Timely detection of leaks is essential for the safe and reliable operation of pressure vessels used in superconducting systems, aerospace, and medical equipment. To address the lack of efficient online leak detection methods for such vessels, this paper proposes a quasi-distributed fiber Bragg grating (FBG) sensing network combined with theoretical stress analysis to diagnose vessel conditions. We analyze the stress–strain distributions of vacuum vessels under varying pressures and examine stress concentration effects induced by small holes; these analyses guided the design and placement of quasi-distributed FBG sensors around the vacuum valve for online leakage monitoring. To improve measurement accuracy, we introduce a vibration correction algorithm that mitigates pump-induced vibration interference. Comparative tests under three leakage scenarios demonstrate that when leakage occurs during vacuum extraction, the proposed system can reliably detect the approximate leak location. The results indicate that combining an FBG sensing network with stress concentration analysis enables initial localization and assessment of leak severity, providing valuable support for the safe operation and rapid maintenance of vacuum pressure vessels.
The influence of mechanical noise on impact identification using fiber Bragg grating (FBG) accelerometers in complex environments remains insufficiently explored and quantitatively characterized. This study investigates how the intensity and propagation path of mechanical noise affect the output stability and impact-detection capability of a self-developed FBG accelerometer. A quantitative evaluation approach is established using wavelength-variation statistics, impact-response amplitude, and signal-to-noise ratio (SNR). Experimental results show that mechanical noise markedly increased the background wavelength fluctuation, with the maximum standard deviation reaching 16.71 pm, approximately 11 times the no-additional-noise value. By contrast, the RMS of the target impact response increased only slightly, from 0.0300 nm to 0.0353 nm, indicating that mechanical noise primarily affects baseline stability rather than the amplitude of the target impact response. Under the highest-intensity disturbance tested, the SNR remained at 22.98 dB, and the impact event remained distinguishable. The proposed multi-indicator approach provides a quantitative characterization of the FBG accelerometer response under different mechanical-noise conditions. A commercial 941B accelerometer was used as an external reference for sensitivity calibration, and its measurements were not used for direct performance comparison with the FBG accelerometer. The results provide experimental evidence of the impact-identification behavior of the proposed sensor under the investigated mechanical-noise conditions.