Fact-checked by Grok 2 weeks ago

Taylor–Couette flow

Taylor–Couette flow is the motion of a viscous, incompressible fluid confined between two concentric cylinders of finite length, with the inner cylinder typically rotating at angular velocity \Omega_i relative to the stationary or slower-rotating outer cylinder at \Omega_o, resulting in a base azimuthal velocity profile v_\theta(r) = A r + B / r (with constants A and B set by the boundary conditions) in the laminar regime.^[1] In Couette's original 1890 viscometer, the outer cylinder rotated while the inner was stationary to measure fluid viscosity through torque; subsequent studies focused on inner-cylinder rotation.^[2] This configuration exhibits a rich variety of instabilities as rotational speeds increase, transitioning from stable circular Couette flow to axisymmetric Taylor vortices—pairs of counter-rotating toroidal rolls aligned along the cylinder axis—beyond a critical Taylor number Ta_c \approx 1708 for narrow gaps.^[1]^[3] Geoffrey Ingram Taylor's seminal 1923 analysis predicted this primary instability through linear stability theory, solving the Navier-Stokes equations perturbed around the base state to derive the critical conditions for onset, confirmed experimentally by observing the appearance of these vortices.^[4] The system's dynamics are governed by dimensionless parameters including the radius ratio \eta = R_i / R_o, the aspect ratio \Gamma = L / (R_o - R_i), and Reynolds numbers Re_i = \Omega_i R_i d / \nu and Re_o = \Omega_o R_o d / \nu (where d = R_o - R_i is the gap width and \nu is kinematic viscosity), with the Taylor number Ta = 4 \Omega_i^2 R_i^2 d^3 / (\nu^2 R_o^2) (for \Omega_o = 0) quantifying centrifugal effects driving the instability.^[5] As Ta exceeds Ta_c, axisymmetric Taylor vortices form, becoming unstable to non-axisymmetric wavy vortices at higher speeds, followed by chaotic or turbulent regimes, providing a model for studying hydrodynamic bifurcations, pattern formation, and subcritical transitions to turbulence under controlled conditions.^[6] The flow's exact solvability in the laminar state and tunable instabilities make it a benchmark for numerical simulations and theoretical developments in nonlinear dynamics.^[7] Taylor–Couette flow holds fundamental importance in fluid mechanics as a paradigm for centrifugal instabilities and shear-driven turbulence, influencing research in geophysical flows like atmospheric jets and oceanic currents.^[6] In engineering, it underpins designs for Taylor vortex flow reactors used in chemical processing for enhanced mass transfer and mixing with minimal axial dispersion, as well as in lubrication systems and bearing analysis.^[8] Astrophysically, it models angular momentum transport in accretion disks around black holes and stars, where differential rotation drives instabilities analogous to magneto-rotational instability.^[9] Ongoing experiments and simulations continue to explore its turbulent statistics, heat transfer enhancement, and behavior in complex fluids like suspensions or non-Newtonians.^[10]

Fundamentals

Geometry and Setup

The Taylor–Couette flow refers to the motion of a viscous incompressible fluid confined between two concentric cylinders, with the inner cylinder of radius R_1 rotating at angular velocity \Omega_1 and the outer cylinder of radius R_2 > R_1 rotating at angular velocity \Omega_2.^[4] In the standard configuration, the outer cylinder is stationary (\Omega_2 = 0), while the inner cylinder rotates to drive the flow.^[4] This setup originated from Maurice Couette's 1890 viscometer experiments, where he rotated the outer cylinder with the inner one fixed to measure liquid viscosities through torque observations. Geoffrey Ingram Taylor extended this in 1923 by rotating the inner cylinder with the outer fixed, investigating the onset of instabilities to probe viscous flow stability.^[4] Key geometric parameters define the system: the radius ratio \eta = R_1 / R_2 (where $0 < \eta < 1), which characterizes the annular gap's relative width; the gap width d = R_2 - R_1; and, for finite-length cylinders of axial height L, the aspect ratio \Gamma = L / d, which influences end effects in experimental setups.^[9] These parameters allow systematic variation in experiments, with narrow gaps (\eta close to 1) approximating plane Couette flow and wider gaps (\eta smaller) enhancing centrifugal effects.^[9] Boundary conditions enforce no-slip at the cylinder walls, where the fluid velocity matches the local wall velocity, ensuring continuity of tangential motion.^[4] In the laminar regime, the flow is predominantly azimuthal, with velocity varying radially between the cylinders and minimal radial or axial components under steady conditions.^[4]

Governing Equations and Dimensionless Parameters

The Taylor–Couette flow is governed by the incompressible Navier–Stokes equations in cylindrical coordinates (r, \theta, z). The continuity equation is

\nabla \cdot \mathbf{v} = 0,

and the momentum equation is

\frac{\partial \mathbf{v}}{\partial t} + (\mathbf{v} \cdot \nabla) \mathbf{v} = -\frac{1}{\rho} \nabla p + \nu \nabla^2 \mathbf{v},

where \mathbf{v} = (v_r, v_\theta, v_z) is the velocity field, p is the pressure, \rho is the constant fluid density, and \nu is the kinematic viscosity. The Laplacian and advective terms are expressed using the standard cylindrical operators: for example, the \theta-component of the Laplacian includes \nabla^2 v_\theta - v_\theta / r^2 - (2/r^2) \partial v_r / \partial \theta. No-slip boundary conditions apply at the cylinder surfaces: \mathbf{v} = (0, \Omega_1 R_1, 0) at r = R_1 and \mathbf{v} = (0, \Omega_2 R_2, 0) at r = R_2, where \Omega_1, \Omega_2 are the angular velocities and R_1, R_2 (with R_1 < R_2) are the inner and outer radii, respectively.^[4] For the laminar base flow, the assumptions of steady, axisymmetric conditions (v_r = v_z = 0, v_\theta = v_\theta(r), independent of \theta and z) simplify the \theta-momentum equation to

$0 = \nu \left( \frac{d^2 v_\theta}{dr^2} + \frac{1}{r} \frac{d v_\theta}{dr} - \frac{v_\theta}{r^2} \right).

