Improve Robustness of Immersed Forcing Wall Function and Method

[hgopalan]

Summary

In this PR, we make several changes to the immersed forcing method to improve robustness and also improve the accuracy of the wall function.

1. The terrain wall-function has been modified to provide wall function capability consistent with the flat terrain.
2. The hybrid version of the Kosovic model has also been modified to be terrain-aware.
3. The restart runs with a user-specified CFL assumes a flat bottom to compute the time step. However, when a terrain is added the initial steps at the terrain interface requires a smaller time step to stabilize. A delay time is added to ensure that for restart runs, the simulation does not start with CFl based dt but instead starts with an initial dt. Once the delay time elapses, the simulation switches back to CFL based dt or a fixed dt, depending on user preference.

The time step of immersed forcing simulations are much higher now when we have multiple refinement levels. It is also possible to have terrain touch the outer x and y boundaries without worrying about time step reduction.

Please check the type of change introduced:

• [X ] Bugfix
• Feature
• Code style update (formatting, renaming)
• Refactoring (no functional changes, no api changes)
• Build related changes
• Documentation content changes
• Other (please describe):

Additional background

The existing regression tests and unit tests will be sufficient.

Issue Number:

[marchdf]

do you have a test you want to add to this? Also there are a lot of tests that still use Kosovic.refMOL, going to need to update those.

[hgopalan]

do you have a test you want to add to this? Also there are a lot of tests that still use Kosovic.refMOL, going to need to update those.

This works with the existing tests. We just modified how the wall function got evaluated. I will update the refMOL part.

[marchdf]

I see diffs in the test but nothing I would not expect from these changes.

[marchdf]

Just for posterity, these are the diffs I see:

         35 - abl_waves_terrain (Failed)                        no_ci regression
         36 - act_abl_joukowskydisk (Failed)                    no_ci regression
         37 - act_abl_uniformctdisk (Failed)                    no_ci regression
        101 - forest_drag (Failed)                              no_ci regression
        122 - abl_kosovic_neutral (Failed)                      no_ci regression
        124 - abl_kosovic_neutral_ib (Failed)                   no_ci regression
        125 - nrel_precursor (Failed)                           no_ci regression
        161 - nrel_terrain (Not Run)                            no_ci regression
        162 - nrel_terrain_amr (Not Run)                        no_ci regression