Automated Table for Inspection and Surveillance

Container inspections are required to ensure that all containers used for nuclear material storage purposes are up to the task of protecting workers, the public, and the environment, following the requirements outlined in DOE Manual 441.1-1. Container inspections require subject matter experts to visually inspect each container to identify concerning areas. The Automated Table for Inspection and Surveillance (ATIS) aims to improve this manual inspection process through an automated scanning system. Using this system, a container can be mounted to the machine to collect detailed information on relevant container surfaces. This allows subject matter experts to improve the quality and throughput of the container inspection process. Additionally, all scans performed are saved to document the history of each container surveilled. The system uses multiple sensors to create a “snapshot” of the current state of the container. A camera is used for inspection of the container’s interior surfaces, including the sidewall, bottom, and bottom radius. A depth sensor then maps the exterior surface to quantify any anomalies (dents) on the container. These sensors are positioned around a container using multiple motion stages and two robotic manipulators to fully automate the data collection process. A machine learning system, based on the U-Net architecture, is utilized to provide immediate feedback to the operator on detected features and their significance to the life of a surveilled container. The system has demonstrated capabilities to detect and quantify dents, corrosion, and pitting on nuclear storage containers.