Dear fipy users and developers,
I have a spherical symmetric problem, so I can use Grid1D mesh, but need to
multiply coefficients in diffusion-convection problem by r^2 and also add
external one like 1/r^2, i.e.
df/dt = 1/r^2 d/dr (r^2 D df/dr) - 1/r^2 d/dr (r^2 u f)
I managed to provide r^2 in "internal" coefficients, but failed to provide
1/r^2 in "external" ones (i.e. something like (1/r**2)*DiffusionTerm). I then
tested to put some CellVariable before the Term as was advised to do earlier
and the script failed again.
Do you have any ideas how to implement this? Below is the test script that
currently works.
Rmax=1.
nsteps=100
dr=Rmax / float(nsteps)
mesh_r = Grid1D(dx = dr, nx = nsteps)
r = mesh_r.getCellCenters()
phi = CellVariable(mesh=mesh_r, name = "variable", value=1.0)
valueLeft = 0
valueRight = 0
BCs = (FixedValue(faces=mesh_r.getFacesLeft(), value=valueLeft),
FixedValue(faces=mesh_r.getFacesRight(), value=valueRight))
convCoeff = FaceVariable(mesh=mesh_r,
value=1.0*mesh_r.getFaceCenters()*mesh_r.getFaceCenters(), rank=1)
diffCoeff = FaceVariable(mesh=mesh_r,
value=1.0*mesh_r.getFaceCenters()*mesh_r.getFaceCenters())
Source = CellVariable(mesh=mesh_r)
Source.setValue(5e-8)
eq = TransientTerm() == DiffusionTerm(coeff = diffCoeff) +
PowerLawConvectionTerm(coeff = convCoeff) + Source
timeStepDuration = 1
time = 1
viewer = Viewer(vars=(phi), limits={'datamin': 1e-15, 'datamax': 1e1})
while time <= t:
timeStep = timeStepDuration
eq.solve(var=phi, boundaryConditions=BCs, dt=timeStep)
time=time+timeStep
viewer.plot()
