Model weldment fracture specimens have been fabricated, tested, and analyzed using finite elements. The specimens consist of an interleaf of commercially pure titanium diffusion-bonded to a harder alloy titanium. A deep edge crack is introduced symmetrically into the interleaf, and the specimens are loaded in pure bending. Variation of the thickness (2h) of the soft interleaf layer provides insight into effects of weld geometry in strongly undermatched weldments tested in plane strain bending. Ductile crack growth (beyond blunting) initiated at loads giving J ? 95 kJ/m2 in all specimens. In the thickest interleaf geometries, stable tearing was obtained, but in the thinnest interleaf (2h ? 3 mm), crack initiation resulted in a massive pop-in of 5.4 mm across an initial ligament of 12 mm. Finite element studies show that the thinnest interleaf geometry had slightly higher peak stress triaxiality at the beginning of cracking, and that the highest triaxiality extended over a larger region than in the thicker interleaf specimens loaded to the same initiation J-values. More importantly, the blockage of plastic straining above and below the crack tip in the 3 mm interleaf specimen forced higher values of plastic strain to spread forward into the ±45‡ sector of highest stress triaxiality directly ahead of the crack tip. The higher strains, in conjunction with the slightly higher stress triaxiality, led to the unstable pop-in initiation.

If a structural material behaves in a linear elastic manner, it is generally assumed that plane-strain fracture toughness (KIc) may be used with applicable solutions to predict conditions for catastrophic fracture. There have been a few instances where the adequacy of this approach has been verified, but these evaluations were conducted for specific applications. This paper evaluates an approach using specimens containing surface cracks, with substantial variations in crack size and shape, and loaded by either tension or bending. Test results are used to compare the critical stress-intensity factor (Kcrit) with KIc. Moiré interferometry is used to obtain KI at the free surface for comparison with the calculated result.

A study of stress wave propagation around localized crack-tip regions in engineering materials is presented. The principal tool of this initial investigation is dynamic moiré interferometry, which yields full-field continuum boundary conditions of surface displacement in a real-time optical processing manner. The displacement sensitivity employed for this study is 3.3 ?m; however, the displacement sensitivity can be as high as 0.2 ?m. The spatial resolution developed is less than 1 mm. After a brief review of the technique of dynamic moiré interferometry, the study of a longitudinal pulse, interacting with an artificial crack in a finite steel bar, is presented. Equations for the near-tip elastic dynamic crack displacements will be reexamined to extract localized elastic dynamic stress-intensity factors in light of the capabilities of dynamic moiré interferometry.

Experimental methods of analysis and corresponding algorithms for converting data into fracture parameters were reviewed. Results obtained from applying the methods to an analysis of boundary layer effects in compact bending specimens and in moderately deep surface flaws under Mode I loading were presented. Finally, a means for incorporating the new results into linear elastic fracture mechanics (LEFM) design rationale was suggested.

After briefly reviewing an experimental optical technique for estimating crack shapes and corresponding stress intensity factor distributions in three-dimensional cracked body problems, this paper lists some basic observations on the three-dimensional aspects of subcritical flaw growth and supports them by citing results from application of the experimental method. Restrictions of the method are listed. It is concluded that, within the restrictions, the method is useful in supporting numerical analysis.

The applicability of fracture mechanics data generated from standard fracture toughness specimens in predicting structural integrity is often a concern. For elastic-plastic conditions, a second parameter identified as "constraint" (hydrostatic stress divided by the equivalent stress) is being used to link different specimen/flaw configurations, thus providing a mechanism for using data generated from standard fracture toughness specimens to predict structural integrity. Much of the literature is concerned with the difference in constraint between specimens and a structure, while little effort is being given to investigating the sensitivity of the crack initiation process to constraint for a given material. This paper presents experimental results relating crack initiation processes to constraint for a ferritic steel and an aluminum-based alloy.