The general solution satisfying the boundary conditions is the circular Couette profile

v_\theta(r) = A r + \frac{B}{r},

with coefficients

A = \Omega_1 \frac{\mu - \eta^2}{1 - \eta^2}, \quad B = \Omega_1 R_1^2 \frac{1 - \mu}{1 - \eta^2},

derived by solving the system at r = R_1 and r = R_2. Here, \eta = R_1 / R_2 is the radius ratio. In the small-gap approximation (d = R_2 - R_1 \ll R_1, so \eta \approx 1), curvature effects are negligible, and the profile approximates a linear shear flow v_\theta(r) \approx \Omega_1 R_1 + (\Omega_2 R_2 - \Omega_1 R_1) (r - R_1)/d.^[4] The nondimensionalization of the problem introduces key parameters controlling the flow behavior. The Reynolds number is defined as \mathrm{Re} = \Omega_1 R_1 d / \nu, representing the ratio of inertial to viscous forces driven by the inner cylinder rotation. The Taylor number \mathrm{Ta} = 4 \mathrm{Re}^2 \frac{1 - \eta}{1 + \eta} (or close variants incorporating additional geometrical prefactors) measures the relative importance of centrifugal and viscous effects, with the critical value for instability onset approaching 1708 in the narrow-gap limit.^[11] The angular velocity ratio \mu = \Omega_2 / \Omega_1 determines the co- or counter-rotation regime, influencing stability (e.g., \mu = 0 for outer cylinder at rest). These parameters, along with \eta, fully characterize the system for fixed aspect ratio.^[4]^[12] To analyze stability, small perturbations are superimposed on the base flow: \mathbf{v} = (0, v_\theta(r), 0) + \delta \mathbf{v}, p = p_0 + \delta p, with \delta \mathbf{v}, \delta p \ll 1. Assuming normal modes \delta \mathbf{v}(r, \theta, z, t) = \hat{\mathbf{v}}(r) \exp(i k z + i n \theta + \sigma t), where k is the axial wavenumber, n is the azimuthal mode number (often n=0 for axisymmetric modes), and \sigma is the complex growth rate, substitution into the Navier–Stokes equations and linearization yields

\sigma \hat{\mathbf{v}} + (\mathbf{U} \cdot \nabla) \hat{\mathbf{v}} + (\hat{\mathbf{v}} \cdot \nabla) \mathbf{U} = -\nabla \hat{p} + \frac{1}{\mathrm{Re}} \nabla^2 \hat{\mathbf{v}},

along with \nabla \cdot \hat{\mathbf{v}} = 0 (nondimensionalized using scales R_1, \Omega_1 R_1, \rho \Omega_1^2 R_1^2), where \mathbf{U} = (0, U_\theta(r), 0) is the base velocity. This forms an eigenvalue problem for \sigma, with neutral stability at \mathrm{Re}(\sigma) = 0.^[4]

Stability Analysis

Rayleigh's Inviscid Criterion

In the inviscid analysis of Taylor–Couette flow, viscosity is neglected (\nu = 0), reducing the governing equations to the Euler equations in cylindrical coordinates, which describe the motion of an ideal fluid confined between two rotating concentric cylinders. This approximation allows for the examination of centrifugal instabilities arising from the base flow's differential rotation, without diffusive effects. Lord Rayleigh developed this framework in his study of revolving fluids, adapting principles from his earlier work on shear flow stability to curved, axisymmetric geometries.^[13] The key quantity in this analysis is the Rayleigh discriminant, defined as

\Phi(r) = \frac{1}{r^3} \frac{d}{dr} \left[ r^4 \Omega(r)^2 \right],

where \Omega(r) = v_\theta(r)/r denotes the angular velocity of the azimuthal base velocity profile v_\theta(r). This expression is equivalent to \Phi(r) = \frac{2 \Omega(r)}{r^3} \frac{d}{dr} [r^2 \Omega(r)], or in terms of the specific angular momentum L(r) = r v_\theta(r) = r^2 \Omega(r), it simplifies to \Phi(r) = \frac{2 L(r)}{r^3} \frac{d L(r)}{dr}. The discriminant captures the radial variation of angular momentum, which governs the balance of centrifugal forces in the flow.^[13] The stability criterion states that the inviscid base flow is stable to axisymmetric perturbations if \Phi(r) > 0 for all radii r in the annular gap, corresponding to \frac{d L(r)}{dr} > 0 everywhere, meaning angular momentum increases radially outward. Conversely, the flow is unstable if \Phi(r) < 0 at any r, indicating regions where angular momentum decreases outward, leading to centrifugal instability. For the standard Taylor–Couette base flow with inner cylinder radius R_i, outer radius R_o > R_i, inner angular velocity \Omega_i, and outer \Omega_o, the condition for inviscid stability is \mu > \eta^2, where \mu = \Omega_o / \Omega_i is the angular velocity ratio and \eta = R_i / R_o < 1 is the radius ratio; instability occurs for \mu < \eta^2. This threshold marks the onset of centrifugal instability when the inner cylinder rotates sufficiently faster relative to the outer one.^[13] Physically, this criterion reflects an analogy to Rayleigh's inflection point theorem for plane-parallel inviscid shear flows, where stability requires no inflection in the velocity profile to prevent inflectional instabilities. In the rotating case, the curved streamlines introduce centrifugal effects, and the increasing angular momentum outward ensures that inner fluid elements experience weaker centrifugal forces than outer ones, suppressing radial exchanges that could amplify perturbations. For counter-rotating cylinders (\mu < 0), the criterion predicts instability since \mu < \eta^2, though such configurations may exhibit modified dynamics in practice due to other effects.^[13]

Taylor's Viscous Criterion

The linear stability analysis of Taylor–Couette flow incorporates viscous effects by considering small perturbations superimposed on the base circular Couette flow. These perturbations are typically assumed to take the form f(r) \exp(i k z + \sigma t), where f(r) is a radial amplitude function, k is the axial wavenumber, z is the axial coordinate, t is time, and \sigma is the complex growth rate determining stability.^[4] This ansatz leads to an eigenvalue problem governed by the Orr-Sommerfeld equation adapted to cylindrical coordinates, which balances inertial, viscous diffusion, and pressure gradient terms to capture how viscosity damps or modifies the perturbations.^[14] In his seminal 1923 work, G. I. Taylor applied a narrow-gap approximation where the relative gap width \beta = d / R_1 \ll 1 (with d = R_2 - R_1 the gap width and R_1 the inner radius), effectively treating the system as a plane channel flow with centrifugal forcing. For the case of a stationary outer cylinder (\mu = \Omega_2 / \Omega_1 = 0), this yields a critical Taylor number Ta_c \approx 1708, defined as Ta = \frac{\Omega_1 d^2}{\nu} \sqrt{\frac{d}{R_1}}, above which axisymmetric disturbances grow. The corresponding critical axial wavelength at onset is \lambda_c \approx 2d, corresponding to toroidal rolls spanning roughly twice the gap width.^[4] For finite gap widths, where \eta = R_1 / R_2 < 1, the full cylindrical eigenvalue problem must be solved numerically, as curvature effects become significant. These solutions provide the neutral stability curve Ta_c(\eta, \mu), showing that the critical Taylor number increases with decreasing \eta (wider gaps), and the most unstable mode shifts toward non-axisymmetric disturbances or longer wavelengths in some regimes. For instance, as \eta \to 0, Ta_c grows substantially, reflecting enhanced stabilization from geometric effects.^[15] Boundary layer analysis within the viscous framework reveals that diffusion primarily acts near the cylinder walls, where no-slip conditions enforce thin shear layers in the perturbation velocity profiles. While these wall-adjacent regions experience localized destabilization due to the steep base-flow shear enhancing centrifugal driving, the overall viscous damping across the gap provides net stabilization relative to the inviscid limit, raising the onset threshold to a finite critical speed.^[14]

Flow Regimes and Instabilities

Laminar Circular Couette Flow

Laminar circular Couette flow constitutes the stable base state in the Taylor–Couette system at low rotation rates, prior to the onset of hydrodynamic instabilities. This regime features a steady, axisymmetric velocity field that is purely azimuthal, with radial and axial components vanishing as v_r = 0 and v_z = 0, while v_\theta = v_\theta(r). The radial momentum balance in the Navier-Stokes equations requires a pressure gradient that precisely counters the centrifugal force generated by the rotational motion. The exact analytical solution for the azimuthal velocity profile derives directly from the steady, incompressible Navier-Stokes equations under these conditions:

v_\theta(r) = A r + \frac{B}{r},

where the constants are

A = \frac{\Omega_1 R_1^2 - \Omega_2 R_2^2}{R_2^2 - R_1^2}, \quad B = \frac{(\Omega_2 - \Omega_1) R_1^2 R_2^2}{R_2^2 - R_1^2}.

Here, \Omega_1 and \Omega_2 denote the angular velocities of the inner and outer cylinders, respectively, and R_1 and R_2 are their radii. This profile satisfies the no-slip boundary conditions v_\theta(R_1) = \Omega_1 R_1 and v_\theta(R_2) = \Omega_2 R_2. The azimuthal shear stress at the cylinder walls is given by \tau = \rho \nu \left( \frac{d v_\theta}{dr} - \frac{v_\theta}{r} \right), evaluated at r = R_1 or r = R_2; substituting the velocity profile yields a stress that varies as $1/r^2, ensuring constant torque transmission across the annular gap. In practical applications, such as Couette viscometers with the outer cylinder stationary, the torque T measured on the inner cylinder enables determination of the fluid's dynamic viscosity via \eta_\text{visc} = \frac{T (R_2^2 - R_1^2)}{4 \pi \Omega_1 R_1^2 R_2^2 L}, where L is the length (for narrow gaps, this approximates to \eta_\text{visc} \approx \frac{T (R_2 - R_1)}{2 \pi R_1^3 L \Omega_1}). This relation holds well for narrow gaps and underpins precise rheological measurements.^[16] This laminar regime persists for Taylor numbers below the critical value, Ta < Ta_c \approx 1708 in the narrow-gap limit with the outer cylinder fixed.

