SNU Planetary Gearbox Dataset

Fault Modes and Severities</p><p>-----------------------------</p><p dir="ltr">The dataset includes one healthy condition and three artificial fault modes.</p><p><br></p><p dir="ltr">- Normal</p><p dir="ltr"> Baseline healthy condition</p><p><br></p><p dir="ltr">- Planet-ring fault</p><p dir="ltr"> Fault on the planet gear tooth meshing with the ring gear</p><p dir="ltr"> Groove diameter: 0.75 mm, 1.00 mm, 1.25 mm</p><p><br></p><p dir="ltr">- Planet-sun fault</p><p dir="ltr"> Fault on the planet gear tooth meshing with the sun gear</p><p dir="ltr"> Groove diameter: 0.75 mm</p><p><br></p><p dir="ltr">- Sun fault</p><p dir="ltr"> Fault on the sun gear tooth</p><p dir="ltr"> Groove diameter: 0.50 mm, 0.75 mm</p><p><br></p><p dir="ltr">Here, groove diameter denotes the nominal diameter of the machined artificial line-shaped fault in millimeters.</p><p><br></p><p dir="ltr">4.

Repository Contents</p><p>----------------------</p><p dir="ltr">This repository contains the following files:</p><p><br></p><p dir="ltr">- SNU Planetary Gearbox Dataset.zip</p><p dir="ltr"> Primary data archive containing vibration and encoder signals.</p><p><br></p><p dir="ltr">- SNU_Figshare_code.zip</p><p dir="ltr"> MATLAB validation code package used to reproduce the technical validation figures.</p><p><br></p><p dir="ltr">- Readme.txt</p><p dir="ltr"> Dataset description, folder hierarchy, acquisition protocol, and usage information.</p><p><br></p><p dir="ltr">- datasummary.csv</p><p dir="ltr"> Machine-readable file-level metadata, including file path, signal type, sample count, amplitude range, file size, SHA-256 checksum, and known anomalies.</p><p><br></p><p dir="ltr">5.

Data Structure and Format</p><p>----------------------------</p><p dir="ltr">All time-series data are stored as tab-delimited text files (.txt) without headers. Each file is provided as a single-column vector with shape (N, 1).</p><p><br></p><p dir="ltr">Folder hierarchy:</p><p dir="ltr">[Fault type] / [Severity level] / [Operating mode] /</p><p><br></p><p dir="ltr">Examples:</p><p dir="ltr">- Normal/Normal/Stationary/Vib.txt</p><p dir="ltr">- Planet-ring fault/Line_0.75/Stationary/Vib.txt</p><p dir="ltr">- Sun fault/Line_0.5/Nonstationary/Encoder.txt</p><p><br></p><p dir="ltr">Files:</p><p dir="ltr">- Vib.txt</p><p dir="ltr"> Raw vibration signal in units of gravity (g)</p><p><br></p><p dir="ltr">- Encoder.txt</p><p dir="ltr"> Encoder pulse signal in volts (V), nominally transitioning between 0 and 5 V for rotational speed analysis</p><p><br></p><p dir="ltr">Each time-series file contains approximately 1.54 x 10^8 samples.

Actual sample counts may vary slightly across files because the recordings include start/stop margins; the exact sample count of each file is listed in datasummary.csv.</p><p><br></p><p dir="ltr">6. Validation Code Package</p><p>--------------------------</p><p dir="ltr">The repository includes MATLAB Live Scripts and supporting functions for reproducing the technical validation figures in the associated Data Descriptor manuscript.</p><p><br></p><p dir="ltr">Main scripts:</p><p dir="ltr">- Data_region_extract.mlx</p><p dir="ltr"> Extracts 5-minute data segments from the original dataset.</p><p><br></p><p dir="ltr">- Figure567.mlx</p><p dir="ltr"> Reproduces Fig. 5, Fig. 6, and Fig. 7.</p><p><br></p><p dir="ltr">- Figure8.mlx</p><p dir="ltr"> Reproduces Fig. 8.</p><p><br></p><p dir="ltr">- Figure9.mlx</p><p dir="ltr"> Reproduces Fig. 9.</p><p><br></p><p dir="ltr">The validation scripts were developed and tested in MATLAB R2025b using the Signal Processing Toolbox.

Detailed dependencies, folder structure, and execution instructions are provided in README_Figure_Reproduction.txt.</p><p><br></p><p dir="ltr">7. Citation</p><p>-----------</p><p dir="ltr">Users should cite this Figshare dataset when using the data. Citation information for the associated Data Descriptor manuscript will be updated after journal publication.</p><p><br></p><p dir="ltr">Related method reference:</p><p dir="ltr">Ha, J. M., Park, J., Na, K., Kim, Y. & Youn, B. D. Toothwise fault identification for a planetary gearbox based on a health data map.

IEEE Transactions on Industrial Electronics 65, 5903–5912 (2017).</p>