J. Mech. Des. Oct 2019, 141(10): 101101
The capabilities of additive manufacturing offer significant potential for revolutionizing existing product development processes by creating products that are rich in shape, material, hierarchical, and functional complexities. However, searching through design solutions in such a multidimensional design space is a challenging task. In this study, the authors propose a holistic approach that applies data-driven methods in successive stages of design search and optimization. More specifically, a two-step surrogate model-based design method is proposed for the embodiment and detailed stages of product design. A Bayesian network classifier is used in embodiment design as the reasoning framework for exploring the design space. A Gaussian process regression model is then used as the evaluation function during optimization, allowing for the exploitation of the design space during the optimization of the detailed design. These models are constructed based on one dataset created using Latin hypercube sampling and then refined using Markov Chain Monte Carlo sampling. This cost-effective data-driven approach is demonstrated by designing a customized ankle brace that has tunable mechanical performance by using a highly stretchable design concept with tailored stiffnesses in different directions.
