This invention relates to the field of sealing performance testing technology, specifically an online performance testing method for
breather valves used in oil and gas storage and transportation. The method includes: acquiring vibration signals from the
breather valve body, multi-band acoustic signals from the
valve port, and total pressure signals within the
storage tank; performing multi-scale complex
wavelet decomposition to construct a time-frequency-energy correlation feature
tensor; inputting the total pressure
signal and the time-frequency-energy correlation feature
tensor as combined observations into a preset continuous
Gaussian mixture
hidden Markov model containing four hidden states: sealed, transient micro-leakage,
continuous leakage, and full opening, to calculate the
posterior probability; and then calculating the real-time
leakage rate. This invention not only accurately determines the opening pressure but also quantitatively calculates the real-time
leakage rate after identifying the leakage state, achieving a comprehensive, accurate, and quantitative
online evaluation of the core performance parameters of the
breather valve, and improving the degree of multi-source
information fusion and anti-interference capability.