Speaker
Description
A major challenge in Additive Manufacturing (AM) is the development of reliable in-situ and online quality monitoring methodologies. Visible and infrared cameras can provide near real-time image data that can be exploited for anomaly detection through Statistical Process Control and Monitoring (SPC/M) methods.
This work investigates image-based monitoring methods for Selective Laser Melting (SLM) processes, aiming to detect shifts from the in-control (IC) to the out-of-control (OOC) state. Two approaches are compared: a partial first-order stochastic dominance methodology and generalized multilinear models for sufficient dimension reduction with tensor-valued predictors. In addition, a hybrid approach combining elements of both methodologies is proposed.
The methods are evaluated using simulated datasets generated from images of a real SLM process, with emphasis on monitoring performance and sensitivity to training sample size. The results highlight the potential of statistically grounded, data-efficient image monitoring methods for next-generation smart manufacturing systems.
| Special/ Invited session | Statistics and data science in the technological field: current issues and new proposals |
|---|---|
| Classification | Both methodology and application |
| Keywords | Non-Parametric, Sufficient Dimension Reduction, Statistical Process Control and Monitoring (SPC/M) |