Optimization-driven conceptual design of truss structures in a parametric modelling environment

Structural optimization methods can be extremely powerful when used at the initial, conceptual, design stage of a building or bridge structure, potentially identifying materially efficient forms that are beyond the imagination of a human designer. This is particularly important at present, given the pressing need to reduce the carbon footprint associated with the built environment in the face of the current climate emergency. In this contribution, a computationally efficient global–local optimization framework is proposed, in which a linear programming-based truss layout optimization step is employed to generate initial (near-)optimal designs, with a non-linear optimization step then used to generate designs that take account of real-world complexity. To facilitate rapid exploration of design concepts, the proposed global–local optimization framework has been made available in the Peregrine plug-in for the popular Rhino-Grasshopper parametric modelling environment. The efficacy of the approach is demonstrated through its application to a range of case study problems.