Saving output at points selected by user when running the IPCS solver

dolfinx
,
1

I’m running the code in jorgensd/dolfinx_ipcs: Simple IPCS solver for dolfinx.

I ran the code as follows:

  • download the docker and git clone the repository;
  • run python3 create_and_convert_2D_mesh.py in command line;
  • run python3 DFG_benchmark.py in command line.

And I make sure I can save output successfully. I changed the DFG_benchmark.py as follows:

def IPCS(outdir: Path, dim: int, degree_u: int,
         jit_options: dict = {"cffi_extra_compile_args": ["-Ofast", "-march=native"], "cffi_libraries": ["m"]}):
    ......
    # Solve problem
    N = int(T / dt)
    if has_tqdm:
        time_range = tqdm(range(N))
    else:
        time_range = range(N)
    
    # Save Velocity and Time.
    velocity = []
    U = uh.x.array[0::2].reshape((-1,1))
    V = uh.x.array[1::2].reshape((-1,1))
    UV = np.hstack((U, V))
    velocity.append(UV)
    time_step = []
    time_step.append(t)

    for i in time_range:
        ......
        with common.Timer("~IO"):
            if i % 5 == 4:
                # print('Current time: ', t)
                # print('Current Iteration: ', i)
                U = uh.x.array[0::2].reshape((-1,1))
                V = uh.x.array[1::2].reshape((-1,1))
                UV = np.hstack((U, V))
                velocity.append(UV)
                time_step.append(t)

    savemat('./DFG-2D3.mat', \
            {'U_star': np.array(velocity), 't': np.array(time_step).reshape((-1,1))}, \
            do_compression=True)
    print('matrix file saved :D')

The above code saves the output at the mesh nodes, which is not free.

Now I want to save the output at points selected by user, such as (0,0), (0.1,0.1), …, how can I achieve this?

2

See for instance the “microphone” implementation in: Implementation — FEniCSx tutorial

3

Thanks for replying.
Now I know the function eval() can help me to evaluate Function at points x, but the eval() has two parameters:

def eval(self, x: npt.ArrayLike, cells: npt.ArrayLike, u=None) -> np.ndarray:
        """Evaluate Function at points x, where x has shape (num_points, 3),
        and cells has shape (num_points,) and cell[i] is the index of the
        cell containing point x[i]. If the cell index is negative the
        point is ignored."""

I have two questions:

  • My example is 2-dimensional, but x has shape (num_points, 3), how can I solve this? Can I set the third number = 0?
  • How can I get the cell for my all points? Can I set cells=np.ones((num_points, ))?

I solved the above questions!!! Just follow the tutorial: Implementation — FEniCSx tutorial