Discrete Kirchhoff Rod Networks for Optimization-Driven Design

Many examples from architecture, mechanical engineering, and material science can be described as networks of elastic rods.
Designing rod networks that exhibit desired characteristics is made difficult by the fact that the mapping between design parameters and performance at equilibrium is highly nonlinear. We propose a computational approach to this problem that leverages simulation and optimization algorithms for rapid design exploration. The technical core of our approach is formed by an efficient computational model for networks of discrete elastic Kirchhoff rods, paired with an inverse problem solver based on sensitivity analysis. We validate our model against standard solid finite element simulations and present applications to compliant mechanisms, deployable pretensioned membranes, structural curve networks, and mechanical metamaterials.