Fundaments for developing an automated structural optimization tool based on strain energy

Design processes are sets of complex, iterative, and multidisciplinary procedures applied to achieve articular solutions. To aid in the understanding of design processes and to support the involved actors, a "Research Engine" (RE) is being developed. The RE is a set of computational processes which work together, in a cyclic manner, to simulate the interaction between spatial design instances and structural solutions within a design process. This paper demonstrates how the RE can generate different structural solutions from a single spatial design and produces relevant information about their structural behavior. This information is then used to (1) lower the total amount of strain energy in a structure and (2) achieve a more homogeneous distribution of strain energy among the structural elements. A set of case studies are presented that have been optimized manually and the findings of these optimizations will be used to suggest underlying principles for an automated optimization tool to be developed. Results show that by adding and relocating structural elements to highly strained sections of the structure and by exchanging the type of structural elements (e.g. walls for columns) both of the previously stated objectives can be achieved. As such, these procedures (addition, relocation, and exchange of structural elements) will be put forward as fundamental rules of the upcoming automated optimization tool.