ABSTRACT
An important question was risen in Castillo et al. (Castillo et al., 2007c): what is damage? Or, equivalently how is damage related to failure? A physical concept of failure, such as crack size, seems to be adequate to define a service limit state, but fatigue failure, as an ultimate limit state, requires a probabilistic framework allowing relating damage levels to probabilities of occurrence. The solution to these question is one of the principal objectives of this paper. Furthermore, several methods to measure damage will be analyzed and discuss to identify which is the best one and which of them are inadequate for measuring damage.