Taylor Vortex Flow

Taylor vortex flow represents the primary instability pattern that emerges in Taylor–Couette flow when the Taylor number exceeds its critical value, Ta_c, for the case of a rotating inner cylinder and stationary outer cylinder (μ=0). This supercritical bifurcation leads to the formation of axisymmetric, toroidal vortices superimposed on the base azimuthal flow v_θ(r). The vortices consist of pairs of counter-rotating toroidal rolls aligned azimuthally along the cylinder axis, with each pair featuring outflow from the inner cylinder at the midplane between rolls and inflow near the inner cylinder at the roll boundaries. The axial wavelength λ of these vortex pairs is approximately 2d at onset, where d is the gap width between cylinders, resulting in an aspect ratio (axial half-wavelength to radial extent) of roughly 1. These rolls span the full azimuthal extent without net axial velocity, forming closed streamlines that enhance radial mixing while preserving azimuthal dominance. The structure was first theoretically predicted and experimentally observed by injecting dye into the flow, revealing the organized secondary circulation. The onset occurs via a supercritical pitchfork bifurcation at Ta_c, where infinitesimal disturbances grow linearly until nonlinear effects saturate the amplitude at a finite value. This weakly nonlinear behavior is captured by the Landau equation, dA/dt = εA - l|A|^2 A, where A is the complex amplitude of the disturbance, ε ∝ (Ta - Ta_c)/Ta_c measures the supercriticality, and l > 0 ensures saturation at |A|^2 = ε/l. Derived for the small-gap approximation, this amplitude equation predicts the steady, finite-amplitude state of the vortices just above threshold. Visualizations of Taylor vortex flow typically employ dye injection or particle image velocimetry (PIV) to map the secondary flow. Early experiments used aluminum dust or dye streaks to trace closed, toroidal streamlines, confirming no net axial flow and highlighting the paired roll geometry. Modern PIV techniques resolve the velocity field, showing symmetric outflow and inflow regions with recirculation confined between cylinders, without axial propagation. The size and shape of Taylor vortices depend on the Taylor number, with the axial wavelength decreasing from ≈2d at Ta = Ta_c to smaller values (≈1.5d) as Ta increases due to enhanced nonlinear interactions. For μ=0 and radius ratio η ≈ 0.95, this axisymmetric vortex state remains stable up to Ta ≈ 10^4, beyond which secondary instabilities may arise, though the precise upper limit varies slightly with η.

Transitions and Nonlinear Phenomena

Wavy Vortex Flow

Wavy vortex flow arises as a secondary instability of the axisymmetric Taylor vortex flow via a supercritical Hopf bifurcation occurring at a Taylor number approximately twice the critical value for the primary instability, Ta_w ≈ 2 Ta_c. This bifurcation introduces non-axisymmetric traveling waves characterized by azimuthal wavenumbers m typically between 1 and 3, with a propagation speed c ≈ 0.3 Ω_1, where Ω_1 denotes the angular velocity of the inner cylinder. In this regime, the steady toroidal Taylor rolls distort into helical wave structures that propagate azimuthally while maintaining periodicity along the axial direction z. The flow retains the basic toroidal topology but develops time-periodic modulations due to the azimuthal traveling waves, leading to enhanced mixing within the vortex cores. These waves exhibit m-fold rotational symmetry and are observed to be stable over a range of supercritical Taylor numbers. Weakly nonlinear analysis provides amplitude equations that describe the saturation of the primary Taylor vortices and the subsequent selection of the azimuthal wavenumber m at the Hopf bifurcation. These equations reveal a band of stable wavenumbers bounded by Eckhaus instabilities, where long-wavelength modulations in the axial direction can destabilize the wavy state, influencing wave selection and pattern formation. Experimental observations confirm that wavy vortex flow persists as the dominant regime up to Taylor numbers around Ta ≈ 10^5, serving as a precursor to more disordered states through the onset of intermittency and quasiperiodic modulations.

Chaos and Turbulence Transitions

As the control parameter increases beyond the wavy vortex flow regime, Taylor–Couette flow exhibits nonlinear routes to chaos, prominently following the Ruelle–Takens–Newhouse scenario. In this pathway, the quasiperiodic motion characterized by two incommensurate frequencies on an invariant two-torus becomes unstable, leading to the emergence of a third incommensurate frequency and subsequent torus breakdown into a strange attractor. This transition typically occurs around a Taylor number of approximately $10^6, marking the onset of low-dimensional chaos with broadband power spectra emerging from the previously discrete spectral lines.^[17]^[18] At higher Reynolds numbers, intermittency routes to chaos become prominent, particularly in counter-rotating configurations where bursts of turbulent-like motion alternate with laminar puffs. This behavior aligns with Pomeau–Manneville type III intermittency, involving a subcritical Hopf bifurcation followed by reinjection mechanisms that sustain chaotic episodes amid predominantly laminar phases. Such intermittent dynamics are observed for Reynolds numbers exceeding approximately 1000, with the turbulent fraction increasing with the control parameter until full spatiotemporal chaos ensues.^[19]^[18] In fully developed turbulent regimes, above Reynolds numbers of about $10^4, the flow transitions to featureless turbulence or spatiotemporal chaos characterized by defects and broadband noise in power spectra, lacking coherent structures like persistent vortices. This state reflects a loss of spatial order, with energy cascading across scales in a manner analogous to homogeneous turbulence, though influenced by the cylindrical geometry.

