Stokes flow

From Wikipedia, the free encyclopedia
Jump to: navigation, search
An object moving through a gas or liquid experiences a force in direction opposite to its motion. Terminal velocity is achieved when the drag force is equal in magnitude but opposite in direction to the force propelling the object. Shown is a sphere in Stokes flow, at very low Reynolds number.

Stokes flow (named after George Gabriel Stokes), also named creeping flow or creeping motion,[1] is a type of fluid flow where advective inertial forces are small compared with viscous forces.[2] The Reynolds number is low, i.e. \textit{Re} \ll 1. This is a typical situation in flows where the fluid velocities are very slow, the viscosities are very large, or the length-scales of the flow are very small. Creeping flow was first studied to understand lubrication. In nature this type of flow occurs in the swimming of microorganisms and sperm[3] and the flow of lava. In technology, it occurs in paint, MEMS devices, and in the flow of viscous polymers generally.

The equations of motion for Stokes flow, called the Stokes Equations, are a linearization of the Navier-Stokes Equations, and thus can be solved by a number of well-known methods for linear differential equations.[4] The primary Green's function of Stokes flow is the Stokeslet, which is associated with a singular point force embedded in a Stokes flow. From its derivatives other fundamental solutions can be obtained.[5]

The fundamental solution due to a point force in a steady Stokes flow was first derived by the Nobel Laureate, Lorentz, as far back as 1896. This solution is now known by the name Stokeslet, although Stokes never knew about it. The name Stokeslet was coined by Hancock in 1953. The closed-form fundamental solutions for generalized unsteady Stokes and Oseen flows associated with arbitrary time-dependent translational and rotational motions have been derived for Newtonian[6] and micropolar[7] fluids.

Stokes equations[edit]

The equation of motion for Stokes flow can be obtained by linearizing the steady state Navier-Stokes Equations. The inertial forces are assumed to be negligible in comparison to the viscous forces, and eliminating the inertial terms of the momentum balance in the Navier–Stokes equations reduces it to the momentum balance in the Stokes equations:[1]

\boldsymbol{\nabla} \cdot \mathbb{P} + \mathbf{f} = 0

where \scriptstyle \mathbb{P} is the Cauchy stress tensor representing viscous and pressure stresses,[8][9] and \scriptstyle \mathbf{f} an applied body force. The full Stokes equations also includes an equation for the conservation of mass, commonly written in the form:

 \frac{d\rho}{dt} +  \rho\nabla\cdot\mathbf{u} = 0

where \scriptstyle \rho is the fluid density and \scriptstyle \mathbf{u} the fluid velocity. To obtain the equations of motion for incompressible flow, it is assumed that the density, \scriptstyle \rho, is a constant.

Furthermore, occasionally one might consider the unsteady Stokes equations, in which the term \scriptstyle \rho \frac{\partial\mathbf{u}}{\partial t} is added to the left hand side of the momentum balance equation.[1]


The Stokes equations represent a considerable simplification of the full Navier–Stokes equations, especially in the incompressible Newtonian case.[2][4][8][9] They are the leading-order simplification of the full Navier–Stokes equations, valid in the distinguished limit Re \to 0.

A Stokes flow has no dependence on time other than through time-dependent boundary conditions. This means that, given the boundary conditions of a Stokes flow, the flow can be found without knowledge of the flow at any other time.
An immediate consequence of instantaneity, time-reversibility means that a time-reversed Stokes flow solves the same equations as the original Stokes flow. This property can sometimes be used (in conjunction with linearity and symmetry in the boundary conditions) to derive results about a flow without solving it fully. Time reversibility means that it is difficult to mix two fluids using creeping flow.

While these properties are true for incompressible Newtonian Stokes flows, the non-linear and sometimes time-dependent nature of non-Newtonian fluids means that they do not hold in the more general case.

