Discrete Elastic Rods with Arbitrary Cross Sections

Abstract

In this conference we explore a new computational model for elastic rods that leverages simulation data to reproduce shell and solid-like efects present in rods that break the assumptions of the classical Kirchhoff rod theory, thus presenting a possible improvement avenue to many states-of-the-art techniques. Our approach consists of taking a data set of simulations from both volumetric solids or shells to train a novel high-order polynomial positive-definite energy model for an elastic rod. This new energy increases the range of material behaviors that can be modeled for the rod, thus allowing for a larger range of phenomena to be captured. In order to propose and test this model, we design an experimental pipeline to test the limits of the linear theory of rods and investigate the interface geometries between the Shell-Rod and Volume-Shell cases to observethe effects of a nonlinear material model and a non-elliptical cross-section in the rod deformation. We also investigate the relation between rod curvature and deformation of the cross-section and curvature to introduce a modification on the bending term of the energy. This allows usto reproduce both the asymmetric bending behavior present in thin volumetric solid and shell beams with non-convex cross-sections. Suggestions for further improvements in models and experimental techniques are also given.

Keith Y. Patarroyo

Researcher Engineer on Chemical Evolution, Digital Chemistry and Unconventional Computation

My research interests include Hierarchical Assembly, Chemical Evolution and Material Computation.