Toward local identification of cohesive zone models using digital image correlation

In recent years, cohesive zone models have been formulated and used to numerically simulate the fracture of solid materials. Cohesive zone models presented in the literature involve a ‘jump’ in the displacement field describing crack onset within a predefined interface network corresponding to interfaces between elements of the finite element (FE) mesh. The introduction of a virtual displacement jump is convenient to numerically manage microcrack or void initiation, growth and coalescence. Until now, the forms of interface laws were mainly chosen semi-empirically in connection with the overall responses of specimens when subjected to standard loadings. In this study, a cohesive zone model identification method is proposed based on the local material behavior derived from kinematical measurements obtained by digital image correlation (DIC). A series of tensile loadings were performed for several damageable elastic-plastic materials on standard tensile specimens. Kinematical data analysis enabled early detection and tracking of the zone where the crack will finally occur. The results of this study highlight the potential of DIC to quantify damage and show how damage assessments can be inserted in cohesive zone model identification.

► Experimental evidence of the cohesive-volumetric decomposition relevance.
► Cohesive law identification directly from overall behavior.
► Detection of loci of crack onset and crack propagation.
► Cohesive law identification procedure on standard dog-bone-like specimens.