Multiscale quasi-time periodic coherent structures in shear flows
Runjie Song 1, Kengo Deguchi 1, Genta Kawahara 2, Yongyun Hwang 3 · 2026
Read Multiscale quasi-time periodic coherent structures in shear flows by Runjie Song 1, Kengo Deguchi 1, Genta Kawahara 2, Yongyun Hwang 3 (2026) 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-2601-18023.txt.
Introduction
Runjie Song, Kengo Deguchi, Genta Kawahara, and Yongyun Hwang look for hidden order inside shear-flow turbulence. Their 2026 preprint studies flows like those in pipes and channels. Many efforts seek simple patterns in the chaos, such as travelling waves or periodic orbits. The authors show that multiscale features need richer states. They focus on solutions that are quasi-periodic in time, repeating with more than one rhythm at once.
Such states can be computed well with a quasi-linear model. That keeps the cost sane while the structure stays rich. Richer building blocks may explain more of what real turbulence does. The work helps the long search for a skeleton beneath turbulent flow.
Catalog record
| CREATOR | Runjie Song 1, Kengo Deguchi 1, Genta Kawahara 2, Yongyun Hwang 3 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2601-18023
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Multiscale quasi-time periodic coherent structures in shear flows Runjie Song 1, Kengo Deguchi 1, Genta Kawahara 2, Yongyun Hwang 3 Abstract Attempts to disentangle shear-flow turbulence often focus on identifying relatively simple solutions, such as travelling waves or periodic orbits. We show, however, that capturing multiscale features requires considering states at least as complex as quasi-time-periodic solutions. Approximations of these states can be computed efficiently using a quasi-linear model, consistent with the large-Reynolds-number asymptotic analysis. The quasi-linear structure is key to producing multiscale critical layers that generate vortices obeying Taylor’s frozen-flow hypothesis. † † affiliation: 1 School of Mathematics, Monash University, VIC 3800, Australia 2 Graduate School of Engineering Science, University of Osaka, 1-3 Machikaneyama, Toyonaka, Osaka 560-8531, Japan 3 Department of Aeronautics, Imperial College London, London, SW7 2AZ, UK 1 Introduction Coherent structures in wall-bounded shear flows play a key role in transporting momentum and mass in turbulent flows ( 24 ) . Understanding and modelling their underlying mechanisms are important for a wide range of engineering applications and environmental sciences, and recent efforts have been driven by two distinct motivations. A natural motivation is to construct reduced models that faithfully reproduce turbulent direct numerical simulations or experimental data. A representative approach of this type is quasi-linear (QL) approximations, in which the flow is typically decomposed into slow- and fast-scale components. The governing equations for the fast-scale component are then linearised about the slow-scale flow, with a suitable closure that often neglects the self-interacting nonlinear terms. Several research groups have pursued this approach. Recently, 4 proposed to use a streamwise average to define the slow component and observed reasonable agreement with direct numerical simulation data at moderate Reynolds numbers. At higher Reynolds numbers, however, the spectral features of the data produced by the model do not match those of the full Navier-Stokes equations, leading to the development of more complicated models; see 14 , which also provides a comprehensive historical review of modelling approaches of this type. Conversely, one can be motivated to extract the minimal mechanisms underlying coherent