Feature analysis acoustic signals for fiber-optic-sensing-based NDE for welded structures

The fiber-optic distributed acoustic sensing (DAS) technique has increasingly become more attractive for structural health monitoring (SHM) and non-destructive evaluation (NDE) purposes. When it comes to traditional acoustic NDE methods, the presence of weldings can present a significant challenge as it can heavily scatter waves resulting in complex data analysis and interpretation. The present work aims to develop an improved understanding and interpretation framework in cases where welds play an important role in the signal with an emphasis on the steel shell of a canister, typically used for Dry Cask Storage Systems (DCSSs) that house spent nuclear waste fuel rods. We also introduce a promising approach in the use of guided ultrasonic waves along with fiber optic sensors that seeks to overcome the challenges that emerge when using traditional acoustic sensing based NDE techniques in welded structures. The study is conducted in a simulation theoretical manner, using a canister model constructed from a representative stainless-steel plate, with different configurations of weldings typically present for DCSS structure. Progressively increasing complexity of the weld physical representation is considered to fully incorporate in physics-based analysis. Furthermore, the acoustic response of these models is obtained from the simulations as a response of an assumed DAS or quasi-distributed acoustic sensing Q-DAS system network. The features originated from the welds are extracted and analyzed, and additional features associated with structure integrity associated with corrosion defects, etc. will also be explored for NDE inspection as in a traditional acoustic NDE approach.