| description abstract | Abstract. 4D Printing, the fabrication of stimulus-responsive systems using Additive Manufacturing (AM), offers significant potential to create active structures and systems that require little or no assembly. 4D Printed Architected Active Composites (4DPAACs), for example, combine two constituent materials into a composite whose properties can differ significantly from those of the constituent materials and can be used to create 4D printed actuators. However, the complex relations between low-level design and system behavior, as well as the tight coupling among geometric, material, and AM processing parameters, make it difficult for designers to exploit this potential. This article proposes a novel method that enables designers to interactively explore the feasible design space offered by 4DPAACs and create designs with targeted properties. The proposed method defines a library of 4DPAAC unit cell architectures and Artificial Neural Networks (ANNs) that map unit cell designs to 4DPAAC behavioral properties. When a designer provides information about the available AM process and materials, the ANNs are used to create design charts, within less than one second, that visualize the feasible properties and can be used to identify unit cell designs that meet target properties as well as explore design tradeoffs. The method's potential is demonstrated on three case studies using different AM technologies and active materials. By providing quickly accessible design space information, the presented framework fills a crucial gap by providing a method to quantitatively explore design spaces in early design phases, making the design possibilities offered by 4D printing and 4DPAACs more accessible. | |
