SNU Planetary Gearbox Dataset: Vibration and Encoder Data for Fault Diagnosis

Approximately 100 minutes of data per configuration were recorded for 14 distinct test settings. This dataset contains raw vibration and encoder signals from a planetary gearbox testbed, designed for fault diagnosis research. It was created by Jinoh Yoo and includes data for multiple speed profiles, fault types, and severities.

Use Cases

• Train fault classification models based on vibration signals from healthy and damaged gears.
• Develop anomaly detection algorithms for non-stationary operating conditions based on sinusoidal speed profiles.
• Benchmark signal processing techniques like time synchronous averaging using the provided encoder data.
• Validate health data map visualization methods for toothwise fault identification.

Strengths

• Contains 14 distinct test configurations covering multiple fault types and severities.
• Data was recorded at a high sampling rate of 25,600 Hz.
• Includes both stationary and non-stationary (sinusoidal) speed modes.
• Accompanied by MATLAB code for angular resampling and health data map visualization.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect the specific experimental bias inherent to the single 2kW testbed used.

Provenance

Source

figshare

Collection Method

Experimental data collected from a 2kW testbed with an accelerometer and encoder.

Freshness

Last updated 2026-05-03 09:02:11; freshness should be verified.