From Event: SPIE Smart Structures + Nondestructive Evaluation, 2023
X-ray computed tomography (XCT) is a promising non-destructive evaluation technique for various advanced manufacturing industries including additive manufacturing (AM). In an AM part, different types of flaws such as lack-of-fusion pores, gas pores, keyhole pores, near-surface pores, and trapped powders can occur. An automated/assisted flaw detection algorithms are expected to be implemented for an automated analysis and inspection. Evaluation and qualification of the algorithm’s flaw capability is a critical aspect of NDT qualification. NIST is developing a computational framework to generate data sets for evaluation of detection algorithms based on realistic XCT simulation. In this presentation, we will describe our workflow to generate complex pore shapes and distribute them in realistic AM parts. Example application and evaluation results will be discussed using various evaluation metrics. Such data sets are expected to help evaluate reliability of detection algorithms.
© (2023) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Felix Kim, Owen V. Hammer, Adam L. Pintar, John Henry J. Scott, and Edward J. Garboczi, "Developing XCT data sets for evaluation of automated flaw detection algorithms (Conference Presentation)," Proc. SPIE PC12491, 8th International Workshop on Reliability of NDT/NDE, PC1249108 (Presented at SPIE Smart Structures + Nondestructive Evaluation: March 14, 2023; Published: 25 April 2023); https://doi.org/10.1117/12.2657270.6323955538112.