foxes-opt

The package foxes-opt provides optimization functionality for the Farm Optimization and eXtended yield Evaluation Software foxes and is based on the optimization interface iwopy.

All three open-source Python packages foxes, foxes-opt and iwopy are provided and maintained by Fraunhofer IWES.

The calculation is fully vectorized and its fast performance is owed to dask. Also the parallelization on local or remote clusters is enabled via dask. The wind farm optimization capabilities invoke the iwopy package which as well supports vectorization.

foxes is build upon many years of experience with wake model code development at IWES, starting with the C++ based in-house code flapFOAM (2011-2019) and the Python based direct predecessor flappy (2019-2022).

Documentation: https://fraunhoferiwes.github.io/foxes-opt/

Source code: https://github.com/FraunhoferIWES/foxes-opt

PyPi reference: https://pypi.org/project/foxes-opt/

Anaconda reference: https://anaconda.org/conda-forge/foxes-opt

Installation

There are multiple ways to install foxes-opt.

Installation as standard user

pip install foxes[opt]

or

pip install foxes-opt

or

conda install foxes-opt -c conda-forge

Installation as developer

As a developer, first clone both repositories, and then install via pip using the -e flag:

git clone https://github.com/FraunhoferIWES/foxes.git
pip install -e foxes

git clone https://github.com/FraunhoferIWES/foxes-opt.git
pip install -e foxes-opt

If you want to contribute your developments, please replace the above repository locations by your personal forks.

Wind Rose Layout Snapshots

Pass --write_layouts to examples/layout_wind_rose/run_pymoo.py or examples/layout_wind_rose/run_slsqp.py to enable WriteLayoutCallback; intermediate output is disabled by default. The callback writes the best first-objective individual of each completed optimizer step to examples/layout_wind_rose/results/, regardless of the working directory. The examples disable CSV output and write one image per generation or accepted iteration. Use --layout_image_type to select its file type; the default is jpg. Each plot title shows the selected layout's first objective name and value. Turbines associated with violated constraints are red; valid turbines are orange, and each image includes an upper-left legend for these colors. The examples write these snapshots without per-file log messages. Early layouts may be infeasible; these are intermediate population snapshots, not final results. The callback options write_csv and write_image independently disable either format; both default to True. The callback accepts both iteration and evaluation events, so it can be used with all callback-enabled iwopy optimizers. The SLSQP and IPOPT examples use LocalFD for finite-difference gradients. The IPOPT example runs through pygmo and does not offer callbacks because pygmo exposes neither exact live IPOPT iterations nor their decision vectors. Running the example again overwrites matching generation numbers. The existing results/ rule in .gitignore excludes these generated files.

Citation

Please cite the JOSS paper FOXES: Farm Optimization and eXtended yield Evaluation Software.

Bibtex:

@article{
    Schmidt2023,
    author = {Jonas Schmidt and Lukas Vollmer and Martin Dörenkämper and Bernhard Stoevesandt},
    title = {FOXES: Farm Optimization and eXtended yield Evaluation Software},
    doi = {10.21105/joss.05464},
    url = {https://doi.org/10.21105/joss.05464},
    year = {2023},
    publisher = {The Open Journal},
    volume = {8},
    number = {86},
    pages = {5464},
    journal = {Journal of Open Source Software}
}