Magnetic Flux Leakage Detection For Long-distance Pipeline Defect

Pipeline transportation has become the main method to transport oil and gas. Due tothe expansion of pipeline construction in China and the durable deadline is closing for partof the pipeline, large amount of safety accidents have emerged in recent years. An effectiveway to prevent pipeline accidents and ensure the safety is the regular pipeline inspection.Among the various non-destructive testing technology of oil and gas pipelines, the magneticflux leakage detection method seems had become the most widely used method, but axialdefects cannot be detected in this way. In order to solve this problem, this paper mainlystudied the pipe circumferential excitation circumferential defect by magnetic flux leakagedetection.In this thesis, the formation mechanism of the leakage magnetic field of pipelinedefects is analyzed, and the magnetic dipole model approximation analysis of commondefect leakage magnetic field is drawn. Considering the insufficiency about the magneticdipole model description for complex defects, the leakage magnetic field finite elementanalysis method is adopted. ANSYS is utilized to analysis leakage magnetic fielddistribution of typical pipeline defects. The pipeline corrosion defective geometricparameter on the signal characteristics of the leakage magnetic field is analyzed using thefeature of signal to describe the geometrical characteristics of the pipeline defects. Thesensor lift-off value on magnetic flux leakage signal is also researched so that a reasonableand effective sensor lift-off could be determined.The thesis also studied the pipeline circumferential excitation magnetic flux leakagedetection principle, and proposed an actual pole separation ring magnetizer structure. Byutilizing the magnetic circuit theory, a magnetic equivalent circuit mathematical model is established. And the magnetic circuit optimization application is done by MATLAB. Theeffect of magnetizer permanent magnets parameters on the magnetic flux leakage signal isalso studied, and a basis for the magnetization parameter design is provided.A detection system of pipeline circumferential direction of excitation MFL is designedby using Hall element as MFL detection sensor in order to avoid the non-uniform sampling,and a displacement sensor module is attached to achieve equidistantly signal sampling. Thisthesis also designed a real-time data acquisition software, and analyzed pipelinecircumferential excitation magnetic flux leakage signal with wavelet method includingsignal wavelet de-noising, signal compression, singular signal point removing, and the weakcharacteristics signal mutation point analysis. The defect signal samples are analyzed andthe signal characteristics of the magnetic flux leakage are extracted. Based on the result, theprinciples of multiple linear regressions are utilized to determine the regression fittingequation of defects parameters and magnetic flux leakage signal characteristics.