Abstract :
[en] The development of new materials and new manufacturing techniques have experienced a rapid development in the past few decades, and today designers have access to a large set of processes and materials to fabricate their designs. However, the conventional trial-and-error or other empirical design methods have become cumbersome and inefficient, while technological design tools such as topology optimization have become a breakthrough in design.
Topology optimization supports the structural design by generating innovative concepts with high performance to weight ratio. The design tool usually proposes highly complex geometries that are difficult or even impossible to manufacture by conventional manufacturing processes. Fortunately, new Additive Manufacturing (AM) techniques provide a greater free-form freedom and enable the production of highly efficient, yet complex, optimized designs. Nonetheless, even the most advanced AM processes have their technological limitations. For instance, it is difficult to print parts with very small details or very large dimensions. It is also difficult to print parts with insufficient mechanical support during the layer-by-layer deposition process. Likewise, processes that deposit large volumes of material present difficulties related to the deposition path.
This thesis introduces manufacturing constraints in density-based topology optimization in order to improve the manufacturability of the optimized designs. Specifically, geometrical restrictions are addressed aiming at imposing minimum member size, minimum cavity size, maximum member size, minimum separation distance between solid members, and minimum part inclination to reduce the use of sacrificial support material. The minimum size of the parts is imposed through filtering techniques. The maximum size is controlled using local volume restrictions, which are gathered into a single global constraint using aggregation functions. The minimum gap between solid members is also imposed through local volume restrictions, but these are applied in regions whose geometry enables to control the separation distance between parts and not their maximum size. The minimum inclination of the parts is imposed through local constraints that compare the surface slope with a critical baseline. The research is conducted within the density-based topology optimization framework and implemented in open-access codes suitable for solving 2D and 3D large-scale design problems.
The assessed constraints demonstrate the ability to influence directly and indirectly the components manufacturability. For example, it is noted that the maximum size restriction can be used to address some limitations of processes depositing large volumes of material, such as the Wire Arc Additive Manufacturing (WAAM). In addition, the proposed methods show the ability to produce solutions with low amounts of intermediate densities and well-defined surfaces, which facilitates the interpretation and manufacture of the optimized designs. In particular cases, designs may feature such reduced complexity that they can even be manufactured by conventional manufacturing processes.
