Radial Basis Function (RBF)-based Parametric Models for Closed and Open Curves within the Method of Regularized Stokeslets

TL;DR

This paper explores RBF-based parametric models for representing elastic curves in fluid flow simulations using the Method of Regularized Stokeslets, demonstrating their accuracy and efficiency for open and closed structures.

Contribution

It introduces RBF and SBF parametric models within MRS for elastic structures, comparing their performance to existing methods for open planar curves.

Findings

RBF and SBF interpolants are competitive with Lagrange-Chebyshev for smooth open curves.

Clustered nodes improve interpolation accuracy for derivatives.

RBF-Stokeslets effectively simulate elastic structures in viscous fluid flow.

Abstract

The method of regularized Stokeslets (MRS) is a numerical approach using regularized fundamental solutions to compute the flow due to an object in a viscous fluid where inertial effects can be neglected. The elastic object is represented as a Lagrangian structure, exerting point forces on the fluid. The forces on the structure are often determined by a bending or tension model, previously calculated using finite difference approximations. In this paper, we study Spherical Basis Function (SBF), Radial Basis Function (RBF) and Lagrange-Chebyshev parametric models to represent and calculate forces on elastic structures that can be represented by an open curve, motivated by the study of cilia and flagella. The evaluation error for static open curves for the different interpolants, as well as errors for calculating normals and second derivatives using different types of clustered parametric nodes, are given for the case of an open planar curve. We determine that SBF and RBF interpolants built on clustered nodes are competitive with Lagrange-Chebyshev interpolants for modeling twice-differentiable open planar curves. We propose using SBF and RBF parametric models within the MRS for evaluating and updating the elastic structure. Results for open and closed elastic structures immersed in a 2D fluid are presented, showing the efficacy of the RBF-Stokeslets method.