Add linear wind condition upsampling method to WindRose #544

Bartdoekemeijer · 2022-11-28T09:02:41Z

Feature or improvement description
This adds a function to the windrose class that allows me to linearly interpolate over the current wind rose dataframe for upsampling. This is useful when, for example, the wind rose is calculated in steps of 5.0 deg yet I am evaluating FLORIS at a resolution of 1.0 deg.

Related issue, if one exists
N/A.

Impacted areas of the software
floris.tools.wind_rose

Additional supporting information
I wonder if this modification fits the way wind_rose was designed. I have been using NearestNDInterpolant and LinearNDInterpolant when working with wind roses and FLORIS to ensure the dimensionality is matched appropriately. Note that wind_rose already has internal_resample and resample functions to take care of these two questions, respectively:

What is the resolution at which the wind rose is calculated from the raw timeseries data (bin edges)?
What is the resolution to which the wind rose is upsampled through linear interpolation?

However, the resample_wind_speed and resample_wind_direction both require a wind rose df as an input and thereby require additional manual labor before one can insert it directly into FLORIS, e.g.:

    wr = wind_rose.WindRose()
    df_wr = wr.make_wind_rose_from_user_data(...)
    df_wr = wr.resample_wind_speed(df_wr, ...)
    df_wr = wr.resample_wind_direction(df_wr, ...)
    freq_floris = np.reshape(df_wr["freq_val"], ...)  # Convert to (n_wind_speeds, n_wind_directions) shape

With the new functionality:

    wr = wind_rose.WindRose()
    wr.make_wind_rose_from_user_data(...)
    freq_floris = wr.interpolate(floris_wind_direction_grid, floris_wind_speed_grid)

Also, the examples with AEP calculations should probably be updated to use the WindRose class.

Finally, perhaps some of the resample functions in WindRose can benefit from the interpolate function to simplify upsampling/resampling.

Test results, if applicable

paulf81 · 2023-02-01T21:23:10Z

@Bartdoekemeijer wanted to check in on this PR, do you think this one is ready to merge if we resolve the above conflict?

Bartdoekemeijer · 2023-02-03T08:42:39Z

@Bartdoekemeijer wanted to check in on this PR, do you think this one is ready to merge if we resolve the above conflict?

Yes, 100% ready to merge!

paulf81 · 2023-02-06T23:18:25Z

hi @Bartdoekemeijer , I just caught this pull request up to the new develop branch with new linting requirements, it should not have changed anything functional but would you mind taking a quick check that this all still looks good to merge?

* Add LinearNDInterpolant to wind_rose * Change from returning interpolant to immediately interpolating to values * Add example to docstring --------- Co-authored-by: Paul <paul.fleming@nrel.gov> Co-authored-by: Rafael M Mudafort <rafmudaf@gmail.com>

