Recorncias e difuso anmala em sistemas Hamiltoneanos caticos IFUSP - - PowerPoint PPT Presentation

recor ncias e difus o an mala em sistemas hamiltoneanos
SMART_READER_LITE
LIVE PREVIEW

Recorncias e difuso anmala em sistemas Hamiltoneanos caticos IFUSP - - PowerPoint PPT Presentation

Recorncias e difuso anmala em sistemas Hamiltoneanos caticos IFUSP - Maro 2010 Eduardo. G. Altmann http://www.tinyurl.com/ifusp2010 Apresentao 1: Torus, mapa padro, ilhas ao redor de ilhas Text Apresentao II:


slide-1
SLIDE 1

Recorências e difusão anômala em sistemas Hamiltoneanos caóticos

IFUSP - Março 2010

  • Eduardo. G. Altmann

http://www.tinyurl.com/ifusp2010

slide-2
SLIDE 2

Torus, mapa padrão, ilhas ao redor de ilhas

Apresentação 1:

slide-3
SLIDE 3
slide-4
SLIDE 4
slide-5
SLIDE 5
slide-6
SLIDE 6
slide-7
SLIDE 7

Text

slide-8
SLIDE 8

Recorrências para detectar rompimento de tori

Apresentação II:

slide-9
SLIDE 9

Interesse em sistemas temporalmente reversíveis (dinâmica quasi-Hamiltoneana pode aparecer)

N=4 Se mais de uma simetria esta presente no sistema a condição de torção (twist) é violada:

Exemplo: Osciladores acoplados

slide-10
SLIDE 10

4 osciladores de fase acoplados: Sistema não Hamiltoneano mas com dinâmica quasi- Hamiltoneana e torus “não torcidos”!

slide-11
SLIDE 11

Reconexões típicas de sistemas não torcionais também ocorrem em sistemas não Hamiltoneanos

slide-12
SLIDE 12

Para o transporte de trajetórias é importante determinar quais parâmetros (ε,ω) o torus existe (divide o espaço de fases). Método: Verificar se uma trajetória que tem de pertencer ao torus (IP) satisfaz o teorema de Slater (e.g., máximo 3 Ts distintos). Particularmente útil quando:

  • Grande número de parâmetros (ε,ω) tem de ser varridos.
  • Sistema de tempos contínuo (difícil integração/sessão de Poincaré)
  • Parâmetros (ε,ω) próximos ao rompimento são escolhidos. Nesse

caso ilhas (“stickiness”) fazem demais métodos muito lentos Limitação:

  • N=4, i.e., mapas bi-dimensionais
slide-13
SLIDE 13
slide-14
SLIDE 14

Rompimento do torus

slide-15
SLIDE 15

Exemplo: mapa bi-dimensional Simetrias: Jacobiano:

slide-16
SLIDE 16

Exemplo: mapa bi-dimensional

slide-17
SLIDE 17

mapa linear por partes espaço de fases hierárquico

Apresentação III:

slide-18
SLIDE 18
slide-19
SLIDE 19
slide-20
SLIDE 20
slide-21
SLIDE 21

Caso genérico / hierárquico

slide-22
SLIDE 22
slide-23
SLIDE 23

Figure 2.3: (Color online) Sticking time distribution ρ(τ) for 100 differ- ent standard maps (2.14) with a con- stant K† added to the y equation: K ∈ [0.5, 0.6], K† ∈ [0, 0.2]. The central green (gray) curve is the average [for fixed ρ(τ)] over all curves, and the red curve (axis on the right) corresponds to the standard deviation of the curves (for fixed ρ(τ) projected to the x-axis). The further parameters are equivalent to those of Fig. 6.1b below.

