Dear all,
I have made a very simple function to export to a format readable by gnuplot.
Gnuplot has a well known mistake, that for the first triangle you have to provide the fourth value two times (so you have 5 lines for the first triangle).
I would like to know, whether this code is well written from the point of view of effectivity, and whether somebody has written something more suitable for higher order spaces (in my case DoFs would be lost, e.g. on facets). Working code is below, you can then visualize in gnuplot using sp "testfile" w l
:
from dolfin import *
# Export to gnuplot
def gnuplot_export(file, mesh, function):
file = open(file, 'w+')
i = 0
for myCell in cells(mesh):
i += 1
myVertices = vertices(myCell)
it = iter(myVertices)
myVertex0 = it.next()
myVertex1 = it.next()
myVertex2 = it.next()
print >>file, '%e %e %e' % (myVertex0.x(0),myVertex0.x(1),function(myVertex0.x(0),myVertex0.x(1)))
print >>file, '%e %e %e' % (myVertex1.x(0),myVertex1.x(1),function(myVertex1.x(0),myVertex1.x(1)))
print >>file, '%e %e %e' % (myVertex2.x(0),myVertex2.x(1),function(myVertex2.x(0),myVertex2.x(1)))
print >>file, '%e %e %e' % (myVertex0.x(0),myVertex0.x(1),function(myVertex0.x(0),myVertex0.x(1)))
if (i == 1):
print >>file, '%e %e %e' % (myVertex0.x(0),myVertex0.x(1),function(myVertex0.x(0),myVertex0.x(1)))
print >>file, ''
NUM_CELL = 24
mesh = UnitSquareMesh(NUM_CELL,NUM_CELL)
h = CellSize(mesh)
n = FacetNormal(mesh)
# Create FunctionSpaces
V = FunctionSpace(mesh, "CG", 3)
W = FunctionSpace(mesh, "CG", 1)
# Boundary conditions
def right(x, on_boundary): return x[0] > (1.0 - DOLFIN_EPS)
def left(x, on_boundary): return x[0] < DOLFIN_EPS
def bottom_center(x, on_boundary):
return x[1] < DOLFIN_EPS and (x[0] > 1./3. - DOLFIN_EPS and x[0] < 2./3.0 + DOLFIN_EPS)
def bottom_lr(x, on_boundary):
return x[1] < DOLFIN_EPS and (x[0] < 1./3. + DOLFIN_EPS or x[0] > 2./3.0 - DOLFIN_EPS)
def top(x, on_boundary):
return x[1] > 1.0 - DOLFIN_EPS
g0 = Constant(0.0)
g1 = Constant(1.0)
bc0 = DirichletBC(V, g1, bottom_center)
bc1 = DirichletBC(V, g0, bottom_lr)
bc3 = DirichletBC(V, g0, top)
bc4 = DirichletBC(V, g0, right)
bcs = [bc0, bc1, bc4, bc3]
# Parameters
epsilon = Constant(0.000000001)
c = Constant(0.)
b = Expression(('-x[1]', 'x[0]'))
f = Constant(0.)
uh = TrialFunction(V)
vh = TestFunction(V)
wh = TestFunction(W)
bb = assemble(dot(b,b)*wh*dx)
bF = Function(W,bb)
tau = 1.
# SUPG (SDFEM) method
a1 = (epsilon*dot(grad(uh),grad(vh)) + vh*dot(b,grad(uh)) + c*uh*vh)*dx
a2 = (h/(2.*sqrt(bF))*tau*inner(dot(b,grad(uh)),dot(b,grad(vh))))*dx
a = a1 + a2
L = f*vh*dx + h/(2.*sqrt(bF))*tau*h*f*dot(b, grad(vh))*dx
# Compute solution
uh = Function(V)
solve(a == L, uh, bcs)
gnuplot_export('./testfile', mesh, uh)