Optionally Map To Physical Space

Similarity profiles are naturally functions of \(\eta\). For comparison against CFD, experiment, or geometry-based data, map the profile to the physical wall-normal coordinate \(y\).

The inverse transform uses the temperature profile \(\tau = T/T_e\) and local edge scales.

Physical-space mapping is optional. If the downstream task only needs similarity profiles, the eta-space solution can be used directly.

Map Onto A Structured Grid

Write the similarity solution as JSON, then generate the baseflow mapping configuration:

1
2

simbl solve simbl_config.toml
simbl baseflow init

Edit simbl_baseflow.toml to identify the solution, edge state, input grid, and output file:

 1
 2
 3
 4
 5
 6
 7
 8
 9
10

[baseflow]
solution_input = "simbl.json"
edge_state_input = "edge_state.json"
grid_input = "grid.hdf5"
output = "baseflow.hdf5"
similarity_transform = "levy_lees"
geometry_transform = "none"
farfield_fill = "edge"
endpoint_tolerance = 1.0e-3
dtype = "f8"

All artifact paths are resolved relative to the configuration file. Run the mapping with:

1

simbl baseflow run simbl_baseflow.toml

The command preserves the input coordinates and writes uvel, vvel, wvel, temp, pres, dens, and visc fields through cfd-io. Cubic splines map the solution inside its eta domain. Grid points beyond the mapped eta_max receive the exact freestream values from the edge FlowState rather than extrapolated spline values.

Farfield filling is allowed only when the similarity endpoint satisfies the configured convergence tolerance for fp, tau, fpp, and taup. The current workflow supports planar Falkner-Skan profiles. Mangler geometry mapping and Falkner-Skan-Cooke crossflow profiles are rejected explicitly until those transformations are implemented.

See the Similarity to Physical Coordinate Transform theory page for definitions.