Hi,
I wrote a code for a time depending system of four pde in a 2-dimensional
space, but I’ve some problem.
I create my .msh file with gmsh and then I run my code on this domain (with a
final time = 50, that is a very little time, I need at least 365) i received
the error message:
*** Error: Unable to successfully call PETSc function 'VecNorm'.
*** Reason: PETSc error code is: 72.
If I use as domain RectangleMesh(-10, -10, 10, 10, 20, 20) at final time = 60 i
received the error message
*** Error: Unable to solve nonlinear system with NewtonSolver.
*** Reason: Newton solver did not converge.
So I’ve two problem:
1) Why on my .xml doesn’t work the code ?
2) How can I increase the final time ? (if it’s possible)
As model I use Cahn Hillard demo.
A piece of my code:
[…]
F = (f1 + f2 + f3 + f4)*dx
dF = derivative(F,u,du)
# Create nonlinear problem and Newton solver
problem = BrainGBM(dF, F, bc)
solver = NewtonSolver()
prm = solver.parameters
prm['linear_solver']='lu' #umfpack
prm["absolute_tolerance"] = 1E-8
prm["relative_tolerance"] = 1E-7
prm["convergence_criterion"] = "incremental"
prm['maximum_iterations'] = 1000
prm["relaxation_parameter"] = 1.0
# Step in time
t = 0.0
T = 50*dt
i = 1
while (t < T):
print i
t += dt
u0.vector()[:] = u.vector()
solver.solve(problem, u.vector())
file << (u.split()[0], t)
i = i+1
[…]
Best regard
Valentina
_______________________________________________
fenics-support mailing list
[email protected]
http://fenicsproject.org/mailman/listinfo/fenics-support
