Tuning of the model parameters

[EddyG93]

Hi everyone,

I am currently involved in a research project comparing the capabilities of different wake models implemented in FLORIS in terms of layout optimization. My reference layout will be the one of the Alpha Ventus Wind Farm and I have processed meteorological data from a neighbouring research platform (FINO1) available.
I am wondering if I can use the default model parameters or if I have to tune them. In case of tuning do you have any ideas on how to start as there is no power data available?
Any help would be greatly appreciated!

Eddy

[Bartdoekemeijer]

Hi Eddy,

Tuning of the model parameters in FLORIS remains a difficult and ambiguous task. We are currently working on this behind the scenes, but this is only its in preliminary stages. Previously, we have tuned FLORIS to high-fidelity simulation data (SOWFA) comparing flow fields and power productions. Also, in field experiments, we may tune certain parameters by hand to match the energy ratios better.

Here is some literature you may find useful:

• Section 2.3 in "Improving wind farm flow models by learning from operational data", Schreiber et al., 2020
  This paper describes both real-time model corrections and a baseline tuning of the model parameters.
• Section 3 in "Closed-loop model-based wind farm control using FLORIS under time-varying inflow conditions
  ", Doekemeijer et al., 2020
  Where I tune a MATLAB-equivalent of FLORIS to SOWFA data. The method is identical, but I will stress that parameters may have a different interpretation between the MATLAB and NRELs Python model. Do not interchangeably use these models.
• And here's a paper explaining energy ratios, which we may use for tuning selective parameters by hand, "Continued results from a field campaign of wake steering – Part 2", Fleming et al., 2020]

Generally, though, if you do not have any or only little data available, then I suggest you stick to the default model parameters. In our experience, They give a good match with SCADA data from offshore and flat-terrain onshore wind farms.

Bart

[bayc]

@mr483 The paper preprint can be found here: https://wes.copernicus.org/preprints/wes-2022-17/.

[mr483]

Sorry, I was actually referring to the paper that is cited in the latest CC paper, with the citation:

Doekemeijer, B. M., Simley, E., and Fleming, P. A.: Comparison of the Gaussian wind farm model with historical data of three offshore wind farms, Energies special Issue: “Fast-running Engineering Models of Wind Farm Flows”, submitted, in review, 2022.

[Bartdoekemeijer]

Hi @mr483 . The paper was formally accepted by the editors and I'm submitting the final proofread/approval today. The official paper should be publicly available not too long from now, hopefully this week.

[mr483]

Great, thank you!

[Bartdoekemeijer]

@mr483 you can find the article here: https://www.mdpi.com/1996-1073/15/6/1964.

[Bartdoekemeijer]

Hi Eddy. Please see my reply to #276 for some preliminary validation results of FLORIS with historical data. We expect to be able to share the scientific articles we have been preparing over the next couple months.