Cohesive fracture and size effect in concrete

The classical Hillerborg's "fictitious crack" concept, which consists in a relationship between stress and crack opening displacement, is a simple but very effective model for the simulation of mode I fracture of concrete specimens. Quite problematic, however, appears to be its experimental calibration through direct methods (direct tension tests). For this reason, in the recent past, indirect methods have been proposed. One of the most appealing is the "size effect" method in which the ultimate nominal stress (normalized failure load or structural strength) of specimens of the same geometry but of different sizes, are used to identify the first part of the softening curve. Objective of this paper is to present the results of recent numerical analyses and experimental investigations aimed at the evaluation of the effectiveness of this approach.