pattern formation in rayleigh-bénard convectionphysbio/activities/lectures/cross/...b) dislocations...
TRANSCRIPT
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 1
Pattern Formation inRayleigh-Bénard Convection
Lecture 3
Chaos
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 2
Outline
• Low dimensional chaos
• Pattern Chaos
• Spatiotemporal chaos
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 2
Outline
• Low dimensional chaos
• Pattern Chaos
• Spatiotemporal chaos
Collaborators: Marc Bourzutschky, Keng-Hwee Chiam, Henry Greenside,
Roman Grigoriev, Pierre Hohenberg, Michael Louie, Dan Meiron, Mark
Paul, Yuhai Tu.
Support: NSF and DOE
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 3
Lorenz Chaos
H o t
C o l d
x
z
T (x, z, t) ' −rz
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 3
Lorenz Chaos
H o t
C o l d
x
z
T (x, z, t) ' −rz+ 9π3√
3Y (t) cos(πz) cos
(π√2x
)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 3
Lorenz Chaos
H o t
C o l d
x
z
T (x, z, t) ' −rz+ 9π3√
3Y (t) cos(πz) cos
(π√2x
)+ 27π3
4Z(t) sin(2πz)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 3
Lorenz Chaos
H o t
C o l d
x
z
T (x, z, t) ' −rz+ 9π3√
3Y (t) cos(πz) cos
(π√2x
)+ 27π3
4Z(t) sin(2πz)
ψ(x, z, t) = 2√
6X(t) cos(πz) sin
(π√2x
)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 3
Lorenz Chaos
H o t
C o l d
x
z
T (x, z, t) ' −rz+ 9π3√
3Y (t) cos(πz) cos
(π√2x
)+ 27π3
4Z(t) sin(2πz)
ψ(x, z, t) = 2√
6X(t) cos(πz) sin
(π√2x
)
(u = −∂ψ/∂z, v = ∂ψ/∂x, r = R/Rc)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 4
Lorenz Model
X = −σ(X − Y )Y = rX − Y −XZZ = b (XY − Z)
b = 8/3 andσ is the Prandtl number.
“Classic” values areσ = 10 andr = 27.
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 5
-10-5
05
10
X
-10
0
10
Y
0
10
20
30
40
Z
-50
510
X
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 6
The Butterfly Effect
The “sensitive dependence on initial conditions” found by Lorenz is often
called the “butterfly effect”.
In fact Lorenz first said (Transactions of the New York Academy of
Sciences, 1963)
One meteorologist remarked that if the theory were correct, one
flap of the sea gull’s wings would be enough to alter the course
of the weather forever.
By the time of Lorenz’s talk at the December 1972 meeting of the
American Association for the Advancement of Science in Washington,
D.C. the sea gull had evolved into the more poetic butterfly - the title of
his talk was
Predictability: Does the Flap of a Butterfly’s Wings in Brazil set
off a Tornado in Texas?
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 7
The Lorenz model does not describe Rayleigh-Bénard convection!
Thermosyphon Rayle igh-Benard Convect ion
Also: Rititake homopolar dynamo
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 8
Small system chaos
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 9
Theoretical Context
Landau (1944) Turbulence develops by infinite sequence of transitions
adding additional temporal modes and spatial complexity
Ruelle and Takens (1971)Suggested the onset of aperiodic dynamics
from a low dimensional torus (quasiperiodic motion with a small
numberN frequencies)
Feigenbaum (1978)Quantitative universality for period doubling route
to chaos
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 10
Some Experimental Highlights
Ahlers (1974) Transition from time independent flow to aperiodic flow at
R/Rc ∼ 2 (aspect ratio 5)
Gollub and Swinney (1975)Onset of aperiodic flow from time-periodic
flow in Taylor-Couette
Maurer and Libchaber, Ahlers and Behringer (1978) Transition from
quasiperiodic flow to aperiodic flow in small aspect ratio convection
Lichaber, Laroche, and Fauve (1982)Quantitative demonstration of the
Fiegenbaum period doubling route to chaos
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 11
Larger System Chaos
In “large” systems presumably the chaos becomes higher dimensional.
