| Abstract: |
| In this talk, I will review our recent progress on enabling the ability to answer, within a mechanical context: "For this dynamical application, what is the best nonlinear elastic constitutive response I could choose?" and, given that, "How can I achieve that nonlinearity?" Using an 1D, FPUT-like chain, we conduct an inverse design algorithm to find polynomial nonlinear springs that optimize, separately, the minimization of peak kinetic energy experienced in response to impact, and a prescribed displacement path executed, both evaluated at the opposite end of the chain. We then show how a second inverse design algorithm can be used to identify solid elastic springs that leverage geometric nonlinearity to achieve the targeted spring law. Simulation and experiment results are compared. Future extensions to irreversible contexts, higher dimensions, and mechanical neural networks will be discussed. |
|