Experimental and Numerical Investigations

Historical Experiments

The foundational experimental investigations of Taylor–Couette flow began with Maurice Couette's 1890 study, where he employed a coaxial cylinder apparatus to measure fluid viscosity. By rotating the outer cylinder at constant speed while holding the inner cylinder stationary, Couette quantified the torque required to maintain the motion, thereby confirming the existence of steady azimuthal (circumferential) flow in the annular gap for sufficiently low rotation rates. These measurements provided early empirical validation of the linear velocity profile expected for viscous Couette flow, though Couette noted anomalous increases in drag at higher speeds, which he attributed to turbulence without identifying the underlying instability mechanism. In 1923, Geoffrey Ingram Taylor extended these experiments to explore the stability limits of Couette flow, using a setup with concentric cylinders of varying gap widths and aspect ratios up to 400. Taylor visualized the flow by suspending fine aluminum dust particles in the fluid, which aligned with streamlines to reveal the onset of instability as axisymmetric toroidal vortices when the inner cylinder's angular velocity exceeded a critical threshold. For narrow gaps (radius ratio η ≈ 0.95), the critical Taylor number—defined as Ta = (Ω₁R₁d/ν)² √(d/R₁) where Ω₁ is the inner cylinder speed, R₁ the inner radius, d the gap width, and ν the kinematic viscosity—was experimentally determined to be approximately 1708, aligning closely with his viscous linear stability analysis. Taylor also tested wider gaps, demonstrating that the critical speed decreased with increasing gap-to-radius ratio, and measured the axial wavelength of the vortices to be roughly twice the gap width.^[4] Subsequent experiments from the 1950s through the 1970s built on Taylor's work, employing enhanced visualization techniques such as dye injection and suspended tracer particles to map vortex structures more precisely. These studies confirmed the axial wavelength of Taylor vortices at approximately 2d near onset, with minor variations depending on the radius ratio and boundary conditions. Key findings included the characterization of the primary instability as a supercritical bifurcation, where the vortex amplitude grows smoothly and continuously from zero as Ta exceeds Ta_c, without evidence of subcritical jumps or hysteresis in standard counter-rotating configurations (inner cylinder faster than outer). For instance, measurements of torque and velocity profiles showed that the finite-amplitude vortices saturate nonlinearly, stabilizing the flow just above criticality.

Gollub–Swinney Experiment

The Gollub–Swinney experiment utilized a Taylor–Couette apparatus consisting of two concentric cylinders with the inner cylinder rotating and the outer cylinder counter-rotating at a speed ratio μ = −0.35. The Reynolds number, defined based on the inner cylinder speed and the gap width, was varied up to approximately 2000 to probe the onset of turbulence. Velocity fluctuations were measured using laser Doppler velocimetry, which provided power spectra of the local radial velocity component at various azimuthal and axial positions.^[20] As the Reynolds number increased, the flow exhibited a sequence of instabilities: from stable laminar circular Couette flow, to the onset of Taylor vortex flow (TVF) with stationary axisymmetric rolls, followed by wavy vortex flow (WVF) featuring azimuthal traveling waves superimposed on the vortices. At higher Reynolds numbers, the system entered a quasiperiodic regime characterized by two incommensurate frequencies—one associated with the axial motion of the Taylor rolls and the other with the azimuthal wave propagation—indicating the presence of an invariant torus in phase space. This quasiperiodic state then transitioned to chaotic motion at a critical Reynolds number Re_c ≈ 1200, where the flow became aperiodic and broadband noise dominated the spectra.^[20] The power spectra from these measurements revealed the gradual broadening and eventual breakdown of the quasiperiodic torus, with discrete peaks merging into a continuous spectrum, providing experimental evidence for the Ruelle–Takens route to chaos through successive Hopf bifurcations leading to a strange attractor. Intermittent bursts of chaotic behavior were observed, visually analogous to "fluid donuts" representing localized turbulent structures amid ordered vortex remnants. These findings contradicted the traditional Landau picture of a single supercritical bifurcation to turbulence and instead supported low-dimensional dynamical models for the transition.^[20] Published in Physical Review Letters in 1975, the experiment marked a pivotal contribution to understanding nonlinear instabilities in shear flows and inspired broader applications of dynamical systems theory to fluid mechanics. In the 1980s, the research group extended these investigations to modulated Taylor–Couette flows, introducing periodic forcing to study pattern formation, multiple frequencies, and additional routes to spatiotemporal chaos.^[20]

