Research

Mesomodeling of failure in laminates

My PhD project at the TU Delft has been on the simulation of failure in laminates. The main innovation was to use X-FEM to prescribe the direction of crack propagation for matrix cracking. This way, it was ensured that matrix cracks that are straight in reality, are also straight in the simulations. This was then combined with cohesive elements for delamination and continuum damage for fiber failure to simulate complex failure mechanisms in which these different processes interact. Much attention was paid to the robustness of the algorithmic framework in order to avoid convergence problems.

Mesomodeling of failure in laminates

A level set model for delamination

One of the limitations in mesolevel modeling of laminates is that the cohesive elements for delamination need to be very small. If elements are not chosen several times smaller than the cohesive band (which is of the order of a millimeter), inaccuracy and severe robustness issues are inevitable. This limits the specimen size that can be analyzed in implicit analysis within a computation time of a couple of days to the order of 10 cm. To avoid this, I have worked in collaboration with the Ecole Centrale de Nantes on the development of a method in which the crack front is represented as a line, rather than as a band. For this purpose, level sets are employed, in combination with a fracture mechanics approach to accommodate crack growth.

A level set model for delamination

Micromechanical modeling of delamination

In 2012 I received a personal grant in the Veni scheme from the Netherlands Organisation for Scientific Research (NWO)  on the micromechanical modeling of delamination. I am developing modeling strategies to simulate the fracture process that constitutes delamination. I have worked with interelement cohesive zones and with the thick level set method for detailed modeling of cusp formation in mode II delamination.

Micromechanical modeling of mode II delamination

Multiscale modeling

I am active in the development of multiscale modeling techniques. Embedding micromechanical models in a multiscale (FE2) framework allows for detailed representation of the physics of nonlinear processes in composite material. Open topics are multiphysics modeling and modeling of fracture.

Micromechanics model for use in multiscale approach

Fatigue

I am involved in several projects that concern the modeling of crack growth under cyclic loading conditions, or fatigue. Level set based methods prove useful for implementing experimental relations between energy release rate and crack growth rate (Paris' relation). In metals, the challenge is to include additional ingredients to capture observed phenomena like crack retardation and corrosion fatigue.

Interfacial thick level set method for fatigue crack growth

Time-dependent shape functions

For my MSc thesis I developed a finite element method with shape functions that change during the analysis. A priori knowledge about the solution was inserted in the solution space to allow for high accuracy with a low number of degrees of freedom. The special functions evolved in time for optimal performance.