Stokes paradox

An interesting property of Stokes flow is known as the Stokes' paradox: that there can be no Stokes flow of a fluid around a disk in two dimensions; or, equivalently, the fact there is no non-trivial solution for the Stokes equations around an infinitely long cylinder.[10]

Demonstration of time-reversibility[edit]

A Taylor–Couette system can create laminar flows in which concentric cylinders of fluid move past each other in an apparent spiral.[11] A fluid such as corn syrup with high viscosity fills the gap between two cylinders, with colored regions of the fluid visible through the transparent outer cylinder. The cylinders are rotated relative to one another at a low speed, which together with the high viscosity of the fluid and thinness of the gap gives a low Reynolds number, so that the apparent mixing of colors is actually laminar and can then be reversed to approximately the initial state. This creates a dramatic demonstration of seemingly mixing a fluid and then unmixing it by reversing the direction of the mixer.[12][13][14]

Incompressible flow of Newtonian fluids[edit]

In the common case of an incompressible Newtonian fluid, the Stokes equations take the (vectorized) form:

 \begin{align} \mu \nabla^2 \mathbf{u} -\boldsymbol{\nabla}p + \mathbf{f} &= 0 \\
 \boldsymbol{\nabla}\cdot\mathbf{u}&=0 \end{align}

Where \scriptstyle \mathbf{u} is the velocity of the fluid, \scriptstyle \boldsymbol{\nabla} p is the gradient of the pressure, \scriptstyle \mu is the dynamic viscosity, and \scriptstyle \mathbf{f} an applied body force. The resulting equations are linear in velocity and pressure, and therefore can take advantage of a variety of linear differential equation solvers.[4]

Cartesian coordinates[edit]

With the velocity vector expanded as \scriptstyle \mathbf{u}=(u,v,w) and similarly the body force vector \scriptstyle \mathbf{f} = (f_x, f_y, f_z) , we may write the vector equation explicitly,

  \begin{align}  \mu \left(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2} + \frac{\partial^2 u}{\partial z^2}\right) - \frac{\partial p}{\partial x} + f_x &= 0 \\
  \mu \left(\frac{\partial^2 v}{\partial x^2}  + \frac{\partial^2 v}{\partial y^2} + \frac{\partial^2 v}{\partial z^2}\right) - \frac{\partial p}{\partial y} + f_y &= 0 \\
  \mu \left(\frac{\partial^2 w}{\partial x^2}  + \frac{\partial^2 w}{\partial y^2} + \frac{\partial^2 w}{\partial z^2}\right) - \frac{\partial p}{\partial z} + f_z &= 0 \\
  {\partial u \over \partial x} + {\partial v \over \partial y} + {\partial w \over \partial z} &= 0 \end{align}

We arrive at these equations by making the assumptions that \scriptstyle \mathbb{P} =\frac{1}{2}(\boldsymbol{\nabla}\mathbf{u} + (\boldsymbol{\nabla}\mathbf{u})^T) - p\mathbb{I} and the density \rho is a constant.[8]

Methods of solution[edit]

By stream function[edit]

The equation for an incompressible Newtonian Stokes flow can be solved by the stream function method in planar or in 3-D axisymmetric cases

Type of function Geometry Equation Comments
Stream function (\psi) 2-D planar \nabla^4 \psi = 0 or \Delta^2 \psi = 0 (biharmonic equation) \Delta is the Laplacian operator in two dimensions
Stokes stream function (\Psi) 3-D spherical E^2 \Psi = 0, where E = {\partial^2 \over \partial r^2} + {\sin{\theta} \over r^2} {\partial \over \partial \theta} \left({ 1 \over \sin{\theta}}  {\partial \over \partial \theta}\right) For derivation of the E operator see Stokes stream function#Vorticity
Stokes stream function (\Psi) 3-D cylindrical L_{-1}^2 \Psi = 0, where L_{-1} = \frac{\partial^2}{\partial z^2} + \frac{\partial^2}{\partial \rho^2} - \frac{1}{\rho}\frac{\partial}{\partial\rho} For L_{-1} see [15]

