An optimization approach for indirect identification of cohesive crack properties

Abstract

A novel computational scheme to deal efficiently and robustly with the solution of a parameter identification problem associated with the cohesive crack model under mode I behavior is proposed. The formulation of this special type of inverse problem takes the form of a challenging optimization problem known in the mathematical programming literature as a Mathematical Program with Equilibrium Constraints (MPEC). The so called “equilibrium constraints” are complementarity constraints involving the orthogonality of two sign-constrained vectors. We propose, in this paper, a smoothing-based optimization algorithm involving data grouping and aggregation. Actual experimental data from both three point bending and wedge splitting tests are used to illustrate applicability of the proposal.