Skip to content

make w2a initialization more accurate, especially for low resolutions #1669

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Jul 21, 2025
Merged

make w2a initialization more accurate, especially for low resolutions #1669

Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
9b6e089
make w2a initialization more accurate, especially for low resolutions
mbkuhn Jun 20, 2025
c080e94
formatting
mbkuhn Jun 23, 2025
00f9db4
Merge branch 'main' into w2a_improve_multiphase_velocity
mbkuhn Jul 18, 2025
4058e2e
avoid dividing by 0, just in case
mbkuhn Jul 21, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
57 changes: 48 additions & 9 deletions amr-wind/ocean_waves/relaxation_zones/waves2amr_ops.H
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -652,6 +652,16 @@ struct InitDataOp<W2AWaves>
const bool has_beach = wdata.has_beach && multiphase_mode;
const bool init_wave_field = wdata.init_wave_field || !multiphase_mode;

amrex::MultiFab phi_tmp_fab(
velocity(level).boxArray(), velocity(level).DistributionMap(), 1,
3);
amrex::MultiFab vel_tmp_fab(
velocity(level).boxArray(), velocity(level).DistributionMap(),
AMREX_SPACEDIM, 3);

const auto& phi_tmp = phi_tmp_fab.arrays();
const auto& vel_tmp = vel_tmp_fab.arrays();

amrex::ParallelFor(
velocity(level), amrex::IntVect(3),
[=] AMREX_GPU_DEVICE(int nbx, int i, int j, int k) noexcept {
Expand All @@ -673,17 +683,13 @@ struct InitDataOp<W2AWaves>
x, problo[0], gen_length, probhi[0], beach_length, wave_sol,
bulk, outlet);

const amrex::Real phi = local_profile[3] - z;
const amrex::Real cell_length_2D =
std::sqrt(dx[0] * dx[0] + dx[2] * dx[2]);
if (phi + cell_length_2D >= 0) {
vel[nbx](i, j, k, 0) = local_profile[0];
vel[nbx](i, j, k, 1) = local_profile[1];
vel[nbx](i, j, k, 2) = local_profile[2];
}
phi_tmp[nbx](i, j, k) = local_profile[3] - z;
vel_tmp[nbx](i, j, k, 0) = local_profile[0];
vel_tmp[nbx](i, j, k, 1) = local_profile[1];
vel_tmp[nbx](i, j, k, 2) = local_profile[2];

if (multiphase_mode) {
phi_arrs[nbx](i, j, k) = phi;
phi_arrs[nbx](i, j, k) = phi_tmp[nbx](i, j, k);
}

// Default w2a values matter for where no updates happen
Expand All @@ -693,6 +699,39 @@ struct InitDataOp<W2AWaves>
w2a_vel[nbx](i, j, k, 2) = quiescent[0];
});
amrex::Gpu::streamSynchronize();

amrex::ParallelFor(
velocity(level), amrex::IntVect(2),
[=] AMREX_GPU_DEVICE(int nbx, int i, int j, int k) noexcept {
const amrex::Real eps = 2. * std::cbrt(dx[0] * dx[1] * dx[2]);
const amrex::Real vof_local =
multiphase::levelset_to_vof(i, j, k, eps, phi_tmp[nbx]);

if (vof_local > 1. - constants::TIGHT_TOL) {
// Entire cell is liquid
vel[nbx](i, j, k, 0) = vel_tmp[nbx](i, j, k, 0);
vel[nbx](i, j, k, 1) = vel_tmp[nbx](i, j, k, 1);
vel[nbx](i, j, k, 2) = vel_tmp[nbx](i, j, k, 2);
} else if (vof_local > 1e-3) {
// Velocity of liquid becomes velocity of entire cell
// Interpolate using cell-centered values to liquid centroid
// Consider only z-direction to find centroid
const amrex::Real z_cell = problo[2] + (k + 0.5) * dx[2];
const amrex::Real z_cell_below = z_cell - dx[2];
const amrex::Real z_cell_bottom = z_cell - 0.5 * dx[2];
const amrex::Real z_liq_center =
z_cell_bottom + 0.5 * vof_local * dx[2];
for (int n = 0; n < AMREX_SPACEDIM; ++n) {
vel[nbx](i, j, k, n) =
vel_tmp[nbx](i, j, k - 1, n) +
(vel_tmp[nbx](i, j, k, n) -
vel_tmp[nbx](i, j, k - 1, n)) *
((z_liq_center - z_cell_below) /
(z_cell - z_cell_below + constants::EPS));
}
}
});
amrex::Gpu::streamSynchronize();
#else
amrex::ignore_unused(data, level, geom, multiphase_mode);
#endif
Expand Down
Loading