This is a detailed final report on a series of investigations made to determine a basis for the evaluation of the ability of weld flaws to initiate brittle fracture. The report is in four parts dealing successively with:

(1) Brittle fracture mechanics based on the Griffith theory and on Irwin?s strain-energy release rate adaptations,
(2) Static tests on flawed butt welds,
(3) Static and dynamic tests on small butt weld flaws with and without residual stress, and
(4) Static tests on weld flaws in a controlled field of high residual stress.

All welding flaws in selected materials were simulated flaws, varied to represent lack of penetration, porosity, lack of fusion, or sharp internal weld cracks.

The major objective was to examine the effect of given flaws in various environments, in order to determine the environment, essential to initiate brittle fracture under low static stress conditions. Low temperature was generally an essential part of the environment, but, low static stress initiation could not be procured below the nominal yield point unless the static stress was augmented by either a dynamic stress or a high previously-incurred residual stress. The residual stress environment proved to be most significant. Brittle fractures were initiated from short internal cracks with as small as 2000 psi of applied static stress at temperatures in the order of 0°F. If total brittle fracture did not result, arrested fractures occurred from small buried flaws, with the arrested crack forming a potential source of fracture initiation.

Finally, this report emphasizes the important hearing that residual stress has on the brittle fracture problem and the need for extended investigations in brittle fracture mechanics based on strain-energy release rates to furnish a complete engineering basis for flaw evaluation.