Linking Dispersion and Stirring in Randomly Braiding Flows

Daniel R. Lester, Michael G. Trefry, Guy Metcalfe · 2024

Read Linking Dispersion and Stirring in Randomly Braiding Flows by Daniel R. Lester, Michael G. Trefry, Guy Metcalfe (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-05407.txt.

Introduction

Daniel Lester, Michael Trefry, and Guy Metcalfe wrote this 2024 study of how fluids mix and spread. Many natural and built flows swirl in random ways as they move. The paths of fluid bits braid around each other in these flows. The authors ask how that braiding links two things: how fast stuff spreads, and how strongly the flow stirs.

They show that a broad class of random flows shares one rule tying spread to stirring. Under the rule, the rate of stirring can be read off from how much the flow spreads sideways. They check the rule on flat flows that change with time and on steady flows in three dimensions. The link gives a simpler way to gauge mixing across many random flows.

Catalog record

CREATORDaniel R. Lester, Michael G. Trefry, Guy Metcalfe
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE48 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-05407

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Linking Dispersion and Stirring in Randomly Braiding Flows Daniel R. Lester, Michael G. Trefry, Guy Metcalfe Abstract Many random flows, including 2D unsteady and stagnation-free 3D steady flows, exhibit non-trivial braiding of pathlines as they evolve in time or space. We show that these random flows belong to a pathline braiding universality class that quantitatively links dispersion and chaotic stirring, meaning that the Lyapunov exponent can be estimated from the purely advective transverse dispersivity. We verify this quantitative link for both unsteady 2D and steady 3D random flows. This result uncovers a deep connection between transport and mixing over a broad class of random flows. Preprint: APS/123-QED Most steady and unsteady flows exhibit complex Lagrangian kinematics, whereby pathlines intertwine in a chaotic manner in space and time [ 1 , 2 ] . Such Lagrangian chaos or chaotic adection [ 3 , 4 ] arises from stretching and folding of fluid elements, leading to highly striated material distributions that profoundly augment many fluid-borne phenomena [ 5 , 6 , 7 ] , including solute transport [ 8 , 9 ] , mixing [ 10 , 11 ] , chemical reactions [ 12 , 13 ] , biological activity [ 14 , 15 ] , particle transport [ 16 , 17 ] and heat transfer [ 18 , 19 ] . Complex stirring arises because the kinematic equation \dot{\mathbf{x}}(t)=\mathbf{v}(\mathbf{x},t) is rich enough to admit chaotic dynamics if the velocity field \mathbf{v}(\mathbf{x},t) possesses sufficient spatio-temporal degrees of freedom ( \text{dof}\geqslant 3 ) [ 20 ] and constraints such as symmetries in Stokes flow [ 21 ] or helicity in Darcy flow [ 22 ] are not imposed. Unsteady 2D flows and stagnation-free steady 3D flows may be topologically unified by considering the former as the latter with unit velocity in time [ 23 ] . For these \text{dof}=3 flows, the intertwining of pathlines may be formalised as a non-trivial braiding process [ 24 ] , and a mathematical framework [ 25 ] and associated tools [ 26 , 27 ] have been developed to quantify the topological complexity of the braiding motions. As pathlines are invariant 1D objects in a 3D continuum, non-trivial pathline braiding stretches and folds the interstitial continuum fluid in a chaotic manner (Fig. 1 a) [ 24 ] . As the topological complexity of the braiding motions approximates that of the fluid flow [ 28 ] , there exist deep connections between braiding complexity and the rate of chaotic stirring [ 29 ] . In addition to stirring, pathline braiding in unbounded random flows is also strongly linked to advective transverse particle dispersion as random pathlines must typically cross one another (with respect to an arbitrary reference direction) to separate without bound. Although there exist particular flows that do not braid but can disperse (e.g. hyperbolic flows) and vice-versa (e.g. confined braiding flows), these specific cases arise in either bounded or deterministic flows. For unbounded random flows, strong links have been observed between non-trivial braiding and particle dispersion in 2D turbulence [ 30 ] and ferrofluids [ 31 ] , while many studies [ 32 ] have uncovered links between particle dispersion and chaotic stirring. Conversely, steady 3D flows with zero helicity density [ 33 , 34 ] \mathcal{H}(\mathbf{x})\equiv\mathbf{v}(\mathbf{x})\cdot\nabla\times\mathbf{v}(\mathbf{x}) are