About this Abstract |
Meeting |
2022 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2022)
|
Symposium
|
2022 Annual International Solid Freeform Fabrication Symposium (SFF Symp 2022)
|
Presentation Title |
Simplified Unsupervised Statistical Anomaly Detection for In-situ Quality Control of Directed Energy Deposition (DED) Additive Manufacturing |
Author(s) |
Ehsan Dehghan Niri, Steven Hespeler, Michael Juhasz, Harold Halliday, Melanie Lang |
On-Site Speaker (Planned) |
Ehsan Dehghan Niri |
Abstract Scope |
The leading reason for parts rejected during metal Additive Manufacturing is the creation of unacceptable defects. Post-process nondestructive testing methods are either time-consuming or impractical for quality control of AM parts with complex geometries. The slow nature of the AM process provides a unique opportunity to collect certain data in real-time to be used for in-situ quality control. The first step is to develop an automated unsupervised statistical anomaly detection algorithm that can detect abnormalities in both measured parameters and sensing features. In this paper, to detect anomalies, a simple unsupervised statistical method is developed to detect outliers in the collected data during the laser-based Directed Energy Deposition process. The results show, while the algorithm could successfully detect the creation of large porosities it could not reliably detect creation of smaller defects. It is concluded that additional sensing data is critical for reliable in-situ quality control to detect small defects. |
Proceedings Inclusion? |
Definite: Post-meeting proceedings |