Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as...
Transcript of Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as...
![Page 1: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/1.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Continuous symmetry reduction forhigh-dimensional flows
Predrag Cvitanovic1 and Evangelos Siminos1,2
1Georgia Institute of Technology2CEA/DAM/DIF, Paris
April 2, 2010
![Page 2: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/2.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Outline
1 Navier-Stokesfluid measurementsbaby Navier-Stokes
2 Kuramoto-Sivashinsky, L = 22, state spacetypes of solutionsPDE’s as dynamical systems
3 Dynamical systems approach to spatially extended systemsLorenz equations examplecomplex Lorenz flow example
4 relativity for cyclistsLie groups, algebras
5 symmetry reductionHilbert polynomial basismethod of slicesslice & dice
6 conclusions - to be done
![Page 3: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/3.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
modern times
amazing data! amazing numerics!
3D turbulent pipe flow
solutions are
rotationally equivariant
translationally equivariant
![Page 4: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/4.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
KSe
Kuramoto-Sivashinsky equation
1-dimensional “Navier-Stokes”
ut + u�u = −�2u − �4u , x ∈ [−L/2,L/2] ,
describes extended systems such as
reaction-diffusion systems
flame fronts in combustion
drift waves in plasmas
thin falling films, . . .
![Page 5: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/5.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Kuramoto-Sivashinsky on a large domain
0 1 2 3 4 5 6 7 8 9 100
20
40
60
80
100
t
x/(2π√
2)
−3
−2
−1
0
1
2
3
turbulent behavior
simpler physical, mathematical and computational settingthan Navier-Stokes
![Page 6: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/6.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
evolution of Kuramoto-Sivashinsky on small L = 22 cell
−10 0 10200
250
300
350
400
x−10 0 10
400
450
500
550
600
x−10 0 10
600
650
700
750
800
x−10 0 10
800
850
900
950
1000
x−10 0 100
50
100
150
200
x
t
horizontal: x ∈ [−11,11]vertical: timecolor: magnitude of u(x , t)
![Page 7: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/7.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
equilibria
E1 E2 E3
10 0 10x
1.5
0.
1.5
u
10 0 10x
1.5
0.
1.5
u
10 0 10x
1.5
0.
1.5
u
E3 invariant under τ1/3.
For any Ei we have a continuous family of equilibria underrotations τ�/L Ei .
![Page 8: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/8.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
symmetries of Kuramoto-Sivashinsky equation
with periodic boundary condition
u(x , t) = u(x + L, t)
the symmetry group is O(2):
translations: τ�/L u(x , t) = u(x + �, t) , � ∈ [−L/2,L/2] ,
reflections: κu(x) = −u(−x) .
translational symmetry → traveling wave solutions
![Page 9: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/9.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
symmetries of Kuramoto-Sivashinsky equation
with periodic boundary condition
u(x , t) = u(x + L, t)
the symmetry group is O(2):
translations: τ�/L u(x , t) = u(x + �, t) , � ∈ [−L/2,L/2] ,
reflections: κu(x) = −u(−x) .
translational symmetry → traveling wave solutions Traveling(or relative) unstable coherent solutions are ubiquitous inturbulent hydrodynamic flows
![Page 10: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/10.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
traveling waves
−10 −5 0 5 10−2
−1
0
1
2
3TW+1
x
u
−10 −5 0 5 10−2
−1
0
1
2
3TW+2
x
u
invariant (as a set) underrotations: relative equilibria.
They live in full space.
![Page 11: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/11.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
traveling waves
−10 −5 0 5 10−2
−1
0
1
2
3TW+1
x
u
−10 −5 0 5 10−2
−1
0
1
2
3TW+2
x
u
invariant (as a set) underrotations: relative equilibria.
They live in full space.
