Searcharxiv⌕ Search

arXiv subjects

R. R. Kerswell

Publications and source records attributed to R. R. Kerswell.

15 recordsLinked to original sources
arXiv2025

Transient Growth in Streaky Unbounded Shear Flow: A symbiosis of Orr and Push-over mechanisms

Transient growth mechanisms operating on streaky shear flows are believed important for sustaining near-wall turbulence. Of the three individual mechanisms present - Orr, lift-up and 'push over' - Lozano-Duran et. al. (J. Fluid Mech. 914, A8, 2021) have recently observed that both Orr and push over need to be present to sustain turbulent fluctuations given streaky (streamwise-independent) base fields whereas lift-up does not. We show here, using Kelvin's model of unbounded constant shear augmented by spanwise-periodic streaks, that this is because the push-over mechanism can act in concert with a Orr mechanism based upon the streaks to produce much-enhanced transient growth. The model clarifies the transient growth mechanism originally found by Schoppa & Hussain, (J. Fluid Mech. 453, 57-108, 2002) and finds that this is one half of a linear instability mechanism centred at the spanwise inflexion points observed originally by Swearingen & Blackwelder (J. Fluid Mech. 182, 255-290, 1987). The instability and even transient growth acting on its own are found to have the correct nonlinear feedback to generate streamwise rolls which can then re-energise the assumed streaks through lift-up indicating a sustaining cycle. Our results therefore support the view that while lift-up is believed central for the roll-to-streak regenerative process, it is Orr and push-over mechanisms that are both key for the streak-to-roll regenerative process in near-wall turbulence.

physics.flu-dyn↗
arXiv2025

Godunov Loss Functions for Modelling of Hyperbolic Conservation Laws

Machine learning techniques are being used as an alternative to traditional numerical discretization methods for solving hyperbolic partial differential equations (PDEs) relevant to fluid flow. Whilst numerical methods are higher fidelity, they are computationally expensive. Machine learning methods on the other hand are lower fidelity but can provide significant speed-ups. The emergence of physics-informed neural networks (PINNs) in fluid dynamics has allowed scientists to directly use PDEs for evaluating loss functions. The downfall of this approach is that the differential form of systems is invalid at regions of shock inherent in hyperbolic PDEs such as the compressible Euler equations. To circumvent this problem we propose the Godunov loss function: a loss based on the finite volume method (FVM) that crucially incorporates the flux of Godunov-type methods. These Godunov-type methods are also known as approximate Riemann solvers and evaluate intercell fluxes in an entropy-satisfying and non-oscillatory manner, yielding more physically accurate shocks. Our approach leads to superior performance compared to standard PINNs that use regularized PDE-based losses as well as FVM-based losses, as tested on the 2D Riemann problem in the context of time-stepping and super-resolution reconstruction.

physics.flu-dyn↗
arXiv2024

Non-equivalence of quasilinear dynamical systems and their statistical closures

It is widely believed that statistical closure theories for dynamical systems provide statistics equivalent to those of the governing dynamical equations from which the former are derived. Here, we demonstrate counterexamples in the context of the widely used mean-field quasilinear (QL) approximation applied to 2D fluid dynamical systems. We compare statistics of QL numerical simulations with those obtained by direct statistical simulation via a cumulant expansion closed at second order (CE2). We observe that, though CE2 is an exact statistical closure for QL dynamics, its predictions disagree with the statistics of the QL solution for identical parameter values. These disagreements are attributed to instabilities, which we term rank instabilities, of the second cumulant dynamics within CE2 that are unavailable in the QL equations.

