[ANN] FLORIDyn.jl initial release v0.5.1

Package Announcements
1

We want to announce the initial release of a Julia version of FLORIDyn.jl, a dynamic wind farm simulation software originally written by Marcus Becker in Matlab. This software allows to:

  • Simulate wind farms dynamically at a low computational cost
  • Estimate the power generated, added turbulence, and wake-induced losses.
  • Apply heterogeneous and time-varying wind speeds and directions
  • Test different modeling approaches

It is very fast, between 6 and 33 times faster than the original Matlab version, and very well documented. A wind park with 54 turbines can be simulated at a speed of more than 1000 times real-time, using only one core of an AMD 7950X CPU. It can be used as foundation for the development of novel wind-farm control algorithms.

Outlook:

  • add the functions, not yet translated from Matlab; this is in particular the ensemble Kalman filter for the estimation spatial distribution of wind speed and turbulence
  • further improve the performance
  • add a turbine data base
  • add different type of wind farm controllers for reducing the wake effects and/or shifting the power production in time
  • we consider the coupling with mesoscale weather models

And my first package where I included videos in the documentation. :grinning_face_with_smiling_eyes:

29 Likes
2

Thanks for sharing. Good to know it exist, even if most of the times I’m just work “downstream” of such model, simply using wind power time series.

The screenshot in the readme and the animation in the doc is very useful to understand the feature scope of the library. In this sense, I would suggest to complement the description by adding the “spatial” and/or “wind field” keywords. By the way, is it a 2D or 3D wind field simulation?

3

It calculates the 3D field, so you can plot a 2D view, default is at the hub height. But the simulation is neither 2D nor 3D, only the values at the observation points are simulated (the white dots behind the turbines in the video). Everything else is interpolated. Wind shear is handled correctly.

I am currently using it to calculate the wake losses as function of wind direction and turbulence intensity. But what it makes unique is the capability to create time series data and to minimize the wake losses for example by yawing the turbines.