Development of a Shape Memory Alloy Spring Network for Shape Control

Abstract

The interest in shape-shifting materials has increased over the past years. A part of this field concerns the shape-shifting of sheets from flat configuration into a double curved configuration, requiring nonuniform strain among the sheet. This functionality can be used for various purposes, in different application fields. For example, shape-changing properties are increasingly used in tangible user interfaces. For activating these sheets, shape memory alloys (SMAs) lend themselves well. That is why multiple shape-shifting sheets that are activated with SMAs already exist. However, a new design is developed and explored in this thesis, in which SMA springs are used as compressive bendable actuators. This is different from the designs that are found in literature, and the distributed low bending stiffness could potentially lead to smoother shapes. The goal of the project is to create a SMA spring network that can deform out-of-plane into a specific shape and to explore possible applications. The idea is to prescribe this shape by prescribing the related nonuniform strain via controlling the springs individually. Through an iterative prototyping process such a network is developed. Flexures are included to impose the out-of-plane bending direction. Antagonist hoses are included to prevent these flexures from interacting with the springs and to draw the network back in plane upon cooling. In the final prototype, the shape of the network can be adjusted via binary activating the springs with Joule heating. More possible future applications of the network are explored through brainstorm sessions. In addition, a finite element model (FEM) is set-up to calculate the resulting shape upon an activation scheme. The important network parameters are tested and integrated in simple elements. The demonstrator can successfully deform out of plane into different shapes upon different activation schemes, although the shape control is low, and the overall shape not very smooth due to excessive local bending of the network struts. This could be diminished by increasing the bending stiffness of the struts. However, this would lead to less overall deflection as well. The control over the strains or the bending should be improved for enhanced shape control. To evaluate the FEM, the prototype is 3Dscanned and compared with the result of the FEM. There is a significant difference in the amount of deformation, which may be due to inaccuracies in the model or in the prototype. The validity of these should be improved before further evaluating the effectiveness of the simplified model.