We might divide up the phenomena into two classes:
Pattern chaos global properties of spatial patterns are important, perhaps
induced by boundary effects
Spatiotemporal chaosstatistical aspects of local units are more
important (0→∞)
In either case we might be interested in
• the mechanism of the dynamics
• characterizing the dynamics
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 12
Pattern chaos: convection in intermediate aspect ratiocylinders
G. Ahlers, Phys. Rev. Lett.30, 1185 (1974)
G. Ahlers and R.P. Behringer, Phys. Rev. Lett.40, 712 (1978)
H. Gao and R.P. Behringer, Phys. Rev. A30, 2837 (1984)
V. Croquette, P. Le Gal, and A. Pocheau, Phys. Scr. T13, 135 (1986)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 13
• First experiments:0 = 5.27 cell, cryogenic (normal) liquidHe4 as
fluid. High precision heat flow measurements (no flow visualization).
• Onset of aperiodic time dependence in low Reynolds number flow:
relevance of chaos to “real” (continuum) systems.
• Power law decrease of power spectrumP(f ) ∼ f−4
• Aspect ratio dependence of the onset of time dependence
0 2 5 57
Rt 10Rc 2Rc 1.1Rc
• Flow visualization (Argon,0 = 7.66)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 14
(from Ahlers 1974)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 15
(from Ahlers and Behringer 1978)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 16
Gao and Behringer (1984)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 17
0
0.005
0.010
0.015
0 1 2 3Reduced Rayleigh Number ε
Noi
se A
mpl
itude
(from Ahlers and Behringer 1978)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 18
Croquette, Le Gal, and Pocheau (1986)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 19
Numerical Simulations
(Paul, MCC, Fischer and Greenside)
• 0 = 4.72,σ = 0.78, 2600. R . 7000
• Conducting sidewalls
• Random thermal perturbation initial conditions
• Simulation time∼ 100τh
– Simulation time∼ 12 hours on 32 processors
– Experiment time∼ 172 hours or∼ 1 week
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 20
0 500 1000 1500 2000time
1.4
1.5
1.6
1.7
1.8
1.9
2
Nu
R = 6949R = 4343R = 3474R = 3127R = 2804R = 2606
50 τh
Γ = 4.72σ = 0.78 (Helium)Random Initial Conditions
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 21
R = 3127(?) R = 6949
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 22
Power Spectrum
• Simulations oflow dimensional chaos(e.g. Lorenz model) show
exponential decaying power spectrum
• Power law power spectrum easily obtained fromstochasticmodels
(white-noise driven oscillator, etc.)
Deterministic Chaos ?⇒? Exponential decay
Stochastic Noise ?⇒? Power law decay
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 23
Simulation Results
Simulations reproduce experimental results....
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 24
Simulation yields a power law over the range accessible to experiment
10-2 10-1 100 101 102
ω10-12
10-11
10-10
10-9
10-8
10-7
10-6
P(ω
)
R = 6949ω-4
Γ = 4.72, σ = 0.78, R = 6949Conducting sidewalls
R/Rc = 4.0
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 25
When larger frequencies are included an exponential tail is found
10-2 10-1 100 101 102
ω10-21
10-19
10-17
10-15
10-13
10-11
10-9
10-7
10-5P
(ω)
R = 6949ω-4
Γ = 4.72, σ = 0.78, R = 6949Conducting sidewalls
R/Rc = 4.0
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 26
1 2 3 4 5 6 7ν (1/τv)
10-20
10-18
10-16
10-14
10-12
10-10
P(ν
)
R = 2804, σ = 0.78, Γ = 4.72Conducting sidewalls
Exponential region, slope = -3.8
Exponential tail not seen in experiment because of instrumental noise floor
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 27
Where does the power law come from?
Power law arises fromquasi-discontinuous changes in the slopeof N(t)
on at = 0.1− 1 time scale associated with roll pinch-off events.
This is clearest to see for the low Rayleigh number where the motion is
periodic , but again the power spectrum has a power law fall off:
• Spectrogram
• Details of time seriesN(t)
• Windowed power spectrum
• Compare periodic and chaotic events
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 28
Periodic dynamicsR = 2804
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 29
Power law spectra
10-3 10-2 10-1 100 101
ν (1/τv)10-25
10-23
10-21
10-19
10-17
10-15
10-13
10-11
10-9
10-7
P(ν
)
R = 2804ν-4
0 0.025 0.05 0.075 0.1ν (1/τv)
0
2E-07
4E-07
6E-07
8E-07
1E-06
1.2E-06
1.4E-06
P(ν
)
Γ = 4.72, R = 2804, σ = 0.78Conducting sidewallsNo overlappingLinear detrending
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 30
Spectrogram
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 31
Time seriesN(t)
600 700time
1.4
1.405
1.41
1.415
1.42
1.425N
u
a) dislocation nucleation
b) dislocations climb
c) both dislocations areat the lateral walls
one dislocation quickly glidesinto a wall foci e) the other dislocation slowly glides
into the same wall focisecondannihilation
Note: Dislocation glide alternates left and right,the portion highlighted here goes right.
