Identification of Weak Gas Leaks Using Dual-Segment Consistency Verification Based on Center Frequency Dispersion Index

Real-time and accurate detection of in-orbit space-craft leakage is crucial for ensuring operational safety but remains a significant challenge due to weak, irregular signals often obscured by intense background noise. Traditional methods, lacking prior frequency information, struggle to consistently identify these subtle leakage signatures. In this paper, we introduce a novel frequency-domain stability analysis framework utilizing an refined Synchrosqueezing Transform (SST). The key contributions of this work are: (1) a refined SST-based spectral processing technique designed to achieve superior time-frequency resolution, and (2) The introduction of the Center Frequency Dispersion Index (CFDI) combined with dual-segment consistency verification enables automatic identification of candidate stable frequency bands. The proposed method is rigorously validated through experiments simulating leakage scenarios on 5A06 aluminum alloy plates (2.5 mm thick), featuring both a precise 0.2 mm micro-hole and irregular leakage geometries. Experimental results demonstrate the method’s outstanding performance, achieving over 90% identification accuracy for different kinds of leaks, while effectively suppressing false positives induced by impact signals. This work provides a significantly more accurate and robust solution for spacecraft leakage monitoring and holds strong potential for broader applications in weak, irregular signal detection within noisy environments.