Toshiba Corp and Microsoft Corpchairman Bill Gates are to worktogether to develop a nextgeneration “traveling-wavereactor”, which could operate forup to 100 years without refueling.[news item - Tokyo, March 23, 2010]
![Page 12: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/12.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
unstable relative periodic orbits
Tp = 16.3, Tp = 33.5, Tp = 47.6, Tp = 10.3 Tp = 33.4�p = 2.86 �p = 4.04 �p = 5.68
−10 0 10
20
40
60
80
100
x−10 0 10
20
40
60
80
100
x−10 0 10
20
40
60
80
100
x−10 0 10
20
40
60
80
100
x−10 0 10
20
40
60
80
100
x
have computed 40,000 unstable periodic and relative periodic orbits.
how are they organized?
![Page 13: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/13.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
types of solutions
symmetries of Kuramoto-Sivashinsky equation
translational symmetry ⇒traveling wave solutions
unstable relative periodic orbits
question
what are the invariant objects that organize phase space in aspatially extended system with translational symmetry and howdo they fit together to form a skeleton of the dynamics?
![Page 14: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/14.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
PDE’s as dynamical systems
state space
the space in which all possible states u’s live
∞-dimensional:point u(x) is a function of x on interval x ∈ L.
in practice:a high but finite dimensional space (e.g. through a spectraldiscretization)
![Page 15: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/15.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
PDE’s as dynamical systems
state space of Kuramoto-Sivashinsky on L = 22
intrinsic dimensionality
dynamics are often captured by fewer variables thanneeded to numerically resolve the PDE.
Lyapunov exponents:(λi) = (0.048, 0, 0, −0.003, −0.189, −0.256, −0.290,−0.310, · · · )‘8-dimensional’ covariant Lyapunov frame? perhapstractable?
how do we exploit such low dimensionality to obtaindynamical systems description?
![Page 16: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/16.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
low dimensional systems:equilibria, periodic orbits organize the long time dynamics.
is this true in extended systems?
![Page 17: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/17.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
from Lorenz 3D attractor to a unimodal map
Lorenz equations
⎡⎣
xyz
⎤⎦ =
⎡⎣
σ(y − x)ρx − y − xz
xy − bz
⎤⎦
withσ = 10,b = 8/3, ρ = 28.
Lorenz attractor
E2E1
E0
x
y
z
![Page 18: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/18.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
from Lorenz 3D attractor to a unimodal map
Equilibria
x = v(x) = 0
Linear stability ofequilibria
Aij =∂vi
∂xj(xEm)
Lorenz attractor
E2E1
E0
x
y
z
![Page 19: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/19.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
from Lorenz 3D attractor to a unimodal map
Linear stability ofequilibria
Aij =∂vi
∂xj(xEm)
Eigenvalues of A:λj = μj ± iνj
Linearly stable ifμj < 0
Linearly unstable ifμj > 0
Lorenz attractor
E2E1
E0
x
y
z
![Page 20: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/20.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
from Lorenz 3D attractor to a unimodal map
E0 :λ1 = 11.83λ2 = −2.66λ3 = −22.83
xy
z
e�1�
e�2�
e�3�
� 1
� 0.5
E010�13
10�12
10�11
10�1010�9
10�8
Lorenz attractor
E2E1
E0
x
y
z
![Page 21: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/21.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
from Lorenz 3D attractor to a unimodal map
E1
λ1,2 = 0.094 ± 10.19λ3 = −13.85
Re e�1�
Im e�1�
e�3�
E1
Lorenz attractor
E2E1
E0
x
y
z
![Page 22: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/22.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
from Lorenz 3D attractor to a unimodal map
Poincaré sectionP : (N-1)-dimensionalhypersurface.
Poincaré return map
0 5 10 15 20 25
5
10
15
20
25
Sn
Sn+
1
WS(E0)
Lorenz attractor
E2
E1
E0
y
z
��
�
![Page 23: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/23.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
Take the hint from low dimensional systems
low dimensional systems:equilibria, periodic orbits organize the long time dynamics.
is this true in extended systems?
