FRACTURE TOUGHNESS VARIATION WITH FLAW DEPTH IN VARIOUS SPECIMEN GEOMETRIES AND ROLE OF CONSTRAINT IN MATERIAL FRACTURE RESISTANCE

A series of pipe tests with circumferential surface cracks has been conducted along with fracture toughness tests using single-edge notch tension (SENT) specimens having similar crack depths and crack orientations as the surface-cracked pipes. This paper presents observation of measured fracture toughness variation due to the crack depth and discusses the effect of constraint on the material resistance to fracture. Crack-tip-opening displacement (CTOD) measurements were obtained with the use of a dual clip-gauge mounted on both the SENT specimens and center of the surface-cracks in the pipes. CTOD was obtained at both the crack initiation and during the crack growth through the ligament. CTOD is a direct measure of the material toughness in the pipe and SENT tests. CTOD at crack initiation and during crack growth can also be related to the material J-Resistance (J-R) curve. Commonly, the material resistance is assumed to be the same for all circumferential surface-crack geometries in a surface-cracked pipe fracture mechanics analyses. However, based on experimental observations on a series of recently conducted surface-cracked pipe tests, the CTOD at the center of the surface crack at the start of ductile tearing and maximum moment changed with the depth of the surface crack. This is believed to be a constraint effect on plasticity in the ligament which depends on crack depth. The CTOD values at crack initiation were decreasing linearly with crack depth. This linear decrease in CTOD trend with flaw depth was also observed in SENT tests. More importantly, the decrease in CTOD with surface crack depth was significant enough that the failure mode changed from being limit-load to elastic-plastic fracture even in relatively small-diameter TP304 stainless steel pipe tests. This toughness drop explains why the Net-Section-Collapse (limit load) analysis overpredicted the maximum moment for some crack geometries, and why the deeper surface cracks tore through the pipe thickness at moments below that predicted by the NSC analysis for a through-wall crack of the same circumferential length. An "Apparent NSC Analysis" was developed in a companion paper to account for the changing toughness with crack depth [1]. Finally, this same trend in decreasing toughness with flaw depth is apparent in surface-cracked flat plates [2] and axial surface flaws in pipes [3]. The leak-before-break behavior for axial surface cracks is also not explained by numerical calculations of the crack-driving force when assuming the toughness is constant for all surface cracks and the through wall cracks, but the change in toughness with surface flaw depth explains this behavior. Previously, axial flaw empirical limit-load solution was developed by Maxey and Kiefner [4], and is consistent with the observations from this paper.