| description abstract | The quantitative translation of physical weld quality into structural integrity prediction depends on accurate characterization of weld material behavior in the presence of fabrication defects. The presence of such defects will, however, significantly influence the response of common material test specimens. If the influence of such defects is fully understood, test specimen data may be interpreted in a more meaningful way. The role of a physically relevant geometric imperfection, in the form of a spherical void defect, on cylindrical tensile specimen response is computationally simulated for HY-100 weld metal. Defect radius and location along the specimen axis are treated as independent parameters. Asymmetry of specimen deformation (in terms of specimen neck location) and specimen ductility (in terms of the reduction of area at failure) are computationally predicted. Results suggest that the neck location does not necessarily coincide with the defect location. Therefore, geometric defects are a sufficient condition for asymmetry of neck location but not a necessary condition for neck formation. In addition, coincidence of the defect and the neck reduces the specimen ductility at failure to a minimum value which depends on defect size. When the defect and neck are separated, the defect free specimen ductility at failure, i.e., the maximum ductility value, is recovered as an upper bound. The transition between these two ductility values is abrupt, despite the continuous nature of the physical problem. Preliminary implications of these results on the assessment of defect criticality are discussed. | |
