Topology optimization (TO) is a well-established methodology for structural design under user-defined constraints, e.g. minimum volume and maximum stiffness. However, such methods have traditionally been applied to static, deterministic loading, in which modulus, position and direction are known and invariant. This is against the probabilistic load combination used in the structural engineering designs, and entails two important shortcomings. The first one is related to maintenance and reliability: static loading fails to consider naturally occurring uncertainties in the loading process, measurements or regular service; also ignoring (quantitatively) unforeseen phenomena such as vibrations, and the material's behavior is assumed linear isotropic, ignoring fatigue, plasticity and anisotropy in functionally-graded materials. The second one concerns optimality itself: often times, the structure presented as "optimal" in fact over-estimates loading and thus wastes material by exceeding its real needs and/or distributing it poorly throughout the design dominion. In this article, a probabilistic framework is presented: uncertain and pseudo-dynamic loading is introduced to create robust topologies via a reinforced SIMP scheme with embedded penalization addressing fatigue damage, layer direction, mechanical response (traction/compression) and yield limit (von Mises equivalent stress). This computationally efficient framework is applied to various loading scenarios, generating diverse designs for the same volume fraction constraint with improved and more realistic performances. Under the proposed method, if loads are permanent and damage isotropic, the methodology converges to the traditional (deterministic) topological optimization results. Future ramifications of this work are pondered, especially regarding metamaterial design.
Irastorza-Valera et al. (Tue,) studied this question.