* Speed up angle wrap functions and improve API (#505) * replace wrap functions with faster implementations * slight change to unit test as 180 -> -180 * Update doc string to reflect accurate return type * Use executable book project for documentation * Add a gh-pages deploy action * Add more content * Add a logo * Add a brief description of each example List examples in numeric order Small cleanup to Examples listing * Add initial API documentation * Add input file reference * Use autodoc and autosummary instead of autoapi Remove TOC from intro page * Locally install the current repo * Configure autodocs better Include undocumented objects as well as all members of an object Add all default autodoc options to config * Link to API docs in references to Python objects * Git-ignore docs build * Add instructions for building documentation And other minor changes * Add Google Analytics tracker * Update README with documentation links * List floris packages in API docs * Enforce building docs from the develop branch * Add a method to show mpl plots This means users don’t have to also import matplotlib.pyplot as plt only to run plt.show() * Update documentation references list (#519) * initial entry for cumulative curl * copy in v2 references * add cc references * add cc refs * replace references to zrefs with references.bib * API improvements * Add a subtract overload to CutPlane Also replaces another method CutPlane.subtract to do something similar. That one made assumptions about the order and expected operations. The new subtract method expects all preprocessing to be down outside of the class. * Fix incorrect yaw angles in fi.calculate_wake In subsequent function calls, the yaw angles are not reset when they are all zeros. This means that if initially the yaw angles are non-zero and then they’re all-zero, the non-zero yaw settings are retained. * Remove requirement to use codecov token to upload reports (#532) * remove potentially error causing line * Update codecov upload action to v3 Co-authored-by: Rafael M Mudafort <rafael.mudafort@nrel.gov> * Update API in examples And clean up examples imports * Use standard conventions in argument case minSpeed, maxSpeed => min_speed, max_speed * Add function to label turbines on CutPlane plots * Rotate turbine lines and ID annotations with wind * Configure isort and apply rules (#535) * Configure isort and apply to floris.simulation * Apply isort to floris.tools * Apply isort to shared modules * Apple isort to examples and tests * Increase line length to 100 * Add isort description in docs * Add weights to AEP wind rose function (#541) * Add weights to get_farm_AEP_wind_rose_class * updating layout passed to floris_interface in example * Remove typecast to int for turbine weighting --------- Co-authored-by: bayc <christopher.j.bay@gmail.com> Co-authored-by: Rafael M Mudafort <rafael.mudafort@nrel.gov> * Add parallel computing interface for farm simulation and yaw optimization (#555) * Add parallel computing class * Simplify and lighten the way floris information is passed to single-core functions * Remove unc_options and only pass unc_pmfs internally: all information already contained in that variable. * Update how wind directions are split up * Add example for parallel computing * Introduce separate variable for number of workers * Expand PCI capabilities with reinitialize, copy. This is needed for some FLASC functionalities like precalculating table of floris solutions * Allow user to reinitialize atmospheric conditions but then also reload how atmospheric conditions are divided by cores * Add toggle to silence print statements from Serial Refine and set default in parallel computing. Namely, with 16 workers, you will see 16 times these print statements. This only becomes more with more workers and quickly clutters up the log file. * Add the option to propagate flow field information from workers to main florisinterface object * Add parallel processing example * Remove duplicate example listing --------- Co-authored-by: Rafael M Mudafort <rafael.mudafort@nrel.gov> * Feature/update layout vis (#496) * Update layout visualization functions * Provide an example of layout visualization * list 23_visaluze_layout in examples.md --------- Co-authored-by: Rafael M Mudafort <rafmudaf@gmail.com> * Add ruff configuration * Fix all E and F codes F are pyflakes E are pycodestyle https://github.com/charliermarsh/ruff#pyflakes-f Note: line length errors are not excluded here * Line length corrections - floris.simulation Max line length setting is 100 * Line length corrections - floris.tools * Line length corrections - floris/ * Line length corrections - tests * Line length corrections - examples * Line length corrections - profiling * Fix W codes Here, it’s only adding a blank line at the end * Fix C4-comprehensions codes * Fix I codes Import formatting. Ruff is not yet at parity with isort, but it is pretty close. * Include Ruff in CI Use pre-commit with Actions Also update the developer dependencies * Add Ruff usage documentation in developer guide * Enable and fix pre-commit formatting hooks (#567) * Enable and fix pre-commit formatting hooks * Add pre-commit instructions to developer documentation * Support heterogeneous wind direction definition via FlorisInterface.reinitialize (#453) * Provide ability to pass het_map to reinitialize * re-order reinitialize inputs * Add a new example * Add an error if het map doesnt match on dimensions to wd * Add error example to example * updating format of example * Add indication final error is on purpose * line length check * Comment out forced error * Add example 16b to docs * use n directions instead of x * Add comment --------- Co-authored-by: Paul <paul.fleming@nrel.gov> Co-authored-by: bayc <christopher.j.bay@gmail.com> * Infrastructure improvements (#569) * Preserve isort formatting - run it last Also enables splitting imports on trailing comma. If a multi-line import statement is short and includes a trailing comma, it will not be refactored into a single line. * Use token for codecov upload * Reduce computational load in an example Initially, this example slowed the Actions workflow by 2x. This commit reduces the amount of work in the example to speed up the Actions workflow. * Restrict PyPI deploy workflow to NREL/floris * Run tests regardless of linting results * Fail the linter * Revert "Fail the linter" This reverts commit d7bfa62. * Add virtual turbine plotting method (#418) * Add func calculate_horizontal_plane_with_turbines * Add demo example * Correct sort order * Improve resolution and force matching range * fix append * Remove calculate with turbine from floris interfa * Delete turbine scan example * Add calc with turbines to visualizations.py * Add vis with turbine to visualization example * Support for yaw angles * Fix handling of yaw angles * fix outdated layout keyword * ruff linting * test pre-commit * test pre-commit * Pre-commit fixes * Fixed doc errors * Clean out commented code * Preserve isort formatting - run it last Also enables splitting imports on trailing comma. If a multi-line import statement is short and includes a trailing comma, it will not be refactored into a single line. * Use token for codecov upload * Reduce computational load in an example Initially, this example slowed the Actions workflow by 2x. This commit reduces the amount of work in the example to speed up the Actions workflow. * Restrict PyPI deploy workflow to NREL/floris * Add info on plotting method * Use coarse mesh for turbine plotting --------- Co-authored-by: Rafael M Mudafort <rafmudaf@gmail.com> * Add linear wind condition upsampling method to WindRose (#544) * Add LinearNDInterpolant to wind_rose * Change from returning interpolant to immediately interpolating to values * Add example to docstring --------- Co-authored-by: Paul <paul.fleming@nrel.gov> Co-authored-by: Rafael M Mudafort <rafmudaf@gmail.com> * Improve the documentation navigation menu and organization (#570) * minor formatting updates * start overhaul of docuemntation structure * convert input reference to markdown for consistency * add a serious placeholder for theory and update the example input file * remove my notes * add top-level workflow description for getting started * add missing period * Add intro notebook * Add developer section to menu * Bug fix: fix toc reference to placeholder page * Warning fix: use alternate cross reference form --------- Co-authored-by: Rafael M Mudafort <rafmudaf@gmail.com> * Feature: External Turbine Library (#568) * add the external library functionality * update tests for new functionality * enable future typing * fix extra white space * remove extra parameters * add missing test file * fix name bug and address #513 * get tests passing and ensure dictionary entries still work * address Raf's comment * add _path to turbine library references * update typo, FI, and tests for _path * fix error on case of using turbine dictionary and missing air density * add additional test for the unique turbine check * Specify coverage file in Codecov upload action * Add a link to Discussions from Issues * Add helpful info to GitHub Issue and PR templates (#597) --------- Co-authored-by: paulf81 <paul.fleming@nrel.gov> Co-authored-by: Rob Hammond <13874373+RHammond2@users.noreply.github.com> Co-authored-by: bayc <christopher.j.bay@gmail.com> Co-authored-by: Bart Doekemeijer <bart.doekemeijer@shell.com> Co-authored-by: pjireland <pireland@windesco.com>