slide-24
SLIDE 24

Efeito de ruído branco e altas dimensões no aprisionamento de trajetórias

Apresentação 1V:

slide-25
SLIDE 25

Coupled standard maps: 2.1 Motivation / model 2.2 Noise perturbation 2.3 High dimensional

slide-26
SLIDE 26

Qual o problema? (do ponto de vista de Mec. Estatística)

slide-27
SLIDE 27

Violate the hypothesis of strong chaos:

  • 1. Ergodicity, i.e., negligible measure of regular components
  • 2. Strong mixing, i.e., fast decay of correlations

Qual o problema? (do ponto de vista de Mec. Estatística)

slide-28
SLIDE 28

Violate the hypothesis of strong chaos:

  • 1. Ergodicity, i.e., negligible measure of regular components
  • 2. Strong mixing, i.e., fast decay of correlations

What happens for increasing phase space dimension? Qual o problema? (do ponto de vista de Mec. Estatística)

slide-29
SLIDE 29

C

Coupled symplectic maps model

slide-30
SLIDE 30

C

Coupled symplectic maps model

slide-31
SLIDE 31

Map is symplectic iff:

Coupled symplectic maps model

slide-32
SLIDE 32

Map is symplectic iff:

Coupled symplectic maps model

slide-33
SLIDE 33

Coupled standard maps: 2.1 Motivation / model 2.2 Noise perturbation 2.3 High dimensional

slide-34
SLIDE 34

2.2 Noise perturbation

slide-35
SLIDE 35

2.2 Noise perturbation

slide-36
SLIDE 36

2.2 Noise perturbation

slide-37
SLIDE 37

2.2 Noise perturbation

slide-38
SLIDE 38

2.2 Noise perturbation

slide-39
SLIDE 39

2.2 Noise perturbation

slide-40
SLIDE 40

2.2 Noise perturbation RW theory

slide-41
SLIDE 41

2.2 Noise perturbation RW theory

slide-42
SLIDE 42

2.2 Noise perturbation RW theory

slide-43
SLIDE 43

Coupled standard maps: 2.1 Motivation / model 2.2 Noise perturbation 2.3 High dimensional

slide-44
SLIDE 44

C

Coupled symplectic maps model

slide-45
SLIDE 45

Coupled symplectic maps model Ergodicity?

C

slide-46
SLIDE 46

Coupled symplectic maps model Ergodicity?

C

slide-47
SLIDE 47

Coupled symplectic maps model Ergodicity?

slide-48
SLIDE 48

N=2-5 show power-law behavior [Kantz, Grassberger (1987), Ding, Bountis, Ott (1990)]

  • 1. Ergodicity, i.e., negligible measure of regular components
  • 2. Strong mixing, i.e., fast decay of correlations

✘ ✔

e.g., zero measure sets on Bunimovich stadium Billiards

?

Coupled symplectic maps model

slide-49
SLIDE 49

Coupled symplectic maps model Strong mixing?

slide-50
SLIDE 50

Coupled symplectic maps model Strong mixing?

slide-51
SLIDE 51

Coupled symplectic maps model Strong mixing?

slide-52
SLIDE 52

Coupled symplectic maps model Strong mixing?

slide-53
SLIDE 53

Coupled symplectic maps model

slide-54
SLIDE 54

Coupled symplectic maps model

slide-55
SLIDE 55

Coupled symplectic maps model

slide-56
SLIDE 56

Coupled symplectic maps model

slide-57
SLIDE 57

Coupled symplectic maps model Strong mixing?

slide-58
SLIDE 58

Coupled symplectic maps model Strong mixing?

