Note: Please check out Version 5 of this dataset since some labels have been corrected.

This dataset is published together with the paper "CARE to Compare: A real-world dataset for anomaly detection in wind turbine data", which explains the dataset in detail and defines the CARE-score that can be used to evaluated anomaly detection algorithms on this dataset. When referring to this dataset, please cite the paper mentioned in the related work section.

The data consists of 95 datasets, containing 89 years of SCADA time series distributed across 36 different wind turbines
from the three wind farms A, B and C. The number of features depends on the wind farm; Wind farm A has 86 features, wind farm B has 257 features and wind farm C has 957 features.

The overall dataset is balanced, as 44 out the 95 datasets contain a labeled anomaly event that leads up to a turbine fault and the other 51 datasets represent normal behavior. Additionally, the quality of training data is ensured by turbine-status-based labels for each data point and further information about some of the given turbine faults are included.

The data for Wind farm A is based on data from the EDP open data platform (https://www.edp.com/en/innovation/open-data/data),
and consists of 5 wind turbines of an onshore wind farm in Portugal.
It contains SCADA data and information derived by a given fault logbook which defines start timestamps for specified faults.
From this data 22 datasets were selected to be included in this data collection.
The other two wind farms are offshore wind farms located in Germany. All three datasets were anonymized due to confidentiality reasons for the wind farms B and C.
Each dataset is provided in form of a csv-file with columns defining the features and rows representing the data points of the time series. Files

More detailed information can be found in the included README-file and in the publication corresponding to this dataset.

Introduction

This dataset is issued from the third and final field campaign of the French national project SMARTEOLE. It consists in data from 7 wind turbines of a single wind farm (Sole du Moulin Vieux, located in France) for which Wind Farm Control field tests were performed to evaluate the performance of a wake steering strategy for improving the power production.

The wind farm consists of 7x Senvion MM82 wind turbines (rotor diameter of 82m, nominal power of 2.05 MW).

Description

The tests were realized between 17 February – 25 May 2020, with wake steering implemented on turbine SMV6. This dataset covers this full period, and it has been pre-processed to facilitate the analysis of the Wind Farm Control experiment. All timesteps when at least one turbine was stopped were removed, and SCADA nacelle position and wind direction signals have been corrected to remove any north alignment issues. Finally, the time resolution has been standardized at 1-min from the raw data recorded at higher frequencies from the different sensors. For more details about the development of the field campaign and the pre-processing steps followed in the data analysis, please consult the related publication : https://wes.copernicus.org/articles/6/1427/2021/wes-6-1427-2021.html. Some information can also be found in the related IEA task 44 wiki page.

The following files can be found in the dataset :

• SMARTEOLE_WakeSteering_SCADA_1minData.csv : the Supervisory Control and Data Acquisition (SCADA) data from the 7 turbines.
• SMARTEOLE_WakeSteering_ControlLog_1minData.csv : logs from the control system located on turbine SMV6, responsible for the application of the wake steering. The applied yaw offset on the turbine at each timestep can be found here.
• SMARTEOLE_WakeSteering_WindCube_1minData.csv : data from the ground based WindCube profiler lidar, located between SMV2 and SMV3. This can be used to assess the ambient environmental wind conditions at the farm.
• SMARTEOLE_WakeSteering_Coordinates_staticData.csv : file listing the coordinates of the wind turbines in the farm and WindCube location in traditional Latitude / Longitude system (WGS84) and XY metric system (French Lambert 93).
• SMARTEOLE_WakeSteering_Map.pdf : the map of the farm showing the location of wind turbines and WindCube. This is the exact same map as the one seen in the paper indicated above.
• SMARTEOLE_WakeSteering_NTF_SMV6_staticData.csv : the transfer function used in the paper to correct the wind speed measured by SMV6 to better match the freestream wind speed at 150m upstream (i.e. approximately 1.8 diameters), derived using WindCube nacelle lidar installed on top of the turbine.
• SMARTEOLE_WakeSteering_correction_factors_SMV1237_staticData.csv : the transfer function used in the paper to derive and correct the reference power and wind speed signals —defined as the mean values of the power and wind speeds from SMV1, SMV2, SMV3, and SMV7— to remove biases from the values at SMV6 as a function of wind direction and wind speed. These corrected reference signals are used for quantifying the impact of the wake steering.
• SMARTEOLE_WakeSteering_GuaranteedPowerCurve_staticData.csv : the warranted power and thrust curves for the standard mode (Mode 0) of the MM82 wind turbine.
• SMARTEOLE_WakeSteering_ReadMe.xlsx : read me file indicating for each dataset the signification of the different variables.

Unfortunately, the WindCube nacelle lidar data on top of SMV6 could not be shared, instead the transfer functions derived thanks to this sensor can be used to correct the SCADA channels. The Wind Energy Science publication describes how these transfer functions were obtained.

Acknowledgement

The creation of this dataset was realized in the scope of French national project SMARTEOLE, supported by the Agence Nationale de la Recherche (grant no. ANR-14-CE05-0034).

Furthermore, we would like to thank ENGIE Green for allowing us to make this dataset publicly available.

How to cite this dataset

When using this dataset in future research, please add the following sentence in the Ackowledgement section of your publication :

"The dataset used in this research has been obtained by ENGIE Green in the scope of French national project SMARTEOLE (grant no. ANR-14-CE05-0034)".

When citing the dataset in the core text of a paper, the reference to Simley et al. can simply be used.

Related datasets and publications

Several field test campaigns were realized in the scope of SMARTEOLE project. Although these data are not made publicly available by default, they can be shared in a per-project basis and under the protection of a dedicated NDA. Please refer to the following publications listed below to get an idea of the content of the different datasets.

SMARTEOLE Field Test 1

• Ahmad T. et al., Field Implementation and Trial of Coordinated Control of WIND Farms, IEEE Transactions on Sustainable Energy, 9(3), 2018, 10.1109/TSTE.2017.2774508.
• Duc T., Optimization of wind farm power production using innovative control strategies, Master’s thesis, DTU Wind Energy-M-0161, 2017.
• Duc T. et al., Local turbulence parameterization improves the Jensen wake model and its implementation for power optimization of an operating wind farm, Wind Energy Science, 4(2), 2019, 10.5194/wes-4-287-2019.
• Torres Garcia E. et al., Statistical characteristics of interacting wind turbine wakes from a 7-month LiDAR measurement campaign, Renewable Energy, 130, 2019, 10.1016/j.renene.2018.06.030.
• Hegazy A. et al., LiDAR and SCADA data processing for interacting wind turbine wakes with comparison to analytical wake models, Renewable Energy, 181, 2022, 10.1016/j.renene.2021.09.019.

SMARTEOLE Field Test 2

• Tagliatti F., Investigation of Wind Turbine Fatigue Loads under Wind Farm Control: Analysis of Field Measurements, Master’s thesis, DTU Wind Energy-M-0302, 2019.
• Göçmen T. et al., FarmConners wind farm flow control benchmark – Part 1: Blind test results, Wind Energy Science, 7(5), 2022, 10.5194/wes-7-1791-2022.

SMARTEOLE Field Test 3

• Simley E. et al., Results from a wake-steering experiment at a commercial wind plant: investigating the wind speed dependence of wake-steering performance, Wind Energy Science, 6(6) 2021, 10.5194/wes-6-1427-2021.

Release Notes

• v1.0 (2022-11-24) : first version of the dataset.

This document list a number of internal database with restricted access hosted by DTU Wind Energy. The databases are accessible with reference to NDA contract between the data provider and DTU Wind Energy.

IMPORTANT Access to these database are restricted to employees at DTU Wind Energy.

Multi-annual 10-minute time series raw SCADA data from a Vestas V52 wind turbine at Dundalk Institute of Technology, Ireland (GPS Co-ords: 53.98352, -6.391390). The period covers from the 30 January 2006 to 12 March 2020. The wind turbine is located in a peri-urban environment and operates as a behind-the-meter system. The wind turbine has a hub height of 60 m and a rotor diameter of 52 m. (It should be noted that a gearbox changeout took place in the period from 04 October 2018 to 27 July 2019 for which there is no positive electrical power output)

This data collection consists of two datasets from a micrometeorological experiment conducted in two distinct operating wind farms in a coastal area of the northeast region of Brazil, called Pedra do Sal Wind Farm (UEPS) and Beberibe Wind Farm (UEBB). These wind farms are located on the northeast coast of Brazil where meteorological conditions are strongly influenced by trade winds and sea breeze. Both datasets represent a full-year of measurements from August/2013 to July/2014.

On both operating wind farms it was commissioned a fully instrumented IEC-compliant 100m met mast, with five levels of first-class calibrated cup anemometers and one level (100m) with 3D sonic anemometer. Additionally at UEPS there's an extra 3D sonic at 20m height on the met mast, as well as a VAISALA LEOSPHERE Windcube8 doppler wind lidar with a range up to 500m height and located 2.5D upwind of one of the wind turbines.

The Pedra do Sal wind farm (UEPS) has an installed capacity of 18MW, with 20 Enercon E-44 installed at 55m a.g.l. At Beberibe wind farm (UEBB) there are 32 Enercon E-48 wind turbines installed at 75m a.g.l. The dataset includes 10min SCADA data for all wind turbines on both wind farms.

This dataset has a high-quality combination of meteorological, SCADA and turbulent flux data of two operating wind farms in Brazil. During a full-year of measurements both datasets had a high data recovery rate (see attached tables). The dataset has already been used to assess the impact of atmospheric stability on the wind farm performance, as well as the effect of mesoscale patterns on the wind profile and wind farm power production. Recirculation of the sea breeze and the development of an internal boundary layer upwind the wind turbines were also characterized.

For more details on the experimental layout, wind turbine locations, meso and microcale wind conditions and any other information not stated in the NetCDF4 files, please refer to the reference material or contact one of the authors.

This dataset contains:

• A kmz file for Penmanshiel wind farm in the UK (for opening in e.g. Google Earth)
• Static data including turbine coordinates and turbine details (rated power, rotor diameter, hub height, etc.)
• 10-minute SCADA and events data from the 14 Senvion MM82's at Penmanshiel wind farm, grouped by year from 2016 to mid-2021, which was extracted from our secondary SCADA system (Greenbyte). Note not all signals are available for the entire period, and there is no turbine WT03
• Data mappings from primary SCADA to csv signal names
• Site substation/PMU meter data where available for the same period
• Site fiscal/grid meter data where available for the same period

The dataset has been released by Cubico Sustainable Investments Ltd under a CC-BY-4.0 open data license and is provided as is. However, please provide any feedback you might have on the dataset and format of the data. I'll try and add or link to additional file formats that might be easier to work with (e.g. for use with specific analysis software), and update this dataset periodically (e.g. twice a year), but please prompt me as required.

Feel free to use the data according to the license, however, it would be helpful to me if you could let me know where, how and why you are using the data, so that I can highlight this to the business (and renewables industry) and hopefully promote similar data sharing initiatives. I am particularly interested in performance analysis/improvement opportunities, how the dataset can be augmented with other (open) datasets, and sharing more generally within the renewables industry.

If you would like to get access to other datasets we may hold (e.g. more recent data, data from our other sites, ~30s resolution data, etc.), please let me know, and, if you have any questions or want to discuss open data and this or other initiatives, please contact me and I will endeavour to help.

I would like to thank Cubico's Senior Legal Advisor & Compliance Officer, IT Director, UK Asset Management Team, Executive Committee and my manager for supporting this initiative, as well as our partners GLIL for agreeing to release this data under an open license. I would also like to thank those I have talked to during the process of releasing this data under an open license and the encouragement and advice I have had on the way.

For contact my email address is charlie.plumley@cubicoinvest.com.

You can also access data from Kelmarsh wind farm here.

This dataset contains:

A kmz file for Kelmarsh wind farm in the UK (for opening in e.g. Google Earth)

Static data including turbine coordinates and turbine details (rated power, rotor diameter, hub height, etc.)

10-minute SCADA and events data from the 6 Senvion MM92's at Kelmarsh wind farm, grouped by year from 2016 to mid-2021, which was extracted from our secondary SCADA system (Greenbyte). Note not all signals are available for the entire period

Data mappings from primary SCADA to csv signal names

Site substation/PMU meter data where available for the same period

Site fiscal/grid meter data where available for the same period

MERRA2 and ERA5 since 2000 up to the same period

The dataset has been released by Cubico Sustainable Investments Ltd under a CC-BY-4.0 open data license and is provided as is. However, please provide any feedback you might have on the dataset and format of the data. I'll try and add or link to additional file formats that might be easier to work with (e.g. for use with specific analysis software), and update this dataset periodically (e.g. twice a year), but please prompt me as required.

Feel free to use the data according to the license, however, it would be helpful to me if you could let me know where, how and why you are using the data, so that I can highlight this to the business (and renewables industry) and hopefully promote similar data sharing initiatives. I am particularly interested in performance analysis/improvement opportunities, how the dataset can be augmented with other (open) datasets, and sharing more generally within the renewables industry.

If you would like to get access to other datasets we may hold (e.g. more recent data, data from our other sites, ~30s resolution data, etc.), please let me know, and, if you have any questions or want to discuss open data and this or other initiatives, please contact me and I will endeavour to help.

I would like to thank Cubico's Senior Legal Advisor & Compliance Officer, IT Director, UK Asset Management Team, Executive Committee and my manager for supporting this initiative, as well as our partners GLIL for agreeing to release this data under an open license. I would also like to thank those I have talked to during the process of releasing this data under an open license and the encouragement and advice I have had on the way.

For contact my email address is charlie.plumley@cubicoinvest.com.

You can also access data from Penmanshiel wind farm here.

These data are supplements for the calculations of the methods from the article "Alignment of scanning lidars in offshore wind farms". The data was used to produce the results from the publication and is intended to be used here as sample data for illustrative purposes.

The dataset file contains data regarding the electrical and mechanical operational actual quantities and parameters obtained and recorded by the internal inverter controller of a Skystream 3.7 small wind turbine (SWT) installed on the roof of the High Voltage Laboratory at the Institute of Energy and Environment (IEE) of the University of Sao Paulo (USP), Brazil, recorded from 2017 to 2022.

The main electrical parameters are the energy, voltages, and currents in the connection grid point and power frequency. Mechanical information can be retrieved, such as the rotation and the wind speed. The temperature, measured in some location points to the nacelle and inverter, is also recorded. Several other parameters concerning the SWT inverter operation, such as the dc voltages on its internal bus, alarms, and flags, are also presented.

The files in the dataset are named as "data_swt_iee_usp_YYYY.csv" where YYYY is the referring year. In the files, the semicolon symbol (;) is used as a column separator, while the dot symbol (.) represents the decimal separator. The first row of the CSV file corresponds to the header row to help identify data as described in the file "data_description.txt". Sampling rate one record per minute.

The first line on each year-based file represents the header table description of the data columns.

