This project is to detect the location of an obstacle responsible for a turbulent flow in the context of tidal turbine.
The code expected as input the koopman mode decomposition of a flow, the following code was used: https://github.com/SkirOwen/ResDMDpy
The code was created using python=3.11, it should support 3.8 but has not been tested, nor it is guaranteed
it will stay compatible.
We recommend using conda or mamba to install
the dependencies if the GPU support is wanted. otherwise pip should suffice.
conda env create -f environment_gpu.ymlconda env create -f environment_no_gpu.ymlYou may want to install it in a virtual environment.
pip install -r requirement.txtNot all the code is accessible through the CLI.
The -m is to specify the model, and the -w is for the
number of window per mode.
Note that the parameter passed as an example would be the default one
if nothing is passed, i.e. not passing -w to a RF would set the -w to
2000 internally.
python -m koolnet -m rf -w 2000python -m koolnet -m xgboost -w 2000python -m koolnet -m cnn -w 4000