CWRU Bearing Fault Dataset — 2HP Motor Vibration, 4 Fault Diameters [12k & 48k Hz]

Overview

The Case Western Reserve University (CWRU) Bearing Data Center dataset is the most widely cited public benchmark for rolling element bearing fault diagnosis in industrial machinery. Experiments were conducted using a 2 HP Reliance Electric motor with bearings seeded with faults using electro-discharge machining (EDM) at four different fault diameters: 0.007, 0.014, 0.021, and 0.028 inches.

Vibration data were measured near and remote from motor bearings using accelerometers, covering normal operation and three fault locations: inner raceway, rolling element (ball), and outer raceway. The motor was tested under loads from 0 to 3 horsepower (motor speeds 1797 to 1720 RPM), yielding a dataset that spans multiple operating conditions and fault severities.

Data are provided at 12,000 samples/second and 48,000 samples/second for drive-end bearing experiments, and at 12,000 samples/second for fan-end experiments. The dataset has been downloaded and cited in thousands of bearing fault diagnosis, deep learning, and vibration signal processing studies.

Column Schema

Key Statistics

• Fault Types: normal, inner raceway, ball, outer raceway
• Fault Diameters: 0.007, 0.014, 0.021, 0.028 inches
• Load Conditions: 0, 1, 2, 3 HP (4 operating conditions)
• Sampling Rate: 12 kHz (DE and FE) and 48 kHz (DE only)
• File Format: MAT (MATLAB) and CSV
• Time Period: Ongoing benchmark, original data collected 2000s

Use Cases

• Bearing fault diagnosis and fault type classification using vibration signals
• Convolutional neural network and deep learning benchmarking for IIoT fault detection
• Transfer learning across motor loads and operating conditions
• Vibration signal feature extraction and frequency-domain analysis (FFT, envelope analysis)

Source & Attribution

The dataset is maintained and freely distributed by the Bearing Data Center at Case Western Reserve University's Case School of Engineering. It is available directly from the CWRU Bearing Data Center website and also mirrored on Kaggle and Zenodo for convenience.