def main():
''' An example of performing a continuation for a 2D lid-driven cavity and detecting a bifurcation point'''
dim = 2
dof = 3
nx = 32
ny = nx
nz = 1
n = dof * nx * ny * nz
# Define a point of interest
poi = (nx // 2 - 1, ny // 4 - 1)
# Define the problem
parameters = {
'Problem Type':
'Lid-driven cavity',
# Problem parameters
'Reynolds Number':
1,
'Lid Velocity':
0,
# Use a stretched grid
'Grid Stretching Factor':
1.5,
# Set a maximum step size ds
'Maximum Step Size':
500,
# Give back extra output (this is also more expensive)
'Verbose':
True,
# Value describes the value that is traced in the continuation
# and time integration methods
'Value':
lambda x: utils.create_state_mtx(x, nx, ny, nz, dof)[poi[0], poi[1], 0,
0]
}
interface = Interface(parameters, nx, ny, nz, dim, dof)
print('Looking at point ({}, {})'.format(
interface.discretization.x[poi[0]],
interface.discretization.y[poi[1]]))
continuation = Continuation(interface, parameters)
# Compute an initial guess
x0 = numpy.zeros(n)
x0 = continuation.continuation(x0, 'Lid Velocity', 0, 1, 0.1)[0]
# Perform an initial continuation to Reynolds number 7000 without detecting bifurcation points
ds = 100
target = 7000
x, mu, data1 = continuation.continuation(x0, 'Reynolds Number', 0, target,
ds)
parameters['Newton Tolerance'] = 1e-12
parameters['Destination Tolerance'] = 1e-4
parameters['Detect Bifurcation Points'] = True
parameters['Maximum Step Size'] = 100
# parameters['Eigenvalue Solver'] = {}
# parameters['Eigenvalue Solver']['Target'] = 3j
# parameters['Eigenvalue Solver']['Tolerance'] = 1e-9
# parameters['Eigenvalue Solver']['Number of Eigenvalues'] = 20
# Now detect the bifurcation point
target = 10000
x2, mu2, data2 = continuation.continuation(x, 'Reynolds Number', mu,
target, ds)
# Compute the unstable branch after the bifurcation
parameters['Detect Bifurcation Points'] = False
parameters['Maximum Step Size'] = 2000
target = 10000
parameters['Newton Tolerance'] = 1e-4
x3, mu3, data3 = continuation.continuation(x2, 'Reynolds Number', mu2,
target, ds)
# Plot a bifurcation diagram
plt.plot(data1.mu, data1.value)
plt.plot(data2.mu, data2.value)
plt.plot(data3.mu, data3.value)
plt.show()
# Add a perturbation based on the eigenvector
interface.set_parameter('Reynolds Number', mu2)
_, v = interface.eigs(x2, True)
v = v[:, 0].real
v = plot_utils.create_state_mtx(v, nx, ny, nz, dof)
# Plot the velocity magnutide
plot_utils.plot_velocity_magnitude(v[:, :, 0, 0], v[:, :, 0, 1], interface)
# Plot the pressure
plot_utils.plot_value(v[:, :, 0, 2], interface)
def main():
''' An example of performing a "poor man's continuation" for a 2D lid-driven cavity using time integration'''
dim = 2
dof = 3
nx = 16
ny = nx
nz = 1
n = dof * nx * ny * nz
# Define a point of interest
poi = (nx // 2 - 1, ny // 4 - 1)
# Define the problem
parameters = {
'Problem Type':
'Lid-driven cavity',
# Problem parameters
'Reynolds Number':
0,
'Lid Velocity':
1,
# Use a stretched grid
'Grid Stretching Factor':
1.5,
# Set a maximum step size ds
'Maximum Step Size':
500,
# Give back extra output (this is also more expensive)
'Verbose':
True,
# Value describes the value that is traced in the continuation
# and time integration methods
'Value':
lambda x: utils.create_state_mtx(x, nx, ny, nz, dof)[poi[0], poi[1], 0,
0],
'Theta':
1
}
interface = Interface(parameters, nx, ny, nz, dim, dof)
print('Looking at point ({}, {})'.format(
interface.discretization.x[poi[0]],
interface.discretization.y[poi[1]]))
x = numpy.random.random(n)
mu_list = []
value_list = []
for mu in range(0, 100, 10):
interface.set_parameter('Reynolds Number', mu)
time_integration = TimeIntegration(interface, parameters)
x, t, data = time_integration.integration(x, 1, 10)
# Plot the traced value during the time integration
# plt.plot(data.t, data.value)
# plt.show()
mu_list.append(mu)
value_list.append(data.value[-1])
# Plot a bifurcation diagram
plt.plot(mu_list, value_list)
plt.show()