The complete and detailed information about the installation, localization,and analysis is in the article published at: https://doi.org/10.3390/wind2040037

Overview

The SUMR-D CART2 turbine data are recorded by the CART2 wind turbine's supervisory control and data acquisition (SCADA) system for the Advanced Research Projects Agency–Energy (ARPA-E) SUMR-D project located at the National Renewable Energy Laboratory (NREL) Flatirons Campus. For the project, the CART2 wind turbine was outfitted with a highly flexible rotor specifically designed and constructed for the project. More details about the project can be found here: https://sumrwind.com/. The data include power, loads, and meteorological information from the turbine during startup, operation, and shutdown, and when it was parked and idle.

Data Details

Additional files are attached:
sumr_d_5-Min_Database.mat - a database file in MATLAB format of this dataset, which can be used to search for desired data files; sumr_d_5-Min_Database.xlsx - a database file in Microsoft Excel format of this dataset, which can be used to search for desired data files; loadcartU.m - this script loads in a CART data file and puts it in your workspace as a Matlab matrix (you can call this script from your own Matlab scripts to do your own analysis); charts.mat - this is a dependency file needed for the other scripts (it allows you to make custom preselections for cartPlotU.m); cartLoadHdrU.m - this script loads in the header file information for the data file (the header is embedded in each data file at the beginning); cartPlotU.m - this is a graphic user interface (GUI) that allows you to interactively look at different channels (to use it, run the script in Matlab, and load in the data file(s) of interest; from there, you can select different channels and plot things against each other; note that this script has issues with later versions of MATLAB; the preferred version to use is R2011b).

Data Quality

Wind turbine blade loading data were calibrated using blade gravity calibrations prior to data collection and throughout the data collection period. Blade loading was also checked for data quality following data collection as strain gauge measurements drifted throughout the data collection. These drifts in the strain gauge measurements were removed in post processing.

General description of wind turbine: The ETH owned wind turbine is Aventa AV-7, manufactured by Aventa AG in Switzerland and was commissioned in December 2002. The turbine is operated via a belt-driven generator and a frequency converter with a variable speed drive. The rated power of the Aventa AV-7 is 7 kW, beginning production at a wind speed of 2 m/s and having a cut-off speed of 14 m/s. The rotor diameter is 12.8 m with 3 rotor blades, and a hub height is 18m. The maximum rotational speed of the turbine is 63 rpm. The tower is a tubular steel-reinforced concrete structure, supported on concrete foundation, while the blades are made of glassfiber with a tubular steel main-spar. The turbine is regulated via a variable-speed and variable pitch control system.

Location of site: The wind turbine is located in Taggenberg, about 5 km from the city centre of Winterthur, Switzerland. This site is easily accessible by public transport and on foot with direct road access right next to the turbine. This prime location reduces the cost of site visits and allows for frequent personal monitoring of the site when test equipment is installed. The coordinates of the site are: 47°31'12.2"N 8°40'55.7"E.

Control and measurement systems and signals: The turbine is regulated via a variable-speed and collective variable pitch control system.

SHM Motivation: Designed and commissioned in 2002, the Aventa wind turbine in Winterthur is soon reaching its end of design lifetime. In order to assess the various techniques of predicting the remaining useful lifetime, a Structural Health Monitoring (SHM) campaign was implemented by ETH Zurich. The monitoring campaign started in 2020, and is still ongoing. In addition, the setup is used as a research platform on topics such as system identification, operational modal analysis, faults/damage detection and classification. We analyze the influence of operational and environmental conditions on the modal parameters and to further infer Performance Indicators (PIs) for assessing structural behavior in terms of deterioration processes.

Data Description: The tower and nacelle have been instrumented with 11 accelerometers distributed along the length of the tower, nacelle main frame, main bearing and generator. Two full bridge strain gauges are installed on the concrete tower based measuring fore-aft and side-side strain (and can be converted to bending moments) – all acceleration and strain signals sampled at 200Hz. Temperature and humidity are measured at the tower base – 1Hz data. In additional we are collecting operational performance data (SCADA), namely: wind speed, nacelle yaw orientation, rotor RPM, power output and turbine status – SCADA signals are sampled at 10Hz. See appendix for further details of the sensors layout.

The measurements/instrumentation setup, type and layout is provided in the pdf files.

The data: the data is provided in zip files corresponding to four use-cases as follows:

• Normal operation data for system identification
• Aerodynamic imbalance on one blade
• Rotor icing event
• Failure of the flexible coupling of the linear drive of the collective pitch system

The data for each of the four uses-cases is organized in zip files. The content of each zip file is as follows:

• Time-series data in HDF5 format
• Metadata:
  • Turbine specification (Aventa-AV-7.json and Aventa-AV-7.yaml)
  • Sensor specification (Aventa_sensors.json )
  • Unstructured description of the Aventa Turbine and the installed sensors (Aventa_Sensors_Specs.xlsx)
• Semantic artifacts:
  • WindIO Wind Turbine YAML schema describing turbine specifications (IEAontology_schema.yaml)
  • Sensor specification JSON schema (sensors_schema.json)
• Media: Pictures of leading edge roughness and a clip of wind turbine operation
• Code: Jupyter notebook containing example code to load metadata from JSON and data from HDF5 files (example.ipynb)