By Green's function: the Stokeslet[edit]

The linearity of the Stokes equations in the case of an incompressible Newtonian fluid means that a Green's function, \mathbb{J}(\mathbf{r}), exists. The Green's function is found by solving the Stokes equations with the forcing term replaced by a point force acting at the origin, and boundary conditions vanishing at infinity:

 \begin{align} \mu \nabla^2 \mathbf{u} -\boldsymbol{\nabla}p &=- \mathbf{F}\cdot\mathbf{\delta}(\mathbf{r})\\
 \boldsymbol{\nabla}\cdot\mathbf{u}&=0 \\ 
|\mathbf{u}|, p &\to 0 \quad \mbox{as} \quad r\to\infty \end{align}

where \mathbf{\delta}(\mathbf{r}) is the Dirac delta function, and \mathbf{F}\cdot\delta(\mathbf{r}) represents a point force acting at the origin. The solution for the pressure p and velocity u with |u| and p vanishing at infinity is given by[1]

 \mathbf{u}(\mathbf{r}) = \mathbf{F} \cdot \mathbb{J}(\mathbf{r}), \qquad
p(\mathbf{r}) = \frac{\mathbf{F}\cdot\mathbf{r}}{4 \pi |\mathbf{r}|^3}


\mathbb{J}(\mathbf{r}) = {1 \over 8 \pi \mu} \left( \frac{\mathbb{I}}{|\mathbf{r}|} + \frac{\mathbf{r}\mathbf{r}}{|\mathbf{r}|^3} \right)

is a second-rank tensor (or more accurately tensor field) known as the Oseen tensor (after Carl Wilhelm Oseen).

The terms Stokeslet and point-force solution are used to describe \mathbf{F}\cdot\mathbb{J}(\mathbf{r}). Analogous to the point charge in Electrostatics, the Stokeslet is force-free everywhere except at the origin, where it contains a force of strength \mathbf{F}.

For a continuous-force distribution (density) \mathbf{f}(\mathbf{r}) the solution (again vanishing at infinity) can then be constructed by superposition:

\mathbf{u}(\mathbf{r}) = \int \mathbf{f}(\mathbf{r'}) \cdot \mathbb{J}(\mathbf{r} - \mathbf{r'}) \mathrm{d}\mathbf{r'}, \qquad
p(\mathbf{r}) = \int \frac{\mathbf{f}(\mathbf{r'})\cdot(\mathbf{r}-\mathbf{r'})}{4 \pi |\mathbf{r}-\mathbf{r'}|^3} \, \mathrm{d}\mathbf{r'}

This integral representation of the velocity can be viewed as a reduction in dimensionality: from the three-dimensional partial differential equation to a two-dimensional integral equation for unknown densities.[1]

By Papkovich–Neuber solution[edit]

The Papkovich–Neuber solution represents the velocity and pressure fields of an incompressible Newtonian Stokes flow in terms of two harmonic potentials.

By boundary element method[edit]

Certain problems, such as the evolution of the shape of a bubble in a Stokes flow, are conducive to numerical solution by the boundary element method. This technique can be applied to both 2- and 3-dimensional flows.

Stokes flow in particular geometries[edit]

Hele-Shaw flow[edit]

Hele-Shaw flow is an example of a geometry for which inertia forces are negligible. It is defined by two parallel plates arranged very close together with the space between the plates occupied partly by fluid and partly by obstacles in the form of cylinders with generators normal to the plates.[8]

Slender-body theory[edit]

Slender-body theory in Stokes flow is a simple approximate method of determining the irrotational flow field around bodies whose length is large compared with their width. The basis of the method is to choose a distribution of flow singularities along a line (since the body is slender) so that their irrotational flow in combination with a uniform stream approximately satisfies the zero normal velocity condition.[8]

Lamb's solution in spherical coordinates[edit]

