Bearing Fault Datasets
Datasets about the full lifetime of bearings in machines under a range of operating conditions.
Asset Description
This asset provides large and well documented datasets covering the full lifetime of bearings in machines under a range of operating conditions. These data are useful for condition monitoring algorithm development, AI model training, testing and validation.
Asset Details
Dataset Information
Fully documented datasets acquired through accelerated lifetime testing of bearings. Data obtained through vibrational condition monitoring (using accelerometer sensors) and through acoustic condition monitoring (using microphone sensors) under a whole range of operating conditions.
- Healthy bearings and bearings with well-defined initial faults: indentations from ø145 µm to ø490 µm + 4 sizes of spall faults on the inner race
- Radial load from 0.7 kN to 9 kN
- Bearing temperatures from 30C to 55C
- Motor speed variations from 600 to 2000 rpm
- Various lubrication options: oil bath, oil recirculation, grease
- Stationary operating conditions versus different stationary speed and loading conditions
Usage
The datasets consist of time series binary files that can be downloaded. A full description of each dataset, as well as some free samples, are available upon request. The datasets can be used for:
- Classification models for diagnostic decision making
- Statistical models for feature selection
- Systems models for feature engineering
- Probabilistic prediction models for remaining useful lifetime estimation
- As a test bed for a machine learning models
Maturity
The datasets are fully documented and available.
Licence
Proprietary
Resources
- Datasets were created by Flanders Make (Belgium). For further information, please contact christophe.bruynseraede@flandersmake.be
- Provided by Flanders Make
Acknowledgement
Most of the datasets were created in the frame of the Living Lab Smart Maintenance project, with the support of VLAIO (Flanders Innovation & Entrepreneurship) in Flanders, Belgium