Simple cases of vorticity flux from moving surfaces

• We examine first the simple,classical, case of sudden tangential acceleration –

Stokes’ first problem.
• Then we examine a related solution – the generation of vorticity flux by an
oscillating wall – Stokes’ second problem.
• These two solutions form the basis of actuation by a moving surface element.
However, we have again to consider the frame of reference and compare the
equation of motion for an actuator in stationary fluid with that for fluid moving
above a stationary surface element.
• We can then consider the effects of a moving surface actuator beneath a boundary
layer – and note that the effects of nonlinearity begin to appear.
• We also need specific mathematical techniques (successive approximations,
perturbation expansions, Taylor series) to provide first-order linear expressions of
non-linear equations.
• Lastly, we come across the related phenomenon (arising through the effects of
nonlinearity) known as “streaming”– in this case, of vorticity.
Sensors & Actuators: Lecture 13&14 1
 The plane wall and sudden tangential acceleration
1 dp
• Initial and boundary conditions with = :P(t)
" dx
t " 0 : u = 0 for all y
t > 0 : u = U(t) for y = 0, u = 0 for y = ∞.
!
"u 1 dp " 2u
! =# + % 2 = #P(t) + & z (2.1)
"t $ dx "y
! "# z " 2# z
=$
"t "y 2 (2.2)
• The sudden
! motion of the plate is not communicated to the whole of the fluid
immediately. Initial conditions ( 0" # t # 0 +) admit an impulse from the plate
acceleration db dt = bo" (t) to the fluid via the pressure gradient, P(t) = "U#(t) :
– this!has a singular solution.
• At wall, strength of initial! vortex sheet, γ = -(U-b ) and σ (t) = -(U-b )δ(t)
0 0 z 0
! y !

U(t)
y=0 b(t)

Sensors & Actuators: Lecture 13&14 2

 The suddenly accelerated plane wall:
Stokes’s first problem.
• A simpler solution that retains the physics is obtained by taking the flow to
be entirely caused by an impulsive start of the wall for all t and P = 0 (initial
vortex sheet soon diffused, γ0 = 0):
!u ! 2u
=" 2
!t !y
such that, for t > 0, u = 0 for all y, u = U0 for y = 0, u = 0 for y = ∞,
and is
U0 & y2 #
* ( y, t ) = exp$$ ' ! (3.1)
1 !
()(t )2 % 4(t "
U0
#S = 1
(!"t ) 2

U0

• N.B. coordinate system fixed with stationary flow at y = ∞.

board›
Sensors & Actuators: Lecture 13&14 3
 Flow near an oscillating flat plate:
Stokes’s second problem.
• If the pressure gradient remains zero, but the wall has a sinusoidal oscillation
u (0, t ) = U 0 cos nt
with other boundary conditions the same, then, we have a further solution of
!u ! 2u
=" 2
!t !y
• The velocity profile is a damped harmonic motion:
u ( y, t ) = U 0e "! cos(nt " ! ) (4.1)

n
that lags the motion of the wall by " = ky = y .
2!
• This constitutes a viscous wave in which the sign of both velocity and vorticity
changes as the wave diffuses outwards.
• The surface vorticity is: 1
& n #2 & '# (4.2)
) S (t ) = U 0 $ ! cos$ nt + !
%( " % 4"
Sensors & Actuators: Lecture 13&14 4
 Stokes’s second problem: solution
Arg º
1 0
2 30
3 60
n
"=y 4 90
2#
5 120
6 150
! 7 180

5 4 3 2 1

n.b. frame of reference at y=∞: u ( y, t ) = U 0e "! cos(nt " ! )

Sensors & Actuators: Lecture 13&14 5
 Oscillating stream above fixed plate - change in
frame of reference
• Consider surface element stationary. Then boundary conditions
are: u0 (0,t) = 0
u0
u0 (",t) = U = U 0 cos nt
• Hence, equation of motion
v0
"u0 " 2 u0 "U
#$ 2 =
" t " y " t (6.1)
!
• with solution: u0 (y,t) = U 0 cos(nt) " u(y,t)
= U 0 cos(nt) " U 0e"# cos( nt " #) (6.2)
!
" 2 u0 " 2u "u0 "U "u
• Note that, by inspection: 2
= # 2
and = #
"y "y "t "t "t
!

• So (6.2) is clearly a solution of (6.1)

! !
Sensors & Actuators: Lecture 13&14 6
 Oscillating stream above fixed plate

Frame of reference on wall: u0 ( y, t ) = U 0 cos(nt ) ! u ( y, t )

Sensors & Actuators: Lecture 13&14 7

 An oscillatory boundary layer -
effects of convection
• Eq. (6.1) with solution (6.2) may be used as a first approximation to the
more realistic problem of a small actuator moving in the surface beneath
a boundary layer. This is physically justified because, at small t, time-
dependent terms and those for the mean shear are much larger than the
convection terms !u0 ! 2u0 !U
"# =
!t !y 2 !t
where, in the freestream (y → ∞) 1 $p $U (8.1)
" %
# $x $t
and u0(0,t)=0, u0 (∞,t)= U(t) = U0cos nt.
• v0 is given by "u0 "v 0
+ =0
"x "y !
• Next we build in the effects of convection by obtaining a second
approximation to !u ! 2u !U !u !u
1 1 0 0
"# 2
=U " u0 (8.2)" v0
! !t !y !x !x !y
where u0 is the first approximation (all terms are known), u1 is the second
approximation, and u1 = 0, y = 0; u1 = 0 at y = ∞.