Modern Numerical Simulations

Modern numerical simulations of Taylor–Couette flow have primarily relied on direct numerical simulation (DNS) techniques that solve the full incompressible Navier-Stokes equations without subgrid-scale modeling, enabling detailed resolution of flow structures across a wide range of Taylor numbers (Ta).^[21] These simulations typically employ spectral methods, such as pseudospectral expansions in Fourier bases, which are particularly effective for periodic domains in the azimuthal and axial directions, allowing efficient computation of instabilities and transitions.^[22] For wide-gap configurations where radius ratios approach unity, finite volume methods on staggered grids have been used to handle non-periodic geometries and capture endwall effects more accurately.^[23] Significant advancements in the 2000s involved pseudospectral codes that achieved resolutions for Ta exceeding 10^8, revealing defect turbulence characterized by intermittent laminar-turbulent regions and large-scale defects in the flow field.^[24] By the 2020s, GPU-accelerated frameworks like nsCouette enabled DNS at Reynolds numbers (Re) greater than 10^5, facilitating studies of high-Re turbulence with reduced computational times through massively parallel implementations. Key results from these simulations include the numerical confirmation of the critical Taylor number Ta_c ≈ 1708 for the onset of Taylor vortex flow, aligning closely with linear stability predictions and early experiments.^[22] Simulations have also demonstrated relaminarization in modulated flows, where traveling wave-like perturbations suppress turbulence and restore laminar states at moderate Re.^[25] Furthermore, the identification of exact coherent structures—unstable, time-independent solutions such as exact vortex pairs—has provided insights into the dynamical skeleton underlying turbulent regimes, including shadowing of turbulent trajectories by these structures.^[26] Despite these advances, challenges persist in the high computational cost of resolving three-dimensional, unsteady flows at extreme Re, often requiring supercomputing resources for sustained accuracy.^[27] Validation against experimental data, such as from the Gollub–Swinney apparatus, remains essential to ensure simulations capture realistic boundary conditions and flow statistics.^[21]

Applications

Classical Applications

One of the primary classical applications of Taylor–Couette flow is in viscometry, where the laminar regime between concentric rotating cylinders serves as the basis for measuring the dynamic viscosity of Newtonian fluids. In a Couette viscometer, the inner cylinder of radius R_i rotates at angular velocity \Omega while the outer cylinder of radius R_o remains stationary, and the torque T required to maintain this rotation is measured over the cylinder length L. The viscosity \eta is then determined from the relation \eta = \frac{T (R_o^2 - R_i^2)}{4 \pi R_i^2 R_o^2 L \Omega}, derived from the exact solution of the Navier-Stokes equations for steady, axisymmetric flow when end effects are negligible. This setup, pioneered in early experiments by Mallock in 1888 and refined by Couette in 1890, provides precise shear rate control and has been widely used for calibrating fluid properties in industrial and laboratory settings.^[28] In rotating machinery, Taylor–Couette-like flows arise in components such as journal bearings and centrifugal pumps, where differential rotation between shafts and housings can induce centrifugal instabilities that increase drag and affect operational efficiency. These instabilities, analogous to the Taylor vortex formation, manifest in wide-gap configurations typical of bearing chambers, leading to enhanced shear stresses and potential vibrations if not mitigated through design optimizations like gap sizing or lubrication adjustments.^[29]^[30] Early studies in the mid-20th century highlighted how such flows contribute to energy losses in turbomachinery, prompting empirical correlations for torque prediction in laminar conditions to guide engineering practice.^[29] Early geophysical models employed the laminar Taylor–Couette flow as a simplified analog for rotating shear in atmospheric and planetary boundary layers, particularly to study zonal wind patterns and vortex formation in limited instability regimes. Pioneering work by Geoffrey Hide in the 1950s and 1960s used rotating heated annular setups, analogous in geometry to Taylor–Couette but including thermal convection, to simulate baroclinic and barotropic instabilities mimicking large-scale atmospheric vortices, such as those in Jupiter's atmosphere, while analyzing pre-turbulent states for tractable predictions.^[31] These models provided foundational insights into angular momentum transport without invoking full nonlinear transitions.^[6] In desalination processes, rotating reverse osmosis membranes configured in a Taylor–Couette geometry enhance mass transfer to mitigate concentration polarization and fouling. By rotating the permeable inner membrane to induce Taylor vortices above the critical Taylor number, the setup improves mixing across the membrane surface, enhancing permeate flux and salt rejection.^[32]^[33] This approach, explored in engineering designs since the early 2000s, leverages the vortical flow to maintain stable operation under high-pressure gradients.

Modern Developments

In recent years, Taylor–Couette flow has been integrated into microfluidic systems to enhance mixing in lab-on-a-chip devices, particularly at low Reynolds numbers where traditional diffusion-limited mixing is inefficient. By exploiting Taylor vortices, these devices generate chaotic advection, improving homogeneity in chemical reactions and biological assays. For instance, PDMS-based microfluidic mixers fabricated in the 2010s utilize the self-excited turbulent-like motion induced by Taylor–Couette geometries to handle wide viscosity ratios, enabling efficient blending without external pumps.^[34] More advanced designs, such as ribbed-wall Taylor–Couette reactors, have demonstrated superior micromixing performance by promoting vortex formation that reduces segregation indices to below 0.1, facilitating applications in point-of-care diagnostics.^[35] In astrophysics, Taylor–Couette flow serves as a laboratory analog for modeling accretion disks around black holes and stars, where differential rotation drives angular momentum transport. The magnetorotational instability (MRI), a key mechanism in these disks, has been experimentally replicated in magnetized Taylor–Couette setups, confirming nonaxisymmetric modes that enhance turbulence and accretion rates.^[36] These studies link MRI growth rates to disk viscosities, with observations showing instability thresholds at magnetic Reynolds numbers around 10–100, providing insights into protoplanetary disk evolution and X-ray binary systems. Recent experiments have further validated the standard MRI, demonstrating its role in sustaining turbulent transport efficiencies up to 0.01 in dimensionless units.^[37] Industrial applications have expanded Taylor–Couette flow into chemical reactors for optimized heat and mass transfer, leveraging controlled vortices to intensify processes without excessive energy input. In these reactors, axial temperature gradients synergistically boost Nusselt numbers by up to 50% compared to laminar flows, enabling precise control in exothermic reactions.^[38] For oil extraction, 2020s simulations using high-speed Taylor–Couette systems assess nanofluid drilling mud performance, revealing drag reductions of 20–30% and energy savings in wellbore operations through enhanced shear-induced mixing.^[39] These configurations outperform conventional stirred tanks in multiphase systems, with Sherwood numbers exceeding 100 for mass transfer enhancement.^[8] Biomedical research employs Taylor–Couette flow to simulate blood dynamics in rotating vessels and pathological conditions like aneurysms, where vortex structures mimic shear stresses on endothelial cells. Couette-type shearing devices, designed for high-throughput analysis, quantify blood trauma indices under controlled rotation, aiding prosthetic valve development with hemolysis rates below 1%.^[40] In aneurysm models, numerical simulations of 2023 reveal counter-rotating vortices akin to Taylor–Couette patterns, correlating wall shear stresses of 0.5–5 Pa with rupture risks via oscillatory indices up to 0.3.^[41] Recent experiments on bacterial suspensions in high-shear Couette flows demonstrate flagella unbundling at shear rates above 100 s⁻¹, preventing collective motion and informing antimicrobial strategies in infected vessels.^[42]

