Flexible lament in timeperiodic viscous ow shape chaos and period three Vipin Agrawal1 2and Dhrubaditya Mitra1 1Nordita KTH Royal Institute of Technology and Stockholm University

2025-04-27 0 0 1.09MB 7 页 10玖币
侵权投诉
Flexible filament in time–periodic viscous flow : shape chaos and period three
Vipin Agrawal1, 2 and Dhrubaditya Mitra1,
1Nordita, KTH Royal Institute of Technology and Stockholm University,
12 Hannes Alfv´ens v¨ag 10691 Stockholm, Sweden
2Department of Physics, Stockholm university, Stockholm.
(Dated: October 11, 2022)
We study a single, freely–floating, inextensible, elastic filament in a linear shear flow: U0(x, y) =
˙γyˆx. In our model: the elastic energy depends only on bending; the rate–of–strain, ˙γ=Ssin(ωt) is
a periodic function of time, t; and the interaction between the filament and the flow is approximated
by a local isotropic drag force. Based on the shape of the filament we find five different dynamical
phases: straight, buckled, periodic (with period two, period three, period four, etc), chaotic and
one with chaotic transients. In the chaotic phase, we show that the iterative map for the angle,
which the end–to–end vector of the filament makes with the tangent its one end, has period three
solutions; hence it is chaotic. Furthermore, in the chaotic phase the flow is an efficient mixer.
Keywords: fluid-structure interactions, spatiotemporal chaos, periodic orbit theory
I. INTRODUCTION
The dynamics of flexible filaments in flows plays a cru-
cial role in many biological and industrial processes [1].
A canonical example is that of cilia and flagella [2, 3]
that takes part in wide variety of biological tasks, e.g.,
swimming of microorganisms, feeding and breathing of
marine invertebrates. In such cases, although the flow
nonlinearities can often be safely ignored, due to its elas-
tic nonlinearities and flow–structure interactions a single
isolated filament can show surprisingly complex dynam-
ics in flows. Both active and passive filament, anchored
or freely floating, in various steady flows have been stud-
ied extensively, see Ref. [4, 5] and references therein. In
steady flows, a single passive filament has quite complex
transient dynamics [6–14]. For active filaments, the focus
has been on how a periodic driving can give rise to sym-
metry breaking, e.g., swimming [15] or whirling [16–18].
This year, three papers have focussed on, how periodic
driving, either of the flow or the filament, can give rise
to secondary instabilities [19] or statistically stationary
state with chaotic/complex dynamics [20, 21]. For the
latter, the shape of the filament, as described by its cur-
vature as a function of its arc length, is a spatiotempo-
rally chaotic function. Henceforth we call this phenom-
ena shape chaos. Such chaotic solutions are particularly
interesting because they have the potential to be used to
generate efficient mixing in microfluidics.
Two effects determine the fate of an elastic filament
in flow. One is the elastic nonlinearity of the filament
and the other is the viscous interaction between the fil-
ament and the flow. The latter, in all its glory, gives
rise to non–local and nonlinear interaction between two
different parts of the same filament. Nevertheless, the-
oretical studies [16, 22–24] have often approximated the
viscous effect as a local, linear, isotropic drag. Can this
local approximation to the flow–structure interaction still
dhruba.mitra@gmail.com
FIG. 1. Sketch of our numerical experiment. Initially
the filament is straight and is aligned vertically. The back-
ground shear flow, Equation (1) is shown as arrow: t=T /4
(top panel) and t= 3T/4 (bottom panel).
capture the shape chaos of a freely-floating filament ? As
we show in the rest of this paper, the answer is yes; we
prove shape chaos using Sharkovskii and Li and Yorke’s
famous result [25] – existence of period orbits of period
three implies not only the existence of orbits of all periods
but also senstive dependence on initial condition.