Sensors & Actuators: Lecture 13&14 8

 Method of successive approximations
• We build a solution u ( x, y, t ) = u0 ( x, y, t ) + u1 ( x, y, t ) where u0 and u1 are
the first and second approximations respectively
• Boundary conditions for the second approximation u1(0,t) = u1(",t)
• Note that the second approximation Eq. (8.2) includes effects of convection.
Therefore, with convection, the freestream (y → ∞) condition, Eq (8.1)
becomes
1 $p $U $U !
" % +U (9.1)
# $x $t $x
and u0 = 0, y = 0; u0 = U(x,t) = U0(x)cos nt at y = ∞

• Original!equation to be solved: "u " 2u "u "u 1 "p

+# 2 + u +v =$
"t "y "x "y % "x
• As a check, add Eq.(8.2) to the solution for the second approximation:
" "2 "u0 "u0 1 "p "U "U
(u0 + u1 ) # $ 2 ( u0 +!u1) + u0 + v 0 %# = +U (9.2)
"t "y "x "y & "x "t "x

Sensors & Actuators: Lecture 13&14 9

!
 Vorticity streaming in an oscillatory
boundary layer II
• First approximation has solution:
u0 ( x, y, t ) = U 0 ( x){cos nt " exp("! ) cos(nt " ! )}
n
where "=y
2!
• This solution is like that for Stokes’s second problem except recall that the
first term derives from the change of coordinate system.
• The terms of the r.h.s. of (8.2) obviously have terms with cos 2 nt which
reduce to terms like cos 2nt and sin 2nt.
• The second approximation has solution:
dU 0 1 ' 2 int
u1 (x, y,t) = U 0 (x) {"1a e + "1b' }
dx n
• where ! 1a denotes the periodic, and ! 1b the steady-state contribution,
respectively, to the second approximation.
board›
!

Sensors & Actuators: Lecture 13&14 10

 Vorticity streaming in an oscillatory
boundary layer III
3
• From the steady-state contribution # 1'b (") = ! so that
4
3 dU 0
u1 ( x, ") = ! U 0
4n dx (6.1)
• Practically, the easiest demonstration of streaming is obtained by using a
circular cylinder undergoing small transverse (or rotary) oscillations.
• Thus fluid flows away from the cylinder in the direction of oscillation and
towards it in the transverse direction.
• This is very a important result: it says that even though the time-averaged
motion of the body is zero, the time average of the resulting fluid motion is
not!
• First approximation comes from the no-slip condition: Stokes’s flow.
• Second approximation comes from the convective terms which produce
streaming. Physically this is an interaction between viscosity and inertia.
board›

Sensors & Actuators: Lecture 13&14 11

Streaming caused by an oscillating cylinder

Sensors & Actuators: Lecture 13&14 12

Sensors & Actuators: Lecture 13&14 13
 Summary
• Rapid accelerations produce, to first order, viscous waves which have zero time
average. They therefore constitute no flux of vorticity.
• The second-order non-linear convection terms become important at longer times
when they lead to inviscid vorticity streaming. The phenomenon also constitutes a
momentum flux.
• Exact solutions to the equations require low Reynolds numbers:
with U d and nd
Re = m St =
! Um
U m2 Re
a streaming Reynolds number is: Re s = =
n! St
• Typically, Re ! 10, St ! 10, Re s ! 1 for boundary layer approximation to be
valid.
• If the oscillations are large, or the Reynolds number is large, these analyses break
down. However, qualitatively, the effect of additional momentum flux that arises
through the non-linear terms will arise at any Reynolds number – Reynolds
stresses.

Sensors & Actuators: Lecture 13&14 14

 References
Lighthill, J. Introduction. Boundary layer theory. Chapter 2. Laminar Boundary Layers (ed. L. Rosenhead).
Dover Publications, 1963.
Lockerby, D. A, Carpenter, P. W. and Davies, C. 2002 Numerical simulation of the interaction of
microactuators and boundary layers. AIAA J. 40, 67.
Lyman, F. A. 1990 Vorticity production at a solid boundary. Appl. Mech. Rev. 43, 157.
Lighthill, J. 1978 Acoustic streaming. J. Sound & Vibration. 61, 391.
Ligrani, P. M. 2001 Flow structure due to dimple depressions on channel surface. Phys. Fluids 13, 3442.
Panton, R. L. 1984 Incompressible Flows. Wiley Interscience.
Schlichting, H. Boundary-Layer Theory. Seventh Edition. McGraw-Hill, 1987.
Stuart, J. T. Unsteady boundary layers. Chapter 7. Laminar Boundary Layers (ed. L. Rosenhead). Dover
Publications, 1963.
Wu, J. Z., Wu, X. H. and Wu, J. M. 1993 Streaming vorticity flux from oscillating walls with finite amplitude.
Phys. Fluids 5 1933.

Sensors & Actuators: Lecture 13&14 15