Defect detection and width estimation in natural gas pipelines using MFL signals

Abstract

Magnetic Flux Leakage (MFL) testing is the most widely used non-destructive techniques for the in-service inspection of oil and gas pipelines. In this study, a novel approach for detecting and estimating the width of defects by employing MFL signals is presented. Estimating the locations of defects and profiles of lengths, widths, and depths of defects from measurements is a typical inverse problem in electromagnetic non-destructive testing. In this study, defect parameters are estimated in two separate consecutive steps. In the first step, a detection algorithm based on image processing approaches is applied on axial flux to estimate the numbers of defects, locations, and orientations of defects. Then, to estimate widths of defects, an inversion procedure based on 2D signal processing is applied on radial flux corresponding to areas detected in previous step. Finally, the efficacy and accuracy of the proposed algorithm is validated through examinations on simulated defects and real experimental MFL data. Simulated defects are generated in presence of multiple uncertainties and noises.