Understanding Magnetic Flux Leakage from Gouges and Dents Containing Gouges
Chen, Jia Dian
MetadataShow full item record
Magnetic flux leakage (MFL) is the most widely used non-destructive evaluation (NDE) technique for operating gas and oil transmission pipelines. The Queen’s Applied Magnetics Group (AMG) has been studying MFL signals from dents for the past 5 years, and recently began considering MFL signals from gouges, and gouge/dent combinations. It is believed that MFL signals from gouges may arise from three factors: the gouge geometry, residual stresses, and a severe plastically deformed layer that is often present at the base of the gouge. The last factor has been studied by another MSc student, Kris Marble. The present thesis work primarily considers the effects of the first two factors – gouge geometry in combination with residual stress. Samples containing realistic gouges, produced under field-like conditions, were obtained from two sources - Stress Engineering Services (SES) in Houston, Texas and GdF Suez in St. Denis, France. At SES, ten different gouges on five pipes (two on each pipe) with increasing severity were produced for the current thesis project. Three gouged pipe samples were obtained from GdF Suez. The bulk of the thesis study involved experimental MFL measurements made for each of the gouges and gouge/dents of SES samples in the study, on both the outer and inner surface, and in some cases when the pipe wall was under pressure. The experimentally obtained MFL signals were interpreted by comparison with the expected MFL geometry-only signals (obtained using magnetic modeling) along with the knowledge of the residual stress information obtained from the neutron diffraction measurements. Various results were obtained and concluded. The neutron diffraction was also done on three GdF Suez samples. The results from one sample were used to model the residual stress distribution around the sample and the modeled MFL signals were compared with the experimental MFL signals. The modeled signals are able to account for most features in the experimental results.