Abstract The initiation and growth of cracks in smooth 2024-T3 aluminum specimens was measured with acetate replicas. Once a small crack had initiated naturally, two tiny indentations were placed across it; when they were illuminated with a laser, the resulting fringes could be monitored to determine the crack opening displacement as a function of load. The crack opening displacements were measured at several positions along the crack as it grew across the thickness of the specimen and became a through-the-thickness crack a few millimeters long. An extensive series of tests was run for R-ratios of 0.5, 0.0 and −1.0. A ‘ standard ’ location for measuring the opening load, Pop, was taken as 50 μm behind the tip of the crack. When the da/dn vs ΔK data was adjusted to account for the closure effects using the measured values of Pop/Pmax (i.e. computing a ΔKeff), the effect of the R-ratio was essentially removed. When ΔKeff was computed using Newman's model, the agreement with the measured values of da/dn was excellent; i.e. that model predicts the closure loads quite well. Even after these adjustments, the small cracks still grow slightly faster than would be predicted by long-crack data, indicating that plasticity-induced closure is not the complete answer.
W. Sharpe, X. Su
Engineering Fracture Mechanics