d) first annihilation
R = 2804, Γ = 4.72, σ = 0.78Conducting
f) process repeats
t = 187 = 8.4 τh
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 32
Windowed power spectra
⟨P(ν)⟩
10-1 100 101
ν
10-22
10-20
10-18
10-16
10-14
10-12
10-10
10-8
P(ν
)t = 594, Nucleationt = 639, Annihilationt = 700, Glideν-4
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 33
Compare periodic and chaotic events
770 775 780 785 790time
1.395
1.4
1.405
1.41
1.415
1.42
1.425
Nu
R = 2804, σ = 0.78, Γ = 4.72Conducting sidewalls
270.5 271 271.5 272time
1.85
1.86
1.87
1.88
1.89
1.9
1.91
1.92
Nu
R = 6949, σ = 0.78, Γ = 4.72Conducting sidewalls
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 34
Mechanism of Dynamics
• Pan-Am texture with roll pinch-off events creating dislocation pairs
(Flow visualization, Pocheau, Le Gal, and Croquette 1985)
• Mean flow compresses rolls outside of stable band (Model equations,
Greenside, MCC, Coughran 1985)
• Theoretical analysis of mean flow (Pocheau and Davidaud 1997)
• Numerical simulation of importance of mean flow (Paul, MCC,
Fischer, and Greenside 2001)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 35
Generalized Swift-Hohenberg Simulations
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 36
Convection Simulations
3 convection cells with different side wall conditions: (a) rigid; (b) finned;
and (c) ramped. Case (a) is dynamic, the others static.
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 37
Spatiotemporal chaos
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 38
Spatiotemporal chaos
Might be defined as the dynamics, disordered in time and space, of a large
aspect ratio system (i.e. one that is large compared to the size of a basic
chaotic element)
Natural examples
• The atmosphere and ocean (weather, climate etc.)
• Heart fibrillation
Spatiotemporal chaos is a new paradigm of unpredictable dynamics.
(What effect does the butterfly really have?)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 39
Spatiotemporal chaos and turbulence both correspond toL→∞
Define length scales:
• Energy injection scaleLI
• Energy dissipation scaleLD
• System sizeL
Spatiotemporal chaosL >> LD ∼ LITurbulence L ∼ LI >> LD
“Spatiotemporal turbulence” L >> LI >> LD
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 40
Spatiotemporal Chaos v. Turbulence
EnergyInjection
EnergyDissipation
ÿ�
ÿ�
Turbulence
Spatiotemporal Chaos
Spatiotemporal Turbulence
ÿ
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 41
Systems
• Coupled Maps
x(n+1)i = f (x(n)i )+ g × (x(n)i−1− 2x(n)i + x(n)i+1)
• PDE simulations
– Complex Ginzburg-Landau Equation
∂tA = A+ (1+ ic1)∇2A− (1− ic3) |A|2A
– Kuramoto-Sivashinsky equation
∂tu = −∂2xu− ∂4
xu− u∂xu
• Physical systems (experiment and numerics)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 42
Examples from convection
T+∆T
T
• Spiral Defect Chaos ( experiment , simulations )
• Domain Chaos (model simulations: stripes , orientations , walls ;
convection simulations )
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 43
Issues
• System-specific questions
• Definition and characterization
– How to narrow the phenomena
– How to decide if theory, simulation, and experiment match
• Transitions to spatiotemporal chaos and between different chaotic
states
• Thermodynamics and hydrodynamics (cf. Granular Flows)
• Control
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 43
Issues
• System-specific questions
• Definition and characterization
– How to narrow the phenomena
– How to decide if theory, simulation, and experiment match
• Transitions to spatiotemporal chaos and between different chaotic
states
• Thermodynamics and hydrodynamics (cf. Granular Flows)
• Control
Ideas and methods from dynamical systems, statistical mechanics, phase
transition theory …
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 44
Spatiotemporal Chaos in RotatingRayleigh-Bénard Convection
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 45
Ray
leig
h N
umbe
r
stripes(convect ion)
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 46
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 47
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 48
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 49
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 50
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 51
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 52
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 53
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
lockeddomainchaos
(?)