![Page 24: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/24.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
Kuramoto-Sivashinsky flow reduced to discrete maps
within the discrete u(x) = −u(−x) invariant subspace
�0.4 �0.3 �0.2 �0.1 0.1 0.2s�n�
�0.4
�0.3
�0.2
�0.1
0.1
0.2
s�n�1�
Christiansen et. al. (1996)
0 0.2 0.4 0.6 0.80
0.2
0.4
0.6
0.8
s
f(s)
Lan and Cvitanovic (2004)
∞− d PDE state space dynamics can be reduced to lowdimensional return maps!
![Page 25: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/25.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lorenz equations example
Kuramoto-Sivashinsky flow reduced to discrete maps
within the discrete u(x) = −u(−x) invariant subspace
�0.4 �0.3 �0.2 �0.1 0.1 0.2s�n�
�0.4
�0.3
�0.2
�0.1
0.1
0.2
s�n�1�
Christiansen et. al. (1996)
0 0.2 0.4 0.6 0.80
0.2
0.4
0.6
0.8
s
f(s)
Lan and Cvitanovic (2004)
∞− d PDE state space dynamics can be reduced to lowdimensional return maps!
BUT! must reduce continuous symmetries first
![Page 26: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/26.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
complex Lorenz flow example
from complex Lorenz flow 5D attractor → unimodal map
complex Lorenz equations
⎡⎢⎢⎢⎣
x1x2y1y2z
⎤⎥⎥⎥⎦ =
⎡⎢⎢⎢⎣
−σx1 + σy1−σx2 + σy2
(ρ1 − z)x1 − ρ2x2 − y1 − ey2ρ2x1 + (ρ1 − z)x2 + ey1 − y2
−bz + x1y1 + x2y2
⎤⎥⎥⎥⎦
ρ1 = 28, ρ2 = 0, b = 8/3, σ = 10, e = 1/10
A typical {x1, x2, z} trajectory of thecomplex Lorenz flow+ a short trajectory of whose initialpoint is close to the relativeequilibrium Q1 superimposed.
attractor
![Page 27: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/27.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
complex Lorenz flow example
from complex Lorenz flow 5D attractor → unimodal map
what to do?
the goal
reduce this messy strange attractor toa 1-dimensional return map
attractor
![Page 28: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/28.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
complex Lorenz flow example
from complex Lorenz flow 5D attractor → unimodal map
the goal attained
but it will cost you
after symmetry reduction; must learnhow to quotient the SO(2) symmetry
1D return map!
0 100 200 300 400 5000
100
200
300
400
500
sns n�
1
![Page 29: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/29.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lie groups, algebras
Lie groups elements, Lie algebra generators
An element of a compact Lie group:
g(θ) = eθ·T , θ · T =∑
θaTa, a = 1,2, · · · ,N
θ · T is a Lie algebra element, and θa are the parameters of thetransformation.
![Page 30: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/30.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Lie groups, algebras
example: SO(2) rotations for complex Lorenz equations
SO(2) rotation by finite angle θ:
g(θ) =
⎛⎜⎜⎜⎜⎝
cos θ sin θ 0 0 0− sin θ cos θ 0 0 0
0 0 cos θ sin θ 00 0 − sin θ cos θ 00 0 0 0 1
⎞⎟⎟⎟⎟⎠
![Page 31: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/31.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
symmetries of dynamics
A flow x = v(x) is G-equivariant if
v(x) = g−1 v(g x) , for all g ∈ G .
![Page 32: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/32.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
foliation by group orbits
group orbits
group orbit Mx of x is the set of allgroup actions
Mx = {g x | g ∈ G}
![Page 33: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/33.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
foliation by group orbits
group orbits
group orbit Mx(0) of state spacepoint x(0), and the group orbit Mx(t)reached by the trajectory x(t) time tlater.
![Page 34: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/34.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
foliation by group orbits
group orbits
any point on the manifold Mx(t) isequivalent to any other.
