Math 705: Fall 2002
HOMEWORK .
Notes and exercises
-
9/9/02 lecture notes [PDF] by Hao Lu and Fabian
Waleffe.
- 9/11/2002: See B3.1 (i.e. Batchelor 3.1) for evolution of line and
surface elements.
- Assume that the scalar function
f(x,t) is conserved along particle trajectories, i.e. Df/Dt=0. The
gradient of f, grad f, is a vector. Does it evolve like a line element,
a surface element or neither?
- We have calculated the time rate of change of the integral over a
material volume. Calculate the time rate of change of the circulation over
a closed material curve. The circulation is the line integral of
v.dx where v is the velocity vector.
- Sketch the streamlines and calculate the vorticity and rate-of-strain tensor for (a) a rigid body
rotation u= (a y, -ax, 0),
(b) a shear flow u= (a y, 0 , 0), (c) a stagnation point flow
u= (a x, -ay, 0). (Note: (x,y,z) are Cartesian coordinates).
Does vorticity imply rotation?
- 9/13/2002: Conservation of mass and momentum. Concept of stress tensor.
- 9/16/2002: Symmetry of stress tensor. Indicial notation. Governing
equations.
- Derive the equation for angular momentum from Newton's law as we derived
for a material volume, conservation of mass and the Reynolds transport theorem.
- Deduce symmetry of the stress tensor from the
angular momentum equation as in Batchelor 1.3 and Acheson exercise 6.14.
- In the `cheap derivation' of the stress tensor symmetry
we assumed that the resultant of the forces on the sides of
our little triangular region applies at the center of each side
and that the moment (torque) of those forces about the middle point
was of higher order. Justify.
- Use conservation of mass to show that the left-hand side of the
momentum equation can be rewritten as: rho Dv/Dt.
- W 9/18/2002: Derivation of momentum equation using fixed control volume.
Macroscopic momentum flux: rho vv, a 2nd order tensor.
Quick look at kinetic theory of gases, distribution function
f(x,v,t), definition of macroscopic density, momentum and
stresses, Boltzmann's equation, Maxwell-Boltzmann equilibrium distribution.
- F 9/20/2002: Parametrization of the stress tensor
in terms of the velocity gradient: Newtonian fluid. Navier-Stokes
equations for an incompressible flow.
- Using the Newtonian flow expression for the stress tensor,
determine the stress on a surface element with normal n
in terms of velocity derivatives and the pressure.
- Show that the divergence of the Newtonian stress tensor
becomes the Laplacian of velocity for an incompressible flow.
- Show that the flows in exercise 3 of 9/11/2002 satisfy the
Navier-Stokes equations and determine the pressure for each flow.
(these flows extend to infinity where the velocity is infinite,
therefore they are physically valid only in a local sense).
- M 9/23/2002: Poisson equation for the pressure. No-slip
boundary condition. Plane Couette flow.
- For flow in a closed domain D, with the velocity prescribed
for all times on the boundary of D, determine the boundary
condition for the pressure. (Hint: take the dot product of
the momentum equation with the normal to the boundary and
deduce a Neumann boundary condition for the pressure. Hence
the pressure is known up to an arbitrary constant.)
- Consider incompressible flow between two infinite parallel plane
separated by a distance H. The velocity is zero then at time t=0
the upper plane moves parallel to itself at constant velocity U.
Find a solution of the Navier-Stokes equation in the form (u(y,t),0,0),
where y is perpendicular to the planes and u is in the direction
of the wall motion. What is the stress on the wall?
- Same as previous problem but adapted to flow between two
concentric cylinder. At time 0, the outer cylinder impulsively starts
rotating at constant speed. What is the stress on the outer cylinder?
the inner cylinder?
- W 9/25/2002: Impulsively started shear flow in a bounded channel
(problem 2 of 9/23/2002). Convective and diffusive time scales,
non-dimensionalization, Reynolds number. Semi-infinite domain: self-similar
solution.
- F 9/27/2002: Impulsively started wall in a semi-infinite domain.
Self-similar solutions.
- Show that the (heat) equation ut= uyy
admits self-similar solutions of the form u(y,t)=t-mu f(eta)
where eta=y/(t)1/2 and mu is any constant.
(Note that "t" stands here for "nu t", i.e. the kinematic viscosity
has been absorbed into "t" which now has units of length-squared!)
Each value of mu leads to a different self-similar solution.
(These various solutions can be obtained by taking derivatives of
the fundamental Gaussian solution
exp(-eta2/4)/ sqrt(4 pi t) that corresponds to
mu=1/2).
- A deeper look:
Do the change of variables u(y,t)=v(eta,t).
- What is the equation
for v(eta,t)? Note that this equation is an advection-diffusion equation
for the passive scalar v . That means that in the (t,eta) plane, v is given by
a balance between advection by the "velocity field" (t,-eta/2) [a stagnation-point
type "flow"] and diffusion in the eta direction, and v (which happens to be
a velocity in the (x,y) plane) does not influence the
"velocity" (t,-eta/2) in the (t,eta) plane, so it is a "passive scalar"
(as opposed to an "active scalar" that influences the flow).
Note that the flow is `squeezing' v toward eta=0, while diffusion wants
to spread v in the eta direction.