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 54
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
lockeddomainchaos
(?)
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 55
Rotat ion Rate
Ray
leig
h N
umbe
r
no pattern(conduction)
stripes(convect ion)
KL
spiralchaos
lockeddomainchaos
(?)
domainchaos
R C
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
∂tA1 = εA1
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
∂tA1 = εA1+ ∂2x1A1
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
∂tA1 = εA1+ ∂2x1A1− A1(A
21
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
∂tA1 = εA1+ ∂2x1A1− A1(A
21+ g+A2
2+ g−A23)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
∂tA1 = εA1+ ∂2x1A1− A1(A
21+ g+A2
2+ g−A23)
∂tA2 = εA2+ ∂2x2A2− A2(A
22+ g+A2
3+ g−A21)
∂tA3 = εA3+ ∂2x3A3− A3(A
23+ g+A2
1+ g−A22)
whereε = (R − Rc(�)/Rc(�)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 56
Onset of Domain Chaos(Tu and MCC)
Amplitudes of rolls at 3 orientationsAi(r , t), i = 1 . . .3
∂tA1 = εA1+ ∂2x1A1− A1(A
21+ g+A2
2+ g−A23)
∂tA2 = εA2+ ∂2x2A2− A2(A
22+ g+A2
3+ g−A21)
∂tA3 = εA3+ ∂2x3A3− A3(A
23+ g+A2
1+ g−A22)
whereε = (R − Rc(�)/Rc(�)Rescale space, time, and amplitudes:
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 57
RescaleX = ε1/2x, T = εt, A = ε−1/2A
∂T A1 = A1+ ∂2X1A1− A1(A
21+ g+A2
2+ g−A23)
∂T A2 = A2+ ∂2X2A2− A2(A
22+ g+A2
3+ g−A21)
∂T A3 = A3+ ∂2X3A3− A3(A
23+ g+A2
1+ g−A22)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 57
RescaleX = ε1/2x, T = εt, A = ε−1/2A
∂T A1 = A1+ ∂2X1A1− A1(A
21+ g+A2
2+ g−A23)
∂T A2 = A2+ ∂2X2A2− A2(A
22+ g+A2
3+ g−A21)
∂T A3 = A3+ ∂2X3A3− A3(A
23+ g+A2
1+ g−A22)
Numerical simulations show chaotic dynamics
Therefore in unscaled (physical) units
Length scale ξ ∼ ε−1/2
Time scale τ ∼ ε−1
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 58
Experimental results
Hu et al. Phys. Rev. Lett.74, 5040 (1995)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 59
0 0.1 0.2ε
100/
ξ2
0
1
2
3
4
5
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 60
Effect of finite size
Hypothesis: deviations from the predicted scaling behavior comes from
the finite size of the experimental system (MCC, Louie, and Meiron 2001)
Use numerical simulations to test a finite size scaling ansatz
ξM = ξf (0/ξ), ξ = ξ0ε−1/2
Generalized Swift-Hohenberg equation (real field of two spatial
dimensionsψ(x, y; t))∂ψ
∂t= εψ + (∇2+ 1)2ψ − g1ψ
3
+ g2z·∇ × [(∇ψ)2∇ψ ] + g3∇·[(∇ψ)2∇ψ ]
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 61
Variation withε
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 62
Variation with system size
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 63
Calculation of domain size
qx
qy
ξΜ
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 64
Domain size v.ε
ε
1/ξ Μ
00
Γ=30πΓ=40πΓ=50πΓ=80π
0.1 0.2 0.3
0.02
0.04
0.06
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 65
Finite size scaling
0
Γ=30πΓ=40πΓ=50πΓ=80π
ε−1/2/Γ
0 0.05 0.10
0.1
0.2
0.3
ε−1/2
/ξΜ
ξM = ξf (0/ξ), ξ = ξ0ε−1/2
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 66
Characterizing Spatiotemporal Chaos
Methods from statistical physics: Correlation lengths and times, etc.
Methods from dynamical systems:Lyapunov exponents and attractor
dimensions.
Quantifies sensitive dependence on initial conditions.
