A Fault Diagnosis Method for Wind Turbine Main Shaft Bearings Based on Lightweight Attention Residual Networks

CN120429771BActive Publication Date: 2026-05-26NORTHEAST GASOLINEEUM UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
NORTHEAST GASOLINEEUM UNIV
Filing Date
2025-04-23
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing bearing fault diagnosis technologies suffer from weak characteristic signals under low-speed, heavy-load conditions, which are easily masked by noise. This makes it difficult to balance diagnostic performance and computational efficiency, resulting in high computational resource consumption.

Method used

A fault diagnosis method for wind turbine main shaft bearings based on a lightweight attention residual network is adopted. The SEDSNet model is constructed by depthwise separable convolution, SE attention mechanism and residual connection, and fault diagnosis is performed by combining Mel spectrum feature extraction.

Benefits of technology

While maintaining high diagnostic accuracy, it significantly reduces computational complexity, improves the model's noise resistance and cross-dataset adaptability, and reduces computational resource requirements.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120429771B_ABST
    Figure CN120429771B_ABST
Patent Text Reader

Abstract

This invention relates to a fault diagnosis method for wind turbine main shaft bearings based on a lightweight attention residual network. It includes: collecting acoustic emission signals from the wind turbine main shaft bearing under different pre-existing defects during operation; extracting Mel-spectrum features from the acoustic emission signals; dividing the generated Mel-spectrum time-frequency image dataset into training, validation, and test sets; constructing a lightweight residual network SEDSNet using depthwise separable convolution, residual connections, and a Squeeze-and-Excitation attention mechanism; training the model using Mel-spectrum features; and verifying the effectiveness of the model for wind turbine main shaft bearings through ablation experiments, comparative experiments, cross-dataset experiments, visualization experiments, and noise experiments. The SEDSNet model is then used for fault diagnosis of the wind turbine main shaft bearing. This invention ensures high diagnostic accuracy while significantly reducing the computational complexity of the model.
Need to check novelty before this filing date? Find Prior Art