Session: 01-02-01: Adaptive and Multifunctional Structures 1

Paper Number: 160354

160354 - Shape Reconstruction and Inverse Design of Highly Multi-Stable Structures 

Multi-stable structures have generated significant interest due to their ability to morph between multiple stable states. These states correspond to strain energy minima and are maintained without continuous actuation. Highly shape-tunable structures can be achieved by creating multi-stable structures with a large number of stable states through a periodic arrangement of multi-stable unit cells. The shape tunability of these structures can be leveraged for applications such as morphing wings and tuning the RF properties of antennas. Consequently, efficient and accurate modelling of multi-stable structures to reconstruct their shape is vital.

Current modelling techniques for multi-stable structures, such as analytical and finite element modelling, fail to meet both efficiency and accuracy criteria. Analytical models yield accurate results with limited computation for simple structures but are not viable for complex multi-stable structures. Finite element models, while accurate for complex structures, demand high computational time and cost. These limitations hinder the practical realization of inverse design of highly multi-stable structures, where specific stable states are algorithmically identified to fit prescribed shapes. Therefore, realizing inverse design requires a shape reconstruction method that is both accurate and efficient.

This study presents a novel shape reconstruction method for multi-stable structures that enables inverse design. The structure used in this study, made from a carbon fiber-reinforced epoxy frame and pre-stretched membrane, is created by combining multiple multi-stable unit cells to form a larger multi-stable structure. Shape reconstruction is achieved via state-to-shape conversion through shape approximation and correction. This accurate and efficient shape reconstruction method underpins the realization of inverse design.

State-to-shape conversion uses the known state of the structure to create an approximate shape based on standardized unit cell building blocks. This approximation is further refined using corrections that account for unit cell interactions and boundary conditions. Inverse design of the multi-stable structure is realized by algorithmically determining the stable state that results in the best fit between the reconstructed shape and the prescribed shape. In this study, the state-to-shape conversion and inverse design will be demonstrated on a 5 x 5 unit cell structure. The accuracy of the reconstructed shape will be validated against high-fidelity finite element models to quantify the RMS shape error. Subsequently, inverse design will be applied across several prescribed shapes, with its efficacy quantified by the RMS error between the determined shape and the prescribed shape.

By leveraging the characteristics of multi-stable structures, this study demonstrates and validates a novel shape reconstruction approach that achieves accurate results with limited computation, while also enabling inverse design.

Presenting Author: Enquan Chew Stanford University

Presenting Author Biography: Enquan is a Mechanical Engineering PhD candidate from Stanford University working under the supervision of Prof. Maria Sakovsky. He received his MSc in Advanced Mechanical Engineering from Imperial College London and his BEng (Hons) in Mechanical Engineering from the National University of Singapore. His current research focuses on enabling mechanical intelligence in highly multi-stable structures through shape reconfiguration and sensing.