Inflow specified from file to enable turbulent inflow#84
Conversation
…anslated point, allow multiple tilt angles
…dentification, change marking routines
…n and how array inputs are unrolled (x-major order)
…k all tables are laid out and tilted correctly
| save_text_interval: .1 #0.01 | ||
| save_xdmf_interval: .1 #0.01 | ||
| fluid: | ||
| velocity_profile_type: uniform |
There was a problem hiding this comment.
Adding for transparency so that user knows what kind of inflow is being applied in this example.
| description: "If true, use the inflow boundary condition to interpolate a non-zero initial velocity throughout the domain (even if that does not respect interior boundary conditions), if false, initial velocity is zero." | ||
| time_varying_inflow_window: | ||
| description: "If true, use the inflow boundary condition to interpolate a non-zero initial velocity throughout the domain (even if that does not respect interior boundary conditions), if false, initial velocity is zero. If `velocity_profile_type` is `specified_from_file`, the velocity at each point is a time-averaged value of that gripoint's velocity data from the .h5 file." | ||
| ramp_window: |
There was a problem hiding this comment.
Changing this variable name to ramp_window since time_varying_inflow_window is now confusing with the addition of the specified_from_file option that is also a time-varying inflow
There was a problem hiding this comment.
Pretty sure this file isn't used, but added special treatment of u_ref for specified_from_file mode just in case
|
Hmm, the tests are failing at the build conda environment step. They all pass locally, so I'm not sure what's going on. |
got it, it seems to hang on #83 too in the same spot (so probably not related to h5py), will try to fix it there. |
eyoung55
left a comment
eyoung55
left a comment
There was a problem hiding this comment.
Looks great, thank you for your patience! Tests are fixed and this is ready to merge.
2 participants
This pull request enables the user to specify a file that defines the time-varying inflow boundary conditions. The primary purpose of this feature is to run PVade with turbulent inflow, but the functionality is general enough that users can specify any inflow data externally and then feed it in as inflow boundary conditions.
Note: You'll need to install
h5pyin your conda/mamba environment to be able to run this.In this example, I use
pyconturbto generate turbulent inflow data of shape (nt, ny, nz) where nt = number of timesteps, ny = number of gridpoints in the spanwise direction, and nz = number of gridpoints in the vertical direction. At each timestep,pvadeselects the y-z slice of turbulent inflow for that timestep and applies it as inflow BCs.This functionality only works for 3D simulations at the moment.
To do:
ramp_windowname and reconcile withtime_varying_inflowinput optionpyconturbdirectly in the code -> Issue Integrate pyconturb directly into PVade #86 has been created