Supervisory control and data acquisition records with stop categories of 47 wind turbines at 4 wind farms in Great Britain and Northern Ireland
Version 1.1.1. This repository holds the data and metadata of the release, file by file: every file is byte-identical to the release, and its SHA-256 is listed in MIRROR_MANIFEST.json. The archival release, with the code that builds and verifies it, is deposited on Zenodo; its DOI is added here when it is published.
Measured 10-minute supervisory control and data acquisition (SCADA) statistics from 47 turbines at 4 wind farms in Great Britain and Northern Ireland, released by 2 custodians under CC BY 4.0, harmonised to one core schema of 25 columns and kept with every channel the source files of each farm have in common (1,265 channel columns over the four farms). Each row carries a binary downtime flag downtime_any derived from the farm's own records. Where it is set, the row also carries the custodian's own category of the stop (source_stop_category, source_stop_code, source_stop_text) and a derived stop_category (forced, scheduled, other, unknown) with its mapping table. The custodians' event logs (Cubico status logs with stops, warnings and informational events; the Hill of Towie alarm log) are included. 5,520,822 ten-minute rows, 201,762 with downtime_any = 1, of which 60,461 forced, 19,483 scheduled, 74,464 other and 47,354 unknown; 903,155 complete 1-hour windows, 33,974 of them with downtime. This is a derived corpus. The measurements and the status, alarm, downtime and operating-state records are the custodians'; this release harmonises them across farms and custodians and derives the downtime flag and the stop category with their provenance. It does not replace the source deposits (see NOTICE.txt).
Yueqiang Xu, University of Oulu. ORCID: https://orcid.org/0000-0002-3528-423X
Use CITATION.cff (GitHub shows it under "Cite this repository"), and cite the source data this release is derived from:
- Plumley, C., & Takeuchi, R. (2025). Kelmarsh wind farm data [Data set]. Zenodo. https://doi.org/10.5281/ZENODO.16807551
- Plumley, C., & Takeuchi, R. (2025). Penmanshiel wind farm data [Data set]. Zenodo. https://doi.org/10.5281/ZENODO.16807304
- Clerc, A., & Lingkan, E. (2026). Hill of Towie wind farm open dataset (Version 2.1.0) [Data set]. Zenodo. https://doi.org/10.5281/ZENODO.22662930
- Clerc, A. (2026). Altahullion wind farm open dataset [Data set]. Zenodo. https://doi.org/10.5281/ZENODO.19948235
cc-by-4.0. The licence text and any attribution notice required by the sources are in the licence and notice files of this repository; they travel with every copy.
| path | bytes |
|---|---|
| data/Altahullion_2025.parquet | 18,899,909 |
| data/Altahullion_2026.parquet | 9,649,303 |
| events/HillOfTowie_events.parquet | 9,258,168 |
| events/Kelmarsh_events.parquet | 1,574,295 |
| events/Penmanshiel_events.parquet | 3,203,249 |
| LICENSE.txt | 20,537 |
| metadata/channels.csv | 307,904 |
| metadata/channels_by_year.csv | 214,613 |
| metadata/columns.csv | 4,071 |
| metadata/events_dictionary.csv | 4,218 |
| metadata/label_derivation.csv | 2,254 |
| metadata/sources.json | 4,688 |
| metadata/stop_category_map.csv | 8,697 |
| metadata/turbines.csv | 6,488 |
| NOTICE.txt | 3,309 |
| proof/FACTS.json | 8,417 |
| proof/MANIFEST.json | 4,330 |
| docs/RELEASE_README.md | 22,102 |
| path | asset name | bytes |
|---|---|---|
| data/HillOfTowie_2019.parquet | data__HillOfTowie_2019.parquet | 895,115,871 |
| data/HillOfTowie_2020.parquet | data__HillOfTowie_2020.parquet | 886,485,527 |
| data/Kelmarsh_2019.parquet | data__Kelmarsh_2019.parquet | 180,857,968 |
| data/Kelmarsh_2020.parquet | data__Kelmarsh_2020.parquet | 263,266,602 |
| data/Kelmarsh_2021.parquet | data__Kelmarsh_2021.parquet | 255,643,361 |
| data/Penmanshiel_2019.parquet | data__Penmanshiel_2019.parquet | 429,458,514 |
| data/Penmanshiel_2020.parquet | data__Penmanshiel_2020.parquet | 605,409,007 |
| data/Penmanshiel_2021.parquet | data__Penmanshiel_2021.parquet | 628,435,757 |
Each large file is attached to the release v1.1.1 under the asset name in the table (GitHub asset names cannot hold folders); save it under its path. For example, in a shell:
curl -L -o data/HillOfTowie_2019.parquet https://github.com/NordaLab/wind-scada-stop-categories/releases/download/v1.1.1/data__HillOfTowie_2019.parquet
6 file(s) of the release are not here: code, build logs, change history and build provenance records. MIRROR_MANIFEST.json lists each one; all are in the archival release.