![Page 35: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/35.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
foliation by group orbits
group orbits
action of a symmetry group endowsthe state space with the structure of aunion of group orbits, each group orbitan equivalence class.
![Page 36: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/36.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
foliation by group orbits
group orbits
the goal:replace each group orbit by a uniquepoint a lower-dimensional reducedstate space (or orbit space)
![Page 37: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/37.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
a traveling wave
x1
x2
x3
τg( )τx( )= x(0)
τg( )
x(0)
g( )tτ
v = c t
v = c relative equilibrium(traveling wave, rotatingwave)xTW(τ) ∈ MTW : thedynamical flow field pointsalong the group tangentfield, with constant ‘angular’velocity c, and the trajectorystays on the group orbit
![Page 38: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/38.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
a traveling wave
x1
x2
x3
τg( )τx( )= x(0)
τg( )
x(0)
g( )tτ
v = c t
v = c relative equilibrium
v(x) = c·t(x) , x ∈ MTW
x(τ) = g(−τ c) x(0) = e−τ c·Tx(
![Page 39: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/39.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
a traveling wave
x1
x2
x3
τg( )τx( )= x(0)
τg( )
x(0)
g( )tτ
v = c t
v = c
group orbit g(τ) x(0)coincides with thedynamical orbit x(τ) ∈ MTW
and is thus flow invariant
![Page 40: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/40.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
a relative periodic orbit
x(T )p
θp
x1
x2x(0)
tv
v t
x3
relative periodic orbit
xp(0) = gpxp(Tp)
exactly recurs at a fixedrelative period Tp, butshifted by a fixed groupaction gp
![Page 41: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/41.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
a relative periodic orbit
x(T )p
θp
x1
x2x(0)
tv
v t
x3
relative periodic orbit startsout at x(0) , returns to thegroup orbit of x(0) after timeTp, a rotation of the initialpoint by gp
![Page 42: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/42.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
a relative periodic orbit
x(T )p
θp
x1
x2x(0)
tv
v t
x3
The group actionparametersθ = (θ1, θ2, · · · θN)are irrational: trajectorysweeps out ergodically thegroup orbit without everclosing into a periodic orbit.
![Page 43: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/43.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
relativity for pedestrians
try a co-moving coordinate frame?
(a)
v1v2
v3
A relative periodic orbit of the Kuramoto-Sivashinsky flow,traced for four periods Tp, projected on(a) a stationary state space coordinate frame {v1, v2, v3};
![Page 44: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/44.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
relativity for pedestrians
try a co-moving coordinate frame?
(b)
v�1v�2
v�3
A relative periodic orbit of the Kuramoto-Sivashinsky flow,traced for four periods Tp, projected on(b) a co-moving {v1, v2, v3} frame
![Page 45: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/45.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
in/equivariance
relativity for pedestrians
no good global co-moving frame!
(b)
v�1v�2
v�3
this is no symmetry reduction at all; all other relative periodicorbits require their own frames,moving at different velocities.
![Page 46: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/46.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
symmetry reduction
all points related by a symmetry operation are mapped tothe same point.
relative equilibria become equilibria and relative periodicorbits become periodic orbits in reduced space.
