Interactive structural topology optimization via skeleton control

Innovative load-bearing designs depend on a careful balance between creative shape and structural performance. While existing preference-based topology optimization methods let designers add subjective preference, they pose high design experience barriers and suffer from inefficient interactions due to reliance on skilled manual sketching. In this study, we propose the bidirectional evolutionary structural optimization with skeleton control (SC-BESO) method. This method begins with extracting the skeleton graph of an existing design, where users can subjectively drag its skeletal nodes to deform the design. The deformed design is then used to guide the generation of its optimized structural design. A design parameter allows users to adjust the optimized result toward either performance-driven or preference-driven outcomes. Based on the proposed method, a cross-platform digital design tool, TouchBESO, is developed for 2D structural designs. Parametric studies show that, even after user interaction that results in stiffness reduction, the SC-BESO method can regain most of the lost performance with only small shape changes. Furthermore, a 3D cantilever example illustrates the scalability of the method, and two potential integrations with drawing systems and the similarity constraint are discussed for future interactive design frameworks. • A topology optimization method with skeleton control is proposed. • Users can intuitively shape structures while keeping high structural performance. • The proposed method can control the importance of preferences on optimized designs. • A digital design tool integrating the proposed method is developed for 2D design.