Nuclear canister integrity monitoring using quasi-distributed fiber acoustic sensors and physics-based modeling

Absence of a final repository for nuclear waste has increased attention on dry cask storage systems (DCSSs) which were originally intended for temporary storage, increasing the need for new structural health monitoring paradigms considering safety and environmental impacts. Current integrity inspection requirements consist of periodic manned inspections due in part to the difficulties with real-time monitoring of internal canister conditions without penetrating the canister surface. Here we overview a new approach to nuclear canister integrity structural health monitoring which combines both quasi-distributed fiber optic acoustic (and other) sensing modalities deployed external to the canister as well as physics-based modeling to enable real-time inference of internal canister conditions, including the identification, localization, and classification of various active or incipient failure conditions. More specifically, we overview the vision for the proposed monitoring approach and describe results to date in theoretical physics-based modeling and artificial intelligence-based analytics to accelerate the development of classification frameworks for rapid interpretation of quasi-distributed acoustic and other complementary fiber optic sensing responses. In addition, we describe early results obtained for a quasi-distributed fiber optic sensor network based upon multimode interferometer sensors using an experimental test bed established for dry-cask storage canister sensing experiments. Future work will be overviewed and discussed in the context of expanded scope of the proposed real-time monitoring system and planned field validations.