Sünek hasar tahmini için gözenekli bir plastisite modelinin sayısal uygulaması ve analizi

Ductile damage and fracture are known to be driven by the microvoid nucleation, growth, and coalescence. Porous micromechanical description of the ductile metals led to many phenomenological material models, which are used to predict the damage and fracture in engineering structures. In this thesis, the assessment of a rate-independent porous plasticity model is done through the representative volume element (RVE) calculations. The model is based on the formalism presented in [1] which is implemented as a user material subroutine through a prediction-correction scheme similar to a classical J2 plasticity framework. In this context, RVE's are taken from a periodic array of spherical voids surrounded by an elastoplastic matrix material with isotropic exponential hardening, and they are deformed under a constant triaxial stress state with a displacement controlled method. The implementation of the model and the method of the RVE calculations are explained in detail. Limitations of the original model are discussed, and a heuristics extensions to the constitutive framework is proposed to obtain a better fit between the porous model and the unit cell results in terms of volumetric void growth and equivalent stress-strain relation. Numerical analyses show the possibility of achieving a compact framework with a straightforward implementation that agrees well with the RVE simulations for a wide range of stress triaxiality values. The present framework is compared with the widely used Gurson-Tvergaard-Needleman (GTN) model and the differences are discussed. A simple void coalescence relation is added to this framework to simulate the final failure phase of ductile deformation. Additionally, tension simulations with smooth and blunt notched specimens are performed with the GTN model, the present porous plasticity model, and the Johnson-Cook uncoupled damage model to address the model's performance in a ductile fracture simulation. Results show that the present framework and the GTN model can yield almost identical results in notched simulations in terms of engineering stress-strain response and the porosity evolution. The thesis is concluded with an outlook and possible future improvements.