Localizing Implicit Gradient Damage Based Modelling of Quasi-Brittle Failure With Non-Planar Crack

Abstract

Localizing implicit gradient damage (LIGD) is a gradient extended model which is equipped with a decreasing internal length scale with damage evolution, Poh and Sun (2017). The model is thermodynamically consistent and resolves the well-known problems of conventional implicit gradient damage (CIGD) model such as artificial diffusion of damage and erroneous predictions of failure initiation and propagation directions. So far, the effectiveness of the model has been demonstrated for two-dimensional quasi-brittle and three-dimensional ductile failure predictions with flat fracture surfaces. It is the aim of this contribution to assess the predictive capabilities of the model for three-dimensional quasi-brittle failures with non-planar cracks. To this end, localizing implicit gradient model is embedded within a tetrahedral element formulation and implemented in commercial finite element package Abaqus through user element (UEL) subroutine. Skew notched prismatic torsion test is modeled and capabilities of the model are assessed in terms of reaction force-displacement curves as well as the resulting crack surfaces, Brokenshire (1996), Jefferson et al. (2004). Comparison of LIGD and CIGD predictions suggest that LIGD is superior to CIGD. Furthermore, as far as capturing the experimental results is concerned, it performs as good as other alternative modeling frameworks, e.g., mixed finite element formulations. © 2024 The Author(s).