families of solutions are mapped to a single solution
![Page 47: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/47.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
reduction methods
1 Hilbert polynomial basis: rewrite equivariant dynamics ininvariant coordinates
2 moving frames, or slices: cut group orbits by ahypersurface (kind of Poincareé section), each group orbitof symmetry-equivalent points represented by the singlepoint
![Page 48: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/48.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
reduction methods
1 Hilbert polynomial basis: rewrite equivariant dynamics ininvariant coordinates: global
2 moving frames, or slices: cut group orbits by ahypersurface (kind of Poincareé section), each group orbitof symmetry-equivalent points represented by the singlepoint: local
![Page 49: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/49.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials
rewrite the equations in variables invariant under thesymmetry transformation
![Page 50: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/50.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials
rewrite the equations in variables invariant under thesymmetry transformation
or compute solutions in original space and map them toinvariant variables
![Page 51: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/51.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials basis
Hilbert basis for complex Lorenz equations
u1 = x21 + x2
2 , u2 = y21 + y2
2
u3 = x1y2 − x2y1 , u4 = x1y1 + x2y2
u5 = z
invariant under SO(2) action on a 5-dimensional state space
polynomials related through syzygies:
u1u2 − u23 − u2
4 = 0
![Page 52: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/52.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials basis
complex Lorenz equations in invariant polynomial basis
u1 = 2σ (u3 − u1)
u2 = −2 u2 − 2 u3 (u5 − ρ1)
u3 = σ u2 − (σ − 1)u3 − e u4 + u1 (ρ1 − u5)
u4 = e u3 − (σ + 1)u4
u5 = u3 − b u5
A 4-dimensional M/SO(2) reduced state space, asymmetry-invariant representation of the 5-dimensional SO(2)equivariant dynamics
![Page 53: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/53.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
state space portrait of complex Lorenz flow
drift induced by continuous symmetry
x1 x2
z
E0
Q1
01
x1 x2
z
E0
W�0�u
W�1�u
Q1
01
A generic chaotic trajectory (blue), the E0 equilibrium, arepresentative of its unstable manifold (green), the Q1 relativeequilibrium (red), its unstable manifold (brown), and one repeatof the 01 relative periodic orbit (purple).
![Page 54: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/54.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials basis
complex Lorenz equations in invariant polynomial basis
u1 = 2σ (u3 − u1)
u2 = −2 u2 − 2 u3 (u5 − ρ1)
u3 = σ u2 − (σ − 1)u3 − e u4 + u1 (ρ1 − u5)
u4 = e u3 − (σ + 1)u4
u5 = u3 − b u5
the image of the full state space relative equilibrium Q1 grouporbit is an equilibrium point, while the image of a relativeperiodic orbit, such as 01, is a periodic orbit
![Page 55: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/55.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
Hilbert invariant coordinates
projected onto invariant polynomials basis
(a) u3
u4
z
Q1
(b)0 100 200 300 400 500
0
100
200
300
400
500
sn
s n�
1
(a) The unstable manifold connection from the equilibrium E0 atthe origin to the strange attractor controlled by the rotationaround the reduced state space image of relative equilibriumQ1;
![Page 56: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/56.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
higher-dimensional invariant bases? an example
first 11 invariants for the standard action of SO(2)
u1 = r1 =√
b21 + c2
1
u3 =b2
(b2
1−c21
)+2b1c1c2
r 21
u4 =−2b1b2c1+
(b2
1−c21
)c2
r 21
u5 =b1b3
(b2
1−3c21
)−c1
(−3b2
1+c21
)c3
r 31
u6 =−3b2
1b3c1+b3c31+b3
1c3−3b1c21 c3
r 31
![Page 57: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/57.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
higher-dimensional invariant bases? an example
first 11 invariants for the standard action of SO(2)
u7 =b4
(b4
1−6b21c2
1+c41
)+4b1c1
(b2
1−c21
)c4
r 41
u8 =4b1b4c1
(−b2
1+c21
)+(
b41−6b2
1c21+c4
1
)c4
r 41
u9 =b1b5
(b4
1−10b21c2
1+5c41
)+c1
(5b4
1−10b21c2
1+c41
)c5
r 51
u10 =−b5c1
(5b4
1−10b21c2
1+c41
)+b1
(b4
1−10b21c2
1+5c41
)c5
r 51
u11 =b6
(b6
1−15b41c2
1+15b21c4
1−c61
)+2b1c1
(3b4
1−10b21c2
1+3c41
)c6
r 61
u12 =−2b1b6c1
(3b4
1−10b21c2
1+3c41
)+(
b61−15b4
1c21+15b2
1c41−c6
1
)c6
r 61
![Page 58: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/58.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials - how to find them?
