Critical review of some methods in nonlinear fracture mechanics

Abstract

Virtually all methods in fracture mechanics seem to be dependent on the existence of an autonomous region near the crack edge, a region in which the processes are dependent on the material rather than on the body and loading geometry. This, usually tacit, assumption is questionable for very large scale yielding. In addition, the most common tool to describe the state of the autonomous region, the J-integral, may lose its path-independence at very large scale yielding, and the assumption of the existence of J- R-curves, that are independent of the scale of yielding, is far from obvious. These circumtances, together with the fact that the J- R-curve method is intended for monotone loading, whereas most structures are subjected to more than one load application during their life-time, calls for more reliable and adequate methods within the realm of non-linear fracture mechanics. One method with such aspirations, the J- N-method, previously suggested by the author is briefly discussed.