Visualizing the complex morphology of fatigue cracks in voxel-based 3D datasets

Abstract

Fatigue cracks are usually characterised by surface sensitive techniques after specimen failure. High resolution micro computed tomography (μCT) based on synchrotron radiation allows the non-destructive visualisation of crack morphology and evaluation of fatigue crack formation/propagation before specimen failure. The visualisation of the complex fracture morphology with characteristic features out of the acquired set of slices is, however, challenging. To obtain a reasonable estimate, two approaches are generally used: the determination of mass centre points in the hollow space and the minimum intensity search in parallel projections. The more sophisticated approach using the elastically deformable contour model, the physical analogy of a rubber band, termed snakes, gives rise to crack morphologies with much less artefacts. The approach was used in the present study for the characterisation of fatigue cracks in poly(methylmethacrylate) (PMMA) and a dental ceramic. The search for the appropriate snake parameters works much better for homogeneous materials, here PMMA, than for inhomogeneous materials, here a dental ceramic. For the ceramic, the regions where the snakes approach provided reasonable results were restricted. Combining μCT with sophisticated computer vision techniques enables the unique characterisation of cracks at the micrometre scale.