Additional data is available upon request, please contact:

• Prof. Dr. Eleni Chatzi (chatzi@ibk.baug.ethz.ch)
• Dr. Imad Abdallah (ai@rtdt.ai , abdallah@ibk.baug.ethz.ch)

For further details or questions, please contact:

Prof. Dr. Eleni Chatzi
Chair of Structural Mechanics & Monitoring

ETH Zürich
http://www.chatzi.ibk.ethz.ch/

SCADA Data from Aventa AV-7 (6kW) Research Wind Turbine, located in Taggenberg (ZH), Lat:47.52 Long:8.68236. The turbine is owned by IET-OST. The data covers time period from 2022-01-01 to 2023-07-20, sampled at 1Hz. The dataset is intended for environmental and operational analysis.
15 SCADA Channels contain timeseries observations (at 1Hz) of: Rotor speed in RPM, Generator speed in RPM, Stator temperature in C, Windspeed in m/s, Converter active power in kW, Wind direction w.r.t. nacelle measured in degrees, System supply voltage 24V , Pitch angle in degrees, Turbine status.
Structured metadata about wind turbine characteristics and SCADA channels is included as JSON files and CSV. Additional information is available upon request.

PaperThis dataset is associated with the paper published in Scientific Data, titled "SDWPF: A Dataset for Spatial Dynamic Wind Power Forecasting over a Large Turbine Array." You can access the paper: https://www.nature.com/articles/s41597-024-03427-5If you find this dataset useful, please consider citing our paper: Scientific Data Paper@article{zhou2024sdwpf, title={SDWPF: A Dataset for Spatial Dynamic Wind Power Forecasting over a Large Turbine Array}, author={Zhou, Jingbo and Lu, Xinjiang and Xiao, Yixiong and Tang, Jian and Su, Jiantao and Li, Yu, and Liu, Ji and Lyu, Junfu and Ma, Yanjun and Dou, Dejing},journal={Scientific Data},volume={11},number={1},pages={649},year={2024},url = {https://doi.org/10.1038/s41597-024-03427-5},publisher={Nature Publishing Group}}Baidu KDD Cup Paper@article{zhou2022sdwpf,title={SDWPF: A Dataset for Spatial Dynamic Wind Power Forecasting Challenge at KDD Cup 2022}, author={Zhou, Jingbo and Lu, Xinjiang and Xiao, Yixiong and Su, Jiantao and Lyu, Junfu and Ma, Yanjun and Dou, Dejing}, journal={arXiv preprint arXiv:2208.04360},url = {https://arxiv.org/abs/2208.04360}, year={2022}}BackgroundThe SDWPF dataset, collected over two years from a wind farm with 134 turbines, details the spatial layout of the turbines and dynamic context factors for each. This dataset was utilized to launch the ACM KDD Cup 2022, attracting registrations from over 2,400 teams worldwide. To facilitate its use, we have released the dataset in two parts: sdwpf_kddcup and sdwpf_full. The sdwpf_kddcup is the original dataset used for the Baidu KDD Cup 2022, comprising both training and test datasets. The sdwpf_full offers a more comprehensive collection, including additional data not available during the KDD Cup, such as weather conditions, dates, and elevation.sdwpf_kddcupThe sdwpf_kddcup dataset is the original dataset used for Baidu KDD Cup 2022 Challenge. The folder structure of sdwpf_kddcup is:sdwpf_kddcup --- sdwpf_245days_v1.csv --- sdwpf_baidukddcup2022_turb_location.csv --- final_phase_test --- infile --- 0001in.csv --- 0002in.csv --- ... --- outfile --- 0001out.csv --- 0002out.csv --- ...The descriptions of each sub-folder in the sdwpf_kddcup dataset are as follows:sdwpf_245days_v1.csv: This dataset, released for the KDD Cup 2022 challenge, includes data spanning 245 days.sdwpf_baidukddcup2022_turb_location.csv: This file provides the relative positions of all wind turbines within the dataset.final_phase_test: This dataset serves as the test data for the final phase of the Baidu KDD Cup. It allows for a comparison of methodologies against those of the award-winning teams from KDD Cup 2022. It includes an 'infile' folder containing input data for the model, and an 'outfile' folder which holds the ground truth for the corresponding output. In other words, for a model function y = f(x), x represents the files in the 'infile' folder, and the ground truth of y corresponds to files in the 'outfile' folder, such as {001out} = f({001in}).More information about the sdwpf_kddcup used for Baidu KDD Cup 2022 can be found in Baidu KDD Cup Paper: SDWPF: A Dataset for Spatial Dynamic Wind Power Forecasting Challenge at KDD Cup 2022sdwpf_fullThe sdwpf_full dataset offers more information than what was released for the KDD Cup 2022. It includes not only SCADA data but also weather data such as relative humidity, wind speed, and wind direction, sourced from the Fifth Generation of the European Centre for Medium-Range Weather Forecasts (ECMWF) atmospheric reanalyses of the global climate (ERA5). The dataset encompasses data collected over two years from a wind farm with 134 wind turbines, covering the period from January 2020 to December 2021. The folder structure of sdwpf_full is:sdwpf_full--- sdwpf_turb_location_elevation.csv--- sdwpf_2001_2112_full.csv--- sdwpf_2001_2112_full.parquetThe descriptions of each sub-folder in the sdwpf_full dataset are as follows:sdwpf_turb_location_elevation.csv: This file details the relative positions and elevations of all wind turbines within the dataset.sdwpf_2001_2112_full.csv: This dataset includes data collected two years from a wind farm containing 134 wind turbines, spanning from Jan. 2020 to Dec. 2021. It offers comprehensive enhancements over the sdwpf_kddcup/sdwpf_245days_v1.csv, including:Extended time span: It spans two years, from January 2020 to December 2021, whereas sdwpf_245days_v1.csv covers only 245 days.Enriched weather information: This includes additional data such as relative humidity, wind speed, and wind direction, sourced from the Fifth generation of the European Centre for Medium-Range Weather Forecasts (ECMWF) atmospheric reanalyses of the global climate (ERA5).Expanded temporal details: Unlike during the KDD Cup Challenge where timestamp information was withheld to prevent data linkage, this version includes specific timestamps for each data point.sdwpf_2001_2112_full.parquet: This dataset is identical to sdwpf_2001_2112_full.csv, but in a different data format.