Lamb's general solution arises from the fact that the pressure p satisfies the Laplace equation, and can be expanded in a series of solid spherical harmonics in spherical coordinates. As a result, the solution to the Stokes equations can be written:

 \begin{align} \mathbf{u} &= \sum_{n=-\infty, n\neq1}^{n=\infty} \left[ \frac{(n+3)r^2\nabla p_n}{2\mu(n+1)(2n+3)} - \frac{n\mathbf{x}p_n}{\mu(n+1)(2n+3)}\right] + \sum_{n=-\infty}^{n=\infty} [\nabla\Phi_n + \nabla \times (\mathbf{x}\chi_n)] \\
p &= \sum_{n=-\infty}^{n=\infty}p_n  \end{align}

where p_n, \Phi_n, and \chi_n are solid spherical harmonics of order n:

\begin{align} p_n &= r^n \sum_{m=0}^{m=n} P_n^m(\cos\theta)(a_{mn}\cos m\phi +\tilde{a}_{mn} \sin m\phi) \\
\Phi_n &= r^n \sum_{m=0}^{m=n} P_n^m(\cos\theta)(b_{mn}\cos m\phi +\tilde{b}_{mn} \sin m\phi) \\
\chi_n &= r^n \sum_{m=0}^{m=n} P_n^m(\cos\theta)(c_{mn}\cos m\phi +\tilde{c}_{mn} \sin m\phi) \end{align}

and the P_n^m are the associated Legendre polynomials. The Lamb's solution can be used to describe the motion of fluid either inside or outside a sphere. For example, it can be used to describe the motion of fluid around a spherical particle with prescribed surface flow, a so-called squirmer, or to describe the flow inside a spherical drop of fluid. For interior flows, the terms with n<0 are dropped, while for exterior flows the terms with n>0 are dropped (often the convention n\to -n-1 is assumed for exterior flows to avoid indexing by negative numbers).[1]

Theorems about Stokes flow[edit]

Stokes' law (Stokes' drag)[edit]

The Stokes' law for the resistance to a moving sphere, also known as Stokes' drag, is a relation describing the drag force on a sphere exerted by the surrounding fluid in a Stokes flow. Given a sphere of radius a, travelling at velocity U, in a Stokes fluid with dynamic viscosity \mu, the drag force F_D is given by:[8]

 F_D = 6 \pi \mu a U

Minimum energy dissipation theorem[edit]

According to the minimum energy dissipation theorem, the Stokes solution dissipates less energy than any other solenoidal vector field with the same boundary velocities.[1]

Lorentz reciprocal theorem[edit]

The Lorentz reciprocal theorem states a relationship between two Stokes flows in the same region. Consider fluid filled region V bounded by surface S. Let the velocity fields \mathbf{u} and \mathbf{u}' solve the Stokes equations in the domain V, each with corresponding stress fields \mathbf{\sigma} and \mathbf{\sigma}'. Then the following equality holds:

 \int_S \mathbf{u}\cdot (\mathbf{\sigma}' \cdot \mathbf{n}) dS = \int_S \mathbf{u}' \cdot (\mathbf{\sigma} \cdot \mathbf{n}) dS

Where \mathbf{n} is the unit normal on the surface S. The Lorentz reciprocal theorem can be used to show that Stokes flow "transmits" unchanged the total force and torque from an inner closed surface to an outer enclosing surface.[1] The Lorentz reciprocal theorem can also be used to relate the swimming speed of a microorganism, such as cyanobacterium, to the surface velocity which is prescribed by deformations of the body shape via cilia or flagella.[16]

Faxén's laws[edit]