In Figure. 1 we show a sketch of our numerical exper-
iment. Initially the filament is aligned vertically. The
background shear flow is given by
U0(x, y) = ˙γyˆx,and ˙γ=Ssin(ωt).(1)
Here T= 2πis the time period of the periodic shear
arXiv:2210.04781v1 [cond-mat.soft] 10 Oct 2022
2
and Sis a constant.
II. MODEL
We model the filament using the bead-spring model [7,
10, 11, 18, 26, 27]: identical spherical beads of diameter
dare connected by over-damped springs of equilibrium
length a. The position of the center of the i-th bead is
Ri, where i= 1 . . . N, the total number of beads. The
equation of motion is:
Rα
i
t =1
3πηd
H
Rα
i
+Uα
0(Ri),(2)
where U0is given in (1). Here ηis viscosity of the fluid,
(·)/∂(·) denotes partial derivative, U0is the velocity of
the background shear, and His the elastic Hamiltonian
of the filament. The Greek indices run from 1 to D,
the dimensionality of the space, and the Latin indices
run from 1 to N. The elastic Hamiltonian [18, 28], has
contributions from bending (HB) and stretching (HS):
H=HB+HSwhere (3a)
HB=aB
N1
X
i=0
κ2
iand (3b)
HS=H
2a
N1
X
i=0
(|Ri+1 Ri| − a)2; where (3c)
κi=1
a|ˆ
ui׈
ui1|and (3d)
ˆ
ui=Ri+1 Ri
|Ri+1 Ri|.(3e)
Here Bis the bending modulus of the filament and His
its stretching modulus. We ignore thermal fluctuations
and torsion. Three dimensionless parameters determine
the dynamics. We call them, the elasto–viscous parame-
ters, the dimensionless frequency and the ratio of stretch-
ing to bending defined respectively as:
µ8πηSL4
B,(4a)
σω
S,and (4b)
KHa2
B.(4c)
In practice, the filaments are inextensible [29], which
we implement by choosing appropriately high value of
K. We evolve Equation (2) using adaptive Runge-Kutta
[30] method with cash-karp parameters [31]. Our code
is freely available [32] and has been benchmarked against
experimental results [20]. A complete list of the param-
eters of the simulation is given in table I. We study the
problem for a large range of µand σall within experi-
mentally realizable range. Note that, with the local ap-
proximation of viscous forces it is possible for the fila-
ment to cross itself. Such unphysical solutions do appear
in our simulations but for values of µother than that
has been considered in this paper. The computational
complexity of the model, where the viscous interaction
is modelled by the non–local Rotne–Pregor tensor [20],
is O(N2) where Nis the number of beads, whereas the
computational complexity of the model with local vis-
cosity is O(N). This allows us to run our simulations for
much longer times than it was possible in Ref. [20].
III. RESULTS
FIG. 2. (A):Phase diagram in the µσplane; We find 5 dif-
ferent qualitatively different dynamical phases: Straight();
periodic (H) with n-period, where n=2(), 3(J), 4(I); com-
plex (F) complex-transients ()(B) Solutions of strobo-
scopic map: The stroboscopic map has many periodic solu-
tions at every point in µσplane. We show time period of
the lowest cycle in Sharvoskii ordering.
A rigid ellipsoid in a periodic shear may show chaotic
three–dimensional rotation under certain conditions [33–
36]. Such behavior emerges due to the nonlinearities
present in the Euler’s equations of rigid body rotation.
Here we consider a filament with no inertia, hence such
chaotic solutions are not present in our system. For a fil-
摘要:

Flexible lamentintime{periodicviscousow:shapechaosandperiodthreeVipinAgrawal1,2andDhrubadityaMitra1,1Nordita,KTHRoyalInstituteofTechnologyandStockholmUniversity,12HannesAlfvensvag10691Stockholm,Sweden2DepartmentofPhysics,Stockholmuniversity,Stockholm.(Dated:October11,2022)Westudyasingle,freely{oa...

展开>> 收起<<
Flexible lament in timeperiodic viscous ow shape chaos and period three Vipin Agrawal1 2and Dhrubaditya Mitra1 1Nordita KTH Royal Institute of Technology and Stockholm University.pdf

共7页,预览2页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!
分类:图书资源 价格:10玖币 属性:7 页 大小:1.09MB 格式:PDF 时间:2025-04-27

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 7
客服
关注