Skip to content

Instantly share code, notes, and snippets.

@Corwinpro
Created June 14, 2019 10:37
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save Corwinpro/0e881f2e22e99c87108e0662238852b0 to your computer and use it in GitHub Desktop.
Save Corwinpro/0e881f2e22e99c87108e0662238852b0 to your computer and use it in GitHub Desktop.
import dolfin as dolf
import numpy as np
LOG_LEVEL = 30
dolf.set_log_level(LOG_LEVEL)
mesh = dolf.UnitIntervalMesh(100)
n = dolf.FacetNormal(mesh)
wall_left = "near(x[0], 0)"
wall_right = "near(x[0], 1.)"
all_bound = "on_boundary"
P2 = dolf.FiniteElement("Lagrange", mesh.ufl_cell(), 2)
P1 = dolf.FiniteElement("Lagrange", mesh.ufl_cell(), 2)
P_mixed = dolf.MixedElement([P2, P1])
W = dolf.FunctionSpace(mesh, P_mixed)
w_trial = dolf.TrialFunction(W)
w_test = dolf.TestFunction(W)
u, p = dolf.split(w_trial)
v, q = dolf.split(w_test)
dt = 1.0e-3
NSteps = int(31.0e2)
def B_Form(trial_func, test_func):
u, p = trial_func
v, q = test_func
return (p * q + u * v) * dolf.dx
def A_Form_viscous(trial_func, test_func):
u, p = trial_func
v, q = test_func
mu = dolf.Constant(1.0e-2)
return (-u.dx(0) * q - (-p + mu * u.dx(0)) * v.dx(0)) * dolf.dx
def unsteady_solve(control_velocity=None, plotting=False):
state_0 = dolf.Function(W)
(u0, p0) = state_0.split(True)
w = dolf.Function(W)
for i in range(NSteps):
a = B_Form((u, p), (v, q)) - dolf.Constant(dt * 0.5) * A_Form_viscous(
(u, p), (v, q)
)
L = B_Form((u0, p0), (v, q)) + dolf.Constant(dt * 0.5) * A_Form_viscous(
(u0, p0), (v, q)
)
L -= dolf.Constant(dt * 1.0) * n[0] * v * dolf.ds
control_right = dolf.DirichletBC(W.sub(0), dolf.Constant(0.0), wall_right)
bcs = [control_right]
tol = 1.0e-5
# How should I specify a goal here?
# M = w * dolf.dx() ?
# My guess was:
(_u, _p) = w.split()
M = (_u**2. + _p**2)*dolf.dx()
problem = dolf.LinearVariationalProblem(a, L, w, bcs)
solver = dolf.AdaptiveLinearVariationalSolver(problem, M)
solver.parameters["error_control"]["dual_variational_solver"][
"linear_solver"
] = "cg"
solver.parameters["error_control"]["dual_variational_solver"][
"symmetric"
] = True
solver.solve(tol)
solver.summary()
(u0, p0) = w.split()
unsteady_solve()
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment