Monitoring And Fault Diagnosis Of Oil Storage Tank Breathing Valve Based On Multi-Sensor

The breathing valve is an important mechanical equipment to reduce the volatilization loss of oil and ensure the safety of oil storage tanks.Due to the long-term working of the breathing valve in an environment where it is easy to corrode,easy to oxidize,and easy to scratch,it is very prone to leakage,rust,stuck and other faults.Therefore,it is of great significance to carry out research on monitoring the abnormal state of breathing valve.This thesis takes the oil storage tank breathing valve as the research object to conduct real-time monitoring and regular fault diagnosis,realize the timely update of the fault status of the oil storage tank breathing valve,ensure the continuous and stable operation of the breathing valve,reduce the probability of damage to the storage tank caused by the breathing valve failure,and ensure the production safety.Firstly,the stability of the movement of the breathing valve disc is proposed as the basis for the fault diagnosis of the breathing valve,and a monitoring system scheme for the breathing valve of the oil storage tank based on the single chip microcomputer is constructed,and the test platform of the monitoring system is established.Driven by the hardware part of the monitoring system of the carrier and the corresponding software,the design of real-time monitoring,data acquisition,real-time storage,power supply integration,data preprocessing and threshold alarm function modules has been completed.Real-time monitoring of humidity and air pollution concentration.Then,the four typical states of the breathing valve,such as non-faulty,leaking,rusted and stuck,are analyzed,and the characteristic signal analysis method of the valve disc movement of the faulty breathing valve is proposed.Based on the collected displacement signal,the acceleration signal in the vertical direction when the valve disc moves is derived,and the signal is used for feature extraction in the time domain,frequency domain and time-frequency domain.The signal features extracted in the time domain are five dimensional parameter indicators such as maximum value,minimum value,variance,peak-to-peak value,and root mean square value,and three dimensionless parameter indicators such as kurtosis,impulse factor,and margin factor.The fault feature extracted in the frequency domain is the frequency standard deviation.In the time-frequency domain,two methods of CEEMD decomposition and wavelet packet transform are used to decompose and extract the fault signal,and make a comparative analysis.The results show that the wavelet packet transform has better signal decomposition effect and faster decomposition efficiency.Therefore,the fault feature extracted in the time-frequency domain is the third-layer band energy value after wavelet packet decomposition.Finally,the research on the fault diagnosis method of breathing valve based on multi-sensor is carried out.The feature-level fusion method and data-level fusion method in the multi-sensor fusion technology are used to combine the eigenvalues of the multi-sensor signals into the eigenvectors.Taking this vector as the input of the fault diagnosis model of the breathing valve,a fault diagnosis model of the least squares support vector machine optimized by the particle swarm algorithm was built,and the fault state of the breathing valve was accurately identified.The paper has 59 figures,8 tables,and 83 references.