Boxes & Loops in Circles & Ovals, Billiards & Ballyards, - - PowerPoint PPT Presentation

boxes loops in circles ovals billiards ballyards
SMART_READER_LITE
LIVE PREVIEW

Boxes & Loops in Circles & Ovals, Billiards & Ballyards, - - PowerPoint PPT Presentation

Boxes & Loops in Circles & Ovals, Billiards & Ballyards, Squircles & Squovals Peter Lynch School of Mathematics & Statistics University College Dublin New Trends in Applied Geometric Mechanics Celebrating Darryl Holms


slide-1
SLIDE 1

Boxes & Loops in Circles & Ovals, Billiards & Ballyards, Squircles & Squovals

Peter Lynch School of Mathematics & Statistics University College Dublin New Trends in Applied Geometric Mechanics Celebrating Darryl Holm’s 70th birthday ICMAT, Madrid, 3–7 July 2017

slide-2
SLIDE 2

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-3
SLIDE 3

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-4
SLIDE 4

Boxes & Loops

The familiar phase portrait of a simple pendulum shows how a separatrix divides the phase plane into two regions: The two regions correspond to libration and rotation. In many dynamical systems there is a similar separation

  • f the phase plane into orbits known as boxes and loops.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-5
SLIDE 5

Boxes & Loops

In many dynamical systems there is a similar separation Into two types of orbits, known as boxes and loops. This is seen in elliptical billiards, astrodynamics, rigid body mechanics and many other systems. We will discuss this phenomenon and illustrate it with a variety of examples.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-6
SLIDE 6

Meeting Darryl: My Good Fortune

I Met Darryl at INI (AOD Programme) in 1996. I Darryl and family in Dublin, July 1999.

  • We worked together on Swinging Spring.

I I Visited Los Alamos in Sep/Oct 2000.

  • Darryl found the 3-wave Equations.

I IMA Workshop, Minnesota, February 2002. I Rock-n-roller. Innumerable emails. I Recently: Numerous visits to Imperial College.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-7
SLIDE 7

Quaternion Plaque on Hamilton’s Bridge

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-8
SLIDE 8

Hamilton’s Bridge in Dublin

Figure : Darryl and Justine in Dublin, 1999?

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-9
SLIDE 9

Sand Sculpture of Hamilton’s Bridge

Figure : Hamilton’s Graffito: i2 = j2 = k2 = ijk = 1.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-10
SLIDE 10

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-11
SLIDE 11

The Swinging Spring

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-12
SLIDE 12

Two distinct oscillatory modes with distinct restoring forces:

  • Elastic or springy modes
  • Pendular or swingy modes

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-13
SLIDE 13

Two distinct oscillatory modes with distinct restoring forces:

  • Elastic or springy modes
  • Pendular or swingy modes

Take a peek at the Java Applet Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-14
SLIDE 14

In a paper in 1981, Breitenberger and Mueller made the following comment: “This simple system looks like a toy at best, but its behaviour is astonishingly complex, with many facets of more than academic lustre.” I hope to convince you of the validity of this remark.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-15
SLIDE 15

The Exact Equations of Motion

In Cartesian coordinates the Lagrangian is

L =T V = 1

2m

⇣ ˙ x2 + ˙ y2 + ˙ Z 2⌘ | {z }

K.E

1

2k(r `0)2

| {z }

E.P.E

mgZ | {z }

G.P.E

The equations of motion are (with !2

Z ⌘ k/m):

¨ x = !2

Z

✓r `0 r ◆ x ¨ y = !2

Z

✓r `0 r ◆ y ¨ Z = !2

Z

✓r `0 r ◆ Z g Two constants, energy and angular momentum: E = T + V h = x ˙ y y ˙ x . The system is not integrable (two invariants, three D.O.F.).

