Development and implementation of a micromechanically motivated cohesive zone model for ductile fracture

Gaining popularity after its coupling with the finite element method, cohesive zone modeling has been used extensively to model fracture, especially in delamination problems. Its constitutive relations, i.e.Traction-separation laws, are mostly derived phenomenologically without considering the underlying physical mechanisms of crack initiation and propagation. The approach could potentially be used for ductile fracture as well where the micromechanics of the phenomenon is explained by nucleation, growth and coalescence of pores. In this context, the objective of this work is to develop and implement a physically motivated cohesive zone modeling framework for ductile fracture in metallic materials. In order to accomplish this, a micromechanically motivated traction-separation relation which considers the growth of a physical pore is developed. Tractions are directly represented as a function of pore fraction, and its evolution is driven by separations. The model is implemented as an intrinsic cohesive zone element in a two-dimensional setting. Implementation steps and methodology including the finite element framework are presented in detail for mode-I, mode-II and mixed-mode fracture cases. The derivation of the mixed-mode case leads to a novel yield function representation of tractions and separations, instead of an explicit expression. Hence, an incremental implicit elasto-plastic numerical integration scheme is utilized to solve mixed-mode system of equations required for crack initiation and propagation. The framework is implemented as a user element subroutine in Abaqus (UEL) and the numerical simulations are conducted with compact tension (CT) and single edge notch (SEN) specimens to test the implementation and influence of the micromechanical parameters such as pore size, shape and spacing on the ductile crack initiation and propagation. The work is concluded by presenting an outlook for the capability of the model to predict crack path. © 2022 Elsevier Ltd.