CN-121141153-B - Breather valve performance on-line inspection method for oil gas storage and transportation

CN121141153BCN 121141153 BCN121141153 BCN 121141153BCN-121141153-B

Abstract

The invention relates to the technical field of tightness test, in particular to an online checking method for the performance of a breather valve for oil and gas storage and transportation, which comprises the steps of obtaining a vibration signal of a valve body of the breather valve, a multi-frequency-band acoustic signal of a valve port and a total pressure signal in a storage tank; the method comprises the steps of performing multi-scale complex wavelet decomposition, constructing a time-frequency-energy correlation characteristic tensor, taking the total pressure signal and the time-frequency-energy correlation characteristic tensor as combined observation values, inputting the combined observation values into a preset continuous Gaussian mixed hidden Markov model which comprises four hidden states of sealing, transient micro-leakage, continuous leakage and full opening, calculating posterior probability, and further calculating to obtain real-time leakage rate. The invention not only can accurately determine the opening pressure, but also can quantitatively calculate the real-time leakage rate after the leakage state is identified, thereby realizing comprehensive and accurate quantitative online evaluation of the core performance parameters of the breather valve and improving the multi-source information fusion degree and the anti-interference capability.

Inventors

Lin haichuan
SUN GANG
YANG CHEN
YANG SHAN
XIE LEI

Assignees

太仓阳鸿石化有限公司

Dates

Publication Date: 20260505
Application Date: 20250916

Claims (9)

1. An on-line inspection method for the performance of a breather valve for oil and gas storage and transportation is characterized by comprising the following steps: S1, obtaining a vibration signal of a valve body of a breather valve, a multi-frequency band acoustic signal of a valve port and a total pressure signal in a storage tank; S2, carrying out multi-scale complex wavelet decomposition on the vibration signal and the multi-band acoustic signal, calculating a cross-correlation function of energy spectrums of the vibration signal and the multi-band acoustic signal under different scales, and constructing a time-frequency-energy correlation characteristic tensor; the total pressure signal and the time-frequency-energy correlation characteristic tensor are used as combined observation values and are input into a preset continuous Gaussian mixed hidden Markov model which comprises four hidden states of sealing, transient microleakage, continuous leakage and full opening, and posterior probability of each hidden state in each time step is calculated through a forward-backward algorithm; When the posterior probability of the transient micro-leakage or continuous leakage state exceeds a preset second threshold, calculating to obtain the real-time leakage rate according to a pre-established functional relation model of the characteristic tensor and the leakage rate in the leakage state; The method comprises the steps of adopting complex wavelets as mother wavelets to carry out multi-scale decomposition on vibration signals and multi-band acoustic signals, adopting sliding time windows to process the decomposed signals, calculating cross-correlation functions of energy spectrums of the vibration signals and energy spectrums of the multi-band acoustic signals under each scale in each time window, extracting zero-delay cross-correlation coefficients of the cross-correlation functions, forming a feature vector by the zero-delay cross-correlation coefficients under all scales, and stacking the feature vectors of continuous time windows to form the time-frequency energy correlation feature tensor.
2. The method of claim 1, wherein the step of acquiring the vibration signal of the valve body of the breather valve, the multi-band acoustic signal of the valve port and the total pressure signal in the storage tank comprises the steps of arranging a piezoelectric acceleration sensor at the center of the top of the shell of the valve body of the breather valve to acquire the vibration signal, arranging an acoustic sensing system at a preset distance right above the valve port of the breather valve to acquire the multi-band acoustic signal, and acquiring the total pressure signal through a pressure transmitter preset at the top of the storage tank.
3. The method of claim 2, wherein the acoustic sensing system covers a 20Hz-60kHz frequency band.
4. The method of claim 1, wherein the complex wavelet is a Morlet complex wavelet.
5. The method of claim 4, wherein multi-scale decomposing the vibration signal and the multi-band acoustic signal comprises decomposing the vibration signal and the multi-band acoustic signal into 128 scales, each scale corresponding to a center frequency.
6. The method according to claim 1, wherein the training process of the continuous Gaussian mixture hidden Markov model comprises the steps of carrying out mixed modeling by using a plurality of Gaussian components for each of four hidden states of sealing, transient micro-leakage, continuous leakage and full-quantity opening, and carrying out iterative training on initial state probability, state transition probability matrix and parameters of the Gaussian mixture model in each state until convergence by using pre-collected standard working condition sample data through a expectation maximization algorithm.
7. The method of claim 6, wherein a hybrid model comprising five gaussian components is configured for each of the four hidden states.
8. The method of claim 1, wherein the first threshold is set to between 0.9 and 0.98 and the second threshold is set to between 0.75 and 0.9.
9. The method of claim 8, wherein the functional relationship model of feature tensor under leak condition and leak rate is a support vector regression model or a neural network regression model.