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-16
SLIDE 16

The Canonical Equations

We consider the case of planar motion. The canonical equations of motion (in polar coordinates) are: ˙ ✓ = pθ/mr 2 ˙ pθ = mgr sin ✓ ˙ r = pr/m ˙ pr = p2

θ/mr 3 k(r `0) + mg cos ✓

These equations may also be written symbolically as ˙ X + LX + N(X) = 0 State vector X is in 4-dimensional phase space: X = (✓, pθ, r, pr)T .

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-17
SLIDE 17

Linear Normal Modes

Suppose that amplitude of motion is small: d dt B B @ ✓ pθ r 0 pr 1 C C A = B B @ 1/m`2 mg` 1/m k 1 C C A B B @ ✓ pθ r 0 pr 1 C C A The matrix is block-diagonal: X = ✓Y Z ◆ : Y = ✓ ✓ pθ ◆ , Z = ✓ r 0 pr . ◆ Linear dynamics evolve independently: ˙ Y = ✓ 1/m`2 mg` ◆ Y , ˙ Z = ✓ 0 1/m k ◆ Z . SLOW FAST

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-18
SLIDE 18

Perturbation Theory

Ratio of rotational and elastic frequencies:

✏ ⌘ ✓!R !Z ◆ = r mg k` .

For ✏ = 0, there is no coupling between the modes. For ✏ ⌧ 1 the coupling is weak. We can apply classical Hamiltonian perturbation theory.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-19
SLIDE 19

Regular and Chaotic Motion

We wish to discuss the phenomenon of Resonance for the spring, and its Pulsation and Precession. Resonance occurs for

✏ ⇡ 1 2 .

This is far from the quasi-integrable case (small ✏). However, for small amplitudes, the motion is also quasi-integrable. We look at two numerical solutions,

  • ne with small amplitude, one with large.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-20
SLIDE 20

Horizontal plan: Low energy case

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-21
SLIDE 21

Horizontal plan: High energy case

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-22
SLIDE 22

The Resonant Case

The Lagrangian (to cubic order) is

L = 1

2

˙ x2 + ˙ y2 + ˙ z2 1

2

  • !2

R(x2 + y2) + !2 Zz2

+ 1

2(x2 + y2)z ,

We study the resonant case:

!Z = 2!R .

The equations of motion are ¨ x + !2

Rx

= xz ¨ y + !2

Ry

= yz ¨ x + !2

Zx

=

1 2(x2 + y 2) .

The system is not integrable.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-23
SLIDE 23

Averaged Lagrangian technique

We seek a solution of the form: x = <[a(t) exp(i!Rt)] , y = <[b(t) exp(i!Rt)] , z = <[c(t) exp(2i!Rt)] The coefficients a(t), b(t) and c(t) vary slowly. The Lagrangian is averaged over fast time: hLi = ⇣!R 2 ⌘h =(a˙ a⇤ + b ˙ b⇤ + 2c ˙ c⇤) +  <(a2 + b2)c⇤) i where  = /(4!R) (we absorb  in t).

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-24
SLIDE 24

The Euler-Lagrange Equations

We derive the Euler-Lagrange equations resulting from this averaged Lagrangian: i ˙ a = a⇤c , i ˙ b = b⇤c , i ˙ c =

1 4(a2 + b2)

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-25
SLIDE 25

The Euler-Lagrange Equations

We derive the Euler-Lagrange equations resulting from this averaged Lagrangian: i ˙ a = a⇤c , i ˙ b = b⇤c , i ˙ c =

1 4(a2 + b2)

We transform to new dependent variables: A = 1

2(a + ib) ,

B = 1

2(a ib) ,

C = c .

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-26
SLIDE 26

The Three-wave Equations

The equations for the transformed amplitudes are:

i ˙ A = B⇤C i ˙ B = CA⇤ i ˙ C = AB

These are the three-wave equations.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-27
SLIDE 27

Invariants

The three-wave equations conserve H =

1 2(ABC⇤ + A⇤B⇤C)

N = |A|2 + |B|2 + 2|C|2 J = |A|2 |B|2 . The three-wave equations are completely integrable.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-28
SLIDE 28

Manley-Rowe Relations

Physically significant combinations of N and J: N+ ⌘ 1

2(N + J) = |A|2 + |C|2 ,

N ⌘ 1

2(N J) = |B|2 + |C|2 .

