Additive manufacturing enables the nearly uncompromised production of optimized topologies. However, due to the overhang limitation, some designs require a large number of supporting structures to enable manufacturing. Because these supports are costly to build and difficult to remove, it is desirable to find alternative designs that do not require support. In this work, a filter is presented that suppresses non-manufacturable regions within the topology optimization loop, resulting in designs that can be manufactured without the need for supports. The filter is based on front propagation, can be evaluated efficiently, and adjoint sensitivities are calculated with almost no additional computational cost. The filter can be applied also to unstructured meshes and the permissible degree of overhang can be freely chosen. The method is demonstrated on several compliance minimization problems in which its computational efficiency and flexibility are shown. The current applications are in 2D, and the proposed method is readily extensible to 3D.
This paper presents an innovative optimisation method for aircraft fuselage structural design. Detailed local finite element analyses of panel buckling are further processed such that they can be applied as failure constraints in the global level optimisation. The high computational costs involved with the finite element analyses are limited by advanced use of surrogate modelling methods. This yields high flexibility and efficiency in the local level optimisation procedure and allows for efficient gradient based search methods as well as more costly direct search optimisations like genetic algorithms (GAs). The method is demonstrated on a composite fuselage barrel design case considering common structural sizing variables like thicknesses and stringer dimensions. Optimised barrel designs are obtained where the constraints that are derived from the panel buckling analyses are active. The total computational cost for the complete local and global level optimisation procedures is in the order of days on common-performance hardware.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.