Some Effects of Specimen Geometry on Crack Propagation and Arrest

Abstract

Fast crack propagation and arrest have been studied in steel specimens of the double cantilever beam type using three different geometries and either a fixed grip or compliant loading system. A finite-difference computer program was used to measure values of the dynamic stress-intensity factor KD. The relationship between values of KD and KS, the stress-intensity factor for a quasi-statically growing crack depended on the specimen geometry and loading system used. Stable crack arrest was easier to obtain in the specimens loaded with fixed grips. This probably resulted from the rigidity of the loading system, which produced specimen oscillations after arrest of insufficient magnitude to reinitiate crack growth. An examination of the energy associated with a moving crack suggests that kinetic energy may be reabsorbed to aid further crack propagation but is not all reabsorbed at arrest. The remaining fraction causes oscillations of the cracked specimen halves which may reinitiate the crack, or be damped out as the stress intensity rings down to its static value. Examination of the data suggests that intermediate arrests, characterized by a minimum in the crack length velocity curve, were a common feature of all tests. The stress intensity for crack arrest appeared to depend on specimen geometry.