On an Adaptive Mask Overlay Topology Synthesis Method

摘要

A stochastic topology design approach is presented that yields binary, well connected continua. Inspired by the well known photolithographic technique used in the fabrication of micro-components, a number of negative-masks are appropriately laid over the design region to simulate voids. A unique feature is the effective use of the masks. In addition to their position and sizes, the number of circular masks is adaptively determined in each step of the optimization process. Thus, not only the void shapes but also their number is varied. The proposed method is significantly efficient compared to the previous implementations [21] and [23] and requires much less computational effort to yield good solutions. The honeycomb parameterization employed eliminates all sub-region connectivity anomalies by ensuring edge connectivity throughout. Boundary smoothening is performed as a preprocessing step to moderate the notches, and to obtain an honest evaluation of a candidate design. Thus, both material and contour boundary interpretation steps are no longer required when post-processing the synthesized solutions. Various features of the method are demonstrated through the synthesis examples of small deformation compliant mechanisms.