N=2,3,4,5 N=2,3,...,15 ξ = 0.05 ξ = 0.05

slide-59
SLIDE 59

Coupled symplectic maps model Strong mixing?

slide-60
SLIDE 60
  • 1. Ergodicity, i.e., negligible measure of regular components
  • 2. Strong mixing, i.e., fast decay of correlations

Coupled symplectic maps

✔ ✔

Non-exponential decay, but sufficiently fast power-law

slide-61
SLIDE 61

fluído incompressível

Apresentação V:

slide-62
SLIDE 62

Passive scalar field θ( x, t) (contaminant), advected by a flow with velocity field given by v(x, t) [Aref,1984] ∂θ ∂t + ∇.( vθ) = Dm∇2θ, where Dm is the molecular diffusion coefficient. The motion of fluid elements (Lagrangian description) is written as d x dt = v( x, t) + η(t), where ηi(t)ηj(t) = 2Dmδi,jδ(t − t). Consider an incompressible ∇. v = 0 2-D fluid x = (x, y). In this case there exist a stream function ψ(x, y, t) such that dx dt = vx = −∂ψ ∂y and dy dt = vy = ∂ψ ∂x .

slide-63
SLIDE 63

Consider a fluid channel infinite in the x direction having the following two flows: Laminar regime: ψ1(x, y) = −v1 sin(πy); Vortex regime: ψ2(x, y) = v2cos(2x)(1 − y2)2

slide-64
SLIDE 64

xn = xn+1 + λ sin(πyn) − 2ρ

π yn(1 − y2 n) cos[2π(xn + 1)] + ξδn,

yn+1 = yn − ρ(1 − y2

n)2 sin[2πxn+1] + ξδ n.

ρ = πv2t0/2 – intensity of the vortex regime; λ = v1t0/2 – intensity of the laminar regime; ξ – intensity of the white noise variable δ (ξ ∼ √Dm); Alternating periodically between the two regimes in a period t0 and mapping the evolution from nt0 → (n + 1)t0 one gets

slide-65
SLIDE 65

xn = xn+1 + λ sin(πyn) − 2ρ

π yn(1 − y2 n) cos[2π(xn + 1)] + ξδn,

yn+1 = yn − ρ(1 − y2

n)2 sin[2πxn+1] + ξδ n.

Espaço misto para dois parâmetros de controle

slide-66
SLIDE 66

10

1

10

2

10

3

10

4

10

5

10

6

t

10

1

10

2

10

3

10

4

10

5

10

6

10

7

x

2

=0.25 =1 =1 (normal diffusion) =2 (ballistic motion)

Transporte super-difusivo

slide-67
SLIDE 67

10 20 30 40 50 60 70 80

x =1 =0.25

  • 0.8
  • 0.4

0.4 0.8

y 100 200 300 400 500 number of iterations - n

  • 0.8
  • 0.4

0.4 0.8

y

trapping trapping flying flying flying flying flying f l y i n g

slide-68
SLIDE 68

10

1

10

2

10

3

10

4

10

5

  • 10
  • 8

10

  • 7

10

  • 6

10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

()

Flights =1 Flights =0.25 Traps =0.25

Estatísitca de aprisionamento e voo

slide-69
SLIDE 69

10

1

10

2

10

3

10

4

10

5

  • 10
  • 8

10

  • 7

10

  • 6

10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

() =0 =0.005 =0.001

=1 flies

10

1

10

2

10

3

10

4

10

5

  • 10
  • 8

10

  • 7

10

  • 6

10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

()

=0.25 flies

10

1

10

2

10

3

10

4

10

5

  • 10
  • 8

10

  • 7

10

  • 6

10

  • 5

10

  • 4

10

  • 3

10

  • 2

10

  • 1

()

=0.25 traps

Efeito da difusão molecular no aprisionamento Tempo final do regime de super-aprisionamento t~ 1/ξ2

slide-70
SLIDE 70

10 100 1000 10000 1e+05 t 1 10 100 D=x

2/t

=0 =0.0004 =0.001 =0.002 =0.01 =0.1 (a) =0.6 =1

Coeficiente de difusão como função do tempo λ=1

slide-71
SLIDE 71

λ=0.25 Coeficiente de difusão como função do tempo

slide-72
SLIDE 72

Difusão total (advecção+molecular) como função da difusão molecular

slide-73
SLIDE 73