def darcy_dual_hybridVEM_example3(**kwargs):
#######################
# Simple 3d Darcy problem with known exact solution
#######################
Nx = Ny = Nz = 7
g = structured.CartGrid([Nx, Ny, Nz], [1, 1, 1])
g.compute_geometry()
kxx = np.ones(g.num_cells)
perm = tensor.SecondOrder(g.dim, kxx)
def funP_ex(pt):
return np.sin(2*np.pi*pt[0])*np.sin(2*np.pi*pt[1])\
* np.sin(2*np.pi*pt[2])
def funU_ex(pt):
return [-2*np.pi*np.cos(2*np.pi*pt[0])\
* np.sin(2*np.pi*pt[1])*np.sin(2*np.pi*pt[2]),
-2*np.pi*np.sin(2*np.pi*pt[0])\
* np.cos(2*np.pi*pt[1])*np.sin(2*np.pi*pt[2]),
-2*np.pi*np.sin(2*np.pi*pt[0])\
* np.sin(2*np.pi*pt[1])*np.cos(2*np.pi*pt[2])]
def fun(pt):
return 12 * np.pi**2 * funP_ex(pt)
f = np.array([fun(pt) for pt in g.cell_centers.T])
b_faces = g.get_boundary_faces()
bnd = bc.BoundaryCondition(g, b_faces, ['dir'] * b_faces.size)
bnd_val = np.zeros(g.num_faces)
bnd_val[b_faces] = funP_ex(g.face_centers[:, b_faces])
solver = hybrid.HybridDualVEM()
data = {'perm': perm, 'source': f, 'bc': bnd, 'bc_val': bnd_val}
H, rhs = solver.matrix_rhs(g, data)
l = sps.linalg.spsolve(H, rhs)
u, p = solver.compute_up(g, l, data)
P0u = dual.DualVEM().project_u(g, u)
if kwargs['visualize']:
plot_grid(g, p, P0u)
p_ex = error.interpolate(g, funP_ex)
u_ex = error.interpolate(g, funU_ex)
np.set_printoptions(linewidth=999999)
np.set_printoptions(precision=16)
errors = np.array(
[error.error_L2(g, p, p_ex),
error.error_L2(g, P0u, u_ex)])
errors_known = np.array([0.1010936831876412, 0.0680593765009036])
assert np.allclose(errors, errors_known)
def darcy_dual_hybridVEM_example2(**kwargs):
#######################
# Simple 2d Darcy problem on a surface with known exact solution
#######################
Nx = Ny = 25
g = simplex.StructuredTriangleGrid([Nx, Ny], [1, 1])
R = cg.rot(np.pi / 6., [0, 1, 1])
g.nodes = np.dot(R, g.nodes)
g.compute_geometry(is_embedded=True)
T = cg.tangent_matrix(g.nodes)
kxx = np.ones(g.num_cells)
perm = tensor.SecondOrder(g.dim, kxx)
def funP_ex(pt):
return np.pi * pt[0] - 6 * pt[1] + np.exp(1) * pt[2] - 4
def funU_ex(pt):
return np.dot(T, [-np.pi, 6, -np.exp(1)])
def fun(pt):
return 0
f = np.array([fun(pt) for pt in g.cell_centers.T])
b_faces = g.get_boundary_faces()
bnd = bc.BoundaryCondition(g, b_faces, ['dir'] * b_faces.size)
bnd_val = np.zeros(g.num_faces)
bnd_val[b_faces] = funP_ex(g.face_centers[:, b_faces])
solver = hybrid.HybridDualVEM()
data = {'perm': perm, 'source': f, 'bc': bnd, 'bc_val': bnd_val}
H, rhs = solver.matrix_rhs(g, data)
l = sps.linalg.spsolve(H, rhs)
u, p = solver.compute_up(g, l, data)
P0u = dual.DualVEM().project_u(g, u)
if kwargs['visualize']:
plot_grid(g, p, P0u)
p_ex = error.interpolate(g, funP_ex)
u_ex = error.interpolate(g, funU_ex)
errors = np.array(
[error.error_L2(g, p, p_ex),
error.error_L2(g, P0u, u_ex)])
errors_known = np.array([0, 0])
assert np.allclose(errors, errors_known)
def darcy_dualVEM_example1(**kwargs):
#######################
# Simple 2d Darcy problem with known exact solution
#######################
Nx = Ny = 25
g = structured.CartGrid([Nx, Ny], [1, 1])
g.compute_geometry()
kxx = np.ones(g.num_cells)
perm = tensor.SecondOrder(g.dim, kxx)
def funP_ex(pt):
return np.sin(2 * np.pi * pt[0]) * np.sin(2 * np.pi * pt[1])
def funU_ex(pt):
return [
-2 * np.pi * np.cos(2 * np.pi * pt[0]) * np.sin(2 * np.pi * pt[1]),
-2 * np.pi * np.sin(2 * np.pi * pt[0]) * np.cos(2 * np.pi * pt[1]),
0
]
def fun(pt):
return 8 * np.pi**2 * funP_ex(pt)
f = np.array([fun(pt) for pt in g.cell_centers.T])
b_faces = g.get_boundary_faces()
bnd = bc.BoundaryCondition(g, b_faces, ['dir'] * b_faces.size)
bnd_val = np.zeros(g.num_faces)
bnd_val[b_faces] = funP_ex(g.face_centers[:, b_faces])
solver = dual.DualVEM()
data = {'perm': perm, 'source': f, 'bc': bnd, 'bc_val': bnd_val}
D, rhs = solver.matrix_rhs(g, data)
up = sps.linalg.spsolve(D, rhs)
u, p = solver.extract_u(g, up), solver.extract_p(g, up)
P0u = solver.project_u(g, u)
if kwargs['visualize']:
plot_grid(g, p, P0u)
p_ex = error.interpolate(g, funP_ex)
u_ex = error.interpolate(g, funU_ex)
errors = np.array(
[error.error_L2(g, p, p_ex),
error.error_L2(g, P0u, u_ex)])
errors_known = np.array([0.0210718223032, 0.00526933885613])
assert np.allclose(errors, errors_known)