there are few elements misbehaving and deforming more than 10000 times bigger than logical model does, there are only 3 of them but as they are very hard to locate i have no luck in deleting them, and it is a converted model from shell to solid so choices are not vast. any idea on what i should do?
Hallo Mr. Ghoroubi,
Elements deforming more than 10000 times bigger tells the following:
- Non-physical loading,
- inappropriate BCs.
''there are only 3 of them but as they are very hard to locate i have no luck in deleting them'' What do you mean by that?
if you know the element number( which is easy to know, either checking .dat file or .msg file or abaqus viewer (distorted elements), then you can delete the element from .inp file, deleting element should be carefully done because you lose stifnees and you may face numerical instability. i would check the reason why the are distorted as previously discreibed and i would correct the jacobian, sweep, and other elements controls.
if you use atomatic solid mesh you need to check elements quality before submetting the job or try to improve the shell mesh to have a better control on solid elements.
hope i was able to give a help.
Warm Regards,
Tamer Qaimari
dear mr.qaimari, thank you for your answer,
there is no chance to edit the mesh by hand(orpahn mesh from micro-ct scan), only possiblity is deleting a certain element,
by locating i mean i have means of deleting mesh by hand, but this is based on seeing the element and i have edited them before, but in this case there is a internal element causing the problem, not to be solved with my conventional methods.
you gave me a nice idea thank you for that and for your time and effort.
might be interesting for people to know that :
""""""
checking .dat file or .msg file or abaqus viewer (distorted elements), then you can delete the element from .inp file, deleting element should be carefully done because you lose stifnees and you may face numerical instability
""""""
worked for the solution of this problem.
Dear Nima,
I have done quite a bit of work on image-based simulations and continue working in this area. Please take a look at my publications, specifically, this one:Processing of fiber architecture data for finite element modeling of 3D woven composites
Processing of fiber architecture data for finite element modeling of 3D woven composites
In this work, we developed several approaches for mesh repair. There are several commercial software packages for meshing imaging data (AMIRA, ScanIP, etc.) which focus on using volumetirc imaging data for generating congruent meshes. If you do not have access to imaging data anymore, it makes this quite difficult. However, you still have an option to do through the data using any of the open-source meshing (Gmsh and alike) packages to identify and maybe manually correct the mesh. If it is surface mesh, MeshLab has really nice tools for this task. If you have access to HyperMesh, it also has some nice tools for mesh editing/repair.
Best of luck!
- Badges
- Science topic
- Similar topics
- Engineering
- Mechanical Engineering