Non-destructive testing of additive manufacturing components for aerospace applications can be as expensive as the actual printing of the component. This research effort, funded by the US. Department of Energy augments existing sensor technology on a Sciaky Electron Beam Directed Energy Deposition system and utilizes machine learning to classify likelihood of defects during deposition. Zone criticality mapping is then applied to the component and a criticality vs. likelihood function maps non-destructive testing processes to reduce overall cost of testing. This presentation will be an overview of approach, report on registration of sensor data, and demonstrate process and sensor signal to defect mapping on coupon scale Ti-6Al-4V test blocks. Quality and thermal scenarios used to develop defect schema will be reviewed as well.
Learning Objectives:
- The participant will understand the complexities in spatial and temporal registration of sensor data for process to defect mapping.
- The participant will be able to define defect likelihood and zone criticality
- The participant will be able to implement a similar approach of implementing sensors and register data for their own defect prediction model.
