1d plot

Can you provide a minimal runnable example that reproduces the error.

Johan

On Friday December 2 2011 11:01:00 Jacopo Lanzoni wrote:
> New question #180690 on DOLFIN:
> https://answers.launchpad.net/dolfin/+question/180690
>
> Dear all,
>
> I wrote a script for a 1-dimensional Dirichlet problem for Helmholtz
> equation but, when I try to plot the solution simply with plot(u), I get
> the following error message
>
> Solving linear variational problem.
> Object cannot be plotted directly, projecting to piecewise linears.
> Traceback (most recent call last):
> File "Helmholtz1D_PML.py", line 84, in <module>
> plot(ur)
> File
> "/Applications/FEniCS.app/Contents/Resources/lib/python2.6/site-packages/d
> olfin/common/plot.py", line 70, in dolfin_plot return
> viper_dolfin.plot(make_viper_object(object), *args, **kwargs) File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> _dolfin.py", line 431, in plot File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> _dolfin.py", line 380, in plot File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> _dolfin.py", line 394, in autoplot File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> _dolfin.py", line 47, in __init__ File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> _dolfin.py", line 72, in plot File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> _dolfin.py", line 150, in plot_genericfunction File
> "/Users/johannr/fenics-11.05/local/lib/python2.6/site-packages/viper/viper
> .py", line 1261, in plot_xy TypeError: function takes exactly 2 arguments
> (1 given)
>
> The same script rewritten for a 2-dimensional case works perfectly fine.
> What should I add to the plot command?
>
> Best wishes
> Jacopo

Dear Johan,

Here is my code.
--------------------------------------------------------------------------------------------------------
from dolfin import *

# Input data
b = 10.0
mesh = Interval(100, -b, b) # domain [-10,10] with 100 elements (h=.2)
p = 1
a = 8 # PML INTERNAL boundary radius
sigmaval = 1
n = 1 # Power order of the PML function
k = 5
potval = 3*k**2 # k_1 = 2*k_0
l = 3 # Obstacle length
f_r = Expression('+cos(k*x[0])', {'k': k}) # Plane wave coming from right
f_i = Expression('-sin(k*x[0])', {'k': k})

# Definition of SubDomainS
class Domain(SubDomain):
    def inside(self, x, on_boundary):
        return abs(x[0]) <= a
class PML(SubDomain):
    def inside(self, x, on_boundary):
        return abs(x[0]) > a
class Obstacle(SubDomain):
    def inside(self, x, on_boundary):
 return abs(x[0]) < l/2

# Function 'pml' is 1 on PML, 0 elsewhere
pml = MeshFunction('uint', mesh, 1)
pml.set_all(0)
pml0 = PML()
pml0.mark(pml, 1)

# Function 'obstacle' is 1 on Obstacle, 0 elsewhere
obstacle = MeshFunction('uint', mesh, 1)
obstacle.set_all(0)
obstacle0 = Obstacle()
obstacle0.mark(obstacle, 1)

# Definition of boundaries
def PML_boundary(x, on_boundary):
    return on_boundary

V = FunctionSpace(mesh,'CG',p)
W = V*V

# Definition of boundary conditions
bcr = DirichletBC(W.sub(0), Constant(0.0), PML_boundary) # Dirichlet condition
bci = DirichletBC(W.sub(1), Constant(0.0), PML_boundary)
bc = [bcr, bci]

V0 = FunctionSpace(mesh, 'DG', n)

sigma = Function(V0)
sigma_values = [0, sigmaval]
for cell_no in range(len(pml.array())):
    subdomain_no = pml.array()[cell_no]
    sigma.vector()[cell_no] = sigma_values[subdomain_no]

Vpot = Function(V0)
V_values = [0, potval] # values of V in the two subdomains
for cell_no in range(len(obstacle.array())):
    obstacle_no = obstacle.array()[cell_no]
    Vpot.vector()[cell_no] = V_values[obstacle_no]

x = Expression('x[0]')
sigma = sigma*(x-a)**n
Ar = 1/(1+sigma**2)
Ai = sigma/(1+sigma**2)
C = Vpot + k**2
(u_r, u_i) = TrialFunctions(W)
(v_r, v_i) = TestFunctions(W)

ar = inner( Ar*grad(u_r) +Ai*grad(u_i), grad(v_r))*dx - inner(C*(u_r -sigma*u_i), v_r)*dx
ai = inner(-Ai*grad(u_r) +Ar*grad(u_i), grad(v_i))*dx - inner(C*(sigma*u_r +u_i), v_i)*dx
Lr = Vpot*(f_r -sigma*f_i)*v_r*dx
Li = Vpot*(sigma*f_r +f_i)*v_i*dx

problem = VariationalProblem(ar+ai, Lr+Li, bc)
u = problem.solve()

ur, ui = split(u)

plot(ur)
---------------------------------------------------------------------------------

Thanks Johan Hake, that solved my question.