Automated non-destructive internal corrosion detection on radioactive drums (ZIKA)

Against the backdrop of extended storage times in German interim storage facilities and the increase in low- and medium-level radioactive waste, monitoring container integrity is of critical importance. The ZIKA (Automated non-destructive internal corrosion detection on radioactive drums/15S9446A) research project, carried out as part of the FORKA funding initiative, presents an automated system for non-destructive testing (NDT) of radioactive waste drums. The primary goal is the reliable early detection of internal corrosion in order to identify safety risks before the integrity of the containers is compromised by externally visible degradation. As a further development of the predecessor project EMOS, the system architecture has been optimized for mobile use in a compact 10-foot container. The novel design integrates complex lifting mechanisms and robotics to ensure complete inspection of the entire drum surface, including the bottom. The multi-sensory approach combines laser scanners for topographic mapping, smart cameras for detecting external defects, and active thermography for identifying internal corrosion. Experimental validations by the Federal Institute for Materials Research and Testing (BAM) showed that laser thermography is more robust and reliable in defect detection than flash thermography. Machine learning algorithms were implemented for data analysis, with random forest models achieving the highest accuracy in defect classification and artifact suppression. The ZIKA system thus represents a significant contribution to increasing long-term safety standards in nuclear interim storage facilities.