The dataset, titled “In situ Visible Light and Thermal Imaging Data from a Laser Powder Bed Fusion Additive Manufacturing Process Co-Registered to X-ray Computed Tomography and Fatigue Data,” is part of the Department of Energy’s Manufacturing Demonstration Facility’s effort to provide comprehensive datasets to support the nation’s additive manufacturing industry. The new dataset aims to establish strong correlations between manufacturing anomalies, internal defects, and resulting mechanical performance.
This dataset includes cutting-edge monitoring data for laser powder bed fusion (L-PBF), a process that employs a laser to melt and fuse metal powder to construct metal parts layer by layer. It encompasses machine process parameters, sensor data, geometries, and detailed images of the 3D printing process captured from various angles and lighting conditions. The dataset integrates high-resolution visible and near-infrared imaging along with X-ray scans of the printed parts.
Luke Scime, a researcher at ORNL, mentioned, “Peregrine captures images during printing and utilizes AI to identify anomalies. This allows for the creation of a three-dimensional map highlighting potential issues, aiding in predicting problematic areas in the final part.”
The Peregrine software’s custom algorithm analyzes pixel values of images to scrutinize edges, lines, corners, and textures, sending alerts to operators regarding any issues during the printing process for prompt adjustments. Through its Dynamic Multilabel Segmentation Convolutional Neural Network (DMSCNN), Peregrine leverages data from multiple sensors to detect problems and issue alerts. For instance, L-PBF prints may encounter spatter, where molten material is expelled as the laser melts the metal powder, potentially affecting the part’s overall quality.
The new dataset includes all DMSCNN segmentation results and fatigue-tested specimens affected by spatter-induced perturbations. This comprehensive set of data supports the development of AI models for digitally qualifying additive manufacturing processes. By utilizing the enhanced open-source Peregrine dataset, researchers and manufacturers can enhance the intelligence of their quality assurance and control systems for 3D-printed components.
Contributors to the new dataset from ORNL include Zackary Snow, Chase Joslin, William Halsey, Andres Marquez Rossy, Amir Ziabari, Vincent Paquit, and Ryan Dehoff.
More information:
Zackary Snow et al, In situ Visible Light and Thermal Imaging Data from a Laser Powder Bed Fusion Additive Manufacturing Process Co-Registered to X-ray Computed Tomography and Fatigue Data, Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States). Oak Ridge Leadership Computing Facility (OLCF); Oak Ridge National Laboratory (ORNL), Oak Ridge, TN (United States) (2025). DOI: 10.13139/ornlnccs/2524534
