General Mechanical

General Mechanical

Why do K-Values not converge when refining the mesh with quadratic tetrahedrons?

    • charlesk

      Hello everybody,

      I’m doing fracture analysis on steel structures mostly using the premeshed-crack option in ANSYS workbench (I used the arbitrary crack before and had the same problems). Due to the fact that I have to do many iterations to calculate the crack propagation, modelling every iteration with crack tip elements does not seem feasible. Using the “semi-elliptical crack” is not an option because my crack will propagate close to the border of the steel part later.

      My main problem is this: when refining my mesh around the crack the SIF (K-Values) keep getting better until a certain element size. From then on, the results get worse and worse again. This does not count for the non-fracture-tool-results (stress, strain, deformation, ...)

      I’m modelling my crack by extruding and then deleting a very thin (0,001 mm) semi-elliptical surface from my body. Afterwards I use the pinch command to ignore the small face within the crack. I’m meshing my whole body using quadratic tetrahedrons.

      Right now I’m working with ANSYS 19 R2, I have encountered the same problem in other versions aswell.

      I added some pictures showing the geometry around the crack and the mesh after deformation. I can upload more information if needed.

      Thank you!


Viewing 0 reply threads
  • You must be logged in to reply to this topic.