Lorenz equations
\[\frac{dx}{dt} = sigma(y -
x)\]
\[\frac{dy}{dt} = rx - y - xz\]
\[\frac{dz}{dt} = xy - bz\]
where:
- sigma : prandtl number (ratio between momentum diffusion and heat
diffusion)
- r : Rayleigh number over critical Rayleigh number
- b : size of the region to be approximated
- (Runge-Kutta and the Lorenz Attractor)
Coding Solution
Core function
nx_lorenz = function (particle, dt, ...) {
sigma = 10. # prandtl number
r = 28. # Rayleigh number
b = 8./3. # Size of the region to be approximated
x=particle$x ; y=particle$y ; z=particle$z
dx = (sigma*(y - x))*dt
dy = (r*x - y - x*z)*dt
dz = (x*y - b*z)*dt
return (c(x+dx, y+dy, z+dz))
}
Initialzing for 2 very close particles
## Integration settings
nt = seq(5000) # Number of timesteps
dt = 0.01 # timestep
## Setup position of the two particles
x0.1 = 0.3000 ; y0.1 = 0.3000 ; z0.1 = 0.3000
x0.2 = 0.3001 ; y0.2 = 0.3001 ; z0.2 = 0.3001
particle.1 = tibble("x" = x0.1, "y" = y0.1, "z" = z0.1)
particle.2 = tibble("x" = x0.2, "y" = y0.2, "z" = z0.2)
Run the calculation with estimating time
## Estimating time using microbenchmark
system.time({
for (i in 2:(length(nt)+1)) {
particle.1 = rbind(particle.1, nx_lorenz(particle.1[i-1, ], dt))
particle.2 = rbind(particle.2, nx_lorenz(particle.2[i-1, ], dt))
}
})
## user system elapsed
## 2.127 0.027 2.155
summary(particle.1)
## x y z
## Min. :-20.000 Min. :-26.719 Min. : 0.2557
## 1st Qu.: -8.384 1st Qu.: -8.612 1st Qu.:20.2868
## Median : -3.936 Median : -3.868 Median :26.0437
## Mean : -1.932 Mean : -1.939 Mean :24.9044
## 3rd Qu.: 4.450 3rd Qu.: 4.281 3rd Qu.:29.3136
## Max. : 21.454 Max. : 29.285 Max. :54.1582
summary(particle.2)
## x y z
## Min. :-19.971 Min. :-26.657 Min. : 0.2558
## 1st Qu.: -7.702 1st Qu.: -8.055 1st Qu.:19.3421
## Median : -1.762 Median : -1.668 Median :25.1407
## Mean : -1.020 Mean : -1.005 Mean :24.6996
## 3rd Qu.: 5.703 3rd Qu.: 5.775 3rd Qu.:30.0023
## Max. : 21.455 Max. : 29.286 Max. :54.1585
Visualization
Tracking the temporal position of the two
particles
plot_ly(x = particle.1$x, y = particle.1$y, z = particle.1$z,
type = 'scatter3d', mode = 'lines', opacity = 1,
line = list(width = 2, color = "blue", reverscale = FALSE)) %>%
add_trace(x = particle.2$x, y = particle.2$y, z = particle.2$z,
line = list(color = "red", width = 2,reverscale = FALSE))
Tracking the temporal distance between the two
particles
plot_ly(x = particle.2$x - particle.1$x,
y = particle.2$y - particle.1$y,
z = particle.2$z - particle.1$z,
type = 'scatter3d', mode = 'lines', opacity = 1,
line = list(width = 2, color = "orange", reverscale = FALSE))
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