Bartdoekemeijer added 2 commits November 28, 2022 09:25

Add LinearNDInterpolant to wind_rose

d8d041c

Change from returning interpolant to immediately interpolating to values

15b9227

Bartdoekemeijer changed the title ~~[Feature] add to_linear_interpolant to wind_rose class~~ [Feature] add interpolate(...) to wind_rose class Nov 28, 2022

rafmudaf added enhancement An improvement of an existing feature floris.tools labels Feb 1, 2023

merge develop

1a6cc0a

rafmudaf changed the title ~~[Feature] add interpolate(...) to wind_rose class~~ Add linear wind condition upsampling method to WindRose Feb 7, 2023

Add example to docstring

7272832

rafmudaf self-requested a review February 7, 2023 20:49

rafmudaf approved these changes Feb 7, 2023

View reviewed changes

rafmudaf self-assigned this Feb 7, 2023

rafmudaf merged commit c99f75a into NREL:develop Feb 7, 2023

Bartdoekemeijer deleted the feature/wind_rose_interpolant branch February 8, 2023 08:59

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add linear wind condition upsampling method to WindRose #544

Add linear wind condition upsampling method to WindRose #544

Bartdoekemeijer commented Nov 28, 2022 •

edited by rafmudaf

Loading

paulf81 commented Feb 1, 2023

Bartdoekemeijer commented Feb 3, 2023

paulf81 commented Feb 6, 2023

Add linear wind condition upsampling method to WindRose #544

Add linear wind condition upsampling method to WindRose #544

Conversation

Bartdoekemeijer commented Nov 28, 2022 • edited by rafmudaf Loading

paulf81 commented Feb 1, 2023

Bartdoekemeijer commented Feb 3, 2023

paulf81 commented Feb 6, 2023

Bartdoekemeijer commented Nov 28, 2022 •

edited by rafmudaf

Loading