Description

Breather valve performance on-line inspection method for oil gas storage and transportation Technical Field The invention relates to the technical field of tightness test, in particular to an online checking method for the performance of a breather valve for oil and gas storage and transportation. Background The breather valve is an indispensable safety accessory on the oil and gas storage tank, and has the core function of automatically opening and closing when pressure difference is generated between the inside and the outside of the storage tank so as to maintain the pressure in the tank within an allowable range, thereby preventing the storage tank from being damaged due to overpressure or negative pressure and ensuring the safety of storage and transportation operation. The performance of the breather valve, particularly the opening pressure and sealing performance thereof, is a key index for guaranteeing the effective function of the breather valve. However, due to long-term exposure in a complex industrial environment, the valve clack of the breather valve and the valve seat may be corroded, worn or attached with impurities due to the influence of various factors such as medium corrosion, environmental temperature change, mechanical vibration and the like, so that performance degradation, such as drift of opening pressure, leakage caused by untimely closing and the like, is caused, and if the faults are not found in time, serious potential safety hazards are formed. In order to ensure the reliability of the breather valve, a method of periodic offline verification is commonly adopted in the industry at present. According to the method, the breather valve is required to be detached from the storage tank and transported to a special verification platform, and the performance test is carried out by using special equipment. The traditional method has the obvious defects that firstly, the disassembly and assembly process is complex, time and labor are consumed, the normal operation of the storage tank is required to be interrupted during the disassembly and assembly process, the production efficiency is influenced, secondly, a long supervision blind area exists between the verification periods, the actual working state of the breather valve cannot be mastered in real time, the breather valve cannot be timely perceived once the breather valve breaks down during the verification period, and in addition, secondary damage to the valve can be caused during the disassembly and assembly process and the transportation process. Therefore, a technology capable of carrying out real-time and online inspection on the performance of the breather valve is developed, and the technology has important significance for improving the intrinsic safety level of the oil gas storage and transportation field. The existing part of online monitoring attempts depend on single pressure or acoustic signals, but under the background of strong noise, the signal characteristics are weak and easy to interfere, and multiple working conditions such as sealing, micro leakage and opening are difficult to distinguish accurately, and especially accurate quantification of early micro leakage cannot be performed, so that an online inspection method which is integrated with multi-source information, high in anti-interference capability and advanced in diagnosis model is urgently needed. Disclosure of Invention The invention provides the following scheme for solving the technical problem of how to improve the fusion degree and the anti-interference capability of multi-source information. A breather valve performance online test method for oil and gas storage and transportation comprises the steps of S1, obtaining a vibration signal of a breather valve body, a multi-frequency band acoustic signal of a valve port and a total pressure signal in a storage tank, S2, carrying out multi-scale complex wavelet decomposition on the vibration signal and the multi-frequency band acoustic signal, calculating cross correlation functions of energy spectrums of the vibration signal and the multi-frequency band acoustic signal under different scales, constructing a time-frequency-energy correlation characteristic tensor, taking the total pressure signal and the time-frequency-energy correlation characteristic tensor as combined observed values, inputting the combined observed values into a preset continuous Gaussian mixed hidden Markov model containing four hidden states of sealing, transient microleakage, continuous leakage and full-open, calculating posterior probability of each hidden state at each time step through a forward-backward algorithm, determining the total pressure signal value corresponding to the moment as opening pressure when the posterior probability of the full-open state exceeds a preset first threshold, and calculating a leakage rate model according to the pre-established characteristic tensor relationship of the feature tensor under the leaka