References

[1]
Viscous-Dominated Flows – Introduction to Aerospace Flight Vehicles
Taylor-Couette flow refers to the flow of a viscous fluid between two concentric cylinders, where one or both cylinders may rotate. This configuration is a ...
[2]
Taylor-Couette Flow - Richard Fitzpatrick
This type of flow is generally known as Taylor-Couette flow, after Maurice Couette and Geoffrey Taylor (1886-1975).
[3]
Taylor-Couette Flow An Experiment on Vortex Flow between Two ...
The Couette apparatus was developed by Maurice Couette in 1890 as a means for measuring the viscosity of a fluid at small imposed angular velocities of the ...
[4]
VIII. Stability of a viscous liquid contained between two rotating ...
All experiments so far carried out seem to indicate that in all cases steady motion is possible if the motion be sufficiently slow.
[5]
Taylor–Couette and related flows on the centennial of Taylor's ... - NIH
Jan 30, 2023 · Taylor's ground-breaking linear stability analysis of fluid flow between two rotating cylinders has had an enormous impact on the field of fluid mechanics.Missing: definition | Show results with:definition
[6]
Taylor–Couette and related flows on the centennial of ... - Journals
Mar 13, 2023 · This two-part issue includes review articles and research articles spanning a broad range of contemporary research areas, all rooted in Taylor's landmark paper.
[7]
Taylor‐Couette reactor: Principles, design, and applications
Feb 3, 2021 · In process technology, the Taylor vortex flow (a) is of great interest, as it combines excellent mixing properties with limited back mixing.FUNDAMENTALS OF... · APPLICATION OF THE TCR IN... · CONCLUSION
[8]
[PDF] Taylor–Couette flow for astrophysical purposes
A concise review is given of astrophysically motivated experimental and theoretical research on Taylor–. Couette flow. The flows of interest rotate ...
[9]
Computational modelling of turbulent Taylor–Couette flow for ...
Apr 4, 2022 · The Taylor–Couette system of interest for this paper involves the fluid flow confined between two concentric cylinders, for which the inner ...
[10]
Taylor–Couette turbulence at radius ratio ${\it\eta}=0.5
Jun 23, 2016 · The two geometrical control parameters of TC flow are the ratio {\it\eta} of the inner and outer cylinder radii and the aspect ratio {\it ...
[11]
On the dynamics of revolving fluids - Journals
On the dynamics of revolving fluids. John William Strutt. Google Scholar · Find this author on PubMed · Search for more papers by this author. John William ...
[12]
High-order meshless global stability analysis of Taylor–Couette ...
Jun 6, 2024 · The classical linear stability theory that led to the Orr–Sommerfeld equations assumes that the base flow is parallel. This assumption holds ...
[13]
Taylor–Couette flow in the narrow-gap limit - Journals
Mar 13, 2023 · The Taylor number T can be expressed as T = 𝛺 ⁢ ( ℛ − 𝛺 ) , where 𝛺 (the rotation number) and ℛ (the Reynolds number) in the Cartesian system ...
[14]
Couette Flow - an overview | ScienceDirect Topics
In 1923, Taylor [2] highlighted the existence of a critical rotation speed ωC. Under this rotational speed, the flow, referred to as Couette flow, is steady ...
[15]
A note on Taylor–Couette style flows with two parallel inner cylinders
Sep 30, 2024 · For a narrow gap (⁠ d ≪ r 1 ⁠), the theory predicts T a = 1708 ⁠, well confirmed by experiments, and its value increases with gap width. When ...
[16]
Dynamical instabilities and the transition to chaotic Taylor vortex flow
Apr 19, 2006 · Swinney, H. L., Fenstermacher, P. R. & Gollub, J. P. 1977a Transition to turbulence in circular Couette flow. In Symposium on Turbulent Shear ...
[17]
Routes to turbulence in Taylor–Couette flow - Journals
Mar 13, 2023 · The competition between the stochastic processes of turbulence decay and turbulence proliferation determines the critical Reynolds number for ...
[18]
An experimental observation of a new type of intermittency
A novel hysteretic form of regular intermittency has been observed experimentally in a variant of the Taylor-Couette problem where the fluid flow is ...
[19]
Influence of global rotation and Reynolds number on the large-scale ...
May 7, 2010 · The presence of vortexlike structures at high shear-Reynolds number ( Re S ≳ 10 4 ) in turbulent Taylor–Couette flow with the inner cylinder ...
[20]
[PDF] Direct numerical simulation of Taylor–Couette turbulent flow ...
Oct 2, 2020 · In wall turbulence, a traveling wave-like control is known to decrease the skin- friction drag and induce the relaminarization phenomenon.
[21]
https://www.cambridge.org/core/journals/journal-of-fluid-mechanics/article/direct-numerical-simulation-of-turbulent-taylorcouette-flow/0378D986DC8B02B793A80F2DF8BC35FB
