Kinematic draping simulation optimization of a composite B-pillar geometry using particle swarm optimization.

The present work presents an algorithmic approach to determine the optimal starting point for any complex geometry draping processes. The time-efficient Kinematic Draping Simulation (KDS) is used to assess the drapability of a geometry depending on many different starting points. The optimization problem is then solved by applying Particle Swarm Optimization (PSO). The proposed methodology is applied to and validated with complex geometry and a common part of the automobile industry: the B-Pillar geometry. The results show that the PSO algorithm may improve random search up to 78 times. After several experiments, PSO particles have discrete coordinates and are located at optimum global and local regions most of the time, leading to solutions for complex objective functions. The global solution is such that the starting point is located near the geometrical centre of the B-Pillar. The novelty of the work is evident: it uses optimization for a real engineering application, and it draws pattern-related conclusions for other geometries. Experimental results are shown to be consistent with simulation results.