Parametric Sensitivities of XFEM Based Prognosis for Quasi-static Tensile Crack Growth
Understanding failure mechanics of mechanical equipment is one of the most important aspects of structural and aerospace engineering. Crack growth being one of the major forms of failure in structural components has been studied for several decades to achieve greater reliability and guarantee higher safety standards.
Conventional approaches using the finite element framework provides accurate solutions, yet they require extremely complicated numerical approaches or highly fine mesh densities which is computationally expensive and yet suffers from several numerical instabilities such as element entanglement or overly soften element behavior. The eXtended Finite Element Method (XFEM) is a relatively recent concept developed for modeling geometric discontinuities and singularities by introducing the addition of new terms to the classical shape functions in order to allow the finite element formulation to remain the same. XFEM does not require the necessity of computationally expensive numerical schemes such as active remeshing and allows for easier crack representation.
In this work, verifies the validity of this new concept for quasi-static crack growth in tension with Abaqus' XFEM is employed. In the course of the work, the effect of various parameters that are involved in the modelling of the crack are parametrically analyzed.
The load-displacement data and crack growth were used as the comparison criterion. It was found that XFEM is unable to accurately represent crack growth in the models in the elastic region without direct manipulation of the material properties. The crack growth in the plastic region is found to be affected by certain parameters allowing us to tailor the model to a small degree. This thesis attempts to provide a greater understanding into the parametric dependencies of XFEM crack growth.