Numerical Simulation and Experimental Investigation of the Fracture Behaviour of an Electron Beam Welded Steel Joint

The 2D Rousselier model, the Gurson-Tvergaard-Needleman (GTN) model and the cohesive zone model (CZM) wereadopted to predict the crack propagation of thick compact tension (CT)specimens.

In this thesis, the author investigates experimentally and numericallythe fracture behavior of an electron beam welded joint made fromtwo butt S355 plates. The 2D Rousselier model, the Gurson-Tvergaard-Needleman (GTN) model and the cohesive zone model (CZM) wereadopted to predict the crack propagation of thick compact tension (CT)specimens. Advantages and disadvantages of the three mentioned modelsare discussed. The cohesive zone model is suggested as it is easy to usefor scientists & engineers because the CZM has less model parametersand can be used to simulate arbitrary crack propagation. The resultsshown in this thesis help to evaluate the fracture behavior of a metallicmaterial. A 3D optical deformation measurement system (ARAMIS) andthe synchrotron radiation-computed laminography (SRCL) techniquereveal for the first time the damage evolution on the surfaceof the sampleand inside a thin sheet specimen obtained from steel S355. Damageevolution by void initiation, growth and coalescence are visualized in2D and 3D laminographic images. Two fracture types, i.e., a flat crackpropagation originated from void initiation, growth and coalescenceand a shear coalescence mechanism are visualized in 2D and 3D imagesof laminographic data, showing the complexity of real fracture. Inthe dissertation, the 3D Rousselier model is applied for the first timesuccessfully to predict different microcrack shapes before shear cracksarise by defining the finite elements in front of the initial notch withinhomogeneous f0-values. The influence of the distribution of inclusionson the fracture shape is also discussed. For the analyzed material, ahomogeneous distribution of particles in the material provides thehighest resistance to fracture.