Expert Parameters

ultraFluidX 2021.1 features additional expert parameters, that can be used to fine-tune the simulations.

When these parameters are not specified, established default values are applied. Moreover, the expert parameters then do not show up in the effective solver deck.

When you specify values for the expert parameters, they are used instead of the default ones, and are also included in the effective solver deck that is written to disk by ultraFluidX.
<simulation> - <general> - <num_ramp_up_iterations>
By default, the mean velocity at the inlet is increased from 0 to the requested value over a certain amount of physical time. Using this parameter, you can also explicitly specify a fixed number of coarsest iterations for the ramp up.
<simulation> - <general> - <convergence_detection>
Activates the ultraFluidX convergence detection algorithm output (still in beta mode). The algorithm analyzes the transient drag and lift signals and estimates if and when the initial transient phase has ended and when the overall simulation has converged; the respective iteration steps are reported in the uFX_summary.txt file, together with averaged force coefficients that are determined between these two points in time
<simulation> - <geometry> - <triangle_mesh_export>
If this parameter is set to true, the triangulated surface used in the code (after triangle_splitting, if applicable) is exported to uFX_meshData/uFX_surfaceMesh.stl.
This surface mesh contains all "physical” surfaces including passive parts, while surfaces needed for the mesh setup, for example, those defining sources or custom zones, are not displayed.
The default value for this parameter is false.
<simulation> - <wall_modeling>
<wall_model_intensity>
The wall model intensity parameter is used together with the <coupling> “one-way” option and determines the strength of the link between the velocity taken as input to the wall model and the information fed to the first near wall cell. For a value of 0.0, the wall model is effectively switched off and a standard no-slip boundary condition is applied.
<donor_interpolation>
The donor position is the point where the wall model is coupled to the flow field, in a sufficiently well resolved part of the boundary layer. This donor position does not necessarily coincide with a voxel center; the information required at this position will be interpolated/derived from the data available in the 27 voxels (3x3x3) around the donor position.
<donor_position_factor>
This factor is applied to a reference length (see below) to determine the exact distance to the wall where the wall model is extracting information from the flow. The default value of 3.5 together with the default reference length corresponding to the near wall voxel size leads to a setup where the donor position is placed in the third layer of voxels above the wall.
<donor_position_use_local_reference_length>
The reference length is coupled to the voxel size at the wall. This parameter controls if a local reference length per refinement level or a global reference length is used.
If set to true (default), the reference length is evaluated per refinement level, which moves the donor position closer to the wall in regions of finer grid.
If set to false, in case the wall touches voxels of different refinement levels, the coarsest voxel size at the wall will be used as a reference length, also in regions of finer grid.
<donor_position_reference_length>
Using this parameter, you can also set the reference length to a specific value in [ m ] , which should, however, be of the order of magnitude of the near wall voxel size. If set to 0.0 (default), the parameter is effectively switched off.
Note: This parameter is only effective in the case that you request a global reference length at the same time (see above). Otherwise, it will be ignored.
<meshing> <general> <seeding_point>
<x_pos>, <y_pos>, <z_pos>
The seeding point is the position in the domain from with the flow region is filled with “fluid” voxels. By default, this position is placed at one of the corners between the inlet plane and the ground.
If the geometric setup is such that a solid is placed at this position, you can indicate another arbitrary point (inside the fluid region) from which the process of filling the fluid region with voxels should start.