Oil-Gas Pipeline Magnetic Flux Leakage Testing Defect Reconstruction Based on Support Vector Machine

Oil-gas pipelines Magnetic flux leakage (MFL) inspection, the reconstruction of defects shape is the key of pipeline inspection integrity evaluation, identifying the shape of defects would lead to a better maintenance plan. In this paper, Support Vector Machine (SVM) method is used in reconstruction of defects shape features. Collecting 450 terms data of 30 defects from field-testing, by training, establish MFL data library based on samples of three different shape defects MFL signal characteristics. MATLAB as a platform is made in the reconstruction experimental, to verify the accuracy of defect classification and defect Identification. The achieved accuracy is 92.7% identifying the class of emulated defects over a set of 210 recordings. The experimental results show: the method has high accuracy and good generalization ability.