Crack-arrest tests on A533 Type B steel compact specimens, as small as 100 mm square by 6.4 mm thick, yielded results that agreed with those reported for 200-mm-square by SO-mm-thick specimens of the same steel. The specimens were trans-verse-wedge loaded and contained both a brittle crack-starter weld and side grooves, as recommended on the basis of results obtained in the ASTM Cooperative Program on Crack Arrest Testing.

Preliminary experiments revealed that the load-line displacement measured at the onset of rapid fracture frequently contained a significant plastic component that, if included in the linear-elastic fracture mechanics (LEFM) computation of K, yielded nonconservative toughness values. An incremental load-cycling technique was used to minimize the effect of plastic displacement prior to crack initiation. Following unloading after a given cycle, the specimen was reloaded until the load-point displacement exceeded that of the prior cycle by 5 percent. This procedure was continued until rapid fracture was initiated. Because of strain hardening, the displacement reached during the crack-initiation cycle was almost entirely elastic and could be used to calculate toughness by LEFM with little error. Use of subsize crack-arrest specimens in combination with incremental load cycling is particularly attractive for studying radiation effects in pressure-vessel steels, because of space limitations in reactors.