Institute of Driveline Technology (IAA)

Ball Bearing Dataset

Project History

The dataset presented in this work was acquired within the funded research projects SEED (Smart Efficient Electric Driveline, FKZ 13FH585KX0) and KI-Werkstatt. As part of the project, a comprehensive experimental database is established to provide high-quality reference data for the analysis and validation of bearing fault diagnosis methods. The dataset includes measurements from SKF 6314 deep-groove ball bearings representing healthy operation, laser-induced inner raceway defects, laser-induced outer raceway defects, and corrosion-induced damage. Data are recorded at rotational speeds of 1000 rpm and 2000 rpm under two representative combined load cases of 3257 N radial / 1556 N axial and 4893 N radial / 2337 N axial. These operating conditions are selected to capture the influence of varying speed and mechanical loading on bearing behavior, resulting in a diverse dataset suitable for investigating fault-related vibration characteristics under representative operating scenarios.

Faults and Test Specification

Rotational Speed [rpm] Inner Race Frequency (BPFI) [Hz] Outer Race Frequency (BPFO) [Hz] Rotational Frequency / Harmonics [Hz]
1000 82,06 51,27 16,67
2000 164,12 102,54 33,33
Bearing Type Fault Fault location Rotational Speed [rpm] Force Radial [N] Force Axial [N] Fault ID / Download
6314 None - 1000 3257 1556  H1
6314 None - 1000  4893  2337  H2
6314 None - 2000  3257  1556  H3
6314 None - 2000  4893  2337  H4
6314 Laser Inner Raceway 1000 3257 1556  IR1
6314 Laser Inner Raceway 1000  4893  2337  IR2
6314 Laser Inner Raceway 2000  3257  1556  IR3
6314 Laser Inner Raceway 2000  4893  2337  IR4
6314 Laser Outer Raceway 1000  3257  1556  OR1
6314 Laser Outer Raceway 1000  4893  2337  OR2
6314 Laser Outer Raceway 2000  3257  1556  OR3
6314 Laser Outer Raceway 2000  4893  2337  OR4
6314 Corrosion - 1000  3257  1556  C1
6314 Corrosion - 1000  4893  2337  C2
6314 Corrosion - 2000  3257  1556  C3
6314 Corrosion - 2000  4893  2337  C4

Applications of the dataset

The dataset can be used for vibration analysis, signal processing, machine learning-based fault diagnosis, and the development and evaluation of condition monitoring algorithms. In addition, it is suitable for Generative Adversarial Network (GAN)-based data augmentation and synthetic data generation, as well as for benchmarking deep learning models for fault detection, classification, and anomaly detection.