American Institute of Aeronautics and Astronautics
1
Shape Optimization and Fluid Dynamic Analysis
of a Translating Flexible Body
Scott L. Thomson
1
Brigham Young University, Provo, UT, 84663
The translation (swimming) of an oscillating, slender, flexible body through an initially
stationary fluid was explored using computational models and optimization routines. The
computational model included fully-coupled two-dimensional fluid and solid finite element
domains. The body leading edge was subjected to a prescribed vertical sinusoidal
displacement, whereas the entire body was free to move horizontally. Large body
deformation resulted in thrust production and horizontal body translation. Gradient-based
optimization methods were used to find the body shape that yielded the maximum horizontal
displacement over one period of vertical oscillation. It is shown that a tapered body with a
“rounded” leading edge achieved significantly greater horizontal velocity than a body of
uniform thickness. The computational domains, numerical models, optimization routines,
and model verification studies are described. The predicted responses of uniform and
optimal bodies are compared, and the sensitivity of horizontal displacement to body shape is
quantified.
I. Introduction
he purpose of this paper is to explore the response of an optimized slender body translating (“swimming”)
through an initially stationary fluid. The research is motivated by previous experimental studies of propulsion
generated by flapping motion. A review of the literature on this topic is not attempted here. Among the most
relevant include studies on a sinusoidally pitching and plunging airfoil
1
, a sinusoidally-plunging airfoil with a
flexible trailing edge
2
, and an oscillating fin
3
.
Heathcote and Gursul
2
performed experiments on a rigid airfoil with a very thin plate attached to the trailing
edge. Various plate thicknesses were studied and the presence of an optimal thickness for thrust coefficient
generation was noted, although identification of a truly “optimal” configuration was not the goal of the study.
In the present paper, a computational model of a body translating through a fluid is presented. The model
included separate, but coupled, fluid and solid domains. Combinations of sliding and deforming meshes allowed for
large fluid and solid domain displacements. Verification studies were performed to explore sensitivity to grid size
and time step size. The model was coupled with optimization routines to search for the optimal shape for
maximizing horizontal displacement. The purpose of the optimization algorithm was not for design; but rather, it
was intended to be used as an efficient tool for exploring the parameter space, from which further models may be
investigated in more detail.
II. Numerical Methods
A. Coupled Fluid-Solid Model
Fully-coupled fluid and solid domains were modeled in two dimensions using the finite element method with the
commerical code ADINA (ADINA R&D, Inc., Watertown, MA). The fluid domain was governed by the unsteady,
incompressible, viscous Navier-Stokes equations. The modeled fluid was water (ρ
fluid
= 1000 kg/m
3
, μ = 0.001 Pa·s).
A laminar flow model was used. The solid domain was based on the large deformation of a linearly-elastic steel
material model (ρ
solid
= 7861 kg/m
3
, E = 210 GPa, ν = 0.3). The solid domain was 0.09 cm long and was of variable
thickness, as discussed below. The fluid domain dimensions and boundary conditions are noted in Fig. 1. No
external fluid forces were applied; thus the only fluid motion was that which resulted from motion of the solid body.
1
Assistant Professor, Mechanical Engineering, 435 CTB, AIAA Member.
T
48th AIAA Aerospace Sciences Meeting Including the New Horizons Forum and Aerospace Exposition
4 - 7 January 2010, Orlando, Florida
AIAA 2010-1432
Copyright © 2010 by Scott Thomson. Published by the American Institute of Aeronautics and Astronautics, Inc., with permission.