Processed subset used in our STR-DDPM experiments on the Paderborn bearing dataset

This repository provides the processed experimental subset used in our paper, “STR-DDPM: Residual-domain diffusion-based data augmentation with seasonal-trend-residual decomposition for few-shot motor fault diagnosis”, derived from the publicly available Paderborn bearing dataset.

Only the subset and operating conditions used in this study are included here. The original MATLAB files were reorganized and converted into CSV format for our experimental pipeline. This record is intended to support reproducibility of the associated paper and does not replace the complete original public dataset.

The original Paderborn bearing dataset is publicly available from the official Bearing Data Center of Paderborn University. According to the official dataset information, the data include synchronously measured motor currents and vibration signals, as well as additional operating-condition variables, and are provided as MATLAB files. The dataset is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License. Non-commercial academic use is allowed, citation of the original source is required, and commercial use requires contacting the original author. Users should therefore comply with the original dataset terms when reusing this processed subset.