Topological Optimization not Joining SPC with Load Points
Hi, I'm trying to perform a topology optimization on a defined solid element volume, already tried making bigger and smaller elements, changed the Load formulation, changed the SPC, played around with the optimization parameters, with minsize and max size too, have tried everything but I'm only getting spheres around the SPC and Load points and it always says the solution has converged and it met all the constraints placed. Can someone help me how to make the solver stop doing this and propose a real geometry?
Attached is a picture of the result and the model if someone can check it out and tell me where is my mistake
Answers
-
Looks like for these loads, the usage of steel is "over".
Looks like the solver is trying to keep everything at a minimum density, as if it would give you a foam equivalent material to withstand these loads.
Frequency is way above the target of 21Hz, and also displacements seems to be quite below targets.
One remark, usually using 2nd order mesh for topology optimization doesn't help so much, as the elements are the 'pixels' of your optimization, and maybe using 1st order elements would enable you to refine more the mesh and get finer definition.
But this wouldnt change your results.
Maybe you'll need to add a minimum mass constraint to avoid the "trivial" solution of all 0.01 densities (minimum) all around. I know this goes against minimizing the mass idea, but as the loads are too small apparently, then adding some minimum mateiral limit would be helpful.
1 -
Adriano A. Koga_21884 said:
Looks like for these loads, the usage of steel is "over".
Looks like the solver is trying to keep everything at a minimum density, as if it would give you a foam equivalent material to withstand these loads.
Frequency is way above the target of 21Hz, and also displacements seems to be quite below targets.
One remark, usually using 2nd order mesh for topology optimization doesn't help so much, as the elements are the 'pixels' of your optimization, and maybe using 1st order elements would enable you to refine more the mesh and get finer definition.
But this wouldnt change your results.
Maybe you'll need to add a minimum mass constraint to avoid the "trivial" solution of all 0.01 densities (minimum) all around. I know this goes against minimizing the mass idea, but as the loads are too small apparently, then adding some minimum mateiral limit would be helpful.
Many Thanks Adriano, I'll try changing the material then to something weaker maybe.
0 -
Rodrigo Padilla_21352 said:
Many Thanks Adriano, I'll try changing the material then to something weaker maybe.
also you could consider running a minimize weighted compliance problem, with constrained volume fraction, and then run with 0.1, 0.2, 0.3 ... to check different results with different volume fractions (fixed material volume).
0
Altair Employee