Francesco Magaletti <[email protected]> writes: > First of all thank you everybody for the fast replies. > > Barry, > I think the solution 1 is a good suggestion but I confess I’m a bit scared of > going deep into the modification of the ODE integrator. > I’m now using TSCN, the classical Crank-Nicolson 2nd order full > implicit time integration.
Unless you have experience for this problem, you should consider whether you'd be better off with an L-stable method (like a BDF or DIRK). > Maybe I can try to implement your solution together with BEULER, as > suggested by Hong. By the way the solution changes rapidly, but not > so fast to justify that small timestep value. How did you measure that? It's changing enough that the nonlinear solve is difficult. If you use almost any other implicit method in PETSc (or -ts_type sundials), you'll get an error-based adaptive controller. I do recommend trying those; you might find that shorter time steps are needed for accuracy too, in which case FAS is moot (this is typical).
signature.asc
Description: PGP signature
