Hostname: page-component-7b7567c8c-7pwfn Total loading time: 0 Render date: 2026-11-16T00:50:39.009Z Has data issue: false hasContentIssue false

Distributed flexibility in inertial swimmers

Published online by Cambridge University Press:  11 February 2020

Daniel Floryan*
Affiliation:
Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, NJ 08544, USA
Clarence W. Rowley
Affiliation:
Department of Mechanical and Aerospace Engineering, Princeton University, Princeton, NJ 08544, USA
*
†Email address for correspondence: dfloryan@alumni.princeton.edu

Abstract

We study a linear inviscid model of a passively flexible swimmer with distributed flexibility, calculating its propulsive performance and optimal distributions of flexibility. The frequencies of actuation and mean stiffness ratios we consider span a large range, while the mass ratio is fixed to a low value representative of swimmers. We present results showing how the trailing edge deflection, thrust coefficient, power coefficient and efficiency vary with frequency, mean stiffness and stiffness distribution. Swimmers with distributed flexibility have the same qualitative features as those with uniform flexibility. Significant gains in thrust can be made, however, by tuning the stiffness such that a resonant response is triggered, or by concentrating stiffness towards the leading edge if resonance cannot be triggered. To minimize power, the opposite is true. Meaningful gains in efficiency can be made at low frequencies by concentrating stiffness away from the leading edge, since doing so induces efficient travelling wave kinematics. We also speculate on the effects of a finite Reynolds number in the form of streamwise drag. The drag adds an offset to the net thrust produced by the swimmer, causing efficiency-maximizing distributions of flexibility to tend towards thrust-maximizing ones, representative of what is found in nature.

Information

Type
JFM Papers
Copyright
© The Author(s), 2020. Published by Cambridge University Press
Figure 0

Figure 1.Schematic of the problem. The varying colour represents the varying material properties.

Figure 1

Table 1.Parameter values used in this work.

Figure 2

Figure 2.Thrust coefficient, power coefficient and efficiency as a function of the mass distribution for a (a) heaving and (b) pitching plate for $\langle R\rangle =0.01$, $S\equiv 1$ and $f^{\ast }=1$. The mass distribution $R$ is linear, and values are normalized by their value when the mass is uniformly distributed.

Figure 3

Figure 3.Thrust coefficient, power coefficient and efficiency as a function of the stiffness distribution for a (a) heaving and (b) pitching plate for $R\equiv 0.01$, $\langle S\rangle =1$ and $f^{\ast }=1$. The stiffness distribution $S$ is linear, and values are normalized by their value when the mass is uniformly distributed.

Figure 4

Figure 4.Trailing edge amplitude as a function of reduced frequency $f^{\ast }$ and stiffness ratio $S$ for a (a) heaving and (b) pitching plate with $R\equiv 0.01$ relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same trailing edge amplitude as the equivalent rigid plate. Under-resolved areas have been whited out. Results are for a uniformly flexible plate. The mean thrust and mean power coefficients are qualitatively the same as the trailing edge amplitude.

Figure 5

Figure 5.Efficiency as a function of reduced frequency $f^{\ast }$ and stiffness ratio $S$ for a (a) heaving and (b) pitching plate with $R\equiv 0.01$ relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same efficiency as the equivalent rigid plate. Under-resolved areas and areas that produce negative efficiency have been whited out. Results are for a uniformly flexible plate.

Figure 6

Figure 6.Analogue of figure 4, for a stiffness distribution $S^{\ast }(x)=1-0.9x$ (stiff leading edge).

Figure 7

Figure 7.Analogue of figure 4, for a stiffness distribution $S^{\ast }(x)=1+0.9x$ (soft leading edge).

Figure 8

Figure 8.Analogue of figure 5, for a stiffness distribution $S^{\ast }(x)=1-0.9x$ (stiff leading edge).

Figure 9

Figure 9.Analogue of figure 5, for a stiffness distribution $S^{\ast }(x)=1+0.9x$ (soft leading edge).

Figure 10

Figure 10.(a) Feasible set of quadratic flexibility distributions. The dark region contains stiffness distributions whose minima are at the leading/trailing edge, and the light region contains stiffness distributions whose minima are at an interior point of the plate. We have drawn some representative distributions in (b), corresponding to the circles in (a).

