Study On Pipeline Leak Detection Method In Well Field-oil And Gas Concentration Station

Pipeline leak detection technology is an important means to ensure safe production of pipelines.Among them,signal processing based leak detection is the most widely used method in the field.With the development of computer technology,various intelligent algorithms are also applied to pipeline information processing to more accurately discover and locate leaks.Based on the introduction of actual oilfield pipeline operation,this paper mainly carries out the following work:1)By analyzing the actual pipeline operation data collected,a modern pattern recognition method for pipeline leak detection is proposed to improve the robustness and accuracy of the detection algorithm.Aiming at the actual situation of missing data in some oilfield field pipelines,a SVM-based leak identification method is proposed.Based on the analysis of the correlation curve of the oilfield production site data,the number of samples in which the SVM identifies the leak is given.Furthermore,the SVM kernel function and parameters suitable for leak identification are further determined,and a leak identification algorithm suitable for running data missing is proposed.2)The paper studies the BP neural network model structure and determines the activation function that can improve the leakage recognition rate.The BP neural network based on the activation function can be used as a supplement to the oilfield pipeline leakage detection method.3)The paper studies the application of cross-correlation principle in pipeline leakage location,and clarifies the relationship between cross-correlation function and pipeline leakage location,and implements the actual algorithm.The field research and verification by the oil field proves that the algorithm developed is feasible and effective.At the same time,based on the multi-scale analysis theory of wavelet transform,this paper proposes to use multi-scale analysis method to shield noise,calculate the relationship between wavelet transform modulus maxima and pipeline leakage position,and propose an operationally strong implementation algorithm.The feasibility of the proposed method is demonstrated by actual calculations.4)The paper analyzes the error of negative pressure wave positioning leakage,and proposes to use temperature to correct the leakage positioning formula.The actual calculationshows that this method can reduce the positioning error.