Robust bearing fault diagnosis for rotating machinery under noisy and variable operating conditions using a Wavelet-Mamba Network
- 1School of Railway Transportation, Shandong Jiaotong University, Jinan, China
Abstract
Bearing fault diagnosis for rotating machinery remains challenging due to strong background noise and variable operating conditions in harsh industrial environments. To address these issues, this paper proposes a robust and lightweight bearing fault diagnosis framework, termed Multi-scale and Dual-scale Wavelet Mamba (MDW-Mamba), for noisy and non-stationary industrial applications. The proposed method employs a multi-stage “noise-feature decoupling” strategy, comprising a statistically guided three-channel input, a Multi-scale Omni-kernel Convolutions (MOM-Conv) backbone, and a Dual-Scale Wavelet Mamba (DSW-Mamba) module. This architecture explicitly separates fault components from broadband noise while enabling efficient long-term dependency modeling. Experimental results on three public datasets under variable-speed conditions demonstrate that MDW-Mamba consistently outperforms state-of-the-art methods. It achieves an average classification accuracy exceeding 87% even at a signal-to-noise ratio of −6 dB, showing exceptional robustness against strong interference. Furthermore, the model maintains an ultra-lightweight footprint of 0.26 M parameters and a fast inference time of 1.285 ms, indicating strong potential for real-time, on-board deployment in resource-constrained industrial edge devices.
Related
- Consistent Modeling of SEI Growth via Electron Diffusion Validated Across Diverse Cell Specifications and Operating Conditions in LFP/Graphite Cells
- Three-dimensional characteristics analysis of maglev train with rotating operation conditions
- An analysis of homogeneous-heterogeneous chemical reactions, entropy generation, joule heating, variable thermal conductivity, Dufour effect, Cattaneo-Christov heat and mass flux, and MHD flow of the casson fluid in a rotating channel
- Investigation by LES of a Rich burn – Quick mix – Lean burn lab-scale combustion chamber operating in high-pressure conditions
- Relativistic Six-Fluid Rotating-Plasma Force-Balance Equilibria for p-11B Fusion Conditions in Spherical Tori
BibTeX
@article{cui2026robust,
title = {{Robust bearing fault diagnosis for rotating machinery under noisy and variable operating conditions using a Wavelet-Mamba Network}},
author = {Dayou Cui and Zhaoyan Xie and Zhixue Wang and Xiaowei Li and Xiangxuan Meng},
journal = {PLOS One},
year = {2026},
month = sep,
doi = {10.1371/journal.pone.0358759},
url = {https://doi.org/10.1371/journal.pone.0358759},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Robust bearing fault diagnosis for rotating machinery under noisy and variable operating conditions using a Wavelet-Mamba Network
AU - Dayou Cui
AU - Zhaoyan Xie
AU - Zhixue Wang
AU - Xiaowei Li
AU - Xiangxuan Meng
PY - 2026
DA - 2026/09/28
JO - PLOS One
DO - 10.1371/journal.pone.0358759
UR - https://doi.org/10.1371/journal.pone.0358759
AB - Bearing fault diagnosis for rotating machinery remains challenging due to strong background noise and variable operating conditions in harsh industrial environments. To address these issues, this paper proposes a robust and lightweight bearing fault diagnosis framework, termed Multi-scale and Dual-scale Wavelet Mamba (MDW-Mamba), for noisy and non-stationary industrial applications. The proposed method employs a multi-stage “noise-feature decoupling” strategy, comprising a statistically guided three-channel input, a Multi-scale Omni-kernel Convolutions (MOM-Conv) backbone, and a Dual-Scale Wavelet Mamba (DSW-Mamba) module. This architecture explicitly separates fault components from broadband noise while enabling efficient long-term dependency modeling. Experimental results on three public datasets under variable-speed conditions demonstrate that MDW-Mamba consistently outperforms state-of-the-art methods. It achieves an average classification accuracy exceeding 87% even at a signal-to-noise ratio of −6 dB, showing exceptional robustness against strong interference. Furthermore, the model maintains an ultra-lightweight footprint of 0.26 M parameters and a fast inference time of 1.285 ms, indicating strong potential for real-time, on-board deployment in resource-constrained industrial edge devices.
SN - 1932-6203
ER -