Research On Contour Reconstruction Technology For Magnetic Flux Leakage Inspection Of Pipeline Defects

Long-distance pipelines,as the most convenient,efficient,and most economical transportation method for oil and gas transportation in my country,have always occupied a major position in China.With the rapid growth of my country’s economy,the demand for oil and natural gas is increasing,and the role of pipelines in oil and gas transportation is becoming more and more important.Previously,four strategic energy channels were specially established for the import of oil and natural gas.In order to ensure the safe operation of oil and gas pipelines,pipeline conditions need to be inspected and maintained.At present,the detection of long-distance pipelines generally adopts pipeline magnetic flux leakage internal detection technology.In order to intuitively and clearly reflect the corrosion status of pipeline defects,this paper studies the 3D contour reconstruction technology of pipeline defects based on the principle of magnetic flux leakage detection.Firstly,using ANSYS finite element simulation software,a three-dimensional simulation model for magnetic flux leakage detection of pipeline defects with different depths,lengths,and widths is established,and the relationship between the shape of the defect and the defect signal is studied.Then,the wavelet multi-scale analysis is used to study the magnetic flux leakage signal of the defect.According to the characteristics of the magnetic flux leakage signal,the wavelet basis and the correlation coefficients,the optimal wavelet basis for multi-scale analysis is selected,the defect signals at different scales are obtained.Then,the wavelet coefficients are thresholded,and the processed wavelet coefficients are used to reconstructed the signal to realize the denoising processing of the magnetic flux leakage signal.Through the feature extraction of the denoising signal,the parameters used to describe the characteristics of the defect waveform are established.In order to accurately realize the identification and quantification of defects,neural networks are used to establish the relationship between the characteristic quantity of the defect signal and the shape and size of the defect.And through experiments and simulation analysis,a large number of basic defect signal data has been obtained,and the defect samples for neural network training and learning are established by extracting the characteristic quantity of the defect signal.When constructing neural networks,different neural networks for defect recognition,depth quantification,length quantification,and width quantification are constructed to achieve different goals.When identifying the defect type,the BP(Back propagation)neural network is used to complete the identification of the defect type;when quantifying the defect,a BP neural network optimized based on genetic algorithm is established,which improves the network prediction accuracy of multi-node output and realizes the accurate quantification of defect depth,length and width.Through the establishment of the defect matrix,the one-to-one correspondence between the intersection of the virtual grid of the pipeline,the waveform signal and the matrix elements is realized,and the quantified defect data is imported into the defect matrix to complete the three-dimensional reconstruction of the defect contour,and realize the Location of defects.According to this reconstruction method,the contour reconstruction of the pipeline defect magnetic flux leakage detection is realized.