MWicke
Mechanical
- Feb 25, 2016
- 5
Dear all,
I want to investigate the role of casting pore morphology and loading direction on the stress concentration factor. The pore geometry is digitalised using a X-ray computed tomography and after a reverse enginnering process imported in Abaqus. The pore is first insert as a cavity in a finite element model of a cylindrical volume to simulate a pull test. In order to investigate the above mentioned effects on the stress concentration factor, rotation angles about the coordinate axes are defined and the pore is rotated with respect to the angles.
So a new FE model and mesh is generated each time using python scripts to be then solved. In my opinion the remeshing each time is both time-consuming and suboptimal. It is not possible to perfom mesh convergence studies for a few hundert configurations to ensure mesh convergence each time. After running the simulations I had to realise, that the mesh and so the stress distribution was not optimal in some cases.
My idea was therefore to use a submodel for this purpose. The global model consists of the pore in a sphere, which are insert in a cylindrical volume (compare added picture for illustration). After analysing the global model one time, the faces of the cylindrical volume are deleted to create the submodel. To investigate the role of pore morphology and loading direction, the submodel is then rotated with respect to the defined angles. This means, that no remeshing is necessary.
My question is now, if my approach is plausibel or if you would suggest another procedure to analyse the pore geometry.
Thanks in advance and best regards
Marcel
I want to investigate the role of casting pore morphology and loading direction on the stress concentration factor. The pore geometry is digitalised using a X-ray computed tomography and after a reverse enginnering process imported in Abaqus. The pore is first insert as a cavity in a finite element model of a cylindrical volume to simulate a pull test. In order to investigate the above mentioned effects on the stress concentration factor, rotation angles about the coordinate axes are defined and the pore is rotated with respect to the angles.
So a new FE model and mesh is generated each time using python scripts to be then solved. In my opinion the remeshing each time is both time-consuming and suboptimal. It is not possible to perfom mesh convergence studies for a few hundert configurations to ensure mesh convergence each time. After running the simulations I had to realise, that the mesh and so the stress distribution was not optimal in some cases.
My idea was therefore to use a submodel for this purpose. The global model consists of the pore in a sphere, which are insert in a cylindrical volume (compare added picture for illustration). After analysing the global model one time, the faces of the cylindrical volume are deleted to create the submodel. To investigate the role of pore morphology and loading direction, the submodel is then rotated with respect to the defined angles. This means, that no remeshing is necessary.
My question is now, if my approach is plausibel or if you would suggest another procedure to analyse the pore geometry.
Thanks in advance and best regards
Marcel