Testing some implementations of a cohesive-zone model at finite strain

Abstract This study shows how the results given by a cohesive-zone model at finite strain may depend strongly on its numerical implementation. A two-dimensional four-node cohesive element is considered, which includes several variants depending on a part of the strain–displacement matrix, on the quadrature rule applied, and on the configuration chosen to perform integration. Finite element simulations combine these variants with a very simple, bilinear, cohesive-zone model, in two tests. The first test involves a single element and illustrates some features of the various implementations. The other test simulates the peeling of an elastomer strip from a rigid substrate.