integrable [ 1 ] (non-chaotic) and admit a pair of streamfunctions \psi_{1}(\mathbf{x}) , \psi_{2}(\mathbf{x}) [ 35 , 36 ] that act as Euler potentials as \mathbf{v}(\mathbf{x})=\nabla\psi_{1}(\mathbf{x})\times\nabla\psi_{2}(\mathbf{x}) . These potentials prohibit pathline braiding as streamlines are confined to level sets of \psi_{1}(\mathbf{x}) , \psi_{2}(\mathbf{x}) and have been proved [ 37 ] to also prevent transverse particle dispersion. While these observations illustrate the strong correlations between chaotic stirring and particle dispersion in randomly braiding flows, lacking is a mechanistic understanding of the underlying link between these phenomena. We examine the link between particle dispersion and braiding via simple random braiding models and show that there exists a pathline braiding universality class to which all 3 dof unbounded random flows belong. This reveals the deep connection between dispersion and chaotic stirring in randomly braiding 3 dof flows; moreover, this uncovers a quantitative link between the Lyapunov exponent \lambda_{\infty} and advective transverse dispersivity D_{T} . We validate the link for two example flows: an unsteady random 2D flow and an steady 3D random flow. The latter case is related to a proof of the ubiquity of chaotic advection in heterogeneous Darcy flow [ 38 ] . (a) (b) (c) Figure 1: (a) Top left: Schematic of the 1D streamline model including streamlines n-1 , n , n+1 in the transverse x_{2}-x_{3} plane and clockwise (black) \sigma_{n} , anti-clockwise (blue) \sigma_{n}^{-1} braid generators . Bottom left: stretching of material elements due to braiding motions (adapted from [ 39 ] ) that evolve with the longitudinal x_{1} direction. Bottom right: braid diagram depicting stretching of material elements (red) around streamlines (black). (b) Braid diagram for N_{p}=20 streamlines over N_{b}=20 braid actions in the x_{1} direction, leading to non-trivial braiding and transverse dispersion. (c) Growth of topological entropy h (black line) and transverse variance \sigma_{x_{2}}^{2} (blue line) with N_{b} , in agreement with ( 1 ) (red line). Inset: Brownian motion of streamlines with increasing N_{b} . To elucidate the link between pathline braiding, chaotic advection and transverse dispersion, we begin with a particularly simple streamline model that, nonetheless, captures the essential features of a mean longitudinal steady chaotic 3D flow: namely braiding of streamlines as they propagate longitudinally in the x_{1} direction. The unsteady 2D analog is a 2D flow with pathlines that braid in time t rather than x_{1} . As shown in Fig. 1 a (top left), the 1D streamline model consists of a 1D array of N_{p} streamlines in \mathbb{R}^{3} which propagate with uniform velocity in the x_{1} direction and repeat periodically with uniform spacing \Delta x_{2}=\ell in the x_{2} direction, where \ell is the velocity correlation length. At integer multiples of longitudinal distance x_{1}=\Delta_{L} , one pair of neighbouring streamlines randomly exchange position via the clockwise or counter-clockwise rotations shown in Fig. 1 a, which respectively are labelled by the braid generators \sigma_{n} and \sigma_{n}^{-1} for the streamline pair (n,n+1) , with n\in 1:N_{p}-1 . The sequence of N_{b} braid generators forms the braid word \mathbf{b}=\sigma_{n_{1}}^{\pm 1}\sigma_{n_{2}}^{\pm 1}\dots\sigma_{n_{N_{b}}}^{\pm 1} that completely defines the braiding of streamlines. Due to periodicity the set of N_{p} streamlines undertake unbounded random walks along the x_{2} coordinate as they propagate longitudinally. Fig. 1 b shows the braid diagram for N_{p}=20 streamlines undertaking N_{b}=20 random braid actions, illustrating braiding actions and transverse dispersion of streamlines. Figure 2: Convergence of scaled mean topological entropy N_{p}/N_{b}\langle h\rangle with braid number N_{b} for different streamline numbers N_{p} (green 50 , light blue 100 , purple 200 , pink 500 , red 1000 ). Dots indicate mean value, error bars indicate \pm standard error. Inset: convergence of scaled mean topological entropy hN_{p}/N_{b} for N_{b}=10^{7} with streamline number N_{p} to random braid entropy \langle\lambda_{\sigma}\rangle . Fig. 1 a shows that streamline braiding can lead to rapid stretching of material elements. If the braiding is non-trivial (pseudo-Anosov) [ 24 ] , this stretching grows exponentially with the number N_{b} of braiding actions, the growth rate of which is quantified by the topological

Read the complete text