Two-stage optimization of infinite rotation-freedom façade systems using machine learning surrogate models

Abstract

Increasing the Degrees Of Freedom (DOFs) of Kinetic Façade Systems (KFS) potentially enhances environmental adaptability but presents challenges in mechanical feasibility and optimization complexity due to high-dimensional design spaces. This paper investigates the mechanism design and optimization strategies for multi-DOF KFS, and assesses the performance trade-offs associated with increased motion and control freedom. An Infinite Rotation Freedom (IRF) prototype is proposed and experimentally validated, and a two-stage surrogate-based optimization framework is developed for multi-DOF façade systems by integrating machine learning-based surrogate models with optimization algorithms for both static feature selection and kinetic motion control. Comparative performance analyses demonstrated that the IRF system significantly improves daylight distribution and thermal regulation compared to conventional louvers, with multi-DOF motion enhancing daylight distribution and increased control freedom enabling more precise glare mitigation. These findings highlight the feasibility and environmental advantages of multi-DOF KFS. Future research should address movement continuity issues to improve operational efficiency.