Metal AM processes such as Directed Energy Deposition (DED) can provide significant benefit to challenging applications, but the transition between concept design, prototyping and creating a production component is demanding. In order to create high-quality, repeatable parts, in-process monitoring is utilized to analyze and control the build process. With DED, various monitoring and control modes are available to reduce parameter development times, improve build quality and limit operator input during a build. Among these control modes are melt pool size and temperature, powder flow, laser power, geometric monitoring and acoustic sensing. These control modes not only significantly reduce the process parameter development cycle, but also result in a higher quality build to include density and material properties. In some AM systems, the data from the builds may not be accessible, yet it has proven to be key to understanding and improving builds. Along with developing and integrating in-process sensors and closed-loop control features, data logging can be employed for understanding the complete build history and variability from build to build. With data-logging capability, a file is automatically generated in real time and contains the time history of all parameters and process-monitoring sensors. The data sets provided can be utilized in post-process analyses to enable machine learning and defect-detection algorithms. This presentation will disclose data that was used to correlate defects to anomalous process parameters using two DOD applications as examples, resulting in a data-driven approach to parameter development on the front end and repeatability and quality in production.
- Analyze what types of sensors might be best employed for their application.
- Understand how sensor data can be used to make real-time and post-build decisions.
- Gain a more thorough understand of the challenges and benefits of DED.