Topological constraints in 2D structural topology optimization

摘要: One of the straightforward definitions structural topology optimization is to design optimal distribution holes and detailed shape each hole implicitly in a fixed discretized domain. However, typical numerical instability phenomena optimization, such as checkerboard pattern mesh dependence, all take form an unexpected number result standard density-type methods, SIMP ESO. Typically, indirectly controlled by tuning value radius filter operator during procedure, which choice one most opaque confusing issues for beginner unfamiliar with algorithm. Based on soft-kill bi-directional evolutionary (BESO) method, model proposed this paper allowed maximal designed structure explicitly specified additional constraint. The digital Gauss-Bonnet formula used count whole iteration. A hole-filling method (HFM) also control existence structure. Several 2D examples illustrate that cannot only limit maximum throughout procedure but mitigate dependence. expected provide designers new way tangibly manage achieve better topological characteristics results.