physics.flu-dyn↗
arXiv2019

Kelvin-Helmholtz billows above Richardson number $1/4$

We study the dynamical system of a forced stratified mixing layer at finite Reynolds number $Re$, and Prandtl number $Pr=1$. We consider a hyperbolic tangent background velocity profile in the two cases of hyperbolic tangent and uniform background buoyancy stratifications. The system is forced in such a way that these background profiles are a steady solution of the governing equations. As is well-known, if the minimum gradient Richardson number of the flow, $Ri_m$, is less than a certain critical value $Ri_c$, the flow is linearly unstable to Kelvin-Helmholtz instability in both cases. Using Newton-Krylov iteration, we find steady, two-dimensional, finite amplitude elliptical vortex structures, i.e. `Kelvin-Helmholtz billows', existing above $Ri_c$. Bifurcation diagrams are produced using branch continuation, and we explore how these diagrams change with varying $Re$. In particular, when $Re$ is sufficiently high we find that finite amplitude Kelvin-Helmholtz billows exist at $Ri_m>1/4$, where the flow is linearly stable by the Miles-Howard theorem. For the uniform background stratification, we give a simple explanation of the dynamical system, showing the dynamics can be understood on a two-dimensional manifold embedded in state space, and demonstrate the cases in which the system is bistable. In the case of a hyperbolic tangent stratification, we also describe a new, slow-growing, linear instability of the background profiles at finite $Re$, which complicates the dynamics.

physics.flu-dyn↗
arXiv2018

Stabilisation and drag reduction of pipe flows by flattening the base profile

Recent experimental observations (Kuehnen et al., 2018) have shown that flattening a turbulent streamwise velocity profile in pipe flow destabilises the turbulence so that the flow relaminarises. We show that a similar phenomenon exists for laminar pipe flow profiles in the sense that the nonlinear stability of the laminar state is enhanced as the profile becomes more flattened. Significant drag reduction is also observed for the turbulent flow when triggered by sufficiently large disturbances. The flattening is produced by an artificial body force designed to mimick a baffle used in the experiments of Kuehnen et al. (2018) and the nonlinear stability measured by the size of the energy of the initial perturbations needed to trigger transition. In order to make the latter computation more efficient, we examine how indicative the minimal seed for transition is in measuring transition thresholds. We first show that the minimal seed is relatively robust to base profile changes and spectral filtering. We then compare the (unforced) transition behaviour of the minimal seed with several forms of randomised initial conditions in the range of Reynolds numbers Re=2400 to 10000 and find that the energy of the minimal seed after the Orr and oblique phases of its evolution is close to that of a localised random disturbance. In this sense, the minimal seed at the end of the oblique phase can be regarded as a good proxy for typical disturbances (here taken to be the localised random ones) and is thus used as initial condition in the simulations with the body force. The enhanced nonlinear stability and drag reduction predicted in the present study are an encouraging first step in modelling the experiments of Kuehnen et al. and should motivate future developments to fully exploit the benefits of this promising direction for flow control.

physics.flu-dyn↗
arXiv2015

Instability driven by boundary inflow across shear: a way to circumvent Rayleigh's stability criterion in accretion disks?

We investigate the 2D instability recently discussed by Gallet et al. (2010) and Ilin \& Morgulis (2013) which arises when a radial crossflow is imposed on a centrifugally-stable swirling flow. By finding a simpler rectilinear example of the instability - a sheared half plane, the minimal ingredients for the instability are identified and the destabilizing/stabilizing effect of inflow/outflow boundaries clarified. The instability - christened `boundary inflow instability' here - is of critical layer type where this layer is either at the inflow wall and the growth rate is $O(\sqrtη)$ (as found by Ilin \& Morgulis 2013), or in the interior of the flow and the growth rate is $O(η\log 1/η)$ where $η$ measures the (small) inflow-to-tangential-flow ratio. The instability is robust to changes in the rotation profile even to those which are very Rayleigh-stable and the addition of further physics such as viscosity, 3-dimensionality and compressibility but is sensitive to the boundary condition imposed on the tangential velocity field at the inflow boundary. Providing the vorticity is not fixed at the inflow boundary, the instability seems generic and operates by the inflow advecting vorticity present at the boundary across the interior shear. Both the primary bifurcation to 2D states and secondary bifurcations to 3D states are found to be supercritical. Assuming an accretion flow driven by molecular viscosity only so $η=O(Re^{-1})$, the instability is not immediately relevant for accretion disks since the critical threshold is $O(Re^{-2/3})$ and the inflow boundary conditions are more likely to be stress-free than non-slip. However, the analysis presented here does highlight the potential for mass entering a disk to disrupt the orbiting flow if this mass flux possesses vorticity.

physics.flu-dyn↗
arXiv2014

An optimisation approach for analysing nonlinear stability with transition to turbulence in fluids as an exemplar

