Search

Multi-material topology optimization has become a popular design optimization discipline since it allows to go a step further in topology optimization of the design of lightweight components and structures. However, it also presents additional challenges in terms of managing a higher complexity in the optimization problem, reliably estimating the manufacturing cost taking into account the cost of joining dissimilar materials, or assessing the manufacturability of the design. This paper proposes a set of methods that solve generic multi-material topology optimization problems, while including several novel aspects such as a comprehensive cost model, specific design rules for multi-material design, and model order reduction techniques to improve the computational efficiency. Two different examples, consisting of a sitting bench and a car cowling, have been solved in order to support the benefits of the approach.

When designing a product, care must be taken to ensure that the product can be assembled later on at the manufacturing floor. This paper presents a methodology that takes in a CAD design to check its ease of assembly. The proposed methodology can calculate feasibility scores for an assembly step, then use these scores to calculate the optimal sequence of such assembly steps to construct the end product from its loose parts. Two assembly step scoring rules are presented: local and global assembly feasibility. Local assembly feasibility determines the directions in which two subassemblies can move apart an infinitesimally small distance based on contact faces between the subassemblies. Global assembly feasibility considers whether this small distance can be extended to infinity, by checking the overlap of projections of the pair of subassemblies in locally feasible directions. For both of these rules, problems are reduced to previously calculated scores whenever possible, in order to reduce calculation time. These rules are integrated in an algorithm which generates a complete assembly sequence. This assembly sequence is guaranteed to be the global optimum for the given rules. The algorithm successfully generates an assembly sequence for a small example case; the simplification rules reduced the number of local and global assembly feasibility checks by a factor 10 and 6, respectively.