The Faxén's laws are direct relations that express the multipole moments in terms of the ambient flow and its derivatives. First developed by Hilding Faxén to calculate the force, \mathbf{F}, and torque, \mathbf{T} on a sphere, they took the following form:

\begin{align} \mathbf{F} &= 6\pi\mu a \left( 1 + \frac{a^2}{6}\nabla^2 \right) \mathbf{v}^{\infty}(\mathbf{x})|_{x=0} - 6\pi\mu a \mathbf{U} \\ 
\mathbf{T} &= 8\pi\mu a^3(\mathbf{\Omega}^{\infty}(\mathbf{x}) - \mathbf{\omega})|_{x=0}  \end{align}

where \mu is the dynamic viscosity, a is thee particle radius, \mathbf{v}^{\infty} is the ambient flow, \mathbf{U} is the speed of the particle, \mathbf{\Omega}^{\infty} is the angular velocity of the background flow, and \mathbf{\omega} is the angular velocity of the particle.

The Faxén's laws can be generalized to describe the moments of other shapes, such as ellipsoids, spheroids, and spherical drops.[1]

See also[edit]


  1. ^ a b c d e f g h i Kim, S. & Karrila, S. J. (2005) Microhydrodynamics: Principles and Selected Applications, Dover. ISBN 0-486-44219-5.
  2. ^ a b Kirby, B.J. (2010). Micro- and Nanoscale Fluid Mechanics: Transport in Microfluidic Devices. Cambridge University Press. ISBN 978-0-521-11903-0. 
  3. ^ Dusenbery, David B. (2009). Living at Micro Scale. Harvard University Press, Cambridge, Mass. ISBN 978-0-674-03116-6.
  4. ^ a b c Leal, L.G. (2007). Advanced Transport Phenomena:Fluid Mechanics and Convective Transport Processes. 
  5. ^ Chwang, A. and Wu, T. (1974). "Hydromechanics of low-Reynolds-number flow. Part 2. Singularity method for Stokes flows". J. Fluid Mech. 62(6), part 4, 787–815.
  6. ^ Shu, Jian-Jun; Chwang, A.T. (2001). "Generalized fundamental solutions for unsteady viscous flows". Physical Review E 63 (5): 051201. arXiv:1403.3247. Bibcode:2001PhRvE..63e1201S. doi:10.1103/PhysRevE.63.051201. 
  7. ^ Shu, Jian-Jun; Lee, J.S. (2008). "Fundamental solutions for micropolar fluids". Journal of Engineering Mathematics 61 (1): 69–79. arXiv:1402.5023. Bibcode:2008JEnMa..61...69S. doi:10.1007/s10665-007-9160-8. 
  8. ^ a b c d e f Batchelor, G. K. (2000). Introduction to Fluid Mechanics. 
  9. ^ a b Happel, J. & Brenner, H. (1981) Low Reynolds Number Hydrodynamics, Springer. ISBN 90-01-37115-9.
  10. ^ Lamb, Horace (1945). Hydrodynamics (Sixth ed.). New York: Dover Publications. pp. 602–604. 
  11. ^ C. David Andereck, S. S. Liu and Harry L. Swinney (1986). Flow regimes in a circular Couette system with independently rotating cylinders. Journal of Fluid Mechanics, 164, pp 155–183 doi:10.1017/S0022112086002513
  12. ^ Dusenbery, David B. (2009). Living at Micro Scale, pp.46. Harvard University Press, Cambridge, Mass. ISBN 978-0-674-03116-6.
  13. ^
  14. ^
  15. ^ Payne, LE; WH Pell (1960). "The Stokes flow problem for a class of axially symmetric bodies". Journal of Fluid Mechanics 7 (04): 529–549. Bibcode:1960JFM.....7..529P. doi:10.1017/S002211206000027X. 
  16. ^ Stone, Howard A.; Samuel, Aravinthan D. T. (November 1996). "Propulsion of Microorganisms by Surface Distorsions". Physical Review Letters. 19 77: 4102–4104. Bibcode:1996PhRvL..77.4102S. doi:10.1103/PhysRevLett.77.4102. 
  • Ockendon, H. & Ockendon J. R. (1995) Viscous Flow, Cambridge University Press. ISBN 0-521-45881-1.

External links[edit]