“Apparent Net-Section-Collapse” Methodology for Circumferential Surface Flaws in Piping

Abstract

The Original Net-Section-Collapse (NSC) analysis was developed in the 1970s for prediction of the maximum (failure) moment for a pipe with a circumferential flaw, and is used widely for circumferentially surface-cracked pipe flaw assessments. A noticeable inaccuracy in the Original NSC analysis was that it predicts that under load-controlled loading the surface-cracked pipe will always have a higher moment capacity than a pipe with a through-wall crack (TWC) of the same length as the surface crack. A large number of past pipe tests show that deep surface cracks in pipes can break through the thickness and result in leaks; so the maximum moment of that surface-cracked pipe was below the maximum moment for the circumferential TWC pipe with the same length. In these cases the applied moment has to be increased for the resulting crack to grow as a TWC. Hence, load-controlled leak-before-break (LBB) fracture behavior has been experimentally observed although it is not predictable by the Original NSC analysis. Furthermore, the loads to develop the leak can be significantly less than the maximum loads predicted by the NSC analysis for the same size flaw. Since it is undesirable to have leakage in many applications, this deficiency in the Original NSC analysis was explored by conducting a matrix of pipe tests explicitly designed to show the experimental differences with the Original NSC equation for actual load-controlled LBB conditions. Circumferential surface-cracked pipe tests were conducted with flaws in the base metal of TP304 stainless steel pipe, as well as in the center of girth welds. Most of the pipe tests were conducted under pure bending, but a few selected surface-crack geometries were conducted with internal pressure. The Original NSC analysis for circumferential surface-cracked pipes under combined bending and axial tension were enhanced through the development of the “Apparent NSC” approach. This modification explained inconsistencies with the Original NSC that has been documented from recent pipe fracture tests conducted, and other pipe fracture data from many countries, as well as implemented in the current ASME Section XI flaw evaluation procedures. The data from the carefully planned circumferential surface-cracked pipe tests also showed that the toughness in the surface-cracked pipe decreases as the flaw depth increases. Additionally, an ovalization/thickness change aspect that increases the moment-carrying capacity for longer cracks was observed compared to the Original NSC equations that assume the pipe is perfectly circular with a constant thickness. The “Apparent NSC” modification accounts for toughness and ovalization/thickness changes and was also validated by stainless steel pipe test data from past EPRI/Battelle and JAERI pipe data from Japan on similar TP304 circumferentially surface-cracked pipe tests also at room temperature. Data from several other past programs with larger diameter pipes and different materials (and test temperatures) were also used to assess the general applicability of the “Apparent NSC” analysis.