Fusion of Distributed Fiber Optics, Acoustic NDE for Spent Fuel Monitoring Towards Physics-Informed Machine Learning

In the results section, we discuss the complementary differences between passive and active monitoring for isolating the differences in the detailed time domain and frequency spectral characteristics of acoustic signatures brought by structural damage features of an idealized nuclear canister geometry. This difference can help us to further process and analyze their signals in future work within the framework of an integrated monitoring system approach. By identifying defect-sensitive and independent features in the resultant signal, such as the center frequency or the amplitude of the wave packet, it is important to help us in further damage location identification and identification of damage types. Both simulated and experimental results can also be incorporated into integrated frameworks to assist with data interpretation such as machine learning and artificial intelligence. In summary, major contributions of the work included: 1. Pressure Impact on the canister inner wall has successfully been detected using simulated optic fiber sensor segment. The feature of pressure-induced vibration is cited and discussed. 2. The frequency of pressure shock-induced vibrations was analyzed. The feasibility of detecting the occurrence of fuel leaks in the tank is demonstrated. 3. The applicability of the passive method in the tank SHM system is confirmed. 4. Comparison and Discussion of Active SHM Systems and Passive SHM Systems using fiber optic acoustic sensing. 5. Discussion and analysis of various signal characteristics using the same optical fiber sensor. Features of active and passive signals can be extracted from resulting signals as input for classification using the same sensor.