Pipeline Leak Aperture Classification And Location Based On Local Mean Decomposition Analysis

Natural gas has gradually become the main energy source that supports industrial production and meets the daily living needs of urban inhabitants. However, there is an increasing risk of small leaks because of ageing pipes, corrosion, welding defects, etc. This issue is a critical potential danger in the safe operation of gas pipelines, therefore, effective methods of determining leak size and leak location must be investigated. In this paper, to recognize leak aperture and location of pipeline leakage information, the acoustic emission techniques in the gas pipeline leak are studied, with the local mean decomposition processing the leakage signals and extracting characteristics information. This paper will focus on the following works:Firstly, on the basis of analyzing the development and research statuses of detection technology method at home and abroad in pipeline leakage, the method of processing non- stationary signals were discussed. The conventional signal processing methods aren’t suited for non-stationary signals, so, local mean decomposition methods is used to process complex leakage signal, the decomposition principle, instantaneous frequency, the characteristics of local mean decomposition method are described, and its advantages and disadvantages are analyzed, simulated signal is given.Secondly, when small leak occurs in the natural gas pipeline, it is difficult to identify the leak scale and aperture. Three kinds of leak aperture recognition methods are proposed, LMD envelope spectrum entropy and SVM、ELMD and Boosting-SVM、the method based on nuclear density estimation. The acquired pipeline leakage signals are decomposed; entropy is used to represent different aperture leakage characteristics, then, studying Boosting-SVM and kernel density estimation of recognition methods. Experimental results show that the proposed methods can effectively identify the different leak apertures, higher recognition accuracy.Finally, aiming at the location problem in pipeline leak detection, local mean decomposition method is used to improves the traditional method of cross-correlation delay; on the basis of analyzing non-stationary signal arrival time difference, it proposed a estimation method of time delay based on Ensemble Local Mean Decomposition(ELMD) and high-order ambiguity function. The leakage signals is decomposed by ELMD and obtained the principal PFs which contained most of leak information. Calculate the time difference of arrival of characteristic frequencies by high-order ambiguity function. Combining time difference with signal propagation speed accomplishes natural gas pipeline leak location. Experiment results show the proposed method can locate the leak and the location accuracy is apparently higher than direct correlation method.