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 67
Lyapunov Exponent
t0
t1 t2
t3
tf
δu0
δufX
Y
Z
λ = lim tf→∞1
tf−t0 ln∣∣∣δufδu0
∣∣∣
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 68
Lyapunov vector for spiral defect chaos
(from Keng-Hwee Chiam, Caltech thesis 2003)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 69
Lyapunov spectrum for spiral defect chaos
(from Egolf et al. 2000)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 70
Dimensions
Most definitions of fractal dimension involve the geometry of the chaotic
attractor in phase space, which is inaccessible in spatiotemporal chaos.
TheLyapunov dimension:
• is accessible to simulations (but probably not to experiment)
• is usually equal to the information dimension
• is expected to be extensive for large system sizes (Ruelle, 1982)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 71
Lyapunov dimension
Defineµ(n) =∑ni=1 λi (λ1 ≥ λ2 · · · ) with λi theith Lyapunov
exponent.
DL is the interpolated value ofn givingµ = 0 (the dimension of the
volume that neither grows nor shrinks under the evolution)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 72
Lyapunov dimension for spiral defect chaos
(Egolf et al. 2000)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 73
Microextensive scaling for the 1d Kuramoto-Sivashinskyequation
78 83 88 93L
15
16
17
18
Lyap
unov
Dim
ensi
on
(from Tajima and Greenside 2000)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 74
Microextensive scaling
Interesting questions:
• Are there (tiny) windows of periodic orbits or chaotic orbits of
non-scaling dimension so that smooth variation is only in theL→∞limit, or is the variation smooth at finiteL
• Can we use this to define spatiotemporal chaos for finiteL?
• Does spatiotemporal chaos in Rayleigh-Bénard Convection show
microextensive scaling of the Lyapunov dimension?
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 75
Conclusions
Small system chaos
• Lorenz model
• experimental demonstration of chaos in continuum systems
Pattern chaos
• aperiodic flow due to pattern dynamics
• power law tail in spectrum due to defect nucleation events
Spatiotemporal chaos
• new paradigm for unpredictable dynamics (cf. chaos, turbulence)
• quantitative predictions for domain chaos in rotating convection, but discrepancies
with experiment remain
• Lyapunov exponent, vector and dimension
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 76
References
Small system chaos
E. Lorenz J. Atmos. Sciences,20, 130 (1963)
M. Gorman, P. J. Widmann, and K. A. Robbins (thermosyphon) Physica19D, 255 (1986);36D, 157(1989)
L. D. Landau Akad. Nauk. Dok.44 339 (1944) (Also in early editions ofFluid Mechanicsby Landau andLifshitz.)
D. Ruelle and F. TakensCommun. Math. Phys.20 167 (1971)
M. J. Feigenbaum J. Stat. Phys.19 25 (1978)
G. Ahlers Phys. Rev. Lett.30, 1185 (1974)
J. P. Gollub and H. L. Swinney Phys. Rev. Lett.35927 (1975)
A. Libchaber and J. Maurer J. Phys. Lett. Paris39 L369 (1978)
G. Ahlers and R.P. Behringer Phys. Rev. Lett.40, 712 (1978)
A. Libchaber, C. Laroche, and S. FauveJ. Phys. Lett. Paris43L211 (1982)
Pattern Chaos
G. Ahlers Phys. Rev. Lett.30, 1185 (1974)
Back Forward
Benasque PHYSBIO 2003:Pattern Formation in Rayleigh-Bénard Convection - Lecture 3 77
G. Ahlers and R.P. Behringer Phys. Rev. Lett.40, 712 (1978)
H. Gao and R.P. Behringer Phys. Rev. A30, 2837 (1984)
V. Croquette, P. Le Gal, and A. PocheauPhys. Scr. T13, 135 (1986)
M. R. Paul, M. C. Cross, P. F. Fischer, and H.§. GreensidePhys. Rev. Lett.87 154501 (2001)
H. S. Greenside, M. C. Cross, and W. M. CoughranPhys. Rev. Lett.60 2269 (1988)
A. Pocheau and F. DaviaudPhys. Rev.E55353 (1997)
Spatiotemporal chaos
Domain Chaos
Y. Tu and M. C. Cross Phys. Rev. Lett.69 2515 (1992)
Y. Hu, R. E. Ecke, and G. Ahlers Phys. Rev. Lett.74 5040 (1995)
M. C. Cross, M. Louie, and D. Meiron Phys. Rev.E6345201 (2001)
Lyapunov exponents and dimension
D. A. Egolf, I. V. Melnikov, W. Pesch, and R. E. EckeNature404736 (2000)
S. Tajima and H. S. GreensidePhys. Rev.E66 017205 (2002)