168

Free floating inline inspection (ILI) tools using magnetic flux density (MFD) measurements have been enhancing pipeline integrity programs for nearly a decade. The use of MFD measurements to infer information on the pipeline wall condition has been demonstrated and with more inspection data becoming available, the capability of using MFD measurements is continuously increasing. This work focuses on anomalous weld identification by applying principal component analysis to magnetic flux density data captured by free-floating ILI tools. The concept of principle component analysis will be explained, how this approach can be applied to unsupervised classification tasks in general, and the results of the specific use case of anomalous weld identification.

Read/Download PaperGradeAvailable after the event