- Separate variables and show that the v(eta,t) equation
admits solutions of the form t-mu f(eta). Show that
the f(eta) equation is a Sturm-Liouville problem. The boundary
conditions are that f(eta) -> 0 "sufficiently fast" as eta -> +/- infinity
(how fast is that?). Deduce that the eigenvalues mu are positive (and
they form a countable infinite set, etc...)
- Find a solution for mu=1/2. Show that the n-th derivative of that
fundamental solution is a solution of the same equation but with mu=(n+1)/2.
This give you all the solutions as derivatives of a Gaussian. Therefore
each of these solutions has the form of a polynomial times a Gaussian.
- So, let f(eta)=g(eta) exp(-eta2/4) and derive the equation for
g(eta). Change variables: eta= 2 xi, and show that g(xi) satisfies
Hermite's equation g'' - 2 xi g' + L g =0, where the primes
now denote derivatives with respect to xi, not eta. Relate L to mu. Show that Hermite's
equation has polynomial solutions (called Hermite polynomials of course)
by looking for a Taylor series solution and showing that it terminates
when L=2n (n a positive integer).
- Take a good look at all exercises in Acheson Chapter 2
and chapter 6.
- M 9/30/2002: Rotational form of NSE, vorticity equation, 2D-3C flows,
2D-2C flows.
- Find a 2D-3C solution of the NSE of the form [u(y),-a y, a z]
with u(y) --> +/- U as y --> +/- infinity, where "a" is a positive constant.
Note the similarity with the v(eta,t)
equation above. Your solution is a stretched vortex sheet. Vortex stretching
(and/or vorticity advection)
balances diffusion. Sketch the flow. This is the planar analog of the axisymmetric
Burgers' vortex (Acheson sect. 5.9). All of these simple flows are
useful in a local sense.
W 10/2/2002: vortex stretching; 2D flows.
F 10/4/2002: Solved and discussed 9/30/2002 exercise + motivated Acheson 2.14
October-November: Kelvin-Helmholtz Instability.
"Unsteady Bernoulli". Linear inviscid stability of
shear flows (Rayleigh's equation, Rayleigh and Fjortoft's theorems).
Vortex dynamics: streamfunction-vorticity formulation, Green's functions
and point vortices. Method of images. Infinite row of vortices.
Stability of an infinite row. Stuart vortices.
Rayleigh-Taylor instability.
Surface gravity-capillary waves.
- Acheson 5.8-14.
- Consider a spherical gas bubble of radius R(t) immersed in an infinite
incompressible fluid. Assuming that R(t) is known, find the velocity induced
in the fluid by the bubble oscillations
and the pressure at the surface of the bubble on the liquid side.
- Verify that the Stuart vortices solution seen in class,
Psi=ln[cosh(y/d)-sqrt(e2-1)/e cos (x/d)], e > 1,
is indeed
a solution of the steady 2D, Euler equations. Show that
the Laplacian of the streamfunction is a function of the
streamfunction only and determine that function.
- Study the Rayleigh equation for the linear stability of the
U(y)=tanh(y/d) profile in the real plane,
where "d" is a length parameter. Show that there
is a neutral mode with x-wavenumber 1/d.
- Acheson 3.1; 3.7; 3.8;
W 11/20/2002: Line tension
F 11/22/2002: Surface tension
- Find the form of an air/water interface in a semi-infinite domain
with one vertical wall and gravity acting in the vertical. The "contact angle"
is given and less than pi/2. How high does the fluid "climb the wall"?
- Same question but for a small channel (i.e. space between two vertical
walls). Assume the walls are close enough that the interface is approximately
cylindrical. The walls dip in the water a little bit only, so there are
two outside interfaces, one on each side of the channel,
that are exactly as in the previous problem. How high does the fluid climb?
Estimate how small the channel must be for the cylindrical approximation
to be a good one. Repeat this exercise for a small cylindrical tube,
with an approximately spherical interface.
- [Exercise from Batchelor]
A rigid sphere of radius R rests on a flat rigid surface, and a small amount
of liquid surrounds the point of contact making a concave-planar lens
whose diameter is small compared with R. The angle of contact of the liquid
with each of the solid surfaces is zero, and the tension in the air-liquid
interface is T. Show that there is an adhesive force of magnitude [4 pi RT]
acting on the sphere. (The fact that this adhesive force is independent
of the volume of liquid is noteworthy).
M 11/25/2002: Rayleigh-Plateau Instability.
M 11/27/2002: Collision of equal and opposite jets: Bernoulli,
sheet thickness, maximal sheet radius; Water bells and sheets.
M 12/02/2002: Intro to instabilities and bifurcations: Landau's equation.
- Consider Landau's equation [dA/dt=sigma A - lambda |A|2 A]
with sigma=(R-Rc) + i omega, and lambda complex. Use the polar
representation A=|A| exp(i phi) where i2=-1 and
phi is real. Derive the evolution equations for |A| and phi. When
omega=0 this is essentially the "normal form" for the pitchfork
bifurcation and when omega is not zero it's the normal form for the
Hopf bifurcation where the steady state A=0 bifurcates to a periodic
solution (limit cycle).