These are the Manley-Rowe relations. The quantities H, N+ and N provide three independent constants of the motion. Constant N+ and constant N correspond to

  • rthogonal circular cylinders in phase-space.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-29
SLIDE 29

Surfaces of Revolution

Motion is on the intersection with plane of constant X.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-30
SLIDE 30

Darryl’s Books on Geometric Mechanics

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-31
SLIDE 31

Ubiquity of the Three-Wave Equations

I Modulation equations for wave interactions in

fluids and plasmas.

I Three-wave equations govern envelop dynamics

  • f light waves in an inhomogeneous material; and

phonons in solids.

I Maxwell-Schrödinger envelop equations for

radiation in a two-level resonant medium in a microwave cavity.

I Euler’s equations for a freely rotating rigid body

(when H = 0).

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-32
SLIDE 32

Analytical Solution of the 3WE

We can derive complete analytical expressions for the amplitudes and phases. The amplitudes are expressed as elliptic functions. The phases are expressed as elliptic integrals. The complete details are given in: Lynch, Peter, and Conor Houghton, 2004: Pulsation and Precession of the Resonant Swinging Spring. Physica D, 190,1-2, 38-62

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-33
SLIDE 33

Original Reference

First comprehensive analysis of elastic pendulum: “Oscillations of an Elastic Pendulum as an Example of the Oscillations of Two Parametrically Coupled Linear Systems” Vitt and Gorelik (1933). Inspired by analogy with Fermi resonance of CO2. Translation of this paper available as Historical Note #3 (1999), Met Éireann, Dublin.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-34
SLIDE 34

Vibrations of CO2 Molecule

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-35
SLIDE 35

1388 667 ⇡ 2

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-36
SLIDE 36

Monodromy in Quantum Systems

It is 80 years since the work of Vitt and Gorelik. “ Remarkably, the swinging spring still has something interesting to offer to the quantum study of the Fermi resonance.” The CO2 molecule as a quantum realization of the 1:1:2 resonant swing–spring with monodromy

Richard Cushman, Holger Dullin, Andrea Giacobbe, Darryl Holm, Marc Joyeux, Peter Lynch, Dmitrií Sadovskií, and Boris Zhilinskií Published in Phys. Rev. Lett. (2004)

“It is now tempting to think of experimental quantum dynamical manifestations of monodromy.”

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-37
SLIDE 37

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-38
SLIDE 38

Springs and Triads

In a Nutshell

A mathematical equivalence with

The Swinging Spring

sheds light on the dynamics of

Resonant Rossby Waves

in the atmosphere.

slide-39
SLIDE 39

Potential Vorticity Conservation

⇣ = Relative Vorticity, f = Planetary Vorticity, h = Fluid Depth. From the Shallow Water Equations, we derive the principle

  • f conservation of potential vorticity:

d dt ✓⇣ + f h ◆ = 0 . Under the assumptions of quasi-geostrophic theory, the dynamics reduce to an equation for alone: @ @t [r2 F ] + ⇢@ @x @r2 @y @ @y @r2 @x

  • + @

@x = 0 This is the barotropic QG potential vorticity equation (BQGPVE) aka the Charney-Hasegawa-Mima Equation.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-40
SLIDE 40

Rossby Waves

Wave-like solutions of the vorticity equation:

= A cos(kx + `y t)

satisfies the equation provided

= k k2 + `2 + F .

This is the celebrated Rossby wave formula Nonlinear term vanishes for single Rossby wave: A pure Rossby wave is solution of nonlinear equation. When there is more than one wave present, this is no longer true: the components interact with each other through the nonlinear terms.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-41
SLIDE 41

Resonant Rossby Wave Triads

Case of special interest: Two wave components produce a third such that its interaction with each generates the other. By a multiple time-scale analysis we derive the modulation equations for the wave amplitudes:

i ˙ A = B⇤C , i ˙ B = CA⇤ , i ˙ C = AB ,

[Canonical form of the three-wave equations].