The Chalmers wind turbine has variable speed operation with a direct driven generator and a frequency converter, it also has a digital control system developed by Chalmers. The wind turbine has a rated power of 45 kW and rated speed of 75 rpm. The wooden tower is 30 m high, the blades of carbon fibres are 7.5 m long, and the turbine diameter is 15.9 m. The individually blade pitch system is electrical. The turbine is situated on the island Björkö at Skarviksvägen, 20 km west of Göteborg city. The coordinates are: 57.71818820625921, 11.683382148764485.

69 SCADA and structural vibration and loads Channels timeseries (sampled at 20 and 100 Hz) such as nacelle accelerations, tower and blades bending moments are included.

Structured metadata about wind turbine characteristics, SCADA, vibration and loads channels are included as JSON files and CSV.

This particular dataset consisting of high frequency sampled data, is intended for condition and structural health analysis.

The data covers:

• the measurements sampled at 100 Hz correspond to the period from 05 July 2022 to 9 June 2023
• the measurements sampled at 20 Hz correspond to the period from 05 July 2022 to 2 August 2023

This repository includes:

Time-series data in csv format:

• B1_CL4_20.csv (this is the data sampled at 20 Hz)
• B1_CL4_100.csv (this is the data sampled at 100 Hz)

Metadata:

• Bjorko_Sensors_Specs_Metadata.csv (Sensors signals specification in csv format)
• Bjorko_modes_mapping.csv (numerical integer value representing the wind turbine controller system mode in csv format)
• Bjorko_modes_mapping.json (numerical integer value representing the wind turbine controller system mode in csv JSON format)
• Bjorko_digital_io_states_mappings.csv (Description of digital input and output states in the wind turbine controller system in csv format)

Media:

• Chalmers-Wind turbine.pdf (description of the wind turbine including pictures)
• Chalmers wind turbine description 220121-short.pdf (description of the wind turbine including pictures)

Semantic artifacts:

• N/A

Other:

• N/A

Additional information is available upon request.

The Wind Turbine Automation Market is projected to grow at 7.2% CAGR, reaching $22.59 Billion by 2029. Where is the industry heading next? Get the sample report now!

The dataset contains two separate files: NREL_Trainset40000.mat and NREL_Testset10000.mat.

The stored input enviormental and operational parameters are:

• Significant wave height, m (Hs), peak period, s (Tp), wave direction, deg (Wave_dir);
• Wind speed, m/s (Vw_mean, Vw_std), wind direction, deg (Wdir_mean, Wdir_std);
• Turbine rotational speed, rpm (Rpm_mean, Rpm_std), blade pitch, deg (Pitch_mean, Pitch_std), turbine yaw angle, deg (Yaw_mean, Yaw_std).

The output of the simulations includes the time series, sampled at 50 Hz, of the reaction force and bending moments at the mudline:

• Fzz, N
• Mxx, Nm
• Myy, Nm

contact: nandar.hlaing@uliege.be

Wind Automation Market size was valued at USD 2.61 Billion in 2024 and is projected to reach USD 2.98 Billion by 2031, growing at a CAGR of 5.44% from 2024 to 2031.

Global Wind Automation Market Drivers

Global Shift Toward Renewable Energy: As the world moves toward sustainability and reducing carbon emissions, the demand for renewable energy sources, particularly wind energy, has surged. Wind power is one of the fastest-growing renewable energy sources, and governments around the world are providing incentives and setting ambitious targets for wind energy generation to meet their climate goals. This growth in wind power capacity directly drives the need for more sophisticated automation solutions to manage and optimize wind farms.