This article introduces, and reviews recent work using, a simple optimisation technique for analysing the nonlinear stability of a state in a dynamical system. The technique can be used to identify the most efficient way to disturb a system such that it transits from one stable state to another. The key idea is introduced within the framework of a finite-dimensional set of ordinary differential equations (ODEs) and then illustrated for a very simple system of 2 ODEs which possesses bistability. Then the transition to turbulence problem in fluid mechanics is used to show how the technique can be formulated for a spatially-extended system described by a partial differential equation (the well-known Navier-Stokes equation). Within that context, the optimisation technique bridges the gap between (linear) optimal perturbation theory and the (nonlinear) dynamical systems approach to fluid flows. The fact that the technique has now been recently shown to work in this very high dimensional setting augurs well for its utility in other physical systems.

physics.flu-dyn↗
arXiv2011

Variational identification of minimal seeds to trigger transition in plane Couette flow

A variational formulation incorporating the full Navier-Stokes equations is used to identify initial perturbations with finite kinetic energy E_{0} which generate the largest gain in perturbation kinetic energy (across all possible time intervals) for plane Couette flow. Two different representative flow geometries are chosen corresponding to those used previously by Butler & Farrell (1992) and Monokrousos et al. (2011). In the former (smaller geometry) case as E_{0} increases from 0, we find an optimal which is a smooth nonlinear continuation of the well-known linear result at $E_{0} = 0$. At $E_{0} = E_{c}$, however, completely unrelated states are uncovered which trigger turbulence and our algorithm consequently fails to converge. As $E_{0} \rightarrow E^{+}_{c}, we find good evidence that the turbulence triggering initial conditions approach a 'minimal seed' which corresponds to the state of lowest energy on the laminar-turbulent basin boundary or 'edge'. This situation is repeated in the Monokrousos et al. (2011) (larger) geometry albeit with one notable new feature - the appearance of a nonlinear optimal (as found recently in pipe flow by Pringle & Kerswell (2010) and boundary layer flow by Cherubini et al. (2010)) at finite $E_{0} < E_{c}$ which has a very different structure to the linear optimal. Again the minimal seed at $E_{0} = E_{c}$ does not resemble the linear or now the nonlinear optimal. Our results support the first of two conjectures recently posed by Pringle et al. (2011) but contradict the second. Importantly, their prediction that the form of the functional optimised is not important for identifying E_{c} providing heightened values are produced by turbulent flows is confirmed: we find the what looks to be the same E_{c} and minimal seed using energy gain as opposed to total dissipation in the Monokrousos et al. (2011) geometry.

physics.flu-dyn↗
arXiv2009

Exchange flow of two immiscible fluids and the principle of maximum flux

The steady, coaxial flow in which two immiscible, incompressible fluids move past each other in a cylindrical tube has a continuum of possibilities due to the arbitrariness of the interface between the fluids. By invoking the presence of surface tension to at least restrict the shape of any interface to that of a circular arc or full circle, we consider the following question: which flow will maximise the exchange when there is only one dividing interface Gamma? Surprisingly, the answer differs fundamentally from the better-known co-directional two-phase flow situation where an axisymmetric (concentric) core-annular solution always optimises the flux. Instead, the maximal flux state is invariably asymmetric either being a `side-by-side' configuration where Gamma starts and finishes at the tube wall or an eccentric core-annular flow where Gamma is an off-centre full circle in which the more viscous fluid is surrounded by the less viscous fluid. The side-by-side solution is the most efficient exchanger for a small viscosity ratio beta less than or equal to 4.60 with an eccentric core-annular solution optimal otherwise. At large beta, this eccentric solution provides 51% more flux than the axisymmetric core-annular flow which is always a local minimiser of the flux.

physics.flu-dyn↗
arXiv2008

Relative periodic orbits in transitional pipe flow

A dynamical system description of the transition process in shear flows with no linear instability starts with a knowledge of exact coherent solutions, among them travelling waves (TWs) and relative periodic orbits (RPOs). We describe a numerical method to find such solutions in pipe flow and apply it in the vicinity of a Hopf bifurcation from a TW which looks to be especially relevant for transition. The dominant structural feature of the RPO solution is the presence of weakly modulated streaks. This RPO, like the TW from which it bifurcates, sits on the laminar-turbulent boundary separating initial conditions which lead to turbulence from those which immediately relaminarise.

