BELYTSCHKO, Ted
Northwestern University
Evanston - USA
BROCKS, Wolfgang
Institute for Materials Research
GKSS
Geesthacht - Germany
DE BORST, Rene
Delft University of Technology
Delft - The Netherlands
R.deBorst@lr.tudelft.nl
HELLMICH, Christian
Institute for Strength of Materials
Vienna University of Technology
Vienna - Austria
Christian.Hellmich@tuwien.ac.at
INGRAFFEA, Tony
Cornell University
Ithaca - New York - USA
MANG, Herbert
Vienna University of Technology
Vienna - Austria
MOES, Nicolas
GeM Institute
Ecole Centrale de Nantes
Nantes - France
nicolas.moes@ec-nantes.fr
Abstract
Generally, two types of fracture analyses can be distinguished. The most elementary is the computation of fracture properties for a given, stationary crack. Typically, this concerns properties like Stress Intensity Factors (SIFs) or J-integrals. Often, Linear Elastic Fracture Mechanics is used as the underlying theory. With the proper knowledge of these identities, fracture mechanics makes it possible to determine if a crack will propagate, in which direction (although here a number of different hypotheses exist) and, for dynamic problems, at which speed the crack will propagate. Since the stress field is singular at the crack tip in Linear Elastic Fracture Mechanics, computational methods have been developed for capturing this singularity, especially for coarse discretizations.
More difficult is the simulation of crack propagation. Originally, this was almost exclusively done in a finite element context. Two approaches can be distinguished. In the discrete approach a Stress Intensity Factor is computed. On basis of this information it is decided if, and if yes, how much, the crack will propagate. Then, the crack is advanced, a new mesh is generated for the new geometry and the process is repeated. Essentially, this approach is a series of computations for a stationary crack. In the smeared approach, the state of stress and the internal variables at an integration point are considered to be representative for the tributary area of the finite element belonging to this integration point. The discrete crack is replaced by a damaged area. Approaches like this essentially follow a damage mechanics format.
The method of simulating the propagation of a discrete crack by a sequence of Linear Elastic Fracture Mechanics calculations is possible by virtue of the linear nature of the theory. In Nonlinear Fracture Mechanics, as with the application of Cohesive-Zone Models, this no longer holds true and methods must be developed that allow for tracing crack propagation in a nonlinear sense. Promising avenues are meshfree methods (e.g. the Element-Free Galerkin method) and finite element methods that exploit the partition-of-unity property of finite element shape functions, such as XFEM. While contributions can pertain to all novel developments in computational fracture, it is envisaged that a large part of the mini-symposium will be devoted to a discussion of computational methods as discussed above for nonlinear fracture mechanics theories.
Apart from sessions that will focus on advances in computational fracture mechanics, special sessions are organized on structural integrity assessment. In these sessions the role that advanced computational fracture models play for predicting and assessing safety and reliability of components and structures will be discussed. Contributions include modern assessment procedures, numerical codes and their validation and applications to structural problems involving e.g., brittle fracture, ductile rupture, creep and fatigue.
COMPUTATIONAL FRACTURE MECHANICS
COMPUTATIONAL AND EXPERIMENTAL METHODS FOR DAMAGE, FRACTURE AND DEFORMATION OF SOLIDS AND STRUCTURES: RECENT ADVANCES