Fast and accurate computation of interactions between linear fiber segments

Abstract

PurposeFiber networks represent a vast class of materials, which can be modeled by representing its microstructure using one-dimensional fiber embedded in three-dimensional space. Investigating the statics, dynamics and thermodynamics of such structures from computational first principles requires the efficient estimation of cohesive-repulsive energies and forces between interacting fiber segments. This study offers a fast, efficient and effective computational methodology to estimate such interactions which can be coupled with Hamiltonian mechanics to simulate the behavior of fibrous systems.Design/methodology/approachThis method preserves the uniformly continuous distribution of particles on the line segments and utilizes adaptive numerical integration of relevant distance-distribution functions to estimate the effective interaction energy and forces.FindingsThis method is found to be cheaper to compute and more accurate than the corresponding discrete scheme. This scheme is also versatile in the sense that any pair-wise interaction model can be used.Research limitations/implicationsThe scheme depends on the availability of a suitable pair-interaction potential, such as a Lennard-Jones potential or Morse potential. Additionally, it can only be used for systems which are purely fibrous in nature. For example, fiber composites with a non-fibrous matrix are not addressed.Practical implicationsPaper, woven and hair can be represented as purely fibrous at some relevant length scales and are thus excellent candidate systems for this scheme.Originality/valueThis paper presents a novel method which allows rapid and accurate implementation of an otherwise computationally expensive process.