Additive manufacturing processes are prone to production errors. Specifically, the unique physical conditions of Laser-Directed Energy Deposition (DED-L) lead to unexpected process anomalies resulting in subpar part quality. The resulting costs and lack of reproducibility are two major barriers hindering a broader adoption of this innovative technology. Combining sensor data with data from relevant steps before and after the production process can lead to an increased understanding of when and w...     »