Learning Forum

[shk1048]

[Imtiaz Gandikota]

What load case did you define for above model? Are you using the PSG algorithm? In general , for unconstrained optimization, the topology is governed by load path of the structure and you may have to include some additional load cases to get a reasonable topology. For hood design, it could be impact load case + some stiffness load cases (bending, torsion, etc) + NVH. Moreover, in above picture, I am assuming you are optimizing material underneath the "skin"? In that case, the structure is not really disconnected, it just added thicker layers in some areas. I would connect the hinge points to the non-design ares (skin of the hood).
One approach to prevent ls-tasc from deleting some regions is to move those elements to non-design part.

[shk1048]

Yes, I am using PSG method. Would it make a difference if I use optimality criteria methods?

[Imtiaz Gandikota]

PSG is more stable compared to optimality criteria. The method is not the cause of the disconnect in the design, unless you are using a very high step size and acceleration factor (in PSG settings) in which case, optimization progresses aggressively and bigger changes to design variables are made in fewer iterations.