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-42
SLIDE 42

The Spring Equations

and the

Triad Equations are

are

Mathematically Identical!

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-43
SLIDE 43

Numerical Example of Resonance

Method of numerical solution of the PDE: @ @t [r2 F ] + ⇢@ @x @r2 @y @ @y @r2 @x

  • + @

@x = 0

  • Potential vorticity, q = [r2 F ] is

stepped forward (with leap-frog method)

  • is obtained by solving a Helmholtz equation

with periodic boundary conditions

  • The Jacobian term is discretized following Arakawa

(to conserve energy and enstrophy)

  • Amplitude is chosen very small.

Therefore, interaction time is very long.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-44
SLIDE 44

Components of a resonant Rossby wave triad All fields are scaled to have unit amplitude.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-45
SLIDE 45

Variation with time of the amplitudes of three components of the stream function.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-46
SLIDE 46

Stream function at three times during an integration of duration T = 4800 days.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-47
SLIDE 47

Precession of Triads

  • Analogies: Interesting — Equivalences: Useful!

Since the same equations apply to both the spring and triad systems, the stepwise precession of the spring must have a counterpart for triad interactions.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-48
SLIDE 48

Precession of Triads

  • Analogies: Interesting — Equivalences: Useful!

Since the same equations apply to both the spring and triad systems, the stepwise precession of the spring must have a counterpart for triad interactions. In terms of the variables of the three-wave equations, the semi-axis major and azimuthal angle ✓ are Amaj = |A1| + |A2| , ✓ = 1 2('1 '2) . Initial conditions chosen as for the spring (by means of the transformation relations). Initial field scaled to ensure that small amplitude approximation is accurate.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-49
SLIDE 49

Polar plot of Amaj versus ✓ for resonant triad.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-50
SLIDE 50

Horizontal projection of spring solution, y vs. x.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-51
SLIDE 51

Polar plots of Amaj versus ✓.

(These are the quantities for the Triad, which correspond to the horizontal projection of the swinging spring.)

  • The Star-like pattern is immediately evident.
  • Precession angle again about 30.

This is remarkable, and illustrates the value of the equivalence: Phase precession for Rossby wave triads had not been noted before. Resonant interactions are important for energy distribution in the atmosphere. They play a central rôle in Wave Turbulence Theory.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-52
SLIDE 52

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-53
SLIDE 53

Candle-holders from Copenhagen

Fireballs (designer: Pernille Vea)

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-54
SLIDE 54

The RnR: a Topless Bowling-ball

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-55
SLIDE 55

Recession

See animated gif of RnR on website.

slide-56
SLIDE 56

Globular Cluster: Messier 54, NGC 6715 Class III Extragalactic Globular Cluster.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-57
SLIDE 57

Box and Loop Orbits: Globular Cluster

Two orbits in a logarithmic gravitational potential. Left: a box orbit. Right: a loop orbit.

Galactic Dynamics. Binney and Tremaine (2008) [pg. 174] Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-58
SLIDE 58

Box and Loop Orbits: Rock’n’roller

Trajectory of the Rock’n’roller in ✓–-plane (✓ radial, azimuthal) with ✏ = 0.1.

slide-59
SLIDE 59

Box and Loop Orbits: Perturbed SHO

Box and Loop orbits for the perturbed SHO.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-60
SLIDE 60

Box and Loop Orbits: Billiards

Box and Loop orbits on a billiard table.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-61
SLIDE 61

The RnR: Main Topics

I Two types of trajectories: boxes and loops. I Simple model: Perturbed 2D harmonic oscillator. I Small-amplitude motion of rock’n’roller. I Equations of motion in quaternionic form. I Recession is associated with box orbits.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-62
SLIDE 62