[22]
Direct numerical simulation of turbulent Taylor–Couette flow
May 2, 2007 · Direct numerical simulation (DNS) is used to investigate turbulent Taylor–Couette (TC) flow. A simulation was run for a Reynolds number of ...
[23]
Simulation of Taylor-Couette flow. Part 1. Numerical methods and ...
Apr 20, 2006 · We use a pseudospectral code in which all of the time-splitting errors are removed by using a set of Green functions (capacitance matrix) that ...
[24]
Numerical study on wide gap Taylor Couette flow with flow transition
Nov 20, 2019 · This study aims to investigate the possible sources of nonaxisymmetric disturbances and their propagation mechanism in Taylor Couette flow ...
[25]
Optimal Taylor–Couette flow: direct numerical simulations
Feb 19, 2013 · We numerically simulate turbulent Taylor–Couette flow for independently rotating inner and outer cylinders, focusing on the analogy with turbulent Rayleigh–Bé ...
[26]
Exact coherent structures and shadowing in turbulent Taylor ...
Jul 23, 2021 · We investigate a theoretical framework for modelling fluid turbulence based on the formalism of exact coherent structures (ECSs).
[27]
Direct numerical simulation of Taylor-Couette flow - ScienceDirect.com
Dec 14, 2022 · It is crystal clear that DNS will play an increasing role in exploring the physics of TC flow because of its high accuracy, however, explicit ...
[28]
[PDF] Taylor-Couette flow: The early days.
11. In 1890 Couette published his thesis, which was a lengthy study of viscosity using a pair of cylinders with the outer one rotating and the inner one ...
[29]
Taylor-Couette flow and transient heat transfer inside the annulus air ...
Dec 1, 2017 · Taylor-Couette flow is a very interesting topic in a number of engineering areas including motors and generators, filters, pumps and bearings ...
[30]
The Three‐Dimensional Velocity Distribution of Wide Gap Taylor ...
Mar 9, 2016 · This activity dates back to 1888 and 1890, when Mallock [3, 4] and Couette [5] conducted independent experiments using concentric rotating ...<|control11|><|separator|>
[31]
[PDF] Geophysical flows as dynamical systems: - Imperial College London
Lorenz (1962, 1963b) chose a highly simplified model of Hide's experiment, in which a peri- odic channel in Cartesian coordinates replaced the annular gap ...
[32]
Control of scale formation in reverse osmosis by membrane rotation
In this study, rotating RO, which takes advantage of Taylor-Couette flow instabilities to reduce concentration polarization and membrane fouling, was ...
[33]
[PDF] Rotating reverse osmosis: a dynamic model for flux and rejection
For a given geometry, a rotational speed suffi- cient to generate Taylor vortices in the annulus is essential to maintain high flux as well as high rejection.
[34]
PDMS-based Turbulent Microfluidic Mixer - PubMed
Apr 7, 2015 · We have developed a Y-shaped, turbulent microfluidic mixer made of PDMS and a glass substrate by strong bonding of the substrates to a ...Missing: Taylor- Couette 2010s
[35]
Micromixing Performance in a Taylor–Couette Reactor with Ribbed ...
Jul 10, 2023 · Taylor–Couette flow (TCF) is a special flow phenomenon, which can offer a superior micromixing performance and have significant potential for ...
[36]
Observation of Nonaxisymmetric Standard Magnetorotational ...
We report that, in a magnetized Taylor-Couette flow, nonaxisymmetric SMRI with an azimuthal mode number can be triggered by a free-shear layer in the base flow.
[37]
Experimental confirmation of the standard magnetorotational ...
Jan 14, 2019 · A ubiquitous source of turbulence is thought to be the magnetorotational instability (MRI) as applied to accretion discs: while purely ...
[38]
Thermo-fluid dynamics and synergistic enhancement of heat transfer ...
Mar 13, 2023 · This study experimentally and numerically investigated the thermo-fluid dynamics of Taylor–Couette flow with an axial temperature gradient from the chemical ...
[39]
Simulation of Wellbore Drilling Energy Saving of Nanofluids Using ...
Jul 8, 2021 · A high-speed Taylor–Couette system (TCS) was devised to operate at speeds 0–1600 RPM to simulate power consumption of wellbore drilling using nanofluids.Missing: recovery | Show results with:recovery
[40]
Design Considerations and Flow Characteristics for Couette-Type ...
Jul 7, 2024 · Couette-type blood-shearing devices are used to simulate and then clinically measure blood trauma, after which the results can be used to assist ...
[41]
Numerical Analysis of Stress Force on Vessel Walls in ...
Nov 24, 2023 · The counter-rotating vortex pair have formed in the same manner as the Taylor–Couette flow [24]. The remaining vortices have originated from ...