Rolling bearing fault diagnosis based on improved deep residual noise reduction network
KONG Guocai
WANG Lie
YUAN Feng
ZHOU Yafeng
JI Xin
YIN Minghu
Abstract：[Objective]Aiming at the limitations of traditional bearing fault diagnosis methods and the noise sensitivity of machine learning methods,an improved deep residual noise reduction(DRNR)network model was proposed.[Methods]Firstly,vibration signals were transformed into time-frequency maps via wavelet transform,and data augmentation technique was applied.Secondly,a deep residual denoising network was designed,and key hyperparameters were optimized using an improved dung beetle optimization(IDBO)algorithm.Finally,the optimized model was tested on the Case Western Reserve University(CWRU)dataset and a real-world dataset from belt conveyor drive motors.[Results]The results show that the proposed model achieves a diagnostic accuracy exceeding 99.3%in low-noise environments and maintains 97.3%under high-noise conditions,demonstrating superior robustness and noise resistance compared to state-of-the-art models.
Keywords：Rolling bearingVibration signal analysisDeep learningDeep residual noise reduction network
Publication Date：2026-04-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：12( 154-164,187 )