# NREL 5 MW Floating Wind Turbine Drivetrain Bearing Damage Vibration Dataset, Subset for Fault Diagnosis (Dibaj and Nejad)

## Facts

- Provider: Ali Dibaj and Amir Nejad
- License: CC-BY-4.0 (https://zenodo.org/records/7674842)
- Size: 3 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/alidi/wind-turbine-5mw-bearing-dataset
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.7674842
- Category: Energy
- Source: https://doi.org/10.1016/j.renene.2022.12.049
- Page: https://scigantic.com/archive/438efa50-b653-56df-bbf2-c120cd342923
- Keywords: Energy, Parquet, Hugging Face

## About

Simulated drivetrain vibration signals for bearing fault diagnosis in wind turbines, from the NREL 5 MW reference drivetrain model installed on a spar-type floating offshore wind turbine. The original Zenodo dataset by Ali Dibaj and Amir Nejad holds one-hour simulated acceleration measurements sampled at 200 Hz for healthy and damaged bearings; this Hugging Face copy is a 12-scenario subset that links the Zenodo record and a deep-learning fault diagnosis code repository. Each row has two vibration time waveforms from accelerometer locations A1A and A3A, the signal length, a fault label (healthy, main bearing, high-speed bearing or low-speed planet bearing damage), the environmental condition (EC1, EC2 or EC3) at wind speeds of 7, 12 and 14 m/s, and the 200 Hz sampling frequency. One Parquet file, about 70 MB (measured 2026-11-21); load with Hugging Face datasets or pandas. Use it for condition monitoring, vibration-based fault classification and transfer learning across operating conditions. CC BY 4.0, as stated on the card and on the Zenodo record (dataset DOI 10.5281/zenodo.7674842); cite Dibaj et al., Renewable Energy 2023, doi:10.1016/j.renene.2022.12.049.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("alidi/wind-turbine-5mw-bearing-dataset")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/train-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/438efa50-b653-56df-bbf2-c120cd342923
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/438efa50-b653-56df-bbf2-c120cd342923
- Over MCP: `get_schema_card` with id `438efa50-b653-56df-bbf2-c120cd342923` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Energy datasets](https://scigantic.com/topics/energy)
