Error with PETSc Nonlinear Solver with No Explicit Boundary Conditions

Errors
1

I am trying to adapt the PETSc Newton solver (here) to allow for adaptive time stepping (as suggested as in a previous post). However, because of the way that I am setting up my problem (elastodynamics with contact), there are no explicit boundary conditions for the elastic body.

When I set the boundary conditions as None, the environment crashes with an MPI error code (I am working in Google Colab). When I set the boundary conditions as [] so that there is an iterable passed to the bc argument, I get the error:

TypeError: apply_lifting(): incompatible function arguments. The following argument types are supported:
    1. (arg0: numpy.ndarray[numpy.float64], arg1: List[dolfinx::fem::Form<double>], arg2: List[numpy.ndarray[numpy.float64]], arg3: List[Dict[Tuple[dolfinx::fem::IntegralType, int], numpy.ndarray[numpy.float64]]], arg4: List[List[dolfinx::fem::DirichletBC<double>]], arg5: List[numpy.ndarray[numpy.float64]], arg6: float) -> None
    2. (arg0: numpy.ndarray[numpy.float32], arg1: List[dolfinx::fem::Form<float>], arg2: List[numpy.ndarray[numpy.float32]], arg3: List[Dict[Tuple[dolfinx::fem::IntegralType, int], numpy.ndarray[numpy.float32]]], arg4: List[List[dolfinx::fem::DirichletBC<float>]], arg5: List[numpy.ndarray[numpy.float32]], arg6: float) -> None
    3. (arg0: numpy.ndarray[numpy.complex128], arg1: List[dolfinx::fem::Form<std::complex<double> >], arg2: List[numpy.ndarray[numpy.complex128]], arg3: List[Dict[Tuple[dolfinx::fem::IntegralType, int], numpy.ndarray[numpy.complex128]]], arg4: List[List[dolfinx::fem::DirichletBC<std::complex<double> >]], arg5: List[numpy.ndarray[numpy.complex128]], arg6: float) -> None

Is there a way to utilize PETSc Newton solvers without passing explicit boundary conditions? Are there ways to pass dirichletbc instantiations that don’t apply any constraints?

2

If you do not have any Dirichlet boundary conditions you do not need to apply lifting. If you want to keep it a possibility for other problems, pass in [[]]
as you can see from the docs: dolfinx.fem — DOLFINx 0.4.2.0 documentation

1 Like