invariant polynomials (Hilbert basis)
![Page 59: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/59.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials - how to find them?
invariant polynomials (Hilbert basis): computationallyprohibitive for high-dimensional flows
![Page 60: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/60.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials - how to find them?
invariant polynomials (Hilbert basis)
Cartan moving frame method / method of slices
![Page 61: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/61.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials - how to find them?
invariant polynomials (Hilbert basis)
Cartan moving frame method / method of slices
![Page 62: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/62.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
Hilbert polynomial basis
invariant polynomials - how to find them?
invariant polynomials (Hilbert basis)
Cartan moving frame method / method of slices:singularities
![Page 63: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/63.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
slice & dice
Lie algebra generators
Ta generate infinitesimal transformations: a set of N linearlyindependent [d×d ] anti-hermitian matrices, (Ta)
† = −Ta, actinglinearly on the d-dimensional state space Mexample: SO(2) rotations for complex Lorenz equations
T =
⎛⎜⎜⎜⎜⎝
0 1 0 0 0−1 0 0 0 00 0 0 1 00 0 −1 0 00 0 0 0 0
⎞⎟⎟⎟⎟⎠
The action of SO(2) on the complex Lorenz equations statespace decomposes into m = 0 G-invariant subspace (z-axis)and m = 1 subspace with multiplicity 2.
![Page 64: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/64.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
slice & dice
group tangent fields
flow field at the state space point x induced by the action of thegroup is given by the set of N tangent fields
ta(x)i = (Ta)ij xj
![Page 65: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/65.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
slice & dice
slice & dice
flow reduced to a slice
Slice M through the slice-fixing point x ′, normal to the grouptangent t ′ at x ′, intersects group orbits (dotted lines). The fullstate space trajectory x(τ) and the reduced state spacetrajectory x(τ) are equivalent up to a group rotation g(τ).
![Page 66: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/66.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
slice & dice
method of moving frames for SO(2)-equivariant flow
flow reduced to a slice
τ
τ
τ
τ
1x
θ1
θ2
2x( )
1x( )
2 2x =x
2
1x( )x(0)
x( )
slice through x ′ = (0,1,0,0,0)group tangent t ′ = (−1,0,0,0,0)Start on the slice at x(0), evolve.Compute angle θ1 to the slicerotate x(τ1) by θ1 tox(τ1) = g(θ1) x(τ1) back into the slice,x1(τ1) = 0. Repeat for points x(τi)along the trajectory.
![Page 67: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/67.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
slice & dice
slice trouble 1
portrait of complex Lorenz flow in reduced state space
(a)
x2
y2
z
W��0�u
01
Q1
(b)
x2
y2
z W��0�u
01Q1
E0
all choices of the slice fixing point x ′ exhibit flow discontinuities/ jumps
![Page 68: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/68.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
slice & dice
slice trouble 2
slice cuts an relative periodic orbitmultiple times
Relative periodic orbitintersects a hyperplaneslice in 3 closed-loopimages of the relativeperiodic orbit and 3images that appear toconnect to a closed loop.
![Page 69: Continuous symmetry reduction for high …chaosbook.org/overheads/continuous/symmReduc.pdfPDE’s as dynamical systems 3 Dynamical systems approach to spatially extended systems Lorenz](https://reader034.fdocuments.net/reader034/viewer/2022050207/5f59f11dc6b24c597a0882e1/html5/thumbnails/69.jpg)
Navier-Stokes KSe, L = 22 Dynamicist’s view of turbulence relativity for cyclists symmetry reduction Summary
summary
conclusionSymmetry reduction: efficient implementation allowsexploration of high-dimensional flows with continuoussymmetry.
stretching and folding of unstable manifolds in reducedstate space organizes the flow
to be doneconstruct Poincaré sections and return maps
find all (relative) periodic orbits up to a given period.
use the information quantitatively (periodic orbit theory).