Previous work has shown that a corrosion assessment more accurate than B31.G or RSTRENG can be made if pipeline stresses are considered. A shell analysis can be carried out if both the corrosion profile and local pipe wall stresses are known. The corrosion profile can be approximated from analysis of magnetic flux leakage (MFL) signals acquired by an inline inspection tool (smart pig), but a measure of pipe wall stress has not been available. Approximations have been made based on pipe curvature, but a more direct measurement is desirable.
Recent work has produced data that show a correlation between multi-level MFL signals from metal-loss defects and the stress in the pipe wall at the defect location. This paper presents the results of MFL scans of simulated corrosion defects in pipe specimens subjected to simultaneous internal pressure and four-point bending. MFL data were acquired at two different magnetic excitations using an internal scanner. The scanner’s sensor array measured axial, radial and circumferential magnetic flux components on the inner pipe surface adjacent to the defect. Comparison of the signals at high and low magnetization yields an estimate of the difference between axial and hoop stresses. If internal pressure is known, the hoop component can be determined, leaving data proportional to axial stress.