Simulating Ductile Fracture in Steel using the Finite Element Method : Comparison of Two Models For Describing Local Instability due to Ductile Fracture.

Abstract

In the shipping industry, there is an increase in the requirement of determining a ship's strength when it comes to collision events. This involves ship-against-ship collisions, but also strength against damage caused by dropped objects, grounding events and collision with rigid objects. There are many methods of determining the damage inflicted to the structure when two objects collide. The most advanced ones make use of numerical methods and particularly the finite element method.

This thesis gives an overview of the theory involved in a ductile failure of an isotropic ductile material such as steel, and explains two different methods of modeling the material behavior related to ductile fracture for use in the finite element method. One model uses the material's true stress/true strain relationship to simulate the structural response due to reduced load-bearing capacity form ductile fracture. The other is a complete fracture model that reduces the load bearing capacity by inflicting damage to the elements used, and is based on the assumed amount of energy it takes to create a crack. The theory behind the two models is explained in this thesis, and material models are developed using a tensile test model in the finite element software package ABAQUS. Then the material models developed are used on a model simulating a steel plate being penetrated by a cone shaped object. The results are compared to earlier material tests done on the same type of structures. Both fracture models are capable of simulating the ductile fracture of a tensile specimen, and no significant differences can be found when monitoring the energy output. When the same material-definition models are used in an analysis of a plate being penetrated, it is however evident that there are differences that are caused by the difference in the way the ductile fracture is simulated. Particularly the effect of reduced stiffness in the elements when using the energy-based fracture model leads to the conclusion that the two methods make the FE-model behave differently when high values of in-plane tensile strain is present.