1
0
Fork 0

Week5: Ex1.1 Almost correct implementation

This commit is contained in:
Eric Teunis de Boone 2020-03-12 16:53:11 +01:00
parent 5ff1bed84f
commit bda69d945f

131
week6/ex1.py Executable file
View file

@ -0,0 +1,131 @@
#!/usr/bin/env python3
import numpy as np
def pdeHyperbolic(a, x, t, f, g = None, dtype=np.float64):
""" Solve a Hyperbolic Partial Differential using finite differences. """
m = len(x) # Amount of objects to track
n = len(t) # Length of Time Vector
# Determine stepsizes
h = 1
if m > 1:
h = x[0] - x[1]
k = 1
if n > 1:
k = t[0] - t[1]
λ_sq = (a*k/h)**2
# Create array to hold the solution
w = np.zeros((n,m), dtype=dtype)
# Create finite difference matrix
A = np.diag(m*[2*(1 - λ_sq)], k=0) + np.diag((m-1)*[λ_sq], k=-1) + np.diag((m-1)*[λ_sq], k=1)
print(A)
# Initialise first two timesteps
w[0] = f(x, t[0])
w[1] = A@w[0]/2 + k*g(x, t[0])
# Calculate for following timesteps
for j in range(2, n-1):
w[j] = A@w[j-1] - w[j-2]
return w
def test_pdeHyperbolic_case1(m=1e2, n=1e3,l=1, T=1):
a = 1 # from the Schroedinger Equation
# Setup spatial and time grids
x = np.linspace(0, l, m)
t = np.linspace(0, T, n)
# Boundary conditions
def f(x,t):
return np.sin(2*np.pi*x)
def g(x,t):
return 2*np.pi*np.sin(2*np.pi*x)
# Solve it
sol = pdeHyperbolic(a, x, t, f, g)
# Plot it with the exact solution
exact_sol = lambda x,t: np.sin(2*np.pi*x)*(np.cos(2*np.pi*t) + np.sin(2*np.pi*t))
t_high_res = np.linspace(0, T, n * 1e4)
from matplotlib import pyplot
from matplotlib import animation as anim
fig, _ = pyplot.subplots()
if False:
for _, x_i in enumerate(x):
pyplot.plot(t_high_res, exact_sol(x_i, t_high_res), label="Exact")
pyplot.plot(t, sol[:,1], label="iter")
pyplot.grid()
pyplot.xlabel("t")
pyplot.ylabel("w")
else:
def animate(i):
pyplot.clf()
pyplot.ylim(-2, 2)
pyplot.grid()
pyplot.plot(x, exact_sol(x, t[i]), label="exact")
pyplot.plot(x, sol[i,:], label="iter")
frames = np.arange(1, n, dtype=np.int)
myAnim = anim.FuncAnimation(fig, animate, frames, interval = 10 )
pyplot.legend()
pyplot.show()
def test_pdeHyperbolic_case2(m=200, n=400, l=1, T=1):
a = 1 # from the Schroedinger Equation
# Setup spatial and time grids
x = np.linspace(0, l, m)
t = np.linspace(0, T, n)
# Boundary conditions
def f(x,t):
return 2*(x < 0.5) -1
def g(x,t):
return 0
# Solve it
sol = pdeHyperbolic(a, x, t, f, g)
from matplotlib import pyplot
from matplotlib import animation as anim
fig, _ = pyplot.subplots()
def animate(i):
pyplot.clf()
pyplot.grid()
pyplot.ylim(-1.5,1.5)
pyplot.plot(x, sol[i,:], label="iter")
frames = np.arange(1, n)
myAnim = anim.FuncAnimation(fig, animate, frames, interval = 5 )
pyplot.legend()
pyplot.show()
if __name__ == "__main__":
#test_pdeHyperbolic_case1()
test_pdeHyperbolic_case2()