Mechanical metamaterials achieve extraordinary physical properties through their internal geometric architectures. While traditional metamaterials are limited by fixed designs, multistable reprogrammable metamaterials can modify these properties post-fabrication and morph into different shapes. However, existing tunable metamaterials often require continuous energy input, rely on global volumetric deformations, or necessitate complex actuation networks. In this talk, I will discuss strategies to overcome these barriers.

First, I will introduce a new architecture comprising a locally stable anti-tetrachiral (LSAT) metamaterial sandwiched between two flexible frames. The frames precisely pre-strain the base metamaterial, inducing local bistability and allowing individual cells to snap independently between soft and stiff states. Spatial patterning of cell states changes the global mechanical properties without altering the structure’s macroscopic volume and shape.

Second, I will present a boundary-driven strategy for controlling mechanical memories utilizing transition waves. Boundary phonon-beating excitations efficiently interact with the translational mode of static kinks pinned at localized defects, enabling active control of the transition-wave location and rewriting of non-volatile mechanical states.

Third, we investigate nonlinear fluid–structure interaction as an actuation mechanism for multistable structures. Specifically, we developed a reduced-order potential-flow model of a clamped–clamped bistable plate immersed in a near-wall von Kármán vortex street (KVS), and study the small- and large-amplitude dynamics of the coupled system. To bridge low-fidelity inviscid models and experiments, we are implementing a structure-preserving machine-learning framework based on the metriplectic formalism. I will discuss how we combine the potential-flow model with wind-tunnel data to learn unmodeled viscous drag and turbulence dissipation, yielding an efficient digital twin.

Ultimately, this talk explores strategies to improve reprogrammable metamaterials through new designs, scalable reprogramming methods, and a framework for efficiently studying fluid-responsive multistable systems.