A Comparison of Notch Tests and Notch-Brittleness Criteria.

Abstract The problem of brittle failure in welded steel storage vessels has many features in common with the failure of welded ships, about which a great deal has been published in recent months. The increasing tendency for a steel structure to fail in a brittle fashion as the temperature is lowered, is known to depend on such factors as the stress distribution at the point of failure, geometrical constraints inherent in the design, the composition, melting and rolling history of the material, and the forming, welding and heat treatment practices employed during fabrication. The precise effects of these factors are individually uncertain and the problem is made more complex by the many variables involved.Notched test specimens are commonly used to obtain a transition from shear to brittle, or cleavage, failure at temperatures in the region of atmospheric temperature, since it has been evident from service failures that this transition, under service conditions, may occur at temperatures at least as high as atmospheric. The manner in which a notch alters the stress condition and so raises the transition temperature in a test-piece has been demonstrated by Gensamerl This principle has been applied to tension, bend, tear and impact tests, but, while variations in each of these forms are currently in use, the only notched tests to be widely used up to the present time have been the notched bar impact tests of the Charpy and Izod types.