Dear All,
I am trying to converge on an optimal solution for a compliance minimization problem using a parametric RBF-based level set method. However, I have faced the following problem(s) for different input files. The blue line is the final area which is supposed to be 5 out of 10 at the end (50% target volume). The black line is the lagrangian which plays a crucial role in convergence. Also, the finite element mesh and the level-set grid are built separately, 50*20 and 80*45, respectively for the following video. Each finite element consists of 9 Gauss Points, and I use an Inverse Distant Weighting interpolation algorithm. I guess the issue is rooted in the way I am normalizing two antagonistic terms in the following equation, V = Energy-Lagrange, where V defines the normal velocity of the surface at each point. I would be happy if someone could help me in solving this issue. Thank you well in advance.
Best regards,
Farzad Tatar