13th International Conference on Fracture June 16–21, 2013, Beijing, China Advanced remeshing techniques for complex 3D crack propagation Vincent Chiaruttini1,*, Vincent Riolo1,2, Frederic Feyel1 1 Onera, DMSM/MNU, Chatillon, France 2 LMS, Ecole polytechnique, Palaiseau, France * Corresponding author: vincent.chiaruttini@onera.fr Abstract This paper is related to the development of an efficient numerical technique to simulate the propagation of 3D cracks that can evolve inside complex industrial structures due to intense fatigue loading. In this prospect, an adaptive remeshing approach is presented: based on an efficient mesh intersection algorithm, and using robust automatic meshers developed at INRIA, this approach allows to insert almost any kind of cracks in very complex meshes. Associated to a fast and accurate stress intensity factor evaluation, a kinking angle criterion and a fatigue propagation law, such remeshing algorithms allow to simulate complex crack growth phenomena (with coalescing multiple cracks, crack fronts splitting, contact in small deformation, etc.) in the context of linear elastic fracture mechanics. Keywords 3D crack growth, adaptive remeshing, conform crack remeshing, linear elastic fracture mechanics. 1. Introduction The accurate prediction of crack propagation becomes increasingly necessary in a wide range of industrial applications (aeronautic, automotive industry, civil engineering, etc.). Mostly due to a better optimisation and a higher level of complexity, manufacturing requires ever more sophisticated design techniques and precise damage tolerance analysis in order to correctly evaluate lifetime assessment. Critical parts, such as rotors in aircraft engines, are actively investigated for cracks, using non-destructive means of detection. Such parts, if cracked, are usually replaced immediately for safety reason, However, crack detection technologies have limitations, and some very small initiated cracks might not be detected during inspections. It is thus necessary to find a way to correctly define when the next inspection should occur. Modeling how such small cracks propagate due to fatigue loading is believed to be a solution. During the last decades, many new approaches have been developed to efficiently deal with complex 3D crack growth simulations under fatigue loading. Two main families of methods have been highly investigated by the computational mechanical community: on the one hand, PUM based methods that allows the presence of a crack inside a discretized component as a possible discontinuities of the unknown fields (like the famous X-FEM/Levelset method [1] with nonconform crack meshing), on the other hand method that describes the crack as evolving boundaries of the discrete domain (boundary elements based methods [2], finite element methods with remeshing [3], etc.). Concerning the standard finite element approach, recent improvements of meshing and remeshing techniques allow the generation of complex 3D meshes that could be refined in local areas of interest where a crack could propagate. Thus, inserting the discontinuity surface in such meshes was one of the last drawback of those methods. Traditional approaches, usually based on a cracked CSG model, or mesh generation with boolean operations, generally lead to a lack of robustness which make them difficult to deal with complex geometries. -1-
RkJQdWJsaXNoZXIy MjM0NDE=