Optimisation of a Cube
Hello,
Im a student.
I am trying to realize a topological optimization of a cube. the goal is volume / mass optimization.
I have a cube (my design space) with a hole that acts as a tapping, a 1mm thick plate that is connected to my cube and is encased (the first boundary condition); My cube is connected to a 2D plate by rigid elements of type RBE2. This 2D plate undergoes a force of 40 on one side (I do not want more to stay in the elastic domain) and is encased on the other side ( second boundary condition )
I impose optimization conditions on the RBE2 rigid point moving along the X, Y, Z axes.
My problem is that no matter what I use, I do not get any shape ... all I have is semi-dense area (density between 0.5-0.7) and the closer I get to 1 and the more I get nothing, and the reduction in volume is 95% 93%, which seems really huge.
At first, I tried to increase the moving constraints of my optimization problem because I had constraint violations when I parsed the .out file. Now I have no more constraint violations but still have a decent topology ... I'm stuck.
I do not know if I have badly defined my optimization problem, if my geometry is not good ( i dont think that can be the problem ), or if my optimization answer is not the right one.
Answers
-
Hi @Ludo045
Try adding volume fraction response and compliance response along with the existing. Make compliance as the objective and check the results.
0 -
Thank you for your answer.
So i tried to minimize my compliance response as the objective to increase the stiffness but i fail to understand why the use of the volume fraction response since it doesnt intervene in the compliance formula. Plus it seems to get the kind of result using volume fraction or not, or even if the volume fraction is 0.5, 0.6, 0.7, 0.8 as upper or lower bounds.
Is there any mathematical or numerical reason why when we want to reduce mass/volume we use the volume fraction as constraint ?
0 -
Hi,
Well there is no mathematical reason, but until the constraint is satisfied mass will be removed from the part. In order to restrict that I suggested using volume fraction.
0