Automated systems can improve the operational efficiency of wind turbines, making it easier to monitor performance and ensure that they are running at optimal capacity. As more wind farms are constructed, the demand for automation technologies to monitor, control, and improve performance is rising, driving the wind automation market.

Increasing Focus on Operational Efficiency: Wind farms face numerous operational challenges, such as managing large numbers of turbines, ensuring maximum efficiency, reducing downtime, and minimizing maintenance costs. Automation technologies can help address these issues by enabling remote monitoring, predictive maintenance, and real-time adjustments to turbine performance.

Automation systems are capable of collecting and analyzing data from sensors on the turbines to detect early signs of wear and tear or potential failures. This helps prevent unexpected breakdowns, optimize energy production, and reduce costly downtime. The drive for operational efficiency and cost reduction in the wind energy sector is one of the major drivers for the adoption of wind automation systems.

Overview

This repository contains data from an offshore measurement campaign conducted during the installation of the offshore wind farm trianel wind farm borkum II (https://www.trianel-borkumzwei.de/). The wind farm consists of 32 Senvion 6XM152 turbines. The installation took place between August 2019 and May 2020.

An offshore wind turbine undergoing installation is interesting from a research point of view for several reasons:

Simple geometry: turbine foundation and tower are both rotationally symmetric steel tubes. Rotational symmetry also leads to (approximate) isotropical structural characteristics in the plane normal to tower and foundation.

High Reynolds number flow: Assuming a tower diameter of 6 m, and average wind speeds ranging from 5 m/s to 12 m/s under installation conditions, Reynolds numbers range from 4.5 million to 10.5 million.

Wave loading under full-scale conditions.

Practical relevance to improving the competetivity of offshore wind.

For fluid mechanics, closely monitoring offshore wind turbines under wind and wave loading thus compares to a full-scale experiment. Monitoring 32 turbines undergoing installation thus enables the measurement a broad spectrum of different states.

The investigation into the data is ongoing, questions and contributions are welcome. The current data release still does not include all data. The dataset will thus be updated again in the future with more data to come. First analytic results can be found here:

Sander, A, Haselsteiner, AF, Barat, K, Janssen, M, Oelker, S, Ohlendorf, J, & Thoben, K. "Relative Motion During Single Blade Installation: Measurements From the North Sea." Proceedings of the ASME 2020 39th International Conference on Ocean, Offshore and Arctic Engineering. Volume 9: Ocean Renewable Energy. Virtual, Online. August 3–7, 2020. V009T09A069. ASME. https://doi.org/10.1115/OMAE2020-18935

Sander, A, Meinhardt, C & Thoben, KD. "Monitoring of Offshore Wind Turbines under Wind and Wave Loading during Installation" Proceedings of the EuroDyn 2020 XI International Conference on Structural Dynamics. Volume 1. Virtual, Online. November 23-26, 2020. https://generalconferencefiles.s3-eu-west-1.amazonaws.com/eurodyn_2020_ebook_procedings_vol1.pdf

Recordings of the conference presentations are available on youtube:

OMAE20: https://www.youtube.com/watch?v=QcAwdv6Z4e4

EURODYN20: https://www.youtube.com/watch?v=iL-jAe0luTw

Physical Background

An offshore wind turbine under installation conditions can be simplified as a cantilevered beam (circular cross-section, rotationally symmetric wall thickness) with an eccentric mass (nacelle with generator) vibrating transversally (fore-aft and side-side in the reference system of the nacelle) under wind and wave loads.

Both wind and wave loads are stochastic and are described using statistical models.

Wave loads are a function of the sea state. For a sea state, the most important parameters are significant wave heigh H_m0 and Wave peak period T_P. To a lesser extend, wave direction, zero upcrossing period and maximum wave height are also important. Different statistical models can be used to describe the sea state and the relationship between significant wave heigh H_m0 and wave peak period. Most prominent in the North Sea is the JONSWAP spectrum.

Wind loads are depending on wind speed, wind direction, shear factor and turbulence intensity. Different statistical models are available to describe the wind spectrum.

Wind and wave loads trigger a structural response of the turbine. The structural response depends on the loading spectrum as well as the transfer function. Furthermore, the structural response is strongly depending on the damping and elasticity of the structure. In turbines, damping is typically very low (~ 0.5 - 1.5 %).

The response is dominated by the first Eigenfrequency of the turbine. As the turbine is assumed to be rotationally symmetric, the fore-aft and side-side mode are extremely close together if not indistinguishable [1].

The response has the characteristics of a narrow-band random vibration. A narrow-band random vibration is characterized by being dominated by a single, narrow frequency peak (here: first Eigenfrequency). The amplitude envelope follows a Rayleigh distribution and the phase angle is equally distributed between 0 and 2 pi.

If viewed from above, the structural response describes a closed curve (orbit) which can be characterized by it shape (eccentricity), mean amplitude and direction. Mathematically speaking, this is a lissajous-figure, where the time series from one response direction is plotted as a function of the time series of the second response direction.

