Fracture toughness comparison between a semielliptical surface crack in a 4PB plate and a through-thickness crack in a 3PB fracture toughness test specimen

Finite-element analyses were conducted for two 4PB plates with different size semielliptical surface cracks. The sizes of the surface cracks were: a/c = 0.357, c/b = 0.175, a/t = 0.125 (smaller surface crack) and a/c = 0.333, c/b = 0.375, a/t = 0.25 (larger surface crack). The dimensions of the plate, b = 101.6 mm and t = 50.8 mm, were the same in both analyses. The results of the analyses were compared to the results of a previously published analysis of a 3PB specimen by Nevalainen & Dodds (1995), with dimensions a/W = 0.5, W/B = 2, and W = 50.8 mm. Differences in the small-scale yielding, the J-Q behaviour and the effective crack-front length (effective specimen thickness in the case of the 3PB specimen) were found between the plate and the 3PB specimen. Two methods were proposed to calculate the fracture toughness of the plate. The experimental verification of the analysis results was carried out by testing the plate specimens with the larger surface crack and the 3PB specimens. The experimental results seemed to support the results of the finite-element analyses, although the test material properties did not match those of the analyses. The ductile-to-brittle transition curve of the test material was determined, and a difference of 29°C in the transition temperature, corresponding to the 100 MPaÖm fracture toughness, was found between the specimen types, the plates behaving in a more ductile way. A combined constraint and thickness correction was applied to the results, and it was found to reduce scatter and bring fracture toughness values to an equally low level.