Research on Classification and Identification of Oil and Gas Pipeline Defects Based on PCC-LPSO-BP
In order to accurately identify the defect type of oil and gas pipeline and analyze the influence of the defect magnetic flux leakage signal characteristics on the identification accuracy,and the oil and gas pipeline defect identification model based on PCC-LPSO-BP was established.The Pearson correlation coefficient(PCC)method was used to analyze the degree of correlation between the characteristics of the defect magnetic flux leakage signal and the defect size,the BP neural network detection model(LPSO-BP model)optimized by particle swarm optimization based on chaos mapping and Levy flight improvement was established,the detection performance of the model was comprehensively compared by using the evaluation index,the detection accuracy of the model for each defect type and the influence of each characteristic on the detection result were analyzed.The results show that the recognition accuracy of LPSO-BP model is increased by 7.47％compared with BP model,and the recognition rate of surface spalling and cracks reaches 100％within the existing data range.The research results have a certain reference value for the identification and quantification of oil and gas pipeline defects.
magnetic flux leakage signaldefect identificationpearson correlation coefficientLPSO-BP mode
NETL
NSTL
万方数据