[1]: under installation condition

Experimental Setup

Several locations were used to record data during the installation of the wind farm. They are listed in the following table:

helihoist-{1,2}: data recorded from the helicopter hoisting platform atop the turbine nacelle. For most installations, two sensor boxes were deployed to ensure data availability.

tp: Measurements from the transition piece

sbitroot: Measurements from the blade lifting yoke's blade root side. The Z-axis is aligned to the blade main axis, X-Axis is perpendicular.

sbittip: Measurements from the tip side of the blade lifting yoke. Z-axis aligned with the blade main axis, X-axis perpendicular

damper: Measurements from the tuned mass damper used during single blade installation

towertop: measurements from inside the turbine tower at the upper lift plattform

towertransfer: measurements from atop the towers during sail out from the base harbour to the installation site

Organization of data

For each turbine installation, a separate folder can be found, e.g. turbine-01 for the first and turbine-16 for the 16th turbine. Turbine numbering follows the order of installation.

Different data sources are organized in subfolders for each turbine dataset. Unfortunately, not every data source is available for each turbine. Data sources are roughly sorted into categories. The following table lists these categories:

location / tom : data from custom build sensor boxes. Data includes acceleration, angular acceleration, magnetic field, gnss recording and rough estimates of the eulerian angles.

waves / wmb-sued : Sea state statistics for the installatin period of the turbine.

waves / fino : Sea state statistics from the german research platform FINO1 located approx. 6 km from the installation site.

waves / waveradar : Sea state statstics, recorded by a wave rider wave laser.

wind / lidar : high fidelity wind data recorded on the installation vessel during the installation of the wind farm.

wind / scada : 10 min. mean wind statistisc recorded on wind turbines in the vicinity of the installation site. This data is used in case no LIDAR data is available.

wind / anemometer : During some of the installations, anemometers were present on the installation vessel. These recordings are sorted into this sub-subfolder.

wind / fino : Additonal wind statistics recorded by the FINO research station. Least recommended for investigations, as these recordings were taken approx. 6 km from the installation site.

The zenodo data set includes 16 zip archives (for 16 turbines) as well as one zip archive including environmental data. The following lists the folder structure of the turbine-04.zip archive (with most of the data files removed for clarity).

└── turbines ├── turbine-04 │ ├── helihoist-1 │ │ └── tom │ │ └── clean │ │ ├── turbine-04_helihoist-1_tom_clean_2019-09-01-11-27-17_2019-09-01-11-54-00.csv │ │ ├── turbine-04_helihoist-1_tom_clean_2019-09-01-11-54-00_2019-09-01-12-20-44.csv │ ├── sbitroot │ │ └── tom │ │ └── clean │ │ ├── turbine-04_sbitroot_tom_clean_2019-09-07-06-48-53_2019-09-07-07-17-16.csv │ │ ├── turbine-04_sbitroot_tom_clean_2019-09-07-07-17-16_2019-09-07-07-45-59.csv │ ├── towertop │ │ └── tom │ │ └── clean │ │ ├── turbine-04_towertop_tom_clean_2000-01-06-18-55-52_2000-01-06-20-30-10.csv │ │ ├── turbine-04_towertop_tom_clean_2000-01-06-20-30-11_2000-01-06-22-04-31.csv │ ├── towertransfer │ │ └── tom │ │ └── clean │ │ ├── turbine-04_towertransfer_tom_clean_2019-08-31-03-11-53_2019-08-31-04-00-09.csv │ │ ├── turbine-04_towertransfer_tom_clean_2019-08-31-04-00-10_2019-08-31-04-48-19.csv │ └── tp │ └── tom │ └── clean │ ├── turbine-04_tp_tom_clean_2019-08-31-18-34-45_2019-08-31-19-07-52.csv │ ├── turbine-04_tp_tom_clean_2019-08-31-19-08-00_2019-08-31-19-40-43.csv

The following list the contents of the environment.zip archive. Note that again most of the data files have been removed for clarity.

└── environment ├── waves │ └── wmb-sued │ ├── wmb-sued_2019-08-15.csv │ ├── wmb-sued_2019-08-16.csv │ ├── wmb-sued_2019-08-17.csv └── wind └── lidar ├── lidar_2019-08-03.csv ├── lidar_2019-08-04.csv ├── lidar_2019-08-05.csv

TOM data description

The abbreviation TOM referes to Tower Oscillation Measurement and the data that was acquired using a specific set of sensor boxes built by university of Bremen for this specific purpose. These Sensor Boxes were initially designed to measure accelerations and GPS tracks of offshore wind turbine towers undergoing installation.

The electricity production and exchange in MW in 5 minutes intervals and updated every 5th minute. NOTE: Data are based on upscaled real-time power measurements from the SCADA system. Errors will occur, and will generally not be corrected. For statistical purposes, see Production and Consumption - Settlement.

The total load, including loses, is calculated as the sum of production from power plants, solar and wind power plus the exchange to Germany, Sweden, Norway, and The Netherlands. The exchange between Bornholm and Price area SE4 is included in the exchange DK2 to Sweden, and should not be taken into account.