Figure 11

Figure 11.Thrust-maximizing linear stiffness distribution as a function of reduced frequency $f^{\ast }$ and mean stiffness ratio $\langle S\rangle$ for a (a) heaving and (b) pitching plate with $R\equiv 0.01$. Under-resolved areas and areas that produce negative thrust have been whited out.

Figure 12

Figure 12.Thrust coefficient of a plate with the stiffness distribution shown in figure 11 relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same thrust coefficient as the equivalent rigid plate.

Figure 13

Figure 13.Same as in figure 11, but with natural frequencies overlaid as curves for $\text{d}S^{\ast }/\text{d}x=-0.9$ (stiff leading edge, green), $\text{d}S^{\ast }/\text{d}x=0$ (uniformly stiff, white) and $\text{d}S^{\ast }/\text{d}x=0.9$ (soft leading edge, purple).

Figure 14

Figure 14.First four quiescent natural frequencies as a function of the stiffness distribution. The natural frequencies have been normalized by the natural frequencies of a uniformly stiff plate, $\text{d}S^{\ast }/\text{d}x=0$.

Figure 15

Figure 15.Thrust coefficient as a function of reduced frequency $f^{\ast }$ and mean stiffness ratio $\langle S\rangle$ for a (a) heaving and (b) pitching plate with $R\equiv 0.01$. The green surface corresponds to a plate with a stiff leading edge ($\text{d}S^{\ast }/\text{d}x=-0.9$), and the purple surface corresponds to a plate with a soft leading edge and matched third natural frequency ($\text{d}S^{\ast }/\text{d}x\in [0.416,0.445]$).

Figure 16

Figure 16.Analogue of figure 12, but compared to a plate with uniformly distributed stiffness instead of a rigid plate.

Figure 17

Figure 17.Analogue of figure 11, but for a linear stiffness distribution minimizing power.

Figure 18

Figure 18.Power coefficient of a plate with the stiffness distribution shown in figure 17 relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same power coefficient as the equivalent rigid plate.

Figure 19

Figure 19.Analogue of figure 18, but compared to a plate with uniformly distributed stiffness instead of a rigid plate.

Figure 20

Figure 20.Analogue of figure 11, but for a linear stiffness distribution maximizing efficiency.

Figure 21

Figure 21.Efficiency of a plate with the stiffness distribution shown in figure 20 relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same efficiency as the equivalent rigid plate.

Figure 22

Figure 22.Analogue of figure 21, but compared to a plate with uniformly distributed stiffness instead of a rigid plate.

Figure 23

Figure 23.Ten snapshots, evenly spaced in time, of a heaving plate with (a) uniform flexibility and (b) $\text{d}S/\text{d}x^{\ast }=0.99$ (soft leading edge) comprising one period of motion. Both cases have the same heave amplitude, $\langle S\rangle =8$ and $f^{\ast }=0.2$.

Figure 24

Figure 24.Ten snapshots, evenly spaced in time, of a pitching plate with (a) uniform flexibility and (b) $\text{d}S/\text{d}x^{\ast }=0.99$ (soft leading edge) comprising one period of motion. Both cases have the same pitch amplitude, $\langle S\rangle =0.5$ and $f^{\ast }=0.3$.

Figure 25

Figure 25.Colour coding of the boundary of the feasible set in figure 10.

Figure 26

Figure 26.Thrust-maximizing quadratic stiffness distribution as a function of reduced frequency $f^{\ast }$ and mean stiffness ratio $\langle S\rangle$ for a (a) heaving and (b) pitching plate with $R\equiv 0.01$. Under-resolved areas and areas that produce negative thrust have been whited out.

Figure 27

Figure 27.Thrust coefficient of a plate with the stiffness distribution shown in figure 26 relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same thrust coefficient as the equivalent rigid plate.

Figure 28

Figure 28.Analogue of figure 26, but for a quadratic stiffness distribution minimizing power.

Figure 29

Figure 29.Power coefficient of a plate with the stiffness distribution shown in figure 28 relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same power coefficient as the equivalent rigid plate.

Figure 30

Figure 30.Analogue of figure 26, but for a quadratic stiffness distribution maximizing efficiency.

Figure 31

Figure 31.Efficiency of a plate with the stiffness distribution shown in figure 30 relative to that of an equivalent rigid plate. Dashed white lines indicate where the flexible plate has the same efficiency as the equivalent rigid plate.

Figure 32

Figure 32.Analogue of figure 31, but compared to a plate with uniformly distributed stiffness instead of a rigid plate.