Research On Leakage Detection And Leakage Point Location Of Long Distance Oil Pipeline

Pipeline transportation is the main mode of transportation in the petroleum transportation industry,and the safety of pipeline transportation plays an important role in the national economy and environmental protection.However,with the increase of the service life of the oil pipeline,the leakage of oil by the lawless elements or the incorrect operation of the construction workers,the pipeline leakage accidents occur.So,the development of pipeline leak detection and leak location technology plays an important role.This paper based on the negative pressure wave leak detection principle,using wavelet denoising method on the negative pressure wave signal denoising,then uses wavelet singular point positioning method to determine the location of the singular point.The negative pressure wave signal acquisition and extraction of the mean value,variance,kurtosis and skewness,as feature vectors,using the method of support vector machine classifier training,finally,the test set is used to predict the effect of classifier.Before the training of the classifier respectively compared the classification results of different kernel functions and different kernel function parameter optimization method,then select the least squares support vector machine(LS-SVM)classification method based on particle swarm optimization(PSO),and the experimental results show the effectiveness of the proposed method.In the process of denoising,the denoising results of different wavelet basis functions and different decomposition levels are compared,the denoising index is used to select the best parameters for wavelet denoising.In the process of support vector machine classification,linear,polynomial and radial basis functions are selected to compare the classification and prediction,At last,the penalty factor and kernel parameters are optimized based on the choice of radial basis function.The experimental results show that particle swarm optimization(PSO)algorithm can optimize the parameters of kernel function.The combination of particle swarm optimization(PSO)method and least squares support vector machine(LS-SVM)finally achieves a good classification result,then it can detect whether there is pipeline leakage.