physics.flu-dyn↗
arXiv2008

Highly-symmetric travelling waves in pipe flow

The recent theoretical discovery of finite-amplitude travelling waves in pipe flow has re-ignited interest in the transitional phenomena that Osborne Reynolds studied 125 years ago. Despite all being unstable, these waves are providing fresh insight into the flow dynamics. Here we describe two new classes of highly-symmetric travelling waves (possessing rotational, shift-&-reflect and mirror symmetries) and report a new family of mirror-symmetric waves which is the first found in pipe flow not to have shift-&-reflect symmetry. The highly-symmetric waves appear at lower Reynolds numbers than the originally-discovered non-mirror-symmetric waves found by Faisst & Eckhardt 2003 and Wedin & Kerswell 2004 and have much higher wall shear stresses. The first M-class comprises of the various discrete-rotationally-symmetric analogues of the mirror-symmetric wave found in Pringle & Kerswell (2007) and have a distinctive double layer structure of fast and slow streaks across the pipe radius. The second N-class has the more familiar separation of fast streaks to the exterior and slow streaks to the interior and looks the precursor to the class of non-mirror-symmetric waves already known.

physics.flu-dyn↗
arXiv2008

Coherent structures in localised and global pipe turbulence

The recent discovery of unstable travelling waves (TWs) in pipe flow has been hailed as a significant breakthrough with the hope that they populate the turbulent attractor. We confirm the existence of coherent states with internal fast and slow streaks commensurate in both structure and energy with known TWs using numerical simulations in a long pipe. These only occur, however, within less energetic regions of (localized) `puff' turbulence at low Reynolds numbers (Re=2000-2400), and not at all in (homogeneous) `slug' turbulence at Re=2800. This strongly suggests that all currently known TWs sit in an intermediate region of phase space between the laminar and turbulent states rather than being embedded within the turbulent attractor itself. New coherent fast streak states with strongly decelerated cores appear to populate the turbulent attractor instead.

physics.flu-dyn↗
arXiv2006

Recurrence of Travelling Waves in Transitional Pipe Flow

The recent theoretical discovery of families of travelling wave solutions in pipe flow at Reynolds numbers lower than the transitional range naturally raises the question of their relevance to the turbulent transition process. Here a series of numerical experiments are conducted in which we look for the spatial signature of these travelling waves in transitionary flows. Working within a periodic pipe of 5D (diameters) length, we find that travelling waves with low wall shear stresses (lower branch solutions) are on a surface which separates initial conditions which uneventfully relaminarise and those which lead to a turbulent evolution. Evidence for recurrent travelling wave visits is found in both 5D and 10D long periodic pipes but only for those travelling waves with low-to-intermediate wall shear stress and for less than about 10% of the time in turbulent flow. Given this, it seems unlikely that the mean turbulent properties such as wall shear stress can be predicted as an expansion over the travelling waves in which their individual properties are appropriately weighted. Rather, further dynamical structures such as periodic orbits need to be isolated and included in any such expansion.

physics.flu-dyn↗
arXiv2006

Critical Behaviour in the Relaminarisation of Localised Turbulence in Pipe Flow

The statistics of the relaminarisation of localised turbulence in a pipe are examined by direct numerical simulation. As in recent experimental data (Peixinho & Mullin, Phys. Rev. Lett. 96, 094501, 2006), the half life for the decaying turbulence is consistent with the scaling (Re_c-Re)^{-1}, indicating a boundary crisis of the localised turbulent state familiar in low-dimensional dynamical systems. The crisis Reynolds number, is estimated as Re_c=1870, a value within 7% of the experimental value 1750. We argue that the frequently-asked question of which Re and initial disturbance are needed to trigger sustained turbulence in a pipe, is really two separate questions: the `local phase space' question (local to the laminar state) of what threshold disturbance at a given Re is needed to initially trigger turbulence, followed by the `global phase space' question of whether Re exceeds Re_c at which point the turbulent state becomes an attractor.

physics.flu-dyn↗