AI's eyes to help with component inspections (newsreleases.sandia.gov)

At Sandia National Laboratories, a transformative inspection workflow has been initiated to enhance defect detection in the manufacturing of ceramic components used in nuclear deterrence. Led by process engineer Jesse Adamczyk, the project integrates new optical and acoustic imaging systems with an AI-assisted review tool, significantly speeding up inspections while keeping human oversight in the loop. The aim is to shift reliance from time-consuming manual microscope inspections to a streamlined digital workflow, allowing operators to review high-resolution images of ceramic billets directly from their workstations. This innovation is especially crucial since identifying defects at the billet stage—before they advance to costly final components—can significantly reduce manufacturing time and costs. The AI augmentation interface enhances human capabilities by highlighting potential anomalies, while trained operators remain essential to verify AI findings. As operators embrace this technology, which is expected to be fully operational soon, the goal extends beyond immediate improvements; it aims to set a precedent for deploying similar workflows across other areas of Sandia and the broader nuclear security enterprise. Funded by the National Nuclear Security Administration's AI for Nuclear Security initiative, this project represents a significant step towards integrating AI into complex scientific challenges.