Monitoring And Identifying Of Gas Pipeline Leakage Signals Based On Interference Distributed Optical Fiber Sensing

In recent years,due to the energy structure adjustment,natural gas pipeline construction has developed very quickly.Because of the natural gas has the characteristics of flammable and explosive,so once the gas pipeline leakage accident happens,huge economic loss will be caused and there will be more serious threat to the lives of the local people.In this dissertation,based on interferometric distributed optical fiber sensing,the monitoring technology of gas pipeline leakage signals based on interference distributed optical fiber sensing are monitored and analyzed.(1)The principle of optical fiber sensing and its interferometry are introduced firstly.Then,the principle of distributed optical fiber sensing based on Sagnac and Mach-Zehnder interferometer is discussed,and the PGC phase demodulation technique is mentioned.Finally,the formula of pipeline leak detection is deduced.(2)Aiming at the problem of signal polarization fading of the system,based on the polarization fading of the signal,the mathematical model of the transmitted light is established by using the Jones matrix analysis method.The effects of the leakage,no leakage and the position of the leakage point on the output interference spectrum are studied.Also,the influence of signal polarization fading on system leakage detection and localization performance is analyzed.(3)To avoid the phenomenon of false positives(false alarm and missing alarm)occurring in the pipeline leak detection system,an carbon steel pipe section with?40 mm diameter and a 3PE pipe section with ?100 mm diameter were selected as testing pipe,and the leakage and leak-The false alarm rate and false alarm rate of the system are calculated.(4)In order to reduce the false alarm rate in the detection system,we use wavelet packet energy feature extraction and cluster analysis based on Euclidean distance method to process the leakage signal,to realize the judgment of pipeline leakage.Null-frequency method for pipeline leakage location.The results show that the method can accurately identify the pipeline leakage,compared with thetraditional null-frequency method based on the false alarm rate of leakage detection decreased by 6.59%.(5)In order to reduce the missing alarm rate in the system,we use the method of the time-domain amplitude statistic to extract the feature of the leakage time-domain signal.Under the pipeline leakage condition,the principal component model is established.The leakage of the pipeline is judged according to the Squared Prediction Error(SPE)statistic,and the Missing Alarm Rate(MAR)is calculated.The result shows that the leakage rate can be accurately estimated by using the SPE statistic,which is about 10 times lower than that of the traditional null frequency method.