Fracture analysis of full-scale pipe experiments on stainless steel flux welds

Abstract

The objective of this paper is to verify the ability of current fracture analysis methods to predict loads and displacement for circumferential, through-wall-cracked, stainless steel, flux-welded (i.e. submerged-arc weld and shielded-metal-arc weld) pipes under pure bending. Crack initiation and maximum load predictions were made using five different J-estimation schemes, the Net-Section-Collapse analysis, and the ASME Section XI IWB-3640 flaw evaluation criteria. The predicted loads were compared with the observed crack initiation and maximum loads from four full-scale pipe fracture experiments involving 152.4 mm (6 in) and 711.2 mm (28 in) nominal diameters at 288 °C (550 °F). In some cases, load-displacement relationships predicted by the estimation methods were also compared with the actual test record. The experiments were conducted under the Degraded Piping and Short Cracks in Piping and Piping Welds Programs. The results show that both LBB.ENG2 and LBB.ENG3 methods, which were previously developed by the authors, provide more accurate predictions than other analysis methods considered in this study. The comparisons with pipe fracture data also reveal that differences in the weld thickness and procedure between the pipes and compact tension specimens can provide widely different fracture toughness characteristics and hence, can significantly influence predictions of the pipe's load-carrying capacity by the J-estimation methods.