structures. Even simple solutions, such as travelling-wave or time-periodic solutions known as exact coherent states (ECS, 25 ), were found to be underpinned by the self-sustaining process ( 13 ) . Although these solutions are unstable, dynamical systems theory allows them to be connected to the statistical properties of chaotic states ( 17 ; 19 ; 26 ) . Attempts to reproduce travelling-wave ECS using reduced models have been made, for example, by 3 , 2 and 20 , which are closely related to the aforementioned QL models. A careful study of why such models succeed ultimately leads to the large-Reynolds-number asymptotic analysis known as vortex-wave interaction ( 12 ; 11 , VWI,) ; for a historical overview of the related study, see 7 . In asymptotic analysis, one must define perturbation expansions for the relevant physical quantities and thoroughly verify that they are consistent with the governing equations. At present, no reduced model exists that simultaneously satisfies these strict criteria and reproduces turbulence statistics. ECS satisfy the governing equations accurately, which is why they are preferred over filtered turbulence data in equation-based approaches such as asymptotic analysis. Nevertheless, a fundamental dilemma remains: the ECS identified so far struggle to capture some important aspects of turbulence. In particular, multiscale vortical structures and their hierarchical features are absent; see e.g. 9 . Our goal is to overcome this limitation by constructing a reduced model that remains consistent with the VWI. The key is that, by incorporating multiple neutral waves into the VWI, each providing feedback to the streaky mean flow, the model can capture more complex temporal behaviour than a travelling wave. The paper is structured as follows. In § 2 , we introduce our proposed method, QL-VWI, and highlight the difference from the existing models discussed above. In § 3 , we first demonstrate that the QL-VWI can approximate a time-periodic Navier-Stokes solutions, and then turn to the more practically relevant plane Poiseuille flow to capture quasi-time-periodic solutions. The latter solutions capture some, though not all, of the characteristic features of coherent structures in shear flow turbulence, as further discussed in § 4 . 2 Formulation of the problem Consider the Navier-Stokes equations in Cartesian coordinates (x,y,z) \displaystyle(\partial_{t}+\mathbf{u}\cdot\nabla)\mathbf{u}=-\nabla p+R^{-1}\nabla^{2}\mathbf{u}+R^{-1}Q\mathbf{e}_{x},\qquad\nabla\cdot\mathbf{u}=0, (1) where t denotes time and \nabla is the gradient operator. The velocity \mathbf{u}=(u,v,w) and the pressure p are defined in a channel with y\in[-1,1] . The combination of the boundary conditions \mathbf{u}=(\pm 1,0,0) at y=\pm 1 and Q=0 corresponds to plane Couette flow, whose base flow is \mathbf{u}=(y,0,0) . Imposing \mathbf{u}=\mathbf{0} at y=\pm 1 and Q=2 yields plane Poiseuille flow, with base flow \mathbf{u}=(1-y^{2},0,0) . For both cases, our Reynolds number R is based on the characteristic velocity of the base flow and the channel half width. We enforce periodicity in z and quasi-periodicity in x and t . More specifically, the flow is assumed to depend on y , z , and the phase variables \theta_{m}=\alpha_{m}x+\omega_{m}t , m=1,2,\dots,M , each taken to be 2\pi periodic. Here, \alpha_{m} and \omega_{m} are real numbers representing the wavenumbers and frequencies, respectively. We use \beta to indicate the 2\pi/\beta periodicity of the flow in z . Travelling wave, periodic, and quasi-periodic solutions in a standard periodic box can all be captured within this setup. Now, let us apply a ‘Reynolds decomposition’ \mathbf{u}=\overline{\mathbf{u}}+\tilde{\mathbf{u}} and p=\overline{p}+\tilde{p} , defined by averaging over all phase variables. Here and hereafter, the overline denotes this averaging operation, while the tilde indicates the fluctuation (or wave ) component. The streamwise mean velocity \overline{u} represents the streaks , while the cross-stream components \overline{v} and \overline{w} represent the rolls . Substituting the decomposition into ( 1 ) and applying the averaging yields: \displaystyle[\overline{v}\partial_{y}+\overline{w}\partial_{z}-R^{-1}(\partial_{y}^{2}+\partial_{z}^{2})]\left[\begin{array}[]{c}\overline{u}\\ \overline{v}\\ \overline{w}\end{array}\right]+\left[\begin{array}[]{c}-R^{-1}Q\\ \partial_{y}\overline{p}\\ \partial_{z}\overline{p}\end{array}\right]=\mathbf{F},~~~\partial_{y}\overline{v}+\partial_{z}\overline{w}=0,~~~~~~~ where \mathbf{F}=-\overline{(\tilde{\mathbf{u}}\cdot{\nabla})\tilde{\mathbf{u}}} is the Reynolds stress. Taking the difference between ( 1 ) and ( 2 ) gives the fluctuation part of the momentum equations \displaystyle(\partial_{t}+\overline{\mathbf{u}}\cdot{\nabla})\tilde{\mathbf{u}}+(\tilde{\mathbf{u}}\cdot{\nabla})\overline{\mathbf{u}}+\{(\tilde{\mathbf{u}}\cdot{\nabla})\tilde{\mathbf{u}}-\overline{(\tilde{\mathbf{u}}\cdot{\nabla})\tilde{\mathbf{u}}}\}=-{\nabla}\tilde{p}+R^{-1}{\nabla}^{2}\tilde{\mathbf{u}},~~~ (9) and {\nabla}\cdot\tilde{\mathbf{u}}=0 . At this stage, no approximations have been made to the equations. The terms in curly brackets in ( 9 ) are the wave-wave interactions, which we hereafter refer to as WW. Removing them makes ( 9 ) linear in the fluctuations. Using the normal mode expansions \tilde{\mathbf{u}}=\sum_{m=1}^{M}\hat{\mathbf{u}}_{m}(y,z)\exp(\text{i}\theta_{m})+\text{c.c.} and \tilde{p}=\sum_{m=1}^{M}\hat{p}_{m}(y,z)\exp(\text{i}\theta_{m})+\text{c.c.} , with c.c. indicating the complex conjugate, the fluctuation part of the governing equations reduces to \displaystyle\mathscr{L}\hat{u}_{m}+\hat{v}_{m}\partial_{y}\overline{u}+\hat{w}_{m}\partial_{z}\overline{u}=-\text{i}\alpha_{m}\hat{p}_{m}, (10a) \displaystyle\mathscr{L}\hat{v}_{m}+\hat{v}_{m}\partial_{y}\overline{v}+\hat{w}_{m}\partial_{z}\overline{v}=-\partial_{y}\hat{p}_{m}, (10b) \displaystyle\mathscr{L}\hat{w}_{m}+\hat{v}_{m}\partial_{y}\overline{w}+\hat{w}_{m}\partial_{z}\overline{w}=-\partial_{z}\hat{p}_{m}, (10c) \displaystyle\text{i}\alpha_{m}\hat{u}_{m}+\partial_{y}\hat{v}_{m}+\partial_{z}\hat{w}_{m}=0, (10d) where \mathscr{L}=\text{i}\alpha_{m}(\overline{u}-c_{m})+\overline{v}\partial_{y}+\overline{w}\partial_{z}-R^{-1}(\partial_{y}^{2}+\partial_{z}^{2}-\alpha_{m}^{2}) . The phase speeds c_{m}=-\omega_{m}/\alpha_{m} are real, so the waves are neutral. At large R , the behaviour of the above equations is well approximated by the Rayleigh equation, which has a singularity at the critical level where \overline{u}=c_{m} (see 11 ). Among the terms in the wave equation ( 10 ), we refer to those involving the roll components \overline{v},\overline{w} as the WR terms. Also, the streamwise component of the Reynolds stress \displaystyle\mathbf{F}=-\sum_{m=1}^{M}\left[\begin{array}[]{c}\partial_{y}(\hat{u}_{m}\hat{v}_{m}^{*})+\partial_{z}(\hat{u}_{m}\hat{w}_{m}^{*})\\ \partial_{y}(\hat{v}_{m}\hat{v}_{m}^{*})+\partial_{z}(\hat{v}_{m}\hat{w}_{m}^{*})\\ \partial_{y}(\hat{w}_{m}\hat{v}_{m}^{*})+\partial_{z}(\hat{w}_{m}\hat{w}_{m}^{*})\end{array}\right]+\text{c.c.}, is referred to as the SR term. The QL-VWI approach solves ( 10 ) and ( 2 ) with the right hand side replaced by ( 2 ). In 2 , 7 , and 20 , the same system was solved but with M=1 ; the approach of 3 is a further simplification in which the WR and SR terms are neglected. The common