Motivation

One of the motivations for studying the Rock’n’roller is the hope of finding an invariant of the motion in addition to the energy. This expectation arises from the symmetry of the body. For the general Chaplygin Sphere, there is a finite angle δ between the principal axis corresponding to I3 and the line joining the centres of gravity and symmetry. For the Rock’n’roller, this angle is zero and the Lagrangian is independent of the azimuthal angle φ. However, we have not found a second invariant and, considering the non-holonomic nature of the problem, its existence remains an open question.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-63
SLIDE 63

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-64
SLIDE 64

The Perturbed Harmonic Oscillator

Unperturbed system: 2D SHO with equal frequencies: L0 = 1

2( ˙

x2 + ˙ y 2) 1

2!2 0(x2 + y 2)

The perturbed system has Lagrangian: L = L0 y 2 ✏r 4 , where ⌧ !2

0 and ✏ ⌧ 1.

The -term breaks the 1 : 1 resonance. The ✏-term is a radially symmetric stiffening.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-65
SLIDE 65

To analyse the system, we assume a solution x(t) = <{A(t) exp(i!0t)} y(t) = <{B(t) exp(i!0t)} and average the Lagrangian over the fast motion. We let A = |A| exp(i↵) and B = |B| exp(i). Defining W = |A|2 |B|2 and = ↵ , we have dW d⌧ = (1 W 2) sin cos d d⌧ = W sin2 1 where = 2✏U/ is a non-dimensional parameter.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-66
SLIDE 66

Again, dW d⌧ = (1 W 2) sin cos d d⌧ = W sin2 1 These are the canonical equations for the Hamiltonian H = 1

2(1 W 2) sin2 + W .

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-67
SLIDE 67

Phase portraits (W– plane) for the perturbed SHO. Left panel: = 0.5. Right panel: = 2.0.

slide-68
SLIDE 68

Box and Loop Orbits: Perturbed SHO

Box and Loop orbits for the perturbed SHO.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-69
SLIDE 69

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-70
SLIDE 70

Sergey Alexeyevich Chaplygin

slide-71
SLIDE 71

Sergey Alexeyevich Chaplygin

Sergey Alexeyevich Chaplygin (1869–1942) was a Russian physicist, mathematician, and mechanical engineer. He is known for mathematical formulas such as Chaplygin’s equation. He graduated in 1890 from Moscow University, and later became a professor. He taught mechanical engineering at Moscow’s Woman College in 1901, and applied mathematics at Moscow School of Technology, 1903. Chaplygin was elected to the Russian Academy of Sciences in 1924. The lunar crater Chaplygin and town Chaplygin are named in his honor. His "Collected Works" in four volumes were published in 1948.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-72
SLIDE 72

The Hierarchy of Spheres

slide-73
SLIDE 73

Schematic Diagram of Chaplygin Sphere

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-74
SLIDE 74

RnR: The Physical System

Consider a spherical rigid body with an asymmetric mass distribution. Specifically, we consider a loaded sphere. The dynamics are essentially the same as for the tippe-top, which has been studied extensively. Unit radius and unit mass. Centre of mass off-set a distance a from the centre. Moments of inertia I1, I2 and I3, with I1 ⇡ I2 < I3.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-75
SLIDE 75

The Lagrangian

The Lagrangian of the system is easily written down: L = 1

2(I1!2 1 + I2!2 2 + I3!2 3) + 1 2( ˙

X 2 + ˙ Y 2 + ˙ Z 2) ga(1 cos ✓) The equations may then be written (in vector form): Σ ˙ θ = ω , K ˙ ω = Pω where the matrices Σ and K are known and Pω = @ (g + !2

1 + !2 2)as + (I2 I3 af)!2!3

(g + !2

1 + !2 2)as + (I3 I1 + af)!1!3

(I1 I2)!1!2 + as(!1 + !2)!3 1 A Note that neither K nor Pω depends explicitly on .

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-76
SLIDE 76

Box and Loop Orbits: Rock’n’roller

Trajectory of the Rock’n’roller in ✓–-plane (✓ radial, azimuthal) with ✏ = 0.1.

slide-77
SLIDE 77

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-78
SLIDE 78

The Routh Sphere: I1 = I2

slide-79
SLIDE 79

Constants of Motion for Routh Sphere

In case I1 = I2, there are three degrees of freedom and three constants of integration. The kinetic energy is K = 1

2[u2 + v 2 + w2] + 1 2[I1(!2 1 + !2 2) + I3!2 3]

The potential energy is V = mga(1 cos ✓) . Since there is no dissipation, E = K + V = constant .

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-80
SLIDE 80

Constants of Motion for Routh Sphere

Jellett’s constant is the scalar product: CJ = L · r = I1s(!1 + !2) + I3f !3 = constant . where s = sin ✓, f = cos ✓ a, = sin and = cos .

[S O’Brien & J L Synge first gave this interpretation.]

Routh’s constant (difficult to interpret physically): CR = q I3 + s2 + (I3/I1)f 2

  • !3 = constant .

Constant CR implies conservation of sign of !3 . . . . . . but this does not automatically preclude recession!

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-81
SLIDE 81

Edward J Routh John H Jellett

1831–1907 1817–1888

slide-82
SLIDE 82

Edward J Routh

Edward John Routh (20 January 1831 to 7 June 1907), an English mathematician, noted as the outstanding coach of students preparing for the Mathematical Tripos examination of the University of Cambridge. He also did much to systematize the mathematical theory of mechanics and created several ideas critical to the development of modern control systems theory. In 1854, Routh graduated just above James Clerk Maxwell, as Senior Wrangler, sharing the Smith’s prize with him. He coached over 600 pupils between 1855 and 1888, 27 of them making Senior Wrangler. Known for: Routh-Hurwitz theorem, Routh stability criterion, Routh array, Routhian, Routh’s theorem, Routh’s algorithm, Kirchhoff-Routh function.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-83
SLIDE 83

John H Jellett

  • J. H. Jellett was a native of Cashel, County Tipperary, the son of a clergyman. He

graduated from Trinity College with honors in mathematics in 1838, and was elected to Fellowship in 1840. In 1847 he was appointed to the newly established chair of Natural Philosophy (Applied Mathematics), which he held until 1870. Jellett was a scholar of considerable eminence and his publications cover the fields of pure and applied mathematics, notably the theory of friction and the properties of optically active solutions, as well as sermons and lectures on religious topics. He was President of the Royal Irish Academy for five years from 1869, received the Royal Society’s Medal in 1881 and an honorary degree from Oxford in 1887. His politics were sufficiently liberal to make him an acceptable candidate to Gladstone who appointed him Provost of Trinity College Dublin in April 1881. He died in office on 19 February 1888.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-84
SLIDE 84

Integrability of Routh Sphere

Using Routh’s constant CR, we have !3 = !3(✓). Then, using Jellett’s constant CJ, we have !2 = !2(✓). Using the energy equation, we can now write:

˙ ✓2 = f(✓) .

For a given ✓, both !2 and !3 are fixed: This confirms that recession is impossible.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-85
SLIDE 85

Invariants of the Rock’n’roller

The only known constant of motion is total energy E. There remains a symmetry: the system is unchanged under the transformation ! + The spirit of Noether’s Theorem would indicate another constant associated with this symmetry; So far, we have not found a “missing constant”.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-86
SLIDE 86

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-87
SLIDE 87

Quaternionic Formulation

The Euler angles have a singularity when ✓ = 0 The angles and are not uniquely defined there. We can obviate this problem by using Euler’s symmetric parameters: = cos 1

2✓ cos 1 2( + )

⇠ = sin 1

2✓ cos 1 2( )

⇣ = cos 1

2✓ sin 1 2( + )

⌘ = sin 1

2✓ sin 1 2( )

These are the components of a unit quaternion q = + ⇠i + ⌘j + ⇣k 2 + ⇠2 + ⌘2 + ⇣2 = 1

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-88
SLIDE 88

William Rowan Hamilton (1805–1865)

slide-89
SLIDE 89

Quaternion Equations

Euler’s symmetric parameters,

  • r

Euler-Rodrigues parameters: = cos 1

2✓ cos 1 2( + )

⇠ = sin 1

2✓ cos 1 2( )

⇣ = cos 1

2✓ sin 1 2( + )

⌘ = sin 1

2✓ sin 1 2( )

The components of angular velocity are !1 = 2[ ˙ ⇠ ⇠ ˙ + ⇣ ˙ ⌘ ⌘ ˙ ⇣] !2 = 2[ ˙ ⌘ ⌘ ˙ + ⇠ ˙ ⇣ ⇣ ˙ ⇠] !3 = 2[ ˙ ⇣ ⇣ ˙ + ⌘ ˙ ⇠ ⇠ ˙ ⌘]

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-90
SLIDE 90

Lagrangian and Hamiltonian

The quaternion equations arise from the Lagrangian L = 1

2(k1 ˙

µ2 + k2 ˙ ⌫2) 1

2(k1 ˜

Ω2

1µ2 + k2 ˜

Ω2

2⌫2) + k1k2(µ ˙

⌫ ⌫ ˙ µ) where (, ⇣, ⇠, ⌘) ! (, ⇣, µ, ⌫). The generalized momenta are pµ = k1( ˙ µ k2⌫) and pν = k2( ˙ ⌫ + k2µ) The Hamiltonian is H = 1

2

p2

µ

k1 + p2

ν

k2 !

  • [k1µpν k2⌫pµ]

+

1 2[k1(k1k2 + ˜

Ω2

1)µ2 + k2(k1k2 + ˜

Ω2

2)⌫2]

slide-91
SLIDE 91

Constants of the Motion

The numerical value of the Hamiltonian (energy) is Eµ+ν = 1

2(k1 ˙

µ2 + k2 ˙ ⌫2) + 1

2(k1 ˜

Ω2

1µ2 + k2 ˜

Ω2

2⌫2)

An additional constant of the motion can be found: K1 ⌘ ⇣

λ2 ˙ µ+β2ν β1λ2β2λ1

⌘2 + ⇣

˙ νβ2λ2µ β1λ1β2λ2

⌘2 = µ2

1 ,

K2 ⌘ ⇣

λ1 ˙ µ+β1ν β1λ2β2λ1

⌘2 + ⇣

˙ νβ1λ1µ β1λ1β2λ2

⌘2 = µ2

2 .

Numerical tests confirm that K1 and K2 are constant.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-92
SLIDE 92

Aspiration

To find an invariant of the motion of the Rock’n’roller in addition to the energy. This expectation arises from the symmetry of the body. In view of the non-holonomic nature of the problem, its existence remains an open question. However, the box and loop orbits suggest that a search would be worthwhile.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-93
SLIDE 93

Aspiration

To find an invariant of the motion of the Rock’n’roller in addition to the energy. This expectation arises from the symmetry of the body. In view of the non-holonomic nature of the problem, its existence remains an open question. However, the box and loop orbits suggest that a search would be worthwhile. In elliptical billiards there is an “extra” invariant: p1 ⇥ p2.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-94
SLIDE 94

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-95
SLIDE 95

Kalejdoskop Matematyczny (1939)

slide-96
SLIDE 96

Billiard Shot leading to a Loop Orbit

slide-97
SLIDE 97

Billiard Shot leading to a Box Orbit

slide-98
SLIDE 98

Box and Loop Orbits: Billiards

Box and Loop orbits on an elliptical billiard table.

Extra invariant: p1 ⇥ p2 is conserved.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-99
SLIDE 99

Billiards and Ballyards

Main idea: Billiard Table with Soft Cushions = ) Ballyard. Playing surface no longer quite flat.

Figure : Potentials 2xN for N 2 {2, 4, 8, 16}.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-100
SLIDE 100

Back to Basics: 1 Dimension

A particle in a parabolic well z = 1

2z1x2 has Lagrangian

L = 1

2 ˙

x2(1 + z2

1x2) 1 2(gz1)x2

The Euler-Lagrange equations are (1 + z2

1x2)¨

x + gz1x + z2

1x ˙

x2 = 0

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-101
SLIDE 101

Back to Basics: 1 Dimension

Figure : Potentials x2 and x2/5.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-102
SLIDE 102

Back to Basics: 1 Dimension

Figure : Potentials x2 and x10.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-103
SLIDE 103

Back to Basics: 1 Dimension

To linearize, make z1 small but keep gz1 fixed: ¨ x + (gz1)x = 0 . This is a geometric/gravimetric approximation. We are flattening the table while turning up gravity.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-104
SLIDE 104

Moving to 2 Dimensions: Circular Table

A particle in a paraboloidal well with axial symmetry, z = 1

2z1r 2 has Lagrangian

L = 1

2[(1 + z2 1r 2)˙

r 2 + r 2 ˙ ✓2] 1

2(gz1)r 2

As before, we let z1 ! 0 with 1

2gz1 = 1.

Since ✓ is a cyclic variable, @L/@ ˙ ✓ = r 2 ˙ ✓ is constant. My Gaffe: Eliminate ˙ ✓ from L using h = r 2 ˙ ✓. Correct: Get E-L equation, then use h = r 2 ˙ ✓.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-105
SLIDE 105

Circular Ballyards

Since the restoring force is central, the angular momentum is conserved. The system is integrable.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-106
SLIDE 106

Potentials with Increasing Power

Figure : Potentials 2x2, 2x4, 2x8, and 2x16.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-107
SLIDE 107

Changing the Potential

Figure : N=2.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-108
SLIDE 108

Changing the Potential

Figure : N=4.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-109
SLIDE 109

Changing the Potential

Figure : N=8.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-110
SLIDE 110

Changing the Potential

Figure : N=16.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-111
SLIDE 111

Changing the Potential

Figure : N=32.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-112
SLIDE 112

Changing the Potential

Figure : N=64.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-113
SLIDE 113

Elliptic Ballyards: N = 4

Left: Angular Momentum varies from 6 to +6. Right: Angular Momentum varies from 4 to 10. Additional Constant of Motion not found (yet!)

slide-114
SLIDE 114

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-115
SLIDE 115

Squircular Ballyard Potential

x4 + y 4 = 1

Figure : Potential, V(x), N = 8. Left: VER-X. Right: HOR-X.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-116
SLIDE 116

Squircular Ballyard

Figure : N = 2. ICs = 1.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-117
SLIDE 117

Squircular Ballyard

Figure : N = 2. ICs = 2.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-118
SLIDE 118

Squircular Ballyard

Figure : N = 2. ICs = 3.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-119
SLIDE 119

Squovular Ballyard

⇣x a ⌘8 + ⇣y b ⌘8 = 1

Figure : Aspect ratio 2 : 1.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-120
SLIDE 120

Outline

Introduction Swinging Spring Potential Vorticity Rock’n’Roller Perturbed SHO Sergey Chaplygin Routh Sphere: I1 = I2 Quaternion Formulation Billiards & Ballyards Squircles & Squovals Conclusion

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-121
SLIDE 121

Conclusion

AIM: To find an invariant of the motion of the Rock’n’roller in addition to the energy. This expectation arises from the symmetry of the body. The box and loop orbits suggest that a search would be worthwhile. The investigation of elliptical billiards may be fruitful. But the goal is not yet reached

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-122
SLIDE 122

Conclusion

“I Still Haven’t Found What I’m Looking For”

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-123
SLIDE 123

Conclusion

“I Still Haven’t Found What I’m Looking For” Let’s hope for success by Darryl’s 75th Birthday.

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-124
SLIDE 124

Thank You All

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin

slide-125
SLIDE 125

Thank You All and Happy Birthday, Darryl!

Intro SS PV RnR SHO